Lecture Notes in Mechanical Engineering

S. Vijayan
Nachiappan Subramanian
K. Sankaranarayanasamy   Editors

Trends in
Manufacturing
and Engineering
Management
Select Proceedings of ICMechD 2019
Lecture Notes in Mechanical Engineering

Series Editors
Francisco Cavas-Martínez, Departamento de Estructuras, Universidad Politécnica
de Cartagena, Cartagena, Murcia, Spain
Fakher Chaari, National School of Engineers, University of Sfax, Sfax, Tunisia
Francesco Gherardini, Dipartimento di Ingegneria, Università di Modena e Reggio
Emilia, Modena, Italy
Mohamed Haddar, National School of Engineers of Sfax (ENIS), Sfax, Tunisia
Vitalii Ivanov, Department of Manufacturing Engineering Machine and Tools,
Sumy State University, Sumy, Ukraine
Young W. Kwon, Department of Manufacturing Engineering and Aerospace
Engineering, Graduate School of Engineering and Applied Science, Monterey,
CA, USA
Justyna Trojanowska, Poznan University of Technology, Poznan, Poland
Lecture Notes in Mechanical Engineering (LNME) publishes the latest develop-
ments in Mechanical Engineering—quickly, informally and with high quality.
Original research reported in proceedings and post-proceedings represents the core
of LNME. Volumes published in LNME embrace all aspects, subfields and new
challenges of mechanical engineering. Topics in the series include:
• Engineering Design
• Machinery and Machine Elements
• Mechanical Structures and Stress Analysis
• Automotive Engineering
• Engine Technology
• Aerospace Technology and Astronautics
• Nanotechnology and Microengineering
• Control, Robotics, Mechatronics
• MEMS
• Theoretical and Applied Mechanics
• Dynamical Systems, Control
• Fluid Mechanics
• Engineering Thermodynamics, Heat and Mass Transfer
• Manufacturing
• Precision Engineering, Instrumentation, Measurement
• Materials Engineering
• Tribology and Surface Technology
To submit a proposal or request further information, please contact the Springer
Editor of your location:
China: Dr. Mengchu Huang at mengchu.huang@springer.com
India: Priya Vyas at priya.vyas@springer.com
Rest of Asia, Australia, New Zealand: Swati Meherishi at
swati.meherishi@springer.com
All other countries: Dr. Leontina Di Cecco at Leontina.dicecco@springer.com
To submit a proposal for a monograph, please check our Springer Tracts in
Mechanical Engineering at http://www.springer.com/series/11693 or contact
Leontina.dicecco@springer.com
Indexed by SCOPUS. The books of the series are submitted for indexing to
Web of Science.

More information about this series at http://www.springer.com/series/11236

S. Vijayan Nachiappan Subramanian
• •

K. Sankaranarayanasamy
Editors

Trends in Manufacturing
and Engineering
Management
Select Proceedings of ICMechD 2019

123
Editors
S. Vijayan Nachiappan Subramanian
Sri Sivasubramaniya Nadar College of University of Sussex
Engineering Brighton, UK
Chennai, Tamil Nadu, India

K. Sankaranarayanasamy
National Institute of Technology Puducherry
Puducherry, India

ISSN 2195-4356 ISSN 2195-4364 (electronic)

Lecture Notes in Mechanical Engineering
ISBN 978-981-15-4744-7 ISBN 978-981-15-4745-4 (eBook)
https://doi.org/10.1007/978-981-15-4745-4
© Springer Nature Singapore Pte Ltd. 2021
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Preface

We are delighted to publish the selected papers presented in the Second

International Conference on Mechanical Engineering Design 2019 (ICMechD
2019) held at Sri Sivasubramaniya Nadar College of Engineering, Chennai, India,
during April 25–26, 2019.
In this book, around 100 quality papers in the areas of advanced manufacturing
technology, development and characterization of metal matrix composite, structures
and processing of polymer matrix composite, material processing technology,
material joining technology, production and operation management, optimization
techniques and energy engineering are included for publication.
We are happy to put together this collection of thoughtful papers on the theme of
“Trends in Manufacturing and Engineering Management.” We are sure that this
book will help to nurture knowledge among the research society.
We are grateful to the reviewers for their valuable suggestions for improving the
quality of the papers. Also, we thank the session chairs and organizing committee
members for their steadfast support and suggestions.

Chennai, India S. Vijayan

Brighton, UK Nachiappan Subramanian
Puducherry, India K. Sankaranarayanasamy

v
Contents

Advanced Manufacturing Technology

Effect of Process Parameters in Electric Discharge Machining
of AZ31 Magnesium Alloy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
M. Somasundaram and J. Pradeep Kumar
Desirability Approach Machining Study on Aluminum Composite
Through Wire-Cut Electric Discharge Technique . . . . . . . . . . . . . . . . . 15
K. Rajkumar, C. Balasubramaniyan, K. Ramraji, A. Gnanavelbabu,
and P. Sabarinathan
A Comparative Study on Abrasive Water Jet Machining
Characteristics of Entry and Exit Layers of Glass and Basalt
Woven Polymer Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
K. Ramraji, K. Rajkumar, M. Rajesh, and A. Gnanavelbabu
Electrolyte and Machining Parameters Optimization of Wire
Electrochemical Cutting of Aluminum/Titanium Diboride
Composite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
K. Rajkumar, M. Rajesh, K. Ramraji, P. Sabarinathan,
and A. Gnanavelbabu
Modeling and Parametric Optimization of Process Parameters
of Wire Electric Discharge Machining on EN-31 by Response
Surface Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Sushant B. Patil, Swarup S. Deshmukh, Vijay S. Jadhav,
and Ramakant Shrivastava
Wire-Cut Electric Discharge Machining on Nickel–Aluminium–
Bronze Using Brass Wire Electrode . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Earnest Beni, Poovazhagan Lakshmanan, and S. C. Amith

vii
viii Contents

Development and Characterization of Metal Matrix Composite

Manufacture, Mechanical Properties and Microstructural
Characterization of Aluminium and Iron Metal Matrix Composite
Manufactured . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
D. Arthur Jebastine Sunderraj, D. Ananthapadmanaban,
K. Arun Vasantha Geethan, and A. John Rajan
Influence of Graphite Particles on Microhardness
and Microstructural Behavior of AA7068 Metal Matrix
Composites Processed by Powder Metallurgy . . . . . . . . . . . . . . . . . . . . 91
K. John Joshua, S. J. Vijay, P. Ramkumar, and D. Philip Selvaraj
Microhardness and Microstructural Behavior of AA7068/SiC Metal
Matrix Composites Synthesized by Powder Metallurgy . . . . . . . . . . . . 101
K. John Joshua, P. Ramkumar, S. J. Vijay, and S. Mohanasundaram
Optimization of CO2 Laser Cutting Parameters for AA6061/B4C/
hBN Hybrid Composites using Taguchi-based Response Surface
Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
A. Gnanavelbabu, V. Arunachalam, K. T. Sunu Surendran, K. Rajkumar,
and E. Anandhababu
Analysis of Corrosion Resistance in Domestic Water Geysers
by Coating Nano-Film Using Thermal Spray Coating . . . . . . . . . . . . . 123
Kabilan Sankar, Karthick Selvam, K. Joy Ashwin,
and S. P. Sathya Prasanth
Effect of Multi-walled Carbon Nanotubes Additions on Its Dispersion
Characteristics in Titanium Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Senzeni Sipho Lephuthing, Avwerosuoghene Moses Okoro,
Masego Mohlala, Noxolo Malgas, Oladeji Ige, and Peter Apata Olubambi
Evaluation of Process Parameters Influence on the Mechanical
Properties of RF Magnetron Sputtered TiC Thin-Film Coating . . . . . . 145
Olayinka Oluwatosin Abegunde, Esther Titilayo Akinlabi,
and Oluseyi Philip Oladijo
Effect of Compaction Loads in Machining of Short Carbon
Fiber-Reinforced Aluminum Composite . . . . . . . . . . . . . . . . . . . . . . . . 157
S. Mohanasundaram, S. J. Vijay, Rajakumar S. Rai, I. Kantharaj,
A. S. Melwyn, and S. Theophilus
Synthesis and Characterization of Al 7072-Al2O3 Metal
Matrix Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
G. Mallesh, R. Pavankumar, V. G. Pradeep Kumar, and L. Laxman Naik
Contents ix

Erosion-wear Behaviour of 304 Stainless Steel Reinforced with TiN

at Elevated Temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Ramokone Marcia Mafafo, Babatunde Abiodun Obadele, Bruno Pilotti,
Walter Roberto Tuckart, and Peter Apata Olubambi
A Concise Review of Nano-enhanced Phase Change Materials
for Passive Cooling Applications in Buildings . . . . . . . . . . . . . . . . . . . . 193
Chukwumaobi K. Oluah, Esther Titilayo Akinlabi, and Howard O. Njoku
Effect of Silicon Carbide in Yttria-stabilized Zirconia for Thermal
Protective Structural Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
S. Subha, Poojari Surendra, Ch. Gowtham Chowdary,
V. Naga Venkata Sai Ram, and Palem Srikanth

Structures and Processing of Polymer Matrix Composite

Optimization of Process Parameter in Abrasive Water Jet Machining
of Blue-Fired Grain-Reinforced Glass Fiber Polymer Composite . . . . . 217
P. Sabarinathan, V. E. Annamalai, and K. Rajkumar
Mechanical Behavior of Raavi and Pineapple Fiber-reinforced
Hybrid Polyester Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
J. Varun Siddharth, I. Daniel Lawrence, and S. Jayabal
Carbon Fiber Surface Treatment for Enhanced Interfacial
Properties: A Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
Kwame Anane-Fenin, Esther Titilayo Akinlabi, and Nicolas Perry
Preparation and Mechanical Property Analysis of Polymer Matrix
Composite Containing Rice Husk and Saw Dust . . . . . . . . . . . . . . . . . 253
Sarojrani Pattnaik, Arnab Sengupta, and Mihir Kumar Sutar
Investigation of the Mechanical Properties of Polyamide 6 Hybrid
Nanocomposites with MWCNT and Copper Nanoparticles . . . . . . . . . 261
T. Anand and T. Senthilvelan
Preparation, Characterization, Image Segmentation and Particle Size
Analysis of Cow Bone Powder for Composite Applications . . . . . . . . . 273
O. M. Ikumapayi, Esther Titilayo Akinlabi, Paul A. Adedeji,
and S. A. Akinlabi
Investigation of Mechanical and Chemical Properties of the Coir
Fiber and Wood Powder Reinforced Hybrid Polymer Composite . . . . 285
T. Prabhuram, D. Elilraja, S. Prathap Singh, and Immanuel Durairaj
An Experimental Study on Hemp/Sisal Fiber Embedded Hybrid
Polymer Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
Akash, Shivakumar Rachoti, Vishwanath Patil, and K. G. Girisha
x Contents

Investigation on Chemical Isolation and Characterization

of Cellulose from Delonix regia Fruit Fibers . . . . . . . . . . . . . . . . . . . . . 303
Kalpit P. Kaurase and Dalbir Singh
Experimental Study and Analysis of Defragmented Carbon
Nanotubes in Polyacrylonitrile Matrix . . . . . . . . . . . . . . . . . . . . . . . . . 315
N. Arunkumar, Joven Job, D. Ananthapadmanaban, N. E. Arun Kumar,
and N. Sathishkumar
Design and Fabrication of Car Door Panel Using Natural
Fiber-Reinforced Polymer Composites . . . . . . . . . . . . . . . . . . . . . . . . . 331
Sodisetty V. N. B. Prasad, G. Akhil Kumar, P. Yaswanth Sai,
and Shaik Varees Basha
Investigation of Physical–Chemical Properties and Evaluation
of Optimal Blend Ratio of Rice Bran Biodiesel: A Mathematical
Regression Analysis Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
M. Selvam, S. Palani, and P. Vaishnavi
Wear and Friction Behaviours of Stainless Steel (SS 316) Wire Mesh
and Carbon Fibre Reinforced Polymer Composite . . . . . . . . . . . . . . . . 361
A. H. Ansari, V. Jayakumar, and S. Madhu
Tribological Properties of PEEK Reinforced with Synthetic
Diamond Composite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
K. Rajkumar, K. Vishal, and P. Sabarinathan
Experimental Investigation on the Mechanical Properties
of American Agave and Glass Fibre Reinforced Polypropylene
Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
M. Indra Reddy, M. Anil Kumar, and Vamsi Inturi
Influence of Magnesium Hydroxide Fillers on Acoustic, Thermal,
and Flame Retardant Properties of Pu Foam . . . . . . . . . . . . . . . . . . . . 393
L. Yuvaraj, S. Jeyanthi, Digvijay D. Kadam, and R. G. Ajai
Development and Analysis of GFRP Conical Springs . . . . . . . . . . . . . . 409
D. Vivek, R. Praveen, A. Sanjay Krishnan, Y. K. Sabapathy, D. Ebenezer,
and M. Selvaraj
Effect of Nano and Microfillers in Basalt/Epoxy Composites . . . . . . . . 419
M. M. Metro and M. Selvaraj
A Feasibility Study on Microwave Joining of GFRP Composite Pipes
with Interlayer Coupling Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
K. Rajkumar, M. Dhananchezian, and S. Aravindan
Comparative Evaluation of Mechanical Properties of GFRP
and Polymer Hybrid Composite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441
R. Raja, Sabitha Jannet, Allen Varughese, and Joby George
Contents xi

Study of Hexagonal Boron Nitride Particulate as Vibration

Behaviour Modifier of Alternate Stacked Glass–Natural Fibre
Polymer Composite Laminate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451
K. Rajkumar and M. Selvaraj
Study of Statistical Distribution and Morphology of Particles
in a Polymer Matrix by Foldscope Imaging Technique . . . . . . . . . . . . 461
P. Kaythry, A. Madhan, and K. Rajkumar
Dynamic Mechanical Analysis of Flax Fiber Stacked Polyurethane
Blend Epoxy Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475
K. Rajkumar, K. Ramraji, M. Rajesh, and M. Rajiv kumar

Material Processing Technology

Comparative Study of Ball Nose and Flat End Milling on A356
Alloy/SiCp Metal Matrix Composite . . . . . . . . . . . . . . . . . . . . . . . . . . . 487
K. Jayakumar
Numerical Modeling of Orthogonal Machining Process Using
Smoothed Particle Hydrodynamics—A Parametric Study . . . . . . . . . . 497
Sarath Babu Thekkoot Surendran, C. S. Sumesh, and Ajith Ramesh
Acoustic Emission-Based Grinding Wheel Sharpness Monitoring
Using Machine Learning Classifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511
J. Revant, Rahul Sree Kumar, K. Rameshkumar, and D. S. B. Mouli
Turning Process Characteristics of Aluminium Matrix Hybrid
Composite Using Grey Relational Surface Methodology . . . . . . . . . . . . 523
A. Gnanavelbabu, V. Arunachalam, K. T. Sunu Surendran, R. Saranraj,
and K. Rajkumar
Corrosion Behaviour of Hot Rolled AA8015 in Natural Seawater
at 1.37 µm Surface Roughness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535
O. Olaogun and Esther Titilayo Akinlabi
Influence of Tool Profiles on Heat Transfer Analysis in Al-6061 Alloy
Using Friction Stir Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 545
M. Bala Chennaiah, A. Sreenivasulu, and K. Ravi Kumar
Tensile, Hardness, and Impact Properties of Amorphous Al–Si–Mg
Cast Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555
Olayide R. Adetunji, Adeniran S. Afolalu, Mufutau A. Mustapha,
Oluwasegun J. Adelakun, Samson O. Ongbali, and Abiodun A. Abioye
Experimental Investigation on the Influence of Tool Geometry
in Minimum Chip Thickness of Microendmilling Using Cutting
Force Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 567
M. Prakash, M. Kanthababu, A. Arul Jeya Kumar,
and V. Prasanna Venkadesan
xii Contents

Temperature and RF Power Effect on the Morphology

and Structural Properties of TiC Thin Film Grown
by RF Magnetron Sputtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 583
Olayinka Oluwatosin Abegunde, Esther Titilayo Akinlabi,
and Oluseyi Philip Oladijo
Effect of Temperature on the Surface Characteristics of Anodized
Aluminium Tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 591
A. L. Sriram Sudhan and A. Brusly Solomon
Experimental Investigation on Heat Pipe-Assisted Cooling During
Milling Process of AISI 1040 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 601
I. Kantharaj, D. J. Hiran Gabriel, Julius Benedict Prakash,
and S. Mohanasundaram
Development of a Modified Magnetic Moulding Set up for Improved
Heat Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 611
B. Anand Ronald, M. Harshal, K. Barath Varadaraj, and G. Gopinath
Quality Assessment Studies on AA7075 Plate in Hot
Rolling Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 619
K. Jayakumar
Artificial Neural Network and Genetic Algorithm-Based Models
for Predicting Cutting Force in Turning of Hardened H13 Steel . . . . . 627
K. Leo Dev Wins, B. Anuja Beatrice, D. S. Ebenezer Jacob Dhas,
and V. S. Anita Sofia
Experimental Study of the Effect of TiN–Zn Coated High-Speed Steel
Cutting Tool on Surface Morphology of AL1060 Alloy During
Machining Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 637
I. P. Okokpujie, C. A. Bolu, O. S. Ohunakin, and Esther Titilayo Akinlabi
Experimental Investigation of Surface Roughness in End Milling
of AA6061 Alloy with Flooded Cooling and Minimum Quantity
Lubrication (MQL) Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 649
D. Nathan, D. Elilraja, T. Prabhuram, and S. Prathap Singh

Material Joining Technology

Performance Study of Dissimilar Alloy Joints of SS321 and SS347
Under MIG Welding Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663
S. Nishanth, S. Prem, T. Prakash, J. Siva, M. Sathish Kumar,
and N. Siva Shanmugam
Effect of Friction Stir Welding Tool on Al–SiC Composites
by Varying Tool Pin Profile and Tool Material . . . . . . . . . . . . . . . . . . 673
P. Jayaseelan, T. V. Christy, and S. J. Vijay
Contents xiii

Fabrication of AISI 304 Austenitic Stainless Steels with TiN

Addition Using Spark Plasma Sintering Method . . . . . . . . . . . . . . . . . 681
Babatunde Abiodun Obadele, Ramokone Marcia Mafafo,
Walter Roberto Tuckart, and Peter Apata Olubambi
Effect of Process Parameters on Bead Width of 202 Grade Stainless
Steel Gas Tungsten Arc Welded Plates Using Response Surface
Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 691
R. Sudhakaran, P. S. Sivasakthivel, M. Subramanian, S. Mahendran,
M. Sathish Kumar, and R. Vijayakumar
Feasibility Study of TIG Welding of AA6063-AA7075 Alloys . . . . . . . . 709
D. Nathan, S. Ashwin Kannan, and P. Krishna Kumar
Effects of Processing Parameters on Temperature Distributions,
Tensile Behaviour and Microstructure of Friction Stir Welding
of Dissimilar Aluminium Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 721
Olatunji P. Abolusoro and Esther Titilayo Akinlabi
Optimizing the Parameters for Friction Stir Welding
of an Aluminium Alloy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 735
M. Selvaraj
Machining of ZE41 Magnesium Alloy in WEDM
Using Taguchi Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 747
G. Selvakumar, S. Ram Prakash, and E. Caleb Kovilpillai
Comparative Study of Friction Stir Welding and Underwater
Friction Stir Welding on Magnesium ZE41 Alloy . . . . . . . . . . . . . . . . . 755
S. Cyril Joseph Daniel and A. K. Lakshminarayanan
Characterizing the Tensile Deformation Behavior of Friction Stir
Welded Dissimilar Joints Using Acoustic Emission Technique . . . . . . . 767
A. Venkatakrishna, A. K. Lakshminarayanan, K. Radhika,
and R. Rajasekaran
Study of Infrared Thermography on Tensile Behavior of Laser Beam
Welded 316LN Austenitic Stainless Steel . . . . . . . . . . . . . . . . . . . . . . . 779
R. Rajasekaran, A. K. Lakshminarayanan, A. Venkatakrishna,
and K. Radhika

Production and Operation Management

Evaluation of Ergonomics Issues in Repetitive Scrap Handling Work
in Automobile Industries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 791
Aravind Babu Dasari and Dhinesh Balasubramanian
Reduction of Terminal Rejections and Failure Cost through Kaizen
at Shop Floor of a Wire Harness Manufacturing Company . . . . . . . . . 817
Pramod Kumar and Jaiprakash Bhamu
xiv Contents

Solving the Flexible Job Shop Scheduling Problem Using a Hybrid

Artificial Bee Colony Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 833
Rylan H Caldeira, A. Gnanavelbabu, and J. Joseph Solomon
Linear Programming in Market Management Using Artificial
Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 845
Shetty Vignesh Uday, Hamritha, and Gaurav Chaudhary
Queueing Theory an Index Factor for Production Inventory Control
in Automotive Industry—A Review . . . . . . . . . . . . . . . . . . . . . . . . . . . 853
Sunday A. Afolalu, Segun Oladipupo, Samson O. Ongbali,
Abiodun A. Abioye, Ademola Abdulkareem, Mfon O. Udo,
and Oluseyi O. Ajayi
Design and Analysis of ASRS Using AGV for Rapid Inventory
Storage System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 861
Pranav Santhosh Nair, Sourabh Nair, Soukhin Sarkar,
Arockia Selvakumar Arockia Doss, and D. Dinakaran

Optimization Techniques
Multi-response Optimization of Inconel 825 Process Parameters
Using LN2 Cooled Zinc-Coated Brass Wire in CNC
Wire-Cut EDM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875
Midthur A. Salman Khan, C. Nandakumar, B. Mohan,
and R. Senthil Kumar
Optimization of Process Parameters During EDM on Inconel
Alloy 625 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 897
K. Jayakumar
Optimization of Laser Trepanning Parameters for Mild Steel
by Taguchi Response Surface Methodology
(T-RSM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905
A. Gnanavelbabu, V. Arunachalam, K. T. Sunu Surendran, V. Dharaniya,
and K. Rajkumar
Parametric Optimization of Cracked Cantilever Beam
Using Genetic Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 915
Mihir Kumar Sutar, Sarojrani Pattnaik, and Pawan Kumar Modi
Study of the Influence of Reinforcement Parameters on Thermal
Conductivity of Magnesium-Based MMCs Through Taguchi’s
Orthogonal Array Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 923
M. R. Shivakumar, N. V. R. Naidu, M. Jai Surya, and D. Indhuja
Finite Element Modelling and Optimisation of Sheet Hydroforming
for Cryo-rolled AA5083 Sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 931
Akhil B. Raj, A. Arun, and Ajith Ramesh
Contents xv

Selection of Parameterization Method for Fitting of Freeform Curves

Using Uniformly Spaced Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 943
G. Rajamohan

Energy Engineering
Experimental Study of an Axial Turbine for Wave Energy
Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 953
Kumud Kumar, Tapas K. Das, R. Srikanth, and Abdus Samad
Investigations into Nonlinear Energy Sinks for a Stochastic
Dynamical Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 963
Anuroop Parvathaneni, Dhritimaan Sharma, Dhruv Vashishtha,
Pradeep V. Malaji, and J. Venkatramani
A Novel Banana Leaf Waste-Based Activated Carbon
for Automobile Emission Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 977
A. John Presin Kumar, S. Sivakumar, R. Balaji, and Mukesh Nadarajan
Analyzing Different Methods to Increase the Natural Period
of a Compact Wave Energy Converter . . . . . . . . . . . . . . . . . . . . . . . . . 991
Vishnu Vijayasankar and Abdus Samad
Comparison of Hydrogen Yield from Ball-Milled and Unmilled
Magnesium Hydride in a Batch System Hydrogen Reactor . . . . . . . . . 1003
J. A. Adeniran, R. S. Fono-Tamo, Esther Titilayo Akinlabi, and T. C. Jen
A Review on the Synthesis of Activated Carbon from Natural
Resources for Mechanical Applications . . . . . . . . . . . . . . . . . . . . . . . . 1013
A. John Presin Kumar, S. Sivakumar, D. Prasanth, B. Guhanesh,
and A. Ijas Ahamed
Combined Casing Groove and Blade Tip Treatment for Wave
Energy Harvesting Turbine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1027
P. Madhan Kumar, Paresh Halder, and Abdus Samad
Design and Development of Wind Tunnel to Study
Smoldering Combustion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1041
S. Sanjana, M. Rakshantha, Yeleti Bunny Venkat, and B. T. Kannan
Spatial Location of Renewable Energy Plants: How Good
Is Good Enough? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1055
Paul A. Adedeji, Stephen Akinlabi, Nkosinathi Madushele,
and Obafemi O. Olatunji
Analysis of Pendulum-Based Nonlinear Energy Sink
for Energy Harvesting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1065
Pradeep V. Malaji
xvi Contents

Effect of Input Velocity on the Output of Vertical Axis Wind

Turbine (VAWT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1075
Sankgeeth Vennila Sigamani, Shami Jose Jose Robin,
Arun Prakash Chandran, and B. Anand Ronald
Radiative Heat Transfer of Magnetic Nanofluid Flow Past a Porous
Inclined Plate: A Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . . . 1085
M. Shanmugapriya
A GA-ANFIS Model for the Prediction of Biomass Elemental
Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1099
Obafemi O. Olatunji, Stephen Akinlabi, Nkosinathi Madushele,
and Paul A. Adedeji
About the Editors

Dr. S. Vijayan, is currently working as Associate Professor in Department of

Mechanical Engineering at Sri Sivasubramaniya Nadar College of Engineering has
20 years of teaching experience. He has received his Bachelor of Engineering (B.E.)
in Mechanical Engineering from the Madurai Kamarajar University in 1997,
completed his post graduate Masters of Engineering (M.E.) in Industrial
Engineering from Thigarajar College of Engineering in 1998 and Doctorate Ph.D.
from the Anna University in the year 2010. To his credit he has completed two
research sponsored projects to the tune of 30 lakh funded by Naval research Board,
DRDO and SSN Trust. He has published 25 International journals papers to his
credit. He has published two books for VDM-Verlag and Lambert, respectively. He
served in administrative responsibilities such as Chairman Valuation in Anna
University and Additional Controller of Examination in SSN Institutions.

Dr. Nachiappan Subramanian is currently working as Professor in Operations

and Logistics Management & Supply Chains, University of Business School,
University of Sussex, United Kingdom. Formerly he served as Associate Professor
in Nottingham University Business School China. In total, he has 20 years of
experience including 18 years in academia and 2 years in a consulting company. He
has received his Bachelor of Engineering (B.E.) in Mechanical Engineering from
the Madurai Kamarajar University in 1995, completed his post graduate Masters of
Engineering (M.E.) in Production Engineering from Thigarajar College of
Engineering in 1998 and Doctorate Ph.D. from the Madurai Kamarajar University
in the year 2006. He has completed his postdoctoral research from Nottingham
University and Royal Melbourne Institute of Technology University. To his credit,
he has completed several sponsored projects to the tune of £ 438,370. To date, he
has published 89 peer-reviewed refereed research papers. His research interest areas
are in sustainable supply chains (environmental and social/humanitarian issues),
risk and resilience, technology enabled operations and marketing interface, and in
performance measurement.

xvii
xviii About the Editors

Dr. K. Sankaranarayanasamy is serving currently as Director of NIT

Pondicherry, formerly a Professor of Mechanical Engineering at National Institute
of Technology Tiruchirappalli, received his B.E. Mechanical Engineering (Hons) in
1981 from PSG College of Technology, India. He received his M.Tech Production
Engineering in 1983 and Ph.D. in 1989 from Indian Institute of Technology
Madras, India. His research targets finite element simulation of welding processes,
Ergonomics Study and Gear Design. He has authored more than 75 papers in
international and national journals and has attended conferences in Slovak
Republic, China, America and Austria. To his credit, he has completed several
sponsored projects (total worth of Rs. 60 lakhs) funded by BHEL, Trichy, NLC,
Neyveli and DST. He served in administrative responsibilities such as Head of the
Department, Dean and Director of NIT.
Advanced Manufacturing Technology
Effect of Process Parameters in Electric
Discharge Machining of AZ31
Magnesium Alloy

M. Somasundaram and J. Pradeep Kumar

Abstract In modern-day scenario, AZ31 magnesium alloy is considered as a suit-

able alternative to iron and aluminum in a wide variety of medical, aerospace and
automobile applications due to high strength-to-weight ratio and other excellent
physical properties. In all the applications, the geometrical accuracy of the products
plays a vital role. Electrical discharge machining (EDM) has emerged as a promi-
nent manufacturing process to achieve geometrical accuracy as per the requirements
and standards. A hole of diameter 10 mm in a flat plate made of AZ31 alloy is
identified as the part feature for investigation in this study. The aim of this work
is to optimize and investigate the effect of both qualitative and quantitative process
parameters such as discharge current (I), pulse-on time (T ON ), pulse-off time (T OFF ),
electrode material (M) on critical-to-quality geometric features such as overcut (OC),
tapercut (TC), circularity (CIR) and cylindricity (CYL). The experiments are carried
out based on Taguchi’s L16 orthogonal array. Pulse-on time and discharge current
are found to have a significant effect on the geometrical accuracy when compared
with other process parameters. Regression models are developed for prediction of
response parameters, and the results predicted by the models are found to correlate
significantly with the results from experiments.

Keywords Electrical discharge machining (EDM) · AZ31 magnesium alloy ·

Discharge current (I) · Pulse-on time (T ON ) · Pulse-off time (T OFF ) · Overcut
(OC) · Tapercut (TC) · Circularity (CIR) · Cylindricity (CYL)

M. Somasundaram (B) · J. Pradeep Kumar

Department of Production Engineering, PSG College of Technology, Coimbatore, India
e-mail: somaattitude@gmail.com
J. Pradeep Kumar
e-mail: jpk.prod@psgtech.ac.in

© Springer Nature Singapore Pte Ltd. 2021 3

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_1
4 M. Somasundaram and J. Pradeep Kumar

1 Introduction

The magnesium and its alloys have been established as the best alternative to iron
and aluminum in a wide variety of applications due to high strength-to-weight ratio
and other excellent physical properties [1]. However, machining magnesium alloys
using conventional methods cause a built-up edge and chatter. The most important
precaution needed to bear in mind, while machining magnesium alloy is that the
formation of fine chips and dust is highly flammable [2]. Due to the development of
new difficulties to machine materials (carbide, composite materials, etc.) and com-
plex geometrical shapes of engineering components, the available traditional fin-
ishing processes alone are incapable of producing required surface finish and other
geometrical features. Even if these processes can be used, they require expensive
equipments and large labor, hence making them economically incompetent. Devel-
opments in advanced machining processes in the last few decades have attributed
to the relaxation of limitations of tool hardness requirement or/and restricted pre-
defined relative motion of cutting edges with respect to the workpiece surface. Use
of alternative source of energy like electrical, chemical, mechanical or thermal to
assist or perform machining relaxed the constraint of using cutting tool harder than
the workpiece. Electric discharge machining (EDM), electrochemical machining
(ECM), ultrasonic machining (USM), abrasive jet machining (AJM), laser beam
machining (LBM), etc. are a few instances of such techniques. Being the most ver-
satile of all the non-conventional machining, electric discharge machining (EDM)
process becomes viable for the machining of these magnesium alloys materials [3–
5]. In all applications, geometrical features of machined parts are more important.
This work is carried out to study the effect of process parameters while machining
a hole feature of diameter 10 mm. Existing research works carried out using AZ31
magnesium alloy focus on the measurement of surface roughness [6]. But the present
work focuses on geometrical features of the holes which is essential to make a proper
hole on AZ31 magnesium alloy while machining using EDM.

2 Details of Experiments

In this work, magnesium AZ31 alloy is chosen as a workpiece material. The work-
piece is in the form of a flat plate with length (50 mm), width (30 mm) and thickness
(6 mm). Tool materials such as copper, brass and graphite are chosen as an electrode.
The electrode in the form of a rod with a diameter of 10 mm and a length of 25 mm
is machined and used to carry out experiments. The workpiece material (anode)
used is AZ31 magnesium alloy. The tool (cathode) materials such as copper, brass
and graphite are used. The experimental trials are conducted using ALTRA ZNC
ORB 5530 EDM machine, based on Taguchi’s L16 orthogonal array. Figure 1 shows
the machined workpiece and their corresponding electrode. The process parameters
such as discharge current (I P ), pulse-on time (T ON ), pulse-off time (T OFF ) and tool
Effect of Process Parameters in Electric Discharge … 5

Fig. 1 Machined workpiece

Table 1 Control factors and their levels

Factors Level 1 (−1) Level 2 (0) Level 3 (1) Level 4 (2)
I P (Amps) 3 6 9 12
T ON (µs) 10 20 30 40
T OFF (µs) 5 6 7 8
M Copper (Cu) Brass (Br) Graphite (Gr) Copper (Cu)

material (M) are selected as control factors, and overcut (OC), tapercut (TC), cir-
cularity (CIR) and cylindricity (CYL) are considered as response variables. Table 1
shows the control factors and their levels. Overcut is equal to half the difference of
the diameter of the hole to the diameter of the tool. The diameter of the hole and
tool is measured using a digital vernier caliper. Tapercut is the difference between
hole diameter of top and bottom surface of the plate. It was measured with the help
of a profile projector. Circularity describes how close an object should be to a true
circle. Cylindricity is used when the hole part features must have good circularity
and straightness. Circularity applies only to cross sections, where the cylindricity
applies simultaneously to the entire surface. Both circularity and cylindricity were
measured using Mitutoyo Coordinate Measuring Machine (CMM).

3 Result and Discussion

Two trials are carried out for each combination of process parameters, and the average
of two values is presented in Table 2.

3.1 Effect of Process Parameters on Response Parameters

Analysis of variance (ANOVA) is used to study the effect of process parameters on

response variables. ANOVA is carried out at a confidence level of 95%. Since it is
always desirable to achieve the lowest overcut, tapercut, circularity and cylindricity,
6 M. Somasundaram and J. Pradeep Kumar

Table 2 Design matrix and their output response parameters

S. No. IP T ON T OFF M OC TC CIR CYL
(Amps) (µs) (µs) (mm) (mm) (mm) (mm)
1 3 10 5 Cu 0.025 0.057 0.0178 0.0428
2 3 20 6 Br 0.055 0.069 0.0105 0.0783
3 3 30 7 Gr 0.14 0.079 0.0092 0.1269
4 3 40 8 Cu 0.085 0.047 0.0068 0.1074
5 6 10 6 Gr 0.07 0.08 0.0416 0.0683
6 6 20 5 Cu 0.035 0.059 0.0063 0.0316
7 6 30 8 Cu 0.065 0.033 0.234 0.0865
8 6 40 7 Br 0.095 0.039 0.0051 0.0801
9 9 10 7 Cu 0.03 0.02 0.0311 0.0635
10 9 20 8 Gr 0.165 0.023 0.0058 0.0501
11 9 30 5 Br 0.15 0.028 0.0451 0.1029
12 9 40 6 Cu 0.205 0.015 0.0068 0.0849
13 12 10 8 Br 0.005 0.039 0.008 0.0766
14 12 20 7 Cu 0.13 0.06 0.025 0.0515
15 12 30 6 Cu 0.165 0.03 0.0321 0.0791
16 12 40 5 Gr 0.27 0.05 0.0647 0.1375

smaller the better characteristic is selected. It is found that pulse-on time shows a
maximum contribution of 48% in the assessment of overcut as shown in Fig. 2. The
increase in pulse-on time resulted in increasing of overcut as shown in Fig. 3, and
this is mainly due to increases in spark energy as pulse-on time increases.
From Fig. 4, it is observed that the discharge current shows a maximum contri-
bution of 28% on tapercut. As discharge current increases, tapercut decreases up to
a certain saturated point as shown in Fig. 5. This phenomenon occurs mainly due to
an increase in spark energy as discharge current increases.
From Fig. 6, it is observed that interaction between discharge current and pulse-
off time shows a maximum contribution of 31% on circularity. Circularity increases

Fig. 2 Contribution of
process parameters on OC
Effect of Process Parameters in Electric Discharge … 7

Fig. 3 Main effects plot for SN ratio OC

Fig. 4 Contribution of process parameters on TC

as an increase in discharge current. Increasing in pulse-off time leads to a decrease

in circularity, and it is mainly due to proper flushing time as an increase in pulse-off
time as shown in Fig. 7.
From Fig. 8, it is observed that pulse-on time shows a maximum contribution of
45% on cylindricity. Cylindricity decreases initially with an increment in pulse-on
time but increased again at a higher value of pulse-on time as shown in Fig. 9. This
is because of the wear of electrode and deposition of the black carbon layer on the
tool surface.
Copper electrode helps to get minimum tapercut and cylindricity as shown in
Figs. 5 and 9. Since it has less tool wear rate and good thermal conductivity, minimum
circularity and overcut are achieved with the help of the brass electrode as shown in
8 M. Somasundaram and J. Pradeep Kumar

Fig. 5 Main effects plot for SN ratio TC

Fig. 6 Contribution of process parameters on CIR

Figs. 3 and 7. Since brass has less thermal conductivity and greater tool wear rate
which made brass suitable for features that can be machined at a particular plane. In
case of geometrical features when compared with copper and brass, graphite found
to be very less significant.
Regression model for prediction of overcut (OC) is developed using the results
from experiments as shown in Eq. (1). Similar models are developed for prediction
of other response variables. The experimental response variables show significant
correlation with predicted response variables as shown in Figs. 10, 11, 12 and 13.
Effect of Process Parameters in Electric Discharge … 9

Fig. 7 Main effects plot for SN ratio CIR

Fig. 8 Contribution of process parameters on CYL

OC (mm) = 0.314 + 0.0659 ∗ IP + 0.0105 ∗ TON + 0.1972 ∗ TOFF

− 0.1273M − 0.1638 ∗ IP ∗ IP − 0.0075 ∗ TON ∗ TON
− 0.0894 ∗ TOFF ∗ TOFF + 0.0656 ∗ M ∗ M + 0.0573 ∗ IP ∗ TON
− 0.0019 ∗ IP ∗ TOFF − 0.0223 ∗ IP ∗ M
− 0.1575TON ∗ TOFF + 0.1675 ∗ TON ∗ M (1)
10 M. Somasundaram and J. Pradeep Kumar

Fig. 9 Main effects plot for SN ratio CYL

Fig. 10 Experimental OC versus predicted OC

4 Conclusion

An experimental setup for significant manufacturing processes, electric discharge

machining, is used, in order to study the effect of discharge current, pulse-on time,
Effect of Process Parameters in Electric Discharge … 11

Fig. 11 Experimental TC versus predicted TC

Fig. 12 Experimental CIR versus predicted CIR

pulse-off time and tool material on the overcut, tapercut, circularity, and cylindric-
ity of a through-hole made in AZ31 magnesium alloy. Based on the results from
experiments, the following conclusions are made.
• Among the mentioned quantitative EDM process parameters, pulse-on time and
discharge current show the significant effect on geometrical features. It was
observed that pulse-off time shows less significant in case of geometrical features.
12 M. Somasundaram and J. Pradeep Kumar

Fig. 13 Experimental CYL versus predicted CYL

• In terms of a qualitative EDM process parameter, tool material also shows less sig-
nificance on the accuracy of geometrical features when compared with other pro-
cess parameters considered in this work. Among the electrode materials, graphite
is found to have a less significant effect when compared with copper and brass.
• As the objective of the present study is to minimize OC, TC, CIR and CYL. The
S/N plots are used to decide the optimal set of process to obtain best response.
The optimal machining parameters for OC are I p = 6 amps; T ON = 10 µs; T OFF
= 8 µs; M = Br, for TC are I P = 9 amps; T ON = 40 µs; T OFF = 8 µs; M = Cu,
for CIR are I P = 3 amps; T ON = 20 µs; T OFF = 7 µs; M = Br, for CYL are I =
6 amps; T ON = 20 µs; T OFF = 5 µs; M = Cu.
• The developed regression models are found to be accurate in the prediction
of response variables within the domain of experiments. The predicted results
correlate well with experimental results.

References

1. Lee ES, Won JK, Shin TH, Kim SH (2012) Investigation of machining characteristics for elec-
trochemical micro-deburring of the AZ31 lightweight magnesium alloy. Int J Precis Eng Manuf
13(3):339–345
2. Abdul-Rani AM, Razak MA, Littlefair G, Gibson I, Nanimina AM (2017) Improving EDM
process on AZ31 magnesium alloy towards sustainable biodegradable implant manufacturing.
Procedia Manuf 7:504–509
3. Kumar P, Parkash R (2016) Experimental investigation and optimization of EDM process
parameters for machining of aluminum boron carbide (Al–B4C) composite. Mach Sci Technol
20(2):330–348
Effect of Process Parameters in Electric Discharge … 13

4. Kathiravan S, Pradeep Kumar J (2017) Experimental investigation on influence of process

parameter in EDM of P20 mold steel. J Adv Nat Appl Sci 8:161–167
5. Balaji, R.L., Kumar, J.P., (2016) Experimental study on electrical discharge machining (EDM)
of monel metal. J Eng Sci Res Appl 10–13
6. Razak MA, Abdul-Rani AM, Rao TVVLN, Pedapati SR, Kamal S (2016) Electrical discharge
machining on biodegradable AZ31 magnesium alloy using Taguchi method. Procedia Eng
148:916–922
Desirability Approach Machining Study
on Aluminum Composite Through
Wire-Cut Electric Discharge Technique

K. Rajkumar , C. Balasubramaniyan, K. Ramraji , A. Gnanavelbabu ,

and P. Sabarinathan

Abstract Aluminum matrix composites (AMCs) are exceptionally designed for

automobile components like brake pads, cylinder body, and pistons owing to high
temperature and wear resistance. Machining is a paramount issue to achieve required
dimensions and tolerance for those products. The present investigation includes fab-
rication of composite and wire-cut electric discharge machining with various param-
eters. AMC designed by B4 C and CNT particles of hybrid composite was fabricated
by liquid metallurgy process. Carbon nanotubes were added to tailor the high tem-
perature withstanding and reduce friction during the sliding. The role of B4 C is to
improve the hardness and wear resistance of the composite. The combined actions of
these reinforcements enhance overall mechanical properties of the composite. The
objective of the present work is to study the machinability aspects of the composite
by wire-cut EDM. Input machining parameters considered are pulse on time, pulse
off time, wire feed, gap in voltage, and servo speed. Kerf width and material removal
rate were studied with response surface method (RSM). Pulse on time and voltage
increase the material removal rate; however, voltage has a more relatively significant
effect on the MRR.

Keywords Electric discharge · Metal composite · Spark erosion · CNT

K. Rajkumar (B) · K. Ramraji · P. Sabarinathan

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, Tamilnadu, India
e-mail: rajkumark@ssn.edu.in
C. Balasubramaniyan
Department of Mechanical Engineering, Sri Venkateswaraa College of Technology,
Sriperumbudur, Tamilnadu, India
A. Gnanavelbabu
Department of Industrial Engineering, College of Engineering, Chennai, Tamilnadu, India

© Springer Nature Singapore Pte Ltd. 2021 15

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_3
16 K. Rajkumar et al.

1 Introduction

Metal matrix composites (MMCs) are now replacing metals in various engineering
applications. Specifically, aluminum matrix composites (AMCs) occupy a position of
great significance in industries, because of their good thermo-mechanical properties,
good damage resistance, and wear resistance. Generally, machining of MMCs is a
laborious work owing to the presence of hard and brittle reinforcement particle and
non-homogenous structure [1]. It was also realized that their machining behavior is
unlike any of the other conventional processes observed so far. It can be machined
through some of the conventional methods. But, severe damage to the workpiece and
the cutting tool are observed. Workpiece damages such as particle cracking, particle
pulling out and particle debonding were occurred [2, 3]. Surface roughness was also
observed to be high, and in order to overcome that, self-lubricating particles are added
to the composites. This reduces wear between the tool and workpiece. Otherwise, it
decreases strength, hardness, and low fracture toughness which are undesirable [4].
To counteract unsatisfactory properties, some novel approach has to be followed,
i.e., adding concurrent lubricants and strength improvement particles. Tribological
investigations are evident that the relationship between work material grain size and
reinforcement particle size has a major impact on the wear behavior [5].
In recent times, carbon nanotubes are considered as reinforcement in aluminum
composites. There are major reasons for selection of CNT as reinforcement that
increases stiffness, strength and vibration absorption properties of the composite.
This also makes the machining process as smoother than ever and facilitates easy
chip breaking [6, 7].
Nowadays, machining of MMCs is viable through unconventional processes [8].
The major advantages of these processes are having a good surface finish and easy
disposal of removed material. Also, regardless of the material properties, it can
machine the component with less tool wear. Shubhra paliwal et al. [9] studied the
EDM process by analyzing pulse on and off time on the machining behavior of the
material. Hewidy et al. [10] conducted EDM experiments using a RSM for the Inconel
601 material. They concluded that RSM method has an advantage of describing the
effect between input variables and response parameters.
Tilekar et al. [11] determined that high electrical conductivity and low melting
point of aluminum are directly affecting machining. As an increase in current, more
melting and vaporization of material are observed, thereby increasing MRR. Bobbili
et al. [12] experimented EDM with hot-pressed boron carbide and observed deep
craters and microcracks at high Ton and peak current which deteriorated surface fin-
ish. Moreover, higher input energy significantly affected the wear rate of the wire and
caused breakage of the wire frequently. Udaya Prakash et al. [13] conducted machin-
ability studies on the Al alloy with fly ash and B4 C composite using a wire-cut EDM.
They determined that gap voltage is the more significant parameter than pulse on
time, pulse off time, wire feed rate, and percentage of reinforcement. However, fabri-
cated Al/boron carbide/CNT hybrid composite is heterogeneous in nature. Therefore,
Desirability Approach Machining Study on Aluminum … 17

it required a non-traditional wire-cut EDM machining approach to maintain the sur-

face integrity. The aim of this study was to investigate the material removal rate and
kerf dimension of the hybrid composite by varying WEDM processes parameters.

2 Experimental Procedure

2.1 Material Selection

Aluminum (Al 6061) was used as matrix element in the composite. B4 C (40 µm) was
used as one of the reinforcing agents and has exotic properties like high hardness,
low specific gravity, and neutron absorption. CNT (74 nm) was used as another
reinforcing fiber and also solid lubricant.

2.2 Composite Fabrication: Al-B4 C-CNT

Fabrication of composite was done using a stir casting apparatus. Aluminum 6061
ingots were melted in the furnace around 800 °C. Flux was used to increase the
wettability between melt matrix and reinforcements and prevents oxidation of melt.
A steel stirrer was used to mix the reinforcements in the melt uniformly and runs at
150 rpm with the help of a motor. Simultaneously, B4 C and CNT were added as the
reinforcements. CNT was functionalized using nitric acid solution before mixing it
with the aluminum melt. The composition of the composite was Al-80 wt% B4 C-
18.5 wt% and CNT-1.5 wt%. After uniform dispersion of the reinforcements in the
aluminum matrix melt, the mixture was poured into a rectangular die of dimension
12 * 100 *10 mm.

2.3 Wire-Cut EDM Process

The machining parameters that influence the metal removal rate and kerf width were
T on , T off , voltage, wire feed (in machine code values), and its levels are presented
in Table 1. A face-centered composite design (FCD) with total of 30 experimental
runs (six center points) has been selected, and experimentation was conducted on
the 4-axis CNC Eco-cut WEDM. The results were analyzed using Design–Expert
software.
The study focuses on the effects of these inputs on the responses, namely MRR,
kerf width, and surface roughness. A wire-cut EDM machine uses bronze-coated
copper wire of 0.25 mm diameter. 15 straight slit cuts were made on the composite.
During the machining process, a EDM oil dielectric fluid was used. The dielectric
18 K. Rajkumar et al.

Table 1 Process parameters and levels

S. No. Process parameter Levels
Units 1 2 3
1 Pulse on time (T on ) (µs) 126 128 130
2 Pulse off time (T off ) (µs) 50 55 60
3 Spark voltage (SV) (V) 20 30 40
4 Wire feed (WF) (mm/s) 2 3 4

fluid flow is perpendicular to the direction of the wire feed. The typical cut job is
shown in Fig. 1.
The kerf width was measured using an optical microscope. The typical cuts on the
workpiece are shown in Fig. 2. Kerf width was measured by using ImageJ software.
Material removal rate was calculated using a mathematical Eq. (1)

MRR = Vc ∗ b ∗ t (1)

where b—kerf width (mm), t—thickness of the workpiece in mm, and surface rough-
ness was calculated in Ra using a tally surf roughness meter with cutoff length
0.25 mm.

Fig. 1 Typical wire EDM processed workpiece

Fig. 2 Typical wire EDM cut on the composite

Desirability Approach Machining Study on Aluminum … 19

3 Results and Discussion

The effects of the process parameters on the MRR, surface roughness, and Kerf width
are shown in Table 2.

Table 2 Experiments results and responses

Std Run SV Ton T off WF MRR Kerf Surface
(V) (µs) (µs) (m/min) (mm3 /min) (mm) roughness
(Ra)
1 1 20.00 126.00 50.00 2.00 0.297 0.364 1.489
9 2 20.00 126.00 50.00 4.00 0.332 0.359 1.412
16 3 40.00 130.00 60.00 4.00 0.451 0.369 1.676
3 4 20.00 130.00 50.00 2.00 0.322 0.382 1.521
13 5 20.00 126.00 60.00 4.00 0.288 0.345 1.405
30 6 30.00 128.00 55.00 3.00 0.341 0.369 1.565
19 7 30.00 126.00 55.00 3.00 0.311 0.374 1.592
27 8 30.00 128.00 55.00 3.00 0.305 0.373 1.543
8 9 40.00 130.00 60.00 2.00 0.42 0.391 1.745
29 10 30.00 128.00 55.00 3.00 0.311 0.369 1.645
23 11 30.00 128.00 55.00 2.00 0.353 0.371 1.622
22 12 30.00 128.00 60.00 3.00 0.313 0.372 1.601
4 13 40.00 130.00 50.00 2.00 0.424 0.382 1.673
28 14 30.00 128.00 55.00 3.00 0.342 0.372 1.556
14 15 40.00 126.00 60.00 4.00 0.389 0.359 1.645
5 16 20.00 126.00 60.00 2.00 0.298 0.353 1.539
15 17 20.00 130.00 60.00 4.00 0.279 0.372 1.422
21 18 30.00 128.00 50.00 3.00 0.358 0.374 1.572
7 19 20.00 130.00 60.00 2.00 0.31 0.378 1.51
12 20 40.00 130.00 50.00 4.00 0.426 0.379 1.562
11 21 20.00 130.00 50.00 4.00 0.342 0.391 1.465
25 22 30.00 128.00 55.00 3.00 0.35 0.372 1.573
24 23 30.00 128.00 55.00 4.00 0.326 0.348 1.532
17 24 20.00 128.00 55.00 3.00 0.322 0.354 1.495
2 25 40.00 126.00 50.00 2.00 0.423 0.349 1.675
6 26 40.00 126.00 60.00 2.00 0.412 0.381 1.645
20 27 30.00 130.00 55.00 3.00 0.373 0.392 1.567
18 28 40.00 128.00 55.00 3.00 0.422 0.364 1.632
26 29 30.00 128.00 55.00 3.00 0.324 0.369 1.541
10 30 40.00 126.00 50.00 4.00 0.382 0.342 1.622
20 K. Rajkumar et al.

In order to understand the effects of process parameters on MRR, contour plots

are shown in Figs. 3 and 4. The effect of variation in Ton and voltage on the MRR is
shown in Fig. 3. It is seen that the MRR increases with increase in voltage. Besides,
it also increases current. As a result of high intensity sparks, more material was
melted and evaporated. It is found that 38 V is sufficient for material removal. T on
also has a significant effect on MRR. Only after a certain value of T on , higher values
of MRR can be achieved. It is found that 128 µs is sufficient for higher material
removal. Higher Ton values mean that longer duration of the spark. This leads to
more material erosion. Fig. 4 also reveals that voltage was also significant, and it
increases MRR. But considering effect of T off on the MRR is a little effect, as longer
T off only reduces the MRR. The low T off (50 µs) is producing higher MRR.
Figures 5 and 6 show the kerf width with effect of voltage, current on and off time.
It is seen that the kerf dimension decreases gradually up to 30 V. Then, it increases
with further increase in voltage. When the cut is initiated, there was no debris and
balance between material erosion and spark intensity. With increase in voltage, the
kerf width increases due to higher order of material removal. This is attributed to the
high intensity sparks. The localized melting results in resolidified layer on both sides
of the cut surfaces. It is also revealed that the kerf dimension increases with increase
in Ton value due to longer time spark eroded side materials. Moreover, Fig. 6 shows
the kerf narrow region at pulse off time 55 and voltage 30.
Figure 7 shows the effect of wire speed and voltage on the surface roughness
of the composite. From the plot, it is clear that when the wire speed increases, the
surface roughness decreases. Due to wire electrode moving fast on the workpiece,

Contour Plot of MRR vs Ton, voltage

130
MRR
< 0.32
0.32 – 0.34
0.34 – 0.36

129 0.36 – 0.38

0.38 – 0.40
0.40 – 0.42
> 0.42

Hold Values
Ton

128 Toff 55
WS 3

127

126
20 25 30 35 40
voltage

Fig. 3 Contour plot of MRR versus T on , voltage

Desirability Approach Machining Study on Aluminum … 21

Fig. 4 Contour plot of MRR versus T off , voltage

Fig. 5 Contour plot of kerf versus T on , voltage

the availability of spark time on the local spot is low, resulting in little erosion of
material. This results in a smoother surface. MRR and surface roughness are inversely
proportional, and hence, as voltage increases, the surface roughness also increases.
In Fig. 8, it can be concluded that similar observation is evident. The combination
of high wire speed and low T on is evident of smoother surface of cut. Moreover, it is
22 K. Rajkumar et al.

Fig. 6 Contour plot of kerf versus T off , voltage

Fig. 7 Contour plot of Ra versus WS, voltage

observed that Ra with slope is steeper than the T on which indicates that voltage has
more significant effect on surface roughness.
Figure 9 shows the desirability values of optimized multiple responses of wire-cut
EDM on the given composite material. The desirability of the cut is having a high
MRR, a low kerf width, and a low surface roughness, and in order to achieve that,
optimization was done, to identify the best balance between the three requirements.
Desirability Approach Machining Study on Aluminum … 23

Fig. 8 Contour plot of Ra versus WS, T on

The corresponding input parameters were found to be exhibiting the best results; the
desirability of the cut was relatively higher when the combination of input param-
eters was of 39.99 V, 128.94 µs-T on , 50 µs-T off , and 4 mm/s-WS, than any other
combinations.

4 Conclusions

AA6061-B4 C-CNT hybrid composite was successfully fabricated, and wire-cut

EDM process was carried out with face-centered composite experiment design. From
the experiment results, the following conclusions are drawn:
1. MRR increases with increase in voltage. It is found that 38 V is sufficient for
material removal. T on also has a significant effect on MRR. It is found that 128 µs
shows off higher material removal. Higher T on leads to more material erosion. It
is observed that T off on the MRR has a little effect, but longer T off reduces the
MRR. By comparing results, voltage has a more significant effect on the MRR
than T on .
2. WEDM experimentation reveals that kerf dimension decreases gradually up to
30 V. But again, it increases with voltage. This could be due to the high order
of material removal. The localized melting generates a resolidified layer on both
sides of the cut surface, affecting the kerf dimension. It is shown that the kerf
dimension increases with increase in T on, but kerf narrow region is obtained at
pulse off time 55 µs and voltage 30.
24 K. Rajkumar et al.

Fig. 9 Desirability values of process parameters

3. It is evident that the smoothness of cut surface is obtained with increasing in wire
speed. The reason is that a fast-moving wire electrode has less time for on a local
spot leading to a reduction of erosion of workpiece material. The combination
of high wire speed and low T on offered smoother surface of cut.
4. The combination of 39.99 V, 128.94 µs-T on , 50 µs-T off, and 4 mm/s-wire speed
provided a relatively more desirable cut than any other combinations of input
parameters.

References

1. Bharath V, Nagaral M, Auradi V, Kori SA (2014) Preparation of 6061Al-Al2O3 MMC’s by stir

casting and evaluation of mechanical and wear properties. Procedia Mater. Sci. 6:1658–1667
2. Sankar M, Gnanavelbabu A, Rajkumar K (2014) Effect of reinforcement particles on the abra-
sive assisted electrochemical machining of Aluminium-Boron carbide-Graphite composite.
Procedia Eng 97:381–389
Desirability Approach Machining Study on Aluminum … 25

3. Yadav RK, Abhishek K, Mahapatra SS (2015) A simulation approach for estimating flank wear
and material removal rate in turning of Inconel 718. Simul Model Pract Theory 52:1–14
4. Kong L, Zhu S, Bi Q, Qiao Z, Yang J, Liu W (2014) Friction and wear behavior of self-
lubricating ZrO2 (Y2O3)–CaF2–Mo–graphite composite from 20 to 1000 °C. Ceram Int
40(7):10787–10792
5. Vangla P, Gali ML (2016) Effect of particle size of sand and surface asperities of reinforcement
on their interface shear behaviour. Geotext Geomembr 44(3):254–268
6. Banks-Sills L, Shiber DG, Fourman V, Eliasi R, Shlayer A (2016) Experimental determination
of mechanical properties of PMMA reinforced with functionalized CNTs. Compos B Eng
95:335–345
7. Moghadam AD, Omrani E, Menezes PL, Rohatgi PK (2015) Mechanical and tribological
properties of self-lubricating metal matrix nanocomposites reinforced by carbon nanotubes
(CNTs) and graphene–a review. Compos B Eng 77:402–420
8. Rajkumar K, Santosh S, Ibrahim SJS, Gnanavelbabu A (2014) Effect of Electrical dis-
charge machining parameters on microwave heat treated Aluminium-Boron carbide-Graphite
composites. Procedia Eng 97:1543–1550
9. Paliwal S, Solanki P (2014) Parameter optimization of wire electrical discharge machining for
minimum surface roughness and kerf width using taguchi method. In: 4th SARC international
conference
10. Hewidy MS, El-Taweel TA, El-Safty MF (2005) Modelling the machining parameters of wire
electrical discharge machining of Inconel 601 using RSM. J Mater Process Technol 169(2):328–
336
11. Tilekar S, Das SS, Patowari PK (2014) Process parameter optimization of wire EDM on
Aluminum and mild steel by using taguchi method. Procedia Mater Sci 5:2577–2584
12. Bobbili R, Madhu V, Gogia AK (2015) An experimental investigation of wire electrical
discharge machining of hot-pressed boron carbide. Defence Technol 11(4):344–349
13. Prakash JU, Moorthy TV, Peter JM (2013) Experimental investigations on machinability
of aluminium alloy (A413)/Flyash/B4C hybrid composites using wire EDM. Procedia Eng
64:1344–1353
A Comparative Study on Abrasive Water
Jet Machining Characteristics of Entry
and Exit Layers of Glass and Basalt
Woven Polymer Composites

K. Ramraji , K. Rajkumar , M. Rajesh, and A. Gnanavelbabu

Abstract Abrasive water jet machining (AWJM) induced kerf and kerf tapper angle
variation is much affected by the properties of the entry and exit layer of fiber
reinforced polymer of composite. In this work, experimental investigation on the
machining characteristics of vinyl ester composite fabricated by basalt and glass
fiber as top and bottom layers separately interlayered with flax fiber. Hybridization
of basalt/glass composite was made by the additional layers of flax woven. Two dif-
ferent stacked composites were fabricated by using hand layup process followed by
static compression loading. The mechanical properties such as tensile and flexural
strength were studied. It was found that top and bottom basalt layers interlayered
flax fiber composite exhibited higher tensile and flexural properties. The tensile and
flexural strength improvements were 7.15% and flexural 13.3% over the top and bot-
tom glass layers interlayered flax fiber composite. The fabricated composites were
machined by abrasive water jet machining (AWJM) using constant cutting parame-
ters with pressure, nozzle speed, and standoff distance. Experimental results reveal
that minimum kerf taper and higher cutting surface quality obtained with entry and
exit layers of glass fiber composite.

Keywords Polymer composite · AWJM · Flax fiber · Kerf · Interlayered fiber

1 Introduction

Nowadays, FRP composite is losing its merits by environmental concerns and gov-
ernment regulation [1]. Natural fibers most widely accepted have many advantage
properties as reinforcement to composite. Owing to relatively low density and high
strength, the natural fiber is popular now. The natural fibers are extracted from the

K. Ramraji · K. Rajkumar (B) · M. Rajesh

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, Tamil Nadu, India
e-mail: rajkumark@ssn.edu.in
A. Gnanavelbabu
Department of Industrial Engineering, College of Engineering, Chennai, Tamil Nadu, India

© Springer Nature Singapore Pte Ltd. 2021 27

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_4
28 K. Ramraji et al.

plants, nutshells, palm shells, wood flour, and animals [2]. Moreover, they have enor-
mous merits over synthetic fiber such as renewable in nature and less abrasive [3,
4]. However, its mechanical properties are much lower than synthetic fiber [5]. This
limits to use in the semi-structural application.
Natural fiber composite exhibits poor resistance to moisture absorption because
of its hydrophilic behavior. Thus, it makes less attractive but rather using suitable
chemical treatments rectifying hydrophilic problem [6]. The reinforcing of natural
and synthetic fibers in a polymer matrix provides the best properties. The hybridiza-
tion of fibers is remarking superiority and reduces relay on the synthetic fibers. Hybrid
composite is being now enough strength, stiffness, and rigidity to satisfy the needs of
semi-structural applications [7, 8]. Apart from that, composite required machining
process, to shape, size, and tolerance for an application.
Conventional machining processes namely drilling, turning, and milling will tend
to pull out fibers from the composite. Additionally, it leads to delamination and
damage to composite. Almost 60% of the composites are rejected due to damages
like fiber pullout and delamination [9]. In order to improve the machining quality of
composite, these works made an attempt with abrasive water jet machining (AWJM)
process. This process is environmentally safe and free from heat-induced detects.
Researchers are more interested in AWJM of the composite to evaluate kerf ratio,
kerf angle, and surface quality parameters [10, 11].
Moreover, the selection of proper machining parameters is more concern to the
researcher and machinist for quality machining. A few researchers revealed that
standoff distance is the most significant factor to reduce the kerf ratio followed by the
traverse speed. The minimum traverse speed and maximum pressure were smoothly
removed material from the FRP composite and also minimize delamination fracture
[12]. The combination of traverse rate and standoff distance increases kerf ratio
but decreases standoff distance resulting in a smooth surface and lowers the kerf
ratio as a result of the increased kinetic energy of jet. Increase in water pressure
and abrasive flow rate improved surface quality of composite [13, 14]. Selvam et al.
[15] investigated AWJM cutting parameters on the surface and kerf taper by response
surface model (RSM). They concluded that kerf and surface roughness are a function
of traverse speed and abrasive flow rate.
A comparative study on the kerf and surface quality by changing entry and exit
layer of glass and basalt woven fiber polymer composite by AWJM process has been
made in this work.
A Comparative Study on Abrasive Water … 29

Table 1 Properties of flax, basalt, and glass fibers [2, 16]

Properties Unit Flax Basalt Glass
Density g/cm3 2.80 1.40 2.55
Tensile module GPa 87 51 73
Tensile strength MPa 4700 1034 2400
Elongation at break mm 3.15 1.5 3

2 Materials and Method

2.1 Materials

The interplay hybrid FRP composite was made by woven fabrics of basalt, flax,
E-glass, and vinyl ester resin of grade VBR 4508 was used as matrix. Regular accel-
erator, promotor, and catalyst were used for curing agent. The properties of flax,
basalt, and glass fabric are summarized in Table 1.

2.2 Fabrication of Composite Laminates

The flax fabric was treated in sodium hydroxide (NaOH) solution with the concen-
tration of 5% in one liter of water for the duration of two hours at room temperature.
NaOH-treated fibers were washed by tap water followed with distilled water. FRP
composite was fabricated by a hand layup process with dead weight. The stacking
sequences of FRP composites were G2F7 and B2F7. G2F7 has two layers of E-glass
woven fabric (top and bottom) and seven layers of flax woven fabric. B2F7 has
two layers of basalt woven fabric (top and bottom) and seven layers of flax woven
fabrics. The ratio of reinforcement weight fraction to the matrix weight fraction is
maintained at 40:60. The size of the laminate was a square of 300 × 300 mm. The
stacking sequence of laminates was shown in Fig. 1. The actual thickness of the
laminate was approximately 7.2 mm for both B2F7 and for G2F7.

2.3 Tensile and Flexural Strength

Universal testing machine (UTM) was used to evaluate the tensile and flexural prop-
erties of the composite. To measure tensile property, the test sample was made in
accordance with ASTM-D 638 and tested at a uniform crosshead speed of 5 mm/min.
Three-point bending tester was performed to the specimen as per ASTM D790 the
standard for the flexural property evaluation. The sample was tested with a strain rate
of 5 mm/min. Typical tensile and flexural fracture specimens are shown in Fig. 2a–d.
30 K. Ramraji et al.

Fig. 1 Stacking layers of B2F7 and G2F7 composite laminates

Fig. 2 a–d Tensile and flexural fractured specimens of B2F7 and G2F7 composites

2.4 Composite Characterization

The surface morphology of tensile fractured and machining surface was studied
through Scanning Electron Microscope SEM (JOEL JSM840A). The failure mech-
anisms of the composite were analyzed through fiber and matrix intact by SEM
images.
A Comparative Study on Abrasive Water … 31

2.5 AWJM Process

A two-axis OMAX-1530(CNC) water jet cutting machine was used for experimen-
tation. AWJM nozzle assembly consists of an orifice of (0.35 mm diameter) with a
sapphire jewel and carbide focusing tube of 0.35 mm internal diameter and length
of 75 mm. The impinging angle of the jet was maintained at 90°. The mixing tube
diameter was 0.762 mm, and the abrasive size was 80 mesh and tool offset of 0.3 mm.
These standard set of AWJM geometry are used widely in most industrial applica-
tions. The photocopy of AWM machining of composite setup is shown in Fig. 3. The
selected parameters and their ranges listed in Table 2.
In this study, surface roughness was measured over the cut by SURFTEST SJ-210.
Three different locations of readings for Ra were measured, and then, the average
value was calculated. The kerf taper was measured using a video measuring system
(VMS), as shown in Fig. 4. Kerf taper was calculated by measuring top, and bottom
of kerf average value of three readings reported in order to minimize the error. The
kerf taper is calculated as given in Eq. (1)

Lt − Lb
Kerf Taper = (1)
2T
Where L t —top kerf width (mm), L b —bottom kerf width (mm), and T —thickness
of the laminate (mm).

Fig. 3 Photocopy of AWJ machining setup and typical machined G2F7 and B2F7 composites

Table 2 AWJM parameters and their ranges

Process parameters Units Levels
Hydraulic pressure (HP) MPa 160 240 320
Standoff distance (SOD) mm 1.5 2.5 3.5
Traverse rate (TR) mm/min 150 225 300
32 K. Ramraji et al.

Fig. 4 Photocopy of VMS for kerf width measurement

3 Results and Discussion

3.1 Tensile and Flexural Strength

Tensile and flexural strength of B2F7 and G2F7 composites are represented in
Fig. 5a, b. From this figure, it clearly shows that the maximum tensile and flexu-
ral strength were obtained with B2F7 composite. The tensile strength and flexural
strength improvement were 7.15% and 13.3%, respectively, over the G2F7 compos-
ite. This could be a typical reaction between vinyl ester resin and basalt fiber reacting a
good interfacial strength. The basalt surface layers provide the load-bearing strength
to the composite. This is certainly delayed fracture and pullout of basalt fibers. As

Fig. 5 a, b Tensile and flexural strength of B2F7 and G2F7 composites

A Comparative Study on Abrasive Water … 33

Fig. 6 Typical SEM images of tensile fracture surface

in flexural strength, the top and bottom layers are so important to redistribute the
bending stress to next interlayers. The sufficient flexibility provided by the basalt
fibers increases the bending strength of the composite.
The typical SEM images of the tensile fracture surface are shown in Fig. 6a, b.
It is observed that good adhesion was seen between basalt fiber and matrix, as in
the case of B2F7composite. G2F7 composite reveals matrix crack, fiber pullout, and
breakage, as shown in Fig. 6b This is due to that changes in fiber surface morphology
during loading which affecting the fiber/matrix interfacial adhesion strength.

3.2 Machining Characteristics of Interplied Composites

3.2.1 Effect of Water Pressure on Kerf Taper and Surface Roughness

by Water Pressure

The effect of water pressure on the kerf taper of B2F7 and G2F7 composites is shown
in Fig. 7a. This figure shows that the kerf taper ratio decreases with an increase in
water pressure for both the B2F7 and G2F7 composites. Moreover, G2F7 composite
exhibited minimum kerf taper. The reduction of kerf taper observed to be 14% at a
pressure of 300 MPa. B2F7 composite is rather produced higher taper angle due to
more brittle nature of basalt fiber. The brittle fracture of the top layer of basalt fiber
increases with increase in water jet energy. This results in larger access by the water
jet next layer of flax fabrics loading to higher kerf taper. As water jet moving in a
transverse direction, the energy of water jet gets lowered due to friction with cutting
surface. It reduces the jet ability to the removal of material at bottom layers. The
effect of surface roughness by water pressure for B2F7 and G2F7 composites is shown
in Fig. 7b. It can be seen that surface quality increases when an increase in water
pressure for both the B2F7 and G2F7 composites. Moreover, G2F7 composite shows
34 K. Ramraji et al.

Fig. 7 a, b Kerf taper and surface roughness with water pressure for the B2F2 and G2F2 composites

a minimum surface roughness with 53.05% lower than B2F7 composite. Owing to
uneven machining in B2F7 composite, the surface roughness was higher.

3.2.2 Effect of Mean Kerf Taper and Surface Roughness by Nozzle

Traverse Speed

The effect of nozzle traverse speed on the kerf taper of B2F7 and G2F7 composites is
shown in Fig. 8a. This figure clearly reveals that kerf taper increases with increasing
the nozzle traverse speed for both the B2F7 and G2F7 composites. Moreover, G2F7
composite shows a minimum kerf taper with 6.2% at a traverse speed of 125 mm/min,
B2F7 composite. This could be due to the energy unbalance between energy require-
ment for cutting and energy supplied to the point of cutting. This unbalanced cutting

Fig. 8 a, b Kerf taper and surface roughness with nozzle traverse speed for B2F2 and G2F2
composites
A Comparative Study on Abrasive Water … 35

Fig. 9 a, b Kerf taper and surface roughness with standoff distance for B2F7 and G2F7 composites

energy increased with traverse speed left larger kerf taper. The effect of nozzle tra-
verse speed on the surface roughness of B2F7 and G2F7 composites is shown in
Fig. 8b. Surface quality decreases with increasing the nozzle traverse speed for both
the B2F7 and G2F7 composites. This is attributed to the un even cutting with var-
ied water jet energy in the traverse direction. Moreover, G2F7 composite shows a
minimum kerf taper with 45.1% lower than B2F7 composite.

3.2.3 Effect of Mean Kerf Taper and Surface Roughness by Standoff

Distance

The effect of standoff distance on the kerf taper of B2F7 and G2F7 composites is
shown in Fig. 9a. Kerf taper ratio increases with every step increment of standoff
distance for both B2F7 and G2F7 composites. This could be to water jet flaring with
distance and loss of water jet energy. Moreover, G2F7 composite shows a minimum
kerf taper with a reduction of 5.9% at a standoff distance of 1.5 mm as compared
with B2F7 composite. The effect of standoff distance on the surface roughness of
B2F7 and G2F7 composites is shown in Fig. 9b. The surface quality of the machined
surface decreases with increasing the standoff distance for both B2F7 and G2F7
composites. This is a reduction in localized material erosion by flared water jet which
results in non-uniform cutting over the surface. Interestingly, G2F7 composite reveals
minimum surface roughness by 49.2% at standoff distance 1.5 mm than the B2F7
composite.
36 K. Ramraji et al.

4 Conclusions

The entry and exit of basalt and glass layers with interlayered flax fiber reinforced
polymeric composite were most influencing the mechanical properties and AWJ
machining process. From the results, the following conclusion were made,
1. The maximum tensile and flexural strength were obtained with entry and exit
basalt layer of B2F7 composite. B2F7 composite shows an improved tensile
strength by 7.15 and 13.3% for flexural strength when compared to the entry and
exit glass layers of G2F7 composite. This is a result of a typical reaction between
vinyl ester resin and basalt fiber, improving the good interfacial strength of the
composite.
2. Water jet pressure was considered as the most significant control factor in influ-
encing surface quality (Ra) and kerf taper. Increase in pressure results in a better
machining performance for both B2F7 and G2F7 composites. G2F7 composite
shows a minimum kerf taper by 14% at a pressure of 300 MPa, as compared to
B2F7 composite. As a result of more brittle nature of basalt fiber (B2F7) increases
taper angle. Whilst, G2F7 composite shows a minimum surface roughness with
53.05% at a pressure of 300 MPa compared with G2F7. This could be uneven
machining in B2F7 composite leads to poor surface quality.
3. Interestingly, another machining parameters, traverse rate, and standoff distance
at low level improved the machining performance of both the composites.
4. These experiments confirm that greater kinetic energy and pressure energy
required for abrasive water jet machining of interlayered composites.

References

1. Frollini E, Silva CG, Ramires EC (2013) Advanced fibre-reinforced polymer (FRP) composites
for structural applications: 2. Phenolic resins as a matrix material in advanced fiber-reinforced
polymer (FRP) composites. Elsevier Inc. Chapters
2. Westman MP, Fifield LS, Simmons KL, Laddha S, Kafentzis TA (2010) Natural fiber com-
posites: a review (No. PNNL-19220). Pacific Northwest National Lab (PNNL), Richland, WA
(United States)
3. Monteiro SN, Lopes FPD, Ferreira AS, Nascimento DCO (2009) Natural-fiber polymer-matrix
composites: cheaper, tougher, and environmentally friendly. JOM 61(1):17–22
4. Kalia S, Kaith BS, Kaur I (2009) Pretreatments of natural fibers and their application as
reinforcing material in polymer composites—a review. Polym Eng Sci 49(7):1253–1272
5. Joshi SV, Drzal LT, Mohanty AK, Arora S (2004) Are natural fiber composites environmentally
superior to glass fiber reinforced composites? Compos A Appl Sci Manuf 35(3):371–376
6. Ramraji K, Rajkumar K, Sabarinathan P (2019) Tailoring of tensile and dynamic thermo-
mechanical properties of interleaved chemical-treated fine almond shell particulate flax fiber
stacked vinyl ester polymeric composites. In: Proceedings of the Institution of Mechanical
Engineers, Part L: Journal of Materials: Design and Applications, p 1464420719849616
7. Gujjala R, Ojha S, Acharya SK, Pal SK (2014) Mechanical properties of woven jute–glass
hybrid-reinforced epoxy composite. J Compos Mater 48(28):3445–3455
A Comparative Study on Abrasive Water … 37

8. Sanjay MR, Arpitha GR, Yogesha B (2015) Study on mechanical properties of natural-glass
fibre reinforced polymer hybrid composites: a review. Mater Today: Proc 2(4–5):2959–2967
9. Ramkumar J, Malhotra SK, Krishnamurthy R (2004) Effect of workpiece vibration on drilling
of GFRP laminates. J Mater Process Technol 152(3):329–332
10. Wang J, Wong WCK (1999) A study of waterjet cutting of metallic coated sheet steels. Int J
Mach Tools Manuf 39:855–870 (1999)
11. Gnanavelbabu A, Saravanan P, Rajkumar K, Karthikeyan S (2018) Experimental investigations
on multiple responses in abrasive waterjet machining of Ti-6Al-4 V alloy. Mater Today: Proc
5(5):13413–13421
12. Irina Wong MM, Azmi AI, Lee CC, Mansor AF (2018) Kerf taper and delamination dam-
age minimization of FRP hybrid composites under abrasive water-jet machining. Int J Adv
Manufact Technol 94(5–8):1727–1744
13. Azmir MA, Ahsan AK (2009) A study of abrasive water jet machining process on glass/epoxy
composite laminate. J Mater Process Technol 209(20):6168–6173
14. El-Hofy M, Helmy MO, Escobar-Palafox G, Kerrigan KM, Scaife R, El-Hofy H (2018) Abra-
sive water jet machining of multidirectional CFRP laminates. Procedia CIRP 68:535–540
(Elsevier)
15. Selvam R, Karunamoorthy L, Arunkumar N (2017) Investigation on performance of abrasive
water jet in machining hybrid composites. Mater Manuf Processes 32(6):700–706
16. Singha K (2012) A short review on basalt fiber. Int J Text Sci 1(4):19–28
Electrolyte and Machining Parameters
Optimization of Wire Electrochemical
Cutting of Aluminum/Titanium Diboride
Composite

K. Rajkumar , M. Rajesh , K. Ramraji , P. Sabarinathan ,

and A. Gnanavelbabu

Abstract A hard to machining characteristics of ceramic reinforced aluminum com-

posite requires unconventional machining methods. Wire electrochemical machining
(WECM) is one of the superior machining processes that produced a smoother and
brighter surface with complex profile. In this investigation, optimization of process
parameters is mainly focused for slit profile cutting on the aluminum 6061/TiB2
composite by a WECM method. Liquid metallurgy route is employed to fabricate
the composite with varied TiB2 weight percentages (5 and 15%). The parameters of
wire electrochemical cutting process are applied voltage (10–14 V), electrolyte flow
rate (2–8 l/min), and electrolyte concentration (12–18%). These were optimized for
achieving optimal material removal rate and surface roughness. The optimization
of process parameters was carried out based on the response surface methodology.
Electrolyte flow rate significantly increases MRR, surface roughness, and slit width.
When considering the optimal MRR and slit width condition, the experiment results
reveal a better surface finish at low-level electrolyte concentration. Inter-electrode
voltage increases the strength of electrochemical ionization resulting in higher MRR,
and on the other hand, it is considerably affected surface roughness.

Keywords Wire electrochemical machining · Surface roughness · TiB2 ·

Electrolyte · Composite machining

1 Introduction

Aluminum metal matrix composites have opened a new window to various indus-
trial sectors like automobile components—drive shaft, brakes, cylinder liners, engine

K. Rajkumar (B) · M. Rajesh · K. Ramraji

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, Tamil Nadu, India
e-mail: rajkumark@ssn.edu.in
P. Sabarinathan · A. Gnanavelbabu
Department of Industrial Engineering, College of Engineering, Chennai, Tamil Nadu, India

© Springer Nature Singapore Pte Ltd. 2021 39

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_5
40 K. Rajkumar et al.

locks, and motor casing [1]. Generally, aluminum metal matrix composites are rein-
forced with mono or combination of reinforcements such as oxides, nitride, carbide,
and borides [2] to increase its mechanical properties. The addition of the reinforce-
ment leads to complexity in machining of composite. Normally, the conventional
machining process cannot help in the context of hard materials like a metal matrix
composites and ceramic materials [3]. A selected machining technology requires to
produce good surface integrity for components made from particulate reinforced
composites. The past and present scenarios show an electrochemical machining
that fulfills the industry requirements without much disturbing the materials being
machined. Another variant of an electrochemical process namely wire electrochemi-
cal machining process is a hybrid unconventional metal cutting process for very hard
materials [4]. The reasons for selecting a wire electrochemical machining process,
specifically for metal matrix composites, are produced complex shapes, profiles, and
slits without residual stress.
Normally, material removal rate (MRR), surface roughness, and slit width are the
essential characteristics for the WEDM machining operations [5]. These character-
istic parameters depend on the machining gap, concentration of electrolyte, type of
electrolyte, wire diameter, and current source characteristics. Senthilkumar et al. [6]
obtained higher MRR and lower surface roughness by performing electrochemical
machining of MMC with different process parameters. It is reported that increases
in voltage and feed rate resulting in lower surface roughness and higher MRR. Fur-
thermore, increasing in concentration and flow rate of electrolyte showed the same
result. Taweel et al. [7] were optimized wire electrochemical turning experimental
results using a response surface method. The input parameters such as feed rate, rota-
tional speed, and overlap distance were optimized to obtain higher material removal
rate and lower surface roughness. In case of ECM optimization, Bhattacharyya et al.
[8] developed a mathematical model to correlate machining response to process
parameters.
In the present study, the WEDM on the Al-5 & 15wt% TiB2 composites has
been done. Process parameters like inter-gap voltage, concentration, and flow rate
of electrolyte were optimized in order to obtain better machining responses such as
surface roughness, material removal rate, and slit width.

2 Materials and Method

Composites made from Al6061 matrix reinforced with TiB2 ceramic particles (size
of 10 µm) with two different proportions of (5% and 15wt%) using a stir casting
method. The fabricated composites were heat treated by T-6 heat treatment method.
Figure 1 shows the experimental machine setup of the wire electrochemical machine.
In a wire electrochemical machining process, copper wire of 0.5 mm diame-
ter can be used as a cathode and the Al6061/TiB2 as an anode. Both electrodes
were immersed in a sodium chloride electrolyte solution with wire feed rate of
1 mm/min. experiments were conducted by varying the input parameters such as
Electrolyte and Machining Parameters … 41

Fig. 1 Experimental machine setup WECM

Table 1 WECM process

Parameters Value
parameters
Tool (mm) Copper wire Ø 0.5
Electrolyte concentration (%) 12, 15, 18
Voltage (V) 10, 12, 14
Feed rate (mm/min) 1
Electrolyte flow rate (l/min) 2, 5, 8
Electrolyte Sodium chloride

DC voltage of (10–14 V), electrolyte concentration (12–18%), and flow rate (2–
8 l/min). The machining output parameters like MRR, slit width, and surface rough-
ness are reported. MRR was calculated using a mass loss method. Slit width was
observed by optical microscope. The surface roughness of machined surface was
measured by using a talysurf meter with cut-off length 0.8 mm. Table 1 shows
complete experimental parameters for WECM.

2.1 Response Surface Methodology

Material removal rate, surface roughness, and slit width were classified as response
functions to the voltage (A), electrolyte concentration (B), and electrolyte flow rate
(C) and were considered as control variables. The experiments were conducted using
a design of experiment technique with central composite design. The experiments
were run as sequence of L19 orthogonal array. Table 2 shows the process parameters
levels as per the central composite design. Table 3 presents an experimental design
matrix and 5 and 15% reinforced composite experimental results.
42 K. Rajkumar et al.

Table 2 Experimental
Process parameters Levels
parameters and their levels
−1 0 1
Voltage (V) 10 12 14
Electrolyte flow rate (l/min) 2 5 8
Electrolyte concentration (%) 12 15 18

Table 3 Experimental design matrix and results

Composition 5% reinforcement 15% reinforcement
(A) (B) (C) MRR Surface Slit MRR Surface Slit
Voltage Electrolyte Electrolyte (gm/min) roughness width (gm/min) roughness width
(V) concentration flow rate (µm) (mm) (µm) (mm)
(%) (l/min)
12 15 5 120 4.311 0.802 96 4.744 0.761
10 15 5 108 3.822 0.772 76 4.322 0.731
10 18 8 129 4.434 0.783 99 4.985 0.745
12 15 8 130 4.115 0.784 106 4.398 0.748
14 15 5 124 4.641 0.819 113 5.349 0.775
10 12 2 79 3.308 0.760 56 3.621 0.734
10 12 8 106 2.724 0.715 84 3.346 0.709
12 15 5 122 4.311 0.802 94 4.823 0.789
12 15 5 117 4.421 0.794 99 4.912 0.795
12 12 5 104 3.311 0.771 80 3.712 0.744
12 15 5 124 4.484 0.814 95 4.658 0.779
12 18 5 126 4.985 0.831 105 5.373 0.785
14 18 2 125 5.581 0.869 105 6.387 0.809
12 15 2 108 4.511 0.819 84 4.856 0.772
14 12 8 125 2.986 0.781 105 3.975 0.745
14 12 2 90 3.689 0.811 80 4.324 0.771
12 15 5 123 4.254 0.811 100 5.01 0.804
10 18 2 107 5.092 0.828 81 5.468 0.761
14 18 8 146 5.225 0.841 127 5.985 0.788

3 Results and Discussion

3.1 Mathematical Modeling

The mathematical model for the material removal rate when considering with various
process parameters for 5 and 15% reinforcement composites is shown in Eqs. 1 and
2,

MRR(5%) = + 120.25 + 8.10 ∗ A + 12.90 ∗ B + 12.70 ∗ C + 0.62 ∗ A ∗ B

+ 0.88 ∗ A ∗ C − 2.38 ∗ B ∗ C − 3.06 ∗ A2 − 4.06 ∗ B2 − 0.057 ∗ C2 (1)
Electrolyte and Machining Parameters … 43

MRR(15%) = +93.95 + 13.40 ∗ A + 11.20 ∗ B + 11.50 ∗ C (2)

The mathematical model for the surface roughness of machined surface of 5 and
15% reinforcement composites is shown in Eqs. 3 and 4,

Surface Roughness (5%) = +4.22 + 0.27 ∗ A + 0.93 ∗ B − 0.27 ∗ C (3)

Surface Roughness (15%) = +4.75 + 0.43 ∗ A + 0.92 ∗ B − 0.20 ∗ C (4)

The mathematical model for the obtained slit width while cutting with wire as
considering various process parameters for 5 and 15% reinforcement composites is
shown in Eqs. 5 and 6,

Slit width (5%) = +0.80 + 0.026 ∗ A + 0.031 ∗ B − 0.018 ∗ C (5)

Slit width (15%) = +0.77 + 0.021 ∗ A + 0.018 ∗ B − 0.011 ∗ C (6)

3.2 Parametric Influence on Material Removal Rate

Analyzing experimental work by DOE method and regression model for MRR, and
the effect caused by control variables is given in ANOVA Table 4. Composite 5
and 15% show that the developed MRR model was well-suited for machining of
composite material, as seen that model is highly significance. Likewise, each control
variable was affected material removal rate within confidence limit of 95%, as clearly
shown in Table 4.
Figure 2a, b shows a 3D model graph of MRR of composite machining with
electrolyte flow rate, concentration, and voltage. Electrolyte flow rate and concentra-
tion were directly affected by MRR. When increased electrolyte flow rate hydrogen
bubbles are removed which were generated at cathodic surface, this increases the
ionic strength of electrolyte in the slit cutting zone. Finally, it leads to high material
removal rate at anode surface. Besides, high flow rate of electrolyte can speed up
ionization process between inter-electrode gaps. This results in more anodic disso-
lution of material and easy flushing of formed debris. Further, electrolyte high flow
rate generates a fresh surface over and over again so that fresh un-machined surface
available for further electrochemical machining. This also resulted in increased MRR
with the flow rate. Similarly, the electrical conductivity of the electrolyte between
the wire and composite is increased by electrolyte concentration. Thus, electrolyte
conductivity between electrodes increased which speed up the electrochemical reac-
tion. This makes a process to steadily improve its cutting ability on the composite. It
is also observed that increases in the voltage and produces high machining current
 44

Table 4 ANOVA table for MRR

Source 5% reinforcement 15% reinforcement
Sum of squares df Mean square F value p-value Prob > Sum of squares df Mean square F value p-value Prob >
F value F value
Model 4179.78 9 464.42 58.91 <0.0001 4372.50 3 1457.5 106.93 <0.0001
A-voltage 656.10 1 656.10 83.22 <0.0001 1795.60 1 1795.6 131.74 <0.0001
B-electrolyte 1664.10 1 1664.1 211.0 <0.0001 1254.40 1 1254.4 92.03 <0.0001
concentration
C-electrolyte 1612.90 1 1612.9 204.5 <0.0001 1322.50 1 1322.5 97.03 <0.0001
flow rate
Residual 70.95 9 7.88 204.45 15 13.63
Pure error 30.80 4 7.70 26.80 4 6.70
Pred R-squared 0.8911 0.9258
Adj R-squared 0.9666 0.9464
K. Rajkumar et al.
Electrolyte and Machining Parameters … 45

Fig. 2 Effect of MRR a electrolyte flow rate and concentration b voltage and electrolyte
concentration

in the inter-electrode gap. It might result in increased current density between the
electrodes that increased material removal rate at anode surface [9].

3.3 Parametric Influence on Surface Roughness

Table 5 shows an ANOVA table for surface roughness of machined surface of compos-
ites. This table indicates developed surface roughness model which perfectly reflects
the control variables behavior to the experimental results. The control variables were
adequately fit into the developed mathematical model.
Figure 3a, b shows the 3D model graph of surface roughness with the electrolyte
flow rate, concentration, and voltage. It is observed from the plot that surface rough-
ness slightly increases with increase in flow rate. Owing to flow streaks on the surface
increases unsteady rotating eddies on the surface. This results in high turbulence
over the surface. Surface roughness increases with increasing electrolyte concentra-
tion. At low concentration, the ion depletion in electrolyte between inter-electrode
gap results in reduced surface roughness [10]. The current density increases with
increase in electrolyte concentration which leads to pitting on the cutting surface. It
is observed that surface roughness affected with inter-electrode potential difference
increases. When voltage increased the metal, dissolution was also increased. This
leads to an uneven dissolution around the cutting surface resulting in poor surface
finish.

3.4 Parametric Influence on Slit Width

Table 6 shows an ANOVA table for slit width of composites. This table clearly
indicates that developed slit width model with consideration of all control variables
is well-fitted by seeing an evidence of less than 5% P-value. All the control variables
were significantly affected the process
 46

Table 5 ANOVA table for surface roughness

Source 5% 15%
Sum of squares df Mean square F value p-value prob > Sum of squares df Mean square F value p-value prob >
F value F value
Model 10.13 3 3.38 130.74 <0.0001 10.72 3 3.57 159.4 <0.0001
A-voltage 0.75 1 0.75 29.12 <0.0001 1.83 1 1.83 81.65 <0.0001
B-electrolyte 8.65 1 8.65 334.93 <0.0001 8.50 1 8.50 379.28 <0.0001
concentration
C-electrolyte 0.73 1 0.73 28.17 <0.0001 0.39 1 0.39 17.26 0.0008
flow rate
Residual 0.39 15 0.026 0.34 15 0.022
Pure error 0.035 4 8.766E−003 0.076 4 0.019
Pred 0.9360 0.9546
R-squared
Adj R-squared 0.9558 0.9635
K. Rajkumar et al.
Electrolyte and Machining Parameters … 47

Fig. 3 Surface roughness plot a electrolyte flow rate and concentration b voltage and electrolyte
concentration

The slit width of the composites as function of different process variables are
shown in Fig. 4a, b. It is observed that high voltage increased order of MRR. Increase
in metal dissolution increases the slit width. Similarly, at high flow rate the burrs were
removed and enlarge the slit width. Hence, the medium flow rate was optimal for
machining of composites.
Slit width was observed using optical microscopy images. Figure 5a, b shows
the slit profiles of machined composites. 5% reinforcement composite shows a line
straight geometric profile (Fig. 5a). This is the effect of lesser reinforcement particle’s
presence in composite. It is understood that these particles slightly interfere the
machining. The presence of larger amount of nonconductive ceramic particles, in case
of 15% reinforcement composite, leads to inhomogeneous dissolution and uneven
distribution of current intensity resulted in an entry side enlarge of slit (Fig. 5b).

4 Conclusions

In this present study, the various compositions of aluminum/titanium diboride com-

posites were machined using a wire electrochemical machining process. The machin-
ing parameters were optimized by using response surface methodology. Mathe-
matical model for the responses such as MRR, slit width, and surface quality was
developed, and the optimal conditions are presented.
1. Electrolyte flow rate and concentration were directly affected MRR of composite
machining. An electrolyte flow rate reduces hydrogen bubbles attachment over
the cathodic surface. This improves ionic reaction of electrolyte in cutting zone;
therefore, high material removal rate is observed at anode surface.
2. The electrolyte electrical conductivity between electrodes increases with elec-
trolyte concentration. This also increases electrochemical reaction resulting in
higher MRR.
3. An increase in electrolyte flow rate substantially increases the surface rough-
ness. Owing to flow streaks over the cutting surface an unsteady rotating eddy
 48

Table 6 ANOVA table for slit width

Source 5% 15%
Sum of squares df Mean square F value p-value prob Sum of squares df Mean square F value p-value prob
>F value >F value
Model 0.020 3 6.708E−003 162.56 <0.0001 9.003E−003 3 3.001E−003 11.12 0.0004
A-Voltage 6.917E−003 1 6.917E−003 167.61 <0.0001 4.326E−003 1 4.326E−003 16.03 0.0012
B-Electrolyte 9.860E−003 1 9.860E−003 238.92 <0.0001 3.422E−003 1 3.422E−003 12.68 0.0028
concentration
C-Electrolyte 3.349E−003 1 3.349E−003 81.15 <0.0001 1.254E−003 1 1.254E−003 4.65 0.0478
flow rate
Residual 6.190E−004 15 4.127E−005 4.049E−003 15 2.700E−004
Pure error 2.552E−004 4 6.380E−005 2.552E−004 4 6.380E−005
Pred R-squared 0.9534 0.5916
Adj R-squared 0.9642 0.6277
K. Rajkumar et al.
Electrolyte and Machining Parameters … 49

Fig. 4 Slit width plot a electrolyte flow rate and concentration b voltage and electrolyte
concentration

Fig. 5 Slit profiles of machined composites. a 5% composite and b 10% composite

increases which results in an uneven machining. Surface roughness increases

with increasing electrolyte concentration. The current density increases with
increasing electrolyte concentration that leads to pitting on the cutting surface.
4. It is observed that high voltage increased slit width dimension. Because of metal
dissolution increases the slit width. Similarly, high flow rate removes burrs but
enlarges the slit width.
5. The optimum process parameters for efficient machining of 5% reinforced com-
posite were voltage (10.2 V), electrolyte concentration (12.8%), and electrolyte
flow rate (8 l/min). The optimum process parameters for efficient machining of
15% reinforced composite were 14 V, 12% electrolyte concentration, and 8 l/min
flow rate.
50 K. Rajkumar et al.

References

1. Hemanth J (2009) Quartz (SiO2 p) reinforced chilled metal matrix composite (CMMC) for
automotive applications. Mater Des 30(2):323–329
2. Shorowordi KM, Laoui T, Haseeb ASMA, Celis JP, Froyen L (2003) Microstructure and inter-
face characteristics of B4C, SiC and Al2 O3 reinforced Al matrix composites: a comparative
study. J Mater Process Technol 142(3):738–743
3. McGeough JA (1988) Advanced methods of machining. Springer Science & Business Media
(1988)
4. Sen M, Shan HS (2005) A review of electrochemical macro-to micro-hole drilling processes.
Int J Mach Tools Manuf 45(2):137–152
5. Sankar M, Gnanavelbabu A, Rajkumar K (2014) Effect of reinforcement particles on the abra-
sive assisted electrochemical machining of Aluminium-Boron carbide-Graphite composite.
Procedia Eng 97:381–389
6. Senthilkumar C, Ganesan G, Karthikeyan R (2009) Study of electrochemical machining
characteristics of Al/SiC p composites. Int J Adv Manufact Technol 43(3–4):256–263
7. El-Taweel TA, Gouda SA (2011) Performance analysis of wire electrochemical turning
process—RSM approach. Int J Adv Manufact Technol 53(1–4):181–190
8. Bhattacharyya B, Sorkhel SK (1999) Investigation for controlled electrochemical machining
through response surface methodology-based approach. J Mater Process Technol 86(1–3):200–
207
9. Bhattacharya B, Sorkhel SK (1997) Response surface methodology based analysis for achieving
controlled electrochemical machining. In: Proceedings of the 17th all India machine tool design
and research conference, pp 307–311
10. Sankar M, Gnanavelbabu A, Rajkumar K, Thushal NA (2017) Electrolytic concentration effect
on the abrasive assisted-electrochemical machining of an aluminum–boron carbide composite.
Mater Manuf Processes 32(6):687–692
Modeling and Parametric Optimization
of Process Parameters of Wire Electric
Discharge Machining on EN-31
by Response Surface Methodology

Sushant B. Patil, Swarup S. Deshmukh, Vijay S. Jadhav,

and Ramakant Shrivastava

Abstract Nowadays, the majority of industry uses nonconventional machines; wire

electric discharge machining is one of them. In this experimentation, optimize pro-
cess parameters of wire electric discharge machining with help of response surface
methodology. Central composite design is used for the design of experiments. The
process parameters considered for this study are a pulse on time, wire feed rate,
pulse off time, and servo voltage. For this experimentation work, EN-31 used as
workpiece material. The high percentage of carbon present in the material due to this
is used for manufacturing punches and dies. To find out significant factors, ANOVA
is calculated. Analysis of variance for MRR clearly shows that pulse on time and
servo voltages are the most significant parameters. From result analysis, the high
value of MRR is obtained at high value of pulse on time and low value of servo
voltage. In the case of surface roughness also pulse on time and servo voltage are the
most significant factors as compared to others; the low value of surface roughness is
obtained at a low value of pulse on time and high value of servo voltage.

Keywords WEDM · EN-31 · MRR · SR · ANOVA · RSM

S. B. Patil (B) · S. S. Deshmukh · V. S. Jadhav · R. Shrivastava

Department of Mechanical Engineering, Government College of Engineering, Karad, Maharashtra
415124, India
e-mail: patilsush96@gmail.com
S. S. Deshmukh
e-mail: deshmukh.swarup@gmail.com
V. S. Jadhav
e-mail: vijay.jadhav@gcekarad.ac.in
R. Shrivastava
e-mail: ramakant.shrivastava@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 51

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_6
52 S. B. Patil et al.

1 Introduction

In wire electric discharge machining electro-thermal mechanism used for cutting

workpiece as same as electric discharge machining. The wire of various materials
like brass, zinc-coated brass, diffused brass, molybdenum is used as an electrode
in case of wire electric discharge machine. The wire electrode continuously travels
along the cutting path generated by the numerical control program. In the wire electric
discharge machining workpiece acts as anode and wire electrode as cathode and the
proper gap is provided between them. Due to this, high-intensity spark is generated
between them; this spark is responsible for melting workpiece. In cutting region, the
temperature goes too high due to this high-intensity spark energy so deionized water
helps to reduce the temperature in that region and helps to remove debris particle
which is produced after cutting workpiece material. This is the non-contact type of
cutting process, and hence no stress is present in the workpiece after machining.
The complex, contour, and irregular shape profiles easily cut from the workpiece
this is one of the major advantages. Kansal et al. [1] optimized process parameters
of PMEDM using response surface methodology. In this experimentation, they used
silicon powder in a dielectric fluid. Central composite design with face-centered
used for the design of experiments. Finally, they concluded that after the addition
of silicon powder in dielectric fluid improvement in value of response variables
takes place. Kung and Chiang et al. [2] carried out parametric optimization of wire
electric discharge machining on aluminum oxide-based ceramic with help of response
surface methodology. After increasing the values of T on and duty factor, MRR and
SR increase up to a certain limit then after it decreases. Patel et al. [3] performed
parametric optimization of process parameters of EDM on Al2 O3 /Sic/Tic ceramic
composite. They finally concluded that pulse on time is a major significant factor as
compared to other factors for surface roughness. After increasing the value of the
duty cycle, surface roughness first increases then decreases. Patil et al. [4] carried out
an analysis of wire electric discharge machining on Al/SiCp MMC using response
surface methodology. The conclusion of that experimentation is that (T on ), (T off ), and
volume fraction (V%) of ceramic reinforcement are the most significant parameters
for cutting rate. For surface roughness (V%) of ceramic, reinforcement and (T on ) are
the most significant parameters. Ojha et al. [5] studied material removal rate and tool
wear rate as response variables after the addition of chromium powder in a dielectric
fluid. They get a high value of material removal rate at higher powder concentration
and tool wear rate is decreases after increasing value of tool diameter. Shandilya
et al. [6] carried out the optimization of process parameters of WEDM on MMC
by response surface methodology. Good surface finish and small size of the crater is
getting at a low level of process parameters. Khanna et al. [7] performed an analysis of
cryogenically treated workpiece on wire electric discharge machining by response
surface methodology. They performed single and multi-objective optimization to
achieve maximum value of MRR and the minimum value of surface roughness. The
error between actual and predicted values for MRR lies between −2 and +4% and for
SR lies between −7 and +7%. Gopalakannan et al. [8] used central composite design
Modeling and Parametric Optimization of Process Parameters … 53

and desirability approach for the optimization of process factors of EDM on MMC.
The low value of EWR is observed at high value of T off (pulse off time) and SR is
influenced by current and T on (pulse on time). Sharma et al. [9] performed single and
multi-objective optimization of process variables of WEDM using response surface
methodology (RSM). For this research work, they used HSLA (High Strength Low
Alloy) steel used as workpiece material and brass wire as an electrode. The result
of ANOVA shows that (CS), T on , and T off are the most significant parameters and
for (DD), (SV), and (T on ) are the most significant parameters. Singh et al. [10]
investigated surface characteristics after powder mixed electric discharge machining
and compare results of surface roughness and recast layer in between powder mixed
and without powder mixed electric discharge machining. The workpiece material
used for this study is aluminum 6061. Finally, they concluded that surface roughness
and recast layer is reduced in the case of powder electric discharge machining as
compared to normal electric discharge machining. Dubey et al. [11] studied the
material removal rate in powder mixed electric discharge machining on AA7075/B4 C
composite. In this experimentation for improving material removal rate chromium
powder is added in a dielectric fluid. From SEM images of the recast, layer shows
that recast layer thickness is more at high value of process variables as compared to
initial values of process variables.

2 Experimental Setup

This experimentation is performed on the ELECTRONICA-ECOCUT WEDM which

is shown in Fig. 1. This is five-axis machine, for normal cutting of workpiece nozzle
move in x- and y-direction and for taper cutting nozzle moves in U and V axis.
The Ra is measured with the help of MITUTOYO SJ-210 surface roughness tester
which is shown in Fig. 2. The EN-31 material is selected as workpiece material
for this study. Due to the high % of carbon majority application is in the punches
and dies industry. The chemical composition of EN-31 is shown in Table 1. Process

Fig. 1 Machine setup

(Electronica-Ecocut)
54 S. B. Patil et al.

Fig. 2 Surface roughness

tester (Mitutoyo SJ-210)

Table 1 Chemical composition of EN-31

Elements C% Mn% S% P% Cr%
Contents 0.90–1.20 0.30–0.75 0.050 Max 0.050 Max 1.00–1.60

parameters and their levels in terms of actual and coded value is shown in Table
2. The zinc (Zn) coated Brass wire of 0.25 mm which is used as an electrode in
wire electric discharge machining. So the weight (Wt.) of the workpiece is measured
with the help of PESCO weight balance machine having least count of 0.001 gm. The
PESCO weight balance machine is shown in Fig. 3. The equation used for calculation
of MRR which is given in Eqs. 1.

Table 2 Process Parameters and their levels in terms of coded and actual value
Factor symbol Parameter Unit Levels
Low (−1) High (+1)
A Pulse on time (T on ) µs 110 (−1) 120 (+1)
B Pulse off time (T off ) µs 45 (−1) 55 (+1)
C Servo Voltage (SV) V 20 (−1) 40 (+1)
D Wire Feed (WF) m/min 2 (−1) 6 (+1)

Fig. 3 PESCO weight

balance machine
Modeling and Parametric Optimization of Process Parameters … 55

Fig. 4 Machined specimen

Weight of workpiece before machining − Weight of workpiece after machining gm

MRR = (1)
Time min

Central composite design is used for design of experiments. Central composite

design is an experimental design in which factors value less or greater than factors
level are tested in a systematic way in order to better relate the factor to the response
in a precise way. In this experimentation full factorial design with all possible combi-
nations of process parameters considered at a low level and high level, in this design
eight axial (star) points at the face of cube and centre point is equal to six. The face-
centered central composite design is selected in which alpha value is equal to one. In
this study, total 30 experiments are conducted according to design matrix given by
central composite design. The 30 specimens cut according to design matrix which
is shown in Fig. 4. The design matrix is developed with the help of “Design Expert
11.0” which is shown in Table 3. The graphical analysis, i.e., (surface plots, contour
plots), analysis of variance (ANOVA), regression analysis is done with the help of
“Design Expert 11.0” software.

3 Results and Discussion

To find out significant factors from model and to check the lack of fit of model,
analysis of variance is calculated. Result of analysis of variance for MRR is shown
in Table 4.

3.1 Analysis of Material Removal Rate

According to the fit summary linear model is significant for analysis of material
removal rate. The value of R2 is 96.07% and the value of adjusted R2 is 95.44% this
clearly indicates that the regression model gives a good relationship between process
factors and the response variable. The factors havingP-value less than 5% (i.e., 0.05)
56 S. B. Patil et al.

Table 3 Design matrix and output response

Exp. Run Process parameters Response variable
A: T on B: T off C: SV D: WF MRR SR
µs µs Volt m/min gm/min Micrometer
1 110 45 20 2 0.220637 1.405
2 120 45 20 2 0.448101 2.870
3 110 55 20 2 0.227062 1.391
4 120 55 20 2 0.407225 3.345
5 110 45 40 2 0.157805 1.390
6 120 45 40 2 0.371077 1.578
7 110 55 40 2 0.14984 1.401
8 120 55 40 2 0.306302 1.490
9 110 45 20 6 0.22588 1.341
10 120 45 20 6 0.45933 3.488
11 110 55 20 6 0.210858 1.296
12 120 55 20 6 0.427424 3.417
13 110 45 40 6 0.16485 1.368
14 120 45 40 6 0.331006 1.620
15 110 55 40 6 0.159317 1.443
16 120 55 40 6 0.31628 1.599
17 110 50 30 4 0.202971 1.404
18 120 50 30 4 0.38827 2.367
19 115 45 30 4 0.286261 1.494
20 115 55 30 4 0.283548 1.519
21 115 50 20 4 0.331577 2.276
22 115 50 40 4 0.215709 1.387
23 115 50 30 2 0.266409 1.469
24 115 50 30 6 0.266356 1.535
25 115 50 30 4 0.254277 1.516
26 115 50 30 4 0.245908 1.492
27 115 50 30 4 0.294183 1.564
28 115 50 30 4 0.285822 1.475
29 115 50 30 4 0.283582 1.555
30 115 50 30 4 0.279677 1.399
Modeling and Parametric Optimization of Process Parameters … 57

Table 4 ANOVA for MRR

Source Sum of DF Mean square F-value p-value Check
squares significance
Model 0.2034 4 0.0509 152.83 <0.0001 Significant
A-T on 0.1674 1 0.1674 502.97 <0.0001 Significant
B-T off 0.0017 1 0.0017 5.24 0.0309
C-SV 0.0343 1 0.0343 103.11 <0.0001 Significant
D-Wire feed 2.601E−06 1 2.601E−06 0.0078 0.9303
Residual 0.0083 25 0.0003
Lack of fit 0.0065 20 0.0003 0.8748 0.6294 Not significant
Pure error 0.0018 5 0.0004
Cor. Total 0.2128 29
Std. Dev. 0.0182 R2 0.9607
Mean 0.2823 Adjusted R2 0.9544
C.V.% 6.46 Predicted R2 0.9415

these factors are most significant. The factor A-Pulse on time, B-Pulse off time, C-
servo voltage has a significant effect. Among them, A-pulse on time, C-servo voltage
these parameters are most significant for MRR. The value of lack of fit is also non-
significant this is desirable. Figure 5 shows that the normal probability plot for MRR
is clearly indicated that residuals are present on a straight line, it means that error
follows a normal distribution. Figure 6 shows the plot for predicted versus actual

Fig. 5 Normal probability

plot for MRR
58 S. B. Patil et al.

Fig. 6 Plot of actual versus

predicted response of MRR

value it shows that model is fitted well. After backward elimination, i.e., eliminates
non-significant terms equation of MRR in coded terms and actual terms is given in
Eqs. 2 and 3 respectively.

(In Coded terms) − MRR = 0.2823 + 0.0964 A − 0.0098B − 0.0437C (2)

(In actual factors) − MRR = −1.70634 + 0.019287 ∗ Ton

− 0.001968Toff − 0.004366 ∗ SV (3)

Figure 7 shows the surface plot for MRR in relation to the process parameters of
the pulse on time and servo voltage. From the surface plot, it is clearly observed that
as a pulse on time increases material removal rate increases, and as servo voltage
increases material removal rate decreases. Hence, the maximum value of the material
removal rate is obtained at high value of pulse on time (120 µs) and a low value of
servo voltage (20 V). The minimum value of the material removal rate is obtained
at a low value of pulse on time (110 µs) and a high value of servo voltage (40 V).
From Fig. 8, it is clearly indicated the effect of the pulse on time and pulse off time
on material removal rate. The material removal rate increases as the value of pulse
off time decreases. In this case, the high value of material removal rate is obtained
at high value of pulse on time (120 µs) and low value of pulse off time (45 µs) and
minimum value material removal is obtained at low value of pulse on time (110 µs)
and high value of pulse off time (55 µs).
Modeling and Parametric Optimization of Process Parameters … 59

Fig. 7 Effect of T on and SV on MRR

Fig. 8 Effect of T on and T off on MRR

60 S. B. Patil et al.

3.2 Analysis of Surface Roughness

In case of surface roughness, fit summary states that the quadratic model is significant.
Result of analysis of variance for surface roughness is shown in Table 5.
The value of R2 is 0.9858 and adjusted R2 is 0.9725 shows a good relationship
between process parameters and the response variable. The process parameters A-
pulse on time, C-servo voltage, and interaction effect of process parameter A with
process parameter C have a significant effect. The lack of fit is also non-significant
which is desirable. Figure 9 shows that normal probability plot for SR it clearly
indicated that residuals are present on a straight line it means that error follows a
normal distribution. Figure 10 shows the plot for predicted versus actual value it
shows that model is fitted well. After backward elimination, i.e., (eliminates non-
significant terms) equation of SR in coded terms and actual terms is given in Eqs. 4

Table 5 ANOVA for SR

Source Sum of squares DF Mean square F-value p-value Check
significance
Model 12.21 14 0.8723 74.34 <0.0001 Significant
A-T on 4.84 1 4.84 412.58 <0.0001 Significant
B-T off 0.0067 1 0.0067 0.5701 0.4619
C-SV 3.17 1 3.17 270.10 <0.0001 Significant
D-WF 0.0328 1 0.0328 2.79 0.1154
AB 0.0045 1 0.0045 0.3826 0.5455
AC 3.06 1 3.06 261.14 <0.0001 Significant
AD 0.0600 1 0.0600 5.12 0.0390
BC 0.0085 1 0.0085 0.7213 0.4091
BD 0.0124 1 0.0124 1.06 0.3196
CD 0.0081 1 0.0081 0.6903 0.4191
A2 0.2252 1 0.2252 19.19 0.0005
B2 0.0184 1 0.0184 1.56 0.2303
C2 0.1503 1 0.1503 12.81 0.0027
D2 0.0204 1 0.0204 1.74 0.2075
Residual 0.1760 15 0.0117
Lack of fit 0.1577 10 0.0158 4.32 0.0600 Not significant
Pure error 0.0183 5 0.0037
Cor. Total 12.39 29
Std. Dev. 0.1083 R2 0.9858
Mean 1.76 Adjusted R2 0.9725
C.V.% 6.14 Predicted R2 0.9042
Adeq. precision 27.7738
Modeling and Parametric Optimization of Process Parameters … 61

Fig. 9 Normal probability

plot for SR

Fig. 10 Plot of actual versus

Predicted response
62 S. B. Patil et al.

and 5 respectively.

(In Coded terms) − SR = 1.53 + 0.5186A − 0.4196C + 0.0427D

− 0.4376AC + 0.0612 AD + 0.2224 A2 + 0.1684C 2
(4)

(In actual factors) − SR = 82.54 − 1.70Ton + 0.86SV − 0.68WF − 0.0087Ton

∗ SV + 0.0061Ton ∗ WF + 0.0088Ton
2
+ 0.0016SV2
(5)

The surface plot in Fig. 11 clearly indicates the variation of surface roughness
with respect to process parameters such as pulse on time and servo voltage. The value
of surface roughness increases as the value of pulse on time increases similarly high
value of surface roughness is obtained at a low value of servo voltage. The maximum
value of surface roughness is obtained at high value of pulse on time (120 µs) and
low value of servo voltage (20 V) and a minimum value of surface roughness is
obtained at low value of pulse on time (110 µs) and high value of servo voltage (40
V). The effect of the pulse on time and wire feed on surface roughness is shown in
Fig. 12 with the help of the surface plot. Surface plot in Fig. 12 clearly indicates that
high value of surface roughness is obtained at a high value of pulse on time (120 µs)
and low value of wire feed rate (2 m/min) and low value of surface roughness is get
at low value of pulse on time (110 µs) and high value of feed rate (6 m/min).

Fig. 11 Effect of T on and SV on SR

Modeling and Parametric Optimization of Process Parameters … 63

Fig. 12 Effect of T on and WF on SR

3.3 Confirmation Tests and Their Comparison with Results

The confirmation test table clearly shows that error is very small lies between ±
4%. Result of confirmation test is shown in Table 6. Finally, confirmation test confirms
the good reproducibility of experimentation results.

4 Conclusions

Modeling and parametric optimization of process parameters of WEDM using

response surface methodology reveals the following conclusions by variation of
process parameters within the specified range. The pulse on time and servo volt-
age affects both material removal rate and surface roughness and pulse off time and
wires feed rate are least significant factors as compared to the pulse on time and
servo voltage. The material removal rate increases with increases in pulse on time
and decreases in the value of servo voltage. Hence, maximum value of material
removal rate is secured at pulse on time is equal to 120 µs and servo voltage is equal
to 20 V and minimum value of surface roughness is got at pulse on time is equal
to 110 µs and servo voltage is equal to 40 V. The ANOVA of surface roughness
after backward elimination clearly shows that pulse on time and servo voltages are
the most significant parameters for surface roughness. The surface plot for surface
 64

Table 6 Confirmation test

Exp. No. Process parameters MRR SR
T on T off SV WF Exp. Predicted Error Exp. Predicted Error
1 110 55 20 2 0.227012 0.219042 3.51 1.391 1.401 −0.71
2 120 50 30 4 0.388270 0.378685 2.46 2.367 2.269 4.14
3 115 50 30 4 0.245908 0.288225 −0.17 1.492 1.528 −2.41
S. B. Patil et al.
Modeling and Parametric Optimization of Process Parameters … 65

roughness clearly shows that surface roughness increases as the value of pulse on
time increases and the value of servo voltage decreases. The minimum value of sur-
face roughness is obtained at a low value of pulse on time and high value of servo
voltage, i.e., (T on = 110 µs and SV = 40 V). From the confirmation test the error
between actual and predicted values for material removal rate and surface roughness
lies within ± 4% range. This error is very small it confirms the good reproducibility
of experimental results.

References

1. Kansal HK, Singh S, Kumar P (2005) Parametric optimization of powder mixed electrical
discharge machining by response surface methodology. J Mater Proc Technol 169:427–436
2. Kung K-Y, Chiang K-T (2008) Modelling and analysis of machinability evaluation in the wire
electrical discharge machining (WEDM) process of aluminium oxide-based ceramic. Mater
Manufact Proc 23:241–250
3. Patel KM, Pandey PM, Venkateswara Rao P (2009) Determination of an parametric com-
bination using a surface roughness prediction model for EDM of Al2 O3 /SiCw /TiC ceramic
composite. Mater Manufact Proc 24:675–682
4. Patil NG, Brahmankar PK (2010) On the response surface modelling of wire electrical discharge
machining of Al/SiCp metal matrix composites (MMCs). J Mach Form Technol 2(1/2):47–70
5. Ojha K, Garg RK, Singh KK (2011) Experimental investigation and modeling of PMEDM
process with chromium powder suspended dielectric. Int J Appl Sci Eng 2:65–81
6. Shandilya P, Jain PK, Jain NK (2012) Parametric optimization during wire electrical discharge
machining using response surface methodology. Proced Eng 38:2371–2377
7. Khanna R, Singh H (2013) Performance analysis for D-3 material using response surface
methodology on WEDM. Int. J Mach Mach Mater 14(1):45–65
8. Gopalakannan S, Senthilvelan T (2014) Optimization of machining parameters for EDM oper-
ations based on central composite design and desirability approach. J Mechan Sci Technol
28(3):1045–1053
9. Sharma N, Khanna R, Gupta RD, Sharma R (2013) Modeling and multiresponse optimization
on WEDM for HSLA by RSM. Int J Adv Manuf Technol 67:2269–2281
10. Singh B, Kumar J, Kumar S (2014) Experimental investigation on surface characteristics in
powder-mixed electrodischarge machining of AA6061/10%SiC composite. Mater Manufact
Proc 29:287–297
11. Dubey V, Singh B (2018) Study of material removal rate in powder mixed EDM of AA7075/B4C
composite in IMME17. Mater Today: Proc 5:7466–7475
Wire-Cut Electric Discharge Machining
on Nickel–Aluminium–Bronze Using
Brass Wire Electrode

Earnest Beni, Poovazhagan Lakshmanan, and S. C. Amith

Abstract Wire-cut electric discharge machining (WEDM) is a widely used non-

traditional subtractive machining process for removing materials from conductive
workpiece to produce parts with intricate shapes/complex profiles. Nickel–alu-
minium–bronze (NAB alloy) is a high-strength, difficult-to-machine nickel-based
alloy. In this work, a trial was conducted to machine NAB alloy using WEDM tech-
nique. The machining voltage, pulse-on time, pulse-off time, and wire feed were
varied for the experiments. The 10 × 10 mm square-shaped samples were cut on
NAB alloy with 0.25-mm-diameter brass wire electrode. Based on the analysis, it
is found that pulse-on time, machining voltage, and wire tension are the significant
parameters that affect the kerf width, material removal rate, and surface roughness.
Scanning electron microscopy (SEM) is used to identify the microstructure and the
surface morphology of the machined work piece. A higher pulse-on time setting
leads to thicker recast layer. At lower value of pulse-on time and higher value of
pulse-off time, the wire deposition on the machined surface is low.

Keywords Brass wire · NAB alloy · Wire-cut EDM · Taguchi’s method · DOE

1 Introduction

Nickel–aluminium–bronze (NAB) alloy is the most widely used group of nickel–

aluminium alloy. NAB alloys can be obtained in both cast or wrought forms. They
possess the following unique combination of properties: high strength, high corrosion
resistance, good wear resistance, good cryogenic properties, and excellent damping
capacity. The commercial uses of NAB alloy range from propellers for ships, landing
gear bushings, bearings for aircraft and sea water pumps, and non-sparking tools in
oil and gas industries [1]. Generally, nickel-based alloys are difficult to machine by
conventional machining processes [2].

E. Beni · P. Lakshmanan (B) · S. C. Amith

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai 603110, Tamil Nadu, India
e-mail: poovazhaganl@ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 67

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_7
68 E. Beni et al.

Electrical discharge machining (EDM) is one of the well-established non-

traditional machining processes intended to machine geometrically complex shapes
and extremely hard conductive materials. EDM is the non-contact-type machining
process, which leads to the following benefits: eliminating vibration, chatter, and
mechanical stresses [3]. EDM is the indispensable machining technology to various
sectors of industries such as automobile, aeronautical, and medical due to its ability
to cut very hard materials [4]. Wire-cut EDM (WEDM) is introduced in the late
1960s and has revolutionized the metal working, tool and die, and mould industries.
Mechanism of material removal in WEDM is based on the thermo-electric princi-
ples. Melting point of the work material is the only property to be considered during
WEDM [5].
Brajesh Kumar Lodhi et al. (2014) used Taguchi L9 orthogonal array to opti-
mize the WEDM parameters for better surface roughness during the machining of
AISI D3 steel. It was concluded that discharge current was the most influencing
machining parameter for surface roughness [6]. Bijaya Bijeta Nayak et al. (2016)
optimized WEDM process parameters while machining deep cryo-treated inconel
718 using hybrid approach [7]. Ashish Goyal (2017) investigated material removal
rate (MRR) and surface roughness of inconel 625 super alloy using cryogenic-treated
zinc-coated wire. ANOVA optimization revealed that current intensity and tool elec-
trodes are dominant parameters that affect the MRR and surface roughness [8].
Amitesh Goswami et al. (2017) optimized the trim-cut WEDM machining param-
eters while machining nimonic alloy 80A. Results show that trim-cut WEDM pro-
duces very good surface finish as compared to the rough-cut WEDM [9]. Coral
reef microstructures with few voids were observed while machining inconel 718 in
WEDM using uncoated brass wire electrode of diameter 200 µm [10]. Inconel 625
materials were machined on a wire-cute EDM by using brass as an electrode; influ-
ence of MRR, SR, discharge current, voltage, and pulse-on time were optimized by
using Taguchi method [11]. The copper electrode with zinc coating and uncoated
brass electrode wire were used for wire EDM machining of IN 718; coated wire
showed better performance like reduction in wire consumption when compared to
uncoated wire electrode [12].
It is apparent from the detailed literature review that no experimental work has
been performed on the machinability, kerf width (KW), and surface roughness (SR)
study of NAB alloy using brass wire in WEDM process. Hence, in this current
research work attempt has been made to examine the machining characteristics (KW,
MRR, and SR) of NAB alloy using untreated brass wire in WEDM.
Wire-Cut Electric Discharge Machining … 69

Table 1 Properties of workpiece material

Material Chemical Hardness (HV) Thickness (mm) Thermal
composition conductivity
(W/m0 K)
NAB alloy Cu-balance, 200 5.5 mm 42
Al-10%, Ni-5%,
Fe-4%, Mn-1.5%

2 Experimentation

2.1 Materials and Methods

Material employed for experimentation is nickel–aluminium–bronze (NAB) alloy

with brass wire as electrode for WEDM process. NAB is a high-strength Cu-Al alloy
with the applications involving heavy loads, abrasive wear resistance, friction, and
corrosion. The addition of nickel increases strength of the alloy without deterio-
rating its outstanding ductility, toughness, and corrosion resistance. The chemical
composition and properties of NAB alloy are given in Table 1.
The wire-cut EDM machine used in this study was Sodick model A280 with the
wire electrode made from brass (Cu—35 wt%Zn). The wire was 0.25 mm in diameter
and can tolerate a tension up to 95 kgf/mm2 . Elektra Maxicut-made CNC-controlled
wire-cut electrical discharge machine is used to conduct wire-cut EDM experiments.
It consists of CNC motion control unit, fixture for holding the workpiece and wire
spool. NAB base plate was of 250 × 150 × 6 mm dimensions that was used for this
study purpose.

2.2 Design of Experiment Based on Taguchi Method

A total of 9 experiments were conducted using Taguchi L9 experimental design

methodology. In our study, the machining parameters were taken as pulse-on time,
pulse-off time, peak current, and wire tension. There are 3 levels of factors in each
process parameters as shown in Table 2. Using these factors and levels, L9 experi-
ments were selected which are shown in Table 3. The responses that are taken into
account are the kerf width, material removal rate, and the surface roughness.
The sample was cut by using brass wire in the wire-cut EDM machine of 10 ×
10 × 6 mm dimension for each set of parameters based on the L9 orthogonal array.
In each experiment, the time taken for machining was noted. After machining the
dimension of the sample and the inner dimension of the base plate from which the
sample was cut are measured for calculating the kerf width and material removal
rate. The EDMed samples were then cleaned using ethanol before SEM analysis.
70 E. Beni et al.

Table 2 Levels of various process parameters

Parameters Designation Levels
L1 L2 L3
Pulse ON (µs)—Ton A 1 3 6
Pulse OFF (µs)—Toff B 4 7 10
Current (A)—I C 2 3 5
Wire tension (N)—WT D 0.98 1.47 2.45

Table 3 Experimental design—L9 orthogonal array (tool: brass wire, workpiece: NAB alloy)
Experiment A B C D Kerf width (mm) MRR (mm3 /min) SR (µm)
number
1 1 4 2 0.98 0.215 2.95 2.245
2 1 7 3 1.47 0.195 3.29 1.958
3 1 10 5 2.45 0.31 6.82 2.256
4 3 4 3 2.45 0.31 5.67 2.289
5 3 7 5 0.98 0.27 6.59 2.494
6 3 10 2 1.47 0.295 4.98 2.175
7 6 4 5 1.47 0.28 6.83 2.277
8 6 7 2 2.45 0.29 4.54 2.587
9 6 10 3 0.98 0.325 5.40 2.556

3 Results and Discussion

3.1 SEM Analysis of WEDM Samples

The SEM microstructures reveal the surface quality for each sample under different
parameters. A smooth surface is obtained in Fig. 1a for the parameters as pulse-on
time: 1 µs, pulse-off time: 4 µs, current: 2 A, and wire tension: 0.98 N. Figure 1c
shows the formation of a rough surface obtained at higher setting parameters as
pulse-on time: 6 µs, pulse-off time: 10 µs, current: 5 A, and wire tension 2.45 N.
The smoother surface as seen in Fig. 1a may be due to the lower intensity discharges
that took place during machining; thus, there is very small debris that can be observed
in this surface. Figure 1c shows the lumps and globules of debris with some deep
craters and microcracks on the surface which may have caused due to higher values
machining parameters. These deeper craters may be the places where the individual
discharge was able to penetrate far into the workpiece.
The recast layers can be clearly seen thicker in Fig. 1d compared to Fig. 1b, as
Experiment No. 1 (pulse-on time: 1 µs, pulse-off time: 4µs, current: 2A, and wire
tension: 0.98 N) was used to machine the sample shown in Fig. 1b and Experiment No.
9 (pulse-on time: 6 µs, pulse-off time: 10 µs, current: 5 A, and wire tension 2.45 N)
Wire-Cut Electric Discharge Machining … 71

a) b)

c) d)

Fig. 1 SEM micrographs of a sample 1 WEDM surface, b sample 1 cross section, c sample 9
WEDM surface, and d sample 9 cross section

was used to machine the sample shown in Fig. 1d. This difference in the recast layer
thickness is due to the resolidified materials formed on the surface layer after the
spark melts a small portion of workpiece and wire electrode during machining.

3.2 Effect of Process Parameters on MRR, Kerf Width,

and Surface Roughness

The higher the MRR, better the machining. Figure 2 shows the effect of MRR for
the various machining parameters. When the pulse-on time increases the MRR also
increases; it is similar to the peak current and wire tension. High current generates
large number of electrons. These electrons striking the work surface in a single
discharge, which increases the erosion process. Therefore kerf width and MRR also
increased. This can be confirmed by Figs. 3 and 2 respectively. Similarly, the surface
72 E. Beni et al.

Fig. 2 Effect plot for material removal rate

Fig. 3 Effect plot for kerf width

roughness also increases when the peak current is high which results in high discharge
frequency, leading to the formation of deeper and larger craters in the surface. Table 4
shows the interaction between various process parameters and their significance
based on the rankings, with percentage contribution for the peak current to be the
highest at 87.2% and the wire tension and the pulse-off time percentage contribution
is the least at 31% and 32.3%, respectively. The pulse-on time is also a significant
factor with percentage contribution of 53.8%.

Table 4 Ranking means for brass wire

Level Pulse on Pulse off Current WT
1 2.249 2.563 2.253 2.561
2 2.786 2.468 2.444 2.476
3 2.787 2.791 3.125 2.786
Delta 0.538 0.323 0.872 0.310
Rank 2 3 1 4
Wire-Cut Electric Discharge Machining … 73

Fig. 4 Effect plot for surface roughness

In all the graphs, the pulse-off time also showed some increase in MRR and kerf
width. This particular phenomenon requires supporting reasons to clarify; thus, the
pulse-off effect is not considered as a major significance causing parameter.
The effect of various parameters like pulse-on time, pulse-off time, peak current,
and the wire tension has been studied through machining of NAB with using brass
electrode wire. The surface roughness was unchanged by using different parameters
as shown in Fig. 3. The SR had been only influenced by the pulse-on time. Wire
tension has the least effect on the surface roughness. This is because the number of
electrons hitting the surface for elapsed time causes more erosion of material, and
thus, surface roughness is increased (Fig. 4).

4 Conclusions

In this study, the effect of machining parameters on the kerf width, MRR, and the
surface roughness of nickel–aluminium–bronze was investigated experimentally by
using brass wire-cut EDM.
1. Peak current and pulse-on time have been found as the most significant factors,
with percentage contribution of 87.2% and 53,8%, respectively. The other two
process parameter interactions were also significant; however, their contribution
in variance is relatively low.
2. The higher values of discharge energy and the elongated period of pulse-on time
increase the machining rate and thus result in large craters and microdamages
that cause the surface to be rough. This problem can be optimized by Taguchi’s
parameter design for WEDM.
3. The effect of various parameters like pulse-on time, pulse-off time, peak current
and the wire tension has been studied through machining of NAB with using
brass electrode wire. The surface roughness was unchanged by using different
parameters. SR had been only influenced by the pulse-on time.
74 E. Beni et al.

4. A linear regression graph is obtained for the MRR, kerf width, and the surface
roughness responses. The result shows that the MRR and kerf width can be highly
influenced by the variation in peak current and the pulse-on time.
5. Based on the results, the untreated brass wire provides good machining perfor-
mance (maximum MRR and better surface roughness) when used for machining
NAB material.
6. In the future, this work can be extended to discuss effect of the cryogenic cooled
electrodes of different materials with NAB and comparison of the same by
discussing the dimensional accuracy and MRR.

References

1. Lia L, Guob YB, Weia XT, Lib W (2013) Surface integrity characteristics in wire-EDM of
inconel 718 at different discharge energy. In: Seventeenth international symposium on electro
machining, Procedia CIRP, vol 6, pp 220–225
2. Amitesh G, Jatinder K (2014) Investigation of surface integrity, material removal rate and wire
wear ratio for WEDM of Nimonic 80A alloy using GRA and Taguchi method. Eng Sci Technol
Int J 17:173–184
3. Kanlayasiri K, Boonmung S (2007) An investigation on effects of wire-EDM machining
parameters on surface roughness of newly developed DC53 die steel. J Mater Process Technol
187–188:26–29
4. Ivan R, Carol P (2016) Guide to nickel aluminium bronze for engineers. Copper Development
Association Publication, No. 22
5. Ho KH, Newman ST (2003) State of the art electrical discharge machining (EDM). Int J Mach
Tools Manufact 43:1287–1300
6. Brajesh KL, Sanjay A (2014) Optimization of machining parameters in WEDM of AISI D3
Steel using Taguchi Technique. In: 6th CIRP international conference on high performance
cutting, HPC
7. Bijaya BN, Siba SM (2016) Optimization of WEDM process parameters using deep cryo-
treated Inconel 718 as work material. Eng Sci Technol Int J 19:161–170
8. Ashish G (2009) Investigation of material removal rate and surface roughness during wire
electrical discharge machining (WEDM) of Inconel 625 super alloy by cryogenic treated tool
electrode. J King Saud Univ Sci 29:528–535
9. Amitesh G, Jatinder K (2017) Trim cut machining and surface integrity analysis of Nimonic
80A alloy using wire cut EDM. Eng Sci Technol Int J 20:175–186
10. Norliana MA, Darius GS, Fuad BMd (2007) A review on current research trends in electrical
discharge machining (EDM). Int J Mach Tools Manuf 47:1214–1228
11. Subrahmanyam M, Nancharaiah T (2019) Optimization of process parameters in wire-cut EDM
of Inconel 625 using Taguchi’s approach. Mater Today: Proc
12. Reolon LW, Henning Laurindo CA, Torres RD, Amorim FL (2019) WEDM performance and
surface integrity of Inconel alloy IN718 with coated and uncoated wires. Int J Adv Manufact
Technol 100:1981–1991
Development and Characterization
of Metal Matrix Composite
Manufacture, Mechanical Properties
and Microstructural Characterization
of Aluminium and Iron Metal Matrix
Composite Manufactured

D. Arthur Jebastine Sunderraj , D. Ananthapadmanaban ,

K. Arun Vasantha Geethan , and A. John Rajan

Abstract The objective of the present work is to investigate the formation of inter-
metallic compounds such as iron aluminides and oxides during the formation of an
alloy by melting commercially pure aluminium (Al) billets and adding iron powder
(Fe) at 750 °C in the molten aluminium. The exothermic reaction during the addition
of ‘Fe’ in the molten ‘Al’ resulted in the formation of more amounts of intermetallic
compounds because of the rise in temperature. The composition of the ‘Al’ and ‘Fe’
powder was varied to obtain various alloys. The optical microscopy examination
revealed the presence of various intermetallic compounds and oxides in the matrix.
The Rockwell hardness survey performed on various samples exhibited a maximum
hardness value of 79 in ‘B’ scale for the alloy composition of 60% Al and 40% Fe. A
maximum shear strength of 288 Mpa was obtained for the samples containing 65%
Al and 35% Fe.

Keywords Intermetallics · Iron aluminide · Microstructure

1 Introduction

Aluminium is a chemical element in the boron group with symbol ‘Al’ and atomic
number 13. It is silvery-white, and it is not soluble in water under normal circum-
stances. Pure aluminium possesses low strength and toughness, and its strength can be
increased with the addition of various elements. Aluminium is alloyed with elements

D. Arthur Jebastine Sunderraj (B) · K. Arun Vasantha Geethan

Department of Mechanical Engineering College, St. Joseph’s Institute of Technology, Chennai
600119, India
e-mail: arthurjebastine@gmail.com
D. Ananthapadmanaban
Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, Tamil Nadu 603110, India
A. John Rajan
Department of Manufacturing Engineering, School of Mechanical Engineering VIT, Vellore,
Tamil Nadu, India

© Springer Nature Singapore Pte Ltd. 2021 77

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_8
78 D. Arthur Jebastine Sunderraj et al.

Fig. 1 Fe-Al phase diagram (Source ASM Hand Book)

like ‘Cu’, ‘Si’, ‘Mg’, ‘Zn’, ‘Fe’ and other elements. The solubility of aluminium in
iron at 400 °C is 11.5% which is as shown in Fig. 1. Addition of excess ‘Al’ in
‘Fe’ results in the formation of intermetallic compounds such as FeAl, Fe3 Al, FeAl2
and FeAl3 . The presence of intermetallic compounds and oxides reinforces the softer
matrix aluminium and increases the mechanical properties. These intermetallic com-
pounds are formed as a result of the reaction between two or more metals which are
mixed together in certain proportions. These intermetallic compounds exhibit neither
metallic nor ceramic properties. Iron aluminides are compounds formed between
‘Fe’ and ‘Al’. The phase diagram drawn between Al–Fe reveals the formation of
intermetallic compounds like Fe3 Al, FeAl, FeAl2 , Fe2 Al5 and FeAl3 . The proper-
ties of iron aluminide are dependent on number of aluminium atoms present in the
compound. The presence of more aluminium atoms in the intermetallic compounds
such as FeAl2 , Fe2 Al5 and FeAl3 is detrimental to strength properties and induces
brittleness. On the other hand, Fe3 Al and FeAl, which have a high iron composi-
tion, are used as structural materials because of their good wear resistance, corrosion
resistance and specific strength properties.
In pure elements and solid solutions, the atoms are bound together with metallic
bonds. In the case of intermetallic compounds covalent bond is formed which alters
the crystal structure, chemical, mechanical and electrical properties. Intermetallic
compounds such as Fe3 Al and FeAl are among a variety of different intermetallic
compounds that can be used in heterogeneous metal (Al)–intermetallic materials and
Manufacture, Mechanical Properties and Microstructural … 79

intermetallics that can be used for corrosion resistance applications at higher tem-
perature. This alloy system containing Al–intermetallic compounds having nearly
homogeneous composition exhibits properties varying from intermetallics to alu-
minium alloys. The literature revealed that the formation of Fe–Al intermetallic
compounds took place during melting of these elements and during powder com-
pacting and sintering of these powders. Wangyu Hu et al. [1] revealed the formation
of nanocrystalline iron aluminide intermetallic compounds by mechanical alloying,
sintering and high-velocity oxy-fuel (HVOF) thermal spraying from the elemental
powders. The strength, ductility and toughness of the samples were increased by
maintaining grain size to the nanocrystalline level. Jiqiang et al. [2] produced an
ultra-fine grained Fe-40Al intermetallic compound. The X-ray diffraction of the Fe-
40Al intermetallic compound revealed the presence of FeAl phase. The Fe-40Al
intermetallic compound presented ultra-fine grains size in the range of 100–600 nm
because of high nucleation rate and the low growth rate.
The fabrication process of the metal–intermetallic materials is one of the most
important areas of research. Several research works presented the capabilities of the
formation of intermetallic compounds, metal–intermetallic compound and metal–
ceramic combination. These include solidification processing, powder metallurgy
(PM), coating technology and layered manufacturing. Among these methods, PM
is the most commonly used because of its ability to produce a wide range of com-
positions, its phase and property control as well as shape-forming capability. In the
present research work, the Al-FeAl intermetallic compound’s fabrication route con-
sists of melting the solid aluminium billets and keeping ‘Fe’ powder in solid state. In
the first method, ‘Al’ solid billets and ‘Fe’ powder are mixed together and then heated
to 750 °C. At this temperature, the molten aluminium reacted with ‘Fe’ powder to
form Al-FeAl rich phase. Solubility of Fe in Al is 10–12%. Since we are adding 20–
40% Fe, it can be seen from the phase diagram that there is a very good possibility
of formation of intermetallics.
This alloy is then heat-treated for 10 h in a muffle furnace. In the second method,
‘Fe’ powder is added with the molten aluminium followed by annealing the alloy
for 10 h at 600 °C to form Al-FeAl system. During the solidification process, oxides
such as Al2 O3 and FeO2 are formed which may have strengthened the Al-FeAl alloy.
Prichard [3] conducted experiments by adding silicon to Fe-15%A1 alloy which
produced faceting on powder particle surfaces during vacuum sintering above
1200 °C. The faceting planes occurred on the low index planes with high sym-
metry. The faceting was due to a change in the solid–vapour surface energy as a
function of crystallographic orientation. The extrusion of Fe–Al–Si alloys produced
a wire texture, which nearly doubled the yield strength of the material in the extru-
sion direction compared to the transverse direction. Although the extruded alloy
was fully dense, the tensile properties in the extruded direction do not represent the
mechanical properties of a sintered microstructure with randomly oriented grains.
The yield strength of sintered powders was comparable to the extruded powders in
the same heat treatment condition. The yield strength and fracture energy of the fine
grain material with approximately 10% porosity exceeded the yield strength and
fracture energy of coarse grain material with approximately 3% porosity. Therefore,
80 D. Arthur Jebastine Sunderraj et al.

the processing parameters are broad enough to develop desirable mechanical prop-
erties with a wide range of microstructures. The next phase of development would
be to incorporate these processing parameters into an MIM process to address the
interaction of the powder with binder systems. The uniform porosity and fine grain
size of the sintered Fe-15A1-2.8Si alloy, which was heat-treated at 600 °C, developed
the best combination of strength and toughness. This indicates the silicon atoms are
more beneficial in a solid solution rather than in an ordered precipitate.
The objective of this work is to produce Al–Fe alloys possessing optimum mechan-
ical properties. Addition of ‘Fe’ in ‘Al’ to very high level decreases strength because
more amounts of brittle phase is formed in the alloy. Therefore, an optimum level
of ‘Fe’ percentage should be fixed for obtaining maximum strength. In the present
work ‘Fe’ content is varied from 20–40% and ‘Al’ content 60–80% to produce various
alloys. In the molten aluminium, ‘Fe’ powder is added which produces an exothermic
reaction. This reaction generates heat, and this heat partially melts the ‘Fe’ powder
to form intermetallic compounds with ‘Al’.

2 Experimental Procedure

2.1 Material

The iron powder 4–8 µm particle size having purity 99.9% is used in this study.
Wrought aluminium having 99.5% purity with 0.2% Si, 0.1% Fe, 0.04% Mg, 0.1%
Mn and 0.05% O is used. The purchased rod is of 20 mm in diameter which is cut to
billets of 20 mm in length.

2.2 Experimental Set-up

The muffle furnace having a heating chamber of 380 mm length and 180 mm width
is used. The heating rate of this furnace can be varied to the required level. The
maximum temperature of the furnace is 1000 °C. Small crucibles made with clay are
used for melting the aluminium metal. The experimental set-up is shown in Fig. 2.

2.3 Experimental Procedure

The billets were cleaned mechanically by turning them in a lathe and finally cleaned
with chemicals like HCl and NaOH. The ‘Al’ billets were weighed and placed inside
the crucibles and heated inside the furnace. The crucibles were heated up to 750 °C
and kept inside the furnace for 1 h. In the molten ‘Al’, ‘Fe’ powder is added to initiate
Manufacture, Mechanical Properties and Microstructural … 81

Fig. 2 Experimental set-up of a muffle furnace

exothermic reaction. Because of this exothermic reaction, the temperature inside the
crucible rises to 1600 °C. Further, this alloy is kept at 750 °C for two hours and given
an annealing heat treatment for 16 h at 550 °C.
The specimens obtained from the furnace were machined in a lathe to obtain
circular discs as shown in Fig. 3a, b. Square and circular specimens of 6 × 6 mm and
4 mm in diameter (for metallography and shear test) were cut using wire-cut electric
discharge machining technique. The WEDM specimens are shown in Fig. 3c. The
metallographic samples were polished using various grades (up to 1200 grade) of

a b

Fig. 3 a and b Al–Fe partial melt turned samples and c wire-cut EDM samples
82 D. Arthur Jebastine Sunderraj et al.

silicon carbide paper to obtain mirror finish. These samples were finished with Al2
O3 slurry to obtain mirror finish without any scratch marks. The polished specimens
were etched with a chemical mixture containing 5 ml HCl + 10 ml HNO3 + 3 ml
HF + 50 ml H2 O for 1 min. The specimens were rinsed in water and dried using a
drier.

3 Result and Discussion

3.1 Optical Microscopy

The optical micrographs shown in Fig. 4a, b correspond to the sample1 processed
at 750 °C and annealed at 550 °C for 19 h inside the furnace. Figure 4a shows
three different regions marked as region ‘A’, region ‘B’ and region ‘C’. These three
regions are rich with intermetallic compounds such as FeAl2 , Fe2 Al5 and FeAl3
and oxides like Al2 O3 and FeO2 . Further, away from these dark regions a yellowish
region with fine grains of ‘Al’ is noted. This region is rich with pure aluminium.

Fig. 4 a and b Optical micrographs of sample 1, d and e optical micrographs of sample 2

Manufacture, Mechanical Properties and Microstructural … 83

Table 1 Annealing temperature for various compositions

S. No Composition in wt. Melting temperature Annealing temperature
percentage 600 °C and time in h
1 80% Al and 20% Fe 750 °C and soaking at this 19
temperature for 2 h
2 80% Al and 20% Fe 750 °C and soaking at this 16
3 75% Al and 25% Fe temperature for 2 h and 16
mechanically agitated
4 65% Al and 35% Fe 16
5 70% Al and 30% Fe 16
6 60% Al and 40% Fe 16

Adjacent to this region on both sides, fine intermetallic compounds which impart
dispersion strengthening effect to this soft region are seen. The micrograph shown
in Fig. 4b reveals three distinct regions containing recrystallized regions rich with
alternate grains of aluminium and intermetallic compounds. The grain boundaries
are precipitated with ‘Fe’. The recrystallization is because of pinning of these ‘Fe’
atoms along the grain boundaries and migration of ‘Fe’ atoms into ‘Al’ matrix and
subsequent formation of newer grain boundaries. In Fig. 4c, a dark region which
resembles like a large spear at the centre of the micrograph with serrations at the
sides is observed and indicates the formation of intermetallic compounds and its
growth into the ‘Al’ grains because of long annealing time and diffusion of ‘Fe’
atoms into the ‘Al’ matrix.
The micrographs shown in Fig. 4a, b correspond to the sample 2 as given in
Table 1. This sample is rich with 20% ‘Fe’ and annealed for 16 h at 550 °C which is
mechanically agitated to obtain homogeneous alloy. The microstructures show dark
regions, which may possibly be intermetallics.
The optical micrographs shown in Fig. 5a, b correspond to the sample 3 processed
at 750 °C and annealed at 550 °C for 16 h inside the furnace. Figure 5a shows two
different regions marked as region A and region B. These three regions are rich with
intermetallic compounds such as FeAl2 , Fe2 Al5 and FeAl3 and oxides like Al2 O3
and FeO2 . Further, proof of intermetallics formation can be obtained by viewing the
XRD patterns as shown in Figs. 9, 10 and 11. The micrographs shown in Fig. 5d–f
correspond to the sample 4 as given in Table 1. This sample is rich with 35% ‘Fe’ and
annealed for 16 h at 550 °C which is mechanically agitated to obtain homogeneous
alloy.
Because of this, the micrographs shown in Fig. 5a, b reveal the presence of
fragmented intermetallic compounds such as FeAl2 , Fe2 Al5 and FeAl3 .
In Fig. 5d, the thick grain boundaries are visible and adjacent to the grain bound-
aries fingerprint-like intermetallic compound formation is noted which resembles
like laves phase (circle mark A). The lengthy intermetallic compounds rich with ‘Fe’
are observed in the published literature by Chakraborty et al. [4]. The circle marks
containing a sphere are possibly intermetallic compounds. The micrographs shown
84 D. Arthur Jebastine Sunderraj et al.

Fig. 5 a and b Optical micrographs of sample 3, d and e optical micrographs of sample 4

in Fig. 5a, b reveal the formation of long-thick-intermetallic compounds rich with

FeAl3 and FeAl6 which is in accordance with the published work by Taylor [5].
The optical micrographs shown in Fig. 6a, b correspond to the sample5 processed
at 750 °C and annealed at 550 °C for 16 h inside the furnace. Figure 6a shows two
different regions marked as region ‘A’ and region ‘B’. These two regions are rich with
intermetallic compounds such as FeAl2 , Fe2 Al5 and FeAl3 and oxides like Al2 O3
and FeO2 . Further, away from these dark regions yellowish bands run across the
micrograph rich with fine grains of ‘Al’ which is noted. This region is rich with pure
aluminium.
Kraishat et al. [6] observed fragmented and lengthy intermetallic compounds
formed in the published literature.

3.2 Rockwell Hardness

The Rockwell hardness values measured across the sample at twelve points were
recorded. The graphical plot of the hardness values is shown in Fig. 7. The hardness
values were on the rise as the composition of ‘Fe’ was on the rise. A maximum
Manufacture, Mechanical Properties and Microstructural … 85

Fig. 6 a and b Optical micrographs of sample 5, d and e optical micrographs of sample 6

Fig. 7 Rockwell hardness values of various samples

86 D. Arthur Jebastine Sunderraj et al.

hardness of 79 in ‘B’ scale was noted for the sample 6. The hardness value was in
the range of 52–78 for the sample 5. A minimum value of 41 in B scale was recorded
for the sample 1. This is due to less amount of ‘Fe’ present in the alloy.

3.3 Shear Strength

The average shear strength values are shown in Table 2. The average shear strengths
of the sample 1 processed at 750 °C and annealed for 19 h was 278 MPa. The average
shear strength of the sample 2 was only 265 MPa. The lowest shear strength was
recorded for the sample 3 (Table 2) which was 250 MPa. The shear strength of
the sample 4 was 280 MPa. For the sample 5, the shear strength was 272 MPa. The
sample 6 exhibited a shear strength of 279 MPa. The shear strength of various samples
is presented as a graph in Fig. 8. The variation in shear strength was attributed to
variation in the composition of ‘Fe’ and difference in annealing time and stirring of
the alloy during the formation of alloy.

Table 2 Shear strength

S. No Composition in wt. Shear strength values in
values of various composition
percentage N/mm2
1 80% Al and 20% Fe 278
2 80% Al and 20% Fe 265
3 75% Al and 25% Fe 250
4 65% Al and 35% Fe 280
5 70% Al and 30% Fe 272
6 60% Al and 40% Fe 279

Fig. 8 Graph showing the shear strength of various samples

Manufacture, Mechanical Properties and Microstructural … 87

3.4 X-Ray Diffraction

The XRD report of the sample 2 as given in Table 1 was shown in Fig. 9 and has
indicated the presence of several intermetallic compounds in the FeAl matrix. The
XRD pattern for the sample 2 is shown in Fig. 9. These compounds are hard and
brittle, which increase the shear strength of the annealed samples. The intermetallic
compounds such as Fe2 Al, Fe3 Al, FeAl, FeAl5 and Fe2 Al3 are present in the alloy.
The secondary phase particles like Al2 O3 and FeO2 are present. These intermetallic
compounds and secondary phase particles strengthened the soft aluminium matrix.
The XRD pattern reveals the presence of pure Al and Fe along with other compounds.
This XRD report of the sample 4 as given in Table 1 is shown in Fig. 10. The
intermetallic compounds such as Fe2 Al, Fe3 Al, FeAl, FeAl5 and Fe2 Al3 are present
in the alloy and indicated the presence of several intermetallic compounds in the
FeAl matrix. It increases the shear strength of the annealed samples. The secondary
phase particles like Al2 O3 and FeO2 were present. These intermetallic compounds
and secondary phase particles strengthened the soft aluminium matrix and make the
increase in strength of the original aluminium.
The XRD report reveals the composition of the various intermetallic compounds
that were present in Fig. 11 of sample 6 and has indicated the presence of several
intermetallic compounds in the FeAl matrix; this intermetallic compounds increase
the shear strength of the annealed samples. The various compositions of alloy are
Fe2 Al, Fe3 Al, FeAl, FeAl5 and Fe2 Al3 which are present in the alloy, and also, the
particles like Al2 O3 and FeO2 are present in the secondary phase particle. These inter-
metallic compounds and secondary phase particles strengthened the soft aluminium
matrix and give more strength to the pure aluminium.

Fig. 9 XRD image of sample 2

88 D. Arthur Jebastine Sunderraj et al.

Fig. 10 XRD image of sample 4

Fig. 11 XRD image of sample 6

4 Conclusions

• The optical microscopy examination revealed the presence of intermetallic com-

pounds in all the alloys. The amount of intermetallic compounds formed was on
the rise for the samples containing more amounts of ‘Fe’.
• The hardness survey is carried out on the samples revealed hardness variation
across the length because of the presence of intermetallic compounds. A maximum
Manufacture, Mechanical Properties and Microstructural … 89

hardness value of 79 in B scale was recorded for the sample 6. The concentration
of Fe was at maximum level for this sample.
• Maximum shear strength of 280 MPa was observed for the sample 4. The minimum
shear strength was observed for the sample 3.
• The formation of intermetallic compounds such as FeAl2 , Fe2 Al5 , FeAl6 and
FeAl3 was observed in all the samples in varying amounts. The oxides such as
FeO2 and Al2O3 were noted in few samples.

References

1. Hu W, Kato T, Fukumoto M (2003) Synthesis and nanocrystalline iron aluminide intermetallic

compounds. Mater Trans 44(12):2678–2687
2. Jiqiang MA, Jun YANG, Qinling BI, Weimin LIU (2010) Preparation of an ultrane-grained
Fe-40Al intermetallic Compound. Acta Metall Sin (Engl Lett) 23(1):50–56
3. Prichard PD (1998) The structure-property relationships of powder processed Fe-Al-Si alloys.
PhD Thesis submitted to Iowa State University
4. Chakraborthy SP, Sharma IG (1998) A study on the preparation of iron aluminium based
intermetallic alloy by aluminothermic smelting technique. J Alloy Comp 280:255–261
5. Taylor DGM (2006) Refinement of second phase dispersion in iron aluminide intermetallics by
high temperature severe plastic deformation. Mater Trans 2092–9098
6. Khraisat W, Abu Jadayil W (2010) Strengthening aluminium scrap by alloying with iron. Jordan
J Mechan Ind Eng 4:372–377
Influence of Graphite Particles
on Microhardness and Microstructural
Behavior of AA7068 Metal Matrix
Composites Processed by Powder
Metallurgy

K. John Joshua , S. J. Vijay , P. Ramkumar , and D. Philip Selvaraj

Abstract Automotive industries require components such as automotive shock

absorbers that should be lightweight and has to withstand high wear resistance.
Monolithic aluminum alloys is not enough to meet the demands. Hence aluminum
alloy (AA7068) reinforced with graphite particles were processed by powder metal-
lurgy technique with different weight percentages of graphite (3, 6, and 9%) along
with base matrix. A low pressure of 318 MPa was applied for compaction and sintered
at a temperature of 560 °C for one hour. Vickers microhardness test was carried out
for measuring the microhardness of the composites. Using pin-on-disc wear-tester,
wear experiments were conducted with a velocity of 1.2 m/s over a sliding distance of
2.5 km of load 5 N. Field Emission Scanning Electron Microscopy (FESEM) anal-
ysis was carried out to investigate the worn surface. Experimental results showed
that the Vickers microhardness number has been reduced to 19 VHN by the addition
of 9% graphite particles. Wear experiments revealed improved wear resistance by
the addition of graphite particles. FESEM analysis exposed that abrasion, delami-
nation, and oxidation were the dominant wear mechanisms for the composites. The
automotive shock absorber produced using such a composite combination would be
lightweight with improved wear resistance.

Keywords Metal matrix composites · Powder metallurgy · Microhardness ·

Sintering

K. John Joshua (B)

Department of Marine Engineering, PSN College of Engineering and Technology, Tirunelveli
627152, Tamil Nadu, India
e-mail: kjohnjoshua@gmail.com
S. J. Vijay · P. Ramkumar · D. Philip Selvaraj
Department of Mechanical Engineering, Karunya Institute of Technology and Sciences,
Coimbatore 641114, Tamil Nadu, India

© Springer Nature Singapore Pte Ltd. 2021 91

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_9
92 K. John Joshua et al.

1 Introduction

Nowadays industries have focused its attention towards aluminum alloys due to its
improved properties such as lightweight, corrosion resistance, improved thermal
conductivity, recycling properties, etc. and are widely used in aviation, automobile
and structural industries [1, 2]. AA7068 aluminum alloy provides high mechanical
strength among the alloying elements and provides improved mechanical proper-
ties [3]. Graphite is a soft reinforcement and due to its self-lubricating property, it
provides a better wear resistance than conventional aluminum alloys for Al-Gr com-
posites, which prevents metal to metal contact by forming a thin layer of graphite
particle [4]. Graphite serves as a solid lubricating layer and reduces the composite
wear without the need for solid and liquid lubrication [5]. The addition of graphite in
aluminum matrix leads to a reduction in hardness and flexural strength [6]. Graphite
powder acts as a friction-reducing element, helps to get stronger bond with metals
and widely used in power metallurgy [7]. Graphite has extrinsic lubricating prop-
erty and hence reduces friction and wear. Large amount of graphite added to cop-
per reduces the hardness and fracture strength of the composites [8]. Addition of
graphite in Al7075/Al2 O3 /Gr composites exhibited superior wear-resistance proper-
ties. Smaller amount of aluminum graphite composite exhibits superior wear prop-
erties over the base alloys [9]. Gheorghe et al. [10] fabricated Al/Al2 O3 /Gr hybrid
composites through powder metallurgy route and showed that Vickers microhardness
was increased by increasing the duration of mechanical milling and also by increasing
the reinforcement content of hard Al2 O3 particles. The presence of graphite particles
act as a solid lubricant and reduced the wear rate. Kaushik et al. [11] produced Al
6082 alloy-SiC-Gr hybrid composites using stir casting technique and reported that
the hybrid composites showed improved wear resistance compared to single SiC
reinforced composite and base alloy. The presence of graphite forms a graphitic film
which acts as a self-lubricant in the hybrid composites. Akhlaghi and Pelaseyyed
[12] developed A365 alloy- graphite composites using in situ powder metallurgy
technique and showed that wear rate reduced as a result of increased graphite con-
tent due to the homogeneous distribution of graphite particles with reduced porosity
and fine-grained base alloy. Akhlaghi and Zare-Bidaki [13] produced Al 2024/5-20
wt% flake graphite composites by in situ powder metallurgy technique and reported
that the hardness of these composites decreased with increasing graphite content. Up
to 5 wt% graphite addition the composites showed improved the wear resistance and
beyond 10 wt% or more graphite particles addition, the wear rate increased. Riahi
and Alpas [14] fabricated A356 Al–10% SiC–4% Gr and A356 Al–5% Al2 O3 –3%
Gr composites using molten metal mixing technique. He reported that the compos-
ites have undergone severe wear regime and the hardness of the composites has been
reduced. Graphitic composites demonstrated mild to severe wear at various load and
sliding speed combinations, which were higher than A356 base alloy and A356–20%
SiC composites. Babu et al. [15] studied the wear behavior of aluminum/graphite
nanofiber/alumina short fiber hybrid composites with a different volume fraction of
fibers (10, 15, and 20%) under dry sliding conditions. The composites with 10 and
Influence of Graphite Particles on Microhardness … 93

15 vol.% of fiber showed the lowest friction and wear loss, and increased friction and
wear loss for 20 vol.% of fiber. Ted and Tsao [16] fabricated aluminum/SiC/graphite
hybrid composites with different volume percentages (2, 5, 8%) of graphite using a
semi-solid powder densification technique and studied the tribological behavior. As
the amount of graphite content increased, the hardness of the composites decreased.
The wear rate of the composites increased up to 5% graphite addition and then
reduced for 8% graphite addition. Rajaram et al. [17] fabricated Al-SiC/graphite
hybrid composites using stir casting technique and studied the wear behavior. The
wear properties showed that by the addition of graphite particles, the wear resistance
of the composites increased compared to Al–Si base alloy for all load ranges.
The produced AA7068/graphite metal matrix composites were tested and their
results were reported.

2 Experimental Procedure

2.1 Composite Preparation

FESEM image of AA7068 and graphite powders used in this research work was
demonstrated in Fig. 1. AA7068 was used as a matrix material and graphite was
used as the reinforcement material to produce metal matrix composites using pow-
der metallurgy technique. AA7068 (Zn-8%, Mg-3%, Cu-2%, Si-0.12%, Fe-0.15%,
Mn-0.1%, Cr-0.05%, Zr-0.1%, Ti-0.01%, Al-bal) powders reinforced with different
weight percentages of graphite (3, 6, 9%) particles were milled in a ball mill for
40 h separately. The average particle size of AA7068 powders and graphite powders
were 1.785 µm and 2.07 µm respectively. Stearic acid is mixed up to 2% to improve
the bonding strength between the metal powders [18]. The ball-milled powders were
compacted using a Universal Testing Machine (UTM) at a low pressure of 318 MPa
and sintered at a temperature of 560 °C for an hour to produce composites of size

Fig. 1 FESEM image of powers a AA7068 after milling b Graphite as received

94 K. John Joshua et al.

Fig. 2 Typical composite

specimen

20 mm diameter and 32 ± 2 mm height. A typical composite specimen of a produced

green compact is as shown in Fig. 2. The composites of AA7068 reinforced with
different weight percentages of graphite (3, 6, 9%) along with plain AA7068 were
produced. The produced composites were polished using 600, 800 and 1200 abrasive
grit paper and with 2 µm and 0.5 µm diamond paste in twin disk polisher and etched
in Keller’s reagent for 10 s for cleaning [19]. Three spots were spotted and the aver-
age values were taken for measuring the Vickers microhardness. Using pin-on-disc
wear-tester, experiments were conducted with a velocity of 1.2 m/s over a sliding
distance of 2.5 km of load 5 N. FESEM analysis was carried out to investigate the
worn surface for the produced composites.

3 Results and Discussion

3.1 Microhardness Test

The hardness values of AA7068/graphite composites were presented in Fig. 3. As

the weight fraction of graphite increases, the hardness decreases. The microhardness
was found to be a minimum of 19 VHN for AA7068 reinforced with a weight
fraction of 9% graphite particles. Incorporation of graphite particles into aluminum
alloys normally reduces the strength, ductility, and hardness of the composites. The
decrease in hardness can be attributed due to the increased amount of the brittleness
of graphite particles with the increased affinity of crack instigation and transmission
at the graphite/base alloy boundary [13]. This result was supported by the results of
other researchers [4, 9].
Influence of Graphite Particles on Microhardness … 95

Fig. 3 Vickers
microhardness of
AA7068/Graphite
composites

3.2 Wear Behavior

The graph of wear rate vs sliding distance and volume loss vs. sliding distance for
AA7068/graphite composites were presented in Fig. 4a, b respectively. It was noticed
that the wear rate and the volume loss of all the composites were lower compared
to that of the matrix material. As the graphite content increased, the wear rate and
the volume loss decreased for all the composites. As the sliding distance increased,
wear rate and the volume loss increased for both the base alloy and the composites.
The graphite particles incorporated in the base alloy act as a solid lubricant, thereby
increasing the wear resistance of the composites. The increase in wear resistance was
due to the direct metal contact which was prevented due to the presence of graphite

Fig. 4 a Wear rate vs sliding distance and b volume loss vs sliding distance of AA7068/Graphite
composites
96 K. John Joshua et al.

particles which formed a rich thin tribofilm between sliding surfaces [9]. Composites
reinforced with 9% graphite showed reduced wear rate and volume loss compared
to other composite combinations.

3.3 SEM Analysis

Figure 5a–d demonstrates the FESEM images of AA7068/graphite metal matrix

composites. Figure 5a demonstrates the FESEM image of AA7068 base matrix.
It was noticed that abrasion, delamination, and oxidation were the dominant wear
modes for the matrix material. Deep, shallow and narrow grooves were noticed on
the surface. Grooves present confirms the abrasive wear mechanism paralleling to
the sliding direction. The evidence of grooves and plastic flow rises the shear stress
which promotes delamination wear. The frictional heat generated and shear stress
produces such deep grooves [20, 21]. Figure 5b demonstrates the FESEM image of
AA7068 with the addition of 3% graphite particles. Shallow and narrow grooves
were found on the worn surfaces paralleling to the sliding direction which shows the
occurrence of plastic deformation. This behavior confirms delamination and abrasion

Fig. 5 FESEM analysis of a AA7068 b AA7068-3% Graphite c AA7068-6% Graphite and

d AA7068-9% Graphite
Influence of Graphite Particles on Microhardness … 97

wear mechanisms [9, 22, 23]. Mixture of rough, smooth regions and large cavities
were found. Cracks and cavities were present on the worn surface due to delamination.
Subsurface deformation, nucleation of cracks, propagation of cracks were involved
in delamination. The nucleation of cracks occurs on the surface of the material at
weak points. Figure 5c, d demonstrate the FESEM images of AA7068 with the
addition of 6% graphite and 9% graphite respectively. Shallow, narrow grooves were
present on the worn surfaces along the sliding direction of the pin. Severely damaged
patches and few rough regions were found on the worn surface. Cracks, cavities, and
small pits were visible on the worn surface. Direct metal contact was prevented
due to the presence of graphite particles which form a rich thin tribofilm between
sliding surfaces [9]. Ploughing and exposure of graphite particles were clearly visible.
Ploughing of materials results in hard wear debris entrapped in between the contact
surfaces due to micro-cutting.

4 Conclusions

• Using the powder metallurgy route, AA7068/graphite composites can be pro-

cessed successfully.
• The microhardness of the composites decreased with an increase of graphite
content and AA7068-9% graphite composite exhibited a minimum of 19 VHN.
The decrease in hardness can be attributed due to the increased amount of the
brittleness of graphite particles with the increased affinity of crack instigation and
transmission at the graphite/base alloy boundary.
• The wear experiments revealed an increased wear resistance by increasing the
graphite content. AA7068-9% graphite composite exhibited a maximum resis-
tance which was due to the direct metal contact which was prevented due to the
presence of graphite particles which formed a rich thin tribofilm between sliding
surfaces.
• FESEM analysis demonstrated that dominant wear modes such as abrasion,
delamination, ploughing, and oxidation were predominant for the produced
composites.
• The automotive shock absorber produced using AA7068-9% graphite composite
combination would be lightweight with improved wear resistance which would
meet the industry demands.

Acknowledgements The authors thank Mr. J. Devamanoharan of Centre for Research in Metal-
lurgy (CRM) Laboratory of Karunya Institute of Technology and Sciences for his assistance to carry
out this experimental work.
98 K. John Joshua et al.

References

1. Sandar D, Chatterjee R, Bharathi RJ, Sreehari P (2019) Investigation on the microstructure and
mechanical properties of AA6082/TiO2 surface composites produced by friction stir process-
ing. In: Lakshminarayanan AK, Idapalapati S, Vasudevan M (eds) Conference 2018, ICEMMM.
Springer, Singapore, pp 19–28
2. Kaku SMY, Khanra AK (2019) Hot deformation studies of Al–Cu–Mg powder metallurgy
alloy composite. In: Lakshminarayanan AK, Idapalapati S, Vasudevan M (eds) Conference
2018, ICEMMM. Springer, Singapore, pp 75–81
3. Amin A, Ali S, Omid N, Hamid F, Saeid (2015) Optimizing consolidation behavior of Al
7068–TiC nanocomposites using Taguchi statistical analysis. Trans Nonferrous Met Soc China
25:2499–2508
4. Pardeep S, Dinesh K, Satpal S (2016) Dry sliding wear investigation of Al6082/Gr metal matrix
composite by response surface methodology. J Mater Res Technol 5(1):29–36
5. Kenneth KA, Kazeem OS (2015) Microstructural characteristics, mechanical and wear
behaviour of aluminium matrix hybrid composites reinforced with alumina, rice husk ash
and graphite. Eng Sci Tech 18:416–422
6. Baradeswaran A, Vettivel SC, Elaya P, Selvakumar N, Franklin IR (2014) Experimental inves-
tigation on mechanical behaviour, modelling and optimization of wear parameters of B4C and
graphite reinforced aluminium hybrid composites. J Mater Des 63:620–632
7. Gan Y, Yih P, Chen CQ (2011) Electroplating graphite powder with metals and its use in
microwave absorbing composite. Mater Sci Eng Compos Mater 3(2):129–133
8. Jing FL, Lei Z, Jin KX, Ke CZ (2015) Sliding wear behavior of copper-based composites
reinforced with graphenenanosheets and graphite. Trans Nonferrous Met Soc China 25:3354–
3362
9. Baradeswaran A, Elaya PA (2014) Study on mechanical and wear properties of Al
7075/Al2 O3 /graphite hybrid composites. Compos Part B 56:464–471
10. Gheorghe I, Valeriu GG, Mihai B, Traian B, Mircea IP (2015) Studies on wear rate and micro-
hardness of the Al/Al2 O3 /Gr hybrid composites produced via powder metallurgy. Compos Part
B 69:603–611
11. Kaushik NC, Rao RN (2016) The effect of wear parameters and heat treatment on two body
abrasive wear of Al–SiC–Gr hybrid composites. Tribol Intl 96:184–190
12. Akhlaghi F, Pelaseyyed SA (2004) Characterization of aluminum/graphite particulate com-
posites synthesized using a novel method termed in-situ powder metallurgy. Mater Sci Eng, A
385:258–266
13. Akhlaghi F, Zare-Bidaki A (2009) Influence of graphite content on the dry sliding and oil
impregnated sliding wear behavior of Al 2024–graphite composites produced by in situ powder
metallurgy method. Wear 266:37–45
14. Riahi AR, Alpas AT (2001) The role of tribo-layers on the sliding wear behavior of graphitic
aluminum matrix composites. Wear 251:1396–1407
15. Babu JSS, Kang CG, Kim HH (2011) Dry sliding wear behavior of aluminum based hybrid
composites with graphite nanofiber–alumina fiber. Mater Des 32:3920–3925
16. edGuo ML, Tsao CYA (2000) Tribological behavior of self-lubricating aluminium/SiC/graphite
hybrid composites synthesized by the semi-solid powder-densification method. Compos Sci
Technol 60:65–74
17. Rajaram G, Kumaran S, Srinivasa Rao T (2010) Sliding wear behavior of Al-Si/graphite
composite. Tribol Trans 54:1(115–121)
18. Angelo PC, Subramanian R (2008) Powder metallurgy: science, technology and application,
1st edn. Prentice Hall, India
19. John Joshua K, Vijay SJ, Philip Selvaraj D (2018) Effect of nano-TiO2 particles on micro-
hardness and microstructural behavior of AA7068 metal matrix composites. Ceram Int
44:20774–20781
20. Ashok KB, Murugan N, Dinaharan I (2014) Dry sliding wear behavior of stir cast AA6061-
T6/AlNp composite. Trans Nonferrous Met Soc China 24:2785–2795
Influence of Graphite Particles on Microhardness … 99

21. Yan TY, Lan J, Gao FF, Li QC (2015) Wear behavior and mechanism of B4C reinforced
Mg-matrix composites fabricated by metal-assisted pressureless infiltration technique. Trans
Nonferrous Met Soc China 25:2543–2548
22. Baradeswaran A, Elaya PA (2013) Wear and mechanical characteristics of Al 7075/graphite
composites. Compos Part B 56:472–476
23. Maa W, Lua J, Wanga B (2009) Sliding friction and wear of Cu–graphite against 2024, AZ91D
and Ti6Al4V at different speeds. Wear 266:1072–1081
Microhardness and Microstructural
Behavior of AA7068/SiC Metal Matrix
Composites Synthesized by Powder
Metallurgy

K. John Joshua , P. Ramkumar , S. J. Vijay ,

and S. Mohanasundaram

Abstract If anybody is missing any arm or leg, a prosthetic limb can be replaced
to take care of their activities by themselves. Manufacturing of prosthetic limbs for
such persons should be lightweight and requires improved hardness. Hence, silicon
carbide (SiC) was incorporated into aluminum alloy (AA7068) to produce compos-
ites by powder metallurgy technique with different weight percentages of SiC (3,
6 and 9%) along with base alloy. A compaction pressure of 318 MPa was applied
and sintered at a temperature of 560 °C for an hour. Vickers microhardness test was
carried out to find out the microhardness values. Wear experiments were conducted
with a velocity of 1.2 m/s over a sliding distance of 2.5 km of load 5 N. Field emission
scanning electron microscopy (FESEM) analysis was carried out to investigate the
worn surface. Experimental results showed that the Vickers microhardness number
has been increased to a maximum of 55 VHN by the addition of 9% SiC particles
which were 66.7% more than the base alloy. Wear experiments exposed an improved
wear resistance by the addition of SiC particles. AA7068-9% SiC composite exhib-
ited maximum wear resistance. FESEM analysis revealed that abrasion and oxidation
were the predominant wear modes for the produced AA7068/SiC composites. The
prosthetic limb produced using such composite combination would be lightweight
with the improved hardness which would serve the persons better to perform their
daily activities.

Keywords Metal matrix composites · Wear · FESEM · Microhardness

K. John Joshua (B)

Department of Marine Engineering, PSN College of Engineering and Technology, Tirunelveli
627152, Tamil Nadu, India
e-mail: kjohnjoshua@gmail.com
P. Ramkumar · S. J. Vijay · S. Mohanasundaram
Department of Mechanical Engineering, Karunya Institute of Technology and Sciences,
Coimbatore 641114, Tamil Nadu, India

© Springer Nature Singapore Pte Ltd. 2021 101

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_10
102 K. John Joshua et al.

1 Introduction

Aluminum matrix composites when reinforced with hard ceramic particles replace
monolithic materials and finds a wide range of application in automobile, transporta-
tion, defense and aerospace industries due to its improved properties such as high
strength, improved wear resistance, high corrosion resistance, low density, heat treat-
ment capability, high elastic modulus, high strength-to-weight ratio and high thermal
conductivity [1–4]. Among the 7xxx series aluminum alloys, AA7068 is one of the
most famous one which provides high strength and improved mechanical properties
[5]. Silicon carbide (SiC) is one of the most used ceramic particulates when intro-
duced into aluminum-based matrix that exhibits various improved advantages such
as high corrosion resistance, improved thermal conductivity, high strength and stiff-
ness, improved wear resistance, good machinability and low cost [6–9]. Mohammad
et al. [10] reported that the hardness and wear resistance improved for Al 5252 alloy
reinforced with micro- and nano-sized SiC particles fabricated using in situ pow-
der metallurgy method. Ali et al. [11] stated that the AA6061 reinforced with SiC
synthesized through semi-solid processing exhibits improved hardness and superior
wear resistance compared to the matrix alloy. Ali et al. [12] concluded that A356
alloy reinforced with nano-SiC composites fabricated through stir casting showed the
improved hardness, yield strength and ultimate strength. Devaraju et al. [13] reported
that aluminum alloy 6061-T6 reinforced with (SiC and Al2O3) hybrid composites
fabricated using stir casting showed the improved hardness and wear resistance due
to the presence of hard SiC and Al2 O3 particles. Nidhi et al. [14] stated that Al–
10 wt%SiC composites fabricated using stir casting technique showed reduced wear
rate and coefficient of friction. Dolatkhah et al. [15] concluded that the Al5052/SiC
metal matrix composite fabricated through friction stir processing showed that the
microhardness has been improved up to 55% and wear rate was decreased up to 9.7
times compared with the base matrix. Powder metallurgy is an economical manufac-
turing method to produce components and offers various advantages such as better
dimensional precision, improved surface finish, ability to produce complex shapes
and cost-effectiveness [16].
The mechanical and microstructural properties are characterized for the fabricated
AA7068/SiC composites, and their results are reported.

2 Experimental Procedure

2.1 Composite Preparation

Figure 1 demonstrates the FESEM image of AA7068 and SiC powders used in this
research work. AA7068 was used as a matrix material, and SiC was used as the
reinforcement material to produce metal matrix composites using a powder metal-
lurgy technique. Power metallurgy provides better bonding of matrix and particles,
Microhardness and Microstructural Behavior … 103

Fig. 1 FESEM image of powers a AA7068 after milling, b SiC as received

better distribution and easier control of matrix structure. AA7068 (Zn-8%, Mg-3%,
Cu-2%, Si-0.12%, Fe-0.15%, Mn-0.1%, Cr-0.05%, Zr-0.1%, Ti-0.01%, Al-bal) pow-
ders reinforced with different weight percentages of SiC (3%, 6%, 9%) particles were
milled in a ball mill for 40 h separately. The average particle size of AA7068 powders
and SiC powders were 1.785 µm and 2.253 µm, respectively. Stearic acid is mixed
up to 2% to improve the bonding strength between the metal powders. The ball-
milled powders were compacted using a Universal Testing Machine (UTM) at a low
pressure of 318 MPa and sintered at a temperature of 560 °C for one hour to produce
composites of size 20 mm diameter and 32 ± 2 mm height. Figure 2 demonstrates

Fig. 2 Typical composite specimen

104 K. John Joshua et al.

the typical composite specimen of a produced green compact. The composites of

AA7068 reinforced with different weight percentages of SiC (3, 6, 9%) along with
plain AA7068 were produced. The produced composites were polished using 600,
800 and 1200 abrasive grit paper and with 2 and 0.5 µm diamond paste in twin disk
polisher and etched in Keller’s reagent for 10 s for cleaning.
Three spots were spotted, and the average values were taken for measuring the
Vickers microhardness using Vickers microhardness tester. Wear experiments were
conducted on DUCOM TR 20-LE pin-on-disk apparatus. The experiments were
conducted at a load of 5 N with a sliding velocity of 1.2 m/s at five sliding distances
at an interval of 500 m for a total distance of 2.5 km. The exact mass was measured
after every sliding distance of 500 m after the wear test. FESEM analysis was carried
out to investigate the worn surface for the produced composites to study the dominant
wear behavior.

3 Results and Discussion

3.1 Microhardness Test

The microhardness values of AA7068/SiC composites are presented in Fig. 3. It is

noted from the graph that the microhardness increases as the SiC content increases.
The microhardness rises from 33 VHN (base alloy) to 55 VHN (9 wt% SiC), showing
an increase of 66.7%. It can be attributed due to the presence of hard ceramic SiC
particles in the base alloy, reduced grain size and the SiC particles act as obstacles
to the motion of dislocation [11, 12]. The increase in microhardness is due to the
presence, pinning effect of the SiC particles and Orowan strengthening. The incor-
poration of SiC particles forms an excellent grain structure due to the grain boundary
restrain and improvement in induced strain [13, 17].

Fig. 3 Vickers
microhardness of
AA7068/SiC composites
Microhardness and Microstructural Behavior … 105

3.2 Wear Behavior

The effect of sliding distance on the wear rate is presented in Fig. 4a. As the SiC
content increases from 0 to 9%, the wear rate decreases for all the composites. As the
sliding distance increases, wear rate increases for both base alloy and composites.
According to Archard’s equation, as hardness increases the wear rate decreases. The
increase in wear resistance is due to an increase in the thermal stability of the base
alloy because of the addition of SiC particles [10]. Composites reinforced with 9%
SiC demonstrate superior wear resistance than other composites. The composites
reinforced with SiC exhibit improved wear resistance compared to that of the base
alloy and also increase with the increase in the SiC content. The wear rate increases
with increasing sliding distance, while as the SiC weight fraction increases, wear
rate decreases.
The effect of sliding distance on the volume loss is presented in Fig. 4b. As the
sliding distance increases, the volume loss also increases for both base alloy and
composites. The volume loss decreases as the SiC content increases from 0 to 9%.
Increase in the SiC content weight percentage improves the hardness which improves
the wear resistance of the composites [18]. Composites reinforced with 9% SiC show
reduced volume loss than other composite combinations. The volume loss of the base
alloy is higher compared to other reinforced composites. The hardness of SiC is more
compared to the base alloy which naturally increases the wear resistance of all the
composites. Hard SiC reinforcement particles provide protection, strength to the
base alloy and also improve the thermal stability and load-bearing capacity to the
composites [19].

Fig. 4 a Wear rate versus sliding distance and b volume loss versus sliding distance of AA7068/SiC
composites
106 K. John Joshua et al.

3.3 SEM Analysis

Figure 5a–d demonstrates the FESEM micrographs of AA7068/SiC composites.

Figure 5a shows the FESEM micrograph of unreinforced AA7068 matrix. It is noticed
that abrasion, oxidation and delamination are the dominant wear modes present in
the matrix. Large number of wider, deeper, shallow and narrow grooves paralleling
to the sliding direction are noticed which confirms abrasion wear behavior. Few
pits, scratches, microcuts and rough regions are visible on the worn surface. Wear
debris is caused due to the hard asperities or pulled out hard particles among the
sliding surfaces. This behavior ensures a severe abrasive wear mechanism [9, 15].
Delamination wear is also noticed from the micrograph which can be attributed to
the quality of the interfacial bond between the particles present in the base alloy [10].
Figure 5b, c demonstrates the FESEM micrographs of AA7068-3%SiC and
AA7068-6% SiC composites, respectively. Abrasion and oxidation are the predom-
inant mechanisms for both the samples. Both worn surfaces exhibit a similar type
of wear pattern. By adding SiC reinforcement particles, the hardness of the compos-
ite increases. Shallow, narrow grooves are present parallel to the sliding direction
which confirms abrasion wear mode. The presence of hard ceramic particles resists

Fig. 5 FESEM analysis of a AA7068, b AA7068-3% SiC, c AA7068-6% SiC and d AA7068-9%
SiC composites
Microhardness and Microstructural Behavior … 107

the delamination process which improves the wear resistance of the composites.
The incorporation of hard SiC content improves the strength and reduces the plastic
deformation of the composites. During deformation, SiC particles initiate microc-
racks and develop a strong interfacial bonding between the reinforcement and the
base alloy. By high-stress concentration, the local strain and dislocation density reach
the critical values, as the SiC particles nucleate microcracks [6, 8]. It is also observed
that less cracks and scratches compared to the base alloy. Addition of SiC particles
protects the base alloy and hence improves wear resistance, hardness and prevents
severe wear [11, 13]. Reduced grain size, improved hardness and uniform dispersion
of SiC particles exhibit improved wear properties. Abrasion wear is predominant at
lower loads. Addition of SiC particles reduces the grain size which reduces plas-
tic deformation and improves the wear resistance [15]. Spherical shaped pores are
evident from the worn surfaces. Pores present are likely to occur in powder metal-
lurgy compared to the liquid melting route which may be due to reduced compaction
pressure applied or different sintering parameters [20].
Figure 5d demonstrates the FESEM micrograph of AA7068-9% SiC composite.
Mild abrasion and oxidation are the predominant wear modes found in the composite.
Few narrow, shallow microgrooves are noticed paralleling to the sliding direction.
The depth and width of the grooves present are very less and the surface is finer
and smoother compared to other composite combinations. This can be attributed to
the microploughing tendency, which is less in the composite. During the eventual
process, first the crack initiation layer is removed and then hard SiC particles are
involved in further wear reduction process [21]. Delamination is eliminated in the
composites reinforced with SiC due to the incorporation of hard SiC particles which
has a strong interfacial bond between the SiC and base alloy particles [10]. It is
evident from the micrograph that the oxide layer is present in all the samples. The
oxide layer serves as a protection layer, thereby minimizing the wear loss by reducing
the contact area between the sliding surfaces [22].

4 Conclusions

• AA7068/SiC metal matrix composites have been produced successfully using the
powder metallurgy route.
• The microhardness has been increased as SiC content increased. AA7068-9% SiC
composite exhibited a maximum hardness of 55 VHN which was 66.7% more than
the base alloy.
• The wear resistance has been improved by increasing the SiC content. Addi-
tion of 9% SiC exhibited superior wear resistance compared to other composite
combinations.
• The FESEM analysis disclosed that abrasion and oxidation were the predominant
wear modes found in the produced AA7068/SiC composites.
108 K. John Joshua et al.

• The prosthetic limb produced using AA7068-9% SiC composite combination

would be lightweight with an improved hardness which would serve the persons
better to perform their daily activities.
• AA7068 was the strongest aluminum alloy having zinc and magnesium as major
elements. Further, other new reinforcements such as silicon oxide, titanium car-
bide, titanium boride and zirconium boride can be incorporated with AA7068
base alloy using different fabrication methods such as powder metallurgy, friction
stir welding and friction stir processing to produce composites. The composites
produced with such combinations would meet industry demands with improved
hardness, high stiffness, improved wear resistance, better thermal stability, etc.,
which can be widely used for various applications.

Acknowledgements The authors thank Mr. J. DevaManoharan, Centre for Research in Metallurgy
Laboratory of Karunya Institute of Technology and Sciences, for his assistance to carry out the
experimental work.

References

1. Shyam Kumar CN, Ranjit B, Devinder Y (2016) Wear properties of 5083Al–W surface
composite fabricated by friction stir processing. Tribol Int 101:284–290
2. Ridvan G, HakanAtapek S, Ahmet K (2017) Influence of preform preheating on dry sliding
wear behavior of 304 stainless steel reinforced A356 aluminum matrix composite produced by
melt infiltration casting. Tribol Int 115:608–618
3. Hayrettin A, Tolga K, Ercan C, Huseyin C (2006) Wear behaviour of Al/(Al2 O3 pCSiCp) hybrid
composites, Tribol Int 39:213–220
4. Ramesh CS, Keshavamurthy R, Channabasappa BH, Pramod S (2010) Friction and wear
behavior of Ni–P coated Si3 N4 reinforced Al6061 composites. Tribol Int 43:623–634
5. Amin A, Ali S, Omid N, Hamid F, Saeid S (2015) Optimizing consolidation behavior of Al
7068 − TiC nanocomposites using Taguchi statistical analysis. Trans Nonferrous Met Soc
China 25:2499–2508
6. Umanath K, Palanikumar K, Selvamani ST (2013) Analysis of dry sliding wear behaviour of
Al6061/SiC/Al2 O3 hybrid metal matrix composites. Compos. Part B 53:159–168
7. Taherzadeh MR, Azari RK, Yazdani S, Brabazon D, Boostani AF (2016) Fabrication of alu-
minum matrix composites reinforced with nano- to micrometer-sized SiC particles. Mat Des
89:58–70
8. Min S, Yue-hui H (2010) Effects of die-pressing pressure and extrusion on the microstructures
and mechanical properties of SiC reinforced pure aluminum composites. Mat Des 31:985–989
9. Uthayakumar MS, Aravindan S, Rajkumar K (2013) Wear performance of Al–SiC–B4 C hybrid
composites under dry sliding conditions. Mater Des 47:456–464
10. Mohammad MG, Farshad A (2016) Wear behaviour of al 5252 alloy reinforced with
micrometric and nanometricSiC particles. Tribol Int 102:28–37
11. Ali M, Mohsen Ostad S (2013) Microstructural and abrasive wear properties of SiC rein-
forced aluminum-based composite produced by compocasting. Trans Nonferrous Met Soc
China 23:1905–1914
12. Ali M, Mohsen Ostad S (2012) Characterization of cast A356 alloy reinforced with nanoSiC
composites. Trans Nonferrous Met Soc China 22:275–280
Microhardness and Microstructural Behavior … 109

13. Devaraju A, Kumar A, Kumaraswamy A, Kotiveerachari B (2013) Influence of reinforcements

(SiC&Al2 O3 ) and rotational speed on wear and mechanical properties of aluminum alloy
6061-T6 based surface hybrid composites produced via friction stir processing. Mater Des
51:331–341
14. Nidhi J, Anshul B, Mondal DP, Das S, Singh M (2011) Sliding wear behavior of aluminum
syntactic foam: a comparison with Al–10 wt%SiC composites. Tribol Int 44:220–231
15. Dolatkhah A, Golbabaei P, BesharatiGivi MK, Molaiekiya F (2012) Investigating effects of
process parameters on microstructural and mechanical properties of Al5052/SiC metal matrix
composite fabricated via friction stir processing. Mater Des 37:458–464
16. Martin F, Garcı C, Blanco Y, Aparicio L (2013) Tribo corrosion behaviour of powder metallurgy
duplex stainless steels sintered in nitrogen. Tribol Int 57:76–85
17. Devaraju A, Kumar A, Kotiveerachari B (2013) Influence of addition of Grp /Al2 O3 p with SiCp
on wear properties of aluminum alloy 6061-T6 hybrid composites via friction stir processing.
Trans Nonferrous Met Soc China 23:1275–1280
18. Jayakumar L, Balamurugan K (2014) Reciprocating wear behavior of 7075Al/SiC in compar-
ison with 6061Al/Al2 O3 composites. Int J Refrac Met Hard Mat 46:137–144
19. Adel MH, Abdalla A, Mohammed TH, Ahmad TM (2009) Wear behavior of Al–Mg–Cu–based
composites containing SiC particles. Tribol Int 42:1230–1238
20. John Joshua K, Vijay SJ, Philip Selvaraj D (2018) Effect of nano-TiO2 particles on micro-
hardness and microstructural behavior of AA7068 metal matrix composites. Ceram Int
44:20774–20781
21. Kaushik NC, Rao RN (2016) The effect of wear parameters and heat treatment on two body
abrasive wear of Al–SiC–Gr hybrid composites. Tribol Int 96:184–190
22. Antony VKC, Selwinrajadurai J (2016) Influence of rutile (TiO2 ) content on wear and
microhardness characteristics of aluminium-based hybrid composites synthesized by powder
metallurgy. Trans Nonferrous Metals Soc China 26:63–73
Optimization of CO2 Laser Cutting
Parameters for AA6061/B4 C/hBN
Hybrid Composites using Taguchi-based
Response Surface Methodology

A. Gnanavelbabu , V. Arunachalam , K. T. Sunu Surendran ,

K. Rajkumar , and E. Anandhababu

Abstract Aluminium hybrid composites due to its tailored properties are being
used in many industrial applications. Machining of metal matrix hybrid composites
with desired quality through conventional methods is a great challenge. As advance-
ment, non-conventional cutting method such as laser cutting is effectively employed
to cut hybrid composites. The present work investigates the cutting quality char-
acteristics of AA6061/B4 C/hBN hybrid composites using pulsed CO2 laser cutting
method. Optimization of machining parameters plays a vital role in improving the
quality characteristics. The parameters involved in this non-contact-type cutting pro-
cess such as laser power (2000–2200 W), gas pressure (12–14 bar), speed of spot
movement (600−1000 mm/s), pulsing frequency (60–100 Hz) and reinforcement
(5–15 Vol.%) were varied according to Taguchi L27 orthogonal array, and the quality
characteristics such as surface roughness, kerf width and edge slope were noted.
Response surface methodology (RSM) was applied to generate quadratic models
and 3D surface plots. Desirability analysis was carried out to obtain the optimal con-
ditions. The optimal parameters obtained are: laser power—2000W, gas pressure—
12.05 bar, speed of spot movement—998.07 mm/s, pulsing frequency—60 Hz and
reinforcement—6.81 Vol.%.

Keywords Aluminium hybrid composites · CO2 laser cutting · Surface response

methodology (RSM) · Desirability analysis

1 Introduction

The needs of structural materials with high specific strength and stiffness have catal-
ysed the emergence of metal matrix composite (MMC). MMCs are intrinsically

A. Gnanavelbabu (B) · V. Arunachalam · K. T. Sunu Surendran · E. Anandhababu

Department of Industrial Engineering, CEG, Anna University, Chennai 600025, India
e-mail: agbabu@annauniv.edu
K. Rajkumar
Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai 603110, India

© Springer Nature Singapore Pte Ltd. 2021 111

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_11
112 A. Gnanavelbabu et al.

smart, high performing and lightweight materials that are being utilized for various
automobile, aircraft and industrial applications [1]. To act as a solution over alu-
minium alloys, the attention shifts to aluminium metal matrix composites (AMMCs)
because of the beneficial properties like lightweight, good castability and machinabil-
ity, durability, good mechanical properties, etc. [2]. Further, properties of AMMCs
can be enhanced by the addition of ceramic (SiC, B4 C, Al2 O3, etc.) and solid lubricant
(CNT, graphite, hBN, etc.) reinforcements [3]. Nowadays, boron carbide (B4 C) is
used as ceramic reinforcement because of its superior properties like high strength,
low density, high hardness and good chemical stability. Hexagonal boron nitride
(hBN) solid lubricant improves the friction and wear resistance of the composite
material, and moreover, it reduces the cutting zone temperature [4–7].
Laser cutting is a technology that uses a laser to cut materials and is typically
used for industrial manufacturing applications. Laser beam machining (LBM) is a
non-traditional subtractive manufacturing machining process, in which a laser is
directed towards the work piece for machining. Chinmaya R. et al. investigated the
experimental evaluation of laser-assisted machining of silicon carbide reinforced
aluminium matrix composites. It is found that laser-assisted machining minimizes
the specific cutting energy, roughness, tool wear and surface damage as compared
to traditional machining process [8]. Adalarasan et al. confirmed the possibility of
obtaining good quality characteristics such as surface roughness, material removal
rate and kerf width using pulsed CO2 laser cutting of Al6061/Al2 O3 composites
[9]. Response surface methodology (RSM) is an easy way to calculate a first-degree
polynomial model, and it uses a fractional factorial design. Adalarasan et al. eluci-
dated that the application of Taguchi-RSM was to optimize the drilling parameters
for Al6061/SiC/Al2 O3 hybrid composite and the experimental trials were designed
using Taguchi’s L18 OA [10].
The literature survey reveals that a less number of research works has been reported
particularly for the optimization of CO2 laser machining parameters on aluminium
metal matrix hybrid composites. Therefore, the main objective of this research work
is to optimize the effects of CO2 laser cutting parameters over surface roughness, kerf
width and edge slope on AA6061-B4 C-hBN hybrid composites using Taguchi-RSM.

2 Materials and Methodology

AA6061 was chosen as the matrix material as it is the most commonly used
precipitation-hardened aluminium alloy for the development of aluminium compos-
ites due to its favourable properties like good mechanical properties and wettability.
B4 C is a hard ceramic commonly used as particulate reinforcements in aluminium
to improve the wear resistance and strength. Hexagonal boron nitride (hBN) is well
known for its excellent lubricating properties, chemical inertness, high-temperature
stability, etc. hBN in composites improve the machinability by acting as a self-
lubricant and reducing the cutting zone temperature due to its good thermal con-
ductivity. In this experiment, B4 C and hBN particles of size 20 µm and 5 µm,
Optimization of CO2 Laser Cutting Parameters for AA6061/B4 C/hBN … 113

Table 1 Parameters and levels

Symbol Parameter Low level Medium level High level
A Laser power (W) 2000 2100 2200
B Gas pressure (bar) 12 13 14
C Speed of spot movement (mm/s) 600 800 1000
D Pulsing frequency (Hz) 60 80 100
E Reinforcement (Vol. %) 5 10 15

respectively, were used as the reinforcements. Three different aluminium hybrid

composites were fabricated by varying volume percentages of B4 C (5–15%) and
hBN (10%) by stir casting technique. CO2 laser cutting was performed on the hybrid
composites using Mitsubishi 3015-eX series (max. power: 3.8 KW) CO2 laser cut-
ting machine. Taguchi’s L27 orthogonal array was utilized to design the experiment
trials. The parameters and its levels are shown in Table 1. Five different controlling
parameters with three levels denoted by low, medium and high were chosen for this
study. Parameters such as stand-off distance and flow rate of assist gas were kept
constant throughout the cutting. The optimization of the laser cutting parameters
was carried out by applying RSM. Desirability graph for all responses was generated
to assess the contribution of input parameters.

3 Results and Discussions

The aluminium metal matrix hybrid composites were successfully fabricated using
stir casting process. The CO2 laser cutting was performed on the composites as per
L27 orthogonal array, and the corresponding responses such as surface roughness
(Ra), kerf width (mm), edge slope (°) are shown in Table 2.
The cutting parameters and their responses were analysed using Design–Expert
software. Polynomial equations of the second order were formed statistically with
the laser cutting parameters and quality characteristics joined with RSM. The math-
ematical model is created with the help of back selection technique and elimination
of higher order to maintain the model hierarchy. This model explains the individ-
ual and interaction of cutting parameters on surface roughness, kerf width and edge
slope. The analysis on the influence of input parameters on responses was carried
out by ANOVA, predicted versus actual graph, response surface plot and quadratic
equation.

3.1 Analysis on Surface Roughness

The ANOVA for the surface roughness is shown in Table 3. SS denotes Sum of
114 A. Gnanavelbabu et al.

Table 2 Parameters, levels and responses of CO2 laser cutting

Trials A (W) B C D E Surface Kerf Edge
(Bar) (mm/s) (Hz) (Vol.%) roughness width slope
(Ra) (mm) (°)
1 2000 12 600 60 5 4.154 0.332 3.16
2 2000 12 600 60 10 5.098 0.227 2.16
3 2000 12 600 60 15 6.473 0.262 2.50
4 2000 13 800 80 5 4.054 0.298 2.84
5 2000 13 800 80 10 4.986 0.262 2.50
6 2000 13 800 80 15 6.331 0.270 2.57
7 2000 14 1000 100 5 3.966 0.321 3.06
8 2000 14 1000 100 10 4.965 0.294 2.80
9 2000 14 1000 100 15 6.295 0.290 2.76
10 2100 12 800 100 5 4.228 0.266 2.54
11 2100 12 800 100 10 5.116 0.252 2.40
12 2100 12 800 100 15 6.654 0.284 2.70
13 2100 13 1000 60 5 4.334 0.282 2.69
14 2100 13 1000 60 10 5.214 0.276 2.63
15 2100 13 1000 60 15 6.659 0.285 2.71
16 2100 14 600 80 5 4.490 0.300 2.86
17 2100 14 600 80 10 5.499 0.266 2.53
18 2100 14 600 80 15 6.681 0.304 2.90
19 2200 12 1000 80 5 4.797 0.253 2.41
20 2200 12 1000 80 10 5.587 0.201 1.91
21 2200 12 1000 80 15 6.836 0.319 3.04
22 2200 13 600 100 5 4.946 0.225 2.14
23 2200 13 600 100 10 5.559 0.223 2.12
24 2200 13 600 100 15 6.908 0.206 1.96
25 2200 14 800 60 5 5.078 0.240 2.29
26 2200 14 800 60 10 5.739 0.240 2.29
27 2200 14 800 60 15 6.968 0.356 3.39

Squares, “D” denotes the Degree of Freedom, M.S. denotes Mean Square and “S”
denotes Significant.
The modelF-value of 269.53 implies the model is significant. Values of “Prob
>F” less than 0.0001 indicates that model terms are significantly contributed. In this
case, A—Laser power, E—Vol.% of reinforcement, AE, E2 are the significant model
terms.
The closeness of the fit determines the actual and predicted values into the sub-
jected desired level. From Fig. 1, it is evident that the actual and predicted values are
Optimization of CO2 Laser Cutting Parameters for AA6061/B4 C/hBN … 115

Table 3 ANOVA for surface roughness

Source SS D M.S. F-value P-value Prob. Remarks
>F
Model 24.40 14 1.74 269.53 <0.0001 S
A 0.70 1 0.70 116.65 <0.0001 S
B 0.026 1 0.026 4.39 0.0579
C 0.024 1 0.024 3.95 0.0703
D 2.772E−003 1 2.772E−003 0.46 0.5103
E 21.69 1 21.69 3602.55 <0.0001 S
AB 4.335E−004 1 4.335E−004 0.072 0.7930
AC 7.633E−003 1 7.633E−003 1.27 0.2822
AD 0.000 0 0.000
AE 0.089 1 0.089 14.80 0.0023
BC 2.427E−003 1 2.427E−003 0.40 0.5374
BD 0.000 0 0.000
BE 0.012 1 0.012 1.94 0.1893
CD 0.000 0 0.000
CE 4.070E−003 1 4.070E−003 0.68 0.4270
DE 2.791E−003 1 2.791E−003 0.46 0.5089
A 4.843E−003 1 4.843E−003 0.80 0.3874
B2 0.000 0 0.000
C2 0.000 0 0.000
D2 0.000 0 0.000
E2 0.35 1 0.35 57.57 <0.0001 S
Residual 0.072 12 6.020E−003
Corr. total 24.47 26
R-Squared = 0.9970; Adj. R-squared = 0.9936; Mean = 5.47

on the same line without any appreciable spreading. It ensures the closeness of the
fit as well as the validity of the fitted model.
Figure 2 illustrates the 3D response surface plot for surface roughness. X-axis
represents laser power, Y-axis represents Vol.% of reinforcement and Z-axis rep-
resents surface roughness. Increase in reinforcement percentage and laser power
increases the surface roughness of the composite. The best surface roughness 4.0197
is achieved with the combination of the lowest laser power and the lowest reinforce-
ment. Even though the minimal interaction was found between the reinforcement and
laser power, laser power emerged to be a dominant parameter in affecting the surface
roughness by means of its linear as well as interaction conditions. The regression
equation for the surface roughness is given in Eq. 1.

Quadratic equation for surface roughness = + 5.28 + 0.36 ∗ A + 0.077 ∗ B

116 A. Gnanavelbabu et al.

Fig. 1 Actual versus predicted values for surface roughness

Fig. 2 Three-dimensional response surface plot for surface roughness

− 0.073 ∗ C − 0.033 ∗ D + 1.10 ∗ E − 0.017 ∗ A ∗ B

+ 0.071 ∗ A ∗ C − 0.086 ∗ A ∗ E + 0.035 ∗ B ∗ C
− 0.031 ∗ B ∗ E E + 0.018 ∗ C ∗ +0.015 ∗ D ∗ E
+ 0.038 ∗ A2 + 0.24 ∗ E 2 (1)

3.2 Analysis on kerf width

The ANOVA for the kerf width is shown in Table 4.

Optimization of CO2 Laser Cutting Parameters for AA6061/B4 C/hBN … 117

Table 4 ANOVA for kerf width

Source SS D M.S. F-value P-value Prob. Remarks
>F
Model 0.030 14 2.141E−003 3.16 0.0263 S
A 6.434E−004 1 6.434E−004 0.95 0.3489
B 4.481E−003 1 4.481E−003 6.62 0.0244 S
C 8.405E−004 1 8.405E−004 1.24 0.2870
D 6.401E−004 1 6.401E−004 .95 0.3500
E 1.934E−004 1 1.934E−004 .29 0.6028
AB 9.074E−005 1 9.074E−005 .13 0.7207
AC 2.308E−003 1 2.308E−003 3.41 0.0897
AD 0.000 0 0.000
AE 7.105E−003 1 7.105E−003 10.50 0.0071 S
BC 2.420E−004 1 2.420E−004 0.36 0.5610
BD 0.000 0 0.000
BE 4.88E−004 1 4.88E−004 0.69 0.4216
CD 0.000 0 0.000
CE 1.261E−003 1 1.261E−003 1.86 0.1974
DE 5.468E−004 1 5.468E−004 0.81 0.3865
A2 1.017E−003 1 1.017E−003 1.50 0.2438
B2 0.000 0 0.000
C2 0.000 0 0.000
D2 0.000 0 0.000
E2 6.913E−003 1 6.913E−003
Residual 8.124E−003 12 6.770E−004 10.21 0.007
Corr. total 0.038 26
R-Squared = 0.7860; Adj. R-squared = 0.5386; Mean = 0.27

The modelF-value of 3.16 implies the model is significant. Values of “Prob >F”
less than 0.0263 indicates that model terms are significantly contributed. In this case,
B, AE are the significant model terms.
Figure 3 represents the predicted and actual values for kerf width. Higher R2 value
indicates the closeness of all the points to the regression line. The obtained R2 value
78.6 indicates a good fitness for kerf width, and the model is validated.
Figure 4 represents the response surface plot for kerf width. Laser power, rein-
forcement percentage and kerf width are on X, Y, Z axes, respectively. Minimum kerf
width 0.200121 was achieved at minimum laser power and nominal reinforcement
%. Increase in reinforcement percentage and laser power increases the kerf width of
the composite. Equation 2 is the corresponding quadratic equation for kerf width.

Quadratic equation for kerf width = + 0.26 − 0.011 ∗ A + 0.032 ∗ B + 0.014 ∗ C

118 A. Gnanavelbabu et al.

Fig. 3 Actual versus predicted values for kerf width

Fig. 4 Three-dimensional response surface plot for kerf width

+ 0.016 ∗ D + 3.278E − 003 ∗ E + 7.778E − 003 ∗ A ∗ B

+ 0.039 ∗ A ∗ C + 0.024 ∗ A ∗ E + 0.011 ∗ B ∗ C
+ 6.250E003 ∗ B ∗ E + 0.010 ∗ C ∗ E − 6.750E
− 003 ∗ D ∗ E − 0.017 ∗ A2 + 0.034 ∗ E 2− (2)

3.3 Analysis on Edge Slope

The ANOVA for the edge slope is shown in Table 5.

Optimization of CO2 Laser Cutting Parameters for AA6061/B4 C/hBN … 119

Table 5 ANOVA for edge slope

Source SS D M.S. F-value P-value Prob. Remarks
>F
Model 2.73 14 0.20 3.17 0.0260 S
A 0.060 1 0.060 0.97 0.3437
B 0.41 1 0.41 6.68 0.0239 S
C 0.079 1 0.079 1.28 0.2802
D 0.060 1 0.060 0.97 0.3445
E 0.016 1 0.016 0.26 0.6172
AB 9.074E−003 1 9.074E−003 0.15 0.7077
AC 0.21 1 0.21 3.44 0.0884
AD 0.000 0 0.000
AE 0.64 1 0.64 10.47 0.0071 S
BC 0.021 1 0.021 0.35 0.5667
BD 0.000 0 0.000
BE 0.042 1 0.042 0.68 0.4247
CD 0.000 0 0.000
CE 0.11 1 0.11 1.79 0.2056
DE 0.051 1 0.051 0.82 0.3819
A2 0.092 1 0.092 1.50 0.2446
B2 0.000 0 0.000
C2 0.000 0 0.000
D2 0.000 0 0.000
E2 0.63 1 0.63 10.27 0.0076 S
Residual 0.74 12 0.062
Corr. Total 3.47 26 3.17
R-Squared = 0.7887; Adj. R-squared = 0.5423; Mean = 2.59

The modelF-value of 3.17 implies the model is significant. Values of “Prob >F”
less than 0.0260 indicates that model terms are significantly contributed. In this case,
B, AE, E 2 are significant model terms.
Figure 5 shows the relationship between actual and predicted values of edge slope.
R2 value for edge slope is 78.87. It indicates the closeness of points to the regression
line, and the model is validated.
Figure 6 represents the response surface plot for edge slope. Laser power, rein-
forcement percentage and edge slope are on X, Y, Z axes, respectively. Minimum
edge slope 1.9107 achieved at minimum laser power and nominal reinforcement %.
Increase in reinforcement percentage and laser power increased the edge slope of the
composites. Equation 3 is the quadratic equation for edge slope.

Quadratic equation for edge slope = + 2.480 − 10 ∗ A + 0.30 ∗ B + 0.13 ∗ C

120 A. Gnanavelbabu et al.

Fig. 5 Actual versus predicted values for edge slope

Fig. 6 Response surface plot for edge slope

+ 0.15 ∗ D + 0.030 ∗ E + 0.078 ∗ A ∗ B + 0.38 ∗ A ∗ C

+ 0.23 ∗ A ∗ E + 0.10 ∗ B ∗ C + 0.059 ∗ B ∗ E + 0.096 ∗ C ∗ E
− 0.065 ∗ D ∗ E − 0.168A2 + 0.32 ∗ E 2 (3)

3.4 Desirability Analysis

Design of expert software was employed to perform the desirability analysis. The
obtained values for parameters and responses are shown in Fig. 7. The desirability
Optimization of CO2 Laser Cutting Parameters for AA6061/B4 C/hBN … 121

Fig.7 Desirability graph for surface roughness, kerf width and edge slope

value 0.990 proves higher proximity of outputs towards the target.

4 Conclusions

CO2 laser cutting on AA6061/B4 C/hBN hybrid composites was successfully done
as per the structured L27 orthogonal array. Optimization of process parameters was
carried out by applying RSM, and the following conclusions were drawn.
1. Minimum surface roughness of 4.0197 Ra was obtained by maintaining the laser
power at 2000 W, pressure 12.03 bar, speed 1000 mm/s, frequency 60.01 Hz and
reinforcement 6.65 Vol.%.
2. Minimum kerf width 0.200121 mm was obtained at laser power 2000 W, pressure
12 bar, speed 999.04 mm/s, frequency 60.16 Hz and reinforcement 6.68 Vol.%.
3. At laser power 2000 W, pressure 12 bar, speed 999.04 mm/s, frequency 60.16 Hz
and reinforcement 6.68 Vol.%, a minimum edge slope 1.9107° was achieved.
4. From the desirability analysis of the parameters, a desirability value of 0.990 was
obtained. It showed that the optimal condition for pulsed CO2 laser cutting on
AA6061/B4 C/hBN composite was laser power = 2000 W, pressure = 12.05 bar,
speed of spot movement = 998.07 mm/s and pulsing frequency = 60 Hz and
Vol.% of % reinforcement = 6.81.

Acknowledgement The financial support for consumables in this research work was provided by
“DST PURSE Phase II”.
122 A. Gnanavelbabu et al.

References

1. Bains PS, Sidhu SS, Payal HS (2016) Fabrication and machining of metal matrix composites.
Mater Manufact Proc 31:573
2. Jojith R, Radhika N (2018) Mechanical and tribological properties of LM13/TiO2 /MoS2 hybrid
metal matrix composite synthesized by stir casting. Part Sci Technol 1–13
3. Hung NP, Zhong ZW Zhong CH (1997) Grinding of metal matrix composites reinforced with
silicon-carbide particles. Mater Manufact Proc 12:1075–1091
4. Yao YT, Chen LQ (2016) B4 C/Al Composites Processed by Metal assisted Pressure less
Infiltration Technique and its Characterization. Mater Manufact Proc 31:1286–1291
5. Santhosh S, Rajkumar K, Gnanavelbabu A (2015) Effect of hBN solid lubricant concentration
on machinability of titanium (Ti-6Al-4V) alloy. Mater Sci Forum 830:87–90
6. Gnanavelbabu A, Rajkumar K, Saravanan P (2018) Investigation on the cutting quality charac-
teristics of abrasive water jet machining of AA6061-B4C-hBN hybrid MMC. Mater Manufact
Process 1313–1323
7. Muthazhagan C, Gnanavelbabu A, Bhaskar GB, Rajkumar K (2014) Influence of graphite
reinforcement on mechanical properties of Al-boron carbide composites. Adv Mater Res
845:398–402
8. Dandekar CR, Shin YC (2012) Experimental evaluation of laser-assisted machining of sil-
iconcarbide particle-reinforced aluminium matrix composites. Int J Adv Manufact Technol
66:1603–1610
9. Adalarasan R, Santhanakumar M, Thileepan S (2016) Selection of optimal machining param-
eters in pulsed CO2 lasercutting of Al6061/Al2 O3 composite using Taguchi-based response
surface methodology (T-RSM). Int J Adv Manufact 93:305–317
10. Adalarasan R, Santhanakumar M (2015) Application of taguchi based response surface method
(TRSM) for optimization of multi responses in drilling Al/SiC/Al2 O3 hybrid composites. J Inst
Eng (India) 96:65–71
Analysis of Corrosion Resistance
in Domestic Water Geysers by Coating
Nano-Film Using Thermal Spray Coating

Kabilan Sankar, Karthick Selvam, K. Joy Ashwin, and S. P. Sathya Prasanth

Abstract The lifespan of domestic water geysers drastically reduces over a period of
time of usage due to environmental challenges such as corrosion and scale develop-
ment. The objective to be achieved here is to increase the property of corrosion resis-
tance of internal components of the geyser and overcoming the challenges posed by
such. A steel plate is to be coated with a nano-film of specified thickness using thermal
spray coating consisting of nano-alumina, nano-zinc oxide and nano-molybdenum.
These elements were chosen for their high corrosion resistance and formation of a
superior alloy coating. The coating elements are mixed in three different ratio com-
binations, and each is coated on three stainless steel sample plates. These sample
plates are tested for corrosion and scale development by performing a salt spray test
for a period of time till visual examination shows the development of corrosion. A
comparison of the results of the three samples is analysed with industry standards
to determine the suitable alternative for geyser coatings. The results have convinc-
ingly proved that there is an increase in the life expectancy of the geyser which can
significantly reduce maintenance expenditure.

Keywords Domestic water geyser. corrosion resistance · Nano-Film coating

1 Introduction

Corrosion is the slow destruction of a material by chemical or electrochemical reac-

tions that occur due to interaction with the environment. They are generally known
to deform and degrade the material, therefore making it non-usable for any purpose.
One such area impacted by corrosion is the domestic water geyser. Geysers used for
this purpose have two types of tanks, a stainless steel tank with no coating and a
mild steel tank coated with porcelain enamel. The most predominant type in India
is the stainless steel tank as it is less expensive to manufacture. However, stainless
steel tanks and porcelain-coated tanks also have a very limited lifetime averaging at

K. Sankar (B) · K. Selvam · K. Joy Ashwin · S. P. Sathya Prasanth

Department of Mechanical Engineering, KCG College of Technology, Chennai, India
e-mail: sankar.kabilan@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 123

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_12
124 K. Sankar et al.

12–18 months. Many geysers undergo corrosion at a rate with respect to the hard-
ness of the water and frequency of usage. The high temperature and hardness of
water combined create a favourable environment for the degradation of the tank at
a gradual pace. McAdam et al. [1] performed experiments to determine the rate of
calcium carbonate formation and how its growth can be controlled. Shi et al. [2] did
a study on the impact of nano-aluminium articles in corrosion resistance coatings.
This experiment provided strong backing for the selection of nano-alumina to be
included in the coating. Kourtidou et al. [3] performed experiments on Ni-Al coat-
ings in high temperature and marine environments. This project aims at developing
a better alternative using nano-technology by providing a nano-film coating of com-
mercially available materials that possess high corrosion resistance properties [4].
A study was conducted to analyze the types of geysers available in the commercial
market before coming to a conclusion on material selection [5].The materials that
fit this category were alumina, zinc and molybdenum. These materials were mixed
in three different ratios to determine the most suitable and effective coating mixture
[6–8] that can withstand the corrosion attacks [9, 10] inside the geyser. Salt spray
test would then be performed to determine the most suitable coating.

2 Methodology

The nano-materials selected for the coating are alumina, zinc and molybdenum. The
three elements were mixed in three different specified ratios. The nano-materials
were in powder form and mixed using a ball grinding machine. After ensuring the
thorough mixing of the materials, each sample ratio was coated onto a stainless steel
plate of dimension 75 mm × 75 mm × 1 mm. The plates had to be pre-treated before
applying the coating. The coating method used here is the plasma thermal spray
coating where the materials were coated onto the plate by blowing at a temperature
of 4000 °C. The plate is then cooled to allow the material to be coated well onto the
plate. The three stainless steel plates coated with different ratios of nano-coating are
tested using a salt spray test for a period of 96 h to determine the impact of corrosion
resistance. Visual examination is performed to determine the impact of corrosion
on the three sample plates, and the results are compared to determine the effect of
corrosion (Fig. 1).

Mixing of na- Thermal Spray Salt Spray Test- Examination

nomaterials by Coating of Na- ing of test sam- and Analysis
ball grinding nomaterials ples of Results

Fig. 1 Experimental set-up

Analysis of Corrosion Resistance in Domestic Water Geysers … 125

3 Experimental Details

Alumina and zinc are commercially feasible and widely used elements. Both elements
possess high corrosion resistance properties. Molybdenum is a rare earth element that
is also used in the steel industry for its ability to impart high corrosion resistance
properties to the alloy it is alloyed to. Hence, they are the reason for the selection of
these nano-materials. The three nano-materials consist of a particle size of 80 nm. The
three nano-materials are mixed in three different ratios. Three test samples are used
here, and the best ratio of material coating is selected from the three. The materials
are weighed as per the ratio required and put into the ball grinding machine. The ball
grinding machine is used to mix the elements thoroughly so as to prevent any large
concentrations of same elements when coated onto the stainless steel plate.

3.1 Reference Samples

Reference Sample 1
Reference sample 1 is one of the two benchmark samples that are considered among
the industry best standards. The salt spray test results of this reference sample will be
compared with the salt spray test results of test samples. Figure 2 shows an image of
the plate sample used. Stainless steel’s resistance to corrosion and staining make it
an ideal material for many applications where corrosion resistance is required. This
reference sample will be tested by salt spray testing for the duration of 96 h. The
stainless steel plate used is of the grade SAE 304 with a dimension of 75 mm in
length, 75 mm in width and 1 mm thickness.

Fig. 2 Reference sample 1

before salt spray testing
126 K. Sankar et al.

Fig. 3 Reference sample 2

before salt spray testing

Reference Sample 2
Reference sample 2 is one of the two benchmark samples that are considered best
according to industry standards whose salt spray test results will be compared with
those of the test samples. The mild steel grade used for this sample is of Indian
Standard 2062 and its dimensions are 75 mm in length, 75 mm in width and a
thickness of 2 mm. The enamel coating is done on top of the steel plate’s surface. It
is a mild steel plate consisting of a coating of enamel coating with a coating thickness
of 125 µm. Figure 3 shows an image of the plate after coating.

3.2 Thermal Spray Coating of Nano-Mixture

The nano-materials after mixing are coated onto the stainless steel plate using thermal
spray coating for a thickness of 125 µm. Initially, the surface of the stainless steel
plate undergoes pre-treatment before coating can be done. The surface is wiped
clean to remove off any dust or foreign particles. The surface undergoes bit blasting
where it is sprayed with aluminium oxide. The surface is then treated with a nickel
chromium bond coating to ensure proper adhesion of the nano-material coating. A
spray gun is used to eject the nano-mixture onto the plate at a speed of 80–90 m/s
with argon gas to prevent external reaction with foreign elements. The amount of
material coated per minute is 30 gm/min and the spray gun is at a distance of 127 mm
from the surface of the stainless steel plate. The details of the three test samples are
described in detail along with an SEM image.

Test Sample 1
Test sample 1 contains nano-alumina, nano-zinc oxide and nano-molybdenum that
are mixed together with the help of ball grinding, and the mixture ratio is as shown in
Table 1. After that process, they are fed to the thermal spray coating machine where
Analysis of Corrosion Resistance in Domestic Water Geysers … 127

Table 1 Ratio of nano-material for test sample 1

Sample Base metal % of alumina % of zinc % of mum Coating
number thickness
Test sample 1 Stainless steel 50 35 15 125 µm

they will be coated onto the stainless steel plate. The stainless steel plate used here
is of the grade SAE 304, and its dimensions are 75 mm in length, 75 mm in width
and 1 mm in thickness. The stainless steel undergoes pre-treatment of its surface for
better adhesion of coating of the nano-material mixture. The nano-material mixture
is coated on top of the stainless steel plate with a coating thickness of 125 µm (Fig. 4;
Table 2).

Test Sample 2
Test sample 2 contains nano-alumina, nano-zinc oxide and nano-molybdenum that
are mixed together with the help of ball grinding, and the ratio is as shown in Table 3.

Fig. 4 SEM image of test

sample 1

Table 2 Test samples before coating and after coating

Test sample Before coating After coating
Test sample 1
128 K. Sankar et al.

Table 3 Ratio of nano-material for test sample 2

Sample Base metal % of alumina % of zinc % of Coating
number molybdenum thickness
Test sample 1 Stainless steel 65 25 10 125 µm

After that process, they are fed to the thermal spray coating machine where they will
be coated onto the stainless steel plate. The stainless steel plate used here is of the
grade SAE 304, and its dimensions are 75 mm in length, 75 mm in width and 1 mm
in thickness. The nano-material mixture is coated on top of the stainless steel plate
with a coating thickness of 125 µm (Fig. 5; Table 4).

Test Sample 3
The three elements are mixed together with the help of ball grinding, and in the
mixture, the ratio is as shown in Table 5. After that, they are fed to the thermal spray
coating machine where they will be coated onto the stainless steel plate. The stainless
steel plate used here is of the grade SAE 304, and its dimensions are 75 mm in length,

Fig. 5 SEM image of test

sample 2

Table 4 Test samples before coating and after coating

Test sample Before coating After coating
Test sample 2
Analysis of Corrosion Resistance in Domestic Water Geysers … 129

Table 5 Ratio of nano-material for test sample 3

Sample Base metal % of alumina % of zinc % of Coating
number molybdenum thickness
Test sample 1 Stainless steel 80 15 5 125 µm

75 mm in width and 1 mm in thickness. The stainless steel undergoes pre-treatment

of its surface for better adhesion of coating of the nano-material mixture. The nano-
material mixture is coated on top of the stainless steel plate with a coating thickness
of 125 µm (Fig. 6; Table 6).

Fig. 6 SEM image of test

sample 3

Table 6 Test samples before coating and after coating

Test sample Before coating After coating
Test sample 3
130 K. Sankar et al.

3.3 Salt Spray Testing of Test Samples and Reference

Samples

The salt spray test is a standardized corrosion test method used to test for corrosion
resistance of materials and surface coatings. It is an accelerated corrosion test that
produces a corrosive attack to coated samples in order to evaluate the suitability of
the coating for use as a protective finish. The apparatus consists of a closed testing
chamber where a saltwater (5% NaCl) solution is atomized by means of spray nozzles
using pressurized air. This produces a corrosive environment of dense saltwater fog
in the chamber so that the test samples are subjected to severely corrosive conditions.
The chamber temperature is maintained at 70.5–71.5 °C with a pH value of 6.65–
6.85. The components are loaded into the chamber at an angle of 30° with air pressure
at 14–18 psi.

4 Examination and Analysis of Results

4.1 Reference Sample 1

Reference sample 2 is a commonly used and commercially feasible stainless steel

plate. It is the most common type of tank available in India due to its reduced selling
price. However, the sample did not last longer than 24 h in the salt spray test. This can
be justified by the possibility of no additional coating on the surface and moderate
corrosion resistance properties. It can be observed through visual examination that
the plate has undergone extensive damage during the salt spray test. Hence, this type
of coating cannot increase the life expectancy of a geyser (Table 7).

Table 7 Test samples before coating and after coating

Before testing After testing
Analysis of Corrosion Resistance in Domestic Water Geysers … 131

Table 8 Test samples before coating and after coating

Before testing After testing

4.2 Reference Sample 2

This sample is considered to be the best in industry standard that is commercially

available. However, the sample did not last long in the salt spray test. It underwent
extensive corrosion damage and did not last longer than 48 h. Though enamel coating
does have corrosion resistance properties, it is not enough to provide a geyser with
the increased life expectancy that needs to be achieved (Table 8).

4.3 Test Sample 1

Test sample 1 image shows it turned the colour of the mixture into a dark silver.
Thus, the appearance of the coating on the test plate is dark in colour compared
to the other three samples. After 96 h of salt spray testing, test sample showed the
endurance and good resistance to corrosion attack from the salt spray. This can be
due to the possibility of an additional percentage of molybdenum and zinc oxide,
both having high corrosion resistant properties. Visual examinations do show slight
traces of corrosion which could have occurred due to the mixture not containing
optimum ratio (Table 9).

4.4 Test Sample 2

The darkness in the colour of the coating has reduced to a minor extent which can
be justified by the reduction in the percentage of molybdenum added in this mixture.
After 96 h of salt spray testing, the test sample showed good endurance and very
good resistance to corrosion attack from the salt spray. The percentage of alumina
in this sample has been increased by 15% with a reduction of 10% in zinc oxide and
132 K. Sankar et al.

Table 9 Test samples before coating and after coating

Before testing After testing

Table 10 Test samples before coating and after coating

Before coating After coating

a 5% reduction in molybdenum. Visual examinations show no traces of corrosion or

erosion, suggesting high corrosion resistance for this coating. This can be because
alumina has bonded well with zinc oxide and molybdenum suggesting this is the
optimum ratio for a stronger anti-corrosion coating (Table 10).

4.5 Test Sample 3

The darkness in colour is the least in test sample 3 among the three test samples. This
can be justified by the very low presence of molybdenum, and a higher percentage
of alumina which is white in colour. Visual examinations showed that the sample
underwent quite extensive corrosion damage but not as severe as the reference test
samples. This can be justified by the increased percentage of alumina and very
few percentages of zinc with traces of molybdenum. Alumina’s corrosion resistance
Analysis of Corrosion Resistance in Domestic Water Geysers … 133

Table 11 Test samples before coating and after coating

Before testing After testing

properties are lesser compared to that of zinc and molybdenum and are the reason
why the corrosion damage is greater among the three samples (Table 11).

5 Conclusion

The three test samples and two reference samples were tested by salt spray testing for
a period of 96 h. The salt spray test provided the simulation of a geyser except that the
saltwater is constantly being sprayed without any intervals. This test provided us with
an insight on how the five samples would survive in what is considered an extreme
environment. Test sample 2 proved to be the optimum combination of nano-materials
that can withstand the extreme environment of the geyser without falling victim to
corrosion. Test results showed that it was expected to last well beyond the 96 h of
testing after performing a visual examination of its surface. Test sample 1 ranked
second in its ability to withstand corrosion. This could be due to the almost correct
ratio of elements but apparently not the optimal ratio to withstand for long. Test
sample 3 ranked third due to a large percentage of alumina and reduced percentages
of zinc oxide and molybdenum. Both the reference samples did not last longer than
48 h. Reference sample 2 ranked fourth lasting for 48 h before giving way to attacks
of corrosion. Reference sample 1 ranked fifth and the last among the five samples
not lasting longer than 24 h providing evidence of why many commercially available
geysers in India have a reduced life expectancy and undergo frequent maintenance
activities. Therefore, it can be concluded that test sample 2 consists of the optimum
ratio of nano-materials to resist attacks of corrosion faced in a domestic water geyser.
This is supported by evidence from the salt spray test.
134 K. Sankar et al.

References

1. McAdam J, Parsons SA (2004) Calcium carbonate scale formation and control. Publ Views
Env Sci Technol 3(2)
2. Shi X, Nguyen TA, Suo Z, Liu Y, Avci R (2009) Influence of Nano-Al concentrates on the
corrosion resistance of epoxy coatings. Elsevier-Surface Coat Technol 204(3):237–245
3. Kourtidou D, Chaliampalias D, Tarani E, Pavlidou E (2019) Deposition of Ni-Al coatings
by pack cementation and corrosion resistance in high temperature and marine environments.
Elsevier Corrosion Sci 148:12–23
4. Gal J-Y, Bollinger J-C, Tolosa H, Gache N (1996) Calcium: a reappraisal of scale formation
and inhibition. Published in Talanta 43(9):1497–1509
5. Venus Home Appliances webpage (2019) https://www.venushomeappliances.com/tank-
comparision.html. Last accessed 12 May 2019
6. Khodair ZT, Khadom AA, Jasim HA (20185) Corrosion protection of mild steel in different
aqueous media via epoxy/nanomaterial coating: preparation, characterization and mathematical
views. J Mater Res Technol
7. Liang Y, Liu F-C, Nie M, Zhao S, Lin J, Han E-H (2013) Influence of Nano-Al concentrates
on the corrosion resistance of epoxy coatings. J Mater Sci Technol 29(4):353–358
8. Koutsomichalis A, Vaxevanidis NM, Psyllaki P (2016) Wear and corrosion behaviour of zinc-
alumina electrodeposited coatings on mild steel. Tribological J Bultrib 6(06)
9. Kadhim MJ, Sukkar KA, Abbas AS, Obaeed NH (2017) Investigation of nano coating for
corrosion protection of petroleum pipeline steel type A106 grade B: theoretical and practical
study in Iraqi Petroleum Sector. Eng Technol J 35(Part A, No. 10)
10. Bhakre VS, Nayak DK (2007) Nanotechnology in corrosion prevention. Published at inter-
national conference CORCON 2007 held at Mumbai focusing on cost of corrosion to the
nation
Effect of Multi-walled Carbon Nanotubes
Additions on Its Dispersion
Characteristics in Titanium Oxides

Senzeni Sipho Lephuthing, Avwerosuoghene Moses Okoro, Masego Mohlala,

Noxolo Malgas, Oladeji Ige, and Peter Apata Olubambi

Abstract The transition metal oxides are inorganic nanomaterials with exclusive
properties. However, metal oxides cannot exceptionally accomplish all requirements
to develop new technologies as they are brittle and have low fracture toughness,
which has motivated researchers to adopt novel techniques in enhancing their prop-
erties. Carbon nanotubes (CNTs) have drawn attention in engineering fields due
to their unique structural, electrical and mechanical properties. This work aims to
achieve a homogeneous dispersion of multi-walled carbon nanotubes (MWCNTs)
by investigating the effect of increased MWCNTs concentration in titanium oxide.
The MWCNTs (0.5, 1, 1.5 wt.%) were dispersed in titanium oxide using high energy
ball milling (HEBM) technique at a lower milling speed of 100 rpm for 6 h. The
composite powders were characterized using scanning electron microscopy (SEM),
transmission electron microscopy (TEM), Raman spectroscopy and X-ray diffraction
(XRD) to understand the uniformity of dispersion, interfacial reaction and structural
integrity of the MWCNTs after dispersion. The results showed that the MWCNTs
detangles at a lower concentration of 0.5 wt%, with an increase in MWCNTs there
is the formation of entanglement. In addition, further analysis on structural integrity
showed minimal structure damage due to lower speed adopted for HEBM technique,
and there were no interfacial reactions favoured.

Keywords Titanium oxide · Multi-walled carbon nanotubes (MWCNTs) ·

Homogeneous · Dispersed

S. S. Lephuthing (B) · A. M. Okoro · M. Mohlala · N. Malgas · O. Ige · P. Apata Olubambi

Department of Metallurgy, School of Mining, Metallurgy and Chemical Engineering, Center for
Nanoengineering and Tribocorrosion, University of Johannesburg, 2006 Gauteng, Republic of
South Africa
e-mail: senzenil@uj.ac.za

© Springer Nature Singapore Pte Ltd. 2021 135

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_13
136 S. S. Lephuthing et al.

1 Introduction

The utilization of CNTs for diverse structural and engineering applications have
spurred more research attention in the recent time. This is due to their attractive,
outstanding and superlative properties [1–4]; high Young’s modulus in the tera-pascal
(TPa) range and excellent strength about 100 GPa to synthesize less dense materials
for diverse applications. Additionally, the exclusive properties of CNTs are ascribed
to their unique morphology and the existence of strong sp2 carbon–carbon (C–C)
bonds in their outermost shells. These sp2 C–C bonds form a fastened honeycomb
lattice and high bond strength in the structure of the material [4].
The development of durable, inert and insoluble composites has been of interest in
scientific research communities. This composite materials demand have encouraged
the integration of CNTs into titanium-based materials to enhance their properties for
composites production. However, during the processing of these composites, CNTs
tend to agglomerate into clusters and this is due to their nanostructural dimension,
high aspect ratio and the strong van der Waal forces that exist between their individual
tubes [5–7]. In many cases, the increase in the content of MWCNTs as reinforcements
improves the properties of the matrix material but poses the tendency of agglomera-
tion and deteriorates the dispersion characteristics of the MWCNTs within the matrix
materials [7, 8] which is ascribed to the corresponding of increase in the strong van
der Waal forces among individual nanotubes.
In a bid to improve the properties of the titanium oxides and achieve uniform
dispersion of MWCNTs in the metal oxides, optimized dispersion techniques are
highly desirable to overcome the strong van der Waal forces and achieved optimal
dispersion characteristics of the MWCNTs in titanium oxides [7]. In this study, high
energy ball milling (HEBM) method with regulated milling parameters was adopted
[9] to 1.0 and 1.5 wt.% of MWCNTs in titanium oxide, and various characterization
techniques were used to investigate the effect of an increase in MWCNTs on its
dispersion.

2 Materials and Methods

2.1 Starting Materials

The starting powder materials used for this study were MWCNTs (NX7100-A-Ultra
purified CNT) with an average diameter of 9.5 nm and an average length of 1.5 μm,
supplied by Nanocyl, Belgium. Titanium (IV) oxide anatase powder with percentage
purity (99.6%) and −325 mesh particle size was supplied by industrial analytical in
Johannesburg, South Africa.
Effect of Multi-walled Carbon Nanotubes Additions … 137

2.2 High Energy Ball Milling (HEBM)

High energy ball mill (Retch 400, Germany) was used to 1 and 1.5 wt.% MWCNTs
into titanium oxide and this was conducted by charging the titanium oxide powder in
an alumina vial (250 ml, inner diameter (ID) = 100 mm) with alumina balls of 5 mm
diameter. The milling was carried out using a ball to powder ratio (BPR) was 10:1 at
100 rpm milling time and speed of 6 h and 100 rpm, respectively. A relaxation time
of 10 min was observed during the milling operation to prevent the samples from
overheating. The admixed powders were collected and characterized after milling.

2.3 Characterization of As-Received and Admixed Powders

Scanning electron microscopy (FESEM, Carl Zeiss Sigma) equipped with energy-
dispersive X-ray spectroscopy (EDX) and transmission electron microscopy (JEOL-
Jem 2100) were used for morphological characterization of the starting and admixed
powders. X-ray diffractometer (XRD, Rigaku D/max-rB) was used for phase identi-
fication using Cu-Kα(λ = 0.154 nm) radiation at a scanning rate of 1°/min over the
angular range of 20–90°. Raman spectroscopy was used to ascertain the structural
integrity of the MWCNTs before and after milling

3 Results and Discussion

3.1 Morphology of As-Received and Admixed Powders

Figure 1(a)–(d) shows the morphology characteristics of as-received powders which

are SEM and TEM analysis. Figure 1(a) shows the SEM image anatase-type titanium
oxide powder (TiO2 ) that has a spherical shape of similar particle sizes. Figure 1(b) is
SEM image of pristine MWCNTs agglomerated together which is attributed by their
large aspect ratio, high surface area strong van der Waal forces between individ-
ual MWCNTs [7]. This is confirmed by TEM image in Fig. 1(c) where MWC-
NTs are bundled into clusters and the high-resolution transmission microscopy
image (HRTEM) in Fig. 1(d) shows the walls of non-destructive individual pristine
MWCNTs.
In order to attain unique properties of MWCNTs in TiO2 , an effective homoge-
nous dispersion of MWCNTs with good interfacial bonding alongside with mini-
mal destruction of their tubular structure has to be achieved [10]. Past works have
emphasized that the addition of high concentration of MWCNTs helps to improve
the properties of the TiO2 [7, 8, 10]. However, the increase in concentration of MWC-
NTs during composite production poses the tendency of increase in agglomeration
of the nanotubes which is as a result of the nanostructural dimension, high aspect
138 S. S. Lephuthing et al.

Fig. 1 SEM images of the as-received powders (a) Titanium oxide, (b) MWCNTs, (c) TEM image
of as-received MWCNTs and (d) HRTEM image showing the walls of MWCNTs

ratio and the strong van der Waal forces that exist between individual MWCNTs [7]
This features of MWCNTs deteriorate the dispersion characteristics in the TiO2 .
The strategies that usually adopted in achieving the detangling of MWCNTs
during the powder processing are to further increase the milling parameter such as
milling time, speed and BPR which could results on defects on the graphitic structure
of the MWCNTs [11] which affect their properties to be attained on TiO2 powder.
Since, it is essential to efficiently transfer these unique properties to TiO2. This can
be achieved by homogeneous dispersion of MWCNTs with good interfacial bonding
without damaging the walls of the nanotubes. Hence, this study adopted the powder
processing using HEBM technique to disperse MWCNTs in TiO2 powder at lower
regulated milling parameters.
Figure 2(a)–(c) shows the SEM images of (0.5, 1.0, 1.5) wt% MWCNTs dispersed
into TiO2 . Figure 2(a) shows the SEM image where 0.5 wt% MWCNTs dispersed in
TiO2 , and it is observed that the MWCNTs are relatively detangled which could be
attributed by less effectiveness of van der Waal forces between individuals. However,
Effect of Multi-walled Carbon Nanotubes Additions … 139

Fig. 2 SEM images of dispersed MWCNTs in titanium oxide (a) 0.5 wt%, (b) 1.0 wt% and (c)
1.5 wt

Fig. 2(b) shows an increase concentration in MWCNTs to 1.0 wt% dispersed in

TiO2 powder and slight entanglement of MWCNTs within TiO2 powder begin to be
noticed. Further increase in MWCNTs concentration to 1.5 wt% dispersed in TiO2
powder is shown in Fig. 2(c). It is observed that MWCNTs are prominently entangled
within MWCNTs with increase in MWCNTs concentration. This could be attributed
by increase in van der Waal forces. In addition, due to lower milling parameter that
were adopted also played a role in dispersion behaviour of the MWCNTs dispersed
in TiO2 powder.
Hence, it was observed that a degree of agglomeration became visible with
increase in MWCNTs wt% concentration due to insufficient impact energy exerted
on the composite during powder processing. Meanwhile, increase in impact energy
has its concerns as it may damage the outer layer of the MWCNT by introducing
open edges and vacancies. This will lead to the destruction of the basal planes of
MWCNTs exposing the prismatic planes, where interfacial products are favoured
[9].
Figure 3(a)–(f) shows the TEM images of (0.5, 1.0 and 1.5) wt% of MWCNTs
dispersed in TiO2 powder, and this was conducted to further ascertain the dispersion
behaviour of MWCNTs as well how the walls were evolved during powder pro-
duction. Figure 3(a)–(c) confirms on what has been observed on SEM images that
MWCNTs are relatively detangled at lower concentration; however, as the wt% con-
centration increases, the entanglement begins until it becomes obvious when 1.5 wt%
is dispersed in TiO2 powder. There is bonding effect of MWCNTs on TiO2 particles,
140 S. S. Lephuthing et al.

Fig. 3 TEM images of dispersed MWCNTs in titanium oxide (a) 0.5 wt% (b) 1 wt% and (c)
1.5 wt% and HRTEM for tubular characteristics (d) 0.5 wt% (e) 1 wt% and (f) 1.5 wt%

and it is observed by wrapping of MWCNTs around that TiO2 particles which may
lead to transfer of unique properties of MWCNTs in TiO2 .
Moreover, HRTEM analysis shown in Fig. 3(d)–(f) for dispersion of MWCNTs
in TiO2 powder. Figure 3(d) displays a slightly evolution on the walls compared to
the pristine MWCNTs. However, when the concentration of MWCNTs dispersed in
TiO2 powder increases to 1.0 and 1.5 wt% in Fig. 3(e)–(f), there is obvious crystalline
structure observed and this confirms that MWCNTs were embedded within titanium
oxide and this could be possibility of good interfacial bonding that leads to the
transfer of the unique properties of MWCNTs to titanium oxide [12].

3.2 Structural Integrity and Phase Analysis of Dispersed

MWCNTs in TiO2

Raman spectroscopy analysis was used to evaluate the structural integrity of MWC-
NTs after dispersion in titanium oxide. This analysis was carried out specifically to
ascertain the D-band which is associated with defects in form of sp3 disorders in the
C–C bonds and the G-band which is related to the plane stretching mode of the C–C
bond to determine the degree of graphitization, orderliness of carbon, metallicity
and crystalline nature of the MWCNTs. Meanwhile, the ID /IG ratio is used to assess
the crystallinity and structural disorder of the MWCNTs during powder processing
[13, 14].
Effect of Multi-walled Carbon Nanotubes Additions … 141

Fig. 4 (a) Raman shift spectrum and (b) XRD patterns Raman shift of dispersed 0.5, 1 and 1.5 wt%
MWCNTs in titanium oxide

Figure 4a shows the Raman spectra of the pristine MWCNTs and the admixed
composite powders. The graphite peaks of the pristine MWCNTs are observed corre-
spondingly, D (1340 cm−1 ) and G (1570 cm−1 ) bands. This was followed by pristine
TiO2 powder which showed anatase structures at B1g (399 cm−1 ), A1g (518 cm−1 )
and Eg (641 cm−1 ), respectively [15]. The dispersion of MWCNTs (0.5, 1.0 and
1.5) wt% shifted the D- and G-band from the pristine of MWCNTs D (1340 cm−1 )
and G (1570 cm−1 ) to (D)1341.3 cm−1 and (G) 1575.6 cm−1 , (D)1345.2 cm−1 and
(G)1575.3 cm−1 and (D)1347.6 cm−1 and (G) 1577.7 cm−1, respectively. This could
have been attributed by the energy extended to the walls of MWCNTs during power
processing. However, there were no interfacial reactions products which are carbide
phases observed on the Raman spectra. Application of lower milling parameters
adopted for MWCNTs dispersion played a role in preserving the walls [16].
The ID /IG ratio of the pristine MWCNTs was calculated to be 0.800, after the dis-
persion of MWCNTs at (0.5, 1.0 and 1.5 wt%) in titanium oxide. There was increase
in ID /IG ratio form 0.800 to 0.914, 0.937 and 1.023, respectively. This characteristic
is an indication of radical shift that broadens the intensity peaks of the D-band and
G-band which is ascribed to higher stress on the walls of MWCNTs during milling.
The stress on the nanotubes may lead to the formation of weak sp3 C–C amorphous
phase from broken sp2 bonds due to harsh milling conditions which favours forma-
tion of carbides [9, 12]. Carbide phases are usually formed when there are severe
damages to the walls of the nanotubes during dispersion.
142 S. S. Lephuthing et al.

The destruction in the walls of the nanotubes usually occurs when high milling
speed, long milling time and extremely high ball to powder ratios are adopted during
milling operation which may further affect the unique properties of the nanotubes
that are transferrable to augment the properties of titanium oxides. From past studies,
HEBM technique is still adjudged as the most effective route for dispersing MWCNTs
in TiO2 [9, 11] especially when regulated milling parameters are adopted to minimize
the destruction to the walls of the MWCNTs during milling operation.
X-ray diffraction technique was conducted to analyse the different phases, struc-
tural changes and the evolution of the crystalline phases in the as-received and
admixed composite powders. The XRD results are shown in Fig. 4b where prominent
peaks are observed at 25.6 and 43.3 which corresponds to (002) and (100) planes of
the MWCNTs. While, in the admixed composite powders, the following planes were
identified (101), (004), (200), (105), (211), (204), (215) and (303) which ascertain the
anatase type of titanium oxide powder. In addition, the XRD results further revealed
that the MWCNT (002) plane was overlaps by the anatase-type titanium oxide (101)
plane in the MWCNTs/titanium oxide composite. Moreover, it was also observed
that with increase in MWCNT concentration, there was a decrease in peak intensity
on corresponding plane (101) of MWCNTs this may be ascribed to the interlock of
the MWCNTs in the titanium oxide.

4 Conclusion

It has been successfully demonstrated that MWCNTs can be dispersed using high
energy ball milling HEBM technique. Though, this dispersion technique has draw-
backs of destruction the MWCNTs wall due to harsh milling conditions during pow-
der processing which poses a challenge on losing their ability to improve TiO2 .
However, milling conditions were controlled to minimize the destruction as shown
in this study. Therefore, the following was drawn from this study.
• The SEM results showed that the lower concentration of 0.5 wt% MWCNTs was
relatively de-agglomeration of MWCNTs. However, when further increases in the
MWCNTs the prominent agglomerated MWCNTs were shown.
• The TEM results confirmed the SEM results, however, at HRTEM analysis. It
was observed that there was good bonding which showed by crystal structures
and this my attribute to good interfacial bonding the leads to transfer of the their
unique properties.
• The structural integrity of MWCNTs showed that indeed, there was good inter-
facial bonding achieved during powder processing. Though, the D- and G-band
were stretched which could defeat of MWCNTs and there was an increase in the
ID /IG ratio that affects the crystallinity. The destruction was minimal as there were
no interfacial reactions observed of the Raman spectra.
Effect of Multi-walled Carbon Nanotubes Additions … 143

Acknowledgements National Research Foundation of South Africa is acknowledged for funding

this research study.

References

1. Mallakpour S, Khadem E (2016) Carbon nanotube–metal oxide nanocomposites: Fabrication,

properties and applications. Chem Eng J 302:344–367
2. Soldano C (2015) Hybrid metal-based carbon nanotubes: novel platform for multifunctional
applications. Prog Mater Sci 69:183–212
3. Mohammadi M (2016) No Title. Dev Pb-MnO2 Compos anodes electrowinning Appl Elec-
trochem Corros Eval. LNCS Homepage, http://www.springer.com/lncs. Last accessed 21 Nov
2016
4. Liu Y, Chae HG, Ho Choi Y, Kumar S (2014) Effect of carbon nanotubes on sintering behavior
of alumina prepared by sol-gel method. Ceram Int 40(5):6579–6587
5. Zapata-Solvas E, Gómez-García D, Domínguez-Rodríguez A (2012) Towards physical prop-
erties tailoring of carbon nanotubes-reinforced ceramic matrix composites. J Eur Ceram Soc
32:3001–3020
6. Lephuthing SS, Okoro AM, Lesufi M et al (2018) Effect of milling parameters on the dispersion
characteristics of multi-walled carbon nanotubes in transition metal oxides. IOP Conf Ser Mater
Sci Eng 430:12002. https://doi.org/10.1088/1757-899X/430/1/012002
7. Munir KS, Li Y, Liang D, Qian M, Xu W, Wen C (2015) Effect of dispersion method on the
deterioration, interfacial interactions and re-agglomeration of carbon nanotubes in titanium
metal matrix composites. JMADE 88:138–148
8. Munir KS, Zheng Y, Zhang D, Lin J, Li Y, Wen C (2017) Improving the strengthening efficiency
of carbon nanotubes in titanium metal matrix composites. Mater Sci Eng, A 696(March):10–25
9. Munir KS, Kingshott P, Wen C (2015) Carbon nanotube reinforced titanium metal matrix
composites prepared by powder metallurgy—a review. Crit Rev Solid State Mater Sci 40(1):38–
55
10. Thostenson ET, Ren Z, Chou T-W (2001) Advances in the science and technology of carbon
nanotubes and their composites: a review. Compos Sci Technol 61(13):1899–1912
11. Hassan MTZ, Esawi AMK, Metwalli S (2014) Effect of carbon nanotube damage on the
mechanical properties of aluminium–carbon nanotube composites. J Alloys Compd 607:215–
222
12. Munir KS, Wen C (2016) Deterioration of the strong sp2 carbon network in carbon nan-
otubes during the mechanical dispersion processing—a review. Crit Rev Solid State Mater Sci
41(5):347–366
13. Adegbenjo AO, Olubambi PA, Potgieter JH, Shongwe MB, Ramakokovhu M (2017) Spark
plasma sintering of graphitized multi-walled carbon nanotube reinforced Ti6Al4V. Mater Des
128(May):119–129
14. Ahmad I, Cao H, Chen H et al (2010) Carbon nanotube toughened aluminium oxide
nanocomposite. J Eur Ceram Soc 30:865–873
15. Inam F, Yan H, Reece MJ, Peijs T (2008) Dimethylformamide : an effective dispersant for
making ceramic—carbon nanotube composites. 195710. https://doi.org/10.1088/0957-4484/
19/19/195710
16. Klein MV, Holy JA, Williams WS (1978) Raman scattering induced by carbon vacancies in Ti
C x. Phys Rev B 17:1546
Evaluation of Process Parameters
Influence on the Mechanical Properties
of RF Magnetron Sputtered TiC
Thin-Film Coating

Olayinka Oluwatosin Abegunde, Esther Titilayo Akinlabi,

and Oluseyi Philip Oladijo

Abstract The mechanical properties of a thin-film coating are very paramount to

the development and applications for both research and industrial usages. Enhanced
mechanical properties give vast options of applications, curb cost and mitigate the
unprecedented failure of materials in services. Optimizing the process parameter is
very crucial in developing thin-film coatings mechanical properties. Taguchi was
used to develop the experimental matrix using three levels and three factors. Three
different sputtering parameters were varied during the experimental process from low
to high level. Mechanical characterizations using nanoindentation, nanoscratch and
wear tests were performed on the samples to determine the optimal process parameter
that yields the best mechanical properties. From Taguchi analysis, the RF power
proves to be the most effective control factor in optimizing the hardness and Young’s
modulus properties with a contribution rate of 70.4% and 84.4%, respectively.

Keywords ANOVA · Taguchi · RF magnetron sputtering

1 Introduction

Mechanical properties have a correlation with other properties like corrosion, opti-
cal properties and surface morphology. Mechanical properties can be improved via
different manufacturing techniques such as composite, heat treatment and surface
modification [1–3]. RF magnetron sputtering is a useful tool technique for improv-
ing these properties by coating the surface with materials of superior qualities. A
detailed analysis of thin-film coating by these processes can be found in Reference

O. O. Abegunde (B) · E. T. Akinlabi · O. P. Oladijo

University of Johannesburg, Johannesburg 2006, South Africa
e-mail: oabegunde@uj.ac.za
E. T. Akinlabi
Covenant University, Ota, Nigeria
O. P. Oladijo
Botswana International University of Science and Technology, Palapye, Botswana

© Springer Nature Singapore Pte Ltd. 2021 145

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_14
146 O. O. Abegunde et al.

[4–7]. Although studies have been done on the applications of sputtering for sur-
face modification [8–10], further understanding of the contribution of each process
parameters of RF magnetron sputtering on the response properties still needs more
research study. Understanding the influence of RF power, sputtering and tempera-
ture on the mechanical properties is important for both industrial application and
research advancement. In this study, Taguchi was used to developing the experi-
mental matrix. The control factors considered are RF power, sputtering time and
temperature. An orthogonal three levels and three factors method (l9 33 ) were used.
The percentage contribution and influence of the independent process parameters’
factors on the hardness and Young’s modulus were analyzed using ANOVA analysis
and signal-to-noise ratio.

2 Experimental Method

2.1 RF Magnetron Sputtering

TiC thin films were deposited in an RF magnetron sputtering coater which was
specifically designed for the deposition process. The substrate (CpTi) was ground
using silica carbide paper and then polished to remove any surface impurities. The
surface of the substrate was further cleaned in acetone and stored in a desiccator
to prevent contamination and surface oxidation. The purity of the TiC target used
is 99.99% pure TiC. The target–substrate distance was kept fixed throughout the
experiments. Evacuation of the sputtering chamber to a base pressure of 1.13 × 10−5
m bar was done for all the coating process before sputtering takes place. The reason
for the low pressure is to provide a long mean free path for a collision between the
target and substrate and allows for the control and minimization of contaminants.
After the evacuation, the system is refilled with argon to a partial pressure of 2.5 ×
10−3 m bar. Argon flow rate of 12 sccm was used, and constant working pressure
was maintained. Maintaining constant pressure helps to place many argon atoms in
the path of the ions and ejected coatings. After the plasma formation, the sputtering
of the target is triggered, and the shutter is removed. Pre-sputtering was done for
all the samples for 5 min before the main sputtering deposition. The substrate was
located directly above the target on a rotating disk to ensure uniform distribution and
condensation of the target on the surface of the substrate. The target material was
cleaned up after each cycle to prevent build up charge, thereby reducing the arcing
effect.
Evaluation of Process Parameters Influence on the Mechanical … 147

2.2 Process Parameters and Optimization

The sputtering experiments were conducted by varying three crucial parameters

of magnetron sputtering, namely RF power, sputtering time and temperature. The
experimental matrix optimization was developed using the Taguchi method. Taguchi
method is a simple, efficient and powerful design tool for statistical analysis and
optimization. It lessens the number of experimental runs to be conducted and curbs
the effects of external factors that cannot be controlled on the system. The flowchart
of the process is shown in Fig. 1, and Table 1 represents the magnetron sputtering
parameters and their levels. The full orthogonal array of Taguchi L9 (33) with three
factors at three levels of low, mid and high and a total of nine deposition runs is
presented in Table 2. The time ranges from 2 to 3 h, power from 150 to 250 W and
temperature from 80 to 1000 °C.

Fig. 1 Flowchart of experimental design

Table 1 Control factors level for experimental setup

Sputtering parameters Symbol Level 1 Level 2 Level 3
RF Power (W) A 150 200 250
Sputtering time (Hrs) B 2.0 2.5 3.0
Temperature (°C) C 80 90 100

Table 2 Orthogonal array of

Experiment no. RF power Sputtering time Temperature
Taguchi L9 (33 )
L1 1 1 1
L2 1 2 2
L3 1 3 3
L4 2 1 2
L5 2 2 3
L6 2 3 1
L7 3 1 3
L8 3 2 1
L9 3 3 2
148 O. O. Abegunde et al.

2.3 Nanoindentation

Hysitron Triboindenter (T1950, USA) was used to investigate the nanomechanical

properties (Young’s modulus and hardness) of thin-film coating under ambient work-
ing condition. Load controlled testing which was modeled to have a trapezoidal flow
pattern of loading-dwelling-unloading was used to generate the indentation profile of
the thin-film coating with a holding time of 15 s at maximum peak load. The loading
and unloading rate of 10 µN s−1 and a maximum load of 300 µN were used. The
diamond indenter of the Triboindenter is made up of Berkovich tip of 100 nm radius
of curvature. Young’s modulus and nanohardness of measured thin films were cal-
culated using Oliver and Pharr analysis [11–13] from the load-displacement curves.
The indentations for each thin film were repeated for 10 times, and the measure-
ment recorded corresponds to the average of the 10 measurements. To ensure that
the measurement taken is actually for the thin film and eliminate the influence of the
substrate on the results, the indenter depth of penetration was maintained at 10% of
the total coating thickness [14, 15].

2.4 Microscratch

The peeling behavior of the coatings was evaluated using a linear displacement
microscratch tester (Anton Paar Micro Combo Tester, Austria). A Rockwell diamond
ball with a radius of 100 µm was used for the nanoscratch test to determine the film
adhesion toughness and failure mode. For each measurement, the applied load was
progressive from 50 mN to 2 N at a sliding speed of 2 mm/min and sliding length of
1 mm. All these experiments were carried out in an ambient atmospheric condition.

2.5 Wear

The tribological properties which represent the macroscratch of the thin-film coat-
ings were achieved on a Rtec Universal ball-on-disk tribometer at ambient condi-
tion. (temperature, pressure and humidity). An E521000 alloy steel grade 25 ball
(6.350 mm in diameter) was used as the counter body while the surface of the thin-
film coating serves as the disk. The movement of the motion is translational with
maximum sliding displacement of 1 mm. The sliding velocity was kept constant at
1 mm/s for 2 min at normal load of 10 N for all the experiments. The friction coef-
ficient was monitored throughout the experiment by a linear variable displacement
transducer and recorded on a data acquisition computer attached to the tribometer.
After completion of the wear, the wear tracks were examined on the microscope for
further analysis of the surface deformation and failure. All process parameters were
repeated three times and the average values were reported.
Evaluation of Process Parameters Influence on the Mechanical … 149

3 Result and Discussion

3.1 Analysis of the Signal-to-Noise (S/N) Ratio for Young’s

Modulus and Hardness

Young’s modulus and hardness of the thin-film coating were characterized using
an experimental approach and the optimization of measured response was analyzed
using signal-to-noise ratio. The signal-to-noise (S/N) ratio is a loss function used for
determining the deviation between the experimental values and desired. S/N ratio
is divided into three distinctive quality characteristics for analysis namely lower the
better, nominal the better and higher the better. The S/N ratio result of the hardness
and young modulus is shown in Table 3. The higher the better equation was used for
the S/N ratio of the hardness. Higher hardness value is a sign of improvement in the
surface quality and improves the resistance to wear and corrosion, thereby reducing
the loss of material due to contact with other surfaces and enhance the overall quality.
Evaluation of young modulus gives in-depth behavior of specific properties of the
coating such as flexibility, bond strength, fatigue life, wear, fracture toughness and
residual stress [16, 17]. Low Young’s modulus yields a more flexible coating behavior
and improves its thermal stress resistance [17]. Higher Young’s modulus results in
high wear rate and short fatigue life. For this reason, the lower the better equation
was used for the Young modulus S/N ratio. The average values of the hardness and
Young’s modulus were calculated after the sputtering deposition and were found
to be 4.54 GPa and 121.32 GPa, respectively. Similar calculation was done for the
S/N ratio of the hardness and Young’s modulus and found to be 12.404 dB and

Table 3 S/N ratio analysis of the hardness and Young’s modulus

Experiment Control factors Hardness S/N Young’s S/N ratio
no. (GPa) ratio for modulus for
hardness E (GPa) Young’s
modulus
A RF B C
power sputtering temperature
(W) time (°C)
(Hrs)
L1 150 2.0 80 4.41 12.889 148.54 −43.4369
L2 150 2.5 90 5.00 13.970 134.19 −42.5544
L3 150 3.0 100 6.51 16.272 135.54 −42.6413
L4 200 2.0 90 2.38 7.532 96.67 −39.7058
L5 200 2.5 100 2.57 8.199 90.79 −39.1608
L6 200 3.0 80 2.37 17.605 89.13 −39.0005
L7 250 2.0 100 7.59 7.495 144.39 −43.1907
L8 250 2.5 80 4.09 12.234 113.21 −41.0777
L9 250 3.0 90 5.91 15.432 139.46 −42.8890
150 O. O. Abegunde et al.

Table 4 Response table for signal-to-noise ratios for hardness and Young’s modulus
Levels Control factors
Hardness (H) Young’s modulus (E)
A B C A B C
Level 1 14.380 12.675 10.873 −42.88 −42.11 −41.17
Level 2 7.742 11.471 12.314 −39.29 −40.93 −41.72
Level 3 15.090 13.066 14.025 −42.39 −41.51 −41.66
Delta 7.349 1.595 3.152 3.59 1.18 0.54
The bold values are the output response with optimal performance

−41.52bB, respectively. The response table for the signal-to-noise ratio and level
values of the input or control factor is illustrated in Table 4. The significant effect of
the input or control factors (RF power, sputtering time, temperature) on the coating
hardness and Young’s modulus obtained from Taguchi analysis was presented. The
optimal sputtering process parameter levels that yield high hardness and low Young’s
modulus values can be easily obtained from the table. The highest S/N ratio of the
control factor at different levels was utilized to determine the optimal setting. From
Table 4, the levels and S/N ratios for the factors yielding optimal hardness values
are factor A (level 3, S/N = 15.090), factor B (level 3, S/N = 13.060) and factor C
(level 3, S/N = 14.025). On the other hand, the levels and S/N ratios for the factors
yielding optimal young modulus values are factor A (level 2, S/N = −39.29), factor
B (level 2, S/N = −40.93) and factor C (level 1, S/N = −41.17), respectively. This
implies that an optimum value for hardness was obtained at RF power of 250 W,
sputtering time 3 h and temperature 1000 C which is graphically illustrated in Fig. 2.
The optimum value for the young modulus was obtained at RF power of 200 W,
sputtering time of 2.5 h and temperature of 800 C graphically presented in Fig. 3.

3.2 Analysis of ANOVA for Hardness and Young Modulus

ANOVA is a statistical technique utilized to study the interaction and contribution of

each control parameter in the design matrix. ANOVA was used to analyze the effect
of RF power, sputtering time and temperature on the behavior of the hardness and
Young’s modulus. Table 5 shows the ANOVA analysis result, and it was calculated
at a 5% significance level and 95% confidence level. The level of significance of
the control or input factors was determined by comparing the F-ratio values and
the percentage contribution rate which depicts the degree of influence each control
parameter on the response is also calculated. From the table, the contribution rate
of A, B and C factors on the hardness are 70.4%, 6.7%, and 19.7%, respectively.
This shows that factor A (RF power) has a major predominant effect on the outcome
of the hardness properties of the coating. A similar trend was also observed for the
young modulus of the coating. The contribution rate for factor A is 84.4%, while
Evaluation of Process Parameters Influence on the Mechanical … 151

Main Effects Plot for SN ratios

Data Means
RF Power Sputtering Time Temperature
16

15

14
Mean of SN ratios

13

12

11

10

7
150 200 250 2.0 2.5 3.0 80 90 100

Signal-to-noise: Larger is better

Fig. 2 S/N ratio graph of the hardness response

Main Effects Plot for SN ratios

Data Means
RF Power Sputtering Time Temperature
-39

-40
Mean of SN ratios

-41

-42

-43
150 200 250 2.0 2.5 3.0 80 90 100

Signal-to-noise: Smaller is better

Fig. 3 S/N ratio graph of the Young’s modulus response

152 O. O. Abegunde et al.

Table 5 ANOVA analysis of the hardness and Young’s modulus

Variance The degree of Sum of Mean square F-ratio Contribution
source freedom squares (SS) (MS) rate (%)
(DoF)
Hardness
A 2 20.2469 10.1234 23.19 70.4
B 2 1.9293 0.9646 2.11 6.7
C 2 5.6579 2.8289 6.20 19.7
Error 2 0.9126 0.4563 – 3.2
Total 8 28.7466 – – 100
Young modulus
A 2 3892.90 1946.45 20.22 84.4
B 2 440.51 220.26 2.29 9.5
C 2 85.74 42.87 0.45 1.9
Error 2 192.57 96.29 – 4.2
Total 8 4611.73 – – 100
The bold values are the output response with optimal performance

factor B and C contribute 9.5% and 1.9%, respectively, to the output response. The
percentage of errors for both ANOVA for hardness and Young’s modulus is 3.2%
and 4.2%, respectively.

4 Tribology

The coefficients of friction (COF) for the microscratch and wear are presented in
Table 6. The range for the friction coefficient of scratch is between 0.031 and 0.075,
while the coefficient of friction range for the wear test is between 0.447 and 0.496.
It is noticeable that there was a substantial improvement in the COF for both wear
and scratch with respect to the process parameters.
Figure 4 shows some selection of the wear scar for the microscratch result micro-
graph obtained. No obvious crack and delamination of the film were noticed around
the crack region. These micrographs revealed good bonding between the film and
the substrate.

5 Conclusion

From the Taguchi analysis of the results using signal-to-noise analysis and ANOVA,
it can be concluded that the RF power is the most effective control factor on both the
Table 6 Friction coefficient for wear and scratch tests
Experimental no L1 L2 L3 L4 L5 L6 L7 L8 L9
Coefficient of friction Scratch 0.031 0.040 0.048 0.071 0.048 0.075 0.036 0.047 0.069
Wear 0.447 0.462 0.484 0.471 0.465 0.496 0.469 0.480 0.462
Evaluation of Process Parameters Influence on the Mechanical …
153
154 O. O. Abegunde et al.

Fig. 4 Macrograph of some selected scratch surface for sample L1, L4 and L7

hardness and Young’s modulus. The coefficients of friction for both microscratch
and wear show considerable improvement, and the process parameters prove to be
major determinant factors in the response outcome.

Acknowledgements The authors will like to acknowledge the Council for Scientific and Industrial
Research (CSIR) for bursary awarded to the first Author.

References

1. Halling J (1986) The tribology of surface coatings, particularly ceramics. Proc Inst Mech Eng
Part C 200(1):31–40
2. Holmberg K, Mathews A (1994) Coatings tribology: a concept, critical aspects and future
directions. Thin Solid Films 253(1–2):173–178
3. Abegunde OO et al (2019) Overview of thin film deposition techniques. AIMS Mater Sci (-
2):174
4. Martin P (2010) Deposition technologies: an overview. In: Handbook of deposition technolo-
gies for films and coatings, 3rd Edn. Elsevier, Oxford
5. Martin PM (2009) Handbook of deposition technologies for films and coatings: science,
applications and technology, William Andrew
6. Wasa K, Hayakawa S (1992) Handbook of sputter deposition technology
7. Seshan K (2012) Handbook of thin film deposition. William Andrew
8. Alexeeva OK, Fateev VN (2016) Application of the magnetron sputtering for nanostructured
electrocatalysts synthesis. Int J Hydrogen Energy, 41(5):3373–3386, 2016
9. Pethe SA et al (2012) Effect of sputtering process parameters on film properties of molybdenum
back contact. Solar Energy Mater Solar Cells 100:1–5
10. Dibyendu R, Shampa A (2013) Effect of sputtering process parameters on the magnetron
sputtered Ni-Ti-Cu thin films. Int J Curr Res 5(01):075–079
11. Oliver WC, Pharr GM (1992) An improved technique for determining hardness and elastic
modulus using load and displacement sensing indentation experiments. J Mater Res 7(6):1564–
1583
12. Oliver WC, Pharr GM (2004) Measurement of hardness and elastic modulus by instrumented
indentation: advances in understanding and refinements to methodology. J Mater Res 19(1):3–
20
13. Fang T, Jian S, Chuu D (2004) Nanomechanical properties of TiC, TiN and TiCN thin films
using scanning probe microscopy and nanoindentation. Appl Surf Sci 228(1–4):365–372
Evaluation of Process Parameters Influence on the Mechanical … 155

14. Bec S et al (1996) Improvements in the indentation method with a surface force apparatus.
Philos Mag A 74(5):1061–1072
15. Mani A et al (2005) Effects of residual stress on the mechanical and structural properties of
TiC thin films grown by RF sputtering. Surf Coat Technol 194(2–3):190–195
16. Janos B, Lugscheider E, Remer P (1999) Effect of thermal aging on the erosion resistance of
air plasma sprayed zirconia thermal barrier coating. Surf Coat Technol 113(3):278–285
17. Karthikeyan S, Balasubramanian V, Rajendran R (2014) Developing empirical relationships to
estimate porosity and microhardness of plasma-sprayed YSZ coatings. Ceram Int 40(2):3171–
3183
Effect of Compaction Loads
in Machining of Short Carbon
Fiber-Reinforced Aluminum Composite

S. Mohanasundaram, S. J. Vijay, Rajakumar S. Rai, I. Kantharaj,

A. S. Melwyn, and S. Theophilus

Abstract Carbon fibers are preferred as reinforcements in metal matrix composites

for their mechanical and metallurgical properties. In the present study, short carbon
fiber-reinforced aluminum composites are fabricated with 2% reinforcement in pure
aluminum powders using a powder metallurgy process. The blended carbon fibers
and aluminum powders are compacted for different loading conditions such as 250,
300 and 350 kN followed by sintering at 580 °C. Abrasive jet machining trials are
conducted on all combinations, by varying the machining parameters. Carbon fibers
are found to be well dispersed in the aluminum matrix when examined through SEM.
The bonding between the carbon fibers and aluminum is also found to be without
any flaw. The mechanical properties showed improvements in wear resistance and
hardness of composites when compared to pure aluminum. The machining parame-
ters are further studied to obtain improved surface finish and minimize delamination
of carbon fibers from the metal matrix.

Keywords Carbon fiber · Aluminum · Sintering · Metal matrix composites ·

Abrasive jet machining

1 Introduction

Aluminum matrix composite materials exhibit great potential in several manufac-

turing industries like aircraft, aerospace, automobiles and various other fields. The
reinforcements in operating temperature should be both stable and non-reactive [1].
Reinforcements can be in the form of short fibers, continuous fibers and particles.

S. Mohanasundaram (B) · S. J. Vijay · R. S. Rai · I. Kantharaj · A. S. Melwyn · S. Theophilus

Department of Mechanical Engineering, Karunya Institute of Technology and Sciences,
Coimbatore 641 114, India
e-mail: johnmohana@yahoo.com
S. J. Vijay
e-mail: vijayjoseph.2001@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 157

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_15
158 S. Mohanasundaram et al.

Aluminum properties are found to be reduced significantly at relatively low tem-

peratures, which subsequently limits its functionality in high–precision mechani-
cal components [2]. Fibers play an important role as reinforcements, fulfilling the
requirements and distributing strength to the constituent matrix, influence their prop-
erties as desired and improve them [1]. Carbon fibers are used to produce materials
that enhance performance in applications such as nanoelectronics, structural and
medical applications. Because of its lightweight, mechanical properties and ther-
mal properties, reinforced carbon fiber (CF) composites were extensively used as
industrial materials [3]. Integrating CF into metal matrices considerably enhances
the hardness, tensile strength, modulus of elasticity and other mechanical properties
such as friction coefficient, wear resistance, corrosion and fatigue strength. They can
also be customized to the corresponding application requirements in metal matrix
composites [4, 5]. CF/Al composites were used in the waveguide mast of the Hubble
Space Telescope antenna [6]. The properties of MMCs are related to matrix prop-
erties, especially at low fiber content or when short fibers or particles are used as
reinforcements instead of continuous fibers [7].
Solid-state processes like powder metallurgy are used to fabricate short fiber-
reinforced aluminum alloys [8] and liquid-state processes such as compo casting [9]
and pressurized liquid-metal infiltration [10]. The liquid processes include melting
and casting, are unable to provide good matrix reinforced homogeneity due to the
high interface of energy between molten metal and reinforcements. Homogeneity
solid-state or mechanical alloy route reinforcements can be produced, clustering of
particle reinforcements. But it is a relatively expensive process [11].
Among the non-traditional processes, abrasive water jet (AWJ) machining has
provided exceptional machining ability for the most difficult materials in which tra-
ditional methods are technically or economically impossible. AWJ finds application
for a wide range of metals and non-metals, concrete, ceramics, metal matrix, lam-
inates and fiber-reinforced composites [12]. The advantages of abrasive water jet
cutting include simple fixtures required, high machining versatility, negligible heat
affected zone, small forces for cutting composite materials, etc. [12, 13].
In the present study, the machining qualities of short carbon fiber-reinforced Al
matrix materials prepared using cold pressing and powder metallurgy by abrasive
water jet machining for two process parameters (standoff distance and transverse
speed) were investigated.

2 Experimental Procedure

2.1 Materials

Aluminum is light, non-magnetic, malleable and ductile. It is widely used in several

industrial applications where light and robust material is needed. Pure metal has
inadequate strength and soft in nature, nevertheless once alloyed with a small amount
Effect of Compaction Loads in Machining of Short Carbon Fiber … 159

of copper, magnesium, silicon, manganese or other elements, it results in a variety

of useful properties. Pure aluminum powder (99.5%) of 16 microns was used for the
fabrication of the composite.
Carbon fibers (CF) have the highest specific rigidity compared to ceramic rein-
forcements. They are also commercially available at less cost and are not hazard to
health. These properties make them to be used as reinforcement for light metals like
aluminum and its alloys. 1 mm chopped carbon fibers were used for the fabrication
of the composite.

2.2 Fabrication of SCF Reinforced al Composite

Ball milling is one of the techniques used to homogeneously blend the reinforcement
particles in the matrix phase [11]. Pure aluminum powder and short carbon fiber
(SCF) of 2 weight percentage were mixed using ball milling for an hour. The Al
matrix powder and SCF mixer are then compacted in the mild steel die, and it is
cold pressed with 250, 300 and 350 kN compressive loads using universal testing
machine. After cold pressing, sintering was carried out in an electric muffle furnace
with an inert atmosphere. The cold-pressed Al/SCF sample is kept inside the furnace
at 580 °C for 30 min and allowed to cool in the furnace. The sintered samples were
of 50 mm in diameter and 25 mm in height.

2.3 SCF/Al Composite Machining

Machining was carried out in a water jet machine at SBB Water Jet (STREAMLINE
SL-V E50V2), Coimbatore, with a 1.1 mm orifice. Experiments were performed at a
constant pressure of 3500 bar and constant mass flow rate of 700 g/min using abrasive
garnet (silicon sand of 100 mesh). The two-process parameters used to perform
the machining are standoff distance and transverse speed. They were considered
in abrasive water jet machining of SCF/Al composite specimens. The machining
process parameters and the compacting load variation are shown in Table 1.

Table 1 Process parameters and the levels for the experiment

Process parameter Unit Level 1 Level 2 Level 3
Compacting load kN 250 300 350
Standoff distance mm 1 1.5 2
Transverse speed mm/min 180 190 200
160 S. Mohanasundaram et al.

Table 2 Process parameter

Process parameter Unit Value
wear test
Sliding velocity m/sec 2
Load N 2
Sliding distance m 1500
Track diameter mm 120
Speed rpm 318.3
Time sec 750

2.4 Wear Test

Wear is the process of friction between two interacting surfaces in a confined envi-
ronment. Wear test was conducted in pin-on-disk test apparatus as per the ASTM
standard G99. The specimens were machined to a size of 10 mm×10 mm×25 mm.
The wear rate was calculated directly as the volume loss of the specimen for the
distance traveled. During the conduct of wear test, the specimen was forced against
the rotating hardened steel disk by applying the load. The test was carried out at
room temperature with the process parameters shown in Table 2.

Sliding velocity = (π DN)/60 (1)

Sliding distance = Sliding Velocity × Time (2)

Wear rate = Volume loss/Sliding distance (3)

2.5 Design of Experiments

Design of experiments (DOE) is the process of carrying out experiments consider-

ing the process parameters at various levels. To evaluate the performance, quality
and cost of experimental design, Taguchi’s method provides a simple, efficient and
systematic approach for an optimal DOE [14]. Based on Taguchi’s method, DOE
with three factors and three levels L9 orthogonal arrays table was selected for the
experimentation as shown in Table 3.
Effect of Compaction Loads in Machining of Short Carbon Fiber … 161

Table 3 Orthogonal array for experiments

S. No Compacting load (kN) Standoff distance (mm) Transverse speed (mm/min)
1 250 1 180
2 250 1.5 190
3 250 2 200
4 300 1 190
5 300 1.5 200
6 300 2 180
7 350 1 200
8 350 1.5 180
9 350 2 190

2.6 Surface Roughness

Surface roughness is the degree of irregularities on the surface which is of three

components, i.e., roughness, waviness and form [15]. Surface roughness is measured
by Mitutoyo make surface tester. In order to reduce the measurement error, three sets
of readings were observed for each sample, and the mean value is taken for SR.

3 Results and Discussion

The 9-run machining was carried out in abrasive water jet machine as per combination
of factors based on the DOE, and the composite’s surface roughness, hardness and
wear rate are measured as in Table 4.

Table 4 Experimental results

S. Compacting Standoff Transverse Surface Hardness (HV) Wear
No load (kN) distance (mm) speed (mm/min) roughness Ra (µm) rate
mm3 /m
1 250 1 180 1.95 20.6 0.080
2 250 1.5 190 2.383
3 250 2 200 2.755
4 300 1 190 2.559 24.6 0.033
5 300 1.5 200 2.712
6 300 2 180 2.327
7 350 1 200 3.32 24.8 0.020
8 350 1.5 180 3.179
9 350 2 190 3.256
162 S. Mohanasundaram et al.

The powder mixer was compacted and sintered, and then, it was kept in the furnace
to maintain the temperature. The sintered composites with various compacting load
have shown an increase in hardness (Fig. 1). The hardness value for the specimen
compacted with 300 kN and specimen compacted with 350 kN are almost similar.
Hence, it may be concluded that increasing the compacting load over 300 kN is
insignificant with respect to the hardness values.
Surface morphology of the sintered composite for various compacting loads is
shown in Fig. 2. The microstructure for the specimen with 250 kN is found to be more
porous than 300 and 350 kN. Hence, the specimen compacted with 250 kN is not as
dense as the other specimens. As the compacting loads are gradually applied to the
powdered mixer, the same leads to the densification of the specimen which reduces
the porosity and increases its hardness. The surface of the specimen compacted with
300 kN (Fig. 2b) and 350 kN (Fig. 2c) is found to have a similar textured surface.
Hence for the higher loads, the change in hardness value is negligible.
Wear tests were conducted with a load of 2 N, by rubbing the sintered composite
specimens. Sliding velocity was kept as constant of 2 m/sec. Wear test results show
the effect of the compacting load on the wear rate. The specimen compacted for the
load of 250 kN displayed a higher wear rate than 300 and 350 kN are shown in Fig. 3.
It shows that the increase in compacting load reduces the wear rate.
The hardness of the composites influenced the resistance to friction and erosion.
Cavities are formed along the composites’ surface due to carbon fiber particles pulling
out from the composites, which in turn increases the surface roughness. The surface
roughness value of the composite increases with an increase in transverse speed from
180 mm/min to 200 mm/min as shown in Fig. 4.
It is observed that material erodes at a higher rate as the transverse speed increases
are shown in Fig. 5. Surface roughness of composite specimens for various standoff
distances. At a load of 300 kN, as the standoff distance increases there is a gradual
reduction in the surface roughness value owing to the porosity of the specimen. Since
the compaction id good at a load of 350 kN the surface roughness value varies linearly
as the specimen is less porous.

Fig. 1 Hardness of composite specimens for different compacting load

Effect of Compaction Loads in Machining of Short Carbon Fiber … 163

Fig. 2 Surface morphology of sintered composite (a) 250 kN compacting load, (b) 300 kN
compacting load and (c) 350 kN compacting load

Fig. 3 Wear rate of specimens for various compacting loads

4 Conclusion

The Al/SCF powder and fiber mixer were compacted for different loads and sintered
at 580 °C. The sintered specimens were machined in abrasive water jet machine for
164 S. Mohanasundaram et al.

Fig. 4 Surface roughness of

composite specimens for
various transverse speeds

Fig. 5 Surface roughness of

composite specimens for
various standoff distances

the various transverse speeds and standoff distances. The following conclusions were
made.
• The sintered composites with various compacting load increased the hardness of
the specimen. For the higher compacting loads, the change in hardness value is
very minimum. As the compacting loads increased, the specimen became denser.
It is found that higher compacting loads play a significant role in enhancing
hardness.
• The wear rate is high for the lower compacting loads and vice versa. Increase in
hardness results in the reduction of wear rate.
• The surface roughness value of the composite increases with an increase in trans-
verse speed. Along the surface, cavities are formed due to pull out of carbon fiber
particles from the composites due to which the surface roughness is increased.
Effect of Compaction Loads in Machining of Short Carbon Fiber … 165

References

1. Vijaya Ramnath B, Elanchezhian C, Annamalai RM, Aravind S, Sri Ananda Atreya T,

Vignesh V, Subramanian C (2014) Aluminium metal matrix composites—a review. RAMS
38(2014):55–60
2. Shirvanimoghaddam K et al (2017) Carbon fiber reinforced metal matrix composites:
fabrication processes and properties. Compos Part Appl Sci Manuf 92:70–96
3. Surappa MK (2003) Aluminium matrix composites: Challenges and opportunities. Sadhana
28(1):319–334
4. Kelly A (2006) Composite materials after seventy years. J Mater Sci 41(3):905–912
5. Ramesh CS, Adarsha H, Pramod S, Khan Z (2013) Tribological characteristics of innovative
Al6061–carbon fiber rod metal matrix composites. Mater Des 50:597–605
6. Miracle DB (2005) Metal matrix composites—from science to technological significance.
Compos Sci Technol 65(15):2526–2540
7. Kavecký Š, Šebo P (1996) Short aluminosilicate fibre reinforced aluminium. J Mater Sci
31(3):757–764
8. Toth IJ, Brentnall WD, Menke GD (1974) Survey of aluminum matrix composites. Metallurgi-
cal Society of the American Institute of Mining, Metallurgical and Petroleum Engineers, Inc,
United States
9. Das S, Dan TK, Prasad SV, Rohatgi PK (1986) Aluminium alloy—rice husk ash particle
composites. J Mater Sci Lett 5(5):562–564
10. Oh S-Y, Cornie JA, Russell KC (1987) Particulate wetting and metal: ceramic interface
phenomena. Ceram Eng Sci Proc 8:912–936
11. Mendoza-Duarte JM, Estrada-Guel I, Carreño-Gallardo C, Martínez-Sánchez R (2015) Study
of Al composites prepared by high-energy ball milling; effect of processing conditions. J Alloys
Compd 643:S172–S177
12. Kovacevic R, Hashish M, Mohan R, Ramulu M, Kim TJ, Geskin ES (1997) State of the art of
research and development in abrasive waterjet machining. J Manuf Sci Eng 119(4B):776–785
13. Vikram G, Ramesh Babu N (2002) Modelling and analysis of abrasive water jet cut surface
topography. Int J Mach Tools Manuf 42(12):1345–1354
14. Davim JP (2003) Design of optimisation of cutting parameters for turning metal matrix
composites based on the orthogonal arrays. J Mater Process Technol 132(1):340–344
15. Choudhury SK, Chinchanikar S (2017) 1.3 finish machining of hardened steel. In: Hashmi M
(ed) Comprehensive materials finishing. Elsevier, Oxford, pp 47–92
Synthesis and Characterization of Al
7072-Al2O3 Metal Matrix Composites

G. Mallesh, R. Pavankumar, V. G. Pradeep Kumar, and L. Laxman Naik

Abstract Composite materials are the combinations of two or more materials which
are different in form and chemical composition. These composite are gradually gain-
ing more importance as a structural material in the present engineering design and
development activities because they offer very attractive mechanical properties such
as high strength-to-weight ratio higher thermal and corrosive resistance. A wide
range of metals and their alloys such as aluminium, titanium and magnesium are
extensively used in combination with different reinforcements such as SiC, Al2 O3 ,
TiC, TiO2 , TiB, ZrO2 and ZrB. Aluminium and its alloys are widely used to develop
lightweight material for construction, transportation, aerospace industries, marine,
automobile and similar engineering applications. In the present study, attempts are
made to synthesize Al 7072-Al2 O3 MMCs using stir casting process by varying the
weight percentage of reinforcement to investigate the mechanical behaviour as per
ASTM standards. EDX and XRD results confirm the presence of Al2 O3 particles in
the composite, and it is observed from the SEM images that alumina particles are
homogeneously distributed in the matrix. Further, results reveal that the mechanical
properties of Al 7072-Al2 O3 were enhanced with the addition of the reinforcement.

Keywords Aluminium Matrix Composites (AMC’s) · Stir Casting · Al2 O3 ·

Mechanical Properties · EDX and XRD

G. Mallesh (B) · V. G. P. Kumar

Department of Mechanical Engineering, Sri Jayachamarajendra College of Engineering, Mysuru,
India
e-mail: mallesh@sjce.ac.in
V. G. P. Kumar
e-mail: mepradeepvg@gmail.com
G. Mallesh · R. Pavankumar
JSS Science and Technology University, Mysuru, Karnataka, India
e-mail: pavanjce2016@gmail.com
L. L. Naik
Department of Mechanical Engineering, Malnad College of Engineering, Hassan, Karnataka, India
e-mail: llnmce@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 167

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_16
168 G. Mallesh et al.

1 Introduction

Nowadays, AMC’s are accepted as a pivotal member in many important industries.

Over the years, AMC’s have been accepted material for different engineering sectors.
Many researchers have reported the use of aluminium in automobile bodybuilding [1],
fabrication industries, aeronautical industries [2], shipbuilding and other structural
applications [3–5].
AMC’s reveal unique properties when reinforcements are added. Determining the
proper reinforcement is an essential task to get the required composite. Commonly
used reinforcements in different grades of aluminium are SiC, Al2 O3 , B4 C, TiC TiO2 ,
TiB2, etc. [6]. Application of AMCs are increasing due to stringent environmental
protocols, and hence, it is inevitable to develop new AMCs.
Auradi et al. [7] developed Al 6061-B4 C composites using pre-heated B4 C par-
ticles and K2 TiF6 flux to overcome the problem of agglomeration and clustering of
the particles for different wt% of the reinforcement particles. XRD, EDX and SEM
analyses were conducted to know the presence and distribution of the reinforcement
in the composites. It is evident from the results that B4 C particles were present and
uniformly distributed in the composites developed. Further, the addition of the B4 C
particles to the melt leads to a decrease in the grain size and hence improves the
mechanical properties.
Stalin et al. [8] synthesized LM4-B4 C (400 mesh) composites for 5, 10 and 15 wt%
of reinforcement by stir casting process to evaluate mechanical behaviour of MMCs.
SEM and XRD tests confirm the presence and uniform distribution of B4 C in the
matrix alloy. It is noticed that mechanical properties were improved with the addition
of 10 wt% of the reinforcement. It is also evident that absorption of energy found to
be increased by increasing percentage of reinforcement.
Kumar et al. [9] fabricated Al6061-SiC and Al7075-Al2 O3 composites using stir
casting route by maintaining a temperature of 720 °C with a stirring speed of 400 rpm
about 10 min to study the microstructure, mechanical and wear properties. It is
observed that the addition of the reinforcement in both the alloy increase mechanical,
tribological properties, but Al7075-Al2 O3 has better properties compared to Al 6061-
SiC.
Kaliselvan et al. [10] synthesized Al 6061-B4 C composites by adding K2 TiF6
flux using stir casting process with argon gas environment to study the mechanical
properties of the composites. It is noticed that the addition of K2 TiF6 results in good
wettability between the particles and matrix. Further, mechanical properties like
hardness increased from 51.3 to 80.8 HV and tensile strength enhanced from 185 to
215 MPa. Gopal Krishna et al. [11] developed AMCs using B4 C using stir casting
process to study the mechanical characteristics of the composites for different wt%
of the reinforcement as per ASTM standards. It is observed from the results that
B4 C particles homogeneously distributed in the matrix and perceived the presence
of particles in the matrix. Further, it is noticed that tensile strength and hardness of
the composites were enhanced with the addition of 12 and 8 wt. of the reinforcement,
respectively.
Synthesis and Characterization of Al 7072-Al2O3 Metal Matrix … 169

Ravi et al. [12] developed aluminium matrix composites using boron carbide
using centrifugal casting technique by varying the wt% of the reinforcement to study
the mechanical properties. It is evident from the results that tensile strength was
enhanced from 117 to 145 MPa, and hardness value increased from 62 to 68 HV.
Zhao et al. [13] developed AA 6061-B4 C composites using a friction stir pro-
cess to study the microstructure and mechanical properties. Results revealed that
with an increasing number of FSP passes, the dispersion of B4 C particles becomes
more uniform, further by way of four FSP passes, the expected AA6061/B4 C sur-
face composite layer is successfully fabricated with homogeneously distributed B4 C
particles.
Vijaya Ramnath et al. [14] proposed that Al6061 B4 C-SiC composite was pro-
duced with stir cast route for different particle size (viz. 37, 44, 63, 105, 250 µ)
by maintaining melt temperature at 800 °C during the process preheated B4 C par-
ticles and degassing agents C2 Cl6 were added into the vortex with a stirring speed
of 300 rpm. It was observed from the XRD analysis the B4 C particles are homoge-
neously dispersed in the matrix. Further, it is noticed that hardness of the composite
was maximum at 250 µ at 12 wt% and an increment in tensile strength at 105 µ for
8 wt% of B4 C.
From the available literature, it is absorbed that many researches are worked
towards the development of lightweight materials for engineering applications using
different reinforcements and found that the addition of the reinforcement enhances
the mechanical and tribological properties. Further noticed that limited work has
been done to study the mechanical characteristics of Al 7072-Al2 O3 AMCs. Hence
in this research work, attempts are made to develop Al 7072-Al2 O3 MMCs to study
the mechanical properties.

1.1 Reasons for Selecting the Matrix and Reinforcement

Material

Al 7072 alloy is an extensively used aluminium alloy which offers a range of good
properties [15]. Al 7072 (Fig. 1) based composites are high strength, good machin-
ability, low thermal and electrical resistivity and easily extractable in nature. Being
stiff and strong material, it finds many applications in different engineering fields
[16, 17].
Among various reinforcements, Al2 O3 (Fig. 2) is selected because of its stag-
gering properties like chemical and thermal stability, availability, relatively good
strength, high hardness strength, high electrical insulation, good corrosion resistance
and biocompatibility.
170 G. Mallesh et al.

Fig. 1 Wrought Aluminium

7072

Fig. 2 Al2 O3 Powder

2 Synthesis of Al 7072-Al2 O3 Composites

Several manufacturing techniques are used to synthesize metal matrix composites.

The most commonly used techniques are liquid, solid and gaseous state [18], and
these processes include squeeze casting, powder metallurgy, spray forming, diffusion
bonding, sinter forging, stir casting and in situ process.
Among these manufacturing processes, stir casting is generally accepted as a
promising route to manufacture metal matrix composites in large volumes with min-
imum cost. Hence in this research work, stir casting technique is used to develop Al
7072-Al2 O3 MMC’s.
Synthesis and Characterization of Al 7072-Al2O3 Metal Matrix … 171

2.1 Stir Casting

Stir casting process is an extensively used method to fabricate MMCs in liquid state
in which molten metal is mixed in the furnace by means of mechanical stirring [19].
It is noticed that almost all the composites were manufactured by this process using
up to 30% volume fraction of reinforcement [20, 21]. It is evident from the literature
that use of stir casting process enables better mixing of matrix and reinforcement in
the MMCs. Different components of stir casting process are shown in Figs. 3 and 4.
Al 7072 rods and 3, 6, 9 and 12 wt% of Al2 O3 are weighed as per the wt% of the
reinforcement and fed into the graphite crucible and heated using an electrical furnace
up to 850 °C and coating was applied inside the crucible to avoid contamination.

Fig. 3 Components of Stir Casting Process

Fig. 4 Stir Casting Process Set-up

172 G. Mallesh et al.

Halide salt, hexaflurotitanate is added to improve wettability between Al2 O3 particles

and Al 7072. Similarly, hexachloroethane (C2 Cl6 ) tablets were fed into the matrix
to remove hydrogen gas in the melt. During stirring, preheated Al2 O3 is gradually
poured into the molten metal and stirring was continued about 10–12 min at the rate
of 250 rpm. Temperature of the molten metal is monitored and regulated by digital
temperature controller installed in the machine. Slag formed with the addition of
hexaflurotitanate was removed from the crucible shown in Fig. 5. The melt was
poured into a preheated (200 °C) cast iron moulds to get the rectangular plates and
circular rods shown in Fig. 6.
After solidification process castings were obtained as shown in Fig. 7 and machin-

Fig. 5 Slag Removal Process

Fig. 6 Pouring of Molten Metal into the Cast iron Moulds

Synthesis and Characterization of Al 7072-Al2O3 Metal Matrix … 173

Fig. 7 Al 7072-different wt.% of Al2 O3 MMC’s

ing of different wt% of Al2 O3 composites using wire cut EDM, CNC turning and
milling was carried out as per different ASTM standards to conduct mechanical tests.

2.2 EDX, XRD and Microstructure Characteristics of Al

7072-Al2 O3 Composites

EDS is an analytical tool used to know the presence of different elements in the
specimen using X-ray spectrum. To stimulate the emission characteristic of X-rays
from a specimen, a high-energy beam of electrons or protons or a beam of X-rays
was focused on the sample being studied.
In the present research, EDS analyses of Al 7072-Al2 O3 composites by varying the
wt% of Al2 O3 were conducted to know the presence of elements in the composites.
The EDS patterns of Al 7072-Al2 O3 with different wt% of Al2 O3 are shown in
Fig. 8(a–e). It is evident from the plots that aluminium exhibits high intensity peaks
which indicate that aluminium is predominant because it is chosen as base alloy.
Further, it is found that low intensity peaks of oxygen were identified which shows
the presence of Al2 O3 reinforcement [22].
The X-ray diffraction patterns (XRD) of Al 7072-Al2 O3 composites were obtained
using Panalytical X-ray diffractometer having 0.15–0.40 nm inter-atomic distances
with photon energies between 3 and 8 keV. Peak values were collected over the 2θ
range of 10–800 with a step size of 0.0170° and step time of 20.3142 s. The crystalline
size measurements were carried out using Debye–Scherrer equation D hkl = β0.98λ cos θ
[23]. It is observed from the XRD patterns (Fig. 9a–e) that all the composites
developed confirm the formation of Al2 O3 particles.
The SEM micrographs of Al 7072-Al2 O3 composites are presented in Fig. 10a–e.
It is noticed from the micrographs that all the castings are free from defects such
as porosity, shrinkages and slag inclusions which highlights the quality of castings
and Al2 O3 particles are distributed homogeneously in the composites which is an
essential requirement to achieve better mechanical properties.
174 G. Mallesh et al.

(a) Unreinforced (b) 3 %Al2O3

(c) 6 % Al2O3 (d) 9 % Al2O3

(e) 12 % Al2O3

Fig. 8 EDS Patterns of Al 7072 and different wt% of Al2 O3 Reinforcement

3 ASTM Standard Mechanical Testing Methods

3.1 Tensile Strength

Tensile strength is the ability or capacity of the material to expand prior to failure.
The tensile tests were conducted on computerized universal testing machine model
TUE400(C) as per ASTM E8M-13a standard shown in Fig. 11.
Synthesis and Characterization of Al 7072-Al2O3 Metal Matrix … 175

(a) Unreinforced

(b) 3 % Al2o3

(c) 6 % Al2o3

(d) 9 % Al2o3

(e) 12 % Al2o3

Fig. 9 XRD Plots of Al 7075 and different wt% of Al2 O3 Reinforcement

176 G. Mallesh et al.

(a) Unreinforced (b) 3 %Al2O3

(c) 6 % Al2O3 (d) 9 % Al2O3

(e) 12 % Al2O3

Fig. 10 SEM Images of Al 7075 and different wt% of Al2 O3 Reinforcement

3.2 Impact Strength

Ability of the material to resist a sudden/shock loads is known as impact strength. Izod
or Charpy impact tests are used to measure the impact energy required to fracture
the specimen. In this study, ASTM E23 standard is used to determine the impact
resistance of Al 7072 reinforced with different wt% of Al2 O3 composite system.
The standard size of impact test specimen is as shown in Fig. 12.
Synthesis and Characterization of Al 7072-Al2O3 Metal Matrix … 177

Fig. 11 ASTM E8M-13a

Tensile Test Specimens

Fig. 12 ASTM E-23 Impact

Test Specimens
178 G. Mallesh et al.

Fig. 13 ASTM E-384 Hardness Test Specimens

3.3 Hardness

Hardness is the property of a material that resist plastic deformation by indentation.

Vickers hardness test (VHN) is the most versatile since only one indenter is used for
a range of loads for different materials. The tests were conducted for different wt%
of Al2 O3 as per ASTM E-384 standard shown in Fig. 13.

4 Results and Discussions

4.1 Tensile Strength

Tensile strength of the composites was tested using universal testing machine, and
it is evident from Fig. 14 that an increase in the wt% of the reinforcement increases
the tensile strength of the specimen. Unreinforced alloy possesses tensile strength
of 51.28 MPa and addition of 3 wt% Al2 O3 tensile strength improved by 10%, i.e.
56.64 MPa. Further increase in the 12 wt% of Al2 O3 , tensile strength reaches its peak

Fig. 14 Tensile Strength of Al 7072-Al2 O3 Composite System

Synthesis and Characterization of Al 7072-Al2O3 Metal Matrix … 179

35
29
30 26
22
Energy, Joules
25
19
20 17
15
10
5
0
Al 7072 3% 6% 9% 12%
Wt. % of Al2O3

Fig. 15 Impact Strength of Al 7072-Al2 O3 Composite System

value of 79.28 MPa, which is approx. 70% improvement [14]. Reason for this expo-
nential improvement in tensile strength is because of grain boundary strengthening
where alumina acts as a grain refiner and decreases the size of the alloy, uniform
distribution of the particle in the matrix, reduction in grain size leads to enhance the
tensile strength of Al 7072-Al2 O3 composite.
It is also evident that increase in the wt% of the reinforcement there was noticeable
decrease in the percentage of elongation due to the addition of hard particles alumina
which leads to the formation of brittle phase in the composite in turn enhances the
load carrying capacity.

4.2 Impact Strength

Izod impact tests were conducted as per ASTM standard to know the impact strength
of the composites. It is observed from the results that a gradual improvement in
the impact strength of the composite as percentage of reinforcement increases. The
improvement in the impact strength is due to the resistance in flowing energy of the
reinforcing particles at the molecule level. It is observed from Fig. 15 that impact
strength of the composite developed is enhanced by 41% compared to base alloy.

4.3 Hardness

Vickers hardness tests were conducted as per ASTM standard to know the hardness
number of Al 7072-Al2 O3 composites. Figure 16 represents hardness number of
3,6,9 and 12 wt% of Al2 O3 reinforcement. It is observed from the results that the
addition of alumina enhances the hardness of the composite. Presence of Al2 O3
particles in the composite helps to resist the movement of dislocation within the
180 G. Mallesh et al.

Fig. 16 Hardness Test Results

matrix, minimization of inter-atomic distance and better bonding between the matrix
and the reinforcement.

5 Conclusions

• Al 7072 reinforced with different wt% of Al2 O3 AMC’s were synthesized using
stir casting process.
• EDX and XRD results reveals the presence of the reinforcement in the composite
developed and SEM images shows the clear and uniform distribution of Al2 O3 in
the matrix alloy.
• Mechanical tests on Al 7072-Al2 O3 AMC’s were carried out using different
ASTM standards to evaluate mechanical behaviour.
• Mechanical properties such as tensile strength, impact strength and hardness of
Al 7072-Al2 O3 composites were enhanced with the addition of reinforcement.

References

1. Miller WS, Zhuang L, Bottema J, Wittebrood AJ, De Smet P, Haszler A, Vieregge A (2000)
Recent development in aluminium alloys for the automotive industry. Mater Sci Eng A
280(1):37–49
2. Heinz A, Haszler A, Keidel C, Moldenhauer S, Benedictus R, Miller WS (2000) Recent
development in aluminium alloys for aerospace applications. Mater Sci Eng A 280(1):102–107
3. Vijaya RB, Elanchezhian C, Annamalai RM, Aravind S, Sri Ananda Atreya T, Vignesh V,
Subramanian C (2014) Aluminium metal matrix composites-a review. Rev Adv Mater Sci
38(2014):55–60
4. Shirvani MK, Abdizadehb H, Baharvandic HR. Mechanical and structural behavior of alu-
minium matrix composite reinforced by B4 C under different processing temperature and
volume fraction of reinforcement
Synthesis and Characterization of Al 7072-Al2O3 Metal Matrix … 181

5. Bindumadhavan PN, Chia TK, Chandrasekaran M, Wah HK, Loh NL, Prabhakar O (2001)
Effect of particle-porosity clusters on tribological behavior of cast aluminium alloy A356-SiCp
metal matrix composites. Mater Sci Eng A 315(1–2):217–226
6. Hong SJ, Kim HM, Chun BS (2003) Effect of clustering on the mechanical properties of
SiC particulate-reinforced aluminium alloy 2024 metal matrix composites. Mater Sci Eng A
347(1–2):198–204
7. Auradi V, Rajesh GL, Kori SA (2014) Processing of B 4 C particulate reinforced 6061 aluminum
matrix composites by melt stirring involving two-step addition. Procedia Mater Sci 6:1068–
1076
8. Stalin B, Murugan C (2016) Evaluation of mechanical behavior of aluminium alloy boron
carbide MMC. In: International conference on emerging engineering trends and science
(ICEETS–2016), pp 2348–8360
9. Kumar GBV, Rao CSP, Selvaraj N, Bhagyashekar MS (2010) Studies on Al6061-SiC and
Al7075-Al2O3 metal matrix composites. J Miner Mater Charact Eng 9(01):43
10. Kalaiselvan K, Dinaharan I, Murugan N (2014) Characterization of friction stir welded boron
carbide particulate reinforced AA6061 aluminum alloy stir cast composite. Mater Des 55:176–
182
11. Gopal Krishna UB, Sreenivas Rao KV, Vasudeva B (2013) Effect of boron carbide rein-
forcement on aluminium matrix composites. Int J Metall Mater Sci Eng (IJMMSE):41–48.
ISSN
12. Ravi B, Balu Naik B, Udaya Prakash J (2015) Characterization of aluminium matrix composites
(AA6061/B4C) fabricated by stir casting technique. Mater Today: Proc 2(4–5): 2984–2990
13. Zhao Y, Huang X, Li Q, Huang J, Yan K (2015) Effect of friction stir processing with B 4 C
particles on the microstructure and mechanical properties of 6061 aluminum alloy. Int J Adv
Manufact Technol 78(9–12):1437–1443
14. Subramanya RP, Kesavan R, Vijaya Ramnath B (2018) Investigation of mechanical properties of
aluminium 6061-silicon carbide, boron carbide metal matrix composite. Silicon 10(2):495–502
15. Eliasson J, Sandström R (1995) Applications of aluminium matrix composites. In: Key
engineering materials. vol. 104, Trans Tech Publications, pp 3–36
16. Viala JC, Bouix J, Gonzalez G, Esnouf C (1997) Chemical reactivity of aluminium with boron
carbide. J Mater Sci 32(17):4559–4573
17. Shorowordi KM, Laoui T, Haseeb ASMA, Celis JP, Froyen L (2003) Microstructure and inter-
face characteristics of B4C, SiC and Al2 O3 reinforced Al matrix composites: a comparative
study. J Mater Process Technol 142(3):738–743
18. Kok M (2005) Production and mechanical properties of Al2 O3 particle-reinforced 2024
aluminium alloy composites. J Mater Process Technol 161(3):381–387
19. Gopalakrishnan S, Murugan N (2012) Production and wear characterisation of AA 6061 matrix
titanium carbide particulate reinforced composite by enhanced stir casting method. Compos B
Eng 43(2):302–308
20. Balasivanandha PS, Karunamoorthy L, Kathiresan S, Mohan B (2006) Influence of stirring
speed and stirring time on distribution of particles in cast metal matrix composite. J Mater
Process Technol 171(2):268–273
21. Reddy AC (2009) Mechanical properties and fracture behavior of 6061/SiCp metal matrix
composites fabricated by low pressure die casting process. J Manufact Technol Res 1(3/4):273–
286
22. Suresh S, Shenbaga Vinayaga Moorthi N, Vettivel SC, Selvakumar N (2014) Mechanical behav-
ior and wear prediction of stir cast Al–TiB2 composites using response surface methodology.
Mater Des 59:383–396
23. Kaviyarasu K, Anand P (2012) One pot synthesis and characterization of caesium doped Sn O2
Nano crystals via a hydrothermal process. J Mater Sci Technol 28:15–20
Erosion-wear Behaviour of 304 Stainless
Steel Reinforced with TiN at Elevated
Temperatures

Ramokone Marcia Mafafo, Babatunde Abiodun Obadele, Bruno Pilotti,

Walter Roberto Tuckart, and Peter Apata Olubambi

Abstract The erosion-wear behaviour of spark plasma-sintered austenitic stainless

steels matrix composite was investigated under high temperature conditions. Erosion-
wear behaviour of the composites was conducted at a constant velocity of 18 m/s
and impinging angle of 90° at different temperatures of 25, 400 and 600 °C. Hard
abrasive alumina particles with particle size of 40 µm were used as erodent. After
SPS, relative density of spark plasma-sintered samples decreased with increasing TiN
content. Furthermore, the results show a reduction in material loss with increased
hardness. However, there was a significant increase in material wastage as tem-
perature increases. Microstructures of the eroded samples indicate that the erosion
damage occurs mainly by plastic deformation and brittle fracture. Large fragments in
the form of wear debris, pitting, ploughing with slip on the side, misplaced material
and a few grooves are observed as well on the SEM images.

Keywords Erosion–corrosion · Spark plasma sintering · 304 stainless steel ·

Microstructure · Wear · Debris

1 Introduction

Austenitic stainless steels (ASSs) are extensively used in nuclear reactors, biomed-
ical implants, as well as in components for chemical and food industries. They are
the most popular corrosion-resistant materials used in various applications due to

R. M. Mafafo · B. A. Obadele (B) · P. A. Olubambi

Centre for Nanoengineering and Tribocorrosion, School of Mining, Metallurgy and Chemical
Engineering, University of Johannesburg, Doornfontein Campus, Johannesburg, South Africa
e-mail: obadele4@gmail.com
B. A. Obadele
Department of Chemical, Materials and Metallurgical Engineering, Botswana International
University of Science and Technology, Palapye, Botswana
B. Pilotti · W. R. Tuckart
Tribology Group, Universidad Nacional del Sur/CONICET, Av. Alem 1253, 8000 Bahía Blanca,
Argentina

© Springer Nature Singapore Pte Ltd. 2021 183

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_17
184 R. M. Mafafo et al.

their excellent corrosion resistance and good mechanical properties [1]. Apart from
applications in corrosive environment, high impact velocity of particles is frequently
encountered in components used in these industries. Examples are coal-fired boilers
used in the power generation industries [2]. Although austenitic stainless steels are
well known for their good corrosion properties, they are famous for poor mechanical
properties such as friction and wear. According to Sun and Bell [3], when ASSs are
in contact with each other or other materials, they experience severe metallic wear as
a result of adhesion junction and severe plastic deformation occurring between the
surfaces of the materials in contact with each other. Furthermore, ASS fails mainly
because of wear rather than corrosion-related material loss problems.
Wear can be defined as the loss of materials due to the mechanical action occurring
between two or more bodies in contact with each other [4]. Wear phenomena include
several mechanisms such as abrasive, adhesive, corrosive and erosion wear. Erosion-
wear mechanism plays an important role in determining the life span of components
used in several applications such as in mining, solids transportation in pipelines,
power generation and marine [5]. Erosion wear is a material deterioration process
due to impacting particles on solid surfaces.
Although there are literatures available on 304 steel reinforcement with ceramics
and nitrides, there is still a need to study the effect of these reinforcements at elevated
temperatures. This is important as there are very few literatures available on erosion-
wear or erosion-corrosion performance of spark plasma-sintered 304 stainless steel
composites at high temperatures. Therefore, this study focuses on the erosion-wear
behaviour of sintered 304 steel under high temperature conditions.

2 Experimental Procedures

Spark plasma-sintered compacts of 304 austenitic stainless steel (ASS) without and
with 2, 4 and 6% TiN were supplied by Centre for Nanoengineering and Tribo-
corrosion. The samples were sintered under high vacuum conditions using graphite
die and punch at an applied pressure of 50 MPa, sintering temperature of 1000 °C,
heating rate of 100 °C/min and holding time of 10 min. After sintering, the sam-
ples were sand-blasted to remove any excess graphite material on them. Thereafter,
the substrates were machined using the robofil 240-SL electrical discharge machine
(EDM) wire cut. The sintered samples were ground and polished to 3 µm diamond
finish. The samples were further polished with fumed silica using a polishing sus-
pension of 0.2 µm to achieve mirror-like surface. The morphology of the sintered
specimen before and after erosion-wear test was investigated using scanning elec-
tron microscope model TESCAN Vega 3 which is equipped with energy-dispersive
spectroscope. The Rigaku Ultima IV X-ray diffractometer XRD equipped with a
graphite-monochromatic Cu Kα radiation source at 40 kV and 30 mA was used to
perform the XRD analysis. Collection of diffractograms took place in the 2θ range
between 3° and 90° with a step size of 0.01 [6].
Erosion-wear Behaviour of 304 Stainless Steel Reinforced … 185

Table 1 Chemical
Erosion conditions
compositions of admixed
powders Erodent material Alumina particles
Particle size (µm) 45
Particle velocity (m/s) 18
Impact angle (°) 90
Sample temperature (°C) 25, 400, 600
Duration of impact (mins) 10

Erosion-wear testing was carried out using a high temperature erosion tester. The
erosion test rig used in the study was designed at the Universidad Nacional Del Sur
(UNS), Argentina. The equipment works similar with that of the ASTM G76-04
standards. Prior to erosion tests, specimens were cut into approximately 35 mm ×
1.5 mm × 1.5 mm dimension, ground and polished using sand papers 180, 220, 400
and 600 and cleaned in acetone for 2 min in an ultrasonic cleaner. The erosion tester
consists of particle feed inlet, air compressor and furnace. The alumina particles were
used as erodent material and were accelerated through the feed inlet by compressed
air at a constant speed of 18 m/s and this cause the abrasive particles to impact on the
sample. Erosion-wear tests were carried out at room temperature, 400 and 600 °C,
and the specimens were eroded for 10 min. Table 1 shows the operating solid particle
erosion-wear parameters used in the study.

3 Results and Discussion

The effects of temperatures (25, 400 and 600 °C) on the rate of erosion of 304
austenitic stainless steel without and with TiN additions at an impact angle of 90°
and an impact velocity of 18 m/s are presented in Fig. 1.
From the graphs, it could be observed that the erosion-wear behaviour of these
materials shows a similar behaviour. The graphs show that material mass loss
increases with increasing temperature and as a result, higher material loss is expe-
rienced at 600 °C while the least material loss is observed at room temperature.
Material loss increases as the temperature increases. This is observed up to a max-
imum temperature where a point of maximum material loss is achieved. The total
material loss between room temperature and 400 °C in all the samples is small in
the same order (in the range of 10–15 mg) while the total material loss between 400
and 600 °C in all the samples rises sharply (in the range of 90–40 mg). In this case,
not only erosion-wear occurs but also the combined effect of erosion and oxidation
takes place at elevated temperatures [7].
The combined effect of temperatures on material loss as the amount of TiN hard
phase is increased in the substrate is presented in Fig. 2. From the graph, it could
be observed that at 25 °C, the materials show almost similar weight loss between
28.9 and 31.4 mg. As the temperature increases, more material loss is experienced
186 R. M. Mafafo et al.

Fig. 1 Temperature versus weight loss of a 304 ASS b 304/2TiN c 304/4TiN and d 304/6TiN

85
304 ASS 2TiN 4TiN 6TiN
75

65
Weight loss (mg)

55

45

35

25

Temperature (oC)

Fig. 2 Combined effect of temperature on material loss

on 304/6TiN. This could be due to the formation of debris that could contain TiN
hard phase which can act as a third body. Since there is a high content of TiN in
304/6TiN, it could translate to higher TiN content as a third body. Therefore, this
may promote higher weight loss. On the other hand, there could be particle loosening
along the grain boundaries at high temperature and as a result, an increase in wear
phenomena on the eroded surface. At 600 °C, it could be observed that 304 ASS
Erosion-wear Behaviour of 304 Stainless Steel Reinforced … 187

with TiN additions display similar trend and closely related weight loss between 50
and 55.15 mg. However, for 304 ASS without TiN at 600 °C, it was observed that
weight loss was the highest (76 mg) compared to 304 ASS with TiN. This could be
due to sintered 304 ASS ductility behaviour as the temperature increase, thereby a
loss in hardness properties [7, 8].
Figure 3 further compares the effect of temperature on the hardness of the sin-
tered composites under erosion-wear test. Higher microhardness values are recorded
at room temperature. At higher temperatures (400 and 600 °C), all the samples
recorded a significant drop in microhardness values. According to studies reported
by Huttunen-Saarivirta et al. [7], when a material is subjected to high temperature
effect, it becomes soft and loses its hardness properties. These findings are in full
agreement with the results indicated in Fig. 3.
Figure 4 compares the effect of temperature (25, 400 and 600 °C) on the erosion-
wear behaviour of the composites. The rate of erosion wear indicates that there is
a proportional relationship between drop in microhardness values (shown in Fig. 3)
and the weight loss, as the percentage of TiN in the material increases.
This indicates that at low temperature, the sintered composites are less prone
erosion-wear. On the other hand, an increase in temperature raises the materials loss.
However, addition of TiN to 304 matrix reduced material loss at 600 °C from 76 mg
(for 304 ASS without TiN) to 55 mg. This shows a correlation between TiN content
in the metal matrix hardness values and weight loss.
Visual aid image of sintered composite surfaces before and after erosion-wear are
presented in Fig. 5. Figure 5a shows the material prior to erosion-wear tests while
Fig. 5b–d shows the materials after being subjected to erosion-wear at 25, 400 and
600 °C, respectively. In every case, erosion-wear region shows a circular shape at
the centre and spread towards the edge of the samples. Furthermore, three layers
could be identified: the first layer is at the centre, where most of the material is
removed; second layer where less material is removed and lastly, the third layer,

Fig. 3 Microhardness values of eroded 304 SS without and with TiN addition at different
temperatures
188 R. M. Mafafo et al.

Fig. 4 Comparison of material loss in spark plasma-sintered 304 stainless steel without and with
TiN

Fig. 5 Visual inspection of sintered specimen; a Before erosion-wear b Erosion-wear at 25 °C

c Erosion-wear at 400 °C and d Erosion-wear at 600 °C

where very small amount of material has been removed. This could be clearly seen
in Fig. 5d. This behaviour is also reported by Kumar and Kanwar [9] and Patel et al.
[10]. Similarly, Kumar and Kanwar [9] found that an erosion study carried out at an
impact angle of 90° results in circular shape on the region where erosion has taken
Erosion-wear Behaviour of 304 Stainless Steel Reinforced … 189

place while at an impact angle of 30°, an elliptical shape is observable on the surface
where erosion has occurred.
SEM micrographs of the eroded composites are displayed in Fig. 6 at 600 °C con-
dition. The SEM of eroded samples provides useful information about the mechanism
of erosion. According to Javaheri et al. [11], the mechanism at which erosion occurs
by solid particle is broadly divided into two mechanisms, namely “cutting” and “de-
formation”. On the damaged surfaces, pitting action is clearly observed. Materials
were predominantly damaged by brittle fracture and plastic deformation. Large frag-
ments of wear debris could be observed in all the samples and plastic deformation
in the form of ploughing, misplaced material, micro-cut and ploughing is present.
The results found in this study are in line with those conducted by Laguna-Camacho
et al. [12] and Qin et al. [13]. Sintered 304 (Fig. 6a) shows an unconventional wear
mechanism which was characterized by large fragments on the surface of the eroded
surface, and this is an indication that the removal mechanism is through brittle frac-
ture. Also, pitting behaviour was observed. Elsewhere, Bahri et al. [14] revealed that
materials from eroded surface of 304 L SS had been removed due to ploughing and
pitting. Patel et al. [10] observed that in a steady-state condition, a simultaneous
action of three phases occurs on the surface of 304 steel at different locations.
During the initial phase, a crater is formed by the impacting particles and the mate-
rial is displaced from the specimen. The second phase involves the material being
displaced due to subsequent impact and this result in lateral material displacement

Fig. 6 SEM micrographs of a 304 ASS b 304 ASS/2TiN c 304 ASS/4TiN and d 304 ASS/6TiN
at 600 °C
190 R. M. Mafafo et al.

accompanied by fracture. The third stage is experienced after few impacts where the
materials become strained and therefore removed from the surface by fracture. Qin
et al. [13] conducted a study on the erosion-wear and intergranular corrosion resis-
tance properties of AISI 304 L austenitic stainless steel after low-temperature plasma
nitriding. It was reported that the most erosion damage taking place on the surface of
the materials includes ploughing action, embedded particles, cutting grooves, large
fragments, pitting action and wear debris on the surface of the sample that was nitride
at 673 K.

4 Conclusion

In this study, solid particle erosion-wear behaviours of spark plasma-sintered 304

ASS, 304+2TiN, 304+4TiN and 304+6TiN were conducted at different temperature
(25, 400 and 600 °C) conditions at an impact angle of 90° and a constant impact
velocity of 18 m/s. We examined the effects of materials and temperatures on erosion-
wear. An increase in temperature resulted in decrease in hardness properties of the
composites. The erosion-wear rate indicated that there is a proportional relationship
between decreasing percentage of hardness and material loss. Macrographs of the
eroded specimens showed three layers in which the first (central) layer represents
high material loss, the second layer showed lesser amount of material wastage and
the third (outer) layer showed least amount of material. The SEM images indicated
that the erosion damage occurs mainly by plastic deformation and brittle fracture.
Large fragments in the form of wear debris, pitting, ploughing with slip on the side,
misplaced material and a few grooves are observed.

Acknowledgements The authors are grateful to Tribology Group, Universidad Nacional del Sur,
Bahía Blanca, Argentina for the use of equipment and facilities. The Institute of Nano Engineer-
ing Research (INER), Tshwane University of Technology, Pretoria, for the use of sintering and
characterization facilities.

References

1. Obadele BA, Andrews A, Shongwe MB, Olubambi PA (2016) Tribocorrosion behaviours of

AISI 310 and AISI 316 austenitic stainless steels in 3.5% NaCl solution. Mater Chem Phys
171:239–246
2. Samandi M, Shedden B, Smith D, Collins G, Hutchings R, Tendys J (1993) Microstructure,
corrosion and tribological behaviour of plasma immersion ion-implanted austenitic stainless
steel. Surf Coat Technol 59:261–266
3. Sun Y, Bell T (2002) Dry sliding wear resistance of low temperature plasma carburised
austenitic stainless steel. Wear 253:689–693
4. Bayer RG (2002) Tribological and wear testing. characterization of materials 1–16
5. Chawla K, Singh K, Saini N, Singh J (2013) Erosion wear behavior of chromium coated steel
304 and grey cast iron. Int J Innovative Res Sci Eng Technol 2:4567–4575
Erosion-wear Behaviour of 304 Stainless Steel Reinforced … 191

6. Dada OR (2015) Characterization and beneficiation of foundry greensand waste from ferrous
foundries. University of Johannesburg
7. Huttunen-Saarivirta E, Antonov M, Veinthal R, Tuiremo J, Mäkelä K, Siitonen P (2011) Influ-
ence of particle impact conditions and temperature on erosion–oxidation of steels at elevated
temperatures. Wear 272:159–175
8. dos Santos JF, Garzón CM, Tschiptschin AP (2004) Improvement of the cavitation erosion
resistance of an AISI 304 L austenitic stainless steel by high temperature gas nitriding. Mater
Sci Eng A 382:378–386
9. Kumar N, Kanwar R (2012) To study erosion behavior of Cr2 O3 Coating on SS-304 boiler
steel tubes in simulated coal fired boiler conditions
10. Patel M, Patel D, Sekar S, Tailor P, Ramana P (2016) Study of solid particle erosion behaviour
of SS 304 at room temperature. Procedia Technol 23:288–295
11. Javaheri V, Porter D, Kuokkala V-T (2018) Slurry erosion of steel–review of tests, mechanisms
and materials. Wear 408:248–273
12. Laguna-Camacho JR, Vite-Torres M, Gallardo-Hernández E, Vera-Cárdenas E (2013) Solid
particle erosion on different metallic materials. Tribology in Engineering. InTech
13. Qin X, Guo X, Lu J, Chen L, Qin J, Lu W (2017) Erosion-wear and intergranular corrosion
resistance properties of AISI 304 L austenitic stainless steel after low-temperature plasma
nitriding. J Alloy Compd 698:1094–1101
14. Bahri A, Guermazi N, Elleuch K, Ürgen M (2016) On the erosive wear of 304 L stainless steel
caused by olive seed particles impact: Modeling and experiments. Tribol Int 102:608–619
A Concise Review of Nano-enhanced
Phase Change Materials for Passive
Cooling Applications in Buildings

Chukwumaobi K. Oluah, Esther Titilayo Akinlabi, and Howard O. Njoku

Abstract Building energy consumption has been on a steady rise over the past
decade. This is due to the ever-growing human population and the effect of climate
change. Space conditioning accounts for over 40% of the final global energy pro-
duction. Thermal energy storage (TES) systems have shown to be quite effective in
reducing buildings energy demand while maintaining the desired comfort. It finds
a useful application in solar thermal systems as and heat recovery systems. Phase
change materials are used in building components for storing excess heat energy
for use when the need arises. The ability of PCMs to store great quantity of heat
in comparatively small volume makes them suitable for this application. Some of
the set backs encountered by using PCMs is its poor thermal conductivity and sub-
cooling characteristics. Nanoparticles have been found very useful in ameliorating
the thermal conductivity of PCMs thereby improving its general performance. This
review work brings to bear recent developments on the use of nanoenhanced PCMs
for passive cooling applications in buildings, and there are very little articles in lit-
erature that emphasizes the application of nanoPCMs in passive cooling systems.
We further discussed the physics of operation and phase transition behaviours with
mathematical representations. Some gaps currently encountered in the research of
nanoPCMs was also highlighted for further research purposes.

Keywords Phase change materials (PCMs) · Thermal energy storage (TES) ·

Nanomaterials · Passive cooling

1 Introduction

Space conditioning in buildings have been identified as the major consumer of the
final global energy production with respect to buildings. This pattern is the same
for both residential and commercial buildings. This figure was reported to be about

C. K. Oluah (B) · E. T. Akinlabi · H. O. Njoku

Department of Mechanical Engineering Sciences, University of Johannesburg, Johannesburg
2006, South Africa
e-mail: chukwumaobi.oluah@unn.edu.ng

© Springer Nature Singapore Pte Ltd. 2021 193

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_18
194 C. K. Oluah et al.

40% of annual global energy production by the international energy agency [1].
A greater majority of the global energy resource is fossil-based and comes with
increased CO2 production. The rapidly growing population of the world and the
effect of climate change account for the rise in energy demand. Some of the adverse
effects of global warmingas listed by the IPCC include but not restricted to the
following: a continuous rise in temperature, increased hurricane, more draught, and
heat waves etc. [2]. It is due to this cause that scientists have sought to reduce global
CO2 emission while maintaining reasonable thermal comfort. Over the past decades,
researches have studied various novel renewable approaches to meet the energy
requirements of residential or commercial buildings while keeping the environment
safe. Thermal energy storage (TES) is one way, that many researchers have proposed,
that can sufficiently reduce building’s energy demand. It finds a useful application in
solar thermal systems as well as heat recovery systems [3]. Thermal energy storage
(TES) mediums are green systems which results in zero-emission of carbon into
the atmosphere. Thermal energy storage (TES) systems absorbs heat, stores it and
releases it when required for use; this helps in load shedding in buildings. TES is
further classified based on their mode of operation as: chemical heat storage (CHTES)
sensible heat storage (SHTES) or latent heat storage (LHTES) [4].

2 Phase Change Materials: Definition, Physics

of Operation and Classification

Phase change materials change their physical form upon absorption or release of
energy. This energy could be in the form of heat of fusion or heat of vaporization;
their ability to store substantial amounts of thermal energy in proportionately small
volume makes them fit for a broad range of applications. When the PCM absorbs
energy from the ambient, the chemical bonds of the PCM breaks up and this leads
to a change of phase. The process of the phase change is an endothermic process;
therefore, heat is absorbed from the ambient causing a drop in the bulk temperature
of the surrounding. When there is a drop in the ambient temperature, the PCM
gives up its energy and returns to the solid state, thereby increasing the ambient
temperature of the room [5]. PCMs are generally classified by considering their
chemical composition as organic, inorganic or eutectic compounds [6]. Summarized
below are some advantages and disadvantages of different classes of PCMs also some
desired properties of PCMs (Table 1).
Some examples of the three classes PCMs applicable to buildings are outlined in
the Table 2.
A Concise Review of Nano-enhanced Phase Change Materials … 195

Table 1 Advantages and disadvantages of PCMS [6]

Types of PCM Advantages Disadvantages
Organic Organic PCMs have the ability to The thermal conductivity of organic
freeze with very little supercooling or PCMs in their solid state are very low
none as the case may be
Ability to be incorporated directly Some organic PCMs are insoluble in
water
Good thermal properties The heat of fusion is low
Uniformity in melting Less dense
Inorganic High thermal conductivity High degree of supercooling
Great heat storage capacity Reduced thermal stability
Low volume expansion upon a Phase segregation
change of phase
Eutectics Definite melting temperature Thermophysical properties of
eutectic PCMs are not well
established yet
High specific heat capacity Low enthalpy

Table 2 PCMs for cooling applications in buildings [7]

Compounds Temp (°C) of fusion Latent heat (KJ/Kg) Ref
Organic Dimethyl sabacate 20.8 120 [10]
Polyglycol (E 600) 22 127 [13]
Paraffin (C13 –C24 ) 22–24 189 [10]
Inorganic CaCl2 ·6H2 O 29 190.8 [13]
Mn(No3 )2·6H2 O 25.8 126 [10]
LiNO3 ·3H2 O 30 296 [10]
Eutectic 33.3% MgCl2 –6H2 O 25 127 [13]
+ 66.6% CaCl2 ·6H2 O
48% CaCl2 + 4.3% 27 188 [13]
NaCl + 0.4% KCl +
47.3% H2 O
60% 30 200 [13]
Na(CH3 COO)·3H2 O
+ 40% CO(NH2 )2

3 NanoPhase Change Materials

The foregoing discussions have revealed how phase change materials play a very
crucial role in latent heat energy storage and dispensing systems. There exist various
solar-deriventhermal technologies that employ the use of PCMs in both active and
passive capacities. Enhancing the heat storage and dispensing capabilities of PCMs
can be achieved by doping pure PCM with nanoparticles [8]. Metals, metallic oxides
196 C. K. Oluah et al.

Fig. 1 Classification of nanomaterials

and carbonic nanoparticles have all been introduced in various quantities to boost the
heat conduction abilities of PCMs]. Babapoor et al. further established that nanopar-
ticles improved the heat transfer abilities of PCMs, and there seem to be a dilemma
in the discourse of the outcome of nanoparticles on the heat of fusion of phase
change materials, while a few authors presented results of positive improvement on
the latent heat of their materials, and some others had shown a gradual decline of
the latent heat value of the nanoPCM with increased quantities of nanoparticles. The
duration of phase change, density and the viscosity have also been improved with
nanoparticles. [9].

3.1 Classification of Nanomaterials

Nanomaterials are broadly classified as organic, inorganic and hybrids just like phase
change materials [8]. Further sub-classifications are identified in the chart below
(Fig. 1).

3.2 Applications of Nanomaterials

Nanomaterials find great applications in basically all spheres of human endeavor,

some of the major applications are outlined in the chart below [8] (Fig. 2).
A Concise Review of Nano-enhanced Phase Change Materials … 197

Fig. 2 Applications of nanomaterials

3.3 Preparation of NanoPCMs

There are two general methods scientists use for the preparation of nanoenhanced
phase change materials; these methods are the one-step method and the two-step
method. The first method being the one-step method entails the concurrent produc-
tion of the nanoparticles into the base PCM while in the two-step approach, the
nanomaterials are first produced before introduction into the base material [10].

3.4 Desired Properties of Nanoenhanced PCM

The essence of doping phase change materials with nanoparticles is to improve its
properties leading to a better performance upon application. Below are some desirable
characteristics of nanoPCMs.

Suitable Phase Change Temperature

The building needs to be maintained at a temperature interval within the human
comfort limits; this means that phase change materials in use should have a phase
change temperature within the interval of application required. [11].

Good Thermal Conductivity

One of the major challenges associated with PCMs are low thermal conductivity,
doping with nanoparticles greatly improves the heat transfer potentials of PCMs
[12].
198 C. K. Oluah et al.

Table 3 Summary of characterization techniques of nanoenhanced PCM [13]

s/n Technique Function
1 Thermal conductivity analyser (TCA) The TCA is used to measure and analyse
the thermal conductivity of nanoPCMs
2 Differential scanning calorimeter (DSC) The DSC is used to ascertain the phase
change temperature and enthalpy of fusion
of nanoPCM/sample materials
4 Scanning electron microscope (SEM) The SEM examines the image of the
morphology nanomaterials
6 Hydrometer Measures the density of the nanomaterial
7 Thermal gravimetric analyser (TGA) Measures and analyses the thermal stability
of nanomaterials
8 Fourier transform infrared (FTIR) The FTIR machine examines the infrared
spectrometer spectrum and evaluates the chemical
affinity of materials being sampled
9 X-Ray diffraction (XRD) The XRD machine measures the crystal
size and structure of nanoparticles

Higher Latent Heat

The potential of PCMs to retain a great quantity of heat in a relatively small volume
is quite desirable. It makes them suitable for quite a number of applications.

3.5 Characterisation of NanoPCM

Nanophase changes materials after preparation are characterized to ascertain their

properties. Various techniques are used to characterize nanophase change materials
some of which are summarised in the Table 3.

3.6 Heat Transfer Models of Nanoparticles in Base PCM

The two-dimensional energy balance equation of heat flow in the PCM containing
nanoparticles is given below
   2 
∂T   ∂T ∂T ∂ T ∂2T
ρc p + ρc p u +v =k + (1)
∂t ∂x ∂y ∂x2 ∂ y2

where u, v, t, ρ, k, C p , T, are the velocity components of the momentum equation in

the cartesian coordinate system, time, density, thermal conductivity and specific heat
at constant pressure and temperature. The phase change in PCMs occurs over a range
of temperature denoted as Tm . Approximations for the thermophysical properties
A Concise Review of Nano-enhanced Phase Change Materials … 199

like density (ρ), heat capacity (ρC p ) and thermal conductivity (k) due to change in
phase presented in Eqs. (2)–(4) below.
⎧
⎨ ρns T < Tm
ρns +ρnl
ρ= T ≤ T ≤ Tm + Tm (2)
⎩ 2
ρnl T ≥ Tm + Tm
⎧
⎨ kns T < Tm
kns +knl
k= Tm ≤ T ≤ Tm + Tm (3)
⎩ 2
knl T ≥ Tm + Tm
⎧
⎨ (ρC p )ns
⎪
 T < Tm
(ρC p )ns +(ρC p )nl ρns +ρnl L h f,n
ρC p = + Tm
T ≤ T ≤ Tm + Tm (4)
⎪
⎩
2 2
(ρC p )n f T ≥ Tm + Tm

Subscripts “ns” and “nl” refers to the state properties of the nanoPCM in solid
and liquid forms. L h f,n in Eq. (4) above represents the latent heat of fusion of the
nanoPCM and expressed mathematically as L h f,n = (1 − ϕ)h s f and ϕ is the weight
percentage of nanoparticles contained in the base materials [14].

4 Nanoenhanced PCMs for Passive Cooling Applications

Introducing small quantities of nanoparticles in base PCMs to improve its ther-

mophysical properties has gained momentum in the recent past. Nanomaterials are
currently being used in virtually all fields of science and technology. The essence of
this review is to highlight works done on various nanomaterials for passive cooling
applications in building. For the want of space, these works have been summarized
in a tabular form below (Table 4).

4.1 Challenges and Scope for Future Research

The use of nanoparticles have greatly improved the thermal conductivity of PCMs,
however, some researchers reported a decline in latent heat as the concentration
of nanodopant increased while some others reported a linear relationship with the
increased concentration of nanomaterials; it is therefore important that a valida-
tion study is carried out. A research on optimum particle loading for best thermal
performance of nanoPCMs should also be investigated. Health risks involved in inte-
gration on nanoparticles in building envelopes should be studied and safety measures
are advised accordingly.
Table 4 Summary of recent studies on nanoenhanced PCM for passive cooling applications
200

Nanoparticle Base material Application Major findings References

TiO2 Mortar bricks Space cooling in buildings Resistance to thermal conductivity [15]
and thermal transmissivity coefficient
for TiO2- coated mortar wall showed
a reduction by 10% and 13%,
respectively
TiO2 Paraffin wax Ceiling roofs for space cooling An optimum value of heat capacity [16]
was attained at 0.7% loading of
TiO2 . Thermal conductivity showed
a liner increment as the concentration
of nanoparticle increased
Aluminiumnanoparticles PEG + Epoxy resins Building material Improved thermal conductivity due [17]
to nanoparticles addition
Al2 O3 + Alumina n-tetradecane +n hexadecane Space cooling in buildings Increasing the amount of [18]
nanoparticles in base PCM resulted
to an improved solidification and
melting rate
Graphite nanosheet Fatty acid Gypsum wallboards A reduction in the energy [19]
consumption by 79% was recorded
for the interior temperature
regulation
AgO and TiO2 Ethyl trans-cinnamate CTES in building 52% improvement in thermal [20]
conductivity, improved energy
efficiency and an overall freezing rate
reduction of 23.9%
AgO ICE block VAV-NTES air conditioning system Achieved energy savings of 36–58% [21]
and 24–51%, during peak load and
everyday basis, respectively
(continued)
C. K. Oluah et al.
Table 4 (continued)
Nanoparticle Base material Application Major findings References
Nanoencapsulation PMMA/capric–stearic acid CTES in buildings The TGA results showed great [22]
thermal conductivity, chemically
stable and excellent thermal
reliability after thermal cycling
Graphene Beeswax CTES in buildings An optimum thermal conductivity of [23]
2.8 W/mK was achieved at 0.3 wt%
of graphene and a 22.5% increase in
heat capacity were recorded
SiO2 Stearic acid, capric acid, palmitic CTES in Buildings The eutectic PCM has a temperature [24]
acid, and (SA-PA -CA) of fusion between the range of
17.16–26 °C. The presence of silicon
dioxide greatly improved its thermal
conductivity and shape stabilized
Activated carbon (AC) Oleic and capric acid CTES in buildings The presented showed that, upon [24]
addition of 0.1% of activated carbon,
a corresponding 55% improvement
in the thermal conductivity of the
pure PCM
A Concise Review of Nano-enhanced Phase Change Materials …

Expanded graphite Capric-lauric acid (CA-LA) CTES in buildings A 114.2% improvement in thermal [25]
conductivity upon addition of
10 wt% of EG
201
202 C. K. Oluah et al.

5 Conclusion

A review of phase change materials doped with varying quantities of nanoparticles

was discussed in this article; research on the use of nanomaterials to improve the
performance of PCMs is relatively not a new concept, but recent advances in the
manufacturing and doping of substances with nanomaterials have made this field of
research gain momentum in the recent past. One major application of nanoenhanced
PCMs is in cool thermal energy storage (CTES) in buildings, this review was geared
towards evaluating studies on nanoPCMs for passive cooling applications in build-
ings; Futhermore, provided an insight into the current challenges of nanoPCMs in
buildings and research gaps.

References

1. IEA Roadmap targets. https://www.iea.org/publications/freepublications/publication/

IEASolar_HC_Roadmap_FoldOutPrint.pdf
2. IPCC (2007) Summary for policymakers. In: Climate change 2007: impacts, adaptation and
vulnerability. contribution of working group II to the fourth assessment report of the inter-
governmental panel on climate change. Cambridge University Press, Cambridge, UK, p
17
3. Oysterman E et al (2012) Review of PCM based cooling technologies for buildings energy and
buildings
4. Hassan N, Mariah B, OSORIO FJB, Marllory I, Xinhai Y, Vignarooban K, Patrick P, Kannan
IAM (2019) Recent developments in phase change materials for energy storage applications:
a review. Int J heat mass transf 129:491–523
5. Hasnain SM (1998) Review on sustainable thermal energy storage technologies, Part I: heat
storage materials and techniques. Energy Convers Manag 39(11):1127–1138. https://doi.org/
10.10106/S0196-8904(98)00025-9
6. Feldman D, Shapiro M, Banu D (1986) Organic phase change materials for thermal energy
storage. Sol Energy Mater 13:1–10
7. Lane GA (1980) Low temperature heat storage with phase change materials. Int J Ambient
Energy 1:155–168
8. Hussain H, Saw C (2013) Afolabi Lreview on nanomaterials for thermal energy storage
technologies. Nanosci Nanotechnol-Asia 3:60–71
9. Babapoor A, Karimi G, Sabbaghi S (2016) Thermal characteristic of nanocomposite phase
change materials during solidification process. J Energy Storage 7:74–81
10. Farid MM, Khudhair AM, Razack SAK, Al-Hallaj S (2004) A review on phase changeenergy
storage: materials and applications. Energy Conv Manag 45(9–10):1597–1615
11. Zhu Y, Qin Y, Wei C, Liang S, Luo X, Wang J, Zhang L (2018) Nanoencapsulated
phasechange materials with polymer- SiO2 hybrid shell materilas: compositions, morphologies
and properties. Energy Conv Manag 164:83–92
12. Yu W, Xie H (2012) A review on nanofluids: preparation, stability mechanisms, and application.
J Nanomaterials. Article ID 435873
13. Sari A, Alkan C, Doguscu DK, Biçer A (2014) Micro/nano-encapsulated n-heptadecane with
polystyrene shell for latent heat thermal energy storage. Solar Energy Mater Solar Cells
126:42e50
14. Harikrishnan MD, Kalaiselvam S (2014) Thermal and rheological properties of hybrid
nanocomposite phase change material for thermal. Energy Convers Manage 86:864–872
A Concise Review of Nano-enhanced Phase Change Materials … 203

15. Carneiro JO, Vasconcelos G, Azevedo S, Jesus C, Palha C, Gomes N, Teixeira V (2014) Energy
Build 81:1–8
16. Wang J, Xie H, Guo Z, Guan L, Li Y (2014) Appl Thermal Eng 73(2):1541–1547
17. Constantinescu M, Dumitrache L, Constantinescu D, Anghel E, Popa V, Stoica A, Olteanu M
(2010) Latent heat nano composite building materials. Eur Polym J 46(12):2247–2254
18. Kalaiselvam S, Parameshwaran R, Harikrishnan S (2012) Analytical andexperimental investi-
gations of nanoparticles embedded phase change materials for coolingapplication in modern
buildings. Renew Energy 39(1):375–387
19. Sayyar M, Weerasiri RR, Soroushian P, Lu J (2014) Experimental and numericalstudy of shape-
stable phase-change nanocomposite toward energy-efficient buildingconstructions. Energy
Build 75:249–255
20. Parameshwaran R, Deepak K, Saravanan R, Kalaiselvam S. Preparation
21. Parameshwaran R, Kalaiselvam S (2014) Energy conservative air conditioning system using
silver nano-based PCM thermal storage for modern buildings. Energy Energy Storage. Appl
Energy 115:320–330. Build 69:202–212
22. Sarı A, Alkan C, Özcan AN (2015) Synthesis and characterization of micro/nano capsules
of PMMA/capric–stearic acid eutectic mixture for low temperature-thermal energy storage in
buildings. Energy Build 90:106–113
23. Amin M, Putra N, Kosasih EA, Prawiro E, Luanto RA, Mahlia T (2017) Thermal properties
of beeswax/graphene phase change material as energy storage for buildingapplications. Appl
Therm Eng 112:273–280
24. Luo Z, Zhang H, Gao X, Xu T, Fang Y, Zhang Z (2017) Fabrication andcharacterization of
form-stable capric-palmitic-stearic acid ternary eutectic mixture/nano-SiO2 composite phase
change material. Energy Build
25. Hussain SI, Dinesh R, Roseline AA, Dhivya S, Kalaiselvam S (2017) Enhancedthermal per-
formance and study the influence of sub cooling on activated carbon dispersedeutectic PCM
for cold storage applications. Energy Build 143:17–24
Effect of Silicon Carbide
in Yttria-stabilized Zirconia for Thermal
Protective Structural Materials

S. Subha, Poojari Surendra, Ch. Gowtham Chowdary,

V. Naga Venkata Sai Ram, and Palem Srikanth

Abstract Modern coating technology development is essentially determined by sys-

tems, which perform extreme and adverse operating conditions. The ceramic coat-
ing technology (CCT) has found applications in space vehicles, missiles, gas turbine
engines, and nuclear reactors for the thermal protective structural material. In present
study, theyttria-stabilized zirconia (YSZ) and with various weight percentages of 0.1
and 0.25% silica carbide nanoparticles were synthesized by a wet chemical process.
The morphological and thermal properties of the prepared hybrid ceramic coating
particles were studied using scanning electron microscope (SEM) and thermogravi-
metric analysis (TGA). The ablation performances of the hybrid coating samples were
tested using oxyacetylene flame test. The grain size and crystal structure of samples
were confirmed by SEM analysis and showed that the synthesize hybrid particles are
composed of agglomerated nanometer sub particles. The results of TGA showed the
major weight mass loss observed around 604 °C for pure YSZ and in the range of
650 °C for hybrid coating cured composites. The linear and mass ablation rates of
the prepared samples were significantly decreased by 43% and 18.4%, respectively,
on adding the 0.25 wt% of SiC hybrid filler.

Keywords Thermal protective coating · Silicon nanocarbide · Thermal stability ·

Ablation rate

1 Introduction

Zirconia-based thermal protective coatings are widely used for the thermal, oxidation,
and hot corrosion protection of high temperature components in gas turbine engines.
These coatings provide insulation to metallic structures in the hot section of turbine
engines and offer increased operating temperature of the engine-enhanced efficiency,
extended durability, and extended life of metallic components subjected to high

S. Subha (B) · P. Surendra · Ch. Gowtham Chowdary · V. Naga Venkata Sai Ram · P. Srikanth
School of Aeronautical Sciences, Hindustan Institute of Technology and Sciences, Chennai, India
e-mail: subhasvkmr@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 205

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_19
206 S. Subha et al.

temperatures [1–3]. Several impurities common in fuels have been identified and
associated with corrosion in the coatings. These impurities include sodium, sulfur,
phosphorous, and especially vanadium. These impurities react with convectional
YSZ turbine blade coating, severely limit the coating lifetime. Therefore, it is of
great interest to develop alternative materials that increase the operating lifetime of
the coating [4].
New zirconia-based ceramic compositions have been under investigation for TBC
applications since many years. Previous work focuses on studying stabilizers such
as Y2 O3 and MgO were susceptible to chemical attacks. Indium-stabilized zirconia
shows relative inertness in vandia containing atmosphere but it volatilizes at a lower
temperature than zirconia [5]. Graphene-based materials are single- or few-layer
platelets that can be produced in bulk quantities by chemical methods. Thermal and
barrier properties of these composites are dependent upon the intrinsic properties
of graphene-based materials and their state of dispersion in the matrix [6]. Nanos-
tructurally stabilized zirconium oxide (NSZ) hard transparent films were produced
without chemical stabilizers by the ion beam-assisted deposition technique (IBAD).
These NSZ films have better thermal stability, probably because of the absence of
porosity due to the nature of the IBAD process [7]. In another study, sol–gel method
has been used to coat multiwalled carbon nanotubes (MWCNTs) with zirconia at
room temperature and the coated tubes were sintered at high temperature. Sinter-
ing has improved the adhesion of the coating with CNTs and thereby improved its
high temperature resistance. It is also observed that the selecting proper sintering
temperature is the key to ensure a uniform coating layer on the CNTs [8].
The aim of this study is to synthesis the yttria-stabilized zirconia with silicon
nanocarbide material through wet chemical method. The morphology and thermal
properties of the prepared hybrid ceramic coating particles were studied by using
scanning electron microscope (SEM) and thermogravimetric analysis (TGA), respec-
tively. The ablation performances of the hybrid coating samples were tested using
oxyacetylene flame test.

2 Experimental

2.1 Materials

SiC of average particle size 50 nm used for composite preparation was obtained
from Sisco Research Chemicals Co. Ltd. Yttria-stabilized zirconia (YSZ) of average
particle size 3 µm was procured from Sigma Aldrich Co Ltd.
Effect of Silicon Carbide in Yttria-stabilized Zirconia for Thermal … 207

2.2 Preparation of Hybrid Coating

The SiC was coated on yttria-stabilized zirconia (YSZ) using sol–gel method.
The solution of YSZ is mixed with distilled water and ethanol in the molar ratio
0.1:3.17:1.75 [9]. The solution was magnetically stirred for 30 min to ensure the
homogenous distribution of particles. Thereafter, the SiC has added to the YSZ solu-
tion and the mixture was pre-reacted for 15 min with ultrasonicator. Finally, the
mixture was dried at room temperature for 24 h. The same process was repeated for
other samples with SiC content (0.1, 0.25 wt%). To improve the interfacial adhesion
of SiC on YSZ, the YSZ/SiC powder was sintered at 1000 °C for 1 h under nitrogen
atmosphere. The pre-determined amount of yttria-stabilized zirconia (YSZ) powder
was added to 150 ml of deionized water in a beaker maintain stirring at 600 rpm.
To ensure the better solid suspension, the mixture was stirred for 5 h. The resultant
mixture was dried in an muffle furnace at 200 °C for 2 h and dried solid mixture was
grinded to form fine powders and poured into a metallic mold (30 mm in diameter
and 10 mm thickness), Then the powder was dry-pressed into a disk-shaped pellet,
under a pressure of 70 MPa [10]. The sintering temperature was set at 1400 °C for
2 h with the heating rate of 1 °C/min as shown in Fig. 1.

Fig. 1 Preparation of hybrid

coating by sol–gel method
208 S. Subha et al.

2.3 Thermogravimetric Analysis

The thermal stability of the prepared samples was investigated using TGA instrument
(SEIKO model TG/DTA 6200). The 5 mg sample was placed in an alumina crucible
and temperature was raised to 900 °C at a heating rate of 10 °C/min.

2.4 Oxyacetylene Flame Test

Ablation performance of the prepared sample was evaluated using oxygen–acetylene

flame test. Each sample was fixed in a sample holder and a distance of 50 mm was
maintained between the nozzle tip and the specimen. The test samples were prepared
with 30 × 10 mm dimension. The oxyacetylene flame temperature was estimated to
be approximately 1400 °C during the test. The linear and mass ablation rates were
calculated using following Eqs. (1) and (2).

d (di − d f )
LR = = (1)
ta ta

where LR is linear ablation rate; d is a total change in thickness of specimen before

and after ablation; d i and d f are the initial and final thickness of the composite sample
before and after ablation test; ta is ablation through time.

M (m i − m f )
MR = = (2)
ta ta

where MR is mass ablation rate; m is a total change in mass of specimen before and
after ablation; mi and mf are initial and final mass of the composite sample before
and after ablation test.

3 Results and Discussion

3.1 Microstructure of Coated YSZ-SiC

Figure 2a shows the surface morphology that YSZ grains blend and become incred-
ibly quick at 1400 °C and the break is about all trans granular because of the grain
development it was seen in Fig. 2b that the microstructure turns out to be progressively
heterogeneous dispersion. The state of micron pores is increased by using higher slur-
ries [11]. When the silicon nanocarbide is added to yttria-stabilized zirconia surface,
strong stacking was diminshed (Fig. 2c) . The heterogeneous dispersion of silicon
nanocarbide was scribed to various sedimentation speeds as shown in Fig. 2d. The
Effect of Silicon Carbide in Yttria-stabilized Zirconia for Thermal … 209

Fig. 2 High magnification of a Surface morphology of YSZ b YSZ with 0.25% silicon nanocarbide
c Surface morphology of YSZ d YSZ with 0.25% silicon nanocarbide

homogeneous dispersion and progressively standard formed pore certainly contribute

to a compressibility of the coating material.

3.2 Thermal Stability

The effect of fillers on thermal stability of YSZ-SiC hybrid coating was studied using
TGA and DTA techniques. Figure 3 shows the weight loss of samples at the low-
temperature stage, below 150 °C, corresponds to the evolution of organic solvents or
gases trapped during the process of synthesizing. This small amount of weight loss
(2–4%) includes formaldehyde, water, cross-linked molecules, methanol and phenol
[12]. However, the evolution of these products was less evident for the YSZ1 and
YSZ2 hybrid coatings. The major weight mass loss was observed around 604 °C for
pure YSZ and in the range of 650 °C for hybrid coating cured composites [13]. The
210 S. Subha et al.

Fig. 3 TGA analysis for YSZ-SiC hybrid coating

thermal decomposition of prepared samples occurs temperature-dependent stages

The decomposition is mainly due to the degradation and oxidation of zirconia [14].
Several researchers have been reported a similar observation on the decomposition
temperatures [15].
In Fig. 4, it can be noticed that the maximum decomposition temperatures of both
stages were increased with the addition of silicon nanocarbide into yttria-stabilized
zirconia.

Fig. 4 DTA analysis for YSZ-SiC hybrid coating

Effect of Silicon Carbide in Yttria-stabilized Zirconia for Thermal … 211

3.3 Ablation Performance

The ablation performance of prepared samples was evaluated using an oxyacetylene

flame test. The linear ablation and mass ablation rates of composites were calculated
by exposing the surface of the sample to oxyacetylene flame at 1200 °C for 60 s. The
images of prepared samples before and after ablation test are shown in Fig. 5a, b. A
comparison of mass and linear ablation rates of the prepared samples are presented
in Figs. 6 and 7, respectively. It was observed that the mass and linear ablation
rates of the pure YSZ were 0.0038 g/s and 0.0028 mm/s, respectively. Whereas, the
mass and linear ablation rates of the ysz1 sample were 0.0036 g/s and 0.0018 mm/s,
respectively, and 0.0031 g/s and 0.0016 mm/s for ysz2 hybrid coating sample .
The addition of 0.1 wt% SiC hybrid filler into YSZ reduced the mass and linear
ablation rates by 5.3% and 35.7%, respectively. The mass and linear ablation rates of
the prepared samples were significantly decreased by 18.4% and 43%, respectively,

Fig. 5 Images of prepared samples: a before ablation test b after ablation test

Fig. 6 Mass ablation performance of prepared sample after oxyacetylene flame test
212 S. Subha et al.

Fig. 7 Linear ablation

performance of prepared
sample after oxyacetylene
flame test

Table 1 Ablative properties

Samples Porosity (%) Mass ablation Linear ablation
of prepared samples
rate (g/s) rate (mm/s)
YSZ 0.338 0.0038 0.0028
YSZ1 0.419 0.0036 0.0018
YSZ2 0.433 0.0031 0.0016

after adding the 0.25 wt% of SiC hybrid filler. The porosity increases as the YSZ-to-
SiC ratio increases, and because the permeability of the material, it directly increases
the thermal properties. The ablative properties of prepared samples are presented in
Table 1.

4 Conclusion

In this work, the effect of silicon nano-carbide coated into yttria-stabilized zirconia
was studied. Thermal studies results show that the YSZ2 sample attained higher ther-
mal stability compared to all other samples. The major weight mass loss was observed
around 604 °C for pure YSZ and in the range of 650 °C for hybrid coating cured
composites. Ablation performance of tested samples linearly reduced as % wt of SiC
nanofillers increased in the composite. The YSZ2 sample showed that the linear and
mass ablation rates were significantly decreased by 43% and 18.4%, respectively.
The surface morphology of hybrid-coated sample revealed the homogeneous disper-
sion and porus structure formed certainly contributed to a compressibility of the
coating material.
Effect of Silicon Carbide in Yttria-stabilized Zirconia for Thermal … 213

References

1. Miller RA (1987) Surface coating technology 30(1)

2. Brindley WJ, Miller RA (1989) Av. Material processes 29(8)
3. Jones RL (1996) Thermal barrier coating: metallurgical and protective coating. London
4. Michael D, Phelphs (2002) Corrosion resistant thermal barrier coating material for industrial
application. Prog Therm Barrier Coating 3–11
5. S Xu (1993) Thermal barrier coating material for industrial gas turbine application. J Ceram
Soc 76(8):2027–2032
6. Jun M, Qingshi M, Izzuddin ZS, Andrew M, Nobuyuki K, Wang CH, Hsu-Chiang K
(2014) Development of polymer composites using modified, high-structural integrity graphene
platelets. Compos Sci Technol 91:82–90
7. Namavar F, Gonghua W (2007) Thermal stability of nano structurally stabilized zirconium
oxide. Nano Technol 18
8. Mo CB, Cha SI, Kim KT, Lee KH, Hong SH (2005) Fabrication of carbon nanotube reinforced
alumina matrix nanocomposite by sol–gel process. Mater Sci Eng A 395:124–128
9. Montanaro L, Jorand Y, Fantozzi G, Negro A (1998) Ceramic foams by powder processing. J
Eur Ceram Soc 18:1339–1350
10. Corbin SF, Apté PS (1999) Engineered porosity via tape casting, lamination and the percolation
of pyrolyzable particulates. J Am Ceram Soc 82:1693–1701
11. Kim J, Lin YS (1999) Porous ion-conducting ceramic membranes. J Am Soc 46:2641–2646
12. Gu Y, Liu X, Meng G, Peng D (1999) Porous YSZ ceramics by water-based gel casting. Ceram
Int 25:705–709
13. Zhang Q, Shen J, Wang J, Wu G, Chen L (2000) Sol-gel derived ZrO 2–SiO2 highly reflective
coatings. Int J Inorg Mater 2:319–323
14. Subha S, Dalbir S, Pappakudi SV, Kavitha N (2018) Experimental investigation of mechanical,
thermal, and ablation performance of ZrO2 /SiC modified C-Ph composites. Int J Appl Ceram
Technol 00:1–11
15. Koo JH, Miller MJ, Weispfenning J, Blackmon C (2010) Siliconepolymer composites
for thermal protection system: fiber reinforcements and microstructures. J Compos Mater
45(13):1363–1380
Structures and Processing of Polymer
Matrix Composite
Optimization of Process Parameter
in Abrasive Water Jet Machining
of Blue-Fired Grain-Reinforced Glass
Fiber Polymer Composite

P. Sabarinathan , V. E. Annamalai, and K. Rajkumar

Abstract Abrasive water jet machining (AWJM) is the non-traditional machining

process which is mostly employed for cutting non-conductive materials. The present
work concentrates on the AWJ machining study of various compositions (5, 15 and
25%) of blue-fired alumina-reinforced polymer composite. The addition of reinforce-
ment improved mechanical properties which may affect the machining characteris-
tics. The experiment was conducted by varying the experimental process parameters
of pressure, abrasive flow rate, and standoff distance. Kerf angle and surface rough-
ness were affecting process parameters which are required to optimize for quality
cutting. The results show that the pressure and standoff distance have a direct corre-
lation with the surface roughness and kerf taper angle. Higher pressure and shortest
standoff distance show a smooth surface roughness due to no appreciable loss of
kinetic energy in cutting wear zone. Similarly, the increase in the reinforcement per-
centage increases surface roughness and kerf angle. Fiber delamination and grain
embedment in the matrix were studied using an optical microscope.

Keywords Machining · Abrasive · Hardness · Surface roughness first section

1 Introduction

Polymer composites have evolved in modern world applications, in various sectors of

automobile, aerospace, construction, and household things [1]. To improve strength
and load bearing capability of polymer composite, various types of fibers and fillers
are added to the matrix. Glass, carbon, and Kevlar are the commercially available
synthetic fibers. Glass fiber is the most used synthetic fiber in polymer composite; this
is due to its low cost, high strength-to-weight ratio, and corrosive resistance property
[2]. Nowadays, researchers are showing interest toward developing new materials

P. Sabarinathan (B) · V. E. Annamalai · K. Rajkumar

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, India
e-mail: sabarinathan14010@mech.ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 217

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_20
218 P. Sabarinathan et al.

from the recycled or recovered spent materials. Recycling of spent materials from
various sectors like agricultural [3], industrial [4], and electrical [5] are explored as
reinforcement substitute to the polymer composites. Addition of this kind of recycled
materials shows good improvement in mechanical properties.
Abrasive water jet machining (AWJM) is one of the unconventional machining
processes and is most suitable to machining of versatile materials [6]. This process
is not evolving with thermal heat leading to no heat-affected zone, and more flex-
ible to cut in all direction [7]. Besides, AWJM process bestows, as exact, required
dimensions with limited tolerance with complex profile.
Conventional machining like drilling of fiber-reinforced polymer composite
exhibited various damages such as fiber pull out, fiber scrolling, and delamination.
These are due effect of variation in cutting forces during drilling operation. AWJM
process produces a minor damage like side wall damage but it is controllable by judi-
cious selection of process parameters. The various process parameters controllable
in AWJM process are pressure, standoff distance, abrasive flow rate and size, and
traverse rate [8]. Cut surface quality was mostly influenced by the above-mentioned
machining parameters. Besides economy of the cut, environment consciousness is
most important issue for cutting operation.
Machining of polymer composite is entirely different form homogenous material
cutting because of discontinuous phases of constituent materials. Azmir et al., have
studied the surface roughness and kerf taper ratio on the glass fiber/epoxy composite
with varying process parameters. Machining quality in terms of surface roughness
was low at higher pressure of 275 MPa, lower standoff distance of 1.5 mm, and lower
traverse rate of 1.5 mm/s [9]. Unde et al., have investigated the abrasive water jet
machining process on carbon fiber-reinforced polymer composite [10]. The results
report that standoff distance and abrasive flow rate were influencing material removal
rate and delamination of fiber. Likewise, hydraulic pressure has majorly influenced
the cut quality. Researchers have studied the machinability studies of with and without
filler-added polymer composites. The results show that the filler added composite
produces a good surface quality. Moreover, the traverse rate and jet pressure were
influential parameters for the machining process [11, 12].
The present work deals influence of cutting quality on recovered blue-fired
alumina-reinforced GFRP composite using an abrasive water jet machining pro-
cess. Various process parameters such as abrasive flow rate, pressure, and standoff
distance were used. Analysis on the response parameters such as kerf taper angle
and surface roughness was done to determine the optimal and influential parameters
on the surface cut quality.
Optimization of Process Parameter in Abrasive Water Jet … 219

2 Materials and Methods

2.1 Preparation of Various Percentage of Filler-Reinforced

Polymer Composite

Vitrified grinding wheel reject of blue-fired alumina with various percentages (5, 15
and 25%) was used as secondary reinforcement for polymer composite production.
Composites were produced using hand layup technique. Epoxy LY556 and hardener
HY951 were raw materials for the matrix. Blue-fired alumina of 120 grit particles
was mixed with resin using magnetic stirring. Mixed resin and glass mat with vari-
ous weight proportions were placed inside wooden mold with cross-section of 310 ×
310 × 5 mm. Five layers of resin in-between four layers of fiber was stacked con-
secutively. Composite samples were cured at room temperature for 24 h. After 24 h,
the samples were taken out from mold and used for machining operations. Figure 1
shows the fabricated composites. These composites filler loading of 5, 15, and 25%
are referred as RBA1, RBA2, and RBA3, respectively. The hardness of fabricated
composites was measured using ASTM-D2240 by D-Shore hardness tester.

2.2 Machining of Polymer Composite

Typical layout of the abrasive water jet machining process is shown in Fig. 2. Two-axis
abrasive water jet CNC machine (OMAX Corporation, Model: 1530) was used for
the machining process. The equipment which is designed to have a nozzle diameter
of 0.762 mm, abrasive mesh size of 80 grit, tool offset of 0.3 mm, and impinging
angle of 90° is maintained constant throughout the experiment.
The machine is capable to work with various process parameters such as pressure,
standoff distance, abrasive flow rate, abrasive mesh size, and traverse rate. From that
effective parameters of water jet pressure (MPa), standoff distance and traverse rate
were chosen. The machining parameters for abrasive water jet cutting were referred

Fig. 1 Fabricated blue-fired alumina-reinforced composites

220 P. Sabarinathan et al.

Fig. 2 Typical abrasive water jet machine

Table 1 Parameters and levels of abrasive water jet machining process

Sl. No. Parameters Level
1 Pressure (MPa) 200 250 300
2 Standoff distance (mm) 1 3 5
3 Abrasive flow rate (g/min) 270 300 330

from the literature [13]. Table 1 shows various process parameters and its level used
for the machining process.

2.3 Characterization of Machined Surface

Kerf width was measured using profile projector (Make: Metzer) with a coordinate
measuring table size of 200 × 200 mm with micrometer. Typical kerf width and taper
angle measurement are depicted Fig. 3. Measurement was done at bottom and top
kerf, and the kerf taper angle was calculated using an Eq. (1),

(K t − K b )
Kerf taper angle ∅ = tan−1 (1)
2t
where t is the thickness of the polymer sample and K t is kerf top width and K b is
kerf bottom width.
Optimization of Process Parameter in Abrasive Water Jet … 221

Fig. 3 Typical geometry of

cut down cross-section using
abrasive water jet machining
process

Surface roughness of the machined surface was analyzed using a Mitutoyo Surf
test SJ-301 stylus-type profilometer. The average surface roughness (Ra) was mea-
sured for all the three composites at different locations. Optical images of the
machined zone were taken using a Leica S6 D stereo microscope.

3 Results and Discussion

3.1 Mechanical Properties of Produced Composites

The hardness of the blue-fired alumina-reinforced polymer composites is shown in

Fig. 4. Incorporation of hard reinforcement in polymer composites shows a linear
increase in hardness. This is due to higher hardness of the added filler and termination
of linear polymer chains. Similar results were reported by Rajesh et al. [14].

3.2 Machining Studies on Polymer Composites

Abrasive water jet machining process was carried by varying the selected process
parameter on the RBA1, RBA2, and RBA3 samples which are shown in Fig. 5.

Effect of Water Pressure

Figure 6 shows the effect of water jet pressure on kerf taper angle and surface rough-
ness of various compositions of blue-fired alumina-reinforced polymer composites.
The experiment was conducted by varying jet pressures of 200, 250, and 300 MPa
and by maintaining the other parameters such as standoff distance (1 mm) and abra-
sive flow rate (330 g/min). The results clearly show that increase in pressure reduces
the kerf taper angle. This could be higher kinetic energy of jet increasing the top kerf
222 P. Sabarinathan et al.

Fig. 4 Hardness value of fabricated composite

Fig. 5 Abrasive water jet-machined polymer composites

width. On the other hand, there is only slight energy loss during cutting at bottom,
resulting in reduction of bottom kerf taper, as shown in Fig. 6a. The higher percent-
age of reinforcement (blue-fired alumina) shows the highest kerf taper angle. This
is due to high pressure garnet particle striking on the composite with higher energy
at entrance. Further jet energy absorbed by reinforcement particles causes them to
dislocate or dislodge from its own sites. This reduces jet energy at exit side. Due
to the interaction of garnet with high pressure, the reinforcement particles in matrix
could be dislodged [15]. As a result, a higher top kerf width is produced.
Optimization of Process Parameter in Abrasive Water Jet … 223

Fig. 6 Effect of water jet pressure on machined surfaces a kerf taper angle b surface roughness

Maximum surface roughness of the machined surfaces is observed at lower pres-

sure range. Increase in pressure decreases the surface roughness value. This is due to
smooth erosion of matrix at higher pressure. In other words, there is no much loss in
jet energy along the jet direction resulting in a uniform cut in both cutting wear zone
and deformation wear zone [16]. Composite RBA3 shows a higher surface roughness
value in all the pressures. This is due to particle pulled out from the matrix at high
pressure. The dislodgement of hard reinforcement can be seen in Fig. 7.

Effect of Standoff Distance

Figure 8a–b shows the kerf taper angle and surface roughness of fabricated polymer
composites by varying standoff distance. The experiment was conducted by main-
taining the constant pressure (300 MPa), abrasive flow rate (330 g/min), and varying

Fig. 7 Machined surface of RBA3 composite

224 P. Sabarinathan et al.

Fig. 8 Effect of standoff distance on machined surfaces a kerf taper angle b surface roughness

the standoff distance (1, 3 and 5 mm). Maximum kerf taper angle was observed
at higher standoff distance and also for the 25% filler composite. An increase in
standoff distance causes an expansion of water jet leading to more interaction with
atmospheric air. It leads to loss in the jet energy, ultimately, increasing kerf taper
angle.
Standoff distance increases dragging of water jet in atmospheric air and impairs
the cutting ability of jet at top and is worst in bottom portion. Owing to uneven
erosion of matrix, the reinforcement particles were protruded from the matrix surface,
increasing the surface roughness. In the case of 25% filler loaded composite, reduced
water jet energy at high standoff distance results in poor interaction of jet with
reinforcement particle. This may lead to deflection of water jet at reinforcement
sites, reducing the machining rate and uneven machining. This adversely affects
surface roughness at the cutting zone.

Effect of abrasive flow rate

Figure 9a–b shows the kerf taper angle and surface roughness of fabricated polymer
composite with respect to various abrasive flow rates. The experiment was conducted
by maintaining the constant pressure (300 MPa), standoff distance (1 mm), and
varying the abrasive flow rates (270, 300 and 330 g/min). The results show that
increase in abrasive flow rate decreases kerf taper angle. At lower abrasive flow rate,
the number of garnet particles interacting with the polymer composite is less. At
higher abrasive flow rate, the penetration depth was higher and uniform machining
was observed at both the top and bottom portions. It reduces kerf taper angle and
also surface roughness of the composite.
Composite with 25% reinforcement shows a higher taper angle and surface rough-
ness in all abrasive flow rates. Limited erosion of polymer matrix could be possibly
due to the presence of more filler particles in the matrix. This increased the taper
angle and surface roughness of the composite machining. Surface of the cut-down
zone was characterized; it shows that the minimum filler added composite (RBA5)
exhibits more uniform cutting as compared to RBA25, as shown in Fig. 10a–b.
Optimization of Process Parameter in Abrasive Water Jet … 225

Fig. 9 Effect of abrasive flow rate on machined surfaces a kerf taper angle b surface roughness

Fig. 10 Machined surface of higher abrasive flow rate a RBA5 b RBA25

4 Conclusion

From the present research work, the following conclusions are arrived.
1. The addition of blue-fired alumina abrasive grain in the epoxy matrix shows a
linear increase in the hardness.
2. Machining on different percentages of (5, 15 and 25%) blue-fired alumina-
reinforced polymer composite was experimentally done and uniform machining
was observed at 5% composite whereas uneven machining was observed for the
25% composite.
3. Better quality cutting of polymer composites was achieved at high water pressure,
low standoff distance, and high abrasive flow rate.
4. Jet pressure has direct co-relation with kerf taper and surface roughness. At higher
pressure, the reinforcement is uniform and deeper machining is observed. At low
pressure, jet capability was reduced to low level, increasing surface roughness
and kerf taper angle.
226 P. Sabarinathan et al.

5. With increase in standoff distance, there is an expansion in the jet size, due to
interaction of jet with the atmospheric air and there is a loss in jet energy. This
results in increase of kerf taper angle.
6. Optical image of machined surface shows that uniform machining has occurred
at minimum filler loaded composite. Maximum filler loaded composite resulted
in increased surface roughness of cut surface.

References

1. Singh N, Hui D, Singh R, Ahuja IPS, Feo L, Fraternali F (2017) Recycling of plastic solid
waste: A state of art review and future applications. Compos B Eng 115:409–422
2. Gnanavelbabu A, Saravanan P, Rajkumar K, Sabarinathan P, Karthikeyan S (2018) Mechan-
ical strengthening effect by various forms and orientation of glass fibre reinforced isopthalic
polyester polymer composite. Mater Today Proc 5(13):26850–26859
3. Sabarinathan P, Rajkumar K, Gnanavelbabu A (2016) Mechanical properties of almond shell-
sugarcane leaves hybrid epoxy polymer composite. Appl Mech Mater 852:43–48
4. Sabarinathan P, Annamalai VE, Kumar SS, Kennedy AX (2019) A study on recovery of alumina
grains from spent vitrified grinding wheel. J Mater Cycles Waste Manag 21(1):156–165
5. Senophiyah-Mary J, Loganath R, Meenambal T (2018) A novel method for the removal of
epoxy coating from waste printed circuit board. Waste Manag Res 36(7):645–652
6. Wang J (1999) Abrasive waterjet machining of polymer matrix composites–cutting perfor-
mance, erosive process and predictive models. Int J Adv Manufact Technol 15(10):757–768
7. Hutyrova Z, Scucka J, Hloch S, Hlavacek P, Zelenak M (2016) Turning of wood plastic
composites by water jet and abrasive water jet. Int J Adv Manufact Technol 84(5–8):1615–1623
8. Gnanavelbabu A, Rajkumar K, Saravanan P (2018) Investigation on the cutting quality charac-
teristics of abrasive water jet machining of AA6061-B4C-hBN hybrid metal matrix composites.
Mater Manuf Processes 33(12):1313–1323
9. Azmir MA, Ahsan AK (2008) Investigation on glass/epoxy composite surfaces machined by
abrasive water jet machining. J Mater Process Technol 198(1–3):122–128
10. Unde PD, Gayakwad MD, Patil NG, Pawade RS, Thakur DG, Brahmankar PK (2015) Exper-
imental investigations into abrasive waterjet machining of carbon fiber reinforced plastic. J
Compos
11. Ramesha N, Akhtar S (2016) Abrasive water jet machining and mechanical behavior of banyan
tree saw dust powder loaded polypropylene green composites. Polym Compos 37(6):1754–
1764
12. Jani SP, Kumar AS, Khan MA, Kumar MU (2016) Machinablity of hybrid natural fiber
composite with and without filler as reinforcement. Mater Manuf Processes 31(10):1393–1399
13. Dhanawade A, Kumar S (2017) Experimental study of delamination and kerf geometry of
carbon epoxy composite machined by abrasive water jet. J Compos Mater 51(24):3373–3390
14. Rajesh S, VijayaRamnath B, Elanchezhian C, Aravind N, Rahul VV, Sathish S (2014) Analysis
of mechanical behavior of glass fibre/Al2 O3 –SiC reinforced polymer composites. Procedia Eng
97:598–606
15. Doreswamy D, Shivamurthy B, Anjaiah D, Sharma NY (2015) An investigation of abrasive
water jet machining on graphite/glass/epoxy composite. Int J Manufact Eng
16. MM IW, Azmi AI, Lee CC, Mansor AF (2018) Kerf taper and delamination damage mini-
mization of FRP hybrid composites under abrasive water-jet machining. Int J Adv Manufact
Technol 94(5–8):1727–1744
Mechanical Behavior of Raavi
and Pineapple Fiber-reinforced Hybrid
Polyester Composites

J. Varun Siddharth, I. Daniel Lawrence, and S. Jayabal

Abstract The raising requirement of plastic materials and their production in

advanced engineering were having significant impact on environment. This investiga-
tion aims to observe the potential of agro waste residues, such as Raavi and Pineapple
fiber, as reinforcement in Polyester GP51 matrix as an alternative to plastic-based
components for structural applications. The composites were developed by using
compression molding technique by varying the fiber content in percentage. Single
fiber test and Fourier transform infrared spectroscopy were used to find the ability
of the fiber. The composites were fabricated and tested as per ASTM standards. The
tensile strength, flexural test, and impact strength were studied and morphological
study was carried out by using scanning electron microscopy and energy dispersive
X-ray analysis. The result shows that the Raavi and Pineapple fiber polymer com-
posites have achieved considerable increasing value of mechanical properties of the
composites. This investigation suggested the possible way of utilizing hybrid fiber
from the agro-based waste in polymer matrix composites.

Keywords Raavi fiber · Pineapple fiber · Polymer composites · Mechanical

behaviors

J. Varun Siddharth (B)

Department of Mechanical Engineering, Anna University, Regional Campus, Madurai, India
e-mail: varunmechanical1993@gmail.com
I. Daniel Lawrence
Department of Mechanical Engineering, Loyola Institute of Technology, Chennai, India
S. Jayabal
Department of Mechanical Engineering, Government College of Engineering, Thanjavur, India

© Springer Nature Singapore Pte Ltd. 2021 227

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_21
228 J. Varun Siddharth et al.

1 Introduction

No the global consumer sectors accounts for major part of the plastic materials, but
commonly these materials are degradable due to a resistance to microbial decay. The
natural fiber-based composite material can be produced by partly replacing polyester
resin with some natural fibers of different compositions with different percentage,
owing to use some conventional resources in our day today life. Van Vuure et al. [1]
have made comparative analysis on three different natural fiber composites based on
bamboo, flax, and coir fiber and he observed the percentage of variations of tensile
properties of natural fibers. Chegdani et al. [2] performed investigation on multiscale
tribological behavior when drilling natural fiber composites and hence the cause by
which the tribo mechanical behavior of drilling operation is affected was studied.
Vitale et al. [3] studied on natural and synthetic fiber-reinforced composite sandwich
panels for attaining a report on failure mode in sandwich panels of the composite.
Torres et al. [4] performed statistical study on comprehensive set of long natural fiber
composite systems to check the variability’s in long natural fiber laminate properties.
Corona et al. [5] observed on natural fiber composite selection without causing an
impact over the environment. Song et al. [6] studied the dynamic tensile properties for
recycled office paper fiber and old corrugated cardboard fiber at different proportions.
Fan et al. [7] determined the thermal decomposition of natural fiber composites and
the behavior and damage profile of natural fiber composite at microscopic and macro-
scopic levels. Pornwannachai et al. [8] determined the flammability and mechani-
cal performance of each fiber about the fire resistance on natural fiber-reinforced
composites. Mejri et al. [9] prepared birch fiber and high density polyethylene com-
posite for replacing nylon which is used in spur gear manufacturing in industries.
Chegdani et al. [10] studied on machinability of natural fiber-reinforced composite
material by investigating the mechanics of chip formation and the multiscale cut
surface by orthogonal cutting experiment. Vignesh G et al. [11] prepared hybrid
composite materials with kenaf, human hair as reinforcements, and epoxy resin
as matrix for investigating its mechanical properties such tensile strength, flexural
strength, and impact strength. Choudhry et al. [12] conducted experimental investi-
gation on polypropylene-based composite material with human hair as reinforcement
by varying the reinforcement and matrix percentage. From the results, they found
the mechanical properties of composite material are good with lower percentage of
reinforcements then the higher. Pothana et al. [13] analyzed the dynamic mechanical
properties of banana fiber-reinforced polyester composites with the effect of fiber
loading, frequency, and temperature. Rana et al. [14] fabricated natural fibers such
sisal, kenaf, hemp, jute, and coir-reinforced polypropylene composites processed by
compression molding using a film stacking method and analyzed the mechanical
properties of composites. Idicula et al. [15] fabricated randomly oriented intimately
mixed short banana and sisal hybrid fiber-reinforced polyester composites and ana-
lyzed the dynamic and static mechanical properties. The major objective of this work
is to bring new classes of fibers under process and to develop an alternative method
toward the use of unconventional and recycling products. In this work, we are going
Mechanical Behaviour of Raavi and Pineapple Fibre-reinforced … 229

to study and compare the mechanical properties, characteristics, and behavior of

natural fiber composites where Pineapple and Raavi serves as reinforcements and
polyester resin serves as matrix, thereby framing applications toward the use of these
composite materials.

2 Materials and Fabrication Procedure

2.1 Materials

In this study, two natural fibers, namely Raavi fiber (ficusreligiosa) is extracted from
the sticks of the branches of peepal tree or bodhi tree by pealing the outer layer of
the sticks by mechanical means and Pineapple fiber (ananascomosus) is extracted
from the leaves of the Pineapple plant by mechanical means and polyester GP51
matrix was used to fabricate natural fiber-reinforced composites to evaluate the fiber
performance under different percentages of composition. The selected fibers are
from different region and they exhibit variability in their physical and mechanical
properties.
The natural fiber composite with 15% Raavi [Ficusreligiosa], 15% Pineapple
[Ananas Comosus], 69.5% polyester GP51 resin, 0.25% of MERP catalyst, and
0.25% of octane solution is kept as composition 1 and the natural fiber composite with
10% Pineapple [Ananas Comosus], 20% Raavi [Ficusreligiosa], 65.5% polyester
GP51 resin, 0.25% of MERP catalyst, 0.25% of octane solution is kept as composition
2 (Fig. 1).

Fig. 1 Pineapple 15%, Raavi 15% as composition 1 and Pineappl e10%, Raavi 20% as composition
2
230 J. Varun Siddharth et al.

Fig. 2 Single fiber test for Pineapple and Raavi fibers

2.2 Single Fiber Strength for Preprocessing

In particular, the preprocessing test is conducted to check the mechanical ability of

the fiber by conducting the single fiber test with respect to ASTM standard D3822
by 100 series modular universal testing machine.
Single Fiber Test for Pineapple Fiber Based on the observations which are
obtained by conducting single fiber test over a Pineapple fiber, we observe that
the stress of the fiber is found to be around 10,100 MPa, with the corresponding
maximum force and elongation of 2.467766 N and 0.945653% by applying gradual
tension of 10 mm/min, respectively. The maximum force and elongation are obtained
just by a step on the point or on the region of the breaking zone (Fig. 2).
Single Fiber Test for Raavi Fiber From the observations which are obtained by
conducting single fiber test over a Raavi fiber, we observe that the stress of the fiber
is found to be around 16,200 MPa, with the corresponding maximum force and
elongation of 3.825889 N and 0.951% by applying gradual tension of 10 mm/min,
respectively. The maximum force and elongation are obtained by just a step on the
point or on the region of the breaking zone.

2.3 FTIR Test

Fourier transform infrared spectroscopy was used to determine the functional groups
in a molecule. The functional group exhibits its vibration by bending, stretching, and
waging at different frequencies.
FTIR Test with Raavi Fiber The functional groups present in the molecules of
Raavi fiber are observed by passing light at different frequencies through the speci-
men. By observing the graph, it is found that the transmittance of IR light is less in
functional group region then in finger print region. The stretching vibration with the
Mechanical Behaviour of Raavi and Pineapple Fibre-reinforced … 231

Fig. 3 FTIR graph for Raavi and Pineapple fibers

transmittance of 88% is obtained for the wave number of 2917.90; the obtained wave
number shows presence of strong alkane (C–H) group. The stretching vibration with
the transmittance of 65% is obtained for the wave number of 3285.88, the obtained
wave number shows the presence of strong and broad alcohol (O–H) group. The
values from 0 to 1800 are neglected since they come under finger print region. It is
observed that Raavi fiber has formaldehyde as its chemical compound. The results
are shown as graphical representation (Fig. 3).
FTIR Test with Pineapple Fiber The results are obtained by observing the graph;
it is found that the transmittance of IR light is less in functional group region then in
finger print region. The stretching vibration with the transmittance of 88% is obtained
for the wave number of 2915.33, and the obtained wave number shows presence of
strong alkane (C–H) group. The stretching vibration with the transmittance of 65%
is obtained for the wave number of 3299.17, and the obtained wave number shows
the presence of strong and broad alcohol (O–H) group. The values from 0 to 1800
are neglected since they come under finger print region. It is observed that Raavi
fiber has formaldehyde as its chemical compound.

2.4 Fabrication Procedure

The fabrication work was executed with the help of compression molding machine
by setting up the composite lamina. The compression molding machine is set up with
5–15 amps as current with the upper and lower die temperatures as 348 K and the
pressure of the hydraulic actuation system is maintained at 2 bar. The lamina plate is
set up with the dimensional specification of 3 mm as its thickness and the 300 mm
as its length and width is taken as a boundary parameter for the experiment, whose
total gross weight is said to be 450 g.

3 Result and Discussion

The prepared composite laminas are introduced to various test method in order to
attain the property and characteristics of the above said compositions. The properties
232 J. Varun Siddharth et al.

and the characteristics of the above said two compositions are compared with each
other in order to observe the changes in the composite behaviour, characteristics,
and properties with respect to change in composition, which is having similar pair
of materials with varying proportion of percentage for each and every material.

3.1 Tensile Test

The tensile test is performed with the help of universal testing machine where the
alternate ends of the testing specimen are held to unidirectional moving supports. The
UTM with the composite specimen is integrated with the computer in order to attain
the results which are obtained through the universal testing machine. The specimen
for the tensile test is grooved using wood grooving machine from the lamina based
on the dimensional specification which is suggested as per ASTM D638 standard.

Tensile Test Characterization of Composition 1 From the visual observation of

the stress–strain graph of the composite specimen were observed that the composite
specimen is capable of withstanding up to the maximum extent of plastic zone, and
the elastic zone is too small where the fracture occurs immediately. The observed
composite specimen is brittle in nature. Based on the visual observation of the graph
represented that the ultimate stress is attained at 32 MPa whose corresponding strain
percentage is 1.99 (Fig.4).

Based on the visual observation of force–length graph, the composite specimen

has elongated up to the length of 1.20 mm. Since the composite specimen is brittle
in nature, the reduction of size along its cross-section is too small. The specimen
attains a maximum force of 1200 N at a length of 1.20 mm and the specimen breaks
at 1.22 mm with a small drop in force of 1998 N for the corresponding thickness of
3 mm.

Tensile Test Characterization of Composition 2 From the visual observation of

the stress–strain graph of the composite specimen were found that the composite
specimen is capable of withstanding up to the maximum extent of plastic zone, and
the elastic zone is too small where the fracture occurs immediately. The observed

Fig. 4 Stress–strain and force–length graphs for composition 1

Mechanical Behaviour of Raavi and Pineapple Fibre-reinforced … 233

Fig. 5 Stress strain and Force length graph for composition 2

composite specimen is brittle in nature. Based on the visual observation of the graph,
we can note that the ultimate stress is attained at 20.1 MPa whose corresponding
strain percentage is 0.67 (Fig. 5).

Based on the visual observation of force to length graph, the composite specimen
has elongated up to the length of 0.68 mm. Since the composite specimen is brittle
in nature, the reduction of size along its cross-section is too small. The specimen
attains a maximum force of 753 N at a length of 0.68 mm and the specimen breaks
at 0.70 mm with a small drop in force of 725 N for the corresponding thickness of
3 mm.

Comparative Analysis of Tensile Test Characterstics Based on the results

obtained through the experimental analysis, the mechanical characteristics of natu-
ral fiber composites for composition 1 and composition 2 are studied. The observa-
tions show that both the composites are exhibiting brittleness and the plastic behav-
ior of the product is comparatively more than that of elastic behavior. In order to
obtain better strength and withstanding ability, comparative analysis is conducted
with composition 1and composition 2 (Table 1).

The above given bar graph and table are obtained by conducting tensile test
between two different compositions. The results obtained through comparative anal-
ysis show that the strength and withstanding ability of the composite composition 1
are greater than that of composition 2 (Fig. 6; Table 2).

Table 1 Comparative table

Mechanical characteristics Composition 1 Composition 2
of tensile test
Stress [MPa] 32 20.1
Strain [%] 1.99 0.67
Maxload [N] 1200 753
Elongation [mm] 1.2 0.68
Break [mm] 1.22 0.7
Break Load [N] 1998 725
234 J. Varun Siddharth et al.

Fig. 6 Stress strain graph for flexural test in composition 1 and composition 2

Table 2 Comparative table

Mechanical characteristics Composition 1 Composition 2
of flexural test
Max force [N] 31.67 32.67
Strength [MPa] 42.22 43.56
Breaking stress [MPa] 33 34.95
Breaking strain [%] 28 18.5

3.2 Flexural Test

The flexural test was conducted with the help universal testing machine by using
three-point bending method by replacing the default setting for tensile test.
Flexural Test Characterization of Composition 1 The results converged through
stress–strain graph show that the maximum force required to break the specimen is
31.67 N and the flexural strength attained by the specimen is 42.22 MPa, respectively.
The stress and strain at which the specimen breaks is 33 MPa and 2.80%, respectively.
Flexural Test Characterization of Composition 2 The results obtained through
stress-strain graph show that the maximum force required to break the specimen is
32.67 N and the flexural strength attained by the specimen is 43.56 MPa, respec-
tively. The stress and strain at which the specimen breaks is 34.95 MPa and 1.85%,
respectively.
Comparative Analysis of Flexural Test Characteristics The table which is repre-
sented below is obtained by conducting flexural test between two different compo-
sitions. The results obtained through comparative analysis show that the maximum
force, flexural strength, and breaking stress are greater in composition 1 and the
breaking strain is greater in composition 2.

3.3 Impact Test

The impact test is done with the help of impact testing machine as per the ASTM
standard D256. The stress–strain curve of the composite specimen pointed that the
specimen is exhibiting an brittle nature.
Mechanical Behaviour of Raavi and Pineapple Fibre-reinforced … 235

Table 3 Comparative table

Mechanical characteristics Composition 1 Composition 2
of impact test
Internal energy [J] 22.449 22.449
External energy [J] 2.797 5.675
Energy absorbed [J] 19.651 21.774
Impact strength [J/M] 22.379 45.400

Impact Test Characterization of Composition 1 Based on the Izod test results

with the test specimen, the internal energy absorbed by the specimen is 22.449 J
and the remaining energy after the specimen stroked is 2.7974 J. Therefore, the total
energy of the fractured specimen is 19.6516 J which is the difference between the
internal energy and the remaining energy. The impact strength which is obtained
in the composite specimen in terms of KJ/m2 is 74.591 and the impact strength in
composite specimen in terms of J/m is 223.797, respectively.
Impact test characterization of Composition 2 Based on the Izod test conducted
with the test specimen, the internal energy absorbed by the specimen is found to be
22.449 J and the remaining energy after the specimen stroke is 5.6750 J. Therefore,
the total energy which is delivered to the composite specimen is 21.774 J which is the
difference between the internal energy and the remaining energy and it is the impact
value of energy in the fracturing specimen. The impact strength which is obtained
in the composite specimen in terms of KJ/m2 is 151.335 and the impact strength in
composite specimen is in terms of J/m is 454.007, respectively.
Comparative Analysis of Impact Test Characteristics The results obtained
through comparative analysis show that the toughness of the composite composition
2 is greater than that of composition 1 (Table 3).

3.4 Analysis of SEM Images

The fractured surface of the Raavi and Pineapple fiber-reinforced composite was
analyzed using scanning electron micro graphs. Observed SEM images show the
fiber and resin compatibility in natural fiber polymer composites. The SEM image of
tensile fractured specimen showed uniform crack propagation due to better mixing
of the fiber and matrix. The fiber segregation was found to be good in the fractured
region due to the improvement in compatibility between the fiber and resin matrix.

3.5 Energy Dispersive X-ray Analysis

Chemical composition (chemical characterization) of the specimens was obtained

by conducting the EDX test by scanning electron microscope instrument.
236 J. Varun Siddharth et al.

Energy Dispersive X-ray Analysis of Composition 1 The obtained results of the

energy dispersive X-ray analysis, carbon element contains the maximum level of
un-normalized concentration and normalized concentration in weight percent of the
element, atomic weight percentage and the error in the weight percent concentration
at the 1 sigma level and also the observed results showed that, Phosphorus element
contains the minimum level of un-normalized concentration in weight percent of the
element, normalized concentration in weight percent of the element, atomic weight
percentage and the error in the weight percent concentration at the 1 sigma level.

Energy Dispersive X-ray Analysis of Composition 2 The obtained results of the

energy dispersive X-ray analysis, carbon element contains the maximum level, and
potassium element contains the minimum level of un-normalized concentration and
normalized concentration in weight percent of the element, atomic weight percentage
and the error in the weight percent concentration at the 1 sigma level (Figs. 7, 8, 9;
Table 4).

Fig. 7 SEM image of composition 1 for 500, 200, and 100 µm

Fig. 8 SEM image of composition 2 for 500, 200, and 100 µm

Mechanical Behaviour of Raavi and Pineapple Fibre-reinforced … 237

Fig. 9 EDAX image of composition 1 and composition 2

Table 4 EDX chemical composition of composition 1 and composition 2 as K-series

El AN Norm.C (wt%) Atom.C (at%) Error (1sigma)
C 6 51.87 60.91 7.67
O 8 40.86 36.02 6.78
Ca 20 4.16 1.46 0.16
Si 14 1.66 0.83 0.11
Al 13 0.86 0.45 0.08
Mg 12 0.45 0.26 0.07
Cl 17 0.14 0.05 0.04
P 15 0.01 0.01 0.03
C 6 57.57 67.78 11.84
O 8 30.06 26.57 7.28
Si 14 6.05 3.05 0.41
Al 13 1.78 0.94 0.17
Ca 20 1.76 0.62 0.11
Mg 12 1.01 0.58 0.13
Fe 26 1.64 0.42 0.11
K 19 0.13 0.05 0.04

4 Conclusion

In this investigation, a systematic procedure to evaluate and analyze the behavior of

Raavi and Pineapple fiber hybrid polymer composites was done under two differ-
ent percentages of compositions. The better tensile, flexural, and impact behaviors
of 32, 42.22 MPa, and 22.379 J/m were obtained for Raavi (15%)- and Pineapple
(15%)-reinforced polymer composites (Composition 1). The maximum tensile, flex-
ural, and impact behaviors of 20.1, 43.56 MPa, and 45.4 J/m were obtained for Raavi
(10%)- and Pineapple (20%)-reinforced polymer composites (Composition 2). The
238 J. Varun Siddharth et al.

microsturcture and chemical composition were studied by scanning electron micro-

scope. The mechanical behaviors studied through real-time experiments witnessed
that both composition of Raavi- and Pineapple-reinforced polymer composites can
be utilized as the replacement of plastics such low load applications because of its
ease of decomposability and low cost.

References

1. Van Vuure AW, Baets J, Wouters K, Hendrickx K (2015) Compressive properties of natural
fibre composites. Mater Lett 149:138–140
2. Chegdani F, El Mansori M (2018) Friction scale effect in drilling natural fiber composites.
Tribol Int 119:622–630
3. Vitale JP, Francucci G, Xiong J, Stocchi A (2017) Failure mode maps of natural and synthetic
fiber reinforced composite sandwich panels. Compos A Appl Sci Manuf 94:217–225
4. Torres JP, Vandi LJ, Veidt M, Heitzmann MT (2017) The mechanical properties of natural fibre
composite laminates: a statistical study. Compos A Appl Sci Manuf 98:99–104
5. Corona A, Madsen B, Hauschild MZ, Birkved M (2016) Natural fibre selection for composite
eco-design. CIRP Ann Manuf Technol 65(1):13–16
6. Song Z, Li S, Lei J, Noble RL, Wang Z (2018) Dynamic tensile properties of ROP/OCC natural
hybrid fibers reinforced composites. Compos Struct 185:600–606
7. Fan M, Naughton A (2016) Mechanisms of thermal decomposition of natural fibre composites.
Compos B Eng 88:1–10
8. Pornwannachai W, Ebdon JR, Kandola BK (2018) Fire-resistant natural fibre-reinforced
composites from flame retarded textiles. Polym Degradation Stab
9. Mejri M, Toubal L, Cuillière JC, François V (2018) Hygrothermal aging effects on mechanical
and fatigue behaviors of a short-natural-fiber-reinforced composite. Int J Fatigue 108:96–108
10. Chegdani F, El Mansori M (2018) Mechanics of material removal when cutting natural fiber
reinforced thermoplastic composites. Polym Testing 67:275–283
11. Vignesh G. Mechanical characterization of kenaf-hair reinforced hybrid composite
12. Choudhry S, Pandey B (2012) Mechanical behaviour of polypropylene and human hair fibres
and polypropylene reinforced polymeric composites. Int J Mech Ind Eng 2(1):118–121
13. Pothan LA, Oommen Z, Thomas S (2003) Dynamic mechanical analysis of banana fiber
reinforced polyester composites. Compos Sci Technol 63(2):283–293
14. Rana AK, Mandal A, Bandyopadhyay S (2003) Short jute fiber reinforced polypropylene
composites: effect of compatibiliser, impact modifier and fiber loading. Compos Sci Technol
63(6):801–806
15. Idicula M, Malhotra SK, Joseph K, Thomas S (2005) Dynamic mechanical analysis of randomly
oriented intimately mixed short banana/sisal hybrid fibre reinforced polyester composites.
Compos Sci Technol 65(7–8):1077–1087
Carbon Fiber Surface Treatment
for Enhanced Interfacial Properties:
A Review

Kwame Anane-Fenin, Esther Titilayo Akinlabi, and Nicolas Perry

Abstract Pristine carbon fibers are intrinsically brittle, fuzzy, non-polar and possess
graphite-based planes which are highly crystallized. These characteristics are limit-
ing to effective interfacial bonding between fiber and matrix. Therefore, the ability to
modify the surface of carbon fibers present an avenue for enhancing surface function-
alities and increasing surface energy which results in improved interfacial properties
(wettability) and mechanical properties. In this paper, the current novel techniques
for carbon fiber surface treatment are presented. The current direction of the field
is reviewed, and emerging technologies are discussed. The comparative study con-
ducted provides insight for optimal selection of treatment approaches depending on
the requirement.

Keywords Interfacial bonding · Surface treatment · Wettability

1 Introduction

Carbon fiber (CF) reinforced composites (CFRPs) are known to be very reliable and
find applications in varying multidisciplinary fields such as the aerospace, automo-
tive and sports equipment manufacturing industries as a result of the excellent and
superior mechanical strength and physical (lightweight) properties [1]. Unmodified
or virgin carbon fiber, however, has some limitations which include a very moderately
active specific surface area, non-polar properties and chemically inert characteristic,
low lipophobic surface, low energy surface and the inherently low weak interfacial
adhesion characteristic of the fibers which limits the efficient transfer of stresses

K. Anane-Fenin (B) · E. T. Akinlabi

Department of Mechanical Engineering Science, University of Johannesburg, Johannesburg,
South Africa
e-mail: lncs@springer.com
N. Perry
Arts et Métiers ParisTech, CNRS, 12M Bordeaux, Esplanade des Arts et Métiers, Talence, France
E. T. Akinlabi
Department of Mechanical Engineering, Coventry University Ota, Ota, Nigeria

© Springer Nature Singapore Pte Ltd. 2021 239

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_22
240 K. Anane-Fenin et al.

from the resin to the carbon fiber that consequently affects the mechanical integrity
of the fabricated composite [2]. Pristine carbon fibers are intrinsically brittle, fuzzy,
non-polar and possess graphite-based planes which are highly crystallized. These
characteristics are limiting to effective interfacial bonding between fiber and matrix
[3]. As a result, many CFRPs tend to demonstrate a propensity for failure within
the fiber-matrix interface due to the relatively weak bonding and eventually results
in the pullout effect which exposes the fiber from the matrix (pullout effect). This
phenomenon compromises the mechanical integrity of the composite material like
the interlaminar shear stress (ILSS).
Over a couple of decades, several techniques for modifying the surface of carbon
fibers to improve the interfacial properties of fabricated composites, include plasma
treatment [4], sizing [1], oxidation treatment [5], graphene oxide modification [6]
coating [7], high energy irradiation [8], chemical grafting [9], electrophoretic depo-
sition (EPD) [10], etc. All the above-stated methods have their own inherent insuffi-
ciencies. An optimal approach to fiber surface treatment is very critical to ensuring
that the right amount of fiber-matrix interface bonding is obtained to prevent issues of
extreme bonding which can result in brittleness [11]. Therefore, a carefully controlled
interface is vital to ensuring that the requisite amount is transferred from the matrix
to fiber reinforcements which ameliorate the induced internal stress concentrations
and also enhance the mechanical properties [12].
The ability to modify the surface characteristics of carbon fiber provides a pow-
erful optimization tool for enhancing the fiber-matrix adhesion properties of carbon
fiber reinforced polymer composites. Therefore, in this study, a thorough review of
the approach currently adopted for the modification of carbon fiber surface is pre-
sented. Comparison of the different methods is discussed, and the advantages and
disadvantages also reported.

1.1 Surface Sizing

In industry, the manufacture of carbon fiber is characterized by the activation of car-

bon fiber to achieve electrolytic surfaces before stabilization and carbonization. The
activation of carbon fiber by anodic oxidation is an important step critical for improv-
ing its surface reactivity due to the development of oxygen-rich functional groups
[13]. This significantly enhances wettability and promotes the implementation of
polymeric sizing, which is required for additional fiber processing. Furthermore, an
overall improvement in bonding between the matrix, polymeric sizing and fiber sur-
face is achieved [13]. Other advantages of anodic oxidation include the simplicity
of the processing methodology and cost-effectiveness. It is an oxidation technique
which inherently limits it to the creation of only oxygen-containing surface functional
groups. The initial intent of applying sizing agents (polymer-based) during commer-
cial carbon fibers manufacture was to provide a 100-nm-thick protective coating
against fiber damage and improve handling [14]. In recent times, the dimension of
exploring and developing sizing agents that also ensure enhanced interfacial bonding
Carbon Fiber Surface Treatment for Enhanced Interfacial … 241

between carbon fiber reinforcement and the matrix is a primary focus for manufac-
turers [15]. Therefore, sizing agents are designed to promote increased fiber-matrix
interlocking, enhanced wettability and improve the density of oxygen-containing
functional groups on the surface of the fiber. To enhance the interfacial adhesion
properties of fiber and matrix, nanoparticle inclusions in sizing agents have been the
most preferred approach; the nanoparticles explored include carbon nanotubes [16],
titanium oxide [17], graphene oxide [18] and zirconium dioxide [15].
Polyurethanes (PU) have also emerged as highly suitable sizing agents because
of advantages such as excellent mechanical properties (strength and toughness), low
viscosity, excellent abrasion resistance, outstanding interlocking adhesion with sev-
eral matrices, cost-efficient and environmentally sustainable [19, 20]. The application
of PUs for sizing has not been thoroughly explored as only a few publications are
available.
The most industrially preferred approach for surface modification is sizing treat-
ment methods and the reasons include simplicity and ease of process operation,
moderate environmental impact, flexibility in tunability and controllability process
is continuous and seamlessly incorporates into the overall manufacture of carbon
fibers [21].

1.2 Plasma Treatment

In contradistinction, conceptually, carbon fiber surface modification using plasma

permits for a wider spectrum of surface functionalization options [13]. During plasma
treatment an interaction between the carbon fiber surface, neutral and excited free
electrons, radicals, ions, atoms or molecules takes place. As a result, four forms
of surface modifications can occur, namely activation, layer deposition, function-
alization or etching. The chemical property and constitution of the precursor gas
for plasma treatment coupled with selected plasma processing parameters determine
the topographic, morphological and chemical characteristics of the carbon fiber. To
improve the fiber-matrix (epoxy) bonding, oxygen or air plasma functionalization
is usually conducted [22]. The resulting fiber surface oxidation is characterized by
the development of oxygen-carrying functional groups equivalent to those observed
after anodic oxidation.
Another promising approach used for enhancing the adhesion between fiber and
matrix is via the modification of the fiber surface through the formation of functional
groups with nitrogen carrying elements. A high chemical interaction between the
amine functional groups on the fiber surface and the epoxy resin groups is envisaged
[23]. The carboxylic and alcohol groups are particularly of great interest as they
are known to react chemically with the epoxy groups within the matrix. In cases
where amine hardeners are used, the fiber surface carbonyl groups are particularly
very reactive to amine groups in general [23] Several studies [24] have shown that
the carbon fiber-epoxy matrix bonding was significantly improved after oxygen or
air plasma functionalization. The drawback of using air or oxygen plasma is the
242 K. Anane-Fenin et al.

deterioration in mechanical strength of the fiber associated as a result of the damage

inflicted on the fiber during the process [25].
Ammonia [26] and nitrogen [27] plasma treatment are often used to introduce
nitrogen carrying functionalities onto the surface of carbon fibers. Ammonia particu-
larly is an attractive precursor gas alternative for achieving nitrogen functionalization
due to its capacity to form very high ratios of amine and imine functional groups on
the surface of the carbon fiber [28] while avoiding the drawback of tensile strength
degradation common with oxygen plasma treatment [27]. An improvement of interfa-
cial bonding of fibers to bismaleimide resin, initially untreated and then treated with
ammonia plasma is reported [25, 29]. Moosburger-Will et al. [30], functionalized
the surface of unmodified carbon fibers, extracted immediately after carbonization
using ammonia plasma in a low pressure environment straight. In comparison with
untreated carbon fibers, an improvement in surface energy and nitrogen functionality
concentration at the surface of treated fibers were reported. The enhanced wettability
is comparable to activated carbon fibers processed using anodic oxidation.

1.3 Electro-grafting/polymerization

Electro-grafting or electro-polymerization is a technique that has been widely used

for depositing functional coatings of polymer homogenously on to the surface of the
carbon fiber. Electro-polymerization has several advantages, these include the ability
to optimize the desired coating through controlling and tuning critical parameters
and conditions such as polymerization time, monomer concentration, homogeneity
and thickness of coating [31]. Subramanian et al. [32] were able to clearly estab-
lish the influence of electro-polymerization parameters like polymerization time,
monomer functionality, selected solvent, post-coating treatment and fiber content on
the mechanical properties of the coated CF. Some studies have observed that electro-
polymerization establishes an interlayer that lies between the matrix and fiber capable
of energy absorption from internal cracks within the composite which consequently
results in a reduction or elimination of the total induced stress concentration [33].
Polymethacrylic acid (PMAA) polymer, when used for electro-grafting of
monomers onto the carbon fiber, is reported to have exhibited the best tensile strength
of fabricated composites [34]. These superior characteristics can be attributed to
hydrogen bonds that form between the hydroxyl groups of the matrix and the car-
boxyl groups of PMAA, thereby forming an improved interfacial adhesion with-
out the drawback of brittleness. In a recent study by Semitekolos et al. [35], the
improvement of interfacial strength between the fiber-matrix bond within carbon
fiber reinforced composites was accomplished through the reaction of carbon fiber
and polymethacrylic acid by a process known as electro-polymerization.
Carbon Fiber Surface Treatment for Enhanced Interfacial … 243

1.4 Grafting

The inclusion of nanoscale materials (e.g., nanoparticles, carbon nanotubes, nan-

oclay, nanofiber, graphene oxide-GO, metals and nonmetals) in several current
research onto the surface of carbon fibers to improve the interfacial strength without
degrading the mechanical properties of composites is a standard [6, 36]. Grafting of
graphene oxide (GO) homogenously onto carbon fiber is an efficient technique for
achieving hierarchical intralaminar and interlaminar reinforcement that introduce
superior fiber-matrix bonding and enhanced overall mechanical properties to the
composite material [37]. The novelty of using graphene oxide (GO) lies in its two-
dimensional configuration coupled with its superior mechanical and thermal prop-
erties [38]. Furthermore, GO can be prepared in combination with several oxygen
functional groups in substantial amounts at a reasonably reduced cost [39].

1.5 Radiation Treatment

Laser radiation treatment has also been successfully adapted for enhancing the fiber-
matrix bond while additionally improving the mechanical properties of the matrix.
In a study by Fischera et al. [40], CO2 and UV lasers were tested for carbon fiber
surface treatment which showed that both lasers equally improved the interfacial
bond strength between the reinforcement and matrix.

1.6 Coating

Studies on the application of polydopamine (PDA) coating have shown it provides a

flexible platform for functionalizing the surface of several substrates without degrad-
ing the mechanical properties. The impressive chemical properties such as the cat-
echol functional group are mainly responsible for superior adhesiveness [41]. PDA
coating treatment has mostly been conducted on glass fibers [42]; however, in a recent
study by Kim and Song [43], the influence of PDA on surface modification of carbon
fiber was assessed. They observed an increase in Young’s modulus of PDA coated
carbon fiber reinforced composite. Carboxyl functional group formation on fiber
surfaces occur during oxygen plasma treatment, while hydrogen plasma treatment
induces the formation of hydroxyl bonds facilitated by sp2 etching which provides
the anchoring platform and also produces superior hydrophilic surfaces [44].
244 K. Anane-Fenin et al.

1.7 Layer-by-Layer (LbL) Self-assembly Technique

Most of the carbon fiber surface modification techniques for the inclusion of
nanoscale materials are not environmentally friendly or green as toxic chemicals
and catalysts are required [45]. In contrast, the layer-by-layer (LbL) self-assembly
technique is generally considered a green approach because it is environmentally
friendly, high flexibility for optimization and tuning, moderate processing condi-
tions and excellent repeatability [45]. This approach basically operates on a cycli-
cally based concept where alternating film deposition of materials that are oppositely
charged takes place.

1.8 Covalent Functionalization

The two main limitations of untreated carbon fiber are the reduced interfacial adhe-
sion of fiber and matrix and relatively high conductivity which is possibly detrimental
to electric machines. Most of the available treatment methods for surface modifica-
tion can resolve the interfacial adhesion issue; however, covalent functionalization
is one of the few methods capable of decreasing carbon fiber conductivity as pre-
sented in the study by [46]. They achieved CF surface modification via covalent
functionalization involving electrochemical reduction of phenyl-diazo species.

1.9 Microwave Irradiation

Microwave irradiation is another emerging tool for modifying the surface of carbon
fiber made possible by carbon fiber’s outstanding microwave absorbability. Wang
et al. [47] developed a microwave irradiation method that was carried out in air
atmosphere, however, some drawbacks reported included some level of burnout and
degradation. A study by [48], avoided the limitations by conducting the irradiation
via immersion of the carbon fiber. The microwave etching process resulted in large
formation oxygen-containing groups on the surface of the carbon fibers.

2 Results and Discussion

A comparison of the different surface treatment is presented in Table 1. In comparing

the various methods, electrochemical and wet oxidation treatments are the least
preferred due to the environmental issues associated with chemical use and the time-
consuming nature of the methods [49]. However, plasma treatment impacts both
physical and chemical properties of carbon fibers without significantly altering the
Table 1 Carbon fiber surface morphological treatments
Surface morphology treatment Treatment agent Enhancement Comment Author
Plasma Ammonia (1) High surface energy Nitrogen functionalization was [30]
(nanoporosity) successfully conducted. Nitrogen
(2) Enhanced wettability functional groups did not
contribute to improvement in
fiber-matrix bonding
Plasma Hydrogen Improvement in mechanical All mechanical properties [49]
properties such as modulus of significantly improved with 30%
elasticity, yield strength, tensile carbon fiber content increase
and compressive stresses
Electro-polymerization Polymethacrylic acid (1) 36% Improvement in The process is cost-efficient and [35]
interlaminar shear strength (ILSS) environmentally friendly
(2) A 3% polymer mass increase
significantly enhanced the
wettability
Chemical grafting GO and CNTs The CF-GO-CNTs composites The excellent interfacial strength [9]
showed significant improvement was attributed to good wettability
Carbon Fiber Surface Treatment for Enhanced Interfacial …

of 48% and 83% for the ILSS and and surface energy, improved
IFSS, respectively fiber-matrix interlocking and
interface chemical bonding
(continued)
245
Table 1 (continued)
246

Surface morphology treatment Treatment agent Enhancement Comment Author

Electrochemical grafting Covalent attachment of small The functionalization of carbon The reduction of carbon fiber [46]
molecules fiber surface using small conductivity was explored
molecules improved the
resistivity by 200% while the
interfacial shear properties
improved by over 19%
Surface sizing Sizing agents Heat treatment at 130 °C and Increased fibers-sizing interaction [15]
150 °C resulted in significant is detrimental to the sizing-matrix
improvement of 71.1% and bonding
15.1%, respectively, on interfacial
shear strength (IFSS)
Layer-by-layer self-assembly GO/SiO2 multilayers films The interfacial shear strength The researchers claim that the [45]
method (IFSS), interlaminar shear approach can be classified as
strength (ILSS) and flexural green and provides a viable
strength of composites improved alternative for the preparation of
by 86.1%, by 89.3% and 30.4%, high interfacial properties
respectively
Coating Polydopamine The mechanical properties such The highest mechanical [43]
as bending strength increased by performance was observed at
11% while significantly improved PDA concentrations of 1.5 and
tensile strength and Young’s 1 g/L and TRIS concentration of
modulus were recorded 40 mM
(continued)
K. Anane-Fenin et al.
Table 1 (continued)
Surface morphology treatment Treatment agent Enhancement Comment Author
Chemical grafting Nylon 6 (PA6) The grafting of Nylon 6 (PA6) Enhanced wettability, fiber-matrix [56]
onto the carbon fiber surface, interfacial adhesion, mechanical
improved the tensile strength of interlocking and chemical
the composites by 30.8% interaction was achieved when
grafting was performed on
oxidized carbon fibers
Microwave irradiation Etching Although the tensile strength of The microwave irradiation was [48]
treated carbon fibers was slightly conducted with the carbon fibers
deteriorated, A significant immersed in water
improvement in the interfacial
shear strength was observed while
the tensile strength degraded
slightly
Carbon Fiber Surface Treatment for Enhanced Interfacial …
247
248 K. Anane-Fenin et al.

bulk mechanical properties [8]. Although oxygen plasma treatment as reported by

Montes-Moran et al. [50] and Fu et al. [51] shows high functional group density
and improved interfacial shear strength, it can cause some fiber degradation. Argon
plasma treatment also has detrimental effects on the fiber; however, argon/ammonia
plasma improves fiber-matrix bonding without degrading the strength of the fiber
[29]. The most preferred method, however, is sizing mainly because of its simplicity
and flexibility [21]. Although a huge amount of research is going on, the surface
modification techniques still require extensive research particularly in areas such as
pre-treatments using infrared- and ultraviolet-lasers [52], solvent cleaning methods
[53], acidic etching [54], peel-ply and abrasion [53] and plasma treatments [55].

3 Conclusion

An exhaustive review of the current methods for carbon fiber surface treatment has
been conducted. A comparison of the techniques is carried out and tabulated. The
advantages and disadvantages of the methods are also presented. All the methods
still require further research and tuning to achieve optimal conditions. The chemical
approaches are efficient but are mostly not environmentally friendly. It is envis-
aged that as technology advances more environmentally sensitive approaches will
be developed.

References

1. Yuan X, Zhu B, Cai X, Liu J, Qiao K, Yu J (2017) Optimization of interfacial properties of

carbon fiber/epoxy composites via a modified polyacrylate emulsion sizing. Appl Surf Sci
401:414–423. https://doi.org/10.1016/J.APSUSC.2016.12.234
2. Lin SP, Han JL, Yeh JT, Chang FC, Hsieh KH (2007) Composites of UHMWPE fiber reinforced
PU/epoxy grafted interpenetrating polymer networks. Eur Polym J 43:996–1008. https://doi.
org/10.1016/j.eurpolymj.2006.12.001
3. Zabihi O, Ahmadi M, Li Q, Shafei S, Huson MG, Naebe M (2017) Carbon fibre surface mod-
ification using functionalized nanoclay: a hierarchical interphase for fibre-reinforced poly-
mer composites. Compos Sci Technol 148:49–58. https://doi.org/10.1016/j.compscitech.2017.
05.013
4. Li W, Yao SY, Ma KM, Chen P (2013) Effect of plasma modification on the mechanical
properties of carbon fiber/phenolphthalein polyaryletherketone composites. Polym Compos
34:368–375. https://doi.org/10.1002/pc.22385
5. Liu J, Tian Y, Chen Y, Liang J, Zhang L, Fong H (2010) A surface treatment technique of
electrochemical oxidation to simultaneously improve the interfacial bonding strength and the
tensile strength of PAN-based carbon fibers. Mater Chem Phys 122:548–555. https://doi.org/
10.1016/J.MATCHEMPHYS.2010.03.045
6. Zhao Z, Teng K, Li N, Li X, Xu Z, Chen L, Niu J, Fu H, Zhao L, Liu Y (2017) Mechan-
ical, thermal and interfacial performances of carbon fiber reinforced composites flavored by
carbon nanotube in matrix/interface. Compos Struct 159:761–772. https://doi.org/10.1016/J.
COMPSTRUCT.2016.10.022
Carbon Fiber Surface Treatment for Enhanced Interfacial … 249

7. Li F, Liu Y, Qu C-B, Xiao H-M, Hua Y, Sui GX, Fu S-Y (2015) Enhanced mechanical properties
of short carbon fiber reinforced polyethersulfone composites by graphene oxide coating. Polym
(Guildf) 59:155–165. https://doi.org/10.1016/J.POLYMER.2014.12.067
8. Tiwari S, Bijwe J (2014) Surface treatment of carbon fibers—a review. Procedia Technol
14:505–512. https://doi.org/10.1016/j.protcy.2014.08.064
9. Gao B, Zhang R, He M, Sun L, Wang C, Liu L, Zhao L, Cui H, Cao A (2016) Effect of a mul-
tiscale reinforcement by carbon fiber surface treatment with graphene oxide/carbon nanotubes
on the mechanical properties of reinforced carbon/carbon composites. Compos Part A Appl
Sci Manuf 90:433–440. https://doi.org/10.1016/j.compositesa.2016.08.012
10. Sui X, Shi J, Yao H, Xu Z, Chen L, Li X, Ma M, Kuang L, Fu H, Deng H (2017) Interfacial and
fatigue-resistant synergetic enhancement of carbon fiber/epoxy hierarchical composites via an
electrophoresis deposited carbon nanotube-toughened transition layer. Compos Part A Appl
Sci Manuf 92:134–144. https://doi.org/10.1016/J.COMPOSITESA.2016.11.004
11. Morgan P (2005) Carbon fibers and their composites. Taylor & Francis. https://doi.org/10.
1016/j.tube.2011.07.006
12. Paul R, Dai L (2018) Interfacial aspects of carbon composites. https://www.tandfonline.com/
doi/full/10.1080/09276440.2018.1439632. https://doi.org/10.1080/09276440.2018.1439632
13. Sharma M, Gao S Mäder E, Sharma H, Wei LY, Bijwe J (2014) Carbon fiber surfaces and com-
posite interphases. https://www.sciencedirect.com/science/article/pii/S0266353814002358.
https://doi.org/10.1016/j.compscitech.2014.07.005
14. Drzal LT, Rich MJ, Koenig MF, Lloyd PF (1983) Adhesion of graphite fibers to epoxy matrices:
II. The effect of fiber finish. J Adhes 16:3–152. https://doi.org/10.1080/00218468308074911
15. Wu Q, Zhao R, Ma Q, Zhu J (2018) Effects of degree of chemical interaction between carbon
fibers and surface sizing on interfacial properties of epoxy composites. Compos Sci Technol
163:34–40. https://doi.org/10.1016/j.compscitech.2018.05.013
16. Kamae T, Drzal LT (2012) Carbon fiber/epoxy composite property enhancement through incor-
poration of carbon nanotubes at the fiber-matrix interphase—Part I: The development of carbon
nanotube coated carbon fibers and the evaluation of their adhesion. Compos Part A Appl Sci
Manuf 43:1569–1577. https://doi.org/10.1016/j.compositesa.2012.02.016
17. Qian J, Wu J, Liu X, Zhuang Q, Han Z (2013) Improvement of interfacial shear strengths
of polybenzobisoxazole fiber/epoxy resin composite by n-TiO2 coating. J Appl Polym Sci
127:2990–2995. https://doi.org/10.1002/app.37970
18. Qin W, Vautard F, Drzal LT, Yu J (2015) Mechanical and electrical properties of carbon fiber
composites with incorporation of graphene nanoplatelets at the fiber-matrix interphase. Compos
Part B Eng 69:335–341. https://doi.org/10.1016/j.compositesb.2014.10.014
19. Loos MR, Yang J, Feke DL, Manas-Zloczower I, Unal S, Younes U (2013) Enhancement
of fatigue life of polyurethane composites containing carbon nanotubes. Compos Part B Eng
44:740–744. https://doi.org/10.1016/J.COMPOSITESB.2012.01.038
20. Pokharel P, Lee DS (2014) High performance polyurethane nanocomposite films prepared from
a masterbatch of graphene oxide in polyether polyol. Chem Eng J 253:356–365. https://doi.
org/10.1016/J.CEJ.2014.05.046
21. Wang Z, Huang X, Xian G, Li H (2016) Effects of surface treatment of carbon fiber: Tensile
property, surface characteristics, and bonding to epoxy. Polym Compos 37:2921–2932. https://
doi.org/10.1002/pc.23489
22. Baghery Borooj M, Mousavi Shoushtari A, Nosratian Sabet E, Haji A (2016) Influence of
oxygen plasma treatment parameters on the properties of carbon fiber. J Adhes Sci Technol
30:2372–2382. https://doi.org/10.1080/01694243.2016.1182833
23. Moosburger-Will J, Jäger J, Strauch J, Bauer M, Strobl S, Linscheid FF, Horn S (2017) Inter-
phase formation and fiber matrix adhesion in carbon fiber reinforced epoxy resin: influence
of carbon fiber surface chemistry. Compos Interfaces 24:691–710. https://doi.org/10.1080/
09276440.2017.1267513
24. Baghery Borooj M, Mousavi Shoushtari A, Haji A, Nosratian Sabet E (2016) Optimization of
plasma treatment variables for the improvement of carbon fibres/epoxy composite performance
by response surface methodology. Compos Sci Technol 128:215–221. https://doi.org/10.1016/
j.compscitech.2016.03.020
250 K. Anane-Fenin et al.

25. Chang T (1999) Plasma surface treatment in composites manufacturing. J Ind Technol 15:1–7
26. Farrow GJ, Jones C (1994) The effect of low power nitrogen plasma treatment of carbon fibres
on the interfacial shear strength of carbon fibre/epoxy composites. J Adhes 45:29–42. https://
doi.org/10.1080/00218469408026627
27. Corujeira Gallo S, Charitidis C, Dong H (2017) Surface functionalization of carbon fibers with
active screen plasma. J Vac Sci Technol A Vac Surf Film 35:021404. https://doi.org/10.1116/
1.4974913
28. Jones C, Sammann E (1990) The effect of low power plasmas on carbon fibre surfaces. Carbon
N Y 28:509–514. https://doi.org/10.1016/0008-6223(90)90046-2
29. Jang BZ (1992) Control of interfacial adhesion in continuous carbon and kevlar fiber rein-
forced polymer composites. Compos Sci Technol 44:333–349. https://doi.org/10.1016/0266-
3538(92)90070-J
30. Moosburger-Will J, Lachner E, Löffler M, Kunzmann C, Greisel M, Ruhland K, Horn S (2018)
Adhesion of carbon fibers to amine hardened epoxy resin: influence of ammonia plasma func-
tionalization of carbon fibers. Appl Surf Sci 453:141–152. https://doi.org/10.1016/j.apsusc.
2018.05.057
31. Bousquet A, Awada H, Hiorns RC, Dagron-Lartigau C, Billon L (2014) Conjugated-polymer
grafting on inorganic and organic substrates: a new trend in organic electronic mate-
rials. https://www.sciencedirect.com/science/article/pii/S0079670014000380. https://doi.org/
10.1016/j.progpolymsci.2014.03.003
32. Subramanian RV, Jakubowski JJ (1978) Electropolymerization on graphite fibers. Polym Eng
Sci 18:590–600. https://doi.org/10.1002/pen.760180708
33. Harris B, Braddell OG, Almond DP, Lefebvre C, Verbist J (1993) Study of carbon fibre surface
treatments by dynamic mechanical analysis. J Mater Sci 28:3353–3366. https://doi.org/10.
1007/BF00354259
34. Kainourgios P, Kartsonakis IA, Dragatogiannis DA, Koumoulos EP, Goulis P, Charitidis CA
(2017) Electrochemical surface functionalization of carbon fibers for chemical affinity improve-
ment with epoxy resins. Appl Surf Sci 416:593–604. https://doi.org/10.1016/j.apsusc.2017.
04.214
35. Semitekolos D, Kainourgios P, Jones C, Rana A, Koumoulos EP, Charitidis CA (2018)
Advanced carbon fibre composites via poly methacrylic acid surface treatment; surface analy-
sis and mechanical properties investigation. Compos Part B Eng 155:237–243. https://doi.org/
10.1016/j.compositesb.2018.08.027
36. Luo W, Zhang B, Zou H, Liang M, Chen Y (2017) Enhanced interfacial adhesion between
polypropylene and carbon fiber by graphene oxide/polyethyleneimine coating. J Ind Eng Chem
51:129–139. https://doi.org/10.1016/j.jiec.2017.02.024
37. Zhang RL, Gao B, Ma QH, Zhang J, Cui HZ, Liu L (2016) Directly grafting graphene oxide
onto carbon fiber and the effect on the mechanical properties of carbon fiber composites. Mater
Des 93:364–369. https://doi.org/10.1016/j.matdes.2016.01.003
38. Steurer P, Wissert R, Thomann R, Mülhaupt R (2009) Functionalized graphenes and ther-
moplastic nanocomposites based upon expanded graphite oxide. Macromol Rapid Commun
30:316–327. https://doi.org/10.1002/marc.200800754
39. Wu Y, Lin X, Shen X, Sun X, Liu X, Wang Z, Kim JK (2015) Exceptional dielectric properties
of chlorine-doped graphene oxide/poly (vinylidene fluoride) nanocomposites. Carbon N Y
89:102–112. https://doi.org/10.1016/j.carbon.2015.02.074
40. Fischer F, Kreling S, Jäschke P, Frauenhofer M, Kracht D, Dilger K (2012) Laser surface pre-
treatment of CFRP for adhesive bonding in consideration of the absorption behaviour. J Adhes
88:350–363. https://doi.org/10.1080/00218464.2012.660042
41. Ran B, Jing C, Yang C, Li X, Li Y (2018) Synthesis of efficient bacterial adhesion-resistant
coatings by one-step polydopamine-assisted deposition of branched polyethylenimine-g-
poly(sulfobetaine methacrylate) copolymers. Appl Surf Sci 450:77–84. https://doi.org/10.
1016/J.APSUSC.2018.04.184
42. Xiong S, Wang Y, Zhu J, Yu J, Hu Z (2016) Poly(ε-caprolactone)-grafted polydopamine par-
ticles for biocomposites with near-infrared light triggered self-healing ability. Polym (Guildf).
84:328–335. https://doi.org/10.1016/J.POLYMER.2016.01.005
Carbon Fiber Surface Treatment for Enhanced Interfacial … 251

43. Kim HJ, Song JH (2019) Improvement in the mechanical properties of carbon and aramid
composites by fiber surface modification using polydopamine. Compos Part B Eng 160:31–36.
https://doi.org/10.1016/j.compositesb.2018.10.027
44. Merenda A, Ligneris E des, Sears K, Chaffraix T, Magniez K, Cornu D, Schütz JA, Dumée
LF (2016) Assessing the temporal stability of surface functional groups introduced by plasma
treatments on the outer shells of carbon nanotubes. Sci Rep 6:31565. https://doi.org/10.1038/
srep31565
45. Fu J, Zhang M, Jin L, Liu L, Li N, Shang L, Li M, Xiao L, Ao Y (2019) Enhancing interfacial
properties of carbon fibers reinforced epoxy composites via Layer-by-Layer self assembly
GO/SiO2 multilayers films on carbon fibers surface. Appl Surf Sci 470:543–554. https://doi.
org/10.1016/j.apsusc.2018.11.168
46. Beggs KM, Randall JD, Servinis L, Krajewski A, Denning R, Henderson LC (2018) Increasing
the resistivity and IFSS of unsized carbon fibre by covalent surface modification. React Funct
Polym 129:123–128. https://doi.org/10.1016/j.reactfunctpolym.2017.06.016
47. Wang B, Duan Y, Zhang J, Zhao X (2016) Microwave radiation effects on carbon fibres interfa-
cial performance. Compos Part B Eng 99:398–406. https://doi.org/10.1016/J.COMPOSITESB.
2016.06.032
48. Yuan JM, Fan ZF, Yang QC, Li W, Wu ZJ (2018) Surface modification of carbon fibers by
microwave etching for epoxy resin composite. Compos Sci Technol 164:222–228. https://doi.
org/10.1016/j.compscitech.2018.05.043
49. Lee ES, Lee CH, Chun YS, Han CJ, Lim DS (2017) Effect of hydrogen plasma-mediated surface
modification of carbon fibers on the mechanical properties of carbon-fiber-reinforced polyether-
imide composites. Compos Part B Eng 116:451–458. https://doi.org/10.1016/j.compositesb.
2016.10.088
50. Montes-Morán MA, Martínez-Alonso A, Tascón JMD, Paiva MC, Bernardo CA (2001) Effects
of plasma oxidation on the surface and interfacial properties of carbon fibres/polycarbonate
composites. Carbon N Y 39:1057–1068. https://doi.org/10.1016/S0008-6223(00)00220-7
51. Fu YF, Xu K, Li J, Sun ZY, Zhang FQ, Chen DM (2012) The influence of plasma surface
treatment of carbon fibers on the interfacial adhesion properties of UHMWPE composite.
Polym Plast Technol Eng 51:273–276. https://doi.org/10.1080/03602559.2011.617406
52. Reitz V, Meinhard D, Ruck S, Riegel H, Knoblauch V (2017) A comparison of IR- and UV-laser
pretreatment to increase the bonding strength of adhesively joined aluminum/CFRP compo-
nents. Compos Part A Appl Sci Manuf 96:18–27. https://doi.org/10.1016/J.COMPOSITESA.
2017.02.014
53. Encinas N, Oakley BR, Belcher MA, Blohowiak KY, Dillingham RG, Abenojar J, Martínez
MA (2014) Surface modification of aircraft used composites for adhesive bonding. Int J Adhes
Adhes 50:157–163. https://doi.org/10.1016/J.IJADHADH.2014.01.004
54. Wetzel M, Holtmannspötter J, Gudladt H-J, Czarnecki JV (2013) Sensitivity of double can-
tilever beam test to surface contamination and surface pretreatment. Int J Adhes Adhes
46:114–121. https://doi.org/10.1016/J.IJADHADH.2013.06.002
55. Gude MR, Prolongo SG, Ureña A (2012) Adhesive bonding of carbon fibre/epoxy laminates:
correlation between surface and mechanical properties. Surf Coatings Technol 207:602–607.
https://doi.org/10.1016/J.SURFCOAT.2012.07.085
56. Hui C, Qingyu C, Jing W, Xiaohong X, Hongbo L, Zhanjun L (2018) Interfacial enhancement
of carbon fiber/nylon 12 composites by grafting nylon 6 to the surface of carbon fiber. Appl
Surf Sci 441:538–545. https://doi.org/10.1016/j.apsusc.2018.01.158
Preparation and Mechanical Property
Analysis of Polymer Matrix Composite
Containing Rice Husk and Saw Dust

Sarojrani Pattnaik, Arnab Sengupta, and Mihir Kumar Sutar

Abstract Nowadays, there is a great demand for natural/wood fiber composites in

different areas. Rice husk (RH) and saw dust (SD) are profusely available in India.
In the present study, RH and SD reinforced polymer matrix composite has been
prepared where both materials shared equal volume fraction in four equal categories
of polymer samples containing 30, 35, 40 and 45% of filler content. The epoxy
resin (HSC 9221) and hardener (EH 411) are mixed in 2:1 ratio. Polymeric samples
prepared for each category have been tested for tensile test, Vickers hardness test
and surface roughness. The results indicated that the tensile strength increased with
percent increase in reinforcing material from 30 to 40%, thereby it deceased slightly.
The highest tensile strength obtained was 57.7 MPa at 40% filler content. The average
hardness of composite was found to be more for 45% filler loading. However, the
hardness of polymeric sample containing 40% filler loading at different locations
was found to be more uniform than that obtained at 45% filler loading and the values
ranged from 7 to 7.4 HV. The surface roughness (SR) was found to be maximum
(0.79 µm) at 45% filler loading due to agglomeration of filler content at few places.
It was found to be minimum (0.64 µm) at 35% filler loading.

Keywords Polymer matrix composite · Saw dust · Rice husk · Hardness · Tensile
strength

1 Introduction

Nowadays, the application of polymer matrix composites (PMCs) in different sectors

such as transportation, aerospace industries, packaging, etc., have been increased
drastically owing to their good corrosion resistance, low density, translucency, high
stiffness, etc [1]. Natural fibers are reinforced with PMCs to enhance their mechanical
and electrical properties. These natural fibers are light in weight, cheaply available,

S. Pattnaik · A. Sengupta · M. K. Sutar (B)

Mechanical Engineering Department, Veer Surendra Sai University of Technology,
Burla, Odisha, India
e-mail: mihirsutar05@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 253

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_23
254 S. Pattnaik et al.

biodegradable and contain no toxic products. Thus, they are frequently inserted in
polymer matrix to form fiber-reinforced PMCs. Till date, commonly used natural
fibers employed in PMCs are flax fiber, jute, date palm leaf fiber, hemp, sisal, coir,
cotton, saw dust, rice husk, etc [2].
Yang et al. [3] used rice husk (RH) flour as filler material in preparing thermo-
plastic composite. The authors did four levels of filler loading and tested the mor-
phology and mechanical properties of the developed composites. It was found that
with increase of filler loading, tensile strengths of the composite decreased, whereas
tensile modulus increased. Chen et al. [4] prepared high loading of RH containing
recycled thermoplastic green composites. The authors found that the fire-retardancy
and mechanical properties of the composites were greatly increased with the usage
of rice husk at higher loadings. However, it adversely affected the dimensional sta-
bility of the composite. Monteiro et al. [5] studied the effects of RH ash and graphite
as fillers separately in polypropylene (PP)-based composites. A good bonding of
aforementioned fillers in the polymeric matrix was found.
Xu et al. [6] prepared perforated high-density polyethylene-based composites
made from recycled rubber and fiber-board sawdust (SD) with different loadings
(18, 40 and 42%) for sound performance testing. It was found that the acoustic prop-
erties of the prepared composite depended on hole-size, cavity depth and perforation
ratio. Suarez et al. [7] studied the tensile properties plus the fracture phenomenon of
sawdust containing polypropylene (PP) or PP together with maleated PP composites.
The tensile strength of the composites containing increased amount of MAPP to the
PP matrix was found to be enhanced and also, they exhibited improved bonding of
sawdust to the PP matrix.
From the above literature survey, it can be seen that the natural fibers, namely rice
husk or saw dust have been proved as a good source of filler materials in preparing
polymer matrix composites. However, till date, the combination of both aforemen-
tioned natural products has not been utilized in preparing a composite material.
Thus, the present work focuses on using both RH and SD as reinforcing materials
for developing epoxy resin base composites.

2 Materials and Methods

The rice husk obtained from the winnowing process was sieved in order to remove
the impurities from it. Saw dust was obtained from the carpentry shop. The epoxy
resin used in the present study was HSC-9221 and the hardener used was EH-411.
Silica powder was also used along with resin and hardener to act as support material
at high temperatures. The composites were made using a compression mold and
made up of mild steel.
First, the volume fraction of fibers (RH & SD) and resin was calculated. The
volume fraction of fibers was equally divided for RH & SD. Then, the epoxy and
Preparation and Mechanical Property Analysis of Polymer Matrix … 255

hardener were mixed in 2:1 ratio and the silica powder was added in such a manner
that it would be 33% of epoxy resin. Composites containing 30, 35, 40 and 45% fiber
volume fractions were made, as shown in Fig. 1 and they were coded as Type 1, 2,
3 and 4, respectively.
Then, the specimens were prepared for conducting tensile tests from each of the
composite types as shown in Fig. 2. The tensile test was performed on panel controlled
digital universal testing machine (UTM) with hydraulic open type grip (Fig. 3). The
sample hardness was measured at different positions using a digital microhardness
tester. The surface roughness of the composites was measured using Taylor Hobson
Talysurf surface roughness tester (shown in Fig. 4). Several readings were taken by
moving the stylus tip on the surface of the workpiece, and finally, the average of the
surface roughness readings of each of the composite types were computed for the
sake of comparison.

Fig. 1 Composites containing variable volume fractions of fibers a 30%, b 35%, c 40% and d 45%

Fig. 2 Tensile test specimens of composite types

256 S. Pattnaik et al.

Fig. 3 Tensile test performed on prepared composite

Fig. 4 Surface roughness testing of composite specimen using Talysurf

3 Results and Discussion

The tensile strengths of the composites of Types I, II, III and IV are shown in Fig. 5
and were found to be 51.7, 52.4, 57.7 and 53.7 MPa, respectively. Figure 5 depicts
the tensile test results of each type of composites. It is seen that the tensile strengths
of the composites increase with increase in filler content from 30 to 40 wt%, and
thereafter, it decreases. The average microhardness test results of all composite types
are depicted in Table 1. It is very obvious from the table that there is an increasing
trend of hardness values with increase in filler percent from 30 to 45 wt%, and Type
IV composite containing 45% filler material possessed maximum Vicker’s hardness
of 9.1 HV (Fig. 6), whereas average hardness value is 8.86 HV.
From the above results obtained for tensile strength and hardness tests, it is evi-
dent that up to 40 wt% of fillers loading, both the tensile strength and hardness of
the composites increased. The Type III composite containing 40% fibers exhibited
maximum tensile strength among all composite types because of highest interfacial
adhesion between the fiber and the matrix which led to efficient load transfer between
them. With rise in fiber content, the reinforcement in the matrix increases causing
increase in hardness of the composites.
It is found that at 45 wt% filler content, the tensile strength of the composite
decreased, whereas its microhardness is increased. The reason behind decline in
tensile strength may be due to the weak chemical reaction at the interface to shift
Preparation and Mechanical Property Analysis of Polymer Matrix … 257

Fig. 5 Tensile test results of a Type I, b Type II, c Type III, d Type IV composite specimen

Table 1 Microhardness test

Sl.No Volume fraction of fibers Vickers hardness (HV)
results of all the composites
(%) avg. value
1. 30 5.56
2. 35 7.00
3. 40 7.26
4. 45 8.86

the tensile stress amid the matrix and the particulates. Also, there is a possibility of
non-uniform distribution of reinforcing fillers in the matrix due to their increased
wt. percent and these fillers would have somewhere agglomerated leading to delam-
ination. Hardness test is done by pressing the filler particles and the epoxy matrix
together so that both have a tight grip which facilitates easy transfer of pressure,
causing increase in hardness of the composite.
The results of surface texture of all the prepared composites are depicted in Table 2.
258 S. Pattnaik et al.

Fig. 6 Microhardness test conducted on composite of Type IV

Table 2 Surface roughness

Type Volume fraction of fibers Average roughness (Ra)
of all the composites
(%) (µm)
I 30 0.66
II 35 0.64
III 40 0.72
IV 45 0.79

It is seen that Type II composite containing 35% filler material showed maximum
surface finish as regards to others, due to uniform distribution of fillers in the matrix.
During the manufacture of the composites, few defects were detected. Few cavities
were seen on the surface of the composite because of the sliding of the upper plate
at the time of loading. Overflow of polymer also took place due to its high percent
in the composition. A line was formed on one side of the composite due to slipping
of the Mylar sheet.

4 Conclusions

From all the different types of tests performed on the fiber-reinforced epoxy compos-
ites containing saw dust and rice husk, final results are accomplished. The following
observations are concluded from the present investigation:
• A good dispersibility of both sawdust and rice husk in the epoxy matrix had
enhanced the tensile strength, hardness and surface texture of the composites.
Preparation and Mechanical Property Analysis of Polymer Matrix … 259

• 40% of reinforcing fillers in the epoxy matrix gave the overall best results among
all the prepared and tested composites.
• A very high surface finish of the composite can be obtained by careful handling
of the specimen during the process.

References

1. Agarwal BD, Broutman LJ (1980) Analysis and performance of fiber composites. Wiley, New
York, pp 3–12
2. Balakrishnan P, John MJ, Pothen L, Sreekala MS, Thomas S (2016) Natural fiber and polymer
matrix composites and their applications in aerospace engineering. Adv Compos Mater Aerosp
Eng. https://doi.org/10.1016/B978-0-08-100037-3.00012-2
3. Yang HS, Kim HJ, Son J, Park HJ, Lee BJ, Hwang TS (2004) Rice-husk flour filled polypropylene
composites: mechanical and morphological study. Compos Struct 63:305–3012
4. Chen RS, Ahmad S, Gan S, Tarawneh MA (2019) High loading rice husk green composites:
dimensional stability, tensile behavior and prediction, and combustion properties. J Thermoplast
Compos Mater. https://doi.org/10.1177/0892705718815536
5. Monteiro AS, Barreira DAS, Bartolomei SS, Rene R. Oliveira and Esperidiana Augusta Barreto
de Moura, Comparative Study of the Use of Rice
6. Xu X, Wang H, Sun Y, Han J, Huang R, Zhan X (2018) Sound absorbing properties of perforated
composite panels of recycled rubber, fiberboard sawdust, and high density polyethylene. J Clean
Prod. https://doi.org/10.1016/j.jclepro.2018.03.174
7. Suarez JCM, Coutinho FMB, Sydenstricker TH (2003) SEM studies of tensile fracture surfaces
of polypropylene-sawdust composites. Polym Test 22:819–824
Investigation of the Mechanical
Properties of Polyamide 6 Hybrid
Nanocomposites with MWCNT
and Copper Nanoparticles

T. Anand and T. Senthilvelan

Abstract The mechanical behaviour of polyamide6 (PA6)-based hybrid nanocom-

posites has been investigated. The polyamide6 (PA6) is blended with multi-walled
carbon nanotube (MWCNT) and copper nanoparticles. The reinforcement mate-
rials in different proportions were blended with PA6 in a co-rotating twin-screw
extruder and the specimens were made in an injection moulding machine. Mechani-
cal tests were conducted on the prepared specimens to study the tensile, flexural and
impact properties. The results suggested that Polyamide6 (PA6) hybrid nanocom-
posites infused with MWCNT and copper nanoparticles have significant increase in
the tensile, flexural and impact properties. An atomic force microscope was used
to examine the microstructure of the composites. Fractured surfaces of Polyamide6
nanocomposites after tensile and impact test were analysed through SEM images.

Keywords Polymer nanocomposites · Polyamide6 (PA6) · MWCNT · Copper

nanoparticles · Atomic force microscope · SEM

1 Introduction

The developments in the field of engineering and technology majorly depend on the
advancement of materials. Increasing demand for lightweight, strong and stiff struc-
ture has paved the way for the development of new materials. Composite materials
have played a vital role in the process of optimization in materials from the beginning
of the 1960s. Composite materials provide flexible designs and tailor-made materials
for an optimum design.

T. Anand (B) · T. Senthilvelan

Pondicherry Engineering College, Pondicherry University, Puducherry, India
e-mail: anand.tvel@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 261

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_24
262 T. Anand and T. Senthilvelan

Nanomaterials as fillers or reinforcements [1, 2] in the composite can produce new

physical properties and novel behaviours. Gheorghe Rusu et al. (2012) have studied
the effect of TiO2 reinforcement in nylon6 and found that the tensile and flexural
characteristics were improved. Vijaya K. Rangari et al. (2012) have addressed the
reinforcement of ZnO/CNTs with nylon6. This study revealed that the tensile stress
and modulus were improved by 73% and 377.4%.
Polyamides (PA6) are semi-crystalline polymers synthesized by the ring opening
polymerization of caprolactam, one of its monomers [3, 4]. The melting point of
polyamide6 (C6 H11 NO)n is about 220 °C, crystalline density at 25 °C is 1.23 g/cm3
and the molecular weight is 113.16 g/mol. Polyamides possess excellent mechanical
properties, particularly toughness, and have good sliding and wear characteristics
[5–8].
Carbon nanotubes are allotropes of carbon with a cylindrical nanostructure.
These cylindrical molecules have extraordinary thermal conductivity, electrical
and mechanical properties. Multi-walled carbon nanotubes (MWCNT) consist of
multiple rolled layers of grapheme
Copper (Cu) is a ductile material with good thermal and mechanical properties [9].
Pure copper is soft and malleable. The atomic weight of copper is 63.546; melting
point is about 1084.620 and it comes under the element category of a transition metal.
The density of copper is 8.96 g/cm3 . Copper nanoparticles are used as a lubricant
additive to reduce the friction and wear and to enhance the mechanical properties.
In this work, twin-screw extrusion and injection moulding processes were used
to prepare polyamide6 nanocomposites with MWCNT and copper nanoparticles as
reinforcements. The weight % of MWCNT was kept as constant and weight % of
copper nanoparticles was varied. The mechanical behaviour of the composites was
studied.

2 Experimental

2.1 Materials

The PA6 (commercial grade) material has been used in the present work. The multi-
walled carbon nanotube is purchased from Applied Sciences Innovations private
limited, Pune. The particle size of MWCNT is 25 nm and it is surface modified with
COOH. Carbon coated copper nanoparticles of size 20 nm were purchased from
Nanoshell Corporation, USA. A silane coupling agent γ -aminopropyltriethoxysilane
has been used to create bonding between the matrix material and reinforcements.
Investigation of the Mechanical Properties of Polyamide … 263

Fig. 1 Specimens for tensile, flexural and impact tests as per ASTM standards

2.2 Preparation of Polyamide6/MWCNT/Nano-Copper

Composites

The materials are kept in an oven at a temperature of 80 °C for four hours to remove
the moisture content present in them. The reinforcement materials are added to PA6
in weight percentage. The surface modified multi-walled carbon nanotube taken is
0.25 wt% and the carbon coated copper nanoparticles taken are 0.2, 0.4 and 0.6 wt%
with PA6. The materials are blended in a co-rotating twin-screw extruder with a
screw diameter of 25 mm and the rotating speed of the screw is kept at 180 rpm.
The operating temperature is 250 °C. The blended material for each composition is
pelletized in a pelletizer.
The blended composite mixture is then injection moulded using the MACI-
FOLD injection moulding machine. The barrel temperature of the injection moulding
machine is 290 °C and 10 MPa pressure. The specimens for the tensile, flexural and
impact tests were prepared as per ASTM standards. Figure 1 shows the specimens
prepared for the testing.
The operating temperature has been selected based on the melting point of the
matrix material, whereas the screw diameter is kept constant. The rotating speed of
the screw and operating pressure has been selected with reference to the machine
catalogue depends on the matrix material. These parameters ensure proper blending
of reinforcements with the matrix material [10].

2.3 X-Ray Diffraction

An XRD analysis [11] was employed to identify the dispersion of MWCNT and
copper nanoparticles present in the matrix material PA6. X-ray diffraction analyses
were performed on the prepared specimens using a Philips X-ray diffraction instru-
264 T. Anand and T. Senthilvelan

ment with a radiation source of Cu-k-β using 30 kV and 100 mA. A step size of 0.02
degree, scan rate of 2 degree per minute and continuous scan mode are the parameters
used for the analysis.

2.4 Atomic Force Microscopy

An atomic force microscope has been widely used for viewing the surface topology
of the polymer composites. It is also applied to study the dispersion of the nano-
reinforcements in the matrix material. AFM images were taken using the NTMDT
atomic force microscope, Ireland. The frequency range is 5 MHz and the resolution
of the atomic force microscope is 0.01 nm.

2.5 Mechanical Property Testing

The tensile test has been conducted in the specimen made as per ASTM D638(Type 1)
with size 165 × 13 × 3 using the INSTRAN 3382, 100 KN universal testing machine
with available load cells 100 N, 1 KN and 10 KN. The flexural test is conducted in
the specimens made as per ASTM D 790 with size 127 × 12.7 × 6.4 mm using the
AUTOGRAPH AG-IS 50 KN universal testing machine with the available range of
loads 0.5,1,2.5,5,10,25 and 50 KN. The impact test is conducted in the specimen
made as per ASTM D256 with size 63.5 × 12.7 × 6.4 mm using the izod impact
testing machine [12]. A notch was provided on the impact test specimen with a depth
of 2, 450 and 0.25 mm end radius. Five similar samples were used for each test
and the average values are considered for the analysis. Figure 1 shows the prepared
specimens for the tensile, flexural and impact tests as per ASTM standards.

3 Results and Discussion

The mechanical properties of PA6, multi-walled carbon nanotube and nano-copper

hybrid composites are given in the following Table 1. The tensile strength, flexural
strength and impact strengths of PA6 nanocomposites are illustrated and compared
with that of pure PA6. The tensile strength, flexural strength and impact strength
of PA6/0.25 wt%/0.6 wt% nano-copper was enhanced by 26.6, 39.47 and 139.46%,
when compared to pure PA6.
In composite materials, the hardness increase with the percentage of the reinforce-
ment materials. Figure 2 shows the shore–D hardness of PA6 and PA6 with various
proportions of reinforcements. PA6 offered a shore–D hardness of 67 and PA6 with
0.25 wt% MWCNT and 0.6% nano-Cu offered a maximum hardness of 72.4 with
8.06% increase in hardness.
Investigation of the Mechanical Properties of Polyamide … 265

Table 1 Mechanical properties of PA6 and its composites

Sl.No Description Tensile % of Flexural % of Impact % of
strength increase strength (MPa) increase strength increase
(MPa) in in (J/m) in
tensile flexural impact
strength strength strength
1 PA6 39.164 – 29.986 – 4837.2 –
2 PA6 + 43.029 9.869 35.609 18.75 5772.8 19.34
0.25%
MWCNT
+ 0.2%
nano-Cu
3 PA6 + 45.182 15.36 38.795 29.38 8589.5 77.69
0.25%
MWCNT
+ 0.4%
nano-Cu
4 PA6 + 49.595 26.6 41.823 39.47 11583.4 139.46
0.25%
MWCNT
+ 0.6%
nano-Cu

74
Hardness (Shore-D)

PA6 + 0.25
72 %MWCNT + 0.2% PA6 +0.25 %
nano Cu PA6 +0.25 %
70 MWCNT + 0.6%
MWCNT + 0.4%
nano Cu
nano Cu
68
PA6
66

64
0 0.2 0.4 0.6
% of Nano Cu Reinforcement

Fig. 2 Hardness (shore–D) of PA6 and its composites

The stress-strain behaviour [13] of PA6 and PA6 composites with 0.25 wt%
MWCNT and 0.2, 0.4 and 0.6 wt% of copper nanoparticles are shown in Fig. 3.
The stress value increases linearly within the elastic limit, beyond which the
strain increases rapidly for a small stress value. The stress-strain behaviour of PA6
nanocomposites was more or less similar to that of virgin PA6.
Figure 4 shows the tensile modulus of PA6 and its composites. The tensile modulus
of PA6 was increased for various proportions of reinforcements. The nanofillers
dispersed evenly in the matrix material and created a strong interaction between the
matrix material and reinforcements [14].
266 T. Anand and T. Senthilvelan

PA 6
60
PA6,0.25% MWCNT,0.2
50 %nano Cu

Stress (Mpa)
PA6,0.25% MWCNT,
40 0.4% nano Cu
PA6,0.25%
30 MWCNT,0.6% nano Cu

20

10

0
30 60 90 120 150 180 210 240 270 300 330 360
Strain (%)

Fig. 3 Stress strain curves of PA6 and its composites

600
Tensile Modulus, MPa

500
400
300
200
100
0
PA 6 PA6 + 0.25 PA6 + 0.25 PA6 + 0.25
%MWCNT +0.2 %MWCNT +0.4 %MWCNT +0.6
% nano Cu % nano Cu % nano Cu

Fig. 4 Tensile modulus of PA6 and its composites

The tensile, flexural and impact strength and percentage of elongation at break
for PA6, PA6 with MWCNT and PA6 with MWCNT and various proportions of
copper nanoparticles (0.2, 0.4 and 0.6%) are depicted in Fig. 5a–d. The addition
of MWCNT and copper nanoparticles with PA6 increased the tensile, flexural and
impact strengths. PA6 with 0.25 wt% MWCNT and 0.6 wt% copper nanoparticles
showed better mechanical properties than the other proportions. The deformability
of PA6 nanocomposites decreases due to the reinforcement materials thus reducing
the elongation at break [15].
Figure 6 shows the X-ray diffraction pattern of PA6 nanocomposites with various
proportions of nano-reinforcements (PA6, 0.25 wt% MWCNT and 0.2 wt% Nano-
Cu, PA6, 0.25 wt% MWCNT and 0.4 wt% Nano-Cu, PA6, 0.25 wt% MWCNT and
0.6 wt% Nano-Cu). The d-spacing was calculated by analysing the peaks of the XRD
patterns using Bragg’s law nλ = 2dsinθ , where θ is the scattering angle, n is a positive
integer and λ is the wavelength of the incident wave and d is the interplanar distance
[16]. An X-ray with a wave length of 1.54 A is used. The XRD pattern shows that
PA6, MWCNT and nano-copper composites have diffraction peaks at 2θ = 21.680
and 43.40 corresponding to a d-spacing of 4.09 and 2.08 A.
The dispersion of nano-copper and multi-walled carbon nanotubes reinforced in
different proportions with PA6 is studied by viewing the structure of PA6 nanohybrid
Investigation of the Mechanical Properties of Polyamide … 267

Tensile Strength, MPa

60

40

20

0
PA 6 PA6 + 0.25 PA6 + 0.25 PA6 + 0.25
%MWCNT +0.2 %MWCNT +0.4 %MWCNT +0.6
% nano Cu % nano Cu % nano Cu
(a)
Flexural Strength, MPa

50
40
30
20
10
0
PA 6 PA6 + 0.25 PA6 + 0.25 PA6 + 0.25
%MWCNT +0.2 % %MWCNT +0.4 % %MWCNT +0.6 %
nano Cu nano Cu nano Cu
(b)
Impact Strength, J/m

15000
10000
5000
0
PA 6 PA6 + 0.25 PA6 + 0.25 PA6 + 0.25
%MWCNT +0.2 %MWCNT +0.4 %MWCNT +0.6
% nano Cu % nano Cu % nano Cu
(c)
Elongation at break, %

400
300
200
100
0
PA6 PA6 + PA6 + PA6 +
0.25 % MWCNT + 0.25 % MWCNT +0.25 % MWCNT +
0.2 % nano Cu 0.4 % nano Cu 0.6 % nano Cu
(d)

Fig. 5 a Tensile, b flexural, c impact strengths, d elongation at break of PA6 and its composites
containing MWCNT and nano-copper

composites using AFM [17]. The height and phase images of the PA6 nanocomposites
with different proportions of reinforcement are shown in Fig. 7.
The copper nanoparticles and multi-walled carbon nanotubes are evenly dispersed
in the PA6 matrix with an average size of 61.49 nm. The microscopy images show,
Fig. 7a PA6 with 0.25% MWCNT and 0.2% nano-Cu consists of an average particle
size of 59.66 nm, Fig. 7b PA6 with 0.25% MWCNT and 0.4% nano-Cu consists of
an average particle size of 61.293 nm, Fig. 7c PA6 with 0.25% MWCNT and 0.6%
nano-Cu consists of an average particle size of 61.48 nm.
268 T. Anand and T. Senthilvelan

PA6, 0.25% MWCNT, 0.2 % Nano Cu

12000 PA6, 0.25% MWCNT, 0.4 % Nano Cu
PA6, 0.25% MWCNT, 0.6 % Nano Cu
11000
10000
9000
8000
7000
Intensity (a.u)

6000
5000
4000
3000
2000
1000
0
0 10 20 30 40 50
2θ ( Degrees)

Fig. 6 X-Ray diffraction pattern of PA6 nano-hybrid composites with various proportions of nano-
Cu and MWCNT

To study the fracture behaviour, fractographs of PA6 and PA6 with MWCNT and
copper nanoparticles were taken using the scanning electron microscope. Figures 8
and 9 show the fracture surface of PA6 and PA6 composites after the tensile and
impact tests. The microstructure of the fractured surface of PA6 appeared rough in
Fig. 8a and PA6 with 0.25 wt% of MWCNT and 0.2 wt% of copper nanoparticles
show small spherical dimples in Fig. 8b indicates a ductile fracture. Figure 8c, d
shows smooth surface and some crazes were seen, indicating that the fracture was
brittle in nature [18].
Figure 9a shows the fractured surface of PA6 after the impact test. The rough
surface observed absorbs energy and that causes ductile fracture [19]. The nanopar-
ticles, MWCNT and copper nanoparticles are found uniformly dispersed in the PA6
matrix material as shown in Fig. 9b, c. The addition of MWCNT and copper nanopar-
ticles made the fractured surface of the composites smooth and brittle fracture was
observed shown in Fig. 9d.
The incorporation of surface modified MWCNTs and carbon coated nano-
copper particles into the PA6 matrix creates good interaction between the nano-
reinforcements and polymeric chains [20–22]. The strong interaction between the
nano reinforcements and PA6 matrix greatly enhanced the dispersion as well as
interfacial adhesion. Hence, the mechanical properties of the composites could be
improved.
Investigation of the Mechanical Properties of Polyamide … 269

(a). Height images of PA6 with 0.25% MWCNT and 0.2% nano Cu

(b) . Height images of PA6 with 0.25% MWCNT and 0.4 % nano Cu

(c) . Height images of PA6 with 0.25% MWCNT and 0.6 % nano Cu

Fig. 7 a. Height images of PA6 with 0.25% MWCNT and 0.2% nano-Cu. b. Height images of PA6
with 0.25% MWCNT and 0.4% nano-Cu. c. Height images of PA6 with 0.25% MWCNT and 0.6%
nano-Cu
270 T. Anand and T. Senthilvelan

Fig. 8 Scanning electron microscopic images of the fractured surface after tensile test a PA6 b PA6,
0.25% MWCNT, 0.2% nano-Cu c PA6, 0.25% MWCNT, 0.4% nano-Cu d PA6, 0.25% MWCNT,
0.6% nano-Cu

4 Conclusion

PA6 nanocomposites reinforced with MWCNT and copper nanoparticles were suc-
cessfully manufactured using twin-screw extrusion and injection moulding pro-
cesses. Tensile, flexural and impact tests were conducted successfully on different
compositions of PA6, PA6 + 0.25% MWCNT + 0.2% nano-Cu, PA6 + 0.25%
MWCNT + 0.4% nano-Cu, PA6 + 0.25% MWCNT + 0.6% nano-Cu. The experi-
ments showed that the tensile, flexural and impact strength values were improved by
26.6%, 39.47% and 139.46% in the composition of PA6 + 0.25% MWCNT + 0.6%
nano-Cu compared to that of virgin PA6.
Investigation of the Mechanical Properties of Polyamide … 271

Fig. 9 Scanning electron microscopic images of the fractured surface after impact test a PA6 b PA6,
0.25% MWCNT, 0.2% nano-Cu c PA6, 0.25% MWCNT, 0.4% nano-Cu d PA6, 0.25% MWCNT,
0.6% nano-Cu

References

1. Rusu G, Rusu E (2012) Anionic Nylon 6/TiO2 composite materials: effects of filler on the
thermal and mechanical behaviour of the composites. Polym Compos 13:1557–1569
2. Rangari V, Mohammad GM (2012) Synthesis of hybrid ZnO/CNTs nano particles and their
reinforcement in Nylon 6 polymer fibers. J Appl Polym Sci 129(1)
3. Liu Q, de. Wijjin JR, Van Blitterswijk CA (1977) Nano—Apatite/Polymer composites:
mechanical and physiochemical characteristics. Biomaterials 18:1263–1270
4. Kransov AP, Aderikha VN, Afonicheva OV, Mit VA, Tikhonov NN, Yu A (2010) Vasil’kov:
categorization system of nanofillers to polymer composites. J Friction Wear 31:68–80
5. Chiu FC, Kao GF (2012) Polyamide 46/multi walled carbon nanotube nano composites with
enhanced thermal, electrical and mechanical properties. Compos Part A 43:208–218
6. Shen Z, Bateman S, Wu DY, McMohan P (2009) The effects of carbon nanotube on mechanical
and thermal properties of woven glass fiber reinforced Polyamide 6 nano composites. Compos
Sci Technol 69:239–244
7. Mohanty S, Nayak SK (2007) Mechanical, thermal and viscoelastic behaviour of Nylon 6/Clay
nano composites with Cotreated Montmorillonites. Poly Plast Technol Eng 46:367–376
8. Liu TX, Liu ZH, Ma KX, Shen L, Zeng KY, He CB (2003) Morphology, thermal and mechanical
behavior of Polyamide 6/layered—silicate nano composites. Compos Sci Technol 63:331–337
272 T. Anand and T. Senthilvelan

9. Luyt AS, Molefi JA, Krump H (2006) Thermal, mechanical and electrical properties of copper
powder filled low-density and linear low-density polyethylene composites. Polym Degrad Stab
91(7):1629–1636
10. So HH, Cho JW, Sahoo NG (2007) Effect of carbon nanotubes on mechanical and electrical
properties of polyimide/carbon nanotubes nano composites. Eur Polym J 43:3750–3756
11. Lee MW, Hu X (2003) Effect of fillers on the structure and mechanical properties of
LCP/PP/SiO2 in-situ hybrid nano composites. Compos Sci Technol 63(3):333–346
12. Chen GX, Kim HS, Park BH, Yoon JS (2006) Multi walled carbon nanotubes reinforced nylon
6 composites. Polymer 27:4760–4767
13. Amrishraj D, Senthilvelan T (2017) Acrylonitrile butadiene styrene composites reinforced with
nano zirconia and PTFE. Polym Compos
14. Sahoo NG, Cheng HKF, Cai J, Li L, Chan SH, Zhao J, Yu S (2009) Improvement of mechanical
and thermal properties of carbon nanotube composites through nanotube functionalization and
processing methods. Mater Chem Phys 117 (2009) 313 – 320
15. Coleman JN, Khan U, Werner J, Blau Yurii K. Gun’ko (2006) Small but strong: a review of
the mechanical properties of carbon nanotube—polymer composites. Carbon 44:1624–1652
16. Berti G, Angelo LD’, Gatto A, Iuliano L. Mechanical characterization of PA—Al2 O3
composites obtained by selective laser sintering. Rapid Prototyping. 16:124–129
17. Pollanen M, Pirinen S, Suvanto M, Pakkanen TT (2011) Influence of carbon nanotube—poly-
meric compatiblizer master batches on morphological, thermal, mechanical and tribologigal
properties of polyethylene. Compos Sci Technol 71:1353–1360
18. Fu SY, Qiao Feng XI , Lauke B, Wing Mai Y (2008) Effects of particle size, particle/matrix
interface adhesion and particle loading on mechanical properties of particulate—polymer
composites. Composites: Part B 39:933–961
19. Saeed MB, Zhan MS. Adhesive strength of nano size particles filled thermoplastic polyimides.
Int J Adhe Adhes 27(4):319–329
20. Kulkarni MV, Elangovan K, Reddy H, Prakash JN (2011) Investigations on the influence of
filler on the impact properties of Polyamides and its composites. I-Managers J Mech Eng 1(4)
21. Yun Zheng L, Tak Lau K, Xin Zhao L (2009) mechanical and thermal properties of nano
Al2O3/Nylon 6 composites. Chem Eng Commun 197:343–351
22. Chitte HK, Bhat NV, Karmakar NS, Kothari DC, Shinde GN (2012) Synthesis and character-
ization of polymeric composites embedded with silver nanoparticles. World J Nano Sci Eng
2:19–24
Preparation, Characterization, Image
Segmentation and Particle Size Analysis
of Cow Bone Powder for Composite
Applications

O. M. Ikumapayi, Esther Titilayo Akinlabi, Paul A. Adedeji,

and S. A. Akinlabi

Abstract Cow bone is a bio-waste material that has caused environmental unfriend-
liness to the people living around the place of production and deposition. In recent
time, researchers have exploited various means of utilizing this bio-waste material
and make it environmentally friendly and economically utilizable, especially in the
area of applications, such as activated carbon, water purification, reinforcement in
composites, filler and additives, and its efficacy has been traced to whether is in
the form of macro-, micro- and nanoparticles. In this study, cow bone head (skull)
was collected, washed and cleaned from meats, processed, sun-dried for six weeks,
and it was then washed again with distilled water to remove impurities and con-
taminants, then placed inside the oven set at 50 °C to dry for 5 h to remove any
trace of moisture content and to ensure absolute dryness. It was then milled into
nanoparticle powder using vibratory disk milling machine and the milling times
were 0 min which was taken to be 150 µm size, 20, 40 and 60 min. The morpho-
logical and physiological characterizations were carried out using scanning electron
microscopy (SEM), energy dispersive X-ray (EDX) as well as X-ray fluorescence
(XRF). MATLAB (R2015a) was employed for image processing into bimodal with
foreground and background pixels using thresholding segmentation method. SEM
images were taken at 1.00 ×, 2.00 × and 5.00 × which resulted to 50 µm, 20 µm
and 10 µm, respectively. The image segmentation was employed to determine the
foreground from a background of cow bone powder (CBP) and to enhance high
resolutions, visual perception as well as to achieve better quality of the final output.

Keywords Cow bone powder · Image segmentation · Milling time

O. M. Ikumapayi (B) · E. T. Akinlabi · P. A. Adedeji

Department of Mechanical Engineering Science, University of Johannesburg, Auckland Park
Kingsway Campus, Johannesburg 2006, South Africa
e-mail: oikumapayi@uj.ac.za
S. A. Akinlabi
Department of Mechanical Engineering, Faculty of Engineering and Technology, Walter Sisulu
University, Butterworth Campus, Butterworth 4960, South Africa
E. T. Akinlabi · S. A. Akinlabi
Department of Mechanical Engineering, Covenant University, Ota, Nigeria

© Springer Nature Singapore Pte Ltd. 2021 273

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_25
274 O. M. Ikumapayi et al.

1 Introduction

Cow bone as organic bio-waste materials have been developed into different particles
sizes from macro, micro as well as to nanoparticles. The development of micropar-
ticles from cow bones was carried out by first carbonized the cow bones at 750 °C
using heat treatable furnace in the absence of oxygen and sieved using different mesh
sizes of 150, 300 and 600 µm and these were then applied to polymer as reinforce-
ment to form polymatrix composites [1] and it has also been used as reinforcement
nanoparticle during friction stir welding of metal composites [2], in this study, cow
bone was milling for 1 h and then carbonized at a temperature of 500 °C inside
muffle furnace for 2 h and allowed for thermal stability inside the furnace to room
temperature. It is a known fact that cow bones are waste products and making the
environment unconducive to live in with the unbearable pollutants. Notwithstand-
ing, researchers are making every effort to convert this agro-wastes to more useful
engineering applications, especially for reinforcement purposes. Cow bone being
natural animal fiber is expected to have good surface compatibility in addition to the
structural compatibility requirements as biomaterials [3, 4].
It is essential to determine the particle size distribution as they affect the pow-
ders dissolution, dispersion and the flow properties. In the recent times, there have
been tremendous works on the techniques to improve and address the dichotomy
between the image backgrounds and foreground, especially when the object in a case
is brighter with much variance [5, 6]. The image segmentation technique requires
knowledge about difference intensity and object size for successful thresholding. To
complement other studies, this study uses a sequence of an algorithm written in MAT-
LAB to estimate the particle size at varying milling time. This is seen as a means of
automating the process, such that particle size can be determined in real time during
milling by using optimal imagery of samples. Thresholding becomes a potent tool
for image segmentation when the variance of the object differs significantly from
that of the background. This is often the case in SEM images and change monitoring
applications [7]. The ideal threshold for segmentation is the intersection point of the
image and its background. This is such that the intersection is far from the class with
large variance [8].

2 Materials and Methods

2.1 Material Collection and Preparation of Cow Bone

Nanoparticle

The skulls of the cow (cow head bone) were collected from the abattoir in the city of
Johannesburg, South Africa as shown in Fig. 1a. It was then washed with de-ionized
water, cleaned from meat and then sun-dried for about six weeks to allow for absolute
dryness as shown in Fig. 1b. After 6 weeks in the sun, it was then washed again with
Preparation, Characterization, Image Segmentation and Particle … 275

Fig. 1 a Fresh cow skull bone obtained, b dried cow bone and c cow bone powder (CBP) in a
150 µm ASTM standard sieve

de-ionized water to remove any impurities and contaminations that may be present.
In order to ensure total dryness before crushing, it was dried again inside the electric
oven set 50 °C to dry for 5 h for 7 days to ensure absolute dryness of moisture
contents, this was further cleaned with acetone before crushing and pulverizing and
sieved into 150 µm size using ASTM meshes standard range by employed KingTest
Sieve, after which it was further milled with digital vibratory disk milling machine
(VDMM) for 60 min before utilization as shown in Fig. 1c. The flowchart of the
study is presented in Fig. 1.

2.2 Characterizations of Cow Bone Powder

The following characterizations were carried out on the cow bone powder scanning
electron micoscopic (SEM), eergy dispersive X-ray (EDX) and X-ray fluorescence
(XRF) to determine the elemental and chemical compositions as well as morphology
of the powder (Fig. 2).

2.2.1 X-Ray Fluorescence (XRF) Spectroscopic Analysis of Cow Bone

Powder

X-Ray fluorescence (XRF) spectroscopic analysis of cow bone powder (CBP) was
carried out using PHILIP PW1404 XRF wavelength disperse spectrometer and it
was revealed that CBP contains the following chemical composition 45.06% (CaO),
34.60% (P2 O5 ), 0.065% (MgO), 0.25% (Fe2 O3 ), 0.17% (Na2 O), 0.16% (Al2 O3 ),
0.08% (Al2 O3 ), 0.12% (K2 O), 0.06% (TiO2 ), 0.02 (ZnO), 0.01% (MnO) and 16.28%
(LOI).
276 O. M. Ikumapayi et al.

Fig. 2 Process flowchart of the study

2.2.2 Scanning Electron Microscopic (SEM) Analysis of Cow Bone

Powder

Morphological and elemental characteristics of developed cow bone powder (CBP)

were carried out using TESCAN model, type VEGA 3 LMH, scanning electron
microscope (SEM) machine which was integrated with energy dispersive X-ray
spectrometer (EDXS), this responsible for elemental composition analysis present in
CBP. To ensure high conductivity and better resolution of the test samples (CBP), the
sample was first sputter coated with a thin layer of carbon before the test and the test
was conducted [9, 10] with the following parametric machine values, beam intensity
of 12, accelerating voltage of 20 kV and a working distance of 15. SEM images
were taken at 1.00 kx, 2.00 kx and 5.00 kx leading to 50 µm, 20 µm and 10 µm,
respectively, for different milling times of 0 min, 20 min, 40 min and 60 min. The
Preparation, Characterization, Image Segmentation and Particle … 277

SEM images at varying magnifications and milling time with respect to SEM image
segmentation are shown in Fig. 3 while the elemental compositions are showed in
Table 1.

Original and Segmented SEM Images at 1000 MAG

a : Micrographs at 50 b : Micrographs at c: Micrographs at 50 d : Micrographs at 50

μm of 0 min 50 μm of 20 μm of 40 mins μm of 60 mins
milled mins milled milled milled

Original and Segmented SEM Images at 2000 MAG

e: Micrographs at 20 f : Micrographs at g : Micrographs at 20 h : Micrographs at 20

μm of 0 min 20 μm of 20 μm of 40 mins μm of 60 mins
milled mins milled milled milled

Fig. 3 SEM images and their segmentations for cow bone powder at varying magnification of 1.00,
2.00 and 5.00 kx at different milling time of 0, 20, 40 and 60 min
278 O. M. Ikumapayi et al.

Original and Segmented SEM Images at 5000 MAG

i: Micrographs at 10 j: Micrographs at k: Micrographs at 10 l: Micrographs at 10

µm of 0 min 10 µm of 20 µm of 40 mins µm of 60 mins
milled mins milled milled milled

Fig. 3 (continued)

Table 1 Elemental compositions of CBP at varying milling time as analyzed by EDXS

Powder Element Elemental composition at different milling times
0 min 20 min 40 min 60 min
CBP C 46.3 45.6 53.9 40.8
O 27.2 30.7 30.0 32.1
Ca 18.3 16.2 10.7 18.3
P 7.5 6.8 4.8 8.0
Mg 0.3 0.5 0.2 0.4
Note 0 min was taken at mesh size of 150 µm

3 Methodology

3.1 Method of Calcination of Cow Bone Powder

for Composites Application

Calcination process is simply mean the method of heating a substance with a con-
trolled temperature and controlled environment [11], i.e., thermal treatment process
of material in the absence or limited supply of air in order to induce thermal stability,
phase transition, thermal decomposition as well as removal of a volatile fraction from
the material before using [12]. This is done to improve the strength, texture, thermal
stability, etc. The milled cow powder cannot be applied directly as reinforcement to
the workpiece that will be processed. If applied without any carbonization process,
it will react with the metal and form slag within the processed metal as a result of a
Preparation, Characterization, Image Segmentation and Particle … 279

thermal generation that could lead to combustion of the particles, thereby leading to
internal and surface defects. On this note, the processed powder was first undergone
calcination process to ensure thermal stability. The calcination of the powder was car-
ried out by putting each of CBP in pure graphite crucible and then put them carefully
inside the muffle furnace set at 500 °C for two (2) hours, after the 2 h, the furnace
was put off and the carbonaceous powder was allowed to cool to room temperature
inside the furnace before it was then brought out for use, and then further grinded was
performed on cow bone powder to ensure smoothness of the carbonaceous powder
before use.

3.2 Image Segmentation and Thresholding

Distinguishing foreground pixels from the background pixels has been used to
enhance visual perception of images [13–15]. One of the common techniques used
to identify morphological maps [16] and image color intensity and is the threshold
segmentation. The technique offers a simplicity in calculation and speed in operation
[17]. Threshold segmentation recognizes image patterns aiming toward enhancing
final image quality and resolutions [18]. This method does not perform well in cases
where there is overlap in the grayscale value of the image [19, 20]. In this study, the
SEM images, which are grayscale, were converted into binary images thus reducing
data complexity.
Otsu’s method [21], a nonparametric and an unsupervised thresholding technique,
was used for image thresholding. The technique selects optimum thresholds by max-
imizing the intraclass variance of binary images. In this study, we assume that the
SEM images are bimodal with foreground and background pixels.
Selecting from a set of local thresholds t1 , t2 , t3 , . . . tn , the thresholds subdivide
the images into n + 1 classes of C1 , C2 , C3 , . . . Cn . This is achieved by maximizing
the objective function of a sum of all threshold variances [6];

J1 (t1 , t2 , t3 , . . . tn ) = σ02 + σ12 + σ22 + σ32 + · · · σn2 (1)

where
t1 −1
 t1 −1
 i pi
σ02 = ω0 (μ0 − μT ) , ω0 =
2
pi , μ0 = . (2)
i=0 i=0
ω0
t2 −1
 t2 −1
 i pi
σ12 = ω1 (μ1 − μT )2 , ω1 = pi , μ1 = . (3)
i=t1 i=t1
ω1
t3 −1
 t3 −1
 i pi
σ22 = ω2 (μ2 − μT )2 , ω2 = pi , μ2 = (4)
i=t2 i=t2
ω2
280 O. M. Ikumapayi et al.

t4 −1
 t4 −1
 i pi
σ32 = ω3 (μ3 − μT )2 , ω3 = pi , μ3 = (5)
i=t3 i=t3
ω3
tn+1 −1 tn+1 −1
  i pi
σn2 = ωn (μn − μT )2 , ωn = pi , μn = (6)
i=tn i=t
ωn
n

and σ02 , σ12 , σ22 , σ32 , . . . σn2 . are the class variances, ω0 , ω1 , ω2 , ω3 , . . . ωn represents
the class probabilities, μ0 , μ1 , μ2 , μ3 , . . . μn represents the mean values of the
segmented classes. The intensity of the whole image is

μT ⊆ μT = ω0 μ0 + ω1 μ1 + ω2 μ2 + ω3 μ3 + . . . + ωm μm ω0
+ ω1 + ω2 + ω3 + ωn = 1.

The segmentation process was performed at different milling times and at different
magnifications. A script for this was computed using MATLAB (R2015a) installed
on a desktop computer workstation with configuration 64 bits, 32 GB RAM Intel (R)
Core (TM) i7 5960X. The mean particle sizes were obtained at different milling time
and different magnifications. A two-way analysis of variance was performed on the
result to determine whether there is a statistically significant difference between the
mean particle sizes across varying milling time and the across the magnifications.

3.3 Results and Discussion

Table 1 and Fig. 4 showed the variability in elemental compositions of cow bone
powder as analyzed by EDXS. It was revealed that the following elements are present,
calcium (Ca), carbon (c), phosphorous (P), oxygen (O) and little traces of magnesium
(Mg). It can be observed that there is variation in composition of the elements as the
milling progresses. The elemental composition varied along with the milling time
from 0–60 min and this was also in agreement with the work carried out Omah et al.
[1].
The mean particle sizes at each milling time and each magnification are as pre-
sented in Table 2. It is expedient to statistically establish whether the difference
between the mean particle sizes at each milling time and magnification are sta-
tistically significant. To establish this, an analysis of variance was performed and
presented in Table 3.
From Table 3, at a confidence interval of 0.05, the difference between the mean
particle sizes obtained at different milling times is statistically significant. Similarly,
the mean particle sizes obtained at different magnifications are statistically signif-
icant. However, the degree of significance in milling time is not as obvious as that
from varying magnifications.
Preparation, Characterization, Image Segmentation and Particle … 281

Fig. 4 EDS of the cow bone powder at varying milling time

Table 2 Mean particle size of segmented SEM images at different magnifications and milling times
Magnification
Milling time ×1000 ×2000 ×5000
0 2.70E+03 1.06E+03 507.996
20 2.98E+03 1.87E+03 526.284
40 3.54E+03 1.89E+03 727.674
60 4.03E+03 2.13E+03 781.975

Table 3 Statistical analysis of variance between mean particle sizes at varying milling time and
different magnifications
Source of variation SS df MS F P-value F crit
Milling time 1,215,747 3 405,248.9 4.464 0.057 4.757
Magnification 14,487,754 2 7,243,877 79.799 4.76E−05 5.143
Error 544,659.7 6 90,776.62
Total 16,248,160 11
282 O. M. Ikumapayi et al.

It can be inferred that the increasing the milling time of cow bone depending on
the use, significantly affect its particle size. Particle sizes at different magnification
are bound to vary due to change in image resolution.

4 Conclusion

The method used in this work permitted the manipulations of micrographs in order
to have a true representation of the powder distribution and at the same time dis-
tinguishing foreground from background. Image processing analysis was investi-
gated in this research through thresholding segmentation at varying milling times
(t = 0, 20, 40, 60 min). It was revealed in the study that progressive milling leads
to increase in particle size distributions and decrease in the area of particle size as
noted in Table 2. It was revealed by EDX analysis that the main constituents of cow
bone powder are calcium, oxygen, magnesium, carbon and phosphorous.

References

1. Omah AD, Okorie BA, Omah EC, Ezema IC, Aigbodion VS (2017) Measurement of dielectric
properties of polymer matrix composites developed from cow bone powder. Int J Adv Manuf
Technol 88:325–335. https://doi.org/10.1007/s00170-016-8759-1
2. Ikumapayi OM, Akinlabi ET (2019) Experimental data on surface roughness and force feed-
back analysis in friction stir processed AA7075-T651 aluminium metal composites. Data Br
23:103710. https://doi.org/10.1016/j.dib.2019.103710
3. Ahmad M, Lee SS, Lim JE, Lee SE, Cho JS, Moon DH et al (2014) Speciation and phy-
toavailability of lead and antimony in a small arms range soil amended with mussel shell, cow
bone and biochar: EXAFS spectroscopy and chemical extractions. Chemosphere 95:433–441.
https://doi.org/10.1016/j.chemosphere.2013.09.077
4. Aramide FO, Ibitoye SA, Oladele IO, Borode JO (2010) Pack carburization of mild steel,
using pulverized bone as carburizer: Optimizing process parameters. Leonardo Electron J
Pract Technol 9:1–12
5. Elbayoumi Harb SM, Isa NAM, Salamah SA (2015) Improved image magnification algo-
rithm based on Otsu thresholding. Comput Electr Eng 46:338–355. https://doi.org/10.1016/j.
compeleceng.2015.03.025
6. Khairuzzaman AKM, Chaudhury S (2017) Multilevel thresholding using grey wolf opti-
mizer for image segmentation. Expert Syst Appl 86:64–76. https://doi.org/10.1016/j.eswa.
2017.04.029
7. Hu QM, Hou ZJ, Wieslaw LN (2006) Supervised range-constrained thresholding. IEEE Trans
Image Process 15:228–240
8. Xu X, Xu S, Jin L, Song E (2011) Characteristic analysis of Otsu threshold and its applications.
Pattern Recognit Lett 32:956–961. https://doi.org/10.1016/j.patrec.2011.01.021
9. Abdulrahman A, Latiff AAA, Daud Z, Ridzuan MB, Jagaba AH (2016) Preparation and char-
acterization of activated cow bone powder for the adsorption of cadmium from palm oil mill
effluent. IOP Conf Ser Mater Sci Eng 136:1–11. https://doi.org/10.1088/1757-899X/136/1/
012045
10. Adewole T (2015) Effect of cow bone ash particle size distribution on the mechanical properties
of cow bone ash-reinforced polyester composites. Chem Mater Res 7:40–47
Preparation, Characterization, Image Segmentation and Particle … 283

11. Mohammed A, Aboje AA, Auta M, Jibril M (2012) A comparative analysis and characterization
of animal bones as adsorbent. Pelagia Res Libr Adv Appl Sci Res 3:3089–3096
12. Ihom PA (2013) Case hardening of mild steel using cowbone as Energiser. Afr J Eng Res
1:97–101
13. Zhu S, Xia X, Zhang Q, Belloulata K (2007) An image segmentation algorithm in image
processing based on threshold segmentation. Third Int IEEE Conf Signal-Image Technol
Internet-Based Syst 2007:673–678. https://doi.org/10.1109/SITIS.2007.116
14. Kuruvilla J, Sukumaran D, Sankar A, Joy SP (2016) A review on image processing and
image segmentation. Int Conf Data Min Adv Comput 2016:198–203. https://doi.org/10.1109/
SAPIENCE.2016.7684170
15. Ikumapayi OM, Akinlabi ET (2019) Image processing and particle size analysis of coconut
shell. Int J Civ Eng Technol 10:2475–2482
16. Rohilla L, Garg V, Mallick SS, Setia G (2018) An experimental investigation on the effect of
particle size into the flowability of fly ash. Powder Technol 330:164–173. https://doi.org/10.
1016/j.powtec.2018.02.013
17. Zaitoun NM, Aqel MJ (2015) Survey on image segmentation techniques. Procedia Comput Sci
65:797–806. https://doi.org/10.1016/j.procs.2015.09.027
18. Aly A, Deris S, Zaki N (2011) Research review for digital image Segmentation techniques. Int
J Comput Sci 3:99–106. https://doi.org/10.5121/ijcsit.2011.3509
19. Kaur D, Kaur Y (2014) Various image segmentation techniques: a review. Int J Comput Sci
Mob Comput 3:809–814
20. Kaur A (2012) A review paper on image segmentation and its various techniques in image
processing. Int J Sci Res 3:2319–7064 ISSN (Online Impact Factor)
21. Otsu N (1979) A threshold selection method from gray-level histograms. IEEE Trans Syst Man
Cybern 9:62–66. https://doi.org/10.1109/TSMC.1979.4310076
Investigation of Mechanical
and Chemical Properties of the Coir
Fiber and Wood Powder Reinforced
Hybrid Polymer Composite

T. Prabhuram, D. Elilraja, S. Prathap Singh, and Immanuel Durairaj

Abstract In many engineering applications, natural fibers are used because of high
strength, low weight, and easy availability. In many composite studies reveals among
the various natural fibers, the coir fiber is used as a reinforcing agent in polymer
composites. In this paper, the mechanical and chemical properties of the hybrid
polymer composites reinforced with coconut coir fiber and wood powder at different
fractions were studied. The hand layup technique is used for sample preparation
as per ASTM D638-03 standards. The mechanical and chemical properties such as
tensile strength, flexural strength, and impact strength and water absorption capacity
of the hybrid polymer composite were tested for different specimens prepared as per
ASTM D638-03 standards. The results of the above tests prove that the coir fiber and
wood powder have a significant influence on the mechanical properties of the hybrid
polymer composite.

Keywords Hybrid polymer composite · Coir fiber · Wood powder · Hand layup
technique

1 Introduction

Natural fiber-reinforced polymer composites have emerged as an environmentally

friendly and alternative to synthetic fiber-reinforced composites. Natural fiber poly-
mer composites play a major role in many engineering applications such as avi-
ation, automotive (bumpers, battery cases, door panels, dashboards, door handles,
and spoilers), construction, and packaging industries due to their unique proper-
ties including lightweight, ease of processing, availability, and cost-effectiveness.
Rajesh et al. [1] checked the mechanical characterization and influence of chemi-
cal modifications and of natural fiber intra-ply fabric polymer composites. Omrani
et al. [2] deal with state of the art on tribological behavior of polymer matrix com-
posites reinforced with natural fibers. Prabhuram et al. [3] deal with mechanical

T. Prabhuram · D. Elilraja · S. Prathap Singh (B) · I. Durairaj

St. Joseph’s Institute of Technology, Chennai, India
e-mail: prathap.singh50@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 285

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_26
286 T. Prabhuram et al.

properties of hybrid composite materials. Balaji et al. [4] used coir fiber and cot-
ton fiber-reinforced unsaturated polyester composites for packaging applications.
Kattimani et al. [5] conducted experimental investigation on coconut coir reinforced
composite polymer. Prathapsingh et al. [6] investigate the tensile properties of alkali-
treated pantandanus odoratissimus fiber-reinforced with polymer matrix composite.
The main objective of this work is the preparation of coconut coir fiber and wood
powder reinforced polymer composite with different weight fractions and compares
the mechanical and chemical properties of the polymer composite.

2 Materials and Methods

2.1 Coconut Coir Fiber

Coir fiber is one of the naturally available fibers extracted from the coconut husk.
The coir fiber is prepared from coconut husk available from rural areas of all over the
southern states of India. The collected coir material is washed with normal water and
dried in sunlight for several days. Hammering is done on dried coconut husk to extract
coir fiber then it will be cut into the small pieces of 10 mm length approximately.
Figure 1 shows the extracted coir fiber from coconut.

2.2 Wood Powder

The wood powder is collected from the local saw mill. The wood powder used in
this work is extracted from the mango tree. The powder is sieved into the same size
and dried in sunlight for several days. Figure 2 shows the processed wood powder.

Fig. 1 Extracted coconut

coir fiber
Investigation of Mechanical and Chemical Properties of the Coir … 287

Fig. 2 Wood powder of

mango tree

2.3 Preparation of Composite

The hybrid composite specimen was prepared by hand layup technique which one
of the common methodologies used for the preparation of composite materials. The
coir fiber and wood powder act as the reinforcing material, whereas the epoxy resin
acts as the base material. The hand layup technique may be divided into four stages
such as mold preparation, gel coating, layup, and finishing. Molds were made of
wood with a dimension of 30 × 30 × 0.5 cm. Gel coating is applied to the surface of
the mold to separate the composite easily after curing. In layup stage, the required
amount of coir and wood powder are measured along with resin and hardener with
the ratio 10:1. Initially, the resin is mixed with coir fiber and wood powder and
stirred for a few minutes to ensure uniform distribution of reinforcement material
with the resin. After that the hardener is mixed with the resin and again stirred well.
Then, the mixture is poured into the mold cavity and allowed to solidify for one day.
After curing, the composite was taken away from the mold and dried is normal room
temperature for a few hours. Table 1 shows the samples and the weight fractions of
coir fiber and wood powder.

Table 1 Samples with

Samples Percentage of Percentage of Percentage of
weight fraction of coir fiber
epoxy wood coir (%)
and wood powder
resin (%) powder (%)
A 70 30 0
B 70 0 30
C 70 15 15
D 70 20 10
E 70 10 20
288 T. Prabhuram et al.

3 Results and Discussion

3.1 Tensile Test

The tensile test was carried out to determine the tensile strength of the prepared
hybrid composite material specimens. The specimen was prepared as per ASTM
standard for tensile test. The prepared tensile test specimens were inspected under
loading in the universal testing machine. The tensile strength of all the samples is
tabulated in Table 2.
From Fig. 3, it was found that the composite with wood powder reinforcement
shows lower tensile strength than the composite with coir reinforcement. The com-
posites having the mixture of 15% of coir and 15% of wood powder reinforcement
shows greater tensile strength than the individual reinforcement composites. This
shows that the combination of coir and wood powder increased the tensile strength
of the composites. Therefore, the increase in the coir percentage gives good tensile
strength than the increase in wood powder percentage.

Table 2 Tensile strength of coir and wood powder reinforced polymer composites
Sample Length (mm) Width (mm) Thickness mm) Area (mm2 ) Load (kN) Tensile
strength (Mpa)
A 175 13.50 4.40 59.40 0.59 9.96
B 175 11.60 4.10 47.56 0.67 14.07
C 190 12.40 4.20 52.08 0.76 14.59
D 175 12.90 5.20 67.08 0.77 11.41
E 185 13.60 4.80 65.28 0.92 14.11

Fig. 3 Tensile strength comparison of samples

Investigation of Mechanical and Chemical Properties of the Coir … 289

3.2 Flexural Test

The flexural test is carried out to determine the flexural properties of the composite
material. The sample was prepared as per ASTM standard. The sample is kept on
bending fixture and the compressive load is given and the curve generated till the
failure of the sample takes place. Table 3 shows the flexural strength of all the samples.
From Fig. 4, it was found that the composite with wood powder reinforcement
shows lower flexural strength, whereas the coir reinforced composite has greater
flexural strength. Therefore, the increase in the coir percentage gives good flexural
strength than the increase in wood powder percentage.

Table 3 Flexural strength of coir and wood powder reinforced polymer composite
Sample Length Width Thickness Area Load (kN) Flexural
(mm) (mm) (mm) (mm2 ) strength
(N/mm2 )
A 100 14.50 4.10 59.45 0.06 36.92
B 100 14.50 4.60 66.70 0.08 39.11
C 100 14.10 4.60 64.86 0.07 35.56
D 100 15.50 4.30 66.65 0.09 47.10
E 100 15.50 4.30 66.65 0.10 52.34

Fig. 4 Flexural strength comparison of samples

290 T. Prabhuram et al.

Table 4 Impact strength of coir and wood powder reinforced polymer composites
Sample Length of span Width (mm) Thickness (mm) Area (mm2 ) Impact strength
(mm) (J)
A 80 15 5 75 2
B 80 15 5 75 4
C 80 15 5 75 2
D 80 15 5 75 2
E 80 15 5 75 2

Fig. 5 Impact strength comparison of the samples

3.3 Impact Test

The impact test is the study of the ability of the material to withstand the sudden
shock loads. This impact test was conducted in the impact testing machine. Table 4
shows the impact strength of all the samples.
From Fig. 5, it was found that the impact test result shows that there is a consid-
erable increase in the impact strength for coir reinforced composite and there is no
considerable change in impact strength of other specimens. It shows that the impact
strength was improved by increasing the coir percentage.

3.4 Water Absorption Test

Water absorption test is used to determine the amount of water absorbed under room
temperature conditions. Initially, the weights of the samples are measured using the
weighing machine and then the samples were immersed in the normal water for 48 h.
Then, the weights of the samples were measured again. The change in the weight of
Investigation of Mechanical and Chemical Properties of the Coir … 291

Table 5 The water

Sample Initial weight Final weight % of water
absorption capacity of coir
(g) (g) absorption
and wood powder reinforced
polymer composite A 7.12 7.15 0.42
B 6.00 6.07 1.17
C 4.66 4.70 0.87
D 7.78 7.84 0.77
E 6.60 6.68 1.2

Fig. 6 Water absorption capacity comparison of the samples

the sample shows the water absorption capabilities of the composite. Table 5 shows
the water absorption capacity of all the samples.
From Fig. 6, it was found that the composite absorbs a very minimum percentage
of water. So, it is used in application such as storage tank and container fabrication.

4 Conclusions

The mechanical and chemical properties such as tensile strength, flexural strength,
and impact strength, and water absorption capacity have been carried out on hybrid
polymer composite samples. The results obtained from the above test conducted are
concluded that the flexural strength of the composites increases with an increase in the
percentage of coir than increases the percentage of wood powder. The coir reinforced
composite has good impact strength than the wood powder reinforced composite.
The water absorption capacity decreases when increasing amount of wood powder
and a decreasing amount of coir fiber.
292 T. Prabhuram et al.

References

1. Rajesh M, Jeyaraj P (2017) Mechanical characterization of natural fiber intra-ply fabric polymer
composites. Influence of chemical modifications. J Reinf Plast Compos 36(22):1651–1664
2. Omrani E, Menezes PL, Rohatgi PK (2016) State of the art on tribological behavior of polymer
matrix composites reinforced with natural fibers in the green materials world. Eng Sci Technol
Int J 19(2):717–736
3. Prabhuram T, Somurajan V, Prabhakaran S (2010) Hybrid composite materials. In: Frontiers in
automobile and mechanical engineering (FAME), pp 27–31
4. Balaji V, Senthil Vadivu KP (2017) Mechanical characterization of coir fiber and cotton fiber
reinforced unsaturated polyester composites for packaging applications. J Appl Packag Res
9(2):12–19
5. Kattimani P, Raikar N, Patil AN, Nandi P (2016) Experimental investigation on coconut coir
reinforced composite polymer. Int J Engineering Trends Technol 41(3):128–133
6. Prathapsingh D, Elilraja D, Prabhuram T, Immanuel Durairaj J (2018) Investigation of ten-
sile properties of Alkali treated pandanus odoratissimus fiber reinforced with polymer matrix
composite. Int J Mech Prod Eng Res Dev 8(8):403–407
An Experimental Study on Hemp/Sisal
Fiber Embedded Hybrid Polymer
Composites

Akash, Shivakumar Rachoti, Vishwanath Patil, and K. G. Girisha

Abstract Environmental concerns encourage the use of artless fibers since bio-
fiber embedded polymer composite crumble easily in nature and cheaper compare
to petroleum-based fibers. Cellulosic fibers are gradually more gaining attention as
their uses are diversified into structural and automobile parts for vehicle where low
weight is necessary. The low density of cellulosic fibers is very beneficial in the
structural and automobile industry. Cellulosic fibers offer many ecological and tech-
nical benefits for its use in incorporating composites. Current research paper inves-
tigates the mechanical strength (tensile, flexural, compression strength and shore-
D hardness) and water absorption performance of different weight percentage of
hemp/sisal fiber embedded polymer resin (epoxy resin) hybrid composites. Hemp
and sisal fiber surfaces were subjected to 10 weight percentage of alkaline (NaOH)
solution to improve the bonding characteristic between hydrophilic cellulosic fibers
and hydrophobic polymer resin. Hybrid composite specimens were prepared by sim-
ple cold pressing technique. A 40 weight percentage of hemp/sisal fiber embedded
hybrid composite sample discloses maximum tensile strength, flexural strength and
compression strength. Shore-D hardness value and absorption of moisture of hybrid
composite samples increases with increase in fiber loading. Morphological analy-
ses are carried out to analyze the fractured surface of hybrid composite samples by
using scanning electron microscope (SEM) and reported. The investigational result
also discloses that the hemp and sisal fibers are promising reinforcements for use in
cheaper bio-composites which have high strength to weight ratio.

Keywords Bio-fiber · Epoxy resin · Alkaline solution · Cold pressing · SEM

Akash (B) · S. Rachoti · V. Patil

Sharnbasava University Kalaburagi, Kalaburagi, KA 585103, India
e-mail: akashbiradar@gmail.com
K. G. Girisha
Akshaya Institute of Technology, Tumkur, KA 585103, India

© Springer Nature Singapore Pte Ltd. 2021 293

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_27
294 Akash et al.

1 Introduction

Cellulosic fibers are increasingly gaining attention as their uses are diversified into
structural and automobile parts for vehicles where low weight is required. The low
density of cellulosic fibers is very helpful in the structural and automobile industry.
Cellulosic fibers offer many ecological and technical benefits for its use in incorpo-
rating composites. Number of types of cellulosic fibers have been investigated for use
in polymers, including bamboo, sisal, hemp, coir, flax, jute, banana leaf extracted,
pineapple leaf extracted fiber etc., and the polymer matrix used for incorporating
the fibers are classified as thermoplastics and thermosetting plastics. Cellulosic fiber
incorporated composites are also claimed to offer ecological reward such as min-
imized dependence on non-renewable material/energy sources, lower green house
gas emission, lower pollutant emissions, enhanced energy recovery and end of life
biodegradability of components. Cellulosic fibers are renewable, discloses high spe-
cific properties, abundantly available with low density as well as they decompose in
soil and less abrasive to costly molds and mixing tools [1]. Due to their poor resistance
to moisture, incompatibility with hydrophobic polymer resin and their tendency to
form aggregate during processing has reduced their potential use as reinforcement
in polymer matrix composites.
The mechanical strength of cellulosic fiber reinforced polymer composites are
lower compare to synthetic-based fibers [2]. To produce rough surface and reactive
hydroxyl groups on cellulosic fiber surface for better bonding with polymeric resin,
cellulosic fibers have to undergo chemical or/and physical treatment to modify the
structure and surface. The hydrophobic epoxy resin—hydrophilic sisal and hemp
fiber interfacial bond was improved by an alkaline treatment.
Hemp fiber and sisal fiber are most suitable cellulosic fibers for use in polymer
composites, due to its greater specific strength and decompose in nature after use.
Due to the more lignin content sisal and hemp fiber is long durable compared to other
cellulosic fibers. With increasing emphasis on fuel efficiency, cellulosic fibers such
as sisal and hemp fiber incorporated composite samples enjoying wider applications
in structural, domestic and automobiles applications. Many aspects of the use of
sisal and hemp fibers as incorporation in polymer resin composites are discussed in
literature. Rout et al. [3] studied the mechanical properties of untreated and treated
hemp fibers, discloses clear signs of the presence of poor bonding, long pulled out
fibers without any polymer resin adhered to the fibers and poor mechanical strength
were obtained compare to treated fiber embedded composites. Alkaline treatment
to cellulosic fibers increases hydrophobic behavior with improving the resistance to
fungi attack and also increases mechanical strength of polymer composites [4, 5].
According to author [5] 45 wt% of fiber loading is considerable reinforcing effect
when the composite specimens are tested for tension. The utilization of cellulosic
fiber alone, in a polymer resin is not enough in satisfactory tackling all the technical
needs of a fiber embedded polymer composite. In an attempt to develop a better but
inexpensive composite, a cellulosic fiber can be combining with another natural fiber
in the same matrix so as to take the best favorable results of both natural fibers. The
An Experimental Study on Hemp/Sisal Fiber Embedded … 295

aim of this research work is to investigate the mechanical properties and moisture
absorption behavior of sodium hydroxide treated hemp and sisal embedded epoxy
resin polymer composites.

2 Experiments

2.1 Materials

Chopped hemp and sisal fibers were used as incorporation materials. These cellulosic
fibers were procured from local dealers. Epoxy resin and corresponding hardener
with brand name AW106 and HV953, respectively, were used as matrix phase to
manufacture hybrid composites. Epoxy resin (AW 106) and hardener (HV953) at a
ratio of 10:1 was mixed to prepare matrix material.

2.2 Fiber Surface Treatment

A 10 wt% of sodium hydroxide treatment removes cementing materials present in

the cellulosic fiber, namely hemicelluloses, waxes, pectin and lignin, thus increase
in the contact area of the cellulosic fiber. Increased fiber surface area leads to the
better bonding of the matrix and cellulosic fiber, thus increasing the wettability and
mechanical strength of the composites [6, 7]. Chopped hemp and sisal fiber were
taken in separate trays, to these trays 10% NaOH solution was added and the both
hemp and sisal fibers were soaked in the NaOH solution for 10 h. The hemp and
sisal fiber were then washed thoroughly with distilled water to remove the excess of
sodium hydroxide sticking to the hemp and sisal fibers. NaOH treated hemp and sisal
fibers were dried in sun light for three days to remove completely moisture content
from the fibers.

2.3 Preparation of Hybrid Composites

Simple cold pressing method was used to manufacture hybrid composite specimens
in 250 * 250 * 6 mm metallic (cast iron) mold at high pressure. Hydraulic pressing
machine was used to manufacture hemp/sisal fiber embedded epoxy resin hybrid
composites. Thin layers of polyvinyl alcohol (PVA) were coated to both male and
female molds to ensure that the sheet will not stick on to the metallic mold. Epoxy
resin (Araldite AW 106) and hardener (HV 953) were mixed in a ratio of 10:1
in a bowl to prepare polymer matrix material. The quantity of fiber (hemp/sisal
fiber) was varied from 10 to 50 wt% in steps of 10 wt%. A well-mixed mixture
296 Akash et al.

of reinforcement and matrix material was poured into PVA coated female mold die
cavity. A PVA coated male die was positioned on the female die and pressure applied
on male die up to 410.4 kg/cm2 from hydraulic pressing machine for three hours.
After drying, fabricated sample was cut to specimen to carry out tensile test, flexural
test and compression test and water absorption behavior according to ASTM standard
dimensions.

3 Results and Discussions

3.1 Mechanical Properties

Tensile test, compression test, flexural test and hardness test were conducted accord-
ing to ASTM D-3039, ASTM D-695, ASTM 790-03 and ASTM D-2240, respec-
tively. Three specimens were tested for individual test and average values are tab-
ulated. Tensile, flexural and compression strength and shore-D hardness values of
hybrid composite samples are shown in Figs. 1, 2, 3 and 4 respectively. The tensile
strength of hemp/sisal fiber incorporated hybrid composites increases with increase
in fiber loading up to 40 wt%, after that decreases is shown in Fig. 1. This indicates
an effective and uniform stress transfer within composite after the incorporation of

Fig. 1 Tensile strength 49

Tensile Strength (MPa)

48
47
46
45
44
43
42
41
40
39
38
10 20 30 40 50
Fiber percentage

Fig. 2 Flexural Strength

Flexural strength (MPa)

80

75

70

65

60

55
10 20 30 40 50
Fiber percentage
An Experimental Study on Hemp/Sisal Fiber Embedded … 297

Fig. 3 Compression

Compression strength (MPa)

180
strength
175
170
165
160
155
150
10 20 30 40 50
Fiber percentage

Fig. 4 Shore-D hardness 84

number Shore-D hardness value 82
80
78
76
74
72
70
68
66
10 20 30 40 50
Fiber percentage

cellulosic fibers in polymer matrix. At higher % of fiber loading, it is difficult for the
polymer to penetrate the reducing spaces between fibers, leading to poor wetting, and
hence, a reduction in the stress transfers efficiency through the fiber-matrix interface.
Drop in tensile strength is attributed to the poor matrix/fiber interfacial bonding. The
results highlight that the fiber weight fraction is the most dominant parameter that
will affect the tensile strength [8].
Flexural strength of the fabricated hybrid composite specimens increases with
increasing fiber loading up to 40 weight percentage, this enhancement in fiber/matrix
strength is due to the ability of cellulose fiber to resist the bonding force. The lower
flexural strength at lower cellulosic fiber content may be attributed to the lower loads
transferred from the matrix to the fibers, thus resulting in lower load carried by fibers
[9]. Significant increases in flexural strength at higher fiber content is because of
the increase in stress transferred to the fibers as a result of increased adhesion at the
fiber-matrix interface.
Compression strength of the fabricated hybrid composite specimens increases
with increasing fiber loading up to 40 weight percentage this is due to the bonding
at the crossing point between hydrophobic matrix and hydrophilic cellulosic fiber is
too strong to transfer strength [10]. Higher cellulosic fiber concentration demands
higher stress for the same deformation; therefore, the reinforcement of hemp fiber
and sisal fiber into resin resulted in increases the compression strength. Further
addition of hemp fiber and sisal fiber, i.e., 50 weight percentage lead to decrease
298 Akash et al.

the compression strength of composite. When the quantity of cellulosic fibers is not
sufficient to restrain the epoxy resin, large stresses will be developed at low strains
and distribution of these stresses will not be uniform [10, 11].
Shore-D hardness of polymer composite refers to its resistance to shape change
after load/force applied on it. For fiber reinforced polymer composites, hardness value
depends on the distribution of fibers in polymer resin. The shore-D hardness value
was taken at four different spots for every sample and average result was considered
as the hardness value of specimen. Figure 4 shows the variation of hardness value
of different wt% of hemp/sisal fiber embedded hybrid composite. Figure 4 discloses
that the hardness value of composites increases with increasing in fiber loading. The
inclusion of hemp fiber and sisal fiber in epoxy resin has decreased the mobility of
polymer chain in the rigid polymer composite [12].

3.2 Moisture Absorption Behavior

Water absorption properties of hemp/sisal fiber incorporated composite specimens

were studied according to ISO 62:1999 standard. Manufactured hybrid composite
samples were dried in oven at 60 °C and refrigerated to atmospheric temperature.
After weighing the specimens to an accuracy of 0.1 mg, they were immersed in
clean water in a bowl at atmospheric temperature. Daily once, the samples were
taken out from the bowl, all the water particles from the surface were removed
with cotton cloth and samples were weighed. Same process was done for 30 days.
The % of moisture absorption of hemp fiber and sisal fiber incorporated epoxy resin
composites increases with increasing fiber loading [13]. The % of moisture absorption
was plotted against the time as shown in Fig. 5. Figure 5 also discloses the moisture
absorption of hybrid composite increases with the increasing in the soaking time.

10
10 wt.% fiber
9
20 wt.% fiber
8
water absorption in %

30 wt.% fiber
7
40 wt.% fiber
6
50 wt.% fiber
5
4
3
2
1
0
0 5 10 15 20 25 30
Time duration in days

Fig. 5 Water absorption of hybrid composites

An Experimental Study on Hemp/Sisal Fiber Embedded … 299

3.3 Morphological Analysis

The surface characteristics of the hemp/sisal fiber incorporated hybrid composite

sample used for the investigation are studied through scanning electron microscope.
Figure 6 discloses the SEM of fractured surface of fabricated hybrid composite
sample. Figure 6 discloses the gap between fibers and matrix and fiber pullouts which
indicates the weak bonding between the hydrophobic matrix and hydrophilic fibers.
But some of the hemp and sisal fiber are fractured, and it is found to have strong
interfacial bonding between matrix and fibers. The presence of voids in fractured
samples, void space around the periphery of fibers confirms the de-wetting effect
that is responsible for poor interface and de-bonding of matrix-fiber [14].

Fiber pullout

Fiber
Fiber
breakage
debonding

Hemp fiber

Sisal fiber

Fig. 6 SEM images of the hemp/sisal fiber reinforced hybrid composite samples
300 Akash et al.

4 Conclusions

The study on mechanical, water absorption and fractured surface of hemp/sisal fiber
incorporated epoxy resin hybrid composite led to the following conclusions.
• Tensile strength, flexural strength and compression strength of hemp/sisal fiber
embedded hybrid composites increases with increasing fiber weight percentage
in composite specimens up to 40 weight percentage, and then it decreases due to
further loading the percentage of hemp/sisal fibers is incompatible with polymer
resin and has tendency to form aggregates.
• Shore-D hardness value of hybrid composite increases with increasing in weight
parentage of fiber due to addition of hemp/sisal fiber in polymer resin has
decreased the mobility of polymer chain in the rigid composites.
• SEM images reveal that the hydrophilic cellulosic fibers (hemp fiber and sisal
fiber) homogeneously mixed with hydrophobic polymer resin (epoxy resin).
• Absorption of water content increases with increase in the weight percentage of
fibers of manufactured hybrid composite specimens.

References

1. Sanjeevamurthy GC, Rangasrinivas G, Manu S (2012) Mechanical performance of natural

fiber-reinforced epoxy-hybrid composites. Int J Eng Res Appl (IJERA) 2(5):615–619. ISSN:
2248-9622
2. Akash, Venkatesh Gupta NS, Sreenivas Rao KV, Prasad CB, Khadka P (2015) Comparative
valuation of mechanical and water absorption properties of pure epoxy resin, coir fiber/epoxy
resin and hemp fiber/epoxy resin composite. Int J App Eng Res 10(55). ISSN 0973-4562
3. Rout J, Misra M, Mohanty AK, Nayak SK, Tripathy SS (2003) SEM observations of the
fractured surfaces of coir composites. J Reinf Plast Compos 22:1083
4. Hill CAS, Khalil HPSA (2000) The effect of environmental exposure upon the mechanical
properties of coir or oil palm fiber reinforced composites. J Appl Polym Sci 77:1322
5. Hill CAS, Khalil HPSA (2000) Effect of fiber treatments on mechanical properties of coir or
oil palm fiber reinforced polyester composites. J Appl Polym Sci 78:1685
6. Gassan J, Bledzki AK (1999) Possibilities for improving the mechanical properties of
jute/epoxy composites by alkali treatment of fibres. Compos Sci Technol 59:1303–1309
7. Sanjeevamurthy GC, Ranga Srinivas G (2012) Sisal/coconut coir natural fibers—epoxy com-
posites: water absorption and mechanical properties. Int J Eng Innov Technol (IJJEIT)
2
8. Joseph K, Varghese S, Kalaprasad G (1996) Influence of interfacial adhesion on the mechanical
properties and fracture behavior of short sisal fiber reinforced polymer composites. Eur Polym
J 32(10):1243–1250
9. Ghosh R, Krishna AR, Reena G, Raju BL (2011) Effect of fibre volume fraction on the tensile
strength of banana fibre reinforced vinyl ester resin composites. Int J Adv Eng Sci Technol
4(1):89–91
10. Akash, Sreenivasa Rao KV, Venkatesha Gupta NS, Arun kumar DS (2016) Mechanical proper-
ties of sisal/coir Fiber reinforced hybrid composites Fabricated by cold pressing method. IOP
Conf Ser Mater Sci Eng 149:1–7
An Experimental Study on Hemp/Sisal Fiber Embedded … 301

11. Venkatesha Gupta NS, Akash, Sreenivasa Rao KV, Arun Kumar DS (2016) Fabrication and
evaluation of mechanical properties of alkaline treated sisal/hemp fiber reinforced hybrid
composite. IOP Conf Ser Mater Sci Eng 149:1–7
12. Mir SS, Nafsin N, Hasan M, Hasan N, Hassan A (2013) Improvement of physico-mechanical
properties of coir-polypropylene biocomposites by fiber chemical treatment. Mater Des
52:251–257
13. Akash, Girisha KG, Venkatesha Gupta NS, Sreenivas Rao KV (2017) A study on flammability
and moisture absorption behavior of sisal/coir fiber reinforced hybrid composites. IOP Conf
Ser Mater Sci Eng 2:1–6
14. Singh NP, Aggarwal L, Gupta VK (2015) Tensile behavior of sisal/hemp reinforced high density
polyethylene hybrid composite. Mater Today Proc 2:3140–3148
Investigation on Chemical Isolation
and Characterization of Cellulose
from Delonix regia Fruit Fibers

Kalpit P. Kaurase and Dalbir Singh

Abstract Lignocellulosic fibers have gained popularity among the research frater-
nity in recent years due to its abundance in nature, biodegradability, high specific
strength, etc. Cellulose can be extracted from biomasses by various chemical and
mechanical methods. In this paper, Delonix regia fruit fibers are presented as a new
source of cellulose, and chemical methods are used for extraction of cellulose from
it. Chemically treated and untreated fibers were characterized by X-ray diffraction
(XRD) and thermogravimetric analysis (TGA) to understand the effect of chemical
treatment and properties of the cellulose yield. The results obtained demonstrated
that a high yield of cellulose is obtained by chemical extraction methods followed
and validated the suitability of Delonix regia fruit fibers for various applications
including biocomposites and bionanocomposites.

Keywords Cellulose · Hemicellulose · Lignin · Bleaching · Alkali treatment ·

Lignocellulosic biomass · Delonix regia fruit fibers

1 Introduction

Lignocellulosic biomass is obtained from all types of plant materials, agricultural

products, urban wastes, etc., and hence, it is abundantly available in nature. Lignocel-
lulosic biomass in the raw and modified forms has been utilized from ancient times
by human beings in various applications such as fodder for animals, food source,
building houses, cloths, energy drinks and medicines. Plant biomass mainly consists
of high energy materials like fat, protein, starch, etc., and plant cell walls which are
made of cellulose, hemicellulose and lignin. Cellulose is the main structural compo-
nent of cell wall which provides strength to the material, and it is very light in weight
[1, 2]. Cellulose extracted from these lignocellulosic biomasses can be used in plenty

K. P. Kaurase (B) · D. Singh

Hindustan Institute of Technology & Science, Hindustan University, Chennai, India
e-mail: kalpitkaurase@gmail.com
D. Singh
e-mail: dsinghcad@gmail.com
© Springer Nature Singapore Pte Ltd. 2021 303
S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_28
304 K. P. Kaurase and D. Singh

of applications such as medical applications, food packing industry, biocomposites,

bionanocomposites, etc.
Major component of lignocellulosic biomasses is water, but on dry basis, the
main constituents are cellulose, hemicellulose and lignin which comprise almost
80% of the total weight, and remaining constituents are pectin, protein, lipid, ash,
carbohydrate and wax. Percentage of these constituents varies in different plants, and
also, in same plant, it varies in different parts and according to different climatic and
growing conditions [3].
Cellulose and hemicellulose together are termed as holocellulose, and it con-
sists of 65–70% of total dry mass. Cellulose is an organic compound with formula
(C6 H10 O5 )n of D-glucopyranose units bonded together by β-glucosidic bonds. Glu-
cose units altogether are frequently called as degree of polymerization (DP), and it
varies from plant to plant. Cellulose is tasteless, odorless, hydrophilic white crys-
talline compound and insoluble in most of the organic solvents. It consists of almost
80% crystalline region and remaining amorphous region [4, 5].
Hemicellulose is a combination of polysaccharide polymers and consists of sugars
of glucose, xylose, mannose, galactose, arabinose and rhamnose. It is a branched
polymer and one of the most complex components of cell wall. It is amorphous
compound having degree of polymerization less than that of cellulose [6].
Lignin is heterogeneous cross-linked phenolic polymers and serves the purpose
of bonding cellulose and hemicellulose through strong hydrogen bonding and stable
lignin–carbohydrate complex. There are basically three types, derived from phenyl
propane, 4-hydroxy 3-methoxyphenylpropane, 3, 5-dimethoxy 4-hydroxy phenyl
propane and 4-hydroxy phenyl propane [7].
Inorganic contents which are generally measured as ash content are also present
in lignocelluloses. Proteins which are polymers of amino acids are available in high
concentrations in some plants. Other contents of the lignocellulosic plants are termed
as extractives which consist of waxes, terpenes, fats, fatty acids, phenols, etc.
Delonix regia is a flowering plant belonging to Fabaceae family of bean trees,
and it is mostly found in tropical forests. In India, it is popularly known as Gulmohar
plant and can be seen all across the country having very high population [8]. Delonix
regia fruits are very hard having high strength and almost unutilized for any kind of
applications in India and are potential biomass being wasted every year which can be
used for cellulose extraction. The main objective of this research work is to extract
cellulose for D. regia fruit fiber using chemical extraction methods and to characterize
the treated and untreated fibers by using thermogravimetric analysis (TGA), X-ray
diffraction (XRD) and differential scanning calorimetry (DSC) to verify the cellulose
extraction process.
Investigation on Chemical Isolation and Characterization … 305

2 Materials and Methods

2.1 Materials

Delonix regia fruits were collected from the local fields of Raipur, Chhattisgarh, India.
Chemicals used were: toluene, ethanol, sodium hydroxide, sodium hypochlorite,
hydrogen peroxide and sulfuric acid. Various chemicals used were of analytical
grades and obtained from Kasliwal Brothers, Raipur, Laxmi Scientific Works, Raipur
and Lab Chemicals, Chennai.

2.2 Experimental Procedures

2.2.1 Sample Preparation

Green fruits of D. regia were cut from the D. regia tree, and seeds of fruit were
removed. Delonix regia fruits were washed several times with deionized water and
chopped into small pieces (approx. 2 cm × 2 cm) by chopper. Chopped fibers were
dried in a convective oven at 80 °C for 24 h and grounded in a mill and sieved
to attend uniform size in powder form and then washed and filtered to remove the
dirt or waxing substances and improving swelling tendency as per NREL standard
procedures [9]. Dirt-free and screened samples were stored in a polyethylene bags
and kept in a desiccator for further use. Figure 1 shows the D. regia fruits biomass
before and after size reduction.

2.2.2 Chemical Composition Determination

Size-reduced and screened samples of D. regia fruits were subjected to subsequen-

tial gravimetric composition analysis methods as per NREL Laboratory Analytical
procedures (LAP) [9, 10].

Moisture Content
Screened sample (8 g) was subjected to convection oven drying at 105 °C for 24 h
and weighed in balance and stored in a desiccator to avoid the moisture absorption.
Total solid in the biomass is obtained by weight difference before and after oven
drying and subtracting it from 100 gives moisture content of the biomass [9–11].

Extractive Content
Dry sample (5 g) was taken in a tared extraction thimble, and Soxhlet apparatus was
assembled, and thimble was inserted in Soxhlet tube. Ethanol (190 ml) was added
as solvent for extraction in the receiving flask and placed on Soxhlet apparatus.
Temperature of the heating plate was kept at 80 °C, and distillation flask was kept in
306 K. P. Kaurase and D. Singh

Fig. 1 Raw biomass (Delonix regia fruit) before and after size reduction; a Delonix regia tree,
b Delonix regia fruits (on tree), c Delonix regia fruits (Cut), d removed seeds, e dried and chopped
fruits, f fruits after size reduction and screening

a water bath for the reflux time of 16 h. After complete extraction, sample was kept in
Buchner funnel with cellulose filter paper and air dried at room temperature [9–11].
Weight difference before and after extraction process gives the extractive content.

Hemicellulose Content
Distilled water and NaOH (150 ml) were added to dry extracted sample (3 g) in
Erlenmeyer flask and boiled for 4 h. Mixture was air cooled, vacuum filtered and
washed to reach neutral pH after completion of boiling time. Residual content was
dried in oven at 105 °C, and weight difference before and after treatment gives
hemicellulose content [9–11].

Ash Content
Dried extracted sample (2 g) was placed in a tared crucible and ashed in a muffle
furnace at 575 °C for 24 h. Crucible was removed and placed in a desiccator (to
avoid moisture absorption) for initial cooling before weighing after complete ashing.
Weight difference before and after complete ashing gives the ash content of the
biomass [9–11].

Lignin Content
72% H2 SO4 (3 ml) was added to dry extracted sample (1 g) and kept at room tem-
perature for 2 h with continuous mixing by magnetic stirrer. After initial hydrolysis,
distilled water (84 ml) was added and kept in autoclave for 1 h at 120 °C to complete
hydrolysis. Then, mixture was air cooled at room temperature and vacuum filtered.
Investigation on Chemical Isolation and Characterization … 307

Acid insoluble lignin was calculated by drying residue at 105 °C and accounting for
ash content. Acid soluble lignin was determined by acid digestion 320 nm. Lignin
content of biomass was obtained by summing acid insoluble and acid soluble lignin
content [11].

Cellulose Content
Cellulose content was estimated by considering that biomass contains moisture, ash,
lignin, hemicellulose, extractives and cellulose and subtracting moisture, ash, lignin,
hemicellulose, extractive percentage from 100%.

2.2.3 Cellulose Extraction

The following techniques were utilized for extracting cellulose from D. regia fruit
fibers: Size-reduced and screened D. regia fruit fiber samples (20 g) were first
subjected to extractive removal procedure by using Soxhlet apparatus. 2:1 (v/v)
toluene/ethanol mixture was added to the round flask and placed on heating mantle,
and Soxhlet apparatus was assembled for the total reaction time of 6 h at 90 °C, fol-
lowed by washing with ethanol and drying [12]. Fibers were pulped chemically with
17.5% (w/w) NaOH and kept at room temperature overnight to improve swelling
tendency of the fiber and partial removal of lignin and hemicellulose, successive
cleaning with tap water, and filtration was carried out afterward. Fibers were deligni-
fied with 72% (v/v) sulfuric acid (H2 SO4 ) at 70 °C for 1 h followed by washing with
distilled water and centrifuged several times to reach neutral pH, and supernatant was
collected [13]. Fibers were subjected to alkaline treatment with 2% (w/w) NaOH at
90 °C for 2 h in a water bath to remove hemicellulose. Fibers were bleached 5 times
with 5% (v/v) sodium hypochlorite at 75 °C for 2 h followed by repetitive rinsing and
filtration. Fibers were further bleached with 10% (v/v) hydrogen peroxide (H2 O2 ) at
80 °C for 1 h, and filtrate was oven dried at 40 °C for 5 h to obtain highly purified
yield of cellulose [14] (Fig. 2).

Fig. 2 Various stages of chemical extraction process a pulping, b delignification, c bleaching,

d extracted cellulose
308 K. P. Kaurase and D. Singh

2.3 Material Characterization

Material characterization was conducted for raw D. regia fruit fiber samples and
cellulose extracted by chemical treatments and termed as sample 1 and sample 2,
respectively. Characterization was carried out using three techniques: X-ray diffrac-
tion (XRD), thermogravimetric analysis (TGA) and differential scanning calorimetry
(DSC).

2.3.1 X-Ray Diffraction (XRD)

X-ray diffraction analysis was performed on Panalytical Xpert-Pro diffractometer

using Cu-Kα radiation with wavelength of λ = 1.5406 nm at generator voltage of
40 kV and tube current of 30 mA. Samples were scanned at 2θ range of 20° to 100°
with the scan step size of 0.026.
Crystallinity index (CrI) was calculated by the equation:
 
(I002 − Iam )
CrI = × 100 (1)
2

where I002 is the intensity of the peak at 2θ angle between 26° and 27° representing
crystalline region, Iam is the intensity of the peak at 2θ angle between 21° and 22°
representing amorphous region [15, 16].

2.3.2 Thermogravimetric Analysis (TGA)

Thermogravimetric analysis (TGA) was conducted by TGA-DTG Stare Mettler-

Toledo system equipped with high throughput autosampler and Stare SW 13.00
software. For sample 1, initially, 7.9850 mg of weighed sample was taken in an
aluminum crucible, and for sample 2, initially, 7.3000 mg of weighed sample was
taken. Heating rate of 10.00 K/min from 25 to 700 °C in a nitrogen environment with
flow rate of 70 ml/min was used for both the samples. Change in mass of the sample
was recorded by microbalance system. First, differential curve (DTG) and TGA curve
were obtained by Stare SW 13.00 software by TGA 1 SF/1100/321 module [17].

2.3.3 Differential Scanning Calorimetry (DSC)

Differential scanning calorimetry (DSC) was conducted by DSC Stare Mettler-Toledo

system equipped with high throughput autosampler and Stare SW 13.00 software.
For sample 1, initially, 4.8000 mg of weighed sample was taken, and for sample 2,
initially, 5.2000 mg of weighed sample was taken in an aluminum crucible. Heating
rate of 10.00 K/min from 25 to 700 °C in a nitrogen environment with flow rate of
Investigation on Chemical Isolation and Characterization … 309

70 ml/min was used for both the samples. Difference in the amount of heat required
to raise the temperature of the sample was recorded, and DSC curve was obtained
by Stare SW 13.00 software by DSC 1/700/2861 module [18].

3 Result and Discussion

3.1 Chemical Composition of the Delonix regia Fruit Fibers

Chemical composition of the D. regia fruit fibers was determined by experimental

procedures. All the experimental procedures were repeated thrice, and the average
of results was determined. Table 1 shows the chemical composition of D. regia fruit
fiber obtained from experimental procedures followed.

3.2 X-Ray Diffraction (XRD)

For analyzing the crystallinity index and crystallite size of the extracted cellulose
with raw D. regia Fruit fiber sample, X-ray diffractometry was conducted. Figure 3
shows diffraction pattern of the extracted cellulose with raw D. regia Fruit fiber

Table 1 Chemical composition of Delonix regia fruit fiber (%w/w)

Cellulose (%) Hemicellulose Lignin (%) Extractives (%) Moisture (%) Ash (%)
(%)
66.9 11.6 6.5 4.5 8.3 2.2

Fig. 3 XRD plot for raw

(untreated) Delonix regia 10000
fruit fiber sample (sample 1)
and chemically treated, i.e., 8000
Intensity (a.u.)

extracted cellulose sample Sample 1 (Untreated)

(sample 2) Sample 2 (Chemically Treated)
6000

4000

2000

20 40 60 80 100
Angle (2*theta) (deg.)
310 K. P. Kaurase and D. Singh

sample. Two peaks were observed for sample 1 at 2θ values of 22.6855° and 186
34.8198° and most of the portion of curve is flat with the absence of sharp peaks which
shows that the major the amorphous nature of the untreated material. For chemically
treated sample, total 24 peaks were observed having narrow and sharp regions which
represent crystalline nature of chemically treated material. Major peak for sample 2
was seen at 2θ value of 29.567°. Crystallinity index (CrI) was found to be 52.24% for
raw D. regia fruit fiber sample (sample 1) and 73.41% for chemically treated sample
(sample 2). Improvement in crystallinity index by chemical treatment is observed
which clearly indicates that all the amorphous material such as hemicellulose, lignin
and wax has been successfully removed leaving behind crystalline material, i.e.,
cellulose [19, 20].

3.3 Thermogravimetric Analysis (TGA)

Chemical structure of various constituents of lignocellulosic biomass such as waxes,

lignin, hemicellulose and cellulose differs from each other, and hence, they degrade
at different temperatures. Figure 4 shows the weight loss and its derivative during
heating for raw (untreated) D. regia fruit fiber sample (sample 1), and Fig. 5 shows the
weight loss and its derivative for chemically treated, i.e., extracted cellulose sample
(sample 2). From these plots, we can observe that sample 1 and sample 2 has first
degradation at 69.5 °C and 63.8 °C, respectively with reduction of almost 7.23% and
28.99% of the total mass which evaporation or loss of moisture content. The second
degradation starts at 173.6 °C and 166.4 °C for sample 1 and 2, respectively, which
represents degradation of hemicellulose, and lignin has started, and major weight loss
is observed at 305.2 °C and 334.6 °C for sample 1 and 2, respectively, with 59.87% and

Fig. 4 TGA and DTG plots for raw (untreated) Delonix regia fruit fiber sample (sample 1)
Investigation on Chemical Isolation and Characterization … 311

Fig. 5 TGA and DTG plots for chemically treated, i.e., extracted cellulose sample (sample 2)

45% reduction in total mass which represents the degradation of cellulose and lignin
in sample 1 and degradation of cellulose in sample 2. This degradation continues
further up to 600 °C till the ashing point after which degradation becomes negligible.
The results indicate that weight loss at first degradation is less in untreated sample as
compared to chemically treated sample due to fact that moisture absorption is less in
untreated sample which is due to the presence of wax and other extractive content.
Weight loss at the second degradation is more in untreated sample as compared to
chemically treated sample because of the presence of lignin in untreated sample
[21–23].

3.4 Differential Scanning Calorimetry (DSC)

Differential scanning calorimetry (DSC) was used to determine the thermal behav-
ior of untreated sample (sample 1) and chemically treated sample, i.e., cellulose
extracted (sample 2). From the results, we can observe that endothermic peak starts
20–130 °C which represents the energy consumed in evaporating the moisture con-
tent; low energy consumption is seen for untreated sample as compared to chemically
treated sample because of the presence of wax and other extractive content. Fur-
ther, endothermic peak was observed between 260 and 370 °C in untreated sample,
due to energy consumption for degradation of lignin, hemicellulose and cellulose,
whereas almost uniform rise in energy consumption 260–400 °C without any peak
was observed in chemically treated sample which represents the presence of pure
crystalline cellulose [24] (Figs. 6 and 7).
312 K. P. Kaurase and D. Singh

Fig. 6 DSC plot for raw (untreated) Delonix regia fruit fiber sample (sample 1)

Fig. 7 DSC plot for chemically treated, i.e., extracted cellulose sample (sample 2)

4 Conclusion and Future Work

From the results obtained following conclusions can be drawn:

1. Cellulose content of the D. regia fruit fiber is 66.9% which is higher than most
of popular cellulose sources, and due to high cellulose yield, it proves to be a
potential candidate for cellulose and nanocellulose extraction and application in
cellulose-based biocomposites and bionanocomposites,
2. XRD, TGA-DTG and DSC results proved that highly purified yield of cellulose
is successfully extracted by the chemical route followed,
Investigation on Chemical Isolation and Characterization … 313

3. Delonix regia fruit fiber is suggested to be a suitable source for cellulose

extraction, and in the future, further studies can be carried out for nanocel-
lulose extraction and its application in cellulose-based biocomposites and
bionanocomposites.

Acknowledgements I would like to thank Bombay Textile Research Association (BTRA), Mum-
bai (India), for providing necessary support and scientific help in compositional analysis. Also, I
would like to thank National Institute of Technology, Raipur and Department of Physics, School of
Sciences, Pt. Ravishankar Shukla University, Raipur for assisting in characterization.

References

1. Krasznai DJ, Champagne Hartley R, Roy HM, Champagne P, Cunningham MF (2018) Com-
positional analysis of lignocellulosic biomass: conventional methodologies and future outlook.
Crit Rev Biotechnol 38(2):199–217
2. Han JS (1997) Chemical composition of fibers, Chapter 5
3. Watkins D, Nuruddin M, Hosur M, Tcherbi-Narteh A, Jeelani S (2015) Extraction and
characterization of lignin from different biomass resources. J Mater Res Technol 4(1):26–32
4. Mohanty AK, Misra M, Hinrichsen G (2000) Biofibres, biodegradable polymers and biocom-
posites: an overview summary of contents 1. Introduction 2. Reinforcing biofibres 2.1 Chem-
ical constituents and structural aspects 2.2 Properties of biofibres 2.3 Degradation properties
of biofibres 2.4 Cost asp
5. Möller J (2014) Comparing methods for fibre determination in food and feed
6. Welker CM, Balasubramanian VK, Petti C, Rai KM, De Bolt S, Mendu V (2015) Engineering
plant biomass lignin content and composition for biofuels and bioproducts. Energies 8(8):7654–
7676
7. Ayeni A, Adeeyo O, Oresegun O, Oladimeji T (2015) Compositional analysis of lignocellulosic
materials: Evaluation of an economically viable method suitable for woody and non-woody
biomass. Am J Eng Res 4(4):14–19
8. Senthilkumar N, Murugesan S, Sumathi R, Babu DS (2015) Tree species in India, vol 6, no 2,
pp 25–28
9. Templeton DW, Scarlata CJ, Sluiter JB, Wolfrum EJ (2010) Compositional analysis of
lignocellulosic feedstocks. 2. Method uncertainties. J Agric Food Chem 58(16):9054–9062
10. Sluiter A et al (2011) Determination of structural carbohydrates and lignin in biomass
determination of structural carbohydrates and lignin in biomass. Natl Renew Energy Lab
2011(July)
11. Giger-Reverdin S (1995) Review of the main methods of cell wall estimation: interest and
limits for ruminants. Anim Feed Sci Technol 55(3–4):295–334
12. Sun RC, Fang JM, Mott L, Bolton J (1999) Extraction and characterization of hemicellu-
loses and cellulose from oil palm trunk and empty fruit bunch fibres. J Wood Chem Technol
19(1):167–185
13. Martins MA, Teixeira EM, Corrêa AC, Ferreira M, Mattoso LHC (2011) Extraction and charac-
terization of cellulose whiskers from commercial cotton fibers. J Mater Sci 46(24):7858–7864
14. Sun JX, Sun XF, Zhao H, Sun RC (2004) Isolation and characterization of cellulose from
sugarcane bagasse. Polym Degrad Stab 84(2):331–339
15. Ibrahim MM, Agblevor FA, El-Zawawy WK (2010) solation and characterization of cellulose
and lignin from steam-exploded lignocellulosic biomass. Biol Syst Eng 5:397–418
16. Singh S et al (2015) Comparison of different biomass pretreatment techniques and their impact
on chemistry and structure. Front Energy Res 2(February):1–12
314 K. P. Kaurase and D. Singh

17. Suryanto H, Marsyahyo E, Irawan YS (2013) Morphology, structure, and mechanical properties
of natural cellulose fiber from Mendong grass (Fimbristylis globulosa). J Nat Fibers (November
2014):37–41
18. Morán JI, Alvarez VA, Cyras VP, Vázquez A (2008) Extraction of cellulose and preparation of
nanocellulose from sisal fibers. Cellulose 15(1):149–159
19. Maheswari CU, Reddy KO, Muzenda E, Guduri BR, Rajulu AV (2012) Extraction and char-
acterization of cellulose microfibrils from agricultural residue e Cocos nucifera L. Biomass
Bioenerg 46:555–563
20. Reddy KO et al (2018) Extraction and characterization of cellulose single fibers from native
african napier grass. Carbohydr Polym 188:85–91
21. Kalita RD, Nath Y, Ochubiojo ME, Buragohain AK (2013) Extraction and characterization
of microcrystalline cellulose from fodder grass; Setaria glauca (L) P. Beauv, and its potential
as a drug delivery vehicle for isoniazid, a first line antituberculosis drug. Colloids Surf B
Biointerfaces 108:85–89
22. Jahan MS, Saeed A, He Z, Ni Y (2011) Jute as raw material for the preparation of
microcrystalline cellulose. Cellulose 18(2):451–459
23. Nuruddin M et al (2011) Extraction and characterization of cellulose microfibrils from
agricultural wastes in an integrated biorefinery initiative. Cellul Chem Technol 45(5–6):5–6
24. Ohwoavworhua F, Adelakun T, Okhamafe A (2009) Processing pharmaceutical grade micro-
crystalline cellulose from groundnut husk: extraction methods and characterization. Int J Green
Pharm 3(2):97
Experimental Study and Analysis
of Defragmented Carbon Nanotubes
in Polyacrylonitrile Matrix

N. Arunkumar , Joven Job, D. Ananthapadmanaban ,

N. E. Arun Kumar , and N. Sathishkumar

Abstract The aim of this research is to elucidate the change in electrical properties
of polyacrylonitrile (PAN) polymer when reinforced with defragmented multiwall
carbon nanotubes. In conventional chopped fiber-reinforced polymer composites,
uniform distributions of fibers throughout the matrix are critical for producing mate-
rials with superior physical and electrical properties. The previous methods have
dispersed carbon nanotubes by aggressive chemical modification of the nanotubes
or by the use of a surfactant prior to dispersion. Here, ultrasonic energy was used to
uniformly disperse and defragment the multiwall nanotubes (MWNTs) in solutions
and to incorporate them into composites without chemical pretreatment. A common
solvent dimethylformamide (DMF) is used for the dissolution process. The film is
formed by solvent casting method which involves evaporation of the homogeneous
solution of CNT and polymer. Electrical characteristic studies were done on the film
samples. The electrical properties such as conductivity and resistivities were found
and compared to the original polymer matrix.

Keywords Carbon nanotubes (CNT) · Dimethylformamide (DMF) ·

Nanocomposite · Single-wall carbon nanotubes (SWNTs) · Multiwall carbon
nanotubes (MWNTs) · Polyacrylonitrile (PAN) · Chemical vapor deposition (CVD)

1 Introduction

Recent theoretical and experimental studies suggest that carbon nanotubes have
remarkable mechanical and electrical properties [1–4]. In this article, the electrical
properties of carbon nanotubes [5] are applied to enhance the conducting property

N. Arunkumar (B) · J. Job · N. E. Arun Kumar · N. Sathishkumar

Department of Mechanical Engineering, St. Joseph’s College of Engineering, Old Mahabalipuram
Road, Chennai, Tamil Nadu 600119, India
e-mail: narunkumar72@gmail.com
D. Ananthapadmanaban
Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Kalavakkam, Kanchipuram, Tamil Nadu 603110, India

© Springer Nature Singapore Pte Ltd. 2021 315

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_29
316 N. Arunkumar et al.

of a poor conductor [6]. The previous study indicated that most polymers are poor
conductors [7]. Our broad objectives were to explore the development of advanced
composites in which carbon nanotubes serve as the reinforcing element in a poly-
mer matrix. Although several studies have focused on producing polymer nanotube
composites, many practical challenges remain before their potential can be fully real-
ized. Dispersing the nanotubes individually and uniformly into the matrix seems to
be fundamental in producing composites with reproducible and optimal properties
[8, 9]. The article focuses on dispersing defragmented MWNTs into the polymer
matrix using solvent casting process [10, 11] and analyzing the change in the elec-
trical properties of the nanocomposite material. Polyacrylonitrile (PAN) was chosen
as the matrix material because its properties are well known; it is soluble in a broad
range of solvents; and its clarity allows dispersion of MWNTs to be observed at the
micron scale [12]. DMF (dimethylformamide) was chosen as a common solvent.
MWNTs were prepared without any aggressive chemical pretreatment, purification,
or modification using CVD method [13]. MWNTs and the polymer (PAN) solutions
were dispersed and defragmented using a probe sonication process. The solution
samples were evaporated to form a film using solvent casting process [14] which
was later characterized by scanning electron microscopy (SEM) to determine disper-
sion and orientation of MWNTs within the PAN matrix. Composite samples were
tested for its electrical properties such as conductivity and resistivity, and the results
were plotted and tabulated.

2 Experimental

2.1 Synthesis of CNT (CVD Method)

See Fig. 1.

2.2 Process

The catalyst employed was a 1:1 mixture of misch metal-based alloy (MmNi3 ) and
graphene oxide (GO). Hydrogen decrepitation was employed for powdering the ingot
of MmNi3 , and a modified Hummer’s method [15] was used for preparing GO as
described by the authors [16, 17]. The catalysts were grinded together in a mortar for
a few minutes and then loaded into the center of a horizontal quartz tubular reactor,
and hydrogen was allowed to enter at 240 °C for 30 min. Hydrogen reduction of GO
to graphene has been observed under these conditions. The sample was heated further
at 700 °C, before introducing a burst of acetylene (C2 H2 ) as the carbon containing
precursor, in addition to the argon (Ar) flow. The as grown samples were purified
by refluxing in conc. HNO3 at 60 °C for 6 h where the end hemispherical distorted
Experimental Study and Analysis of Defragmented Carbon … 317

Fig. 1 Schematic illustration of the synthesis of graphene-coated CNT composites

carbon layers were removed. Acid-treated samples were washed with deionized water
until pH 7, filtered using a 0.2 micron polytetrafluoroethylene (PTFE) membrane,
and dried at 60 °C in a vacuum. The graphene-wrapped CNT samples grown at
240 °C and 500 °C of H2 flow were labeled as graphene–CNT-240 and graphene–
CNT-500, respectively. The growth mechanism of the graphene-wrapped CNT is
now considered. At 500 °C, GO exfoliated into graphene sheets in the presence of
hydrogen gas. The reaction was further carried out by allowing acetylene at 700 °C.
When the catalyst particle (MmNi3 ) is saturated with carbon, carbon leaves the
particle in another form. Saturated carbon atoms nucleate into six membered rings
and finally lead to CNT formation which is described in Fig. 1. The magnified
images using FESEM and TEM are described in Fig. 2. This is the well-reported
vapor liquid solid (VLS) mechanism for the growth of carbon nanotubes by chemical
vapor deposition. The graphene nanosheets located at a distance of ≥0.35 nm from
a nanotube get affixed to the surface of the nanotube by van der Waals interaction.

2.3 Observation and Inferences

The catalyst (MmNi3 ) must be crushed into very fine powder. The growth of the
carbon nanotubes takes place on the surface of the catalyst. Hence, finer the particles
of the catalyst more surface area is formed for nanotube growth. The tubular structure
of CNTs depends upon the insertion of gases at optimal temperature. Hydrogen is
passed at 2400 °C to reduce the graphene oxide into graphene. The oxide group
reacts with hydrogen and leaves as water vapor from the furnace. Graphene is the
main source of carbon to form the carbon nanotubes. Acetylene is passed at 7000 °C
to separate the carbon nanotubes from the misch metal catalyst.
318 N. Arunkumar et al.

Fig. 2 a Low, b high-magnification FESEM images and e TEM images of graphene–CNT-500;

and c low, d high-magnification FESEM images f TEM images of graphene–CNT-240

2.4 FESEM and TEM Images of Purified CNT

The CNTs formed by CVD method were checked for morphology using field elec-
tron scanning electron microscope (FESEM) and transmission electron microscope
(TEM) at different magnifications. The tubular structures are the carbon nanotubes
indicated in Fig. 2.

2.5 Carbon Nanotube Production Device by Chemical Vapor

Deposition

Three independent temperature adjustors are applied to control the temperature at

different positions of quartz tube furnace to gain high temperature balance so that the
created carbon tube has high crystallizability. Various carbon nanotubes or nano-coil,
Experimental Study and Analysis of Defragmented Carbon … 319

micro-carbon coil, and hexagonal crystal lattice may be created and form film. Put
silicon, quartz, porcelain, oxide, and other carriers in the external import device with
catalyzer precursor which is prepared at advance, catalyst and carbon nanotube will
be created at the same time, without the necessity to prepare catalyzer separately.
The device can easily prepare highly dispersed CNT compound with CNT lamella
structure/core structure of compound material. The catalyst is placed in the center
of the furnace to ensure that there is uniform heating and uniform chemical vapor
deposition. The gases and other chemicals are passed as mentioned earlier, and
various formation processes of the carbon nanotubes take place at their respective
temperatures.

2.6 Preparation of Polyacrylonitrile Polymer (PAN)

The polymer used in this research was purchased from Sigma-Aldrich. The polymer
has 99% purity and a molecular weight ranging from 1,50,000 to 2,00,000 mol.
Preparation method adopted to prepare this polymer is described by the authors of
[18] in Fig. 3.

Fig. 3 Flowchart of the preparation process of polyacrylonitrile

320 N. Arunkumar et al.

2.7 Preparing the Polymer Solution

Polyacrylonitrile (PAN) powder was first weighed and placed in the beaker described
in Fig. 4. It is then dissolved in DMF using a probe sonicator which produces ultra-
sonic waves to form a homogeneous polymer solution described in Fig. 5. The time
used for obtaining a complete homogeneous solution of PAN is about 12 min having
an OFF and ON cycle of 5 s and with a pulse of 50 Hz.

Fig. 4 Polyacrylonitrile
polymer (PAN) weighed in a
beaker before preparation of
PAN solution

Fig. 5 Prepared PAN

solution after probe
sonication process
Experimental Study and Analysis of Defragmented Carbon … 321

Fig. 6 Prepared CNT

solution after probe
sonication process

2.8 Preparing the Carbon Nanotube Solution

The multiwall carbon nanotubes were weighed using the electronic weighing balance.
Solutions were created with wt% varying from 1 to 10. The solvent used in this
application is the same dimethylformamide (DMF). 60 ml of the solvent was taken
in 100 ml beaker, and the carbon nanotubes were added. The amount of carbon
nanotubes added for every sample preparation varies from 10 to 100 mg, respectively.
A total of 10 samples were created with weight percentage 1–10 %wt. The probe
sonicator was used to disperse and defragment the carbon nanotubes in DMF. The
complete dispersion takes place in 20 min and having an ON cycle of 15 s and
OFF cycle of 5 s and with a pulse of 50 Hz. The final homogeneous solution of
defragmented CNT and DMF is shown in Fig. 6.

2.9 Mixing the Polymer and Carbon Nanotubes Solutions

It was ensured that the solvent chosen (DMF) would be a common solvent for both
PAN and CNTs, thereby assuring no chemical reactivity between solvents. The two
solutions formed in 2.7 and 2.8 were mixed using a probe sonicator for about 30 min
with a pulse of 50 Hz and an ON and OFF cycle of 5 s. A homogeneous solution of
defragmented CNT and PAN was formed which is described in Fig. 7. Evaporation
process is carried out by pouring the solution in a 4-inch petri dish and placing it on
a hotplate as shown in Fig. 8. The hot plate is maintained at a constant temperature
of 90 °C. A vacuum furnace can also be used for evaporation purpose.
The evaporation process takes place till all the solvent has evaporated and a thin
film (50 µm thick) of composite material is formed in the petri dish as depicted
in Fig. 9. The film was then peeled from the petri dish carefully without tearing
as shown in Fig. 10. This method adopted for preparing a thin film by evaporating
322 N. Arunkumar et al.

Fig. 7 Prepared
homogeneous solution of
CNT and PAN by probe
sonication process

Fig. 8 Homogeneous
composite solution placed on
a hot plate for evaporation
process

Fig. 9 Nanocomposite films

which is formed after
evaporation (solvent casting
process)
Experimental Study and Analysis of Defragmented Carbon … 323

Fig. 10 Composite films

when taken out from petri
dish

the common solvent is known as solvent casting process explained in detail with an
example by the authors of [19, 20]. The concentration of the nanotubes was varied
with respect to the polymer concentration. The same process was repeated for all the
samples and was checked for its morphology using a scanning electron microscope.
SEM observations at low resolution were used to assess the degree of dispersion of
the defragmented carbon nanotubes in the polymer matrix.

3 Result and Discussion

The tubular structure of carbon nanotubes has been distorted into flake-like struc-
tures during probe sonication. According to references [1–4], carbon nanotubes have
excellent electrical conducting properties. Our discussion focused on using this prop-
erty of CNTs to enhance the electrical conductivity of insulators, namely polymers by
incorporating CNTs into them. Polymers have two major drawbacks which restrict
its usage. One being its mechanical strength and other its electrical and thermal
properties. By overcoming its drawbacks and with addition to its advantages (light
in weight, high corrosion resistant, etc.), it would have a huge impact in replacing
conventional materials in structural and electrical domains.
The samples were observed under a scanning electron microscope, and images of
the morphology were obtained. The following are the SEM images for a few samples
of the nanocomposite film. It is evident that the carbon nanotubes are completely
defragmented into carbon layers or strands of carbon. This carbon is represented by
the dark shaded area in the image. The lighter strands represent the polymer polyacry-
lonitrile strands which are spread throughout the matrix. Fig. 11 is the SEM image
of 4 %wt CNT (which is 40 mg of CNT in 960 mg of polymer). The magnifica-
tion was taken to 1 µm, and the analysis was conducted. While the defragmentation
process using the probe sonicator, it is noticed that complete defragmentation into
324 N. Arunkumar et al.

Fig. 11 Low-resolution
SEM images of composite
films prepared using the
probe sonicator

pure carbon will occur only during extensive use of the device. Figure 11, 12, and
13 indicate a lesser level of defragmentation of the CNTs which can be noticed by
the flake-like structure indicating the graphene distorted rings merged with polymer.
Hence, longer the sonication process, more the defragmentation of CNTs occurs.
The flake-like structure indicated in Fig. 12 has graphene layers fused with the poly-
mer which was the result of defragmenting the CNTs during solvent casting it with
the polymer. The magnification was taken at 500 nm. Partial flakes and dispersion
of PAN and defragmented CNTs are noticed in Fig. 13. The sample consisted of

Fig. 12 SEM image of

nanocomposite containing
5% wt CNT
Experimental Study and Analysis of Defragmented Carbon … 325

Fig. 13 SEM image of

nanocomposite containing
8 %wt CNT

8 %wt CNT and was magnified at 1 µm. The broad flake-like structure represents
the carbon/graphene structure merged with the polymer which is similar to Figs. 11
and 12 (Fig. 14 and Table 1).
Table 2 indicates the values which have been measured using a multimeter. This
multimeter has a very high sensitivity of measuring resistivity of sample. The resis-
tance and conductivity have been calculated using basic formula of electrical resis-
tance and resistivity formula. The graph shows the variation of various electric prop-
erties of the nanocomposite with the variation of carbon nanotube weight percentage.

Fig. 14 SEM image of

nanocomposite containing
9 %wt CNT
326 N. Arunkumar et al.

Table 1 Concentration of
S. No. Percentage of Weight of CNT Weight of
various samples
CNT (%) added (mg) polyacrylonitile
(mg)
1 2 20 980
2 3 30 970
3 4 40 960
4 5 50 950
5 6 60 940
6 7 70 930
7 8 80 920
8 9 90 910

Table 2 Electrical characteristics measured

S. No. CNT (wt%) Resistivity (m) Resistance () Conductivity (S/m)
1 2 210.17 2526.45 3.96 × 10−4
2 3 87.16 1045.43 9.57 × 10−4
3 4 33.26 399.12 2.50 × 10−3
4 5 8.66 103.92 9.62 × 10−3
5 6 3.75 45 0.022
6 7 1.51 18.12 0.055
7 8 0.57 6.84 0.146
8 9 0.16 1.92 0.52

The electrical resistivity of the material decreases gradually with increase in concen-
tration of CNT as shown in Fig. 15. Similarly, resistance decreases by a large extent
with increase in CNT as shown in the graph. Conductance increases drastically with
the increase in CNT as shown in the graph. Resistivity is measured in m (ohm

Fig. 15 Resistivity,
resistance, and conductivity
versus CNT concentration
Experimental Study and Analysis of Defragmented Carbon … 327

Fig. 16 Resistivity versus

CNT concentrations

meter). Weight of the CNT in the composite is measured in mg. Figure 16 shows that
with the increase in carbon nanotubes in the polymer matrix, the resistivity decreases.
This happens because the free electrons travel through a defined path along the carbon
nanotubes. Therefore, there is less opposition to the flow of electrons. Hence, there
is decrease in resistivity with more number of CNTs added in the polymer matrix.
Resistance is measured in ohms (). Concentration of the CNT is measured in mg.
Resistance is defined as the opposition to the flow of electric current. This opposi-
tion involves protons which oppose the flow of electric current by creating a barrier.
When CNTs are added to the polymer matrix, the potential barrier space is reduced
drastically and the concentration of current along the carbon nanotubes increases;
hence, the opposition is less. Therefore, there is a decrease in resistance with more
number of CNTs added in the polymer matrix as shown in Fig. 17. Conductivity is
measured in siemens/meter. The concentration of the CNTs is measured in mg. Con-
ductivity is defined as the flow of electrons through a material for a specified area of
cross section and length. The conductivity of the material increases drastically with
the increase in CNT concentration as depicted in Fig. 18. Hence, conductance also
increases. The flow of electrons is aligned along the carbon nanotube walls and acts
as a carrier of current which is present all through the polymer matrix.

Fig. 17 Resistance versus

CNT concentrations
328 N. Arunkumar et al.

Fig. 18 Conductivity versus

CNT concentrations

4 Conclusion

It has been experimentally proved that by inserting the carbon nanotubes into the
polymer matrix, the electric properties of the polymer are improved drastically. The
original resistance of the polymer was found to be in the range of 1016 . After
adding carbon nanotubes into the matrix, it was found to be in the range of 10 to 103
. This proves that a significant increase in the electrical properties of a polymer
can be improved by addition of carbon nanotubes. With further doping of carbon
nanotubes, the conductance can be compared with the present materials which are
used for electrical purposes such as copper and silver. The PAN nanocomposite is
a unique material which consists of good electrical properties, high tensile strength
which is inherited from the carbon nanotubes and also good corrosion resistant which
is inherited from the polymer polyacrylonitrile.

Acknowledgements The authors would like to thank:

Dr. Ramaprabhu S, Dept of physics, Alternative Energy and Nanoscience Laboratory, IITM,
for permitting to use the laboratory facilities and conducting the research.
Dr. Eswariah, Dept of physics, Alternative Energy and Nanoscience Laboratory, IITM, for his
guidance through the course of the research work.

References

1. Bhushan B (2010) Introduction of carbon nano tubes. In: Springer handbook of nanotechnology,
Part A, pp 47–118
2. Perie T, Brosse A-C, Tence-Girault S (2012) Mechanical and electrical properties of multi
walled carbon nanotubes/ABC block terpolymer composites. Carbon 50(8):2918–2928
3. Bai J, Huang Y (2010) Fabrication and electrical properties of graphene nanoribbons. Mater
Sci Eng: R: Rep 70(3–6):341–353
4. Shokrieh MM, Rafiee R (2010) Mechanics of composite materials. Chem Mater Sci 46(2):155–
172
5. Marinho B, Ghislandi M, Tkalya E, Koning CE, de With G (2012) Electrical conductivity
of compacts of graphene, multi-wall carbon nanotubes, carbon black, and graphite powder.
Powder Technol 221:351–358
Experimental Study and Analysis of Defragmented Carbon … 329

6. Inam F, Reece MJ, Peijs T (2012) Shortened carbon nanotubes and their influence on the
electrical properties of polymer nanocomposites. J Compos Mater 46(11):1313–1322
7. Zhang Y, Rutledge GC (2012) Electrical conductivity of electrospun polyaniline and
polyaniline-blend fibers and Mats. Macromolecules 45(10):4238–4246
8. Hou H, Reneker DH (2004) Carbon nanotubes on carbon nanofibers: a novel structure based
on electrospun polymer nanofibers. Adv Mater 16:69–73
9. Rzepka M, Bauer E, Reichenauer G, Schliermann T, Bernhardt B, Bohmhammel K (2005)
Hydrogen storage capacity of catalytically grown carbon nanofibers. J Phys Chem B
109:14979–14989
10. Greiner A, Wendorff JH (2007) Electrospinning: a fascinating method for the preparation of
ultrathin fibers. Angew Chem Int Ed 46:5670–5673
11. Aravind SS, Eswaraiah V, Ramaprabhu S (2011) Facile synthesis of one dimensional graphene
wrapped carbon nanotube composites by chemical vapor deposition. J Mater Chem 21:15179
12. Jeong JM, Choi MY, Ahn YT (2006) Morphology and properties of polyacrylonitrile
nanocomposite prepared viain-situ polymeriszation with macroazoinitiator. Macromol Res
14(3):312–317
13. Lu C, Liu J (2006) controlling the diameter of carbon nanotubes in chemical vapor deposition
Method by carbon feeding. J Phys Chem 110(41):20254–20257
14. Cantoro M, Hofmann S, Pisana S, Scardaci V, Parvez A, Ducati C, Ferrari AC, Blackburn
AM, Wang K-Y, Robertson J (2006) Catalytic chemical vapor deposition of single-wall carbon
Nanotubes at lowtemperatures. Nano Letters 6(6):1107–1112
15. Marcano DC, Kosynkin DV, Berlin JM, Sinitskii A, Sun Z, Slesarev A, Alemany LB, Lu W,
Tour JM (2010) Improved synthesis of graphene oxide. ACS Nano 4(8):4806–4814
16. Arai T, Yamazaki H, Tsujimoto T (2006) Solvent casting process. US Patent No. 7,132,074
17. Eswaraiah V, Sankaranarayanan V, Ramaprabhu S (2011) Inorganic nanotubes reinforced
polyvinylidene fluoride composites as low-cost electromagnetic interference shielding materi-
als. Nanoscale Res Lett 6(1):137
18. Kim JH, Ganapathya HS, Hong S-S, Gal Y-S, Lim KT (2008) Preparation of polyacrylonitrile
nanofibers as a precursor of carbon nanofibers by supercritical fluid process. J Supercrit Fluids
47:103–107
19. Safadi B, Andrews R, Grulke EA (2002) Multiwalled carbon nanotube polymer composites:
synthesis and characterization of thin films. J Appl Polym Sci 84(14):2660–2669
20. Moniruzzaman M, Winey KI (2006) Polymer nanocomposites containing carbon nanotubes.
Macromolecules 39(16):5194–5205
Design and Fabrication of Car Door
Panel Using Natural Fiber-Reinforced
Polymer Composites

Sodisetty V. N. B. Prasad, G. Akhil Kumar, P. Yaswanth Sai,

and Shaik Varees Basha

Abstract For a lot of years, extensive research was carried on the use of natu-
ral fibers in automotive sector to meet the environmental regulations as well as to
increase the profit margins of company. Natural fibers being of low cost, recyclable
and lightweight provide industry with the best alternative for traditional materials
such as steel, aluminum and synthetic fiber. Applications such as car door panels,
sun visors, seat covers and rearview mirrors have provided excellent scope for usage
of natural fiber-reinforced polymer composites. Various car manufacturers have suc-
cessfully implemented natural fiber-reinforced polymers in some of their car models
in recent years and have yielded desirable results in terms of lightness, cost saving
and reduction of environmental pollution. Various factors such as combinations of
fibers, layer sequence, fiber-to-matrix ratio and manufacturing method influence the
properties of the composite specimens. The natural fibers used are flax, hemp and
kenaf. Epoxy is selected as a matrix material. Hand layup method was used to fab-
ricate the specimens as well the car door panel. The mechanical testing was carried
out to determine the tensile strength, flexural strength, hardness and impact strength
according to ASTM standards. The composite with most desirable mechanical prop-
erties is chosen to fabricate the final car door panel. This paper also discusses about
the optimization of influencing parameters for drilling the natural fiber-reinforced
epoxy plate using Taguchi method. Spindle speed, feed rate and drill bit diameter
are selected as input parameters. These parameters which influence the thrust force
are measured by using drill tool dynamometer.

Keywords Hand layup molding · Door panel · Mechanical testing · Taguchi

method

S. V. N. B. Prasad (B) · G. Akhil Kumar · P. Yaswanth Sai · S. V. Basha

SRM Institute of Science and Technology, Kattankulathur, Chennai, India
e-mail: prasadsodisetty.m@ktr.srmuniv.ac.in

© Springer Nature Singapore Pte Ltd. 2021 331

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_30
332 S. V. N. B. Prasad et al.

1 Introduction

Recently, the growing need of alternative materials in the field of aerospace, marines
and construction, etc., had led to in-depth research in natural fiber-reinforced polymer
composites [1, 2]. Natural fiber-reinforced polymer composites consist of polymer
matrix reinforced with natural fibers such as flax, kenaf, sisal, hemp and jute. These
fibers possess high specific strength and high modulus and are used in various appli-
cations [3]. Natural fibers offer advantages over synthetic fibers such as high specific
strength, renewable, recyclable, less damage to processing equipment and easily
available. Srinivas et al. [4] investigated the properties of various natural fiber com-
posites such as jute, sisal, banana, bamboo, flax, kenaf, coir, hemp, curaua, cotton,
abaca and Ramie and compared it with synthetic fibers such as glass, carbon and
Kevlar. Hemp, flax, kenaf, abaca, etc., were found to have comparable properties
of tensile strength and Young’s modulus to those of synthetic fibers. Natural fibers
were seen as potential replacements due to their lower cost and recyclable in nature.
Various studies have been carried out on comparison of using hand layup method
and vacuum infusion method for preparation of composites [5]. In the present study,
hand layup method is selected based on various factors such as flexibility of design,
low volume of production and good surface finish. It is also an easier method for
production of heavy components.
Mechanical testing of natural fiber-reinforced polymer composites is quite essen-
tial to determine its tensile strength, flexural strength, impact strength, hardness and
water absorption property. Numerous studies have been carried out on mechanical
testing on natural fiber-reinforced polymer composites [6–12]. Mechanical proper-
ties of natural fiber-reinforced polymer composites depend on various factors such
as type of fiber, fiber-to-matrix ratio, stacking sequence of fiber, fiber orientation,
manufacturing method and number of layers. Elanchezhian et al. [13] evaluated the
mechanical properties of kenaf, name and banana hybrid composites. The ratio of
resin to hardener was found in the ratio of 10:1 to provide adequate interfacial bond-
ing between fibers. It was found that kenaf fiber possessed the highest tensile strength
and hardness among all the combinations and can be used for automotive compo-
nents using car hood and mudguard. The ratio of hardener to resin should be 1:10
to provide suitable interfacial strength, and the critical ratio of fiber to matrix was
found to be 40:60 for best mechanical properties. Girisha et al. [14] studied the prop-
erties of sisal, coconut fiber and ridge guard fiber. The mold was prepared by using
hand layup method. The specimens of the three fibers with epoxy are made with
varying the ratio of fiber to matrix. Increase in strength was observed with increase
in fiber volume fraction. However, natural fibers are hydrophilic in nature because
of cellulose content present in them. Hydrophilic fibers offer insufficient adhesion
with hydrophobic polymer matrices. This may lead to improper adhesion and bond
breakage between matrix and fibers. Various types of chemical treatment methods
such as alkali treatment and saline treatments are available to remove the moisture
from the natural fibers [15]. Alkali treatments such as treatment with 10% NaOH
for 24 h improve the fiber–matrix adhesion due to removal of moisture. This along
Design and Fabrication of Car Door Panel Using Natural … 333

with increased surface roughness, increased aspect ratio and reduction in moisture
absorption properties of fibers results in increased mechanical properties.
The applications of NFRCs in automobile industry include door panels, floor
panels, sun visor, door cladding, mudguards and in aerospace applications such as
aircrafts nose, doors, struts, outboard and indoor flaps and stabilizers [16]. Holbery
et al. [17] studied the process of making thermoplastic polypropylene with natural
fibers like compounding process, extrusion, kneading, injection molding, compres-
sion molding and sheet molding compound. The applications for these composites in
automotive sector include door cladding, seatback lining, cushions, floor panel, floor
body panels, head restraints and sound proofing. The improvement of these compos-
ites from plastic composites is increased stability, enhanced mechanical properties,
excellent electrical properties, good impact strength. dos Santos et al. [18] postu-
lated that the traditional materials used in automotive application can be replaced by
natural reinforced composites. Experimentation was carried on modification of nat-
ural fiber to improve their properties like stretching, calendaring, electric discharge
method, alkaline treatment and copolymerization. The fiber used for this study is
curaua. The application of this natural fiber is sun visor, sun visor hinge pin and
car interiors. The major problems which occur in drilling of natural fibers are fiber
pullouts from matrix and delamination of composite. So, various parameters have to
be optimized during drilling of natural fibers for higher surface finish and low thrust
force [19]. Various experimental analyses were carried out to determine the torque
and thrust force during drilling operation of NFRCs [20].

2 Methodology

Flax, hemp and kenaf fiber mats were selected as reinforcements for epoxy (LY556)
matrix. Hy951 Hardener was used. Hand Layup was used for fabricating the compos-
ite specimens. Totally six different composite plates with layers each were fabricated.
All the natural fibers were chemically treated with 10% NaOH solution to remove the
moisture content present in the fibers. Flax fiber-reinforced epoxy composite plate
with layers of flax fiber is fabricated with size 350 * 350 mm as shown in Fig. 1a.
Similarly, with the same size, kenaf and hemp are fabricated as shown in Fig. 1b, c,
respectively. Three hybrid plates are fabricated using three different combinations of
fiber mats such as kenaf + hemp, flax + hemp and flax + kenaf as shown in Figs. 1d,
e, f, respectively.
334 S. V. N. B. Prasad et al.

a b c

d e f

Fig. 1 a Flax [F6 ], b Kenaf [K6 ], c Hemp [H6 ], d Kenaf + hemp [(K/H)3 ], e Flax + hemp [(F/H)3 ],
f Flax + kenaf [(F/K)3 ]

3 Experimental Testing

3.1 Tensile Test

Tensile test is carried out on all the six specimens as per ASTM D-3039 standard. In
this test, the specimen is subjected to a constant tensile force until it fails. The stress–
strain plot for various combinations of composite plates is shown in Fig. 2. Properties
that can be determined by a tensile test are ultimate tensile strength, breaking strength,
maximum elongation and area reduction.

Fig. 2 Stress versus strain

graph
Design and Fabrication of Car Door Panel Using Natural … 335

Fig. 3 Flexural testing graph

3.2 Flexural Test

Flexural test is carried out as per ASTM D-790 standard. The three-point flexural or
bending test is used to determine the modulus of elasticity, flexural stress and flexural
strain. The main advantage of a three-point bending test is the ease of the specimen
preparation and its testing. The main disadvantage of this method is that the results
are sensitive to specimen, the loading geometry and the strain rate of the specimen
under the load. The load versus displacement graph is shown in Fig. 3.

3.3 Izod Impact Test

Impact test is carried out as per ASTM D-256 standard that was used for testing. The
energy absorbed by the specimens is calculated from the height from which the arm
swings to after hitting the specimen which is shown in Fig. 4.

3.4 Hardness Test

ASTM D-2583 standard was followed to carry out this testing. The Barcol hard-
ness test studies the indentation hardness of materials by measuring the depth of
penetration made by an indenter, loaded on a material sample and compared to the
penetration in a reference material. The values of hardness are shown in Fig. 5.
336 S. V. N. B. Prasad et al.

Fig. 4 Impact testing graph

Fig. 5 Hardness testing graph

Mechanical testing was carried out on all the specimens; the results are shown
in Table 1. The composite made up of all six flax fibers showed better mechanical
properties over others. So, this combination of six flax fibers is chosen for fabrica-
tion of car door panel. It is observed that individual fibers possess higher mechanical
properties than the hybrid combinations. This unusual trend is attributed to the pres-
ence of a large number of voids on the surface of hemp and kenaf fiber, which leads
to poor interfacial bonding and hence poor mechanical properties among the hybrid
fibers.

3.5 Comparison of Flax Fiber with ABS Plastic

Flax fiber-reinforced epoxy composite is selected among the five combinations. The
properties of ABS plastic and the flax fiber-reinforced composite plate are compared
as shown in Table 2.
Design and Fabrication of Car Door Panel Using Natural … 337

Table 1 Mechanical testing results

Fiber Tensile strength Flexural Izod impact Hardness
combination (MPa) strength (MPa) strength (kJ/m2 ) (Barcol
(ASTM (ASTM D-790) (ASTM D-256) hardness)
D-3039) (ASTM
D-2583)
Kenaf [K6 ] 22.3 36.7 5.86 36.1
Flax [F6 ] 28.0 62.7 9.0 45.1
Kenaf + hemp 17.0 49.6 6.6 34.0
[(K/H)3 ]
Kenaf + flax 21.7 46.4 3.8 37.0
[(F/H)3 ]
Hemp [H6 ] 21.0 60.6 8.7 29.0
Hemp + Flax 15.0 66.2 5.6 35.1
[(H/F)3 ]

Table 2 Comparison between ABS plastic and flax

Properties Flax fiber composite ABS plastic
Flexural strength (MPa) 62.7 51.986470
Tensile Strength (MPa) 28.0 46
Izod impact strength (kJ/m2 ) 9.0 9.01
Hardness (Barcol hardness) 45.1 97
Thickness of specimen (mm) 5 5

On comparing flexural strength of the flax fiber and ABS plastic, flax has more
flexural strength. But impact strength is almost same for both flax and ABS plastic.
So, flax can be used as alternative material for ABS plastic while fabricating the car
door panel.

4 Machining

CNC machining center is used for drilling the natural fiber-reinforced epoxy com-
posites using HSS twist drill tool. Drill tool dynamometer is connected to CNC
machining center which is used to measure the thrust force as shown in Fig. 6.
Thrust force values are measured for flax specimen. L9 orthogonal array is used in
Taguchi analysis. Three influencing parameters are chosen in three levels (3 * 3). The
three influential parameters are drill bit size, spindle speed and feed rate. The main
objective function is to reduce the thrust force. So smaller is better is chosen while
undergoing Taguchi analysis using Minitab software. Residual plots for thrust force
338 S. V. N. B. Prasad et al.

Fig. 6 Experimental Setup

for drilling

are obtained after Taguchi as shown in Fig. 8 by selecting four in one. All the param-
eters are taken with three levels as shown in Table 3. Thrust force is the response
variable recorded for all nine runs. General linear model was performed in analysis
of variance using Minitab software to obtain the % of contribution of parameters on
thrust force. Spindle speed plays a major role in varying the thrust force from the %
of contribution in Table 5. Spindle speed, drill bit size and feed rate are contributing
13.33, 53.08 and 30.86, respectively. The signal-to-noise ratio shows the optimized
parameters from the given parameters. The highest signal-to-noise ratio value will
be the optimized value. At 5 mm drill bit size, 50 mm/min feed rate and 3000 rpm
spindle speed are shown in Fig. 7. Response table for means is obtained as shown in
Table 4.

4.1 Regression Equation

The following regression equation is obtained when analysis of variance is performed.

Thrust force (N) = 63.31 − 18.52 Drill bit diameter (mm)_ 5

+ 11.73 Drill bit diameter (mm)_ 6 + 6.79 Drill bit diameter (mm)_ 7
− 12.64 Feed rate (mm/min)_ 5(1) + 0.55 Feed rate (mm/min)_ 100
+ 12.09 Feed rate (mm/min)_ 150 + 9.21 Spindle speed (rpm)_ 1000
− 3.00 Spindle speed (rpm)_ 2000 − 6.21 Spindle speed (rpm)_ 3000
(1)
Design and Fabrication of Car Door Panel Using Natural … 339

Main Effects Plot for SN ratios

Data Means
Drill bit diameter (mm) Feed rate (mm/min) Spindle speed (rpm)
-33

-34
Mean of SN ratios

-35

-36

-37

-38
5 6 7 50 100 150 1000 2000 3000
Signal-to-noise: Smaller is better

Fig. 7 Response graph

Residual Plots for Thrust force (N)

Normal Probability Plot Versus Fits
99
4
90
Residual
Percent

2
50
0
10
-2
1
-8 -4 0 4 8 40 60 80
Residual Fitted Value

Histogram Versus Order

6.0 4
Frequency

4.5
2
Residual

3.0
0
1.5
-2
0.0
-2 -1 0 1 2 3 4 1 2 3 4 5 6 7 8 9
Residual Observation Order

Fig. 8 Residual plots for thrust force

340 S. V. N. B. Prasad et al.

Table 3 Experimental design with orthogonal array

S. No. Drill bit Feed rate Speed (rpm) Thrust force SNRA1 MEAN1
diameter (mm/min) (N)
(mm)
1 5 50 1000 45.59 −33.1774 45.59
2 5 100 3000 36.69 −31.2910 36.69
3 5 150 2000 52.10 −34.3368 52.10
4 6 50 2000 56.95 −35.1099 56.95
5 6 100 1000 83.02 −38.3837 83.02
6 6 150 3000 85.15 −38.6037 85.15
7 7 50 3000 49.48 −33.8886 49.48
8 7 100 2000 71.89 −37.1334 71.89
9 7 150 1000 88.95 −38.9829 88.95

Table 4 Response table for means

Level Drill bit diameter (mm) Feed rate (mm/min) Speed (rpm)
1 44.79 50.67 72.52
2 75.04 63.87 60.31
3 70.11 75.40 57.11
Delta 30.25 24.73 15.41
Rank 1 2 3

Table 5 Analysis of variance

Source DF Seq SS Adj SS Adj MS F-value P-value %
Contribution
Drill bit 2 1579.96 1579.96 789.98 19.46 0.049 53.08
diameter (mm)
Feed rate 2 918.49 918.49 459.24 11.31 0.081 30.86
(mm/min)
Speed (rpm) 2 396.86 396.86 198.43 4.89 0.170 13.33
Error 2 81.19 81.19 40.60 2.73
Total 8 5726.01 100.00

4.2 Confirmability Test

Conformability test is performed for the optimized parameters of 5 mm drill bit size,
50 mm/min feed rate and 3000 rpm spindle speed. Least thrust force of 33.82 N is
obtained.
Design and Fabrication of Car Door Panel Using Natural … 341

5 Fabrication of Mold and Door Panel

Original door panel was purchased from the automobile spare parts vendor. To fab-
ricate the mold using hand layup, firstly, applied wax on the specimen to clean the
surface and then applied the PVA on the specimen for easy removal of mold after
curing. Resin, catalyst and chalk powder are mixed and applied over the PVA. The
glass fiber mat is placed, and resin mixture is transferred on to it. Repeated this pro-
cedure for three layers and cured it for 48 h. Mold is removed from the original door
panel. The final mold obtained from hand layup is as shown in Fig. 9. Flax fiber is
used to fabricate the final door panel. The same technique was employed to fabricate
the door panel using hand layup method as shown in Fig. 10.

Fig. 9 Mold made of glass

fiber

Fig. 10 Door panel

fabricated by flax
342 S. V. N. B. Prasad et al.

6 Conclusion

Flax fiber-reinforced epoxy composite possesses higher values of tensile strength,

impact strength and hardness among various combinations tested. The flexural
strength of hemp and flax combination is higher than the individual flax fiber. Hence,
flax fiber is chosen to fabricate the door panel. It is observed that individual fibers
possess higher mechanical properties than the hybrid combinations. This unusual
trend is attributed to the presence of a large number of voids on the surface of hemp
and kenaf fiber, which leads to poor interfacial bonding and poor mechanical prop-
erties among the hybrid fibers. For low volume of production, hand layup method is
most suitable. Natural fibers can be extensively used in automobile applications as
alternatives for synthetic fibers and ABS plastic due to their comparable mechanical
properties, low cost and recyclability. The optimized parameters to obtain less thrust
force to drill flax fiber-reinforced epoxy specimen are drill bit of size 5 mm, feed
rate of 50 mm/min and spindle speed of 3000 rpm. Spindle speed plays a major role
in varying thrust force.

References

1. Haydar Zaman U, Mubarak Khan A, Ruhul Khan A, Arifur Rahman M, Lily Das R, Al-Mamun
MD (2010) Role of potassium permanganate and urea on the Improvement of the mechanical
properties of jute polypropylene composites. Fibers Polym 11(3):455–463
2. Ratna Prasad AV, Mohana Rao K, Gupta AVSSKS, Reddy BV (2011) A study on flexural
Properties of Wild cane grass Fiber-reinforced polyester composites. J Mater Sci 46:2627–2634
3. Yang Yuqiu, Ota Tomoko, Morii T, Hamada H (2011) Mechanical property and hydrothermal
aging of injection molded jute/polypropylene composites. J Mater Sci 46:2678–2684
4. Srinivas K, Lakshumu Naidu A, Raju Bahubalendruni MVA (2017) A review on chemical
and Mechanical properties of natural fiber reinforced polymer composites. Int J Perform Eng
13(2):189–200
5. MohdYuhazri Y, Phongsakorn PT, Sihombing H (2010) A Comparison process between vacuum
infusion and Hand lay-up method toward kenaf/polyester composites. Int J Basic Appl Sci
IJBAS-IJENS 10(3)
6. Jamrichova Z, Akova E (2013) Mechnical testing of natural fiber composites for automotive
industry. Univ Rev 7(3):20–25
7. Dan-mallam Y, Abdullah MZ, MegatYusoff PSM (2012) Predicting the tensile properties of
woven kenaf/polyethylene terephthalate (PET) fiber reinforced polyoymethylene (POM) hybrid
laminate composite. J Mech Civil Eng (IOSRJMCE)
8. Muthukumar V, Venkatasamy R, Sureshbabu A, Arunkumar D (2011) A study on mechanical
properties of natural fiber reinforced Laminates of Epoxy (Ly 556) polymer matrix composites.
Int J Prod Technol Manag Res 2(2)
9. Barile C, Casavola C, De Cillis F (2019) Mechanical comparision of new composite materials
for aerospace applications. Elsevier J Compos Part B 162:122–128
10. Alagarraja K, Dhamodharan A, Gopinathan K Mathan Raj R, Ram Kumar K (2014) Fabrication
and testing of fiber reinforced polymer composites material. IOSR J Mech Civil Eng (IOSR-
JMCE)
11. Arun Kumar, Abishek T, Jegatheesh Raja S, Gokul Krishnan A (2016) A review on mechanical
properties of natural fiber Reinforced hybrid polymer composites. Int J Adv Sci Tech Res 6
Design and Fabrication of Car Door Panel Using Natural … 343

12. Rihayat T, Suryani S, Fauzi T, Agusnar H, Wirjosentono B, Alam PN, Sami M (2017) Mechan-
ical properties evaluation of single and hybrid composites polyester reinforced bamboo, PALF
and coir fiber. IOP Conf Ser Mater Sci Eng 334(2018):012081
13. Elanchezhian C, VijayaRamnath B, Kaosik R, Nellaiappan TK, Santhosh Kumar K, Kavirajan
P, Sughan MU (2017) Evaluation of mechanical properties of kenaf based hybrid composite
for automotive components replacement. ARPN J Eng Appl Sci 10(13)
14. Girisha C, Sanjeevamurthy, Rangasrinivas G (2012) Tensile properties of natural fiber
reinforced epoxy hybrid composites. Int J Mod Eng Res (IJMER)
15. Suardana NPG, Piao Y, Lim JK (2011) Mechanical properties of hemp fibers and hemp/PP
composites: effects of chemical surface treatment. Mater Phys Mech 11:1–8
16. Leao AL, Rowell R, Tavares N (1998) Application of natural fibers in automotive industry. Sci
Technol Polym Adv Mater
17. Holbery J, Houston D (2006) Natural-fiber-reinforced polymer composites in automotive
applications. J Mater
18. dos Santos PA, Giriolli JC, Amarasekera J, Moraes G, Natural fibers plastic composites for
automotive applications. SABIC Innovative Plastics
19. Palani Kumar K (2011) Experimental investigation and optimization in drilling of GFRP
composites. Measurement 44:2138–2148
20. Chandramohan D, Marimuthu K (2011) Drilling of natural fiber particle reinforced polymer
composite material. Int J Adv Eng Res Stud 1(1):134–145
Investigation of Physical–Chemical
Properties and Evaluation of Optimal
Blend Ratio of Rice Bran Biodiesel:
A Mathematical Regression Analysis
Approach

M. Selvam , S. Palani, and P. Vaishnavi

Abstract Our research work mainly focusses on the examination of properties,

both physical and chemical, of biodiesel blends among neat diesel at different volu-
metric proportions. Properties like viscosity; calorific value (CV); pour point (PP);
cetane number; could point (CP); flash point and fire point are measured for each
blending ratios, and it is examined against the relevant standards (USA ASTM D
6751; Europe EN:14214 as well as India IS:15607). Plot between the properties
and different blending ratios is drawn, and regression equations (R2) are formulated
from which biodiesel properties are predicted and compared by mineral diesel. From
this investigation, the optimal blending ratio is projected considering the different
properties of biodiesel.

Keywords Rice bran oil · Cetane number · Flashpoint · Calorific value ·

Mathematical regression analysis

1 Introduction

The fossil fuel is a great power source, but it has many drawbacks. It is the main
contributor to the increase of global warming due to its emission which leads cli-
matic changes and environmental effects like acid rain, depletion of the ozone layer,
increase in greenhouse phenomena and formation of smog. This is a renewable source
of energy, but the formation of fossil fuel takes over a million of years to renew when
compared to other sources like solar and wind energy, so it is considered to be a
non-renewable form of energy. Biodiesel be able to be extracted from a variety of
raw materials similar to vegetable oil, animal fat, used cooking oil, etc. Biodiesel is
said to be a good alternative because it is biodegradable, less toxic in nature, superior

M. Selvam (B) · S. Palani

Anna University, Chennai, Tamil Nadu 600025, India
e-mail: mselvam@veltechmultitech.org
M. Selvam · P. Vaishnavi
Department of Mechanical Engineering, Vel Tech Multitech, Avadi, Chennai, Tamil Nadu
600062, India

© Springer Nature Singapore Pte Ltd. 2021 345

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_31
346 M. Selvam et al.

lubricity and the abundant availability of various raw materials around us. Biodiesel
enhances the efficiency of diesel engines because it has high oxygen content, and it
also has high carbon content which increases the combustion rate. Along with the
efficiency, the emission characteristics of biodiesel are found to be better than diesel.
It emits a relatively less amount of carbon monoxide, hydrocarbons, particulate mat-
ter than diesel. However, biodiesel tends to increase the emission of nitrogen oxide
(NOx ), and also, utilization of biodiesel (B100%) within diesel engine will result in
reduced power, torque and fuel economy. Many researchers have reported that vari-
ous blends of diesel and biofuel could increase the power, torque and fuel economy
of the vehicle.
In the recent environmental research scenario regarding biofuel, the investigation
rice bran oil is considered for production of biodiesel due to the vast availability, and
India produces of about 8 lakhs tonnes rice bran oil from rice husk (which contains
nearly 10–23% of oil). Rice bran oil is removed from rice husk with help of huller
and sheller mills. Such extracted rice bran oil consists of several compositions of
fatty acids like 1.297% stearic, 14.883% palmitic, 42.013% oleic, 38.561% linoleic,
1.396% linolenic, 0.122% myristic. Production of biodiesel starts with transesteri-
fication process which involves to get fatty acid methyl ester through rice bran oil.
Fatty acids from the oil are removed so as to minimize oil viscosity for meeting the
standard value of mineral diesel.
Lin et al. (2009) produced biodiesel using transesterification process and optimum
reaction conditions to analyze the fatty acid composition. The fuel properties obtained
are well suited with the standards of ASTM D6751-02 and DIN V51606 [1]. Dennis
et al. (2010) created biodiesel by means of catalyzed process of transesterification at
atmospheric temperature and pressure which avoids the necessity of high temperature
and pressure for animal fats, and waste cooking oils increase the molar ratio of
triglycerides to alcohol of 6:1 [2]. Atabani et al. (2013) suggested Sterculia foetida
oil biodiesel blend to increase the effectiveness of properties of maximum kinematic
viscosity of 75.826 mm2 /s at 40 °C [3]. Evangelos (2013) found that the cold flow
properties have higher impact on the biodiesel than that of conventional diesel [4].
Ashrafu et al. (2014) stated that the fatty acid composition plays a major role in
the fuel properties such as viscosity, calorific value and cetane number. They also
established that monosaturated fatty acid present in the biodiesel blend at minimum
temperature increases the ignition quality and also increases the fuel properties [5].
Subhash et al. (2015) indicated that the biodiesel blend of B15 does not lead to
the wall impingement, whereas the B20 has better boundary of wall impingement
(within uncertainty ±1.3%). [6]. Bhaskor et al. (2016) carried out the three diverse
compression ratios of 18, 17.5 and 17, at injection timing 23 °C at various loading
conditions. In maximum load, the most extreme brake thermal efficiencies (BTEs)
with respect to double fuel mode (DFM) are observed to be 20.27%, 19.97% and
18.39% for CRs of 18, 17.5 and 17 [7]. Mayank et al. (2016) found that the fatty
acid content in crude rice bran oil is elevated with respect to oil quality since of over
fatty content. The two-step transesterification process is followed which includes
acid-catalyzed transesterification and base-catalyzed transesterification [8]. Baharak
et al. (2016) found that the edible oil has various drawbacks like deforestation,
Investigation of Physical–Chemical Properties and Evaluation … 347

increased vegetable oil price, to overcome which extraction can be done with non-
edible oils(rice bran oil, rubber seed, mahua seed, silk cotton, babassu tree) [9].
Mohamed Shameer et al. [10] used the B20 blend because of its low emission and
increased engine performance. Also, they used advanced injection timing as well as
injection pressure for reducing emission [10].
Maria et al. (2017) found that rice bran is the by-product of rice which is under-
utilized and could be used in biodiesel production and also rich in protein. They
also stated that the rice bran oil is more captivating for its more nutraceutical char-
acteristics and balanced fatty acid content [11]. Ertan et al. (2017) stated that the
oxygenated fuel blends reduce NOx emission but increase CO2, CO [12]. Evangelos
et al. (2018) analyzed the properties of fuel with respect to the fatty acid composition
of the biodiesel. They also analyzed the properties such as the cetane number, den-
sity, kinematic viscosity with respect to the fatty acid content present in the biodiesel.
[13]. Ertan et al. (2019) stated that the fuel properties play a major role in the per-
formance, combustion and injection characteristics of the diesel engine. They also
stated that the combustion and performance of methyl and ethyl ester show better
values when compared with diesel at various engine conditions such as engine speed
and at different load conditions (BMEP 3.3, 5.0, 6.6, 8.3 bar) [14]. Jeyaprakash
et al. (2019) investigated rice bran biodiesel in a four-stroke, single-cylinder, direct-
injection engine which runs at 1500 rpm is made in which the biodiesel blend of
20%(B20) has greater values when compared with other blending ratio. Their results
show that the rice husk contains about 16–20% of crude rice bran oil, and it has high
in-cylinder pressure of about 72.1 bar and heat release rate of about 70.826 J/CA
[15].
In our research work, consider rice bran oil for production of biodiesel due to
the vast availability. Rice bran oil is removed from rice husk with help of huller and
sheller mills. Such extracted rice bran oil consists of several compositions of fatty
acids like stearic, palmitic, oleic, linoleic, linolenic, myristic. Production of biodiesel
starts with transesterification process involves to get fatty acid methyl ester (FAME)
from rice bran oil.

2 Extraction of Rice Bran Oil

Rice bran oil is edible oil which is used for cooking purpose. Rice husk is a main
source of rice bran oil. India produces about 80 million tonnes of paddy from which 5
million tons of rice bran and to the extent of 8 lakh tonnes of rice bran oil are obtained.
2 lakh tonnes is currently produced and about 10–12% percentage is of edible grade.
Around 5–10 kg of rice bran is produced from 100 kg of paddy. About 10–23%
of rice bran oil is produced from rice bran. Rice bran is a mixture of substances
which contains protein, fat, ash and crude fiber. 70% of the paddy is processed by
huller mills, and about 30% is produced by sheller mills. Rice bran oil production
comes under three production phases. The first and foremost is the primary phase
which includes needing, feeding, separating of fogging material (killing insects that
348 M. Selvam et al.

involves using fine pesticide spray). The separated material is fed into the cooker for
steam processing, and then, the cooked material is fed into the palletizer to make
material into pallets. Now, the pallets are allowed to cool down; then they are sent
to the conveyor for the extraction. The second phase is started by pouring pallets
into hexanoate in order to remove the hexane. Then, the final stage is the distillation
process; the product from the distillation process is fed into the missile tank for
storage, and now, the stored product is sent to the heater to evaporate the hexane in
order to produce crude rice bran oil. The crude rice bran oil consists of dissolved
gum which is removed by degumming process. Further, it is sent to bleaching section
to refine the crude oil; following the bleaching section, it is sent to the de-waxing
section where the wax is removed from the oil. At last, deodorizing process is carried
out at high vacuum and temperature to remove the fatty acids. Thus, the required
edible rice bran oil is obtained.

2.1 Production Methodology

The raw rice bran oil was extracted from conventional process. The oil was then
filtered to remove minuscule residues. The raw rice bran oil was transformed to
biodiesel by transesterification technique. Initially, the oil was heated to a temperature
of about 55–60 °C. By the side, methanol (10%) and potassium hydroxide (KOH)
were titrated and added to the oil. Then, the concoction is stirred thoroughly at
600 rpm for about an hour. The mixture is kept nearly for 24 h for the glycerol to
settle down. At the end of the process, glycerol is settled down, and after its removal,
fatty acid methyl ester is obtained. Warm water was added to the methyl ester and
kept for 24 h in order to take away suspended impurities. The above said process is
clearly shown in Fig. 1.
The properties like kinematic viscosity, density, flash point, fire point, cetane
number, cloud point, pour point were tested using standard apparatus and examined
for both raw rice bran oil and biodiesel.

3 Extraction of Rice Bran Oil

3.1 Acid Composition of Biodiesel and Crude Oil

Rice bran oil as well as rice bran biodiesel is composed of 17.892% and 18.849%
of saturated fats and 82.102% and 81.152% of unsaturated fats, respectively. The
level of free fatty acid of rice bran oil as well as rice bran biodiesel is evaluated by
means of gas chromatography (GC) test. Gas chromatography (GC) is a typical kind
of chromatography utilized for isolating and breaking down composites that can be
vaporized without disintegration. The free fatty acid compositions are described in
Investigation of Physical–Chemical Properties and Evaluation … 349

Fig. 1 Production methodology of raw rice bran oil

terms of C:D ratio, where C represents the number of carbon atoms and D represents
the number of double carbon atom chains. The succeeding acids have the major part
in rice bran oil and rice bran biodiesel. Rice bran oil and rice bran biodiesel consist
of 42.013% and 41.414% of oleic acid C18:1 which is also used as an emulsifying
agent as it withstands the structure of chemical compounds. The linoleic acid C18:2
has volumes of 38.236% and 38.032% in rice bran oil and biodiesel. The palmitic
acid C16:0 serves as release agent and also used to increase the volatility. It is present
about 14.883% and 15.897% in rice bran oil and biodiesel, respectively.

3.2 FTIR Analysis

The vibrational motion of the functional group is determined by the Fourier trans-
formation infrared spectroscopy of the diesel sample and biodiesel blends of B30 as
shown in Fig. 2 and B100 as shown in Fig. 3. In the infrared region, the light rays pass
through the interferometer and the beam splitter to generate absorption frequencies.
350 M. Selvam et al.

Fig. 2 FTIR analysis in the spectrum of B30

Fig. 3 FTIR analysis in the spectrum of B100

The peaks in the spectrum band are the absorbent of the vibration molecule and the
concentration of the mixture. MIR (Mid-range) valDB; ues for this apparatus are in
range of 400–4000 cm−1 and as a resolution of about 1 cm−1 . The highest peak is
found for blend of B30. The double bond stretching vibration related to C–H alkyl
methylene group which frame olefin and to extend fluctuation and reduce at methyl
group and the second peak clearly at 1745 cm−1 C=O carbonyl group it frames esters
the C–H group’s tendency in –CH3 and CH2 assignment and as same as the B100
Investigation of Physical–Chemical Properties and Evaluation … 351

Fig. 4 FTIR analysis in the spectrum of D100

sample as shown in Fig. 3. In between 3007 and 2926 cm−1 stretch combined with
2855 cm−1 and yield stretch has shown –CH2 and the C–H alkyl group is present at
another peak got at 1743 cm−1 carbonyl C=O aldehyde are identified both sample
bending has approximately at 723 cm−1 . The sample of diesel shown in Fig. 4 has
similar functional groups at the approximate region of wave number in the spectro-
metric image, and the appropriate values were given as well as the functional group
families and type of vibration motion.

3.3 CHN Elemental Analysis

The technique CHNS/O analysis is used to determine element weight percentage in

fuels for diesel, rice bran biodiesel and rice bran oil that is mainly described in carbon,
hydrogen, nitrogen, sulfur and oxygen, and the elemental analysis is synthesized by
the Eagar smart data handling software automatic calculation. The high carbon diesel
to RBD and RBO, than it shows the main element for high burn source oxygen RBO
having more than other fuels and its sulfur content has very less in RBO and RB
biodiesel than diesel which has 0.21% so that its concluded maximum Sulfur free
and high burn source capacity elements are abundantly available in the raw material
and proposed biodiesel.
352 M. Selvam et al.

3.4 Physical and Chemical Properties of Rice Bran Oil

and Biodiesel

Blending the biofuel with mineral diesel is mostly preferred in many countries
because the use of 100% of biodiesel in the engine may have some drawbacks like less
availability, poor cold flow properties and comparatively reduced calorific value. For
this purpose, a detailed investigation was done to analyze the deviation, and required
information is gained in the physical and chemical properties of diesel and biodiesel
blend. Rice bran oil is blended with mineral diesel at different blending ratios from
10 to 90% to determine the properties like kinematic viscosity, density, calorific
value, fire point, flash point, cloud point, pour point. The above said properties are
measured using the instruments compared with different bending ratios, and graphs
are generated to determine the deviation of properties for each blending ratio for the
same. The mathematical regression equations are determined from which values of
properties for different blending ratios are established. This considerably reduces the
time and cost.

3.5 Examination of Properties of Diesel and Biodiesel Blends

Graphs for different properties of rice bran biodiesel blends are established. Mathe-
matical regression equation is developed for these properties using analysis software
know as sigma plot software, which is mainly used to predict the values for the dif-
ferent properties of various blending ratios. Here, in the mathematical equation, ‘x’
represents the blending ratio which is defined as the mixture of mineral diesel and
biofuel.

3.5.1 Linear Regression

Equation:

Y = ax + b (1)

‘a’ coefficient

xi yi − nxi yi

A= (2)
xi2 − nxi2

‘b’ coefficient

xi xi yi − xi2 yi

b= (3)
xi2 − nxi2
Investigation of Physical–Chemical Properties and Evaluation … 353

Linear correlation coefficient

nxi yi − xi yi

r x y =       (4)
nxi − xi2 nyi2 − xi2
2

Coefficient of determination

R 2 = r x2y (5)

Standard error of the regression

 
1  yi − ȳi 
Ā =   × 100 (6)
n yi

4 Results and Discussion

In our research work, rice bran oil properties have been evaluated and compared
with various standards like American Standard (ASTM), European Standard (EN)
and Bharat Standard (BIS).

4.1 Density

Density can be defined as ratio of mass per volume which is mostly communicated
as kilogram per meter cubic. Density plays a major role in the engine output and
performance since the change in density has a direct influence on the mass of the fuel,
which has an effect on the fuel injection in the engine. Also, density has greater effect
on emission like particulate matter (PM) and NOx . When the density is increased by
10%, it leads to loss of combustion heat by 4%. Thus, in order to increase the heat
of combustion, the density of rice bran is reduced. The main reason to reduce the
density of biofuel is to reduce the viscosity which has influences on the fuel injection
characteristics. The density of mineral diesel is in the range of about 820 kg/m3 .
It is clearly detected that the density values increase in a linear manner in which
the density increases with the increase in blending ratio because of the increase in
the biofuel ratio in the biodiesel mixture. Thus, the density range for B30 is almost
nearer to the density value and is also in range mentioned in the EN-590 (reference
paper no). The density value reaches up to a range of 874 kg/m3 for the B90 blends.
These values are shown in Table 1.
 354

Table 1 Properties of rice bran biodiesel for different blending ratios

Properties Unit B0 B10 B20 B30 B40 B50 B60 B70 B80 B90 B100
Density kg/m3 820 826 831 838 844 851 856 862 868 874 878
Kinematic viscosity mm2 /s 2.4 2.6 2.8 3 3.3 3.5 3.7 3.9 4.2 4.4 4.75
Calorific value kj/kg 43,000 42,850 42,250 42,050 41,750 41,250 40,950 40,650 40,050 39,700 39,500
Flash point °C 55 68 79 89 100 112 122 135 145 156 165
Fire point °C 62 73 85 98 111 125 136 149 161 175 182
Cetane number 49 49.9 50.6 51.3 52.1 53 53.9 54.7 55.5 56.1 57
Cloud °C 1 2 4 5 6 8 9 10 12 13 14
M. Selvam et al.
Investigation of Physical–Chemical Properties and Evaluation … 355

Fig. 5 Density vs Biodiesel blending ratio and kinematic viscosity vs Biodiesel blending ratio

4.2 Kinematic Viscosity

Viscosity is the oil’s resistance to flow and shear. Thus, the kinematic viscosity was
determined using Redwood Viscometer. This kinematic viscosity has an important
influence in the fuel flow and fuel injection. Kinematic viscosity should be in the
range of 2–5 mm2 /s which will affect the fluidity of the fuel at low temperature.
When the kinematic viscosity increases, it will affect the combustion rate since the
atomization of the fuel will be delayed. Spray cone angle and spray penetration are
reduced with increase in viscosity and results in bigger droplets. High viscosity can
reduce fuel flow rates which results in inadequate fueling of engine. On the other
hand, low viscosity results in fuel leakage and loss of fuel system calibration. As in
the density, the viscosity value also has a linear increase in value with the increase
in the volumetric ratio of the rice bran biofuel with the mineral diesel as shown in
Fig. 5. In this, the viscosity range for raw biodiesel (B100) is about 30.26 mm2 /s
whereas the value for B30 blend is almost nearer at 3 mm2 /s which satisfies the
engine specification and also helps to increase the combustion rate.

4.3 Calorific Value

The amount of heat liberated during the combustion at steady pressure and tempera-
ture is known as calorific value. It is one of the major factors that has an effect in the
combustion rate and also affects the output power of the engine. The calorific value
is measured using the bomb calorimeter, which is measured by the heat produced by
combustion of a little quantity of oil. This is usually expressed in kilojoules’ per kilo-
gram. Figure 6 shows the variation of calorific value with the various blending ratios
of rice bran biodiesel. In contrast to density and kinematic viscosity, the calorific
value shows a decrease in value with increasing blending ratio. The calorific value
for raw rice bran oil is almost around 36,283 kJ/kg which is less when compared to
356 M. Selvam et al.

Fig. 6 Calorific value vs Biodiesel blending ratio and Flash point vs Biodiesel blending ratio

mineral diesel. But the biodiesel blends of B20 and B30 have a calorific value nearer
to diesel.

4.4 Flash Point

The minimal temperature at which the fuel starts to vaporize and forms as an ignitable
mixture in the presence of atmospheric air around the fuel surface is termed as flash
point. Ignition source temperature does not have effect on flash and fire point, but
its value is far higher than that of flash point and fire point. However, lower the flash
point easier to ignite the fuel mixture. The flash point of diesel is 55, and flash point
of biofuel is 165. In order to reduce the flash point value, the biofuel is blended
with mineral diesel. The fuel is termed as flammable if the flash point value is less
than 37.8 °C, and if value is greater than 37.8 °C, then the fuel is called combustible.
Thus, crude rice bran oil and biodiesel come under combustible fuel. Figure 7 clearly
shows that value of flash point gradually increases from B0 to B100.

Fig. 7 Fire point vs Biodiesel blending ratio and Cetane number vs Biodiesel blending ratio
Investigation of Physical–Chemical Properties and Evaluation … 357

Fig. 8 Cloud point vs Biodiesel blending ratio and Pour point vs Biodiesel blending ratio

4.5 Fire Value

Fire point can be defined as the minimal temperature at which the fuel vapor starts
to burn for a specified time after it is ignited. In general, the fire point will be 10 °C
more than the flash point. In Fig. 8, same as that of the density and flash point, the
fire point also has a linear increase in values for increasing values of blending ratio.
It is observed that the raw rice bran oil has a fire point of about 242 °C. The blending
ratio B30 and B20 has optimal fire point when compared to the mineral diesel.

4.6 Cetane Number

Cetane number is a marker or indicator which gives the info about the ignition
delay and also indicates the ignition quality of the fuel. For the increasing value of
cetane number, the ignition delay period reduces subsequently the rate of combustion
increases and vice versa. As per the ASTM standards, the cetane number should have
a minimal value of 40 to a maximum value of 60. If the value of cetane number falls
below 40, it leads to smoke emission and fails to obey the emission laws. The cetane
number value of biodiesel lies in the acceptable range, and it also enhances the
combustion rate and reduces the ignition delay period alternatively. In the same way
as before, it also shows a linear increase in value with increase in volumetric ratio of
rice bran biofuel with the mineral diesel.

4.7 Cloud Point, Pour Point

Cloud point can be referred to as the temperature at which the wax crystal is created
in the fuel which has a direct impact at the low condition application. Pour point is
also an important property for the low temperature application. It is defined as the
358 M. Selvam et al.

temperature at which the fuel losses its fluidity and flow characteristics. The cloud
point also shows constant increase with increase in blending ratio where the cloud
point of mineral diesel is around 1 °C. It is recommended that the cloud point and
the pour should be low as possible. In the same way, as that of the cloud point, the
pour point also shows an increase in readings. The pour point for the raw rice bran
biodiesel is around +5 °C whereas for mineral diesel is about −9 °C.

5 Conclusion

In this investigation, the physical and chemical properties of rice bran oil are evaluated
and compared with the mineral diesel as per ASTM-D-6751, EN-14214 and IS-
15607 standards. From this investigation, it is clearly known that different properties
of rice bran biodiesel are nearer to that of mineral diesel at lower blending ratios,
and the deviation increases as the blending ratio increases. Mathematical regression
calculation is used to determine the properties for different blending ratios. Sigma
plot software is used to draw the graph using the values determined by regression
values. From the graph, it is evaluated that B30 is optimum blending ratio. At this
blending ratio, it is found that viscosity of oil is higher than that of mineral diesel,
and other properties like density, cetane number, calorific value, flash point and fire
point are closer to that of mineral diesel. Cold flow properties like cloud point and
pour point do not show major deviation to the value of mineral diesel.

References

1. Lin L, Ying D, Chaitep S, Vittayapadung S (2009) Biodiesel production from crude rice bran
oil and properties as fuel. Appl Energy 86:681–688
2. Dennis Leung YC, Xuan W, Leung MKH (2010) A review on biodiesel production using
catalysed transesterification. Appl Energy 87:1083–1095
3. Atabani AE, Mahlia TMI, Masjuki HH, Badruddin IA, Yussof HW, Chong WZ, Lee KT (2013)
A comparative evaluation of physical and chemical properties of biodiesel synthesized from
edible and non-edible oils and study on the effect of biodiesel blending. Energy 58:296–304
4. Evangelos Giakoumis G (2013) A statistical investigation of biodiesel physical and chemical
properties and their correlation with the degree of unsaturation. Renew Energy 50:858–878
5. Ashraful AM, Masjuki HH, Kalam MA, Rizwanul Fattah IM, Imtenan S, Shahir SA, Mobarak
HM (2014) Production and comparison of fuel properties engine performance emission charac-
teristics of biodiesel from various non-edible vegetable oils A review. Energy Convers Manag
80:202–228
6. Lahane S, Subramanian KA (2015) Effect of different percentages of biodiesel–diesel blends
on injection spray combustion performance and emission characteristics of a diesel engine.
Fuel 139:537–545
7. Bhaskor Bora J, UjjwalSaha K (2016) Experimental evaluation of a rice bran biodiesel e biogas
run dual fuel diesel engine at varying compression ratios. Renew Energy 87:782–790
8. Chhabra M, Sharma A, Dwivedi G (2016) Performance evaluation of diesel engine using rice
bran biodiesel. Egypt J Petrol 26(2):511–518
Investigation of Physical–Chemical Properties and Evaluation … 359

9. Sajjadi B, Aziz A, Raman A, Arandiyan H (2016) A comprehensive review on properties of

edible and non-edible vegetable oil-based biodiesel Composition specifications and prediction
models. Renew Sustain Energy Rev 63:62–92
10. Mohamed Shameer P, Ramesh K, Sakthivel R, Purnachandran R (2017) Effects of fuel injection
parameters on emission characteristics of diesel engines operating on various biodiesel: a
review. Renew Sustain Energy Rev 67:1267–1281
11. Maria Capellini C, Giacomini V, Maitê Cuevas S, Christianne Rodrigues EC (2017) Rice bran
oil extraction using alcoholic solvents Physicochemical characterization of oil and protein
fraction functionality. Ind Crops Prod 104:133–143
12. Alptekin E (2017) Emission injection and combustion characteristics of biodiesel and
oxygenated fuel blends in a common rail diesel engine. Energy 119:44–52
13. EvangelosGiakoumis G, Christos Sarakatsanis K (2018) Estimation of biodiesel cetane number
density kinematic viscosity and heating values from its fatty acid weight composition. Fuel
222:574–585
14. Alptekina LE, Sanlib H, Canakcia M (2019) Combustion and performance evaluation of a
common rail DI diesel engine fueled with ethyl and methyl esters. Appl Thermal Eng 149:180–
191
15. Dharmaraja J, Nguyen DD, Shobana S, Saratale GD, Arvindnarayan S, Atabani AE, Chang
SW, Kumar G (2019) Engine performance, emission and bio characteristics of rice bran oil
derived biodiesel blends. Fuel 239:153–161
Wear and Friction Behaviours
of Stainless Steel (SS 316) Wire Mesh
and Carbon Fibre Reinforced Polymer
Composite

A. H. Ansari, V. Jayakumar, and S. Madhu

Abstract There is a growing demand for the development of new polymer matrix
composite materials that exhibits high strength, high wear resistance, good rigidity
and less weight. In this work, an attempt has been made to develop a novel car-
bon fibre reinforcement polymer (CFRP) composite with stainless steel (SS 316)
wire mesh and to investigate its tribological characteristics with the objective of
improving the wear resistance. The pin-on-disc experiment was performed on the
fabricated CFRP-SS316 fibre metal laminate by using EN31 steel (high carbon alloy)
pins. Wear rate and coefficient of friction have been examined by conducting four
experiments on CFRP-SS316 laminate. The process parameters used for this inves-
tigation includes applied load, sliding velocity, sliding distance and speed. Experi-
mental results showed that the wear rate increases on a CFRP-SS316 disc laminate
when the load increases gradually from 10 N to 40 N. The minimum wear rate of
16.34 mm3 /N was obtained for 10 N applied load.

Keywords Carbon fibre reinforced polymer · Stainless steel 316 wire mesh · Wear
rate · Coefficient of friction

A. H. Ansari
Department of Mechanical Engineering, Saveetha School of Engineering, Saveetha Institute of
Medical and Technical Sciences, Chennai, Tamil Nadu, India
V. Jayakumar (B)
Department of Mechanical Engineering, Amrita School of Engineering, Chennai, Amrita Vishwa
Vidyapeetham, Tamil Nadu, India
e-mail: v_jayakumar@ch.amrita.edu
S. Madhu
Department of Automobile Engineering, Saveetha School of Engineering, Saveetha Institute of
Medical and Technical Sciences, Chennai, Tamil Nadu, India

© Springer Nature Singapore Pte Ltd. 2021 361

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_32
362 A. H. Ansari et al.

1 Introduction

Polymer matrix composites potentially used in automobile, medical and engineering

industries. They offer high stiffness, strength, abrasion resistance, creep resistance
and non-flammability. These composites have been widely used for many applica-
tions because they are light in weight and ease of manufacturing. As compared with
other engineering materials, polymer matrix composites have high brittleness. Nowa-
days, hybrid composites were fabricated which consist of different fibre with suitable
matrix using various methods [1]. As compared with other engineering materials,
polymer matrix composites have high brittleness. In order to increase the ductil-
ity, hybrid FRP composites (CFRPP and SRPP) were fabricated [2]. Carbon fibre
is widely used as reinforcing materials in advanced polymer composites because of
their unique properties. In order to control the surface characteristics, the carbon fibre
has undergone for surface treatments [3]. Composite laminates have susceptible to
damage induced by impact as compared with traditional metals [4]. Fibre reinforced
metal laminate composites offer better damage tolerance to fatigue crack growth and
impact damage [5].
Tensile behaviour of hybrid polymer composites fabricated with carbon fibre and
three different sizes of stainless steel wire mesh were investigated and reported that
addition of SS wire mesh increased the tensile strength [6]. Aluminium reinforced
CFRP hybrid polymeric composite have been fabricated and investigated the friction
and wear behaviour. This investigation concluded that the applied load is the signif-
icant parameter that affects the friction coefficient [7]. In the study, fibre reinforced
polymer composite fabricated with aluminium thin sheets using embossing process.
The effect of fibre orientation and polymeric matrix on the feasibility of hybrid joint
was investigated [8]. Much previous research works investigated about thermoset-
ting fibre reinforced polymer composite in terms of wear, tensile load, shear load,
displacement and absorption energy were studied. Several studies on hybrid steel
wire mesh reinforced fibre composite were also performed [9, 10].
Cryogenic-treated carbon fabric epoxy composite was prepared by hot compres-
sion moulding technique. Wear characteristics of treated and untreated carbon fab-
ric/epoxy (CF/E) composite were investigated using pin-on-disc method at room
temperature for varying load and sliding speed. In this work, morphology of worn
surfaces of the wear test composite specimens was studied by scanning electron
microscope (SEM) [11]. Carbon fibre reinforced epoxy composites containing 42
and 52 wt% carbon fibres were fabricated using moulding technique. Wear and fric-
tion characteristics were investigated using a pin-on-flat plate configuration. In this
work, static and dynamic coefficients of friction and wear rates of the fabricated
epoxy composites under different loading conditions were experimentally deter-
mined. Microstructures of the fabricated composites revealed that applied load and
speed are the significant parameters that affect the wear properties [12]. Effect of the
normal load and sliding speed on wear rate of two-body abrasive wear of aluminium
reinforced polymer composite was investigated [13]. Recently, several studies on the
wear behaviours of the hybrid composites were also performed [14–19].
Wear and Friction Behaviours of Stainless Steel (SS 316) … 363

The objective of this current research is to fabricate the stainless steel wire mesh
(SS316) reinforced carbon fabric epoxy composite using hand layup process, and the
tribological properties such as wear rate, frictional force and coefficient of friction
were investigated on CFRP-SS316 composite sample under different applied load,
sliding velocity, sliding distance and speed.

2 Materials and Method

2.1 Friction Material Processing

Stainless steel reinforced carbon fibre reinforced polymer composite is fabricated

using hand layup process which is used for this investigation. Four layers of carbon
fibre reinforced polymer composite arranged in unidirectional manner and stainless
steel 316 wire mesh with the size of 50 microns material are combined with CFRP
to fabricate a new composite material. Epoxy and hardener with 10:1 ratio is used
as a matrix to strength both CFRP and SS316. The 30 N load was applied on the
fabricated laminate for 24 h for uniform distribution of matrix and fibre. Figure 1a
shows the fabricated CFRP-SS316 composite. Abrasive water jet machining is used
for making the pin-on-disc specimen with inner and outer diameter 6 mm and 55 mm,
respectively. Hence, a disc (CFRP-SS316) is formed with its initial weight 7.527 g.
EN31 steel is a high carbon alloy manufactured by following forging, annealing,
stress reviling, hardening and tempering to find the wear and frictional property of a
composite. The height and thickness of the pin is 1 inch and 8 mm, respectively, as
shown in Fig. 1b CFRP-SS316 sample and EN31 steel pins. Figure 1c demonstrated
the microstructure fabricated CFRP-SS polymer composite.

2.2 Wear Testing Characterization

The pin-on-disc wear test is carried out to predict the wear performance of the
fabricated carbon fibre reinforced SS polymer composite. A pin-on-disc wear tester
is used to find the wear rate and coefficient of friction of the samples. A pin is located
against a flat rotating disc specimen such that a circular wear path is described by the
machine. The machine is used to evaluate wear and friction properties of materials
under pure sliding conditions. The disc can serve as specimen, while the other as
counter face. Pin with various geometry is used. Figure 2a shows the two-dimensional
model of pin and disc. Figure 2b shows the CAD model of wear mechanism.
The disc placed above the flat rotating surface and a pin is just above the disc
with sliding diameter as shown in Fig. 2a. The wear track formed depending upon
the load applied on the material through pin as shown in Fig. 2b. This also shows the
metal removal rate and wears depth of cut of material.
364 A. H. Ansari et al.

Fig. 1 a Fabricated CFRP samples. b Wear test specimen. c SEM image of the fabricated CFRP-SS
polymer composite

(a) 2D model of pin and disc ( b) 3D drawing of wear mechanism

Fig. 2 Wear mechanism

Wear and Friction Behaviours of Stainless Steel (SS 316) … 365

2.3 Pin-on-Disc Wear Parameters

The wear test parameters used for this investigation is applying load (N), sliding
velocity (m/s), sliding distance (m) and speed (rpm). The above three parameter
values are considered depending upon the material properties. Table 1 shows the
wear parameters used for this investigation, and Fig. 3a, b shows the pin-on-disc
apparatus and sample.
The specific wear rate is calculated as loss in volume loss (V ) divided by applied
load (P) and sliding distance (L) as shown in the following Eq. 1

k = m/P · L (1)

where ‘m’ is mass loss, ‘P’ is the applied load and ‘L’ is the sliding distance.

3 Results and Discussions

The wear tests have been done on CFRP-SS316 material specimen, and its average
values are given in Table 2. With the help of software and arrangement made in the
wear equipment, it was possible to record reading on every 1 s, and for the 16.664 min

Table 1 Wear test parameters

S. No. Parameter Range
1 Applied load (N) 10, 20, 30, 40
2 Sliding velocity (m/s) 1, 1.5, 2, 2.5
3 Sliding distance (m) 1000, 1500, 2000, 2500
4 Speed (RPM) 623, 637, 1038, 1274

Fig. 3 a Pin-on-disc apparatus. b Wear test specimen

366 A. H. Ansari et al.

Table 2 Wear rate and COF of CFRP-SS316 polymer composite

Exp Load Speed Sliding Sliding Wear rate Frictional Coefficient
(N) (rpm) velocity distance (mm3 /N-min) force (N) of friction
(m/s) (m)
1 10 623 1 1000 16.34 1.008 0.1008
2 20 637 1.5 1500 118.1 1.92 0.096
3 30 1038 2 2000 379.2 7.38 0.246
4 40 1274 2.5 2500 633.9 10.76 0.269

test duration, 998 readings were recorded for the wear rate and coefficient of friction.
Table 2 shows the average values of wear rate and coefficient of friction.

3.1 The Effect of Pin-on-Disc Parameters on Wear Rate

Figure 4 shows the effect of pin-on-disc parameters on wear rate of CFRP-SS316

composite material. Figure 4a shows the effect of applied load on wear rate. From the
figure, it was observed that the wear rate increases on CFRP-SS316 material when

700
700
600
600
Wear rate (mm3/N-m)
Wear rate (mm3/N-m)

500
500
400 400

300 300

200 200
100 100
0 0
0 10 20 30 40 50 0 1 2 3
Load (N) Sliding velocity (m/s)
(a) Effect of applied load on wear rate (b) Effect of sliding velocity on wear rate

700 700
600 600
Wear rate (mm3/N-m)
Wear rate (mm3/N-m)

500 500
400 400
300 300
200 200
100 100
0 0
700 1200 1700 2200 2700 500 700 900 1100 1300 1500
Sliding distance (m) Speed (rpm)
(c) Effect of sliding distance on wear rate (d) Effect of speed on wear rate

Fig. 4 Effect of pin-on-disc parameters on wear rate

Wear and Friction Behaviours of Stainless Steel (SS 316) … 367

the applied load increases. The minimum wear rate obtained from this experiment
is 16.34 mm3 /N-m (10 N) and 633.9 mm3 /N-m was obtained for 40 N applied load
as shown in Fig. 4a. Also, the minimum wear rate is obtained for sliding velocity of
1 m/s and sliding distance of 1000 m. The minimum speed is 623 rpm for less wear
rate and for maximum speed 1274 rpm, the wear rate is high as shown in Fig. 4d.
The applying load has a significant influence on wear rate. When the load increases,
the friction between the pin and the work surfaces increases which in turn increase
the wear rate of the fabricated composite [19].
The reason behind high wear loss of a material is that due to the applied load
which is high on the material, there was no possibility of the fibre detachment due
to deep embedding of fibres in the epoxy resin. The need of stainless steel 316 mesh
inside the material is the wear rate that decreases due to less friction. The sliding
velocity is a circumferential speed of the rotating face. When the sliding velocity
increases on the material, the contact pressure between the pin on material increases
and affects the unidirectional fibres for more material loss. The high speed also is
one of the reasons, why the removal rate of material is high.

3.2 The Effect of Pin-on-Disc Parameters on Coefficient

of Friction (COF)

The effect of wear parameters on COF for the fabricated CFRP-SS316 composite
material is shown in Fig. 5. The effect of COF increases on CFRP-SS316 material
when load increases. The minimum COF is 0.1008 for 10 N applied load and maxi-
mum 0.269 for 40 N applied load as shown in Fig. 5a. In the same way, the minimum
COF is 0.1008 for sliding velocity and sliding distance (i.e., 1 m/s and 1000 m) and
maximum 0.269 for 2.5 m/s and 2500 m as shown in Figs. 5b, c. For minimum speed,
623 rpm less COF was obtained, and for maximum speed 1274 rpm, COF is high
as shown in Fig. 5d. The effect of coefficient of friction increases, because at low
speed, the friction is high, whereas at high speed the friction is less.

4 Conclusions

The tribological properties of carbon fibre reinforced polymer—stainless steel 316—

is investigated in this work. The effect of wear rate and coefficient of friction on load,
sliding velocity, sliding distance and speed was investigated. The SS reinforced car-
bon fabric reinforced epoxy composite shows relatively good sliding wear resistance.
The addition of stainless steel 316 material in carbon fabric improved the wear resis-
tance and acts as solid lubricant during sliding. The developed composite material
can be used in brake disc to improve the wear resistance.
368 A. H. Ansari et al.

0.3 0.3

0.25 0.25

0.2 0.2

COF
COF

0.15 0.15

0.1 0.1

0.05 0.05
10 20 30 40 50 0.75 1.25 1.75 2.25 2.75
Load (N) Sliding Velocity (m/s)
(a) Effect of applied load on COF (b) Effect of sliding velocity on COF
0.3
0.3
0.25
0.25
0.2
0.2
COF
COF

0.15
0.15
0.1

0.1 0.05

0.05 0
500 1000 1500 2000 2500 3000 0 500 1000 1500
Sliding distance (m) Speed (rpm)
(c) Effect of sliding distance on COF (d) Effect of speed on COF

Fig. 5 Effect of pin-on-disc parameters on COF

References

1. Sathishkumar TP, Naveen J, Satheeshkumar S (2014) Hybrid fiber reinforced polymer

composites—a review. J Reinf Plast Compos 33(5):454–471
2. Swolfs Y, Meerten Y, Hine P, Ward I, Verpoest I, Gorbatikh I (2015) Introducing ductility
in hybrid carbon fibre/self-reinforced composites through control of damage mechanisms.
Compos Struct 131:259–265
3. Jin F-L, Park S-J (2015) Preparation and characterization of carbon fiber-reinforced thermoset-
ting composites: a review. Carbon Lett 16(2):67–77
4. Davim JP, Marques N, Baptista AM (2001) Wear 251:1100–1104
5. Sinmazçelik T, Avcu E, Bora MÖ, Çoban O (2011) A review: fibre metal laminates, background,
bonding types and applied test methods. Mater Des 32:3671–3685
6. Truong GT, Choi K-K (2019) Tensile behavior of hybrid composites of carbon fibres—steel
wire mesh reinforced polymer. Mech Adv Mater Struct
7. Caliman R (2016) Friction behaviour of aluminium composites mixed with carbon fibers with
different orientations. Mater Sci Eng 145:022035
8. Huang Z, Sugiyama S, Yanagimoto J (2014) Adhesive–embossing hybrid joining process to
fibre-reinforced thermosetting plastic and metallic thin sheets. Procedia Eng 81:2123–2128
9. Sakthivel Munisamy, Vijayakumar Subramani (2017) Influence of stainless-steel wire mesh on
the mechanical behaviour in a glass-fibre-reinforced epoxy composite. Mater Technol/Mater
Technol 51:455–461
10. Hasselbruch H, von Hehl A, Zoch H-W (2014) Properties and failure behavior of hybrid wire
mesh/carbon fibre reinforced thermoplastic composites under quasi-static tensile load. Mater
Des
11. Shivamurthy B, Murthy K, Anandhan S (2018) Tribology and mechanical properties of carbon
fabric/MWCNT/epoxy composites. Adv Tribol, Article ID 1508145
Wear and Friction Behaviours of Stainless Steel (SS 316) … 369

12. Sahinand Y, De Baets P (2017) Tribological behaviour of unidirectional carbon fibre reinforced
epoxy composites. Mater Sci Eng, 174
13. Odabas D (2018) Effects of load and speed on wear rate of abrasive wear for 2014 Al alloy.
Mater Sci Eng, 295
14. Suresha B, Chandramohan G, Renukappa NM, Siddaramaiah (2007) Mechanical and tribolog-
ical properties of glass-epoxy composites with and without graphite particulate filler. J Appl
Polym Sci 103:2472–2480
15. Khalladi A, Elleuch K (2017) Tribological behavior of wheel rail contact under different
contaminants using pin-on disk methodology. J Tribol 139:011102–011108
16. Rudas JS, Gomez LM, Toro A, Gutiérrez JM, Corz A (2016) Wear rate and entropy generation
sources in a Ti6Al4V—WC/10Co sliding pair
17. Uvaraja VC, Natarajan N, Rajendran I, Sivakumar K (2013) Tribological behavior of novel
hybrid composite materials using Taguchi technique. J Tribol 135:021101–021109
18. Shibo W, Niu C, Teng B (2017) Tribological behavior of polytetrafluoroethylene: effect of
sliding motion. J Tribol 139:011301–011306
19. Xu S, Liu T, Tangpong XW, Zhong W-H (2012) Wear and friction of carbon nanofiber-
reinforced HDPE composites. J Tribol 134:041602–041606
Tribological Properties of PEEK
Reinforced with Synthetic Diamond
Composite

K. Rajkumar , K. Vishal, and P. Sabarinathan

Abstract Tribological behaviour of synthetic diamond reinforced poly ether ether

ketone was studied for polymer self-lubricate bearing components. Addition of rein-
forcement to the matrix improved its mechanical, thermal and tribo properties of
composite. Synthetic diamond provides many advantageous properties which is
responsible for higher hardness and lowered the thermal degradation. This paper
deals with synthetic diamonds of various volume fractions reinforced with PEEK.
This thermo polymer matrix composite was manufactured by powder metallurgy
process. The characterization of synthetic diamond and polymer matrix composites
was done by scanning electron microscopy. Wear and tribo coefficient of developed
PEEK matrix composite was found by tribometer. Diamond particles were used to
prevent severe galling during higher loading.

Keywords Wear · Synthetic diamond · Galling · Polymer matrix composite

1 Introduction

Tailor to the multifunctional properties of the material, one potential direction is to

develop composite materials that possess combined properties of two or more materi-
als. Metals are being replaced by polymeric materials, because of its inexpensiveness.
They are also becoming materials of choice for many engineering applications. Poly-
meric materials have low density-to-strength ratio making them as more preferable.
Semi-crystalline thermoplastics compose regions of both amorphous and ordered
molecular structure which let them show a greater performance. As a result, these
polymers have chemical resistance as well as wear resistance [1]. Polymer materials
with low softening temperature have been replaced by high-temperature polymer like
poly ether ether ketone (PEEK). It is known as a semi-crystalline high-performance
thermoplastic with excellent mechanical properties. It is the most popular polymeric

K. Rajkumar (B) · K. Vishal · P. Sabarinathan

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, India
e-mail: rajkumark@ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 371

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_33
372 K. Rajkumar et al.

material. It has greater elastic modulus, mechanical strength, wear resistance and
inert to chemical reactivity. It is easy processable characteristics.
Recently, the usage of PEEK in the engineering and medical fields is increasing to
multifold. It has good resistance against biological actions [3, 4]. It is a material for
high temperature-bearing application since it has high creep property up to 250 °C
[2]. Various researches have been done to test the tribo behaviour of PEEK and its
composites. Various carbon allotropes are being reinforced with thermoplastics due
to their high mechanical properties and low manufacturing cost. Some of such mate-
rials are graphite, diamond and graphitize carbon (long, short) fibre. Carbon fibre
as a reinforcement in PEEK brings in good effect in its strength and tribo nature
[5, 6]. PEEK composites are being used in many areas like electrical, automobile,
biomedical and aerospace. Advantageous properties like lightness, self-lubrication
and easy maintenance can be achieved by using this polymer during the application.
Tribological performance of PEEK and PTFE has been researched. The results stated
that PEEK is used for the production of précise-machined custom bearings which
are suitable for market needs. PEEK’s properties such as self-lubrication, high cor-
rosion resistance and high melting temperature (340 °C) make it to be considered
as one of the high-performance materials among polymers [7]. The stable opera-
tion of the polymer and polymer composites were influenced by the frictional wear
[9]. Thermal conductivity of the composite was increased by the addition of the
nano-particle [10]. Dispersion of zirconia fillers in the polymer matrix increases the
mechanical property, thermal stability and certain combinations of their physical and
mechanical properties. The amount of nano-filler affects the thermal stability of the
polymer composite [11]. Diamonds exhibit properties like high thermal conductivity,
high corrosion resistance, high wear resistance, low coefficient of thermal expansion
and high hardness. Due to higher thermal conductivity and higher modulus of syn-
thetic diamond, it is used as reinforcement in high temperature polymer [12]. By the
addition of nano-sized TiO2 , the wear rate of the PEEK can be reduced [13]. Wear
rate of the polymer matrix can be reduced in large sum by the addition of diamond
as reinforcement. These literatures were reporting tribological properties of PEEK
composite reinforced with ceramic and carbon nano-materials. Synthetic diamond
is another form of high-strength crystalline carbon reinforcement which enhances
wear resistance of polymer materials.
In this work, pure peek with different weight percentages of synthetic diamond
were added in PEEK matrix, and its tribological properties were studied. Wear loss
and friction coefficient of synthetic diamond reinforced PEEK were reported as
function of sliding velocities.
Tribological Properties of PEEK Reinforced with Synthetic … 373

Table 1 Properties of poly

Properties Values
ether ether ketone (PEEK)
Size 20 µm
Density 1.32 g/cm3
Glass transition temperture 143 °C
Yield stress 107 MPa

2 Experimental Procedure

2.1 Materials

The received poly ether ether ketone of grade 2000FP powder form properties is
shown in Table 1. In this, PEEK makes the matrix phase, and synthetic diamonds
(SD) are the reinforcing material. The size of PEEK and synthetic diamond were
20 µm and 5 µm, respectively. Poly vinyl alcohol (PVA) was used as a binder agent.

2.2 Preparation of Composites

Powder metallurgy process was used to fabricate the polymer composite. The syn-
thetic diamond was varied from 10 to 30 volume % to reinforce PMC. To get a
homogeneous mixture, pure PEEK powder and different composition of synthetic
diamond were mixed in ball milling technique. This mixture was mixed for three
hours. The composites were prepared by pressing the powder hydraulically and sin-
tering it. The powder was pressed at 100 MPa to get a cylindrical shape of diameter
10 mm and height 10 mm. During sintering process, temperature ramp of 100 °C
was maintained. The polymer composites were sintered at 3250 °C and soaked for
two hours. P/M process produces high potential homogeneous composite materials,
it involves lesser time, and it is cost efficient. Cylindrical specimens were fabricated
in the form of pellets, as shown in Fig. 1.

2.3 Micro-Structural Analysis

To confirm the structure of diamond, XRD was carried out for the synthetic diamond
powder. From the fabricated samples, particle distribution in the composite was
evaluated by optical microscope. SEM with 15 kV of accelerating voltage is applied
for image observation. Before SEM analysis, the specimen was cleaned with alcohol
and then gold sputtered over the specimen to make it surface conductive.
374 K. Rajkumar et al.

Fig. 1 Fabricated PEEK

composites

2.4 Density

The density of the polymer matrix composite (PMC) was measured by Archimedes’
principle. The fabricated composite was submerged in water, and its equivalent
weight is calculated. Weight of the PMC was also calculated in the air. From these
values, the density of the PMC was obtained.

2.5 Hardness

Hardness of PMC was measured by Vickers hardness testing machine NOVA

330/360. The diamond intender in the form of right pyramid with the angle of 136°
was used. The hardness was measured with a loading time of 15 s. Load of 100 kgf
is applied for about 10–15 s. The average value of 20 measurements was taken to
measure the hardness of the polymer matrix composite.

2.6 Fourier Transform Infrared (FTIR) Spectroscopy

Fourier transform infrared (FTIR) spectroscopy was conducted in BRUCKER, Ger-

many, 3000 Hypersion. It was observed that the sample absorbed the IR radiations
that were passed through it. The usefulness of infrared spectroscopy is to evaluate
the difference in chemical structures (molecules) of PMC.
Tribological Properties of PEEK Reinforced with Synthetic … 375

2.7 Tribological Test

Pin-on-disc wear testing is a method of characterizing the wear rate and coefficient
of friction, and frictional force between two sliding materials. The tests have been
conducted under normal atmospheric condition in PIN ON DISC FRICTION &
WEAR TEST RIG–TE-165-SPOD. The PMC pin specimen was normally loaded
over the steel disc. Test was conducted on a flat surface of cylindrical pin of diameter
10 mm and height of 10 mm. Tribo test was conducted for a constant normal load
of 100 N using different sliding velocities of 0.6, 1.2, 1.8, 2.4 m/s. Each test was
conducted for a period of 30 min which ensures a steady-state coefficient of friction
for all the tested samples. An average of three trials was done for each sample. All
the tribological sliding tests were conducted at 20 °C in a dry-sliding condition.

3 Results and Discussion

3.1 Material Characterization

Figure 2 shows the XRD profile of synthetic diamond. From the figure, peeks at 430
and 750 show the presence of carbon which confirms that the compound is diamond.
The crystal structure was found to be cubic in nature.
Figure 3 shows the typical optical image of the fabricated composite. The image
clearly shows the synthetic diamond particles that were fairly dispersed. However, the
presence of agglomerates is obvious. No micro-cracks are seen from the fabricated
sample.

Fig. 2 XRD pattern of SD

376 K. Rajkumar et al.

Fig. 3 Optical micrograph

of PPEK-30% SD composite

Synthetic diamond particles

Fig. 4 FTIR spectrums

(Black-10%SD, Red-20%SD
and Green-30%SD)

3.2 FTIR

The sample of PEEK/synthetic diamond composite was analysed using the FTIR
technique. A typical spectrum of PEEK/synthetic diamond composite can be seen
in Fig. 4. The spectrum is trace distinctive of amorphous PEEK/synthetic diamond
composite. The finger-print region of the spectrum (1432–1024 cm−1 ) exhibits the
distinctive spectral features typical to an IR-recorded spectrum of PEEK.

3.3 Density

The effect of synthetic diamond on the density of peek is shown in Fig. 5. Due to
the high density of diamond, the density of the composite material increased as the
reinforcement increased.
Tribological Properties of PEEK Reinforced with Synthetic … 377

Fig. 5 Density of PEEK

composites

3.4 Mechanical Testing

As per the ASTM E92 standard, Vickers hardness test was conducted for the polymer
composite. The hardness of peek and synthetic diamond is shown in Fig. 6. Hardness
of the peek is improved when reinforcement is added, due to the high strength of
synthetic diamond. Although there was an agglomeration of diamond particle beyond
10%, the hardness of peek composite was observed to increase slightly. This is
because of the termination and pinning mechanism.

Fig. 6 Hardness of PEEK

composites
378 K. Rajkumar et al.

3.5 Tribo Analysis

As seen from the previous section, the 30% SD reinforced composite shows a high
mechanical property. Therefore, the tribological studies are limited to this composi-
tion only when compared to pure PEEK. In this section, tribo behaviour of 30% SD
reinforced composite was done with different sliding speed at constant normal load.
The wear performance is influenced by sliding velocities according to the volume
fractions and environment. Figure 7 shows the wear rate of virgin PEEK and PEEK
composite. This figure shows that wear rate of fabricated composite and pure PEEK
is decreasing with sliding velocity. This is because the contact region of the pin
and the counterpart gets minimized [7]. PEEK composite shows a lower wear rate
as compared to pure PEEK. SD in PEEK makes this composite stronger to sliding
wear loss. As a result, the load-bearing constituent of SD provides a direct contact
to counter surface. Greatest hardness of SD also shares the contact load and reduces
tearing of PEEK [6]. It can be seen from SEM image Fig. 8a that PEEK surface was
torn out along with some micro-cutting grooves. However, PEEK composite was
observed to be the only counter surface with hard asperities scoring marks over the
surface, as shown in Fig. 8b.
Under normal loading, there is a decrease in coefficient of friction at room temper-
ature with sliding velocity increasing, as seen from Fig. 9. The graph clearly shows
the reduction of coefficient of friction from 0.047 for the pure PEEK to 0.035 for
the PEEK reinforced with synthetic diamond composite. The reduction of COF is
around 50%. This is because of the addition of the SD which renders the rolling or
mending effect of smaller-sized SD particles between the sliding surfaces [8].

Fig. 7 Wear rate versus sliding velocity

Tribological Properties of PEEK Reinforced with Synthetic … 379

Fig. 8 Worn surface SEM images (a) pure PEEK (b) 30%SD-PEEK

Fig. 9 Coefficient of friction versus sliding velocity

4 Conclusions

1. The effects of several volume fractions of fillers in a PEEK polymer matrix were
studied. The mechanical and wear rate of PEEK and different composition of
synthetic diamond powder have been studied in this paper.
2. It was observed that the dispersion of synthetic diamond particles in PEEK matrix
is fair at low volume of fraction.
3. Density of PEEK composite increased due to the addition of synthetic diamond
powder, due to the higher density of diamond.
4. The hardness of the PEEK polymer composite is higher with increase in the
reinforcement when compared with pure PEEK.
5. Under constant loading condition of 100 N, tribological tests were performed at
different sliding velocities of 0.6 m/s, 1.2 m/s and 1.8 m/s. Friction coefficients
and wear rates were evaluated, and from scanning electron microscopy, worn
surfaces were examined.
380 K. Rajkumar et al.

6. As the amount of the synthetic diamond powder increased in the polymer matrix,
the wear rate of the composite decreased.
7. With the increase in reinforcement, coefficient of friction decreases. The
coefficient of friction of PEEK-based composites was reduced by nano- and
micro-particle which becomes more significant at higher contact pressure.

References

1. Simsiriwong J, Shrestha R, Shamsaei N, Lugo M (2015) Effect of microstructural inclusion on

fatigue life of poly ether ether ketone (PEEK). J Mech Behav Biomed Mater 51:388–397
2. Werner P, Altstadt V, Jaskulka R, Jacobs O, Sandlerd JKW, Shaffer MSP, Windle AH
(2004) Tribological behaviour of carbon-nanofibre-reinforced poly(ether ether ketone). Wear
257:1006–1014
3. Patel P, Richard Hull T, Richard E, Stanislav I, Richard N, Crowley S, Natallia S (2011)
Investigation of the thermal decomposition and flammability of PEEK and its carbon and
glass-fiber composites. J Polym Degrad Stab 96(1):12–22
4. Xie GY, Sui GX, Yang R (2011) Effects of potassium titanate whiskers and carbon fibers on
the wear behavior of poly ether ether ketone composite under water lubricated condition. J
Compos Sci Technol 71(2):828–835
5. Avci A, Arikan H, Akdemir A (2004) Fracture behavior of glass fiber reinforced polymer
composite. J Cem Concr Res 34(1):429–434
6. Ovaert TC, Cheng HS (1991) Counter-face topographical effects on the wear of PEEK and a
PEEK-carbon fiber composite. Wear 150(2):150–157
7. Koike H, Kida K, Santos EC, Rozwadowska J, Kashima Y, Kanemasu K (2012) Self-lubrication
of PEEK polymer bearing in rolling contact fatigue under radial loads. Tribol Int 49:30–38
8. Avanzini A, Donzella G, Mazzu A, Petrogalli C (2013) Wear and rolling contact fatigue of
PEEK and PEEK composites. Tribol Int 57:22–30
9. Koike H, Kida K, Mizobe K, Shi X, Oyama S, Kashima Y (2015) Wear of hybrid radial bearing
(PEEK ring –PTFE retainer and alumina ball) under dry rolling contact. Tribol Int 90:77–83
10. Shang LG (2002) Characterisation of interphase nanoscale property variations in glass fibre
reinforced polypropylene and epoxy resin composites. Composites 33(4):559–576
11. Mishra TK, Kumar A, Pandey VV, Kumar V (2013) PEEK composites reinforced with zirconia
nano filler. Compos Sci Technol 72(1):1627–1632
12. Zhao Z, Liu Y, Feng W, Zhang Q, Cheng L, Zhang L (2017) Improvement on the thermal
conductivity of Diamond/CVI-SiC Composites using large diamond particles. Diam Relat
Mater 74:1–8
13. Kurdi A, Wang H, Chang L (2018) Effect of nano-sized TiO2 addition on tribological behaviour
of poly ether ether ketone composite. Tribol Int 117:225–235
Experimental Investigation
on the Mechanical Properties
of American Agave and Glass Fibre
Reinforced Polypropylene Composites

M. Indra Reddy, M. Anil Kumar, and Vamsi Inturi

Abstract The utilization of natural fibres has become extensive instead of conven-
tional synthetic fibres because of the issues with the greenhouse effect and conscious-
ness towards the environment. Exploring natural fibres has several advantages such as
economically viable, availability in fibrous form readily and ease of extraction from
the plant leaves. In this study, an attempt has been made to investigate the mechanical
properties such as tensile, impact and flexural strength of the composites made by
reinforcing American agave fibre into a polypropylene resin. In addition to that, the
mechanical properties of hybrid composites that are fabricated by reinforcing Amer-
ican agave and glass fibre as 1:1 proportion into a polypropylene resin are studied.
The proportion of fibre content is varied from 10 to 30% by volume percentage, and
the behaviour in the mechanical properties in each case is examined. It is observed
that the American agave and glass fibre reinforced polypropylene hybrid compos-
ite has exhibited better mechanical properties than American agave fibre reinforced
polypropylene composite.

Keywords American agave fibre · Polypropylene · Mechanical properties

1 Introduction

1.1 A Subsection Sample

Over the years, natural fibres are serving mankind and find their application in vari-
ous fields. Eventually, natural fibres are contending with the man-made fibres, espe-
cially, as far as sustainability, quality and economy of production are concerned [1].

M. Indra Reddy (B) · M. Anil Kumar

Department of Mechanical Engineering, S R K R Engineering College, Bhimavaram, India
e-mail: reddy.indra.m@gmail.com
V. Inturi
Department of Mechanical Engineering, BITS Pilani, Hyderabad Campus, Hyderabad, Telangana,
India

© Springer Nature Singapore Pte Ltd. 2021 381

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_34
382 M. Indra Reddy et al.

Because of their biodegradable nature, the utilization of natural fibres has become
expensive. Alongside, they often have low densities, beneath processing costs and
sustainable than the synthetic materials [2]. Conventionally, natural fibres have been
used as reinforcements in polypropylene composites due to their recyclability nature
[3]. The alkali and acrylic acid treatment of the banana fibres could enhance the
water resistance property and improves the mechanical properties of the composites.
However, the addition of compatibilizer will further intensify the above-mentioned
properties. It was reported that the addition of compatibilizer improves the impact
and flexural strength which further yields to alleviate the water absorption capac-
ity [4]. Zaman et al. [5] considered irradiated banana fibres and treated them with
monometer in MeOH solution blended with 2% benzyl peroxide. The authors have
used thermal curing method and observed a significant enhancement in the mechan-
ical properties because of the monometer treatment. Rana et al. [6] examined the
influence of compatibilizer on jute fibre reinforced polypropylene composites. The
authors observed that there is a reduction in the water absorption capacity and a sud-
den increase in the mechanical properties. This can be attributed to the fact that the
addition of compatibilizer developed a linkage between the hydroxyl groups of jute
and the carboxyl groups of the compatibilizer [6]. On the other hand, the properties
such as modulus of composites and tensile strength are affected by fibre composition,
liquid of immersion and number of immersion cycles. In the case of kenaf fibre rein-
forced polypropylene composites, there is a depression in the tensile strength values
because of the number of immersion cycles and conditioning. However, continuous
and cyclic immersion in bleach resulted to alter the tensile strength of the composites
significantly. However, there is a minimal variation in the values of tensile modulus
after a number of immersion cycles [7].
Oksman et al. [8] examined the effect of fibre morphology on the mechanical prop-
erties of natural fibres and on the fibre breakage of the composites due to the extru-
sion process. It was observed that the sisal composites had displayed the favourable
impact properties and the longest fibres after the extrusion among the considered
natural fibres. In general, the flexural stiffness of the composites increases with fibre
content, being highest for flax-type composites. Perhaps, the addition of fibres has no
effect on the flexural strength property because of the low compatibility [8]. Based on
the exhaustive literature review, it has been found that the mechanical properties of
American agave, polypropylene composites have not explored. The main objective
of this study is to prepare the composite by incorporating the American agave and
glass fibres at various volume percentages into a polypropylene resin matrix. Finally,
the composites are examined and characterized to evaluate the tensile, impact and
flexural properties.

2 Experimental Procedures

The following section elaborates the experimental procedure carried out, and the
steps involved are:
Experimental Investigation on the Mechanical Properties … 383

1. Materials
2. Extraction of fibres
3. Fabrication of composites
4. Investigating the mechanical properties of composites.

2.1 Materials

The materials used in this experimental work are:

1. Reinforcement: American agave fibre and glass fibre
2. Matrix: Polypropylene resin
3. Hardener/Catalyst: Methyl ethyl ketone peroxide
4. Accelerator: Cobalt naphthenate.

2.2 Mater Extraction of Fibres

American agave fibre is extracted directly from the American agave plant. Initially,
teen-aged American agave plants are taken, and then, the leaves are plucked. Later,
the edges of the leaves are trimmed, and then, these edge-trimmed American agave
plant leaves are allowed to soak in water for about 10–15 days. After soaking for
sufficient time, the residual from leaves gets deposited in the water. Remaining leaf
is taken into hand and rubbed, and final fibre is obtained (these fibres appear like
threads) and then dried in sunlight. For removing additional organic waste from the
fibre, we go for alkali treatment. In alkali treatment, we soak the fibre in 1% NaOH
solution for sufficient time, and later, the fibres are again dried in sunlight to obtain
the agave fibre. The American agave plant is shown in Fig. 1.

2.3 Fabrication of Composites

Polymer matrix composites are made using polypropylene resin matrix to evaluate
the reinforcing ability of American agave fibres. The accelerator (1.5% by volume)
and catalyst (1.5% by volume) are supplemented to the resin for curing at room tem-
perature. The appropriate mixture of polypropylene resin and American agave fibres
is filled up in the prepared mould and fed to injection moulding machine. In order
to ensure good quality fibre reinforced composites, the deformation and movement
of the fibre should be minimized. Hence, a compressive pressure of 0.05 MPa is
enforced on to the mould, and the composite specimens are allowed to cure for 24 h.
In addition to the curing at room temperature, post-curing at 70 °C for about two
384 M. Indra Reddy et al.

Fig. 1 American agave fibre

hours is carried out onto the specimens. The fabricated samples are represented in
Fig. 2. Five composite samples are prepared with five different volume percentages
of American agave Fibres.

2.4 Tensile Test

In order to investigate the tensile properties of the American agave and glass fibre
reinforced polypropylene composites, the specimens are prepared according to the
ASTM D 638M standards (Standard, 2010). The composite specimens are prepared
with a length of 165 mm, width of 12.7 mm and thickness of 3 mm. Three identical
specimens are tested for each percentage volume fraction of fibre. These composite
specimens are tested using a tensile testing machine operating with a cross-head
speed of 2.5 mm/min.

2.5 Flexural Test

In order to examine the flexural properties of the specimens, three-point bend tests
are carried out in accordance with ASTM D 790M (Standard, 2010) test standards.
Three-point bend test is chosen, as it provides the flexibility in determining the centre
point deflections with the equipment. In addition to that, it operates with the less
amount of material for each test cycle. The specimens are made with 125 mm length,
12.7 mm width and 3 mm thickness. During the test, the outer rollers are kept 64 mm
apart, and the specimens are examined under the strain rate of 0.2 mm/min. For each
composition, three identical specimens are tested using tensile testing machine, and
all the experiments are carried out at a cross-head speed of 2.5 mm/min. Flexural
Experimental Investigation on the Mechanical Properties … 385

Fig. 2 Fabrication of composites

strength (σ ) and modulus (E f ) are calculated using the following formulae:

σ = 3Pl/2bt 2
E f = l3m/bt 3

where
σ Bending strength (N/mm2 )
P Maximum load (N)
L Distance between the supports (mm)
B Width of the specimen
T Thickness of the specimen
M Slope of the initial straight portion of the load-deflection curve.
386 M. Indra Reddy et al.

2.6 Impact Test

The test specimens for carrying the impact test are prepared according to the standards
of ASTM D256. These specimens are having a length of 55 mm, width of 10 mm
and thickness of 10 mm. To achieve the notch onto the specimen, a sharp file having
an included angle of 45° is drawn across the cut at an angle of 90° to the sample axis.
These specimens are fractured in a plastic Izod impact testing machine (capacity—
21.68 J). For each composition, three specimens are tested, and the value of the
absorbed energy is noted. Finally, the impact strength for a specimen is calculated
by dividing the measured absorbed energy value with the cross-sectional area.

3 Results and Discussions

3.1 Tensile Strength

The variation of mean tensile strength and tensile modulus of composite with vary-
ing fibre content is illustrated in Figs. 3 and 4. It can be observed that, the tensile
strength is increasing with the increase in the fibre content of the polypropylene
matrix. This can be attributed to the fact that as the polypropylene resin dissem-
inates, it dispenses the applied stress to the American agave fibres which further
resulted to higher strength value. The tensile strength of the composite with maxi-
mum fibre content (30% volume of the fibre) is found to be 27.03 MPa. The tensile
strength and the tensile modulus of American glass fibre reinforced polypropylene
composite considered in this study are far better than that of American agave fibre

Fig. 3 Variation of tensile strength of composites with volume of fibre

Experimental Investigation on the Mechanical Properties … 387

Fig. 4 Variation of tensile modulus of composites with volume of fibre

reinforced polypropylene composites. From the tensile test, we obtained load versus
displacement graphs at 30% volume of fibre, which are shown in Fig. 5. The tensile
tested samples are represented in Fig. 6.

Fig. 5 Load versus displacement curve of composite at 30% volume of fibre

388 M. Indra Reddy et al.

Fig. 6 Tensile tested specimens

3.2 Flexural Strength

Figures 7 and 8 describe the flexural behaviour of American agave fibre reinforced
composites and American agave–glass fibre reinforced composites. The trend is
similar to the trend observed for the case of tensile properties with the same above-
mentioned conviction. The flexural strength value of the composite with maximum
fibre (30% volume of the fibre) content is about 84 MPa. The American agave–glass
fibre reinforced composite considered in this study has shown improved flexural
strength value than that of American agave fibre reinforced composites. The flexural
tested samples are represented in Fig. 9.

Fig. 7 Variation of flexural strength of composites with volume of fibre

Experimental Investigation on the Mechanical Properties … 389

Fig. 8 Variation of flexural modulus of composites with volume of fibre

Fig. 9 Flexural tested specimens

3.3 Impact Strength

Figure 10 describes the results of Izod impact test. From below figure, it is evident
that as the percentage of volume increases, the impact strength value increases. It
can be noted that the impact strength value of the composite with maximum fibre
(30% volume of the fibre) content is found to be 5.8 J. The impact strength of
390 M. Indra Reddy et al.

Fig. 10 Variation of impact strength of composites with volume of fibre

American agave–glass fibre composite is better than American agave fibre reinforced
composites. The impact tested samples are represented in Fig. 11.

Fig. 11 Impact tested specimens

Experimental Investigation on the Mechanical Properties … 391

4 Conclusions

In this study, American agave fibre reinforced polypropylene composite and Ameri-
can agave–glass fibre reinforced polypropylene hybrid composites are fabricated with
different volume of fibres, and the mechanical properties, namely tensile, impact and
flexural properties, are examined. After evaluating the obtained experimental results,
the following conclusions are made:
• The tensile strength and tensile modulus of the American agave and agave–glass
fibre reinforced polypropylene composite increases with the increase in volume
of fibre.
• The flexural strength and flexural modulus of the American agave and agave–glass
fibre reinforced polypropylene composites increases with the increase in volume
of fibre.
• The impact strength of the American agave and agave–glass fibre reinforced
polypropylene composites increases with the increase in volume of fibre.
• American agave and glass reinforced polypropylene composites show bet-
ter mechanical properties than American agave fibre reinforced polypropylene
composites.

References

1. Kozłowski RM (ed) Handbook of natural fibres: types, properties and factors affecting breeding
and cultivation. Elsevier
2. Heng JY, Pearse DF, Thielmann F, Lampke T, Bismarck A (2007) Methods to determine surface
energies of natural fibres: a review. Compos Interfaces 14(7–9):581–604
3. Joshi SV, Drzal LT, Mohanty AK, Arora S (2004) Are natural fiber composites environmentally
superior to glass fiber reinforced composites? Compos A Appl Sci Manuf 35(3):371–376
4. Prasad N, Agarwal VK, Sinha S (2016) Banana fiber reinforced low-density polyethy-
lene composites: effect of chemical treatment and compatibilizer addition. Iran Polym J
25(3):229–241
5. Zaman HU, Khan MA, Khan RA (2013) Banana fiber-reinforced polypropylene composites: s
study of the physico-mechanical properties. Fibers Polym 14(1):121–126
6. Rana AK, Mandal A, Mitra BC, Jacobson R, Rowell R, Banerjee AN (1998) Short jute fiber-
reinforced polypropylene composites: effect of compatibilizer. J Appl Polym Sci 69(2):329–338
7. Haniffah WH, Sapuan SM, Abdan K, Khalid M, Hasan M, Hoque ME (2015) Kenaf fibre
reinforced polypropylene composites: effect of cyclic immersion on tensile properties. Int J
Polym Sci
8. Oksman K, Mathew AP, Långström R, Nyström B, Joseph K (2009) The influence of
fibre microstructure on fibre breakage and mechanical properties of natural fibre reinforced
polypropylene. Compos Sci Technol 69(11–12):1847–1853
Influence of Magnesium Hydroxide
Fillers on Acoustic, Thermal, and Flame
Retardant Properties of Pu Foam

L. Yuvaraj, S. Jeyanthi, Digvijay D. Kadam, and R. G. Ajai

Abstract Polyurethane foam is a versatile material for many applications like acous-
tic, thermal insulation, as well as for energy absorption. The main aim of this paper is
to improve acoustic properties by adding weight percentage of 2, 4, and 6 of magne-
sium hydroxide (Mg(OH)2 ) micro-particles in polyurethane (Pu) foam. To investigate
the influence of micro-particles on acoustic properties, the samples are tested in an
impedance tube to measure sound absorption coefficient. The experimental results
are compared to finite element results predicted from Johnson–Champoux–Allard
model. For thermal properties, the samples are experimented in thermal conductivity
and fire retardant test. The results indicate that the significant improvement in the
acoustic and thermal properties due to the addition of magnesium hydroxide.

Keywords Polyurethane foam · Magnesium hydroxide · Sound absorption

coefficient · Thermal conductivity · Flame retardancy

1 Introduction

The noise produced by vehicles not only affects the quality of the automobile but
also decreases the comfort of driving. Eliminating automotive interior noise has
become an important and predominant topic in noise control engineering. This noise
reduction depends on the coefficient of sound absorption. It is necessary to identify the
coefficient of sound absorption of polyurethane at different frequency ranges. These
porous materials such as polyurethane foam have been widely used in various fields
like acoustic absorption thermal insulation, aeronautics, and building construction.
Cao et al. [1] have discussed Delany–Bazley model for sound absorption of porous
material. Ekici et al. [2] improved coefficient of sound absorption of polyurethane
foam by the addition of tea leaf and luffa-cylindrical. Tao et al. [3] explained the

L. Yuvaraj · S. Jeyanthi (B) · D. D. Kadam

School of Mechanical and Building Sciences, VIT, Chennai, Tamil Nadu 600127, India
e-mail: jeyanthi.subramanian@vit.ac.in
R. G. Ajai
Department of Mechanical Engineering, SVCE, Sriperumpudur, Tamil Nadu, India
© Springer Nature Singapore Pte Ltd. 2021 393
S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_35
394 L. Yuvaraj et al.

effect of rice straw and wheat straw fiber in polyurethane foam on sound absorption
property. Chen et al. [4] added bamboo chips and bamboo stem in polyurethane
foam in order to enhance the coefficient of sound absorption and also carried out
airflow resistivity test. Tiuc et al. [5] have discussed the addition of textile waste in
polyurethane foam with detailed microstructure studies they noticed improvement
in the acoustic properties of developed hybrid foams. Benkreira et al. [6] investigated
porosity, airflow resistivity, and sound absorption of porous material. Verdejo et al.
[7] have discussed the effect of the addition of carbon nanotube in polyurethane
foam on the acoustic property as well as on Young’s modulus. Orfali et al. [8] added
silicon oxide and carbon nanotubes in polyurethane foam to enhance the coefficient of
sound absorption. Rasheed et al. [9] have explained compressive strength, thermal
conductivity, and moisture absorption test of polyurethane foam with glass wool.
Chen et al. [10] have discussed the effect of variation of reactants like percentage of
blowing agent of polyurethane foam on sound-absorbing capacity.
Sripathy and Sharma [11] described the flammability and moisture absorption
test of polyurethane foam. Mazraeh-shahi et al. [12] explained method to calibrate
thermal diffusivity and thermal conductivity of porous material. Moretti et al. [13]
explained the influence of granule size of aerogels on acoustic and thermal proper-
ties and also explained transmission loss. Xu et al. [14] discussed thermal stability
and flame retardancy property of pure polyurethane and the addition of organoclay
nano-composites. Nedkov et al. [15] and Ahmadinejad et al. [16] explained finite
element modeling of the impedance tube and calibration of the coefficient of sound
absorption. Wang et al. [17] have enhanced the coefficient of sound absorption by
using rice hull in polyurethane foam. Matyka et al. [18] have described the con-
cept of tortuosity, porosity, and the relationship between these factors of porous
material. Lim et al. [19] discussed the sound-absorbing property of knife fibers and
also discussed the effect of thickness of the sample and air gap behind sample on
the coefficient of sound absorption. Othmani et al. [20] and Santoni et al. [21] pre-
sented different mathematical models for porous absorption. The explained concept
of porosity, touristy, flow resistivity calibrated coefficient of sound absorption for
recycled sugarcane wastes.
Putra et al. [22] explained the effect of air gap on sound-absorbing property of
porous material by using pineapple-leaf fibers. Gle et al. [23] explained the acoustic
properties of hemp concrete as a green building material. Piegay et al. [24] discussed
sound absorption and transmission loss of flax wool and hemp wool samples, and
carried out validation with Delany–Bazley and Tarnow perpendicular random model.
Curtu et al. [25] used wood fiber, jute, and wood chips to prepare biodegradable com-
posite samples and evaluated the acoustic properties of these samples. Hurrell et al.
[26] explained mathematical models of airflow resistivity for non-woven materials.
Berardi et al. [27] explained the comparative study of sound absorption of various
fibers like coconut, straw, cane, sheep wool, cork, hemp, and kenaf. Kuranska et al.
[28] explained microstructure details of pure and composite polyurethane foams and
investigated their compressive strength performance. Or et al. [29] explained oil
palm empty fruit bunch fibers as the sustainable acoustic absorber and carried out
an experiment of sound absorption to explain the effect of air gap behind the sample
Influence of Magnesium Hydroxide Fillers on Acoustic … 395

and also explained the effect of density and thickness of sample on sound absorption
property of porous material.
Fouladi et al. [30] explained the Biot–Allard model for sound absorption of coir
fiber and projected comparative results of the impedance tube, Delany–Bazley model,
and Biot–Allard model. Koruk et al. [31] experimented with pure luffa fiber, luffa
fiber with perforated lines, and luffa composite with epoxy with varying volume
fraction in order to evaluate the coefficient of sound absorption and transmission
loss. Martellottaa et al. [32] explained the acoustic-absorbing property of samples
made from olive pruning wastes and also projected the effect of air gap and the effect
of thickness of samples.
In this study, the different weight percentage of magnesium hydroxide is added to
Pu foam. The acoustic capabilities of the foam samples measured in impedance tube
test, and other physical parameters are tested in the individual experimental setup.
These physical parameters are given as input to the FEA analysis to predict the
numerical sound absorption. For thermal properties, the sample is tested in thermal
conductivity and flame retardant test.

2 Sample Preparation

The flexible foams are a simple mixture of isocyanides and polyol in the ratio of
100:65. First of all, pure foam with no additives is prepared by adding the two
compounds in a circular mold with one face open and letting the reaction take place.
The reaction is exothermic, and carbon dioxide gas is released during the blowing
process. The mixture is allowed to expand and is then removed from the mold and
is prepared according to the required dimensions. Now, three different samples were
prepared with different weight percentages of magnesium hydroxide, i.e., 2%, 4%,
6% with average particle size, 6.8 (µm), reactants, and their weight percentage are
mentioned in Table 1. Composition of magnesium hydroxide filled hybrid foam is
mentioned in Table 2. Samples are shown in Fig. 1

Table 1 Reactants used for

Chemicals %wt
pure PUF
Isocyanate 43.55
Polyol 53.6
Blowing agent 1.6
Catalyst 0.85
Surfactants 0.4
396 L. Yuvaraj et al.

Table 2 Composition for

Sample name %wt of PuF %wt of Mg(OH)2
hybrid foam
a 100 0
b 98 2
c 96 4
d 94 6

a) Pure PUF b) 98% PUF + 2 % Mg(OH)2

c) 96% PUF + 4 % Mg(OH)2 d) 94% PUF + 6 % Mg(OH)2

Fig. 1 Test samples of pure and hybrid polyurethane foam

3 Experimental Testing and FEA Analysis

3.1 Porosity Measurement

In order to characterize and better prediction of sound absorption coefficient, porosity

is an essential physical parameter. Porosity is defined as the amount of fluid that
occupied in the pores. Porosity can be tested by pressure/mass method introduced
by Salissou et al. [33]. Foam sample placed in air-sealed container and the mass
is calculated in high- and low-pressure condition. A simple formula is stated by
Salissou which gives us porosity from measured values. The schematic working and
the existing test setup are shown in Fig. 2.
Influence of Magnesium Hydroxide Fillers on Acoustic … 397

Fig. 2 Experimental setup

for porosity measurement

3.2 Flow Resistivity Measurement

The airflow resistivity is major parameter for governing acoustic behavior of porous
material. The airflow resistivity depends on the direction of airflow through the
material. The flow resistance (Rf ) is the ratio of the pressure difference between
two sides of the sample to the air velocity, which is stably passed through a porous
medium. The flow resistivity (σ ) flows resistance per unit thickness of the sample.
Experimental setup is shown in Fig. 3. By using this method, we can measure airflow
resistivity of porous material ranging from 100 to 10,000 Ns/m4 , and air velocity can
vary up to 50 mm/s. The pressure differential across samples range from 0.1 to 250 Pa.

P
Rf = (1)
V

Fig. 3 Experimental setup

of flow resistivity
measurement
398 L. Yuvaraj et al.

Fig. 4 Experimental setup

for tortuosity measurement

Rf P
σ = ,σ = (2)
d V ×d

The test samples with 100 mm diameter and 40 mm thickness are held between
the holder and mounting plate. Compressed air with 50 mm/s is supplied to one side
of the sample, and pressure difference (P) is measured by a digital pressure gauge.

3.3 Tortuosity Measurement

Tortuosity is a dimensionless structural parameter, which indicates the effects of

pore structure on the velocity of fluid flow. It is defined as the ratio of actual path
length through the material to the linear path length. It shows how well the porous
material prevents direct flow through the porous medium. The more complex the
path, the more time a wave is in contact with the absorbent, and hence, the more
energy dissipation and more absorbing capability. Using the below test setup, we
can find the actual length of the twisted path. This experimental setup constitutes
transmitter (T) and receiver (R) ultrasonic sensors, where the signal allow to pass
from T to R and the travel time is measured. The experiment carried out in two
conditions as presence and absence of sample and the existing test setup is shown in
Fig. 4.

3.4 Sound Absorption Measurement

In this setup, confined acrylic tube with different diameter based on cut-off frequency
is used to find the sound absorption coefficient as per ASTM E-1050. The test setup is
Influence of Magnesium Hydroxide Fillers on Acoustic … 399

Fig. 5 Measurement of sound absorption by impedance tube

equipped with two microphones, loudspeaker, and DAQ system. A loudspeaker that
excites the sound in quantified frequency and microphone senses the sound pressure
at position one and two. The two-channel DAQ systems were used to record the time
domain data for post-processing. The MATLAB® software is used to estimate the
transfer function across two microphone data. Based on the transfer function, normal
incidence sound reflection and absorption obtain following Eqs. (4) and (5)

P1
H= (3)
P2
H − e− jks j2k(l+s)
R= e (4)
e jks − H

α = 1 − |R|2 (5)

S is the distance between two microphones, and l is the distance between the
speaker and first microphone (Fig. 5).

3.5 FEA Modeling of Impedance Tube

The acoustic body of impedance tube is presented with parameters like the speed
of sound in air and volume density. For meshing 3D brick, acoustic fluid element
is selected. The Johnson–Champoux–Allard model is selected for FEA analysis.
Harmonic mass source (loudspeaker) is inserted at one end of the tube, and another
400 L. Yuvaraj et al.

Fig. 6 FEA modeling of

impedance tube

end sample is placed. Source of sound generates plane waves which incident normally
on the sample to be tested. The frequency range is selected up to 5000 Hz. Mesh
sensitivity was carried, and the maximum element length must satisfy the following
equation (Fig. 6).

C
L= (7)
8f

where L is the maximum element length, C is the speed of sound in air, and f is the
maximum frequency of the model. Parameters used for FEA modeling of impedance
tube are listed below.
1. Impedance tube length = 600 mm
2. Impedance tube diameter = 33 mm
3. Speed of sound in air = 343 m/s
4. Density of air = 1.2 kg/m3 .
Other than above for FEA analysis, other parameters are also required like flow
resistivity, touristy, and porosity which is listed in the result section.

3.6 Measurement of Thermal Conductivity

Thermal conductivity depends on density, porosity, shape, and size of pores. The addi-
tion of fillers should not increase the thermal conductivity, or at least, the material
should maintain the existing value. So, it is essential to ensure the thermal conduc-
tivity of polyurethane filled with magnesium hydroxide. The principle used in heat
flow meter method is represented by the schematic diagram in Fig. 7. This method is
based on Fourier’s law of heat flow. The sample is placed between two plates. One
plate is heated, and the other is kept cool. The temperature gradient across sample
and heat flux calibrated by the sensor and thermal conductivity is calculated by the
equation of Fourier’s law. Thickness of all sample is maintained as 30 mm.
Influence of Magnesium Hydroxide Fillers on Acoustic … 401

Fig. 7 Schematic representation of thermal conductivity measurement

T
Q = KA (7)
d

3.7 Flame Retardancy Test

Basically, polyurethane foam has excellent mechanical, acoustic, and thermal

insulation properties, but its poor fire resistance leads to restriction for some
application.
So, the samples underwent fire retardant test to study the influence of magnesium
hydroxide as filler in Pu foam. The test is carried out by allowing to burn the sample
for different periods of 10 s, 15 s, and 20 s (Fig. 8).

4 Results

4.1 Porosity, Flow Resistivity, and Tortuosity of Sample

The experimental test results of porosity, flow resistivity, and tortuosity are listed in
Table 3. The results prove that the physical parameters are interconnected to each
other. Addition of magnesium hydroxides increases the porous nature of Pu foam
which means the number of pores is increased. The increasing number of pores leads
to more number of cell walls, which makes the porous path as resistive and complex.
One can notice that as the percentage of magnesium hydroxide increased, the values
of porosity, flow resistivity, and tortuosity is also increased.
402 L. Yuvaraj et al.

Fig. 8 Sample subjected to

flame retardancy test

4.2 Sound Absorption Results

As mentioned previously, magnesium hydroxide enhances all parameters and the

same resembles in the sound absorption result as shown in Fig. 9. The percentage of
Mg(OH)2 filler in polyurethane foam increases the coefficient of absorption which
increases over all frequency ranges for all samples. However, in case of 6% sample,
the sound absorption performance is poor between the range of 0–1000 Hz, and
it is mainly due to the addition of more magnesium hydroxide causes the foam to
behave rigid by compared to other samples. The other fact is that pores turn to void
if excess addition of fillers and severely collapse the foam beyond 6% of magnesium
hydroxide.

4.3 Comparison of Experimental and FEA Result

Figure 10 shows the comparison of experimental results to numerical results which is

predicted from experimental results of porosity, flow resistivity, and tortuosity. There
is a minimal gap notice at some region in the comparison plot which is mainly due

Table 3 Porosity, flow resistivity, and tortuosity factor of samples

Samples Porosity Flow resistivity (Ns/m4 ) Tortuosity factor
Pure PUF 0.857 3126.56 1.08
98% PUF + 2% Mg(OH)2 0.869 3463.21 1.13
96% PUF + 4% Mg(OH)2 0.914 3768.21 1.14
94% PUF + 6% Mg(OH)2 0.943 3974.1 1.16
Influence of Magnesium Hydroxide Fillers on Acoustic … 403

Fig. 9 Comparison of the coefficient of sound absorption of pure and hybrid polyurethane foam

to non ideal condition of experimental setup such as air leakage, loose fit of sample
and so on.

4.4 Thermal Conductivity Measurement

There is no significant increase in thermal conductivity of pure and hybrid

polyurethane foam. Thus, it maintains its insulation property with the addition of
Mg(OH)2 filler content in polyurethane foam (Table 4).

4.5 Flame Retardancy Test

The flame retardant property of samples b, c, and d increased significantly, based on

the percentage of addition. The enhancement occurs due to the magnesium hydroxide
is a non-flammable material. During burning, the magnesium hydroxide starts to
liberate water vapors. The layer of water vapors starts to envelop the flame, and
thus, flame retardancy property of these hybrid foam can be enhanced. Magnesium
hydroxide Mg(OH)2 acts as smoke suppressor as well as acid-free flame retardant
(Fig. 11).
404 L. Yuvaraj et al.

a) Pure Polyurethane foam

b) 98% PUF + 2 % Mg(OH)2

c) 96% PUF + 4 % Mg(OH)2

Fig. 10 Comparison of experimental and FEA result of the coefficient of sound absorption of
different foams
Influence of Magnesium Hydroxide Fillers on Acoustic … 405

d) 94% PUF + 6 % Mg(OH)2

Fig. 10 (continued)

Table 4 Parameters of measurement of thermal conductivity

Sample (W/m) T1 (°C) T2 (°C) K (W/m)
Pure PUF 58.01 80 30 0.035
98% PUF + 2% Mg(OH)2 58.3 80.2 30.34 0.0358
96% PUF + 4% Mg(OH)2 58.2 80.15 31.91 0.0362
94% PUF + 6% Mg(OH)2 58.15 80.12 32.32 0.0365

5 Conclusion

It is concluded that this study investigated the acoustic and thermal properties dif-
ferent weight percentage addition of magnesium hydroxide as filler in Pu foam. The
filler enhances the physical parameters of foam samples as well as thermal properties.
The measured sound absorption results compared with numerical FEA showing fairly
good agreement to each other. The insulating properties of foam samples remain con-
stant, which is noticed in the result of a thermal conductivity test. Non-flammable
nature of magnesium hydroxide improves fire retardant property of polyurethane
foam.
406 L. Yuvaraj et al.

Fig. 11 Flammability Test of various samples

Acknowledgements We would like to thank the Department of Science and Technology (DST-
SERB; File no: ECR/2015/000111) for providing us the necessary facilities and funds for conducting
this study.

References

1. Cao L, Fu Q, Si Y, Ding B, Yu J (2018) Porous materials for sound absorption. Compos

Commun 31(10):25–35
2. Ekici B, Kentli A, Küçük H (2012) Improving sound absorption property of polyurethane foams
by adding tea-leaf fibers. Arch Acoust 37(4):515–520
3. Tao Y, Li P, Cai L (2016) Effect of fiber content on sound absorption, thermal conductivity, and
compression strength of straw fiber-filled rigid polyurethane foams. BioResources 11(2):4159–
4167
Influence of Magnesium Hydroxide Fillers on Acoustic … 407

4. Chen S, Jiang Y (2018) The acoustic property study of polyurethane foam with addition of
bamboo leaves particles. Polym Compos 39(4):1370–1381
5. Tiuc AE, Vermeşan H, Gabor T, Vasile O (2016) Improved sound absorption properties of
polyurethane foam mixed with textile waste. Energy Procedia 85:559–565
6. Benkreira H, Khan A, Horoshenkov KV (2011) Sustainable acoustic and thermal insulation
materials from elastomeric waste residues. Chem Eng Sci 66(18):4157–4171
7. Verdejo R, Stämpfli R, Alvarez-Lainez M, Mourad S, Rodriguez-Perez MA, Brühwiler PA,
Shaffer M (2009) Enhanced acoustic damping in flexible polyurethane foams filled with carbon
nanotubes. Compos Sci Technol 69(10):1564–1569
8. Orfali WA (2015) Acoustic properties of polyurethane composition reinforced with carbon
nanotubes and silicon oxide nano-powder. Phys Procedia 70:699–702
9. Rasheed AK, Marhoon II (2015) Mechanical and physical properties of glass wool-rigid
polyurethane foam composites. Alnahrain J Eng Sci 18(1):41–49
10. Chen S, Jiang Y, Chen J, Wang D (2015) The effects of various additive components on the
sound absorption performances of polyurethane foams. Adv Mater Sci Eng
11. Sripathy M, Sharma KV (2013) Flammability and Moisture absorption test of rigid
polyurethane foam. Int J Sci Eng Res 4(2)
12. Mazraeh-shahi ZT, Shoushtari AM, Bahramian AR (2015) A new method for measuring the
thermal insulation properties of fibrous silica aerogel composite. Procedia Mater Sci 11:583–
587
13. Moretti E, Merli F, Cuce E, Buratti C (2017) Thermal and acoustic properties of aerogels:
preliminary investigation of the influence of granule size. Energy Procedia 111:472–480
14. Xu Z, Tang X, Zheng J (2008) Thermal stability and flame retardancy of rigid polyurethane
foams/organoclay nanocomposites. Polym-Plast Technol Eng 47(11):1136–1141
15. Nedkov T, Pleshkova-Bekjarska S (2015) Modeling of 3D impedance tube with a complex
termination impedance using finite element method. International Scientific Conference on
Information, Communication and Energy Systems and Technologies, 2015 june, pp 170–173,
Sofia, bulgaria
16. Ahmadinejad P, Monazzam MR (2018) 3D simulation of impedance tube: effect of mesh size
on sound absorption characteristic. Int J Comput Sci Netw Secur 18(1):179–182
17. Wang Y, Zhang C, Ren L, Ichchou M, Galland MA, Bareille O (2013) Influences of rice hull in
polyurethane foam on its sound absorption characteristics. Polym Compos 34(11):1847–1855
18. Matyka M, Khalili A, Koza Z (2008) Tortuosity-porosity relation in porous media flow. Phys
Rev E 78(2):026306
19. Lim ZY, Putra A, Nor MJ, Yaakob MY (2018) Sound absorption performance of natural kenaf
fibres. Appl Acoust 15(130):107–114
20. Othmani C, Taktak M, Zain A, Hantati T, Dauchez N, Elnady T, Fakhfakh T, Haddar M (2017)
Acoustic characterization of a porous absorber based on recycled sugarcane wastes. Appl
Acoust 120:90–97
21. Santoni A, Bonfiglio P, Fausti P, Marescotti C, Mazzanti V, Mollica F, Pompoli F (2019) Improv-
ing the sound absorption performance of sustainable thermal insulation materials: natural hemp
fibres. Appl Acoust 150:279–289
22. Putra A, Or KH, Selamat MZ, Nor MJ, Hassan MH, Prasetiyo I (2018) Sound absorption of
extracted pineapple-leaf fibres. Appl Acoust 136:9–15
23. Glé P, Gourdon E, Arnaud L (2011) Acoustical properties of materials made of vegetable
particles with several scales of porosity. Appl Acoust 72(5):249–259
24. Piégay C, Glé P, Gourdon E, Gourlay E, Marceau S (2018) Acoustical model of vegetal wools
including two types of fibers. Appl Acoust 129:36–46
25. Curtu I, Stanciu MD, Cosereanu C, Vasile O (2012) Assessment of acoustic properties of
biodegradable composite materials with textile inserts. Materialeplastice 49:68–72
26. Hurrell AI, Horoshenkov KV, Pelegrinis MT (2018) The accuracy of some models for the
airflow resistivity of nonwoven materials. Appl Acoust 15(130):230–237
27. Berardi U, Iannace G (2017) Predicting the sound absorption of natural materials: best-fit
inverse laws for the acoustic impedance and the propagation constant. Appl Acoust 115:131–
138
408 L. Yuvaraj et al.

28. Kuranska M, Prociak A (2012) Porous polyurethane composites with natural fibres. Compos
Sci Technol 72(2):299–304
29. Or KH, Putra A, Selamat MZ (2017) Oil palm empty fruit bunch fibres as sustainable acoustic
absorber. Appl Acoust 119:9–16
30. Fouladi MH, Ayub M, Nor MJ (2011) Analysis of coir fiber acoustical characteristics. Appl
Acoust 72(1):35–42
31. Koruk H, Genc G (2015) Investigation of the acoustic properties of bio luffa fiber and composite
materials. Mater Lett 15(157):166–168
32. Martellotta F, Cannavale A, De Matteis V, Ayr U (2018) Sustainable sound absorbers obtained
from olive pruning wastes and chitosan binder. Appl Acoust 141:71–78
33. Salissou Y, Panneton R (2007) Pressure/mass method to measure open porosity of porous
solids. J Appl Phys 101:124913
Development and Analysis of GFRP
Conical Springs

D. Vivek, R. Praveen, A. Sanjay Krishnan, Y. K. Sabapathy, D. Ebenezer,

and M. Selvaraj

Abstract Increasing competition and innovation in automobile sector tends to mod-

ify the existing products or replace old products by new and advanced material prod-
ucts. A suspension system of vehicle is also an area where these innovations are
carried out regularly. Nowadays, the automobile industry has shown much interest in
using glass fibre reinforced plastics (GFRP) components replacing conventional steel
components due to its ‘high strength to low weight’ ratio. Therefore, replacement of
steel conical suspension springs (in heavy automobiles) with GFRP conical springs
with an aim to reduce its weight was attempted. A manual pultrusion and braiding
process is approached for fabricating the FRP open coil springs using e-glass and
epoxy resin. Its total deformation, equivalent stress and equivalent elastic strain were
analysed using GFRP conical spring at different static loads of 20, 40, 60 and 80 N.

Keywords Glass fibre reinforced plastics (GFRP) · Epoxy resins · Manual

pultrusion · Creo 3.0 · ANSYS

1 Introduction

Generally, automobile parts are made up of steel and its alloys. Steel is an alloy con-
sisting mainly of iron with carbon content of 0.2–2.1% by weight. Steel has many
physical properties such as density, elasticity, strength and hardness. Among these
properties, the strength property is very important, which can be achieved by the addi-
tion of more materials like manganese, niobium and vanadium. But, steel has some
limitations like having more weight, and its corrosion resistance is low. By using FRP
components, fuel efficiency and endurance of various components in machines can
be achieved. FRP is slowly replacing steel, especially in the automobile sector. FRP is

D. Vivek · A. Sanjay Krishnan

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, India
R. Praveen · Y. K. Sabapathy · D. Ebenezer (B) · M. Selvaraj
Sri Sivasubramaniya Nadar College of Engineering, Chennai, India
e-mail: ebenezerd@ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 409

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_36
410 D. Vivek et al.

a composite material made up of polymer matrix reinforced with fibres [6]. The fiber
is usually glass (fiber glass), carbon (in carbon-fiber-reinforced polymer), aramid
or basalt. The polymer is usually an epoxy, vinyl ester or polyester thermosetting
plastic, though phenol formaldehyde resins are still in use. FRP composite material
has excellent mechanical properties such as impact resistance, strength, stiffness and
flexibility. Main advantages of this FRP are corrosion resistance, low weight and
high formability [7, 8]. FRP materials are anisotropic and can be tailor made to suit
any requirement to counter stresses in any desired direction and location. The com-
ponents made of FRP are thermally and dimensionally stable and well suited to work
at high temperatures. FRP is widely used in many industries like automobile suspen-
sion system, railway wagons, electrical applications, aerospace and marine, etc [8].
The conical springs are cone shaped springs to provide near constant spring rates
with lower solid height than traditional compression springs. The improved solid
height is achieved by nesting active springs within each other during compression.
It can also achieve the objective of reducing the weight of the engine component,
thus increasing shock absorption, reducing spring weight and improving strength of
springs [9].

2 Literature Review

Conical compression springs designed by Krishna et al. [3] is having the character-
istics of a cylindrical spring of much greater free length.
The spring is characterised by wire whose diameter diminishes from base to apex
end and by coils whose pitch diminishes from base to apex end such that the ratio
of mean coil diameter to wire diameter is constant from coil to coil. Further, each
coil diameter is preferably dimensioned to fit within the next larger coil permitting
a spring to be designed which exhibits a small solid compressed height [5].
Kudapane [4] invented a conical spring wherein the diameters of the coils vary
progressively and the spaces separating adjacent coils vary in relation to the dimen-
sions of the coils to provide pre-determined force displacement spring characteristics.
A conical spring is provided wherein the diameters of the coils thereof vary progres-
sively, and the spaces separating adjacent coils vary in relation to the dimensions
of the coils to provide pre-determined force displacement spring characteristics has
been referred.
Optimisation of steel helical spring by composite spring was presented by Mehdi
Bakhshesh, and Majid Bakhshesh [2] proposed springs that can reserve high level
of potential energy and have undeniable role in industries. Helical spring is the most
common element that has been used in car suspension system. In their research, steel
helical spring related to light vehicle suspension system under the effect of a uniform
loading has been studied.
Arularasan and Sabapathy [1] fabricated GFRP helical spring with E-glass and
epoxy resin using semi-mechanised pultrusion process, and braiding process has
Development and Analysis of GFRP Conical Springs 411

found that a compression load of 1205.327 N with the displacement corresponding

as 101.986 mm was studied.

3 Experimentation

The equipments required for GFRP fabrication process consists of fibre rolls, tray,
clamps, steel tube, PVC hose pipe and wooden mould. By using marker pen, the pitch
is marked in the periphery of the wooden mould and hammered with number of nails
in a marked area in order to mention the pitch gap. The E-glass fibres are made with
different loops. Now, the resin and hardener are poured into a tray and thoroughly
mixed until it attains heat reaction. This process is called exothermic reaction.
After, the fibres are dipped and completely submerged with resin and hardener
solution. A steel tube is rigidly fixed by holding with clamps. A PVC hose tube is
inserted over one side of steel tube and another side is kept open.
Steel wire is inserted into hose tube through steel tube. Now, steel wire is tied
with fibres having more number of loops, and other side of steel wire ending is used
for pulling. Slowly, the steel wire is pulled through steel tube.
The fibre with solution is doped into a hose tube. Finally, the hose tube containing
solution and fibres are rolled on the mould with preferred pitch of 20 mm. After
a period of 40–48 h, the hose tube is removed by hammering on mould. Cut the
hose tube periphery, and FRP conical spring is manufactured. By using this data, the
analytical solving is done by using software called ANSYS.
The procedure for analysing GFRP conical spring for different loads is explained
in following, open the ANSYS software. Click and drag the static structural module.
Select the engineering data and upload the composite material as epoxy E-glass
with having unidirectional. Update the constant properties values like density (2000
kgm−3 ), Young’s modulus (x-direction = 4500 Mpa, y-direction = 10,000 Mpa and
z-direction = 10,000 Mpa), shear modulus (xy = 5000 Mpa, yz = 3846.2 Mpa and xz
= 5000 Mpa). Import the geometry from Creo 3.0 software in which it is converted
into IGES file format. Generate the model in ANSYS Workbench software. Now,
fix the one end of spring and other end as different loads (i.e., 20 N, 30 N, 60 N,
80 N) through setup option. Update and generate a solution which provides the total
deformation, equal elastic strain and equivalent stress using von-Misses formulation.
Finally, result is obtained and noted down the maximum values of total deformation,
equal elastic strain and equivalent stress for plotting the graphs.

3.1 3D Modelling

Figure 1 describes how to create a tapered spring in Creo 3.0. The helical sweep
option is used to create a tapered spring, define internal helix profile, select suitable
axis and make profile for axis of revolution. Now, draw a cross line with preferred
412 D. Vivek et al.

Fig. 1 Schematic line diagrams for creating tapered spring using Creo 3.0

dimensions, and after clicking ok, edit the profile to draw a wire diameter circle at
lowest point and apply the pitch required in millimetre. Finally, the conical spring
is generated and converted it into initial graphic exchange format (IGES) file. The
generated 3D model is shown below in Fig. 2.

Fig. 2 3D model
representations in Creo 3.0
Development and Analysis of GFRP Conical Springs 413

Fig. 3 Total deformation

4 Analysing of GFRP Conical Springs

4.1 Total Deformation

Total deformation is the elongation of conical spring. A compressive force is applied

on the top and bottom of the plate. So, the acquired maximum total deformation for
loads such as 20 N, 40 N, 60 N and 80 N are 22.74 mm, 45.48 mm, 68.22 mm and
90.96 mm. Figure 3 shows total deformation for 20 N load.

4.2 Equivalent Elastic Strain

The compressive loads such as 20 N, 40 N, 60 N and 80 N are applied on the spring

plates. The maximum values acquired such as 1.172, 2.345, 3.518 and 4.691 are
available. Figure 4 shows equivalent elastic strain for 20 N load.
414 D. Vivek et al.

Fig. 4 Equivalent elastic strain

4.3 Equivalent Stress

Von-Misses equation is used to calculate for various loads such as 20 N, 40 N, 60 N

and 80 N. The maximum values acquired are 10.708 MPa, 21.416 MPa, 32.124 MPa
and 42.832 MPa. Figure 5 shows the equivalent stress for 20 N load.

4.4 Strain Energy

Strain energy for different loads such as 20 N, 40 N, 60 N and 80 N is maximum of

0.4419 MJ, 1.768 MJ, 3.9776 MJ and 7.0713 MJ. Figure 6 shows the strain energy
for 20 N load (Table 1).

4.5 Graphical Analysis of Various Loads

Figure 7 shows the graphical representation between equivalent stress and equivalent
Development and Analysis of GFRP Conical Springs 415

Fig. 5 Equivalent stress

elastic strain. In this graph, equivalent stress is noted in x-axis, and equivalent elastic
strain is noted in y-axis. The maximum stress and strain values mentioned in the
result are used in x-axis and y-axis. After jointing the points, the curve is gradually
increased as shown in Fig. 7.
Figure 8 is a graphical representation between load and strain energy. The curve
clearly shows that increasing in load causes gradual increase in strain energy.

4.6 Fabrication of GFRP Conical Springs

Figure 9a shows the wooden mould which is hammered with nails in a pre-determined
dimensions before winding the hose tube. The pre-determined dimensions are also
said to be pitch of spring. Figure 9b shows the final product after completely cut and
removed hose tube.
416 D. Vivek et al.

Fig. 6 Strain energy

Table 1 Software analytical results of FRP conical spring

Load (N) Maximum total Strain energy Maximum Maximum
deformation (mm) (MJ) equivalent stress equivalent elastic
(MPa) strain (× 10−3 )
20 22.74 0.4419 10.708 1.172
40 45.48 1.7678 21.416 2.345
60 68.22 3.9776 32.124 3.518
80 90.96 7.0713 42.832 4.691

Fig. 7 Equivalent stresses

Equivalent elastic

versus equivalent elastic 0.005

strain 0.004
strain

0.003
0.002
0.001
0
0 10 20 30 40 50
Equivalent stress(MPa)
Development and Analysis of GFRP Conical Springs 417

Strain Energy(MJ)
6

0
0 20 40 60 80 100
Load(N)

Fig. 8 Load versus strain energy

Fig. 9 a Wooden mould with conical spring, b fabrication of conical spring

5 Conclusion

In the present paper, conical springs or tapered springs using GFRP composites as
epoxy resin and glass fibre were developed experimentally by the application of the
manual pultrusion technique that can be used effectively to improve weight saving.
In general, this study considered that the composite open coil springs can be used
for light commercial vehicle, together with saving of weight of the component.
In case of automobiles, the weight reduction is one of the important factors and
may not cause any technological impact. By using the geometric data, the conical
springs are investigated by the aid of analysis software called ANSYS.
The result shows that for such loads (20 N, 40 N, 60 N and 80 N), total defor-
mation decreases gradually, but elastic strain and stress are increased by applying
compressive force. It is found that maximum values for each load are obtained.
418 D. Vivek et al.

References

1. Arularasan R, Sabapathy YK (2014) Fabrication and testing of FRP open coil springs. Appl Mech
Mater 592–594:1065–1069. https://doi.org/10.4028/www.scientific.net/amm.592-594.1065
2. Bakhshesh M, Bakhshesh M (2012) Optimization of steel helical spring by composite spring.
Int J Multidiscip Sci Eng 3:47–51
3. Krishna J, Manoj M, Reddy P, Kumar S (2014) Design and optimization of leaf spring using
composites. IJMCA 2:163–168
4. Kudapane SA (2017) Modeling and finite element analysis of hybrid FRP leaf spring, pp 551–556
5. Mahanthi DL, Murali CVS (2017) Design and analysis of composite leaf spring for light weight
vehicle. Int J Adv Eng Res Sci 4:147–152. https://doi.org/10.22161/ijaers.4.3.23
6. Pozhilarasu V, Parameshwaran Pillai T (2013) Comparison of performance of glass fibre
reinforced plastic leaf spring with steel leaf spring. Int J Chem Tech Res 5:1339–1345
7. Sun W, Guan Z, Li Z (2017) Compressive failure analysis of unidirectional carbon/epoxy com-
posite based on micro-mechanical models. Chinese J Aeronaut 30:1907–1918. https://doi.org/
10.1016/j.cja.2017.10.002
8. Zarif Karimi N, Heidary H, Fotouhi M, Minak G (2017) Experimental analysis of GFRP lam-
inates subjected to compression after drilling. Compos Struct 169:144–152. https://doi.org/10.
1016/j.compstruct.2017.01.017
9. Zhou L, Zhao L, Liu F, Zhang J (2018) A micromechanical model for longitudinal compressive
failure in unidirectional fiber reinforced composite. Results Phys 10:841–848. https://doi.org/
10.1016/j.rinp.2018.08.003
Effect of Nano and Microfillers
in Basalt/Epoxy Composites

M. M. Metro and M. Selvaraj

Abstract There is a growing demand emerging from industries towards usage of

composite materials. Remarkable properties of basalt fibre reinforced polymer com-
posites are being tailored for its application in various fields. Effect of epoxy/basalt
fibre with nano and microparticles as filler has been studied in this work. Owing to
easier manufacturing process, low cost and exceptional properties, basalt fibre finds
its wider usage. Nano graphene platelets are added as filler to epoxy resin to enhance
physical and mechanical properties of matrix. Synergistic effect of zinc borate with
alumina trihydrate has also made them to incorporate with the resin as a flame
retardant. Influence of nano titanium dioxide with graphene platelet nanopowder as
reinforcement in basalt fibre-epoxy composite was also determined by appropriate
tests. Mechanical properties such as tensile and flexure strength were investigated in
accordance with ASTM standards. SEM image of failure pattern is also studied.

Keywords Basalt fibre · Epoxy · Nanographene platelets · Alumina trihydrate ·

Zinc borate · Titanium dioxide

1 Introduction

Recent research has its focus in energy conservation and environmental issues
for green living. Booming polymeric composite materials attracts the inventors in
research vicinity. Basalt fibre possesses high tensile strength when compared to glass
fibre and better fracture strain than do carbon fibre. Abundant availability of basalt
rock and its simple fabrication process makes the fibre a desired research component.
Exceptional behaviour of basalt fibre in chemical, weather and heat resistance made
itself a promising fibre [1]. Surface modification of basalt fibre improves interfacial
interaction [2].

M. M. Metro · M. Selvaraj (B)

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, India
e-mail: selvarajm71@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 419

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_37
420 M. M. Metro and M. Selvaraj

Current research results indicate that composites with micro and nano-sized mate-
rials reinforced in polymer matrix perform better mechanical properties than do con-
ventional fibre reinforced composites [3–6]. The fillers are added to polymer in order
to reduce the cost and also to improve physical or chemical properties of polymer.
Addition of inorganic particulate fillers such as nano/microglass, CNT, Al2 O3 ,
CaCO3 , silicates into polymers improves stiffness and strength of polymer matrix,
thus expanding their applications in diverse field. Particulate fillers also modify the
physical and mechanical properties of polymer composites in many ways.
In the present work, basalt fibres were treated by alkaline agent, sodium hydroxide
solution before fabrication of the composite. Effects of surface-treated basalt fabric
with the addition of nano and microfillers on mechanical behaviour were studied
by suitable tests. The commercially available nanoparticles are generally incorpo-
rated into polymers such as epoxy, polypropylene and polystyrene to form a matrix.
Introduction of nano fillers like carbon nanotubes, graphene into micrometal matrix
enhances mechanical properties and reduction in weight. High specific surface area
and 2-D structure of graphene increase tensile strength of the composite [3, 4]. These
nanoparticles enhance the mechanical, thermal and electrical properties of composite
matrix [5, 6].
Nano graphene platelets (GnPs) are similar to carbon nanotubes but possess excel-
lent performance characteristics when compared to other nano materials. Graphene
nanoplatelets are two dimensional (2-D) thin sheets made up of tiny stacks of
graphene, which replaces carbon nanotubes, carbon fibre, nanoclays in various com-
posite applications. Carbon atoms of graphene nanoplatelets are arranged in a honey-
comb structure with sp2 hybridization. Thickness of graphene nanoplatelet is of a sin-
gle carbon [7]. Two-dimensional platelets have capability of sliding over one another,
allowing higher loading and lower viscosities. GnPs are widely used as an ideal filler
in industry owing to its excellent mechanical and physical properties, including
exceptional conductivity and high specific strength. Incorporation of GnPs in very
small quantity enhances the performance of composite [8]. Addition of graphene also
has a considerable effect on flammability properties of the composites [9]. Outstand-
ing performance of graphene nanoplatelets is due to stability of sp2 bonds that form
a hexagonal lattice [10]. Surface modified CNT enhanced the performance of tensile
strength and Young’s modulus compared to plain CNT/basalt/epoxy composites [11,
12]. Silane modified CNT increases fracture toughness of basalt/CNT/epoxy by 40%
than acid modified CNT composites [13]. Addition of graphene platelets nanopowder
enhances thermo-mechanical and flame-retardant properties of polymer composite.
Flame retarders are substances that are added into polymer molecule or blended
physically into the polymers. Non-halogen additives are used with polymers for its
flame-retardant property. The most commonly used non-halogen additives are metal
hydroxides, like aluminium trihydroxide (ATH) due to their low cost and toxicity,
ease of handling and minimal corrosion effects. Alumina trihydrate (ATH), the most
extensively used flame retardant, is in usage for decades [14]. Fire retardant property
of alumina trihydrate flags an important role in epoxy resin. Alumina trihydrate itself
has the potential to retard flame inevitably. Synergistic effect of alumina trihydrate
could endure the flammable property of multicomponent epoxy nano composite [15].
Effect of Nano and Microfillers in Basalt/Epoxy Composites 421

Metal borates gained significant interest and extensive usage in industrial appli-
cation due to their excellent structural chemistry [16]. Zinc borates are being used
as a polymer additive for more than decades. Addition of zinc borate significantly
improves the flame-retardant property. Synergistic effect of zinc borates in addition
to polymer/metal hydroxide provides feasible properties [17].
Incorporation of aluminium trihydroxide and zinc borate in epoxy asphalt binders
increases flame retardancy of neat cold-mix epoxy asphalt binders. Synergistic effect
of aluminium trihydroxide and zinc borate in epoxy asphalt binders enhances glass
transition temperature of the composite. Halogen-free flame retardants disperse uni-
formly and increase the tensile strength [18]. Metal fillers in polymers enhance
mechanical performance of composite such as hardness. It also acts as a reinforce-
ment and increases the interfacial bond between fillers and polymers [19]. Mechan-
ical properties of metal matrix fabricated into epoxy depend mainly on size of parti-
cle. Tensile modulus of composite laminate fabricated with metal matrix reinforce-
ment increases with increase in filler [20]. The size of particles added as fillers in
epoxy concludes hardness value of fabricated composite. Aluminium particles when
added to epoxy resin increases the hardness of composite with decrease in particle
size [21]. Tensile modulus and density of metal matrix filled epoxy depend on rule
of mixtures [22]. Interfacial strength between fibre and composite is improved by
incorporation of titanium dioxide [23–25]. The interfacial shear strength is more for
graphene–titanium than that of graphene–aluminium [25].
Several studies have been conducted to explore the application of basalt fibre
as a reinforcing substance in various composites. However, performance analysis
of graphene nano platelets with microfillers and graphene nano platelets with nano
titanium dioxide on basalt/epoxy composite has not been studied elsewhere till date
to the maximum knowledge of author.

2 Materials and Procedures

2.1 Materials

The commercially available woven-type basalt fabric with surface density of

300 g/m2 was used as a reinforcing material in this study. Woven fabric has an average
fibre diameter of 12 µm. A solution of sodium hydroxide (NaOH) was used for sur-
face treatment of the fabric. Epoxy resin of bisphenol A diglycidylether (Araldite LY
556) and hardener (HY 951) were used for fabrication. Graphene platelet nanopow-
der of thickness ranging between 6 and 8 nm with surface area 150 m2 /g was used.
The purity of carbon was 99.5%. Alumina trihydrate, Al(OH)3 or Al2 O3 · 3H2 O, a
white amorphous powder with average particle size as 1.5–5 microns were used. Zinc
borate (2ZnO · 3B2 O3 _3.5H2 O), a white amorphous powder with an average particle
size less than 7 µm, was also used as microfiller. The weight ratio was considered by
422 M. M. Metro and M. Selvaraj

varied alumina trihydrate (ATH) and constant zinc borate. Titanium dioxide (TiO2 –
rutile), an ultrapure white nanopowder with an average particle size 250 nm, was
used with basalt/epoxy matrix and graphene platelet nanopowder.

2.2 Composite Preparation

Basalt fabrics were sprayed with 1 (N) NaOH solution and dried at room temperature.
Surface-treated basalt fabrics were then cleaned with deionized water to maintain
neutrality [25]. Finally, basalt fibres were made to dry in open air for 24 h at a
room temperature of 30 °C. The nano filler, graphene platelet nanopowder of 0.2%,
0.4% and 0.6% by weight were added to resin and stirred using a magnetic stirrer
of 600 rpm. ATH and zinc borate of 3% and 1%, respectively, by weight were then
added to resin as microfillers followed by magnetic stirring. Epoxy resin with fillers
was then mixed with hardener in the weight ratio of 10:1 and was mixed thoroughly
using a mechanical stirrer continuously. Vacuum bagging procedure was adapted for
fabrication of basalt fibre reinforced composite. Basalt fibre reinforced composite
excluding nanofiller was considered as control specimen. The composite laminates
were fabricated to the dimensions of 300 mm × 300 mm × 3 mm.
Similar procedure was repeated for graphene platelet nanopowder and titanium
dioxide nanopowder reinforced basalt epoxy composite laminate. The schematic
representation of fabrication process is illustrated in Fig. 1.
Various combinations of composite specimens are enlisted in Table 1.
Basalt/epoxy composite laminates with nano and microfillers were fabricated with
varying weight percentages of ATH as 0, 1%, 3% and 5% with a constant 1 wt% of
zinc borate as microfillers. GnP nano fillers were also included by 0, 0.2%, 0.4% and
0.6%, respectively. Another set of basalt/epoxy composite laminates were fabricated
with different nano fillers. Based on the previous work, 0.5 wt% of nano titanium
dioxide was kept constant and varying graphene platelet nanopowder as 0, 0.2, 0.4
and 0.6 wt%, respectively.

2.3 Surface Morphology

The morphology and surface textures of basalt fibre reinforced with nano graphene
platelet powder, titanium dioxide, alumina trihydrate and zinc borate were investi-
gated by scanning electron microscope (SEM). However, long cylindrical smooth
surface is the common morphological characteristics of fibres after treatment. Fail-
ure mechanisms related with dispersion of nano and microfillers in epoxy resin were
also detailed using SEM images.
Effect of Nano and Microfillers in Basalt/Epoxy Composites 423

Fig. 1 Schematic diagram of fabrication process

Table 1 Combinations for

Combination for composite % by wt of fillers as
composite specimen
specimen reinforcements
Epoxy + BF 0
Epoxy + BF + ATH + ZB 0, 1, 3 & 5 (ATH), 1(ZB)
Epoxy + BF + ATH + ZB + 0, 1, 3 & 5 (ATH), 1(ZB), 0.2
GnP (GnPs)
Epoxy + BF + ATH + ZB + 0, 1, 3 & 5 (ATH), 1(ZB),
GnP 0.4(GnPs)
Epoxy + BF + ATH + ZB + 0, 1, 3 & 5 (ATH), 1(ZB),
GnP 0.6(GnPs)
Epoxy + BF + TiO2 0.5 (TiO2 )
Epoxy + BF + TiO2 + GnP 0.5 (TiO2 ), 0.2(GnPs)
Epoxy + BF + TiO2 + GnP 0.5 (TiO2 ), 0.4(GnPs)
Epoxy + BF + TiO2 + GnP 0.5 (TiO2 ), 0.6(GnPs)
424 M. M. Metro and M. Selvaraj

2.4 Tensile Test

Tensile test samples were prepared according to ASTM D 3039, the international
standard. Specimens were tested using universal testing machine, Shimadzu AG–X,
of load capacity 50 kN. The cross head speed for tensile tests was maintained at
2 mm/min.

2.5 Flexure Test

Flexural test for composite specimen was performed with a cross head speed of
2 mm/min. Samples with and without GnPs were prepared according to ASTM
D 790 specifications and subjected to three-point bending load. The load cell was
equipped with 50kN.

2.6 Impact Test

Impact of composite specimens was tested using Charpy impact testing machine,
AIT 300-N. Specimens for impact test were prepared for ASTM D 256 standard.
The Charpy impact test machine was equipped with a 300 J hammer.

3 Results and Discussions

3.1 Tensile Strength

Basalt fibre reinforced composite specimens with multiscale fillers were subjected to
standard tensile test. Variations in tensile properties of composites based on addition
of nano and microfillers in matrix are discussed here. Incorporation of varied per-
centages of GnPs triggered the mechanical characterization of composites to certain
extent.
Stress–strain values of varying weight % of GnPs in multiscale composite explain
the deviation of strength after a sudden drop in load at 0.4 wt%. Tensile modulus of
composite specimen fabricated with varying wt% of GnPs, ATH and constant zinc
borate is shown in Fig. 2. It is clearly observed from Fig. 2 that tensile modulus of
composite specimen with wt% of 0.4 GnPs with a combination of 3 wt% (ATH) and
1 wt% (Zinc borate) named as A3Z1 marks highest value, whilst A0Z1 without GnPs
possesses the lowest as 18.97 MPa and 10.65 GPa, respectively. Incorporation of nano
Effect of Nano and Microfillers in Basalt/Epoxy Composites 425

Fig. 2 Tensile modulus 25

A0

T ensile modulus (GPa)

versus wt% of nano filler Z1
20 A1
Z1
15

10

0
0 % GnP 0.2% GnPs 0.4% GnPs 0.6% GnPs
Wt of nano fillers (%)

Table 2 Tensile test results of composite specimen with micro and nano fillers (A3Z1)
GnPs content (%wt) Tensile strength Tensile modulus Tensile strain (mm)
(MPa) (GPa)
Epoxy/Basalt 212 12.53 0.023
Epoxy/Basalt/ATH/ZB 216 14.12 0.027
0.2 234 15.76 0.029
0.4 244 18.97 0.030
0.6 219 16.81 0.028

filler along with microfillers enhances the tensile property of composite specimen.
Nano fillers being able to slide over one another in its structural arrangement create
a good interfacial bond with the matrix. Presence of microfillers in matrix grooves
the surface and widens gap between fibre and matrix.
Table 2 shows the obtained result in terms of tensile properties for A3Z1 alone
(i.e. 3% ATH with 1% zinc borate and varying GnPs as 0, 0.2%, 0.4%, 0.6% by
weight). Tensile strength and modulus of control specimen were 216 MPa and 14.12
GPa, respectively. An enhancement in tensile strength and modulus was observed for
0.4 wt% GnPs with an increment of 14.03% and 40.89%, respectively, than that of
control sample. It may be noted that tensile strength and modulus of hybrid multiscale
composites with GnPs were observed to have higher values than composites without
GnPs.
Weak Vander walls force lowers homogeneity of material due to the presence
of multiscale fillers. Elongation at break decreased dramatically with almost linear
tensile behaviour after 0.4% GnP. Tensile strength of multiscale composite specimen
increases with increase of nano graphene platelets up to a certain percentage and then
decreases with increase of it. Inclusion of GnPs up to 0.4 wt% exhibited a yielding
behaviour which shows typical plastic, and the sudden fall for 0.6 wt% of GnPs
exhibited a brittle behaviour. A decline in elastic modulus was observed, thereby
indicating softening of material with an increase in nano filler of the composites.
It was stated from previous study that the addition of GnPs at 0.1% by weight
reaches critical value [26]. Addition of graphene enhances toughness of epoxy with
an enhancement of 65% in fracture toughness [27, 28]. A decreasing trend for tensile
426 M. M. Metro and M. Selvaraj

strength was noticed after 0.4% of nano filler. This enhancement in tensile strength
may be due to the inclusion of multiscale fillers into the polymers. Stiffness of
composite specimen increases with 0.4 wt% of GnPs and decreases for 0.6 wt% of
GnPs. Elasticity decreases from the addition of 0.6 wt% of GnPs to the polymer.
However, there is a decrease in tensile strength and tensile strain values at break,
by 10% and 6.67%, respectively, for same loading of GnPs. SEM images of tensile
specimen of A3Z1 with 0.4 wt% of GnP are illustrated in Fig. 3.
From SEM image shown in Fig. 3, it is observed clearly that incorporation of nano
fillers has made interfacial bond between fibre and matrix closer, thus reducing the
gap. The surface roughness (Fig. 3a) is due to protrusion of microparticles on matrix.
Presence of microfillers may result in stress concentrations around its surface creat-
ing weak bondage in matrix and hence makes the matrix to crack. Agglomeration of
nanoparticles is also seen on the surface of composite laminates. These agglomera-
tions might have caused interfaces to weaken and in turn to fracture. Fibres visualized
without matrix coat in Fig. 3b is seen and may be due to error in fabrication. This
space created by improper fill of matrix leads the composite specimen to fail.
Table 3 enlists tensile strength and modulus of composite specimen fabricated with
0.5 wt% of nano titanium dioxide and varied wt% of (0, 0.2, 0.4, 0.6%) graphene
platelet nanopowder. It is observed that composite specimen with 0.4 wt% of GnPs
marked maximum tensile strength and modulus, whilst e minimum is with exclu-
sion of GnPs. Addition of 0.5 wt% of nano titanium dioxide has increased tensile
strength and modulus of composite by 21.85% and 41.09%, respectively, than that
of epoxy/basalt plain composite. Increase of 13.77% and 21.89% on tensile strength
and modulus with 0.2 wt% of GnPs than 0.4 wt% of GnPs has an increase in strain
rate with softening of composite. Strain at 0.4 wt% reaches ultimate and had sudden
decline for 0.6 wt% making the composite brittle.
Ease of mobility of nano fillers into matrix makes the interaction better than that of
multiscale fillers. Interfacial interaction of nano fillers with polymer forms a strong
bondage and hence strong matrix. Mobility of polymer-coated nano fillers interlocks

Fig. 3 a and b SEM images of tensile specimen (A3Z1)

Effect of Nano and Microfillers in Basalt/Epoxy Composites 427

Table 3 Tensile test results of specimen with nano TiO2 and GnPs
%wt of fillers Tensile strength Tensile modulus Tensile strain (mm)
(MPa) (GPa)
Epoxy/Basalt 212 12.53 0.023
Epoxy/Basalt/0.5 225 13.26 0.026
(TiO2 )
0.5 (TiO2 ) & 0.2 230 15.26 0.031
(GnPs)
0.5 (TiO2 ) & 0.4 264 19.01 0.035
(GnPs)
0.5 (TiO2 ) & 0.6 240 14.15 0.029
(GnPs)

into fabric consequently resulting in a strong interaction of fibre/matrix. Synergis-

tic effect of nanoparticles as filler into matrix has made composite stronger than
unfilled specimen [29, 30]. Inclusion of titanium carbide enhances tensile property
of composite [30].

3.2 Flexural Strength

Flexural test specimens were subjected to three-point bending tests. The flexural
performance of multiscale composite specimen fabricated by varying wt% of GnPs,
ATH with constant zinc borate was evaluated. Discrepancies in flexural modulus of
composite specimen are depicted in Fig. 4. An enhancement in flexural performance
of composite laminates was observed with an inclusion of 0.2 wt% nano filler with
microfiller filled resin than that of epoxy/basalt composite. Multiscale composite
with 3 wt% of ATH and 1 wt% of zinc borate in addition to 0.4 wt% of GnPs has

14

12 A0ZB1
Flexural modulus (GPa)

A1ZB1
10 A3ZB1
A5ZB1
8

0
0 % Gn P 0 . 2 % Gn P s 0 . 4 % Gn P s 0 . 6 % Gn P s

Wt . Of nano fillers (%)

Fig. 4 Flexural values of composite specimen with micro and nano fillers
428 M. M. Metro and M. Selvaraj

Table 4 Flexural values of composite specimen with micro and nano fillers (A3Z1)
% weight of GnPs Flexural strength Flexural modulus Flexural strain (mm)
(MPa) (GPa)
Epoxy/Basalt 208.92 4.03 0.022
Epoxy/Basalt/ATH/ZB 243.24 7.05 0.028
0.2 329.73 9.93 0.026
0.4 395.45 12.45 0.027
0.6 295.54 10.61 0.024

good effect on flexure modulus of composite specimen. A decremental effect in

flexure strength and modulus after addition of 0.4% GnPs reveals that composite has
better interfacial bond at 0.4% GnPs, thus making composite stronger. It is clearly
observed that it makes the material more ductile beyond the addition of 0.4% GnPs
by weight.
Delamination of composite formed on surface of specimen with 0.2 wt% GnPs is
more than do 0.4 and 0.6 w% of GnPs composite laminates. The action of this type
of behaviour features the brittle nature of the composite. This delamination might
have formed due to weak interfacial adhesion of matrix and fibres.
Flexural values of tested composite specimen with micro and nano fillers for ATH
and zinc borate with weight ratio of 3:1 and varied wt% of GnPs are enlisted in Table 4.
Basalt/epoxy composite with microfillers attained more strength than basalt/epoxy
composites by 15%. Presence of multiscale fillers increases surface roughness with
the reaction of intraparticle fusion. Nano fillers increase the smoothness of composite
surface.
Flexural modulus of multiscale composite has an incremental value with 0.4 wt%
of GnPs to matrix. Addition of microfillers with nano matrix has improved flexural
strength and modulus of composite by 18.13% and 22%, respectively, for 0.4 wt%
of GnPs than that of composite without nano filler. Incorporation of nano material
beyond 0.4 wt% makes the material more brittle. Inclusion of 0.6% of GnPs by
weight has a decline in strength and modulus by 28.92% and 15.95%, respectively,
from the values of composite specimen with maximum value. It was mentioned by
earlier research that addition of 0.1 wt% of GnPs in epoxy with basalt fabric improved
strength resulted in further decrease in interfacial bonding [26].
SEM images of flexural specimen describe the fracture surface as given in Fig. 5.
Orientation of nano and microfillers along basalt fabric and resin indicates nature
of fracture surface. Distribution of nano fillers formed a neat and smooth surface.
Smooth delaminated surface is seen in Fig. 5a. Weak matrix interface between fibres
causes matrix failure which in turn pulls out the fibres as shown in Fig. 5b. Non-
uniformity in distribution of nano fillers clearly depicts poor interfacial bond between
fibre and matrix which in turn ruptures fibre. Fibre pull outs are initiated due to weak
interfacial bonding between epoxy and fibre. It is evident that more number of fibre
pull outs result in the weak and brittle specimen.
Effect of Nano and Microfillers in Basalt/Epoxy Composites 429

Fig. 5 a and b SEM images of flexural test specimen

SEM images of flexural specimen of nano graphene platelets and TiO2 are illus-
trated in Fig. 6. The matrix that bounds fibre itself demonstrates clear and smooth
matrix formed with inclusion of different nanoparticles. Presence of nanofillers
grounds smooth matrix.
Flexural test results of composite specimen fabricated with 0.5 wt% of TiO2 with
varying wt% of GnPs are tabulated in Table 5. It is evident that specimen with 0.4
wt% has marked maximum, whilst epoxy/basalt scored the minimum. Increase in
strain rate from plain specimen (epoxy/basalt) till composite specimen with 0.4 wt%
GnPs measures higher tensile and flexure properties. Composite has proved its yield
capacity till 0.4 wt% of GnPs with a limit. The specimen bounds its stiffness to
0.4 wt% of GnPs into 0.5 wt% of TiO2 filled epoxy/basalt.

Fig. 6 SEM images of flexural test specimen fabricated with nano GnPs and TiO2
430 M. M. Metro and M. Selvaraj

Table 5 Flexural test results for composite specimen with TiO2 and GnPs
% weight of GnPs Flexural strength Flexural modulus Flexural strain (mm)
(MPa) (GPa)
Epoxy/Basalt 208.92 4.03 0.022
Epoxy/Basalt/0.5 236.66 7.98 0.025
(TiO2 )
0.5 (TiO2 ) & 0.2 335.23 11.87 0.026
(GnPs)
0. 5(TiO2 ) & 0. 400.64 14.41 0.028
4(GnPs)
0.5 (TiO2 ) & 0.6 358.75 12.69 0.025
(GnPs)

3.3 Impact Strength

Energy absorbed and lost by specimen during Charpy impact test gives impact resis-
tance and impact strength of the specimen. The striking velocity of pendulum is
2.45 m/s. Impact strength of specimen subjected to impact energy of 300 J at room
temperature is shown in Fig. 7. It is evident from Fig. 7 that incorporation of 0.4 wt%
GnPs increased the impact strength of specimen with a reduction on further addition
of GnPs. Addition of 0.6 wt% of GnPs declined in its impact strength. Impact strength
of composite specimen fabricated with nano fillers is higher than composite with nano
and microfillers. This might have been effected from interfacial interaction of fillers
which enhances interfacial bonding between fibre and matrix. Impact strength of
composite falls from 1517 to 1505 kJ/m2 after the addition of 0.6 wt% GnPs into
nano TiO2 matrix. The composite fabricated with multiscale fillers recorded a value
of 1507 kJ/m2 for 0.4 wt% of GnPs from 1435 kJ/m2 for plain epoxy/basalt compos-
ite. Though the impact strength values are similar, further addition of GnPs lowers
impact strength and thus leads to delamination.

Fig. 7 Impact strength of 1530

multiscale and nano fillers 1520

A3ZB1
Impact strength (kJ/m2)

1510 TiO2-GnP

1500

1490

1480

1470

1460
0 % GnP 0.2% GnPs 0.4% GnPs 0.6% GnPs
wt % of GnP
Effect of Nano and Microfillers in Basalt/Epoxy Composites 431

4 Conclusion

The experimental investigations on mechanical properties of basalt fibre reinforced

composite were studied. The following are highlighted as remarks.
1. Incorporation of nano graphene platelets and microfillers such as aluminium
trihydrate, zinc borate improved mechanical properties of multiscale composite
with 0.4 wt% of GnPs.
2. Addition of 0.4 wt% graphene nano platelets to nano titanium dioxide enhances
tensile strength of the composite. Increment in flexural strength signifies the
strength of composite.
3. The impact strength of multiscale and nano composite remarks with 0.4 wt% of
graphene nano platelets for higher loading.

References

1. Dhand V, Mittal G, Rhee KY (2015) A short review on basalt fiber reinforced polymer
composites. Compos Part B Eng 73:166–180. https://doi.org/10.1016/j.compositesb.2014.
12.011
2. Espano (2013) Investigation of the effect of different silane coupling agents on mechanical
performance of basalt fiber composite laminates with biobased epoxy matrices. Polym Compos
34
3. Dasari BL, Morshed M, Nouri JM (2018) Mechanical properties of graphene oxide reinforced
aluminium matrix composites, pp 136–144
4. Muhammad R, Fusheng P, Aitao T (2014) Effect of graphene nanoplatelets addition on
mechanical properties of pure aluminum using a semi-powder method, pp 101–108
5. Park JH, Jana SC (2003) The relationship between nano- and micro-structures and mechanical
properties in PMMA-epoxy-nanoclay composites. Polymer (Guildf) 44:2091–2100. https://
doi.org/10.1016/S0032-3861(03)00075-2
6. Koricho EG, Khomenko A, Haq M (2015) Effect of hybrid (micro- and nano-) fillers on
impact response of GFRP composite. Compos Struct 134:789–798. https://doi.org/10.1016/j.
compstruct.2015.08.106
7. Sengupta R, Bhattacharya M, Bandyopadhyay S, Bhowmick AK (2011) A review on the
mechanical and electrical properties of graphite and modified graphite reinforced polymer
composites. Prog Polym Sci 36:638–670. https://doi.org/10.1016/j.progpolymsci.2010.11.003
8. Weihua K, Yuuki H, Lei Hua YI (2008) Thermal and mechanical properties of a poly(« -
caprolactone)/graphite oxide composite. J Appl Polym Sci 107:1395–1400
9. Huang G, Gao J, Wang X (2012) How can graphene reduce the flammability of polymer
nanocomposites? Mater Lett 66:187–189. https://doi.org/10.1016/j.matlet.2011.08.063
10. Papageorgiou DG, Kinloch IA, Young RJ (2017) Mechanical properties of graphene and
graphene-based nanocomposites. Prog Mater Sci 90:75–127. https://doi.org/10.1016/j.pmatsci.
2017.07.004
11. Lee JH, Rhee KY, Park SJ (2010) The tensile and thermal properties of modified CNT-reinforced
basalt/epoxy composites. Mater Sci Eng, A 527:6838–6843. https://doi.org/10.1016/j.msea.
2010.07.080
12. Liang M, Wong KL (2017) Study of mechanical and thermal performances of epoxy resin filled
with micro particles and nanoparticles. Energy Procedia 110:156–161 https://doi.org/10.1016/
j.egypro.2017.03.121
432 M. M. Metro and M. Selvaraj

13. Kim MT, Rhee KY, Park SJ, Hui D (2012) Effects of silane-modified carbon nanotubes on
flexural and fracture behaviors of carbon nanotube-modified epoxy/basalt composites. Compos
Part B 43:2298–2302. https://doi.org/10.1016/j.compositesb.2011.12.007
14. Lynch TJ, Chen T, Riley D (2003) Advances in ATH benefit Aluminium hydroxide is a widely
used flame retardant for reinforced plastics. Materials E, 44–46
15. Elbasuney S (2017) Novel multi-component flame retardant system based on nanoscopic.
Powder Technol 305:538–545. https://doi.org/10.1016/j.powtec.2016.10.038
16. Schubert DM, Alam F, Visi MZ, Knobler CB (2003) Structural characterization and chemistry
of the industrially important zinc borate, Zn [B3 O4 (OH)3 ]. Chem Mater 4:866–871
17. Bourbigot S, le Bras M (2003) European meeting on fire retardancy and protection OF
Materials, pp 327–335
18. Chen R, Gong J, Jiang Y (2018) Halogen-free flame retarded cold-mix epoxy asphalt binders:
rheological, thermal and mechanical characterization. Constr Build Mater 186:863–870. https://
doi.org/10.1016/j.conbuildmat.2018.08.018
19. Tanwyn M, Khushairi M, Sharif S (2015) Evaluation of mechanical properties of filled epoxy
composite for improving mould performance—a review. Appl Mech Mater 735:13–18. https://
doi.org/10.4028/www.scientific.net/AMM.735.13
20. Bhagyashekar MS (2014) J Reinf Plast Compos, 28–42. https://doi.org/10.1177/
0731684408095034
21. Chung S, Im Y, Kim H, Jeong H (2005) Evaluation for micro scale structures fabricated using
epoxy-aluminum particle composite and its application
22. Tilbrook MT, Moon RJ, Hoffman M (2005) Study on the mechanical properties of alumina—
epoxy composites with an interpenetrating network structure. Mater Sci Eng A 393:170–178.
https://doi.org/10.1016/j.msea.2004.10.004
23. Ma L, Zhu Y, Li X (2018) The architecture of carbon fiber-TiO2 nanorods hybrid structure in
supercritical water for reinforcing interfacial and impact properties of CF/epoxy composites.
Polym Test 66:213–220. https://doi.org/10.1016/j.polymertesting.2018.01.021
24. Yi C, Bagchi S, Dmuchowski CM (2018) Direct nanomechanical characterization of carbon
nanotube—Titanium interfaces. Carbon N Y. https://doi.org/10.1016/j.carbon.2018.02.069
25. Lipatov YV, Gutnikov SI, Manylov MS (2015) High alkali-resistant basalt fiber for reinforcing
concrete. J Mater 73:60–66. https://doi.org/10.1016/j.matdes.2015.02.022
26. Bulut M (2017) Mechanical characterization of Basalt/epoxy composite laminates containing
graphene nanopellets. Compos Part B 122:71–78. https://doi.org/10.1016/j.compositesb.2017.
04.013
27. Silva F, Njuguna J, Sachse S (2013) The influence of multiscale fillers reinforcement into impact
resistance and energy absorption properties of polyamide 6 and polypropylene nanocomposite
structures. Mater Des 50:244–252. https://doi.org/10.1016/j.matdes.2013.02.041
28. Prasad V, Joseph MA, Sekar K (2018) Investigation of mechanical, thermal and water absorption
properties of flax fibre reinforced epoxy composite with nano TiO2 addition. Compos Part A
115:360–370. https://doi.org/10.1016/j.compositesa.2018.09.031
29. Mahesha CR, Mohan N, Rajesh M (2017) Role of nanofillers on mechanical and dry sliding
wear behavior of basalt-epoxy nanocomposites. Mater Today Proc 4:8192–8199. https://doi.
org/10.1016/j.matpr.2017.07.161
30. Thotsaphon T, Katsuyoshi K, Hisashi I (2019) Titanium matrix composite reinforced with
in-situ formed TiC using carbon black nano particles via a wet process. Trans JWRI 38:13–17
A Feasibility Study on Microwave
Joining of GFRP Composite Pipes
with Interlayer Coupling Agents

K. Rajkumar, M. Dhananchezian, and S. Aravindan

Abstract This paper investigates the feasibility of energy, economic, and envi-
ronment way to joining the GFRP pipes. These GFRP pipes are used to transport
chemicals, fuel, oil, and natural gases. Microwave joining is economic and energy-
efficient method for bonding thermoset composite with suitable intercoupling layers.
Microwave joining was carried out in 2.45 GHz furnace with suitable susceptor and
various coupling agents such as molten polypropylene and polyethylene, epoxy resin
liquid, and sodium silicate. These interlayered GFRP pipes were tested to evaluate
their lap shear strength and bond strength. Among the various coupling agents used,
polyethylene interlayered coupling agent joints exhibited a good shear strength. The
abstract should summarize the contents of the paper in short terms, i.e., 150–250
words.

Keywords Microwave joining · GFRP · Composite bonding · Inter layer

1 Introduction

In the electromagnetic spectrum, microwave energy is in the frequency range of 0.3–

30 GHz. Microwaves are used to process the materials for many years. Its application
is wide in material-processing industries in particular such as joining of thermoplastic
composite (TPC) and joining of ceramic to ceramic [1–4]. The joining of thermo-
plastics using microwave heating has been exploited much in industry for a number
of years. Thermoset materials having highly desirable properties are now readily
available at economically attractive prices; use of such materials for the production
of large-shaped articles has been highly restricted. This is particularly true if the arti-
cles have complex shapes or reinforced with glass fibers and too large to be molded

K. Rajkumar · M. Dhananchezian (B)

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, India
e-mail: dhananchezianm@ssn.edu.in
S. Aravindan
Department of Mechanical Engineering, Indian Institute of Technology, Delhi, India

© Springer Nature Singapore Pte Ltd. 2021 433

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_38
434 K. Rajkumar et al.

economically. An attempt has been made on solving this problem by microwave

joining of glass fiber reinforced plastic pipe (GFRP). GFRP, an advanced compos-
ite material, is widely used in a variety of applications, including aircraft structural
member, robots, and transportation of corrosive liquid [5, 6]. This study focuses on
the effect of microwave exposure time, temperature, and different coupling agents
on joining of GFRP pipes.

2 Materials and Microwave Joining

GFRP pipes were made using the resin composition of isophthalic (50%) and
vinylester (50%). The volume fraction of the materials is 65:35 (resin/glass). The
hand lay-up technique was used to manufacture the GFRP pipe with the dimensions
of 30 mm and 55 mm of inner and outer diameters, respectively. The glass fiber
orientation angle in matrix was 90°.
The term “joining” refers to a zone bounded by the GFRP substrates, in which
the entanglement of polymer chains has progressed beyond simple, mechanical,
and chain entanglement. The process comprises generating sufficient heat energy at
interface of an assembly of thermosets substrates to bond them together using an
“interface strip,” referred to as a “coupling agent.” Prior to joining, the surfaces of
the pipes were cleaned, and surface contaminations were removed by fine SiC paper.
The different coupling agents used were liquid sodium silicate, epoxy resin, molten
polypropylene, and polyethylene. Pipe to be joined was placed in multimode cavity
of industrial microwave furnace (2.45 GHz and 3.2 kW). Normally, GFRP thermoset
substrates are transparent to the microwave radiation, and the only part that can be
heated up is the coupling agent implanted at the joint. The microwave power and
heating rate were 320 W and 10 °C/min, respectively. The temperature range fixed
was based on the glass transition temperature T g of the coupling agent. The lap shear
strength of the processed joints was evaluated. Effect of microwave exposure was
assessed through buckling strength testing of the GFRP pipes.

3 Results and Discussion

Initially, liquid sodium silicate was used as the coupling agent which was applied
between the surfaces to be joined. The microwave power was adjusted to 320 W.
The holding time, also referred to as “exposure time at maximum temperature,”
was varied between 5 and 10 min. Table 1 shows the results of microwave joining
of GFRP pipe with different coupling agents. The samples prepared using sodium
silicate as coupling agents gave a successful weld at 150 °C. When the temperature
was increased to 200 °C, the weld exhibited bad strength, but joints have been made.
But once the temperature was increased to 250 °C, both the pieces separated out on
the application of slightest additional force and hence the joint failed at 250 °C. The
A Feasibility Study on Microwave Joining of GFRP Composite … 435

Table 1 Microwave heating history and different coupling agent

Coupling Maximum Holding Heating rate Weld Shear
agent temperature time (min) (°C/min) success strength
(°C) (MPa)
Liquid sodium 150 5 10 Yes 40.8
silicate
Liquid sodium 150 10 10 Yes 40.8
silicate
Liquid sodium 200 5 10 Yes –
silicate
Liquid sodium 200 10 10 Yes –
silicate
Liquid sodium 250 5 10 No
silicate
Liquid sodium 250 10 10 No
silicate
Epoxy resin 150 10 10 No
Epoxy resin 200 10 10 No
Epoxy resin 250 10 10 No
Polypropylene 150 10 10 No
Polypropylene 200 10 10 No
Polypropylene 250 10 10 No
Polyethylene 250 10 10 No
Polyethylene 350 2 12 No
Polyethylene 300 2 12 Yes 122.4

variation in the holding time also had not affected the failure at 250 °C. Epoxy resin
could not provide the joint since there was no bonding at all. Polypropylene paste
was applied between the joints to be made. The samples failed to produce joint at
any of the temperatures tested. Although there was some bonding between the two
surfaces of joint but with a slight rotation, the joints separated away. Polyethylene
thin film was then tested as a coupling agent. The polyethylene was cut in the shape
of rings, and a double film was placed at the junction of two thermoset pipes.
The temperature was varied from 250 to 350 °C with 2–10 min holding time.
The samples at 250 and 350 °C failed on application of slight force. Same samples
processed at 300 °C and exhibited good strength. Figure 1 shows the successful
joint made by using polyethylene as a coupling agent. Heat generated along the
bond interface, i.e., coupling agent is transferred from inside to the outer surface. At
the softening temperature, the coupling agent softens and joins together. When the
temperature of coupling agent increases, it absorbs/couples with microwave energy.
When amount of microwave energy is absorbed, heat generation also is increased.
This cumulative effect leads to burnouts as noticed on samples. Samples prepared
with sodium silicate at 250 °C suffered due to burnouts at the outer layer of the
436 K. Rajkumar et al.

Fig. 1 Joint prepared with

using polyethylene as
coupling agent

Fig. 2 Burnouts due to high

temperature heating

substrates. Samples prepared with polyethylene at 350 °C suffered due to the burnouts
on the whole of the outer surface of the substrate. Figure 2 shows the burnout sample
of joint with polyethylene. From Table 1, it is understood that higher temperature
and longer duration of microwave heating degrade the material properties. A very
narrow weldability window is available, for joining these thermoset pipes.

3.1 Microstructure Analysis and Mechanism of Joint

Formation

Figure 3a shows the SEM image of the joint made using liquid sodium silicate was
a coupling agent (150 °C and 10 min). The SEM image shows well-defined bound-
ary between the two pipes. The amorphous glass phase (sodium silicate) absorbs
microwaves, and thus, a selective heating was observed [4]. Sodium silicate is con-
verted into solid films or bonds between the two GFRP pipes by evaporation of water
(dehydration). During microwave exposure, the viscosity of the glass drops down to
A Feasibility Study on Microwave Joining of GFRP Composite … 437

a b
Pipe 1

Coupling agent Pipe 2

Coupling agent
10 µm
10 µm

c
Coupling agent

Pipe 2
20 µm

Fig. 3 SEM image of successful microwave joint with coupling agent a liquid sodium silicate,
b polyethylene, c magnified image of polyethylene

the lower order and spreading of this material over the interface. Microwave exposure
makes it easier to inter-diffusion of coupling agent over the GFRP pipes. Further-
more, attainment of crystalline state at the processing temperature can also facilitate
effective bond. Sodium silicate films are having an affinity to moisture pickup, and
then, degradation of joints is also possible. However, microwave heating can com-
pletely remove the water particles from the sodium silicate which is helpful to form
good bond.
In the case of polyethylene coupling agent, the interface is observed very narrow
as seen in Fig. 3b. To understand more clearly, the interface of the joint is magnified,
and the corresponding image is shown in Fig. 3c. It is interesting to observe that there
is no presence of voids or discontinuities at the interface. The role of the coupling
agent is to absorb the microwave energy, generate local heat at the interface at the butt
faces of thermoset material. The dielectric property of polymers to external electric
field depends mainly on its polarity. Moreover, in unpolar polymers (polyethylene),
only induced (electron or ion) polarizations play a role when interact with electro-
magnetic wave [7], whereas in polar polymer orientation, dipole polarization can
contribute for heating. During microwave heating, due to localized interdiffusion
occurred [8] between the coupling agent to adjacent substrate materials the bond
formation occurred.
438 K. Rajkumar et al.

3.2 Joint Shear Strength and Compressive Strength

After the elimination of the failed samples, successful joints were tested for their joint
strength. The mechanical strength of the joint was tested through shear strength.
The successful joint was held in position with the help of the designed fixture.
From Table 1, it can be understood that sodium silicate and polyethylene are typical
coupling agents for bonding. Successful joints made from coupling agent such as
sodium silicate with 150 °C and 5 min and 150 °C and 10 min exhibited same shear
strength of 40.8 MPa. Polyethylene coupling agent with 300 °C and 2 min exhibited
strength of 122.4 MPa. Polyethylene produced better joint at 300 °C, and the shear
strength was quite high when compared with that of sodium silicate. In fact, the
shear strength of the former was almost triple than that of the latter. At an exposure
time of 10 min and higher temperature 350 °C, the substrate material was weakened,
and when it was subjected to a shear strength test, failure occurred at the substrate
material. The surface skin also been affected during the microwave exposure when
exceeding the optimal parameters. The buckling test was carried out on microwave
exposed at 150 °C and 10 min and unexposed samples. For carrying out the buckling
test, the sample was placed in the compressive testing machine, and the load was
steadily applied. Both the samples showed the same load of failure (2 kN). It is
understood from this compressive strength data that the microwave exposure had not
induced any detrimental effect on the GFRP pipe.

4 Conclusions

Microwave heating technology is successfully demonstrated for joining of ther-

moset substrate with coupling agent. Coupling agent such as liquid epoxy resin
and polypropylene had not yielded positive results. Samples got successfully joined
using polyethylene and liquid sodium silicate coupling agent. Joints made using
polyethylene exhibited maximum shear strength compared to the liquid sodium sil-
icate samples. Buckling strength of microwave exposed and unexposed thermoset
samples had not exhibited any variation in compressive strength, which confirms that
no thermal degradation occurred.

References

1. Ku HS, Siores E, Ball JAR, Horsfield B (1999) Microwave processing and permittivity
measurement of thermoplastic composite at elevated temperature. J Mater Process Technol
89:419–424
2. Siores E (1994) Microwave technology for welding and joining. Mater World 2(10):526
3. Liu F, Turner I, Siores E, Groombridge P (1996) A numerical and experimental investigation
of the microwave heating of polymer materials inside a ridge waveguide. J Microwave Power
Electromagn Energy 31:71–82
A Feasibility Study on Microwave Joining of GFRP Composite … 439

4. Aravindan S, Krishnamurthy R (1999) Joining of ceramic composites by microwave heating.

Mater Lett 38:245–249
5. Ramkumar J, Aravindan S, Malhotra SK, Krishnamurthy R (2004) An enhancement of the
machining performance of GFRP by oscillatory assisted drilling. Int J Adv Manuf Technol
23:240–244
6. Aravindan S, Naveen Sait A, Noorul Haq A (2008) Machinability study of GFRP pipes using
statistical techniques. Int J Adv Manuf Technol 37:1069–1081
7. Svorcik V, Ekrt O, Rybka V, Liptak J, Hnatowicz V (2000) Permittivity of polyethylene and
polyethyleneterephtalate. J Mater Sci Lett 19:1843–1845
8. Rajkumar K, Aravindan S, Kulkarni MS (2012) Wear and life characteristics of microwave
sintered copper-graphite composites. J Mater Eng Perform 21:2389–2397
Comparative Evaluation of Mechanical
Properties of GFRP and Polymer Hybrid
Composite

R. Raja, Sabitha Jannet , Allen Varughese, and Joby George

Abstract Synthetic fiber polymer laminates are becoming competitive material for
engineering and structural applications. In this work, an attempt is made to syn-
thesize a laminate with natural as well as synthetic fibers to get the advantage of
cost-effective natural fibers. Coir is a natural fiber available in significant amount
in south India. The present work aims at manufacturing a strong, cost effective,
light and low impact on environment hybrid material. Both GFRP laminates and
hybrid (GFRP + coconut Coir) laminates were fabricated by hand layup method and
compression molding. Physical and mechanical properties of this laminates such as
tensile strength, impact strength and density were evaluated. Microscopy images of
this laminates were arrived at through scanning electron microscope. Even though
the performance of hybrid (GFRP + coconut Coir) laminates is relatively lesser than
GFRP laminates, these hybrid laminates may be adopted in the industries where cost
is an important design factor.

Keywords GFRP · Coir · Impact strength · Density · SEM · Tensile strength

1 Introduction

Biocomposites are an attractive new age alternative for synthetic composites [1]. The
increase in fiber reduces ductility but when treated with epoxy resin produces inho-
mogeneity however hardness increases [2]. Energy absorption and failure modes
were significantly affected by strain rate [3]. Coir fibers obtained from disregarded
coconut shells were used for reinforcement up to 80%. Up to 50% strength increased
further addition resulted in agglomeration [4]. Highest tensile strength of 30.7 MPa
was reported for cotton unsaturated polymer composite and tensile strength increased
linearly [5]. It was found that 30 volume percentage of coir reinforcement displayed
ballistic performance as good as Kevlar [6]. The coir reinforced epoxy matrix treated

R. Raja · S. Jannet (B) · A. Varughese · J. George

Mechanical Engineering Department, Karunya Institute of Technology & Sciences, Coimbatore
641114, India
e-mail: sabithajannet@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 441

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_39
442 R. Raja et al.

with alkali showed 15% increase in impact strength when compared with untreated
samples [7]. Coconut shell powder as filler with fiber reinforcement in epoxy exhib-
ited reduction in wear rate due to the filler powder [8]. Epoxy resin was reinforced
with untreated coconut fiber leading to increase in tensile strength and higher impact
strength which was influenced by the high aspect ratio of the fibers [9]. The inclu-
sion of cow dung and corn starch fibers increases the friction coefficient. Overall, 6%
weight cow dung showed superior wear properties and thus proved that cow dung
fiber is efficient in improving the friction properties of composites. [10] There was
improved mechanical properties in treated fiber compared with raw fibers [11]. There
was 25% and 28% increase in flexural and tensile strength, respectively, after the coir
was treated in sodium carbonate for 96 h. Coir polyester achieved superior mechan-
ical properties compared to coir epoxy derivatives composites. The solution also
led to increase in the superfluous fiber roughness attributing to increased mechan-
ical properties [12]. Coconut fibers pre-treated with alkali solution such as NaOH
exhibited superior mechanical properties. And the length of the fibers subsequently
affected the tensile properties of the composites [13]. The hygroscopic swelling and
absorption effect on the mechanical properties of biocomposites revealed that the
direction of the fiber played a major role in the improved mechanical properties
[14]. Surface modification of fibers resulted in improved mechanical properties [15].
The flexural strength and tensile properties improved with longer coir [16]. The
effect of immersion of coir in water leads to a drastic decrease was preceded with
modulus stabilization [17]. The interaction effect between the curing time and coir
reinforced volume decided the tensile strength properties of coir fiber reinforced
epoxy composites [18].

2 Materials and Methods

Epoxy resins have to be cured for a minimum of 24 h at room temperature. 30 volume

percentage of fiber reinforcement with polymer based materials are popular with hand
layup method. Volume of fiber and matrix was first calculated. Epoxy content was of
30%, and glass fiber was 70%. Epoxy resin LY556 (density 1.15–1.20 g/cm3 at 25 0 C)
and hardener HY951 (density 1 g/cm3 at 20 °C) were used. A ratio of 10:1 by weight
of resin and hardener were employed. OHP sheet is placed beneath the wooden mold.
OHP sheet is used because it is a polymer plastic, so the resin will not stick to the
sheet. The sides of the mold were covered using a tape. The glass fiber sheets were
laid into the mold layer by layer, and the resin was applied gradually using a brush.
This process was repeated until the required thickness was achieved. The mold was
then covered using another OHP sheet, and weight was applied on it. Curing was
done for first 24 h at room temperature and was then exposed to sunlight. Hybrid
is made using the same method as hand layup GFRP. Alternate layers of coir and
GFRP were laid, and resin was applied. Hybrid sample consisted of 30% epoxy, 35%
glass fiber and 35% coconut fiber. Compression molding refers to a process where
up to 2000 psi pressure and temperature from 120 to 200 °C is used in order to cure
Comparative Evaluation of Mechanical Properties of GFRP … 443

the part. This technique is used to rapidly cure large quantities fiberglass-reinforced
polymer parts. Matched metal dies are placed in large molding press. The materials
to be molded are placed between the hot dies and pressed. Fiber arrangement and
volume percentage are similar to the hand layup GFRP. Only difference is that the
fiber and resin are compressed in between hot dies at 200 °C. Fiber arrangement and
volume percentage are similar to the hand layup hybrid. Only difference is that the
fiber and resin are compressed in between hot dies at 200 °C. The size of the specimen
for the impact test was as per ASTM A370. Brinell hardness test was carried out on
a specimen with measurements 50 * 50 * 10 mm.

3 Results and Discussion

3.1 Density

From Fig. 1, it can be seen that the density of GFRP and hybrid produced using
compression molding method is less dense when compared to hand layup method.
It may be due to the removal of excess epoxy through vaporization or through side
leaks when pressure is applied. It can therefore be inferred that GFRP and hybrid
composite fabricated using compression molding method are suitable for production
of low weight structural models and components (Fig. 2).

Fig. 1 Effect of GFRP and hybrid composite on the density

444 R. Raja et al.

Fig. 2 Tensile samples according to ASTM D3039 after the test. a Hand layup GFRP. b Hand
layup hybrid. c Compression molding hybrid. d Compression molding GFRP

3.2 Tensile Test

The tensile test of the composites was performed as per the ASTM D3039
(200 * 15 * 6) standards. The test was done using a universal testing machine.
The above figure shows tensile stress samples after the test.
From Fig. 3, it is evident that the tensile strength of compressed GFRP is higher
than that of hand layup method, this is because during compression the gap between
different layers is reduced due to the high pressure applied as well as the heat treatment
given, and this is also true when we look in hybrid composite. The hybrid composite
possesses less tensile strength due to that the GFRP slab contains more glass fibers
when compared to hybrid composite, and from the experiment, it is provident that the
tensile strength increases with increased glass fiber content as glass fiber is stronger
than natural fiber (coir). Therefore, we could conclude that the composite materials
fabricated by compression molding method possess more tensile strength than those
fabricated with usual hand layup method.
Comparative Evaluation of Mechanical Properties of GFRP … 445

Fig. 3 Effect of GFRP and hybrid on the ultimate tensile strength (MPa)

3.3 Brinell Hardness

The typical test uses a 10 mm (0.39 in) diameter steel ball as an indenter with a
1000 kgf force. For softer materials, a smaller force is used; for harder materials, a
tungsten carbide ball is substituted for the steel ball (Fig. 4).
From Fig. 5, it is evident that the indentation of compressed GFRP is lower
than that of hand layup method, this is because during compression the gap between
different layers is reduced due to the high pressure applied as well as the heat treatment
given, and this is also true when we look in hybrid composite. Therefore, we could
conclude that the composite materials fabricated by compression molding method
possess more hardness than those fabricated with simple hand layup method.

3.4 Charpy Impact Test

The specimens were cut as per ASTM A370 standard. A V-notch is provided at
28 mm, i.e., at half the length of the specimen. The depth of the notch was around
2 mm, and 10 * 10 was its width and thickness, respectively.
From Fig. 6, it can be noticed that the impact strength of compression molding
is found to be way higher than that of the hand layup method, and this is due to
better adhesiveness of different layers in composite slabs. The compression molding
uses pressure as well as temperature to give better compression of the material,
446 R. Raja et al.

Fig. 4 Experimental values of Brinell hardness test obtained in this study are plotted in graph

Fig. 5 Charpy impact test specimens a Hand layup GFRP. b Hand layup hybrid. c Compression
hybrid. d Compression GFRP

and as a result, the bonding between various atom particles increases on comparing
with hand layup method. Among compression molded composite, the glass fiber
composite provides better impact strength than hybrid composite because the fiber
content in GFRP composite is found to be higher as well as the strength of glass fiber
Comparative Evaluation of Mechanical Properties of GFRP … 447

Fig. 6 Effect of fabrication technique and GFRP and hybrid polymer on the impact strength

strand is better when compared with a single coir fiber strand. Therefore, we can
conclude that the compressed glass fiber composite possesses more impact strength
when compared with other three samples.
Failures in composites are mainly due to voids, improper fiber bonding, poor
adhesiveness between the fiber and matrix, etc. The SEM images of the four samples
were taken at a magnification of 300× without any gold sputtering. The size of the
image measured 50 microns. The surface topography of each sample was studied,
and the conclusion obtained is as follows. In Fig. 7a and Fig. 7b, lots of voids are
visible. Weak interface between the fiber and matrix due void leads to premature
failure. These weak points lead to stress concentrations which cause de-bonding of
fiber. Cracks propagate from voids to voids which eventually lead to matrix failure,
fiber failure and ultimate failure of composites. Unlike the hand layup samples, the
compression molding processed samples as shown in Fig. 7c, d have higher interfacial
bonding between fire and matrix which will improve the ability of composite material
to undergo loading. The matrix clings to individual strands of fibers which suggest
that the fiber and matrix interfacial strength are exceptionally high. Thus, it can be
concluded that compression molded composites have higher bond strength and can
withstand higher loads without failing when compared to hand layup method.
448 R. Raja et al.

a b

c
d

Fig. 7 SEM images of different samples. a SEM image of hand layup GFRP 1-Fiber breakage.
b SEM image of hand layup hybrid. c SEM image of compression GFRP, 1-Fiber breakage. d SEM
image of compressed hybrid 2-voids

4 Conclusion

In the present investigation, a glass fiber coir hybrid composite was fabricated using
hand layup and compression molding method, and its mechanical properties were
tested. It was observed that compression molding showed enhanced mechanical prop-
erties when compared to hand layup. The effect of the addition of coconut fibers has
been studied, and the following conclusions have been drawn:
The tensile strength of hybrid is 44% of GFRP. Tensile strength of compression
molded composite is slightly higher than hand layup. Hybrid is 87% as hard as
GFRP, and when fabricated using compression molding, it showed slight increase in
hardness. Compression molded hybrid was 92% as tough as the compression molded
GFRP according to the impact test results.
Compression molding had also significantly raised the toughness of the compos-
ites. It was observed from the SEM images that the compression molded composites
had better interfacial bonding between matrix and fibers.
Comparative Evaluation of Mechanical Properties of GFRP … 449

The mechanical properties of hybrid are obviously lower than GFRP. But the
addition of coconut fibers leads to cost and weight reduction of the regular glass fiber
reinforced composites. Cost and strength of materials should always be maintained
at an optimum level. In this case, hybrid materials are able to provide reasonable
strength compared to purely synthetic materials at a lower cost, reduced weight and
also efficient usage of natural eco-friendly resource.

References

1. Khan A, Ahmad MA, Joshi S (2015) A systematic study for electrical properties of chemically
treated coir fiber reinforced epoxy composites with ANN model. Int J Sci Res 4:410–414
2. Deshpande S, Rangaswamy T, Engg M (2014) Effect of fillers on E-glass/jute fiber reinforced
epoxy composites. Int J Eng Res Appl 4:118–123
3. Fazita MRN, Khalil HPSA, Izzati ANA, Rizal S (2019) 3 - Effects of strain rate
on failure mechanisms and energy absorption in polymer composites. Elsevier Ltd.
https://doi.org/10.1016/B978-0-08-102293-1.00003-6
4. Monteiro SN, Terrones LAH, Almeida JRMD (2008) Mechanical performance of coir
fiber/polyester composites. Polym Test 27:591–595. https://doi.org/10.1016/j.polymertesting.
2008.03.003
5. Balaji V, Senthil Vadivu K (2017) Mechanical characterization of coir fiber and cotton fiber
reinforced unsaturated polyester composites for packaging applications. J Appl Packag Res
9(2), Article 2. https://scholarworks.rit.edu/japr/vol9/iss2/2
6. Santos F, Neves S, Sousa E, Pereira É, Júnior L (2017) Ballistic application of coir fiber
reinforced epoxy composite in multilayered armor. Mater Res 20:23–28
7. Karthikeyan A, Balamurugan K (2012) Effect of alkali treatment and fiber length on impact
behavior of coir fiber reinforced epoxy composites. JSIR 71(09):627–631. http://nopr.niscair.
res.in/handle/123456789/14634
8. Kumar TV, Chandrasekaran M, Santhanam V (2017) Characteristics analysis of coconut shell
husk. J Eng Appl Sci 12:2401–2406
9. Biswas S, Kindo S, Patnaik A (2011) Effect of fiber length on mechanical behavior of coir
fiber reinforced epoxy composites. Fibers Polym 12:73–78. https://doi.org/10.1007/s12221-
011-0073-9
10. Ma Y, Wu S, Zhuang J, Tong J, Qi H (2019) Tribological and physio-mechanical characteri-
zation of cow dung fibers reinforced friction composites: an effective utilization of cow dung
waste. Tribiol Int 131:200–211. https://doi.org/10.1016/j.triboint.2018.10.026
11. Hasan M, Mir SS (2015) Manufacturing of coir fiber-reinforced polymer composites by hot,
Chapter 15. https://doi.org/10.1007/978-3-319-07944-8
12. Cesar J, Luiz R, Machado L, Vieira G, Teixeira R, Freire S, Mano V, Hallak T (2018) Effects
of sodium carbonate on the performance of epoxy and polyester coir-reinforced composites.
Polym Test 67:533–544. https://doi.org/10.1016/j.polymertesting.2018.03.043
13. Nazeer A (2014) To Study the mechanical properties of coconut coir fiber reinforced with
epoxy resin AW 106 & HV 953 IN. Int J Mod Eng Res 4:38–47
14. Péron M, Célino A, Castro M, Jacquemin F, Le Duigou A (2019) Study of hygroscopic stresses
in asymmetric biocomposite laminates. Compos Sci Technol 169:7–15. https://doi.org/10.1016/
j.compscitech.2018.10.027
15. Cunha G, Filho DO, Carlos R, Mota DS, Claudia A, Antunes M, Olimpio O, Filho DA (2019)
Effects of hybridization on the mechanical properties of composites reinforced by piassava
fibers tissue. Compos Part B 162:73–79. https://doi.org/10.1016/j.compositesb.2018.10.050
16. Obada DO, Kuburi LS, Dauda M, Umaru S, Dodoo-arhin D, Balogun MB, Iliyasu I, Iorpenda
MJ (2018) Effect of variation in frequencies on the viscoelastic properties of coir and coconut
450 R. Raja et al.

husk powder reinforced polymer composites. J King Saud Univ Eng Sci. https://doi.org/10.
1016/j.jksues.2018.10.001
17. Habibi M, Laperrière L, Mahi H (2019) Effect of moisture absorption and temperature on quasi-
static and fatigue behavior of nonwoven flax epoxy composite. Compos Part B 166:31–40.
https://doi.org/10.1016/j.compositesb.2018.11.131
18. Romli FI, Nizam A, Shakrine A, Rafie M, Laila D, Abdul A (2012) Factorial study on the tensile
strength of a coir fiber-reinforced epoxy composite. Procedia Soc Behav Sci 3:242–247. https://
doi.org/10.1016/j.aasri.2012.11.040
Study of Hexagonal Boron Nitride
Particulate as Vibration Behaviour
Modifier of Alternate Stacked
Glass–Natural Fibre Polymer Composite
Laminate

K. Rajkumar and M. Selvaraj

Abstract The vibration behaviour of polymers can be modified by the particu-

lates and natural fibres which are interesting study to the design engineers. A study
has been carried out to investigate the tensile, flexural and vibrational behaviour
of vinyl ester composites reinforcing with woven kenaf and glass fibres. Vibration
behaviour of composite was further enhanced by the addition of hexagonal boron
nitride (2&4% hBN) particles. The fibres were chemically treated to improve its
adhesion to matrix. The tensile strength of fabricated composites was marginally
reduced with increase in hBN particles. The vibration behaviours of composite were
studied by electro-dynamic shaker. The experimental results were clearly showed
that natural frequency of composite decreases with increasing particulate content.
The hBN particles inclusion in the treated kenaf laminate has increased damping
ratio by dissipating by more local vibration energy. This is due to combined effect of
chemical treatment of fibre and local relaxing behaviour of secondary hBN particles.
The two fold increase in damping ratio obtained by adding four times of hBN parti-
cle. Dynamic mechanical tests proved that hBN, a secondary particle, is improving
vibration damping characteristics. Hence, it potentially benefits the composite as an
engineered material.

Keywords Vibration energy · Electro-dynamic shaker · Natural frequency ·

Polymer composite

1 Introduction

In recent years, it is a known fact that the environment is completely messed up. As
a result, the researchers are probing a solution to fix it using naturally acquired mate-
rials. There comes the natural fibre as an alternate for synthetic fibre. The studies, till

K. Rajkumar · M. Selvaraj (B)

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Kalavakkam, Chennai, Tamilnadu, India
e-mail: selvarajm71@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 451

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_40
452 K. Rajkumar and M. Selvaraj

the date, have undoubtedly proved that the natural polymeric substances are substi-
tuted for synthetic polymers. In order to achieve the desired properties of the current
applications, the evolution of composites with various natural fibres as reinforcement
and fillers do meet the requirements. Various studies have reported a better result on
the hybridization of natural fibres along with synthetic fibres reinforced with ther-
mosetting polymers [1–3]. Woven fibres with particulates improve the properties of
the composites. Glass fibres are hybridized with natural fibres like kenaf, sisal, jute,
coir, banana, flax, date palm wood, basalt.
Researchers are attracted to the fibres acquired from nature to utilize in poly-
mer composites due to their eco-friendly nature and sustainability [1]. The lignin,
hemicellulose, cellulose-based composites are suitable for the electrical, mechani-
cal, thermal and biodegradation applications [1, 3]. The elastic modulus of fibres
and the interface properties of the matrix have a large effect on the stress and strain
to failure initiation [2]. The thermal property of individual low conductivity kenaf
fibres proves that a reduction in 16% of the effective thermal conductivity of the
composite [3]. The addition of kenaf fibres reduces the degree of crystallinity of the
recycled polypropylene/polyethylene sourced from rejected–unused diapers whose
plastic and absorbent gel were separated [4]. The tensile strength and flexural strength
of kenaf fibre were improved considerably when reinforced with vinyl ester, epoxy
and polyurethane matrixes [5]. The double-layered kenaf fibres increase the thermal
property and the tensile strength of the composite than single-layered kenaf fibre
[6]. The thermal and dynamic mechanical properties of silane-treated pineapple leaf
fibre and kenaf fibre have better than that of untreated fibres [7]. Kenaf fibres absorb
sound and thus used for acoustic purpose. The sound absorption of kenaf fibres is
similar to that of synthetic rock wool [8].
The dynamic flexural modulus has no significant effect on the radial configuration
of the kenaf yarns, whilst the glass fibres have on damping factor. The bilateral
configuration of glass yarns has considerable effects on the static flexural modulus,
strength and the strain at break [9].
The addition of microsize calcium carbonate with kenaf fibres reinforced vinyl
ester increases the tensile and flexural properties of the composite [10]. The
flax/basalt hybrid reinforced with vinyl ester laminates enhances the durability and
water repellence behaviour [11]. The inclusion of kenaf fibre with hexagonal boron
nitride (hBN) and epoxy increases the thermal conductivity of the composite than
with hBN and epoxy alone [12].
Hexagonal boron nitride also termed as white graphite is chemically stable and
possesses good lubrication properties, resistant to oxides, high thermal conduc-
tivity and excellent electrical insulation [13]. Recently, the application of boron
nitride nanomaterials (nanoparticles, nanotubes, nanosheets, etc.) in medical areas
is increasing rapidly because of its unique properties and suitable to the other
applications [14] also.
A study on the damping behaviour of composite is necessary since it reduces
the vibration by dissipating energy. Various types of transducers have been used to
measure the response of the vibrations of composite. An impact hammer attached
with a force transducer is used as an exciter and a mini accelerometer to obtain the
Study of Hexagonal Boron Nitride Particulate … 453

Table 1 Physical and

Polymer Vinyl ester
Mechanical properties of
thermoset resin Density (gm/cm3 ) 1.14
Tensile strength (MPa) 86
Elastic modulus (MPa) 3200
Elongation at break (%) 5–6

response of the simply supported specimen [15]. Flat coils were attached to the ends
of a beam within the fields of permanent magnets. An alternating current through
one coil was used as vibrators whilst the motion of the other coil as transducer. The
composite specimen with high flexural modulus lowers specific damping capacity
[16].
From the above review, it is concluded that the work presented in this paper has not
been done anywhere in the past. This paper provides an overview of the characteristics
of treated and untreated fabrics of kenaf and glass reinforced vinyl ester with boron
nitride particle. The contribution of the boron nitride with the thermosetting polymer
has a significant role on the mechanical and vibration properties of polymer laminates.

2 Materials and Methods

2.1 Materials

The E glass woven was chosen for its low cost and high performance in strength,
stiffness, heat and chemical resistance. Kenaf being one of the lignocellulosic fibres
was selected due to its low cost, low density, high specific strength and sustainability.
Woven kenaf fibres were soaked in 6% (wt) of NaOH solution at 25 °C for 24 h
maintaining a water solution ratio of 15:1. They were then washed well with distilled
water and after neutralizing with dilute acetic acid. The kenaf fabrics were dried for
48 h. The boron nitride particle of size 5 µm is used due to its lamellar structure and
high thermal conductivity. The vinyl ester resin of grade HPR 8711 with catalyst
(1% MEKP), promoter (Benzyl peroxide) and accelerator (0.5% Cobalt Napthanate)
to form a matrix was used. Table 1 refers the physical and mechanical properties of
the thermoset polymer used.

2.2 Composite Fabrication

The hybrid composite laminates were fabricated at room temperature by hand layup
technique. The treated and untreated fabrics of kenaf along with glass fibres of size
30 × 30 cm2 were placed in the thermoset resin mixed with hexagonal boron nitride
454 K. Rajkumar and M. Selvaraj

Fig. 1 Specimen used for

tensile test

of various proportions. It was then cured with the promoter, catalyst and accelerator.
Five sets of hybrid composites were prepared by placing the fabrics of treated kenaf
and glass alternately for 0, 2 and 4% of hBN. Similarly, the untreated kenaf fibres
were arranged alternately with glass fibres for 2% and 4% hBN mixed with resin,
respectively.

2.3 Tensile Test

The five samples prepared for each formulation were cut according to ASTM D-3039
(1995) with a size of 250*25*4 mm to carry out the tensile test in an universal testing
machine (Instron 4206) at a feed rate of 0.1 mm/min. Figure 1 shows the test samples.
During the tension test, the grips were moved apart at a constant rate to stretch the
specimen. The results were averaged for the five specimen of each formulation for
calculation purpose. Scanning electron microscope (SEM) was used to evaluate the
fracture surface of the polymer matrix and to illustrate the fibrillose morphology of
the kenaf fibres.

2.4 Flexural Test

The prepared flexural samples of each formulation with dimensions of 64*16*4 mm

according to ASTM D790 were tested in universal testing machine. The bending
strength of the composite was measured by a three point bending test.
Study of Hexagonal Boron Nitride Particulate … 455

2.5 Vibration Test

A multirecorder frequency processing unit was used to determine the vibration of the
hybrid composite laminates. Vibration induced on the specimen by electro-dynamic
shaker. The phase analyser (FFT analyser) has four input channels and was used
as a data acquisition system. The accelerometer measures the acceleration of the
vibrations. The impact hammer was used for impacts of varying amplitude. The
real-time frequency analysis software, DEWESOFT, was used for the data analysis
and verification. The one end of the specimen was hold by gripper, and another
end was free, as cantilever mode. The specimens were excited at the free end by
impact hammer, and the corresponding responses were recorded. This vibration test
determines the natural frequency and damping ratio.

3 Results and Discussions

3.1 Tensile Strength

Figure 2 shows a comparative analysis of tensile strength for the composite laminates
fabricated with treated and untreated kenaf/glass fabrics reinforced with 0, 2 and 4%
of hBN. The fabricated specimen with treated kenaf fabric hybridized with glass
fibres for the addition of 2% hexagonal boron nitride (hBN) mixed with vinyl ester
resin has the highest tensile strength. The tensile strength gradually increases for the
samples fabricated with the treated kenaf fibres hybridized with glass fibres matrix
with 0–2% addition of hBN and decreases for 4% of hBN.

Fig. 2 Tensile strength of composites

456 K. Rajkumar and M. Selvaraj

Fig. 3 Flexural strength of the composites a Untreated b treated

The tensile strength obtained for the untreated kenaf fibre with the addition of
2% hBN gives the maximum value, whilst the addition of 4% hBN gives the mini-
mum value. The notable improvement in tensile strength of the treated kenaf fibres
hybridized with glass fibres can be due to strong interface adhesion. The proper addi-
tion of hBN, that is, 2% resulted a very good adhesion property for the fibres with
the matrix than 4%. The fall of tensile strength value for 4% addition of hBN does
not attribute high bond between fibre and matrix. Figure 3 shows fracture surface of
untreated and treated kenaf fibre hybrid composites.
The untreated fibre composite shows an inefficient transfer of resin along the fibre
direction leading to poor intact between fibre-matrix interfaces. The used NaOH
treatment has removed impurities from the external surface of kenaf fibres, so their
surface appears clean, and it gives good adhesion to the polymer matrix.

3.2 Flexural Strength

The flexural strength of the laminated hybrid composite samples is shown in Fig. 4.
It is clearly noted that the flexural strength of the sample fabricated with untreated
kenaf fabric hybridized with glass fabric reinforced in 2% hBN mixed with vinyl
ester resin gives the maximum, whilst the minimal value is seen for 4% hBN. The
composite laminates fabricated with treated fabrics of kenaf and glass reinforced in
thermoset resin gradually increase from 0 to 2% hBN and a fall in value for 4% hBN.
Study of Hexagonal Boron Nitride Particulate … 457

Fig. 4 Flexural strength of the composites

3.3 Vibrational Behaviour

The graphical output of vibrational behaviour polymer composite is shown in Fig. 5.

It is clearly seen from Fig. 5 that natural frequency of laminates decreases with
increasing particulate content. Chemical treatment of kenaf fibre increases the natural
frequency of laminates. The natural frequency of the laminate without hBN was
higher due to continuous phase of matrix.

Fig. 5 Vibrational behaviour of hybrid composites

458 K. Rajkumar and M. Selvaraj

Table 2 Natural frequency

Samples Natural frequency Damping ratio
and damping ratio
(Hz)
Untreated, 2% hBN 50 0.033
266 0.008
Treated, 0% hBN 46 0.043
262 0.022
Treated, 4% hBN 34 0.076
190 0.045
Untreated, 4% hBN 53 0.031
296 0.013
Treated, 2% hBN 45 0.041
245 0.025

The two modal frequencies of each sample are enlisted in Table 2. The damping
ratio was obtained using the DEWESOFT software. In higher modal frequency,
treated fibre laminate shows the higher damping than that of untreated kenaf fibre
laminate. The strong adhesion of matrix and fibre mostly relaxes the vibration energy.
The hBN particles inclusion in the treated kenaf laminate increases the damping ratio
by dissipating by more vibration energy, as shown in Fig. 6. This is a synergizing
effect by the strong adhesion between fibre-matrix and local relaxing behaviour of
secondary hBN particles. The weak interface was observed from the untreated fibre
reinforced composite, despite being the presence of hBN particle, leading to lower
damping ratio. This is due that lower dissipation of vibration energy by the kenaf
fibres.
The damping ratio of composite without hBN particle exhibited a minimum value.
The four times of increasing particle content in matrix observed to be increased the
two times of damping ratio.

Fig. 6 Damping ratio of

composite at high modal
frequency 0.04
Damping ratio

0.03

Untreated fiber
0.02 Treated fiber

0.01

0 2 4
hBN %
Study of Hexagonal Boron Nitride Particulate … 459

4 Conclusions

Vinyl ester hybrid composites were fabricated with kenaf fibre mat, glass fibre mat
and different wt. fraction of hBN particles. The composite with treated fibre-2%
hBN showed a higher tensile strength, but 4% hBN addition shows a lower value.
Fracture surface of untreated fibre composite shows an incomplete resin transfer
to fibre leading poor intact between fibre-matrix interfaces. The NaOH treatment
has removed impurities from the surface of kenaf fibres paved the way for good
adhesion property. It is clearly showed that the flexural strength of the treated kenaf
fabric hybridized with glass-2% hBN exhibited a maximum, whilst the minimal value
observed for 4% hBN content composite.
It is clearly seen that natural frequency of laminates decreases with increasing
particulate content. Fibre treatment increases the natural frequency of laminates due
to undisturbed continuous phase of matrix. The hBN particles inclusion in the treated
kenaf laminate increases the damping ratio by dissipating by more vibration energy.
This is a synergizing effect of strong adhesion between fibre-matrix and local relaxing
behaviour of secondary hBN particles. The damping ratio of composite without hBN
particle exhibited a minimum value. The four times increasing hBN particle content
in the matrix increases the two times of damping ratio.

References

1. Mohammed L, Ansari MNM, Pua G, Jawaid M, Saiful Islam V (2015) A review on natural
fiber reinforced polymer composite and its applications. Int J Polym Sci. Article ID 243947,
15 pages http://dx.doi.org/10.1155/2015/243947
2. Asp LE, Berglund LA, Talreja R (1996) Effects of fiber and interphase on matrix Initiated
Transverse failure in polymer composites. Compos Sci Technol 56:657–665
3. Guillou J, Lavadiya DN, Munro T, Fronk T, Ban H (2018) From lignocelluloses to biocomposite:
multi-level modeling and experimental investigation of the thermal properties of kenaf fiber
reinforced composites based on constituent materials. Appl Thermal Eng 128:1372–1381
4. Taufiq MJ, Mansor MR, Mustafa Z (2018) Characterization of wood plastic composite
manufactured from kenaf fibre reinforced recycled-unused plastic blend. Compos Struct
189:510–515
5. Sharba MJ, Salman SD, Leman Z, Sultan MTH, Ishakand Mohamad MR, AzmahHanim A
(2016) Effects of processing method, moisture content and resin system on physical and
mechanical properties of woven kenaf plant fiber composites. BioResources 11(1):1466–1476
6. Islam MR, Beg MDH (2016) Effects of reinforcing arrangement of kenaf fibers into unsaturated
polyester for improved properties. J Mech Eng Sci 10(2):2020–2030
7. Asima M, Paridah MT, Saba N, Jawaid M, Alothman OY, Nasir M, Almutrai Z (2018) Thermal,
physical properties and flammability of silane treated kenaf/pineapple leaf fibers phenolic
hybrid composites. Compos Struct 202:1330–13338
8. Lim ZY, Putra A, Nor MJM, Yaakob MY (2018) Sound absorption performance of natural
kenaf fibers. Appl Acoustics 130:107–114
9. Hashemi F, Brancheriau L, Thahir PM (2018) Hybridization and yarns configuration effects
on flexural dynamic and static properties of pultruded hybrid kenaf/glass fibre composites.
Compos Part A 112:415–422
460 K. Rajkumar and M. Selvaraj

10. Fairuz AM, Sapuan SM, Zainidin ES, Jaafar CNA (2016) Effect of filler loading on mechan-
ical properties of pultruded kenaf fiber reinforced vinyl ester composites. J Mech Eng Sci
10(1):1931–1942
11. Ramraji K, Rajkumar K, Sabarinathan P (2019) Tailoring of tensile and dynamic thermo-
mechanical properties of interleaved chemical-treated fine almond shell particulate flax fiber
stacked vinyl ester polymeric composites. In: Proceedings of the Institution of Mechanical
Engineers, Part L: Journal of Materials: Design and Applications, p 1464420719849616
12. Xia C, Garcia AC, Shi SQ, Warner N, Wu Y, Cai L, Rizvi HR, D’Souza NA, Nie X (2016)
Hybrid boron nitride—natural fiber composites for enhanced thermal conductivity. Sci Rep
6:34726. https://doi.org/10.1038/srep34726
13. Kivnc M, Barutca B, Koparal AT, Goncu Y, Bostanci SH, Ay N (2018) Effects of hexagonal
boron nitride nanoparticles on antimicrobial and antibiofilm activities, cell viability. Mater Sci
Eng C. https://doi.org/10.1016/j.msec.2018.05.028
14. Gnanavelbabu A, Rajkumar K, Saravanan P (2018) Investigation on the cutting quality charac-
teristics of abrasive water jet machining of AA6061-B4C-hBN hybrid metal matrix composites.
Mater Manuf Process 33(12):1313–1323
15. Rao MD, Crocker MJ (2009) Vibrations of bonded beams with a single lap adhesive joint. J
Sound Vibration 92(2):299–309
16. Kadioglu F (2009) Measurement of dynamic properties of composites in vibration by means
of a non-contact mechanism. J Reinforced Plast Compos 28, No. 12/2009 https://doi.org/10.
1177/0731684408089514
Study of Statistical Distribution
and Morphology of Particles in a Polymer
Matrix by Foldscope Imaging Technique

P. Kaythry, A. Madhan, and K. Rajkumar

Abstract Design and material engineers are focusing on strengthening and tough-
ening of the polymer matrix by introducing secondary particles. However, particle
distribution is a control parameter to ensure the properties of polymer materials. Fur-
ther, the agglomeration of the fine particle affects polymer composite performance.
The chosen process may affect the statistical distribution and aggregation of sec-
ondary particles in the polymer composite. There is a lack of simple and affordable
technique to evaluate the particle distribution, morphology and degree of agglomera-
tion of particles in the polymer matrix. In this paper, the origami crafted microscope
called foldscope is used to find out the particle distribution in the polymer matrix. Ini-
tially, the fillers/particles are mixed in the epoxy matrix through an ultra-sonication
technique. The experiments are conducted by varying the size of filler (30, 80 and
120 grit) with sonication parameters such as power rate (70, 80 and 100%) and pro-
cessing time T on (45, 90 and 135 s) with constant 3% filler in weight. With the above
parameters, Taguchi L17 array was constructed, and the process parameters were
optimized to the high degree of uniform distribution. From the particulate polymer
composite, the images captured by foldscope were further analyzed by ImageJ tech-
nique for determination of the statistical distribution of particles. The processing
parameters of 80% power, T on 135 s for 120 grit size exhibited excellent particle
distribution and higher tensile strength.

Keywords Polymer matrix · Foldscope · Ultra-sonication · Particle distribution

P. Kaythry (B) · A. Madhan

Department of Electronics and Communication Engineering, Sri Sivasubramaniya Nadar College
of Engineering, Chennai, India
e-mail: kaythryp@ssn.edu.in
K. Rajkumar
Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, India

© Springer Nature Singapore Pte Ltd. 2021 461

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_41
462 P. Kaythry et al.

1 Introduction

The modern applications like aerospace, automobiles, renovation of building and

bridges [1] are using polymer composite materials. The dispersion and interfacial
adhesion between the matrix and filler are required to achieve the desired mechanical
properties [2, 3]. The structural feature of polymer composite dictates by tensile and
flexural strength.
Visual inspection is the first level evaluation to select any polymer composite in
the applications [4]. It is used to identify the particle distribution, agglomeration,
porosity and discontinuity by conventional microscopes like compound microscope,
optical microscope and scanning electron microscope (SEM) [5–7]. These techniques
require different sample preparation methods, which are more expensive and tedious.
To overcome difficulties, foldscope is used for the visual study of the polymer matrix.
It also provides flexibility over the other conventional microscopes. For using a
foldscope, the images are captured by attaching a mobile phone camera over it
[8]. It produces quality (3024 × 4032 pixels) microscopic images at any instance.
Also, it does not require any special working conditions [9]. Hence, foldscope is a
highly effective method for microscopic level image capturing. However, an image
processing is further required to analyze the captured images [10, 11].
Specific to the composite, the dispersion of particles depends on the processing
methods [3]. Generally, the particle distribution in a polymer matrix is done by
centrifuging, vibratory table, magnetic particulate distribution and ultra-sonication
method. Microscopic image defines the relative amount of mass, number of a particle
according to size and mean distance between the particles in a spatial matrix [12].
In this work, an ultrasonic technique was used to disperse almond particles in the
polymer matrix and correlated spatial distribution of particles to the tensile strength.

2 Materials and Methods

2.1 Basic Components and Preparation

In this work, dry almond shells were reinforced to obtain the polymer composite.
Initially, the impurities from almond shells were removed. Then, it is crushed by
using Willy’s mechanical mill. The crushed particles were sieved to 30, 80 and 120
grit size by laboratory shaker and used as a filler material for the polymer matrix.
Araldite HY550 epoxy resin and hardener HY951 were used as polymer matrix to
fabricate the polymer composite. A constant filler weight of 3% is used to fabricate
the composite laminate.
Study of Statistical Distribution and Morphology … 463

Fig. 1 a Sonication setup and b ice-bath setup

2.2 Ultra-Sonication of Particle Dispersion

Ultra-sonication is a technique used for dispersing particle in a solution. When the

probe inserted into the solution, ultrasonic waves are formed [13]. It creates a large
number of micro-bubbles, followed by high and low pressure in the solution. The
difference in pressure leads to sudden and violent collapse of micro-bubbles. Under
this circumstance, even dispersion of particles would be occurred [14].
Sonication and ice-bath setup are shown in Fig. 1a, b. It consists of a probe,
processing container and programmable logic controller (PLC). The ultrasonic pro-
cessor operates at 20 kHz frequency with the maximum power of 2000 W. The power
is varied from 60, 70, 80 and 100% of maximum power. The variable operating time
(T on ) and constant interval time (T off = 3 s) were controlled by PLC to study the
particle dispersion in the polymer composites. The sonication process increases the
temperature of the resin, which changes viscosity. So, the processing container is
surrounded by the ice-bath to minimize the temperature variation.

2.3 Fabrication

The hand layup method was used for the fabrication of polymer composite because
of its low cost, maximum effectiveness and simplicity. The particulate laminate has
dimensions of 300 mm × 300 mm × 4 mm, which is divided into three segments
along its length. Each segment has a size of 300 mm × 100 mm × 4 mm. The almond
shell powder of different grits was mixed with resin by ultra-sonication process. For
the particle dispersion study, Taguchi experimental design was constructed, and the
464 P. Kaythry et al.

Table 1 Selected parameters

Particle size in grit Power % in 2000 W T on (s)
range and levels used for
Taguchi design 30 70 45
80 80 90
120 100 135

Table 2 Formation of L17

Experiment Particle size T on (s) Power (% in
array–Box–Behnken method
number (Grit) 2000 W)
of experimentation
1 80 135 70
2 120 135 80
3 80 90 80
4 120 90 70
5 80 45 70
6 80 45 100
7 30 90 100
8 80 90 80
9 80 90 80
10 30 135 80
11 80 90 80
12 120 90 100
13 30 45 80
14 80 90 80
15 80 135 100
16 30 90 70
17 120 45 80

range and level of the selected input parameters are given in Table 1. Table 2 shows
the formation of L17 array by using a Box–Behnken method of experimentation.
The fabricated almond particle dispersed in polymer composites is shown in Fig. 2.

2.4 Tensile Test

The tensile test was carried out to evaluate the tensile strength by universal testing
machine (Instron 4206) at a feed rate of 0.2 mm/min. A standard dimension of
250 mm × 25 mm × 4 mm was prepared according to American Society for Testing
and Materials (ASTM) D638 standard. Abrasive water jet machine (AWJM) was used
to make the standard specimen. The prepared specimens and tested tensile samples
are shown in Figs. 3 and 4, respectively.
Study of Statistical Distribution and Morphology … 465

Fig. 2 Almond particle dispersed in polymer composite

Fig. 3 Prepared specimen for tensile test

466 P. Kaythry et al.

Fig. 4 Tensile tested samples of Foldscope

A foldscope is an optical microscope designed by Manu Prakash. It costs less than

one US dollar. The foldscope weighs 8 grams and provides magnification from 140
× to 2000 × . The foldscope can be attached to a smartphone with the help of mag-
nets, permitting the user to take magnified pictures. It is compact and light, especially
when compared with other microscopes. Due to its portability and maneuverability,
the foldscope displayed greater ease of access.
The fundamental components like punched sheets, magnetic couplers and spher-
ical glass lens, along with a battery that powers the LED as a light source are used
to assemble a foldscope. The lens forces light waves to bend as they pass through an
object. As the light waves bend, the object is magnified based on the principles of
ray optics. The magnification usually depends on how curved the lens, the distance
between the lens, object and mobile camera. The picture of an assembled foldscope
is shown in Fig. 5. The linear movement along X- and Y-axis used to produce images
over various point and fully constrained X–Y panning over a 20,620 mm2 region.
Z-axis movement is used to provide controlled space between optics stage relative
to the sample-mounting stage by flexure-based focusing. This movement helps to
capture clarity images and attached scale used to measure distance travelled across
the sample slide with 0.5 mm resolution.
Study of Statistical Distribution and Morphology … 467

Fig. 5 An assembled foldscope [8]

2.5 Foldscope Image Analysis

Foldscope images of fabricated composites are captured and visually analyzed.

ImageJ software is an image processing tool used to evaluate the degree of agglom-
eration, particulate counts, particle distribution, mean, the standard deviation of par-
ticles, histogram and an area of filler distribution [14]. Figure 6 shows the processed

Fig. 6 Processed image

468 P. Kaythry et al.

image of polymer composite which is an experiment number 10, as presented in

Table 2. It describes the percentage of filler occupied area from the total available
area.

3 Results and Discussion

3.1 Foldscope Visual Images

Table 3 represents the processing parameters and results for particle distribution and
tensile strength. As the grit size increases, the mean particle size decreases. From the
experiment number 2, 4 and 12, it is observed that smaller particle size shows a higher
particle distribution with very less mean free path between the particles. Since the
surface area to the volume of the particle is higher for smaller particles, the number of
particles present in the fixed area is high. On the other hand, the experiment numbers
7, 10 and 13 with larger particle size reveal particle distribution with low surface area
to the volume of the particle. As results, the number of particles in the fixed area was
low. Thereby mean free path between the particles is also high. The foldscope images

Table 3 Image processing results of the tensile strength of composites

Experiment Figure Particle size T on (s) Power (% in Particle Strength
No. (Grit) 2000 W) distribution (MPa)
(% area)
1 (1,2) 80 135 70 45.514 25.4
2 (1,1) 120 135 80 48.72 42.2
3 – 80 90 80 42.691 26.185
4 (2,1) 120 90 70 36.764 25.4
5 (2,2) 80 45 70 42.784 15.56
6 (3,2) 80 45 100 52.086 32.5
7 (2,3) 30 90 100 50.835 22.7
8 – 80 90 80 42.691 26
9 – 80 90 80 42.691 26
10 (1,3) 30 135 80 38.734 19.7
11 – 80 90 80 42.691 26
12 (3,1) 120 90 100 62.071 30.4
13 (3,3) 30 45 80 71.128 24
14 – 80 90 80 42.691 26
15 – 80 135 100 59.543 14.8
16 – 30 90 70 42.691 23.32
17 – 120 45 80 33.689 29
Study of Statistical Distribution and Morphology … 469

of polymer composites are given by a column–row matrix in Fig. 7. The grit size of
120, 80 and 30 is mentioned for each column. It also evidences that the strength of
the composites partly depends on the degree of the particle distribution.
The statistical analysis of the experimental results is performed using Design
Expert statistical software. A separate analysis of variance (ANOVA) is carried out
for the particle distribution and tensile strength. The optimum level of variables for the
uniform distribution is obtained using the Design Expert program. For a significant
model, the P value should be < 0.05. From ANOVA for particle distribution in Table 4,
the Model P value of 0.0098 implies the model is significant. There is an only chance
of 0.98% error in Model F value occurs due to noise in experimentation. Power (B),
Particle size × T on (AC) were significant model terms.
Increasing the power of the sonication process increases the intensity of the sonic
waves, which is shown in Fig. 8. There is no effect on the particle distribution when

Column – 120 grit Column – 80 grit Column – 30 grit

Fig. 7 Foldscope image of polymer composites with different process parameters

470 P. Kaythry et al.

Table 4 Analysis of variance for particle distribution

Source Sum of df Mean square F value P value Prob
squares >F
Model 1145.78 6 190.96 5.41 0.0098 Significant
A-Particle 66.77 1 66.77 1.89 0.199
size
B-Power 375.12 1 375.12 10.63 0.0086
C-T on 19.67 1 19.67 0.56 0.4725
AB 68.47 1 68.47 1.94 0.1938
AC 595.49 1 595.49 16.87 0.0021
BC 5.59 1 5.59 0.16 0.6991
Residual 352.91 10 35.29
Pure error 1.45 4 0.36
Cor total 1498.69 16

Fig. 8 Contour plot showing the effect of particle size and power on particle distribution

the threshold level of power is above 93%. Moreover, 120 grit particle size requires
less energy to disperse the particle with high distribution percentage due to the mass–
acceleration relationship. If the power is below 93%, then the maximum distribution
of particles occurs for all range of particle size. For 30 grit particle size, it requires
a higher power level of 85% for the dispersion of particulates to cover a maximum
of 47.85% distribution area.
The 120 grit particle size with medium power (85%) gives the higher tensile
strength. The higher particle size with all range of power reduces the strength of the
composites due to mass–acceleration relationship as shown in Fig. 9.
Study of Statistical Distribution and Morphology … 471

Fig. 9 Correlation of tensile strength of particle size and power

The ANOVA for tensile strength is given in Table 5. In this Model, P value of
0.0224 implies the model is significant. So, there is only a 2.24% chance that a model
deviates, which could be due to the noise in the experimentation. Particle size (A),
Particle size × T on (AC), Power × T on (BC) are significant model terms.
Figure 10 shows the relationship between tensile strength, particle size and power.
The particle size affects the mechanical strength of the polymer matrix composites.
In a smaller particle (120 grit), the lower mean free path of particles increases the
cross-linking of polymer chain. Hence it results in higher tensile stress of particulates

Table 5 Analysis of variance for tensile strength

Source Sum of df Mean square F value P value Prob
squares > F
Model 543.6581 9 60.40645 5.027942 0.0224 Significant
A-Particle 173.7248 1 173.7248 14.46002 0.0067
size
B-Power 12.6418 1 12.6418 1.052243 0.3391
C-T on 0.077477 1 0.077477 0.006449 0.9382
AB 8.154704 1 8.154704 0.678758 0.4372
AC 68.21096 1 68.21096 5.677552 0.0487
BC 189.6129 1 189.6129 15.78246 0.0054
Aˆ2 55.42087 1 55.42087 4.612966 0.0688
Bˆ2 52.70503 1 52.70503 4.386913 0.0745
Cˆ2 0.302417 1 0.302417 0.025172 0.8784
Residual 84.09905 7 12.01415
Pure error 1.16392 4 0.29098
Cor total 627.7571 16
472 P. Kaythry et al.

Fig. 10 Tensile Strength

in polymer composite. The degree of cross-linking is low for large particle size (30
grit), which leads to lower tensile strength.
Figure 11 shows the relationship between tensile strength, power and T on . The
tensile strength of composites increases with the T on due to the higher degree of
particulate disperse with all the range of particle size. It is directly derived that
power of the sonication process increases the tensile strength of composite. This is
because of higher degree particle distribution with all the range of particle size.

Fig. 11 Correlation of tensile strength of T on and power

Study of Statistical Distribution and Morphology … 473

Fig. 12 Average tensile strength to ratio of distribution and particle size

3.2 Tensile Strength-Particle Factional Image Area

Figure 12 depicts the average tensile strength to the ratio of distribution and particle
size. In zone-I, when the ratio of distributions to particle size is low, there is a steep
rise in tensile strength due to the higher order of cross-linking. Zone-II is called as
a transition zone, where the tensile strength begins to decrease. In zone-III, there is
no improvement in the tensile strength due to the saturation in cross-linking.

4 Conclusions

Distribution of particles is analyzed by visual study through foldscope images for

the various polymer composites with varying grit sizes and other process parameters.
The foldscope captured images of polymer composites are evaluated using ImageJ
software. Analysis of variance identified the desired parameters for the particle dis-
tribution and tensile strength of the polymer composites. From the results, the smaller
particulates with a higher distribution increase tensile property even at lower T on and
power. Foldscope imaging is simple and least cost technique for characterization and
visual inspection of polymer-based products.

Acknowledgements Authors are thankful to the Department of Biotechnology, Govt. of India—

Foldscope Research Grant (BT/IN/Indo-US/Foldscope/39/2015) for supporting this work.
474 P. Kaythry et al.

References

1. Friedrich K, Abdulhakim A, Almajid (2013) Manufacturing aspects of advanced polymer

composites for automotive applications. Appl Compos Mater 20(2):107–128
2. Sabarinathan P, Rajkumar K, Gnanavelbabu A (2016) Mechanical properties of almond shell-
sugarcane leaves hybrid epoxy polymer composite. Appl Mech Mater 852:43–48
3. Fu S-Y (2008) Effects of particle size, particle/matrix interface adhesion and particle loading on
mechanical properties of particulate–polymer composites. Compos Part B: Eng 39(6):933–961
4. Ramakrishna S (2001) Biomedical applications of polymer-composite materials: a review.
Compos Sci Technol 61(9):1189–1224
5. Saunders RA, Lekakou C, Bader MG (1999) Compression in the processing of polymer com-
posites 1. A mechanical and microstructural study for different glass fabrics and resins. Compos
Sci Technol 59(7):983–993
6. Ramraji K, Rajkumar K, Sabarinathan P (2019) Tailoring of tensile and dynamic thermo-
mechanical properties of interleaved chemical-treated fine almond shell particulate flax fiber
stacked vinyl ester polymeric composites. In: Proceedings of the Institution of Mechanical
Engineers, Part L: Journal of Materials: Design and Applications, p 1464420719849616
7. Qaiss A, Bouhfid R, Essabir H (2015) Characterization and use of coir, almond, apricot, argan,
shells and wood as reinforcement in the polymeric matrix in order to valorize these products.
In: Agricultural biomass-based potential materials. Springer, Cham, pp 305–339
8. Cybulski JS, Clements J, Prakash M. Foldscope: origami-based paper microscope. PLoS ONE
9(6):98781. https://doi.org/10.1371/journal.pone.0098781
9. Kaythry P, Sangeetha P, Madhan A (2019) Study on shape and texture of different types of
fine aggregate used in concrete using Foldscope In: Proceedings of national conference on
‘MICROCOSMOS 2019’, Port Blair, Andaman & Nicobar Islands, p 11
10. Igathinathane C (2008) Shape identification and particles size distribution from basic shape
parameters using ImageJ. Comput Electron Agric 63(2):168–182
11. Papadopulos F (2007) Common tasks in microscopic and ultrastructural image analysis using
ImageJ. Ultrastructural Pathol 31(6):401–407
12. Loganathan D, Gnanavelbabu A, Rajkumar K, Ramadoss R (2014) Effect of microwave heat
treatment on mechanical properties of AA6061 sheet metal. In: Procedia engineering, vol 97,
pp 1692–1697
13. Kaboorani A, Riedl B, Blanchet P (2013) Ultrasonication technique: a method for dispersing
nanoclay in wood adhesives. J Nanomater 3
14. Singh AK, Sandun D, Fernando, Hernandez R (2007) Base-catalyzed fast transesterification
of soybean oil using ultrasonication. Energy Fuels 21(2):1161–1164
Dynamic Mechanical Analysis of Flax
Fiber Stacked Polyurethane Blend Epoxy
Composites

K. Rajkumar , K. Ramraji , M. Rajesh , and M. Rajiv kumar

Abstract Dynamic mechanical properties of polymer blend composites reinforced

with flax woven were investigated. Polyurethane prepolymer was used as a modifier
with the varied composition of 5–20 wt% within the epoxy resin. The ultrasonic
stirring process was employed to mix the polyurethane with epoxy resin. Polymer
blend composite was fabricated by using hand layup technique followed compres-
sion pressing. Tensile and flexural test was carried out for the fabricated composite.
Experimental results revealed that the increasing polyurethane composition gradually
decreases the mechanical properties of the blended polymer composite. Additionally,
polyurethane blending has reduced the brittleness of composites. Dynamic mechan-
ical properties of fabricated blend composites like storage modulus, loss modulus,
and loss factor (tan δ) were improved only with 10% polyurethane composition.

Keywords Polymer blend · Polyurethane · Mechanical properties · Flax fiber

1 Introduction

Natural fiber-reinforced polymer (NFRP) composites are recently playing an impor-

tant role in designing of structural and semi-structural components due to its superior
mechanical and dynamic mechanical properties. The use of natural fibers has highly
accepted because of increase environmental concern and eco-friendly demands. Nat-
ural fibers are greatly reducing the application of glass fiber in the polymer composites
[1–3]. Natural fiber is basically extracted from the plant, animal, and mineral, namely
flax, kenaf, hemp, sisal, basalt, etc. [4]. Natural fibers are improving the damping
characteristics of the composite [5]. Flax is a bast fiber with comparable mechanical
properties to the glass fiber [6]. The natural fibers are normally hydrophilic nature
which results in high moisture absorption and weak adhesion with the matrix. The

K. Rajkumar (B) · K. Ramraji · M. Rajesh · M. Rajiv kumar

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, Tamil Nadu, India
e-mail: rajkumark@ssn.edu

© Springer Nature Singapore Pte Ltd. 2021 475

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_42
476 K. Rajkumar et al.

chemical treatment method has reducing hydrophilic behavior of the natural fiber.
Flax fiber woven was treated with 5% NaOH solution providing better mechanical
properties [7]. However, its low degradation temperature (∼200 °C) makes them
incompatible to the thermoset polymer composite. This also restricts the selection
of manufacturing method for natural fiber composite processing [8].
Epoxy (EP) resin is one of the most important matrixes for the polymer compos-
ites include surface coatings, insulating materials, and structural adhesives. But its
high cross-linking network during curing produces undesirable brittle characteristic.
This leads to poor resistance to crack growth, low toughness, and low impact strength
[9, 10]. In order to change its brittle nature by blending with compatible polymers
such as an elastomer or rubbery, thermoplastic is required. These are providing addi-
tional interpenetrating polymer networks (IPNs) with an epoxy cross-link network.
Polyurethane and the epoxy resins are NCO (nitrogen, carbon, and oxygen) and OH
group. Epoxy blending with polyurethane makes it strong and flexible. Also, it gives
a stable epoxy structure by polyurethane IPNs [11, 12]. The properties of the blends
are the same as those synthesized by batch polymerization. Moreover, it requires the
least time and energy input for the polymerization process [13]. Polyurethanes (PUs)
are a unique family of polymers that possess some of the properties of plastics and
rubbers. PU/EP IPNs show excellent thermal and mechanical properties as derived
from the advantage characteristic of both the PU and EP [14, 15]. During the past
decades, the great deals of efforts have been made on the evaluation of mechanical
properties, and morphology of PU/EP IPNs and optimal ratio of PU IPNs in EP [16].
Viscoelastic polymeric materials are used as a medium of attenuating sound and
damping the vibration. Those capabilities of the polymer are determined by how
much dissipating sound or the vibrational mechanical energy into heat. This is indi-
cated by intensity and broadness of the loss modulus or tan δ peaks at a particular
range of temperatures. As temperature or testing frequency increases, the loss mod-
ulus (E  ) and loss tangent (tan δ) show an intensity peak when damping reaches
a maximum value [17]. The storage modulus (E  ) or dynamic modulus is typically
related to Young’s modulus of a material. It depends on the stiffness of a material
and determines the sample whether it is stiff or flimsy. E  is a material property that
having an ability to store energy applied to it. The changes in strength and stiffness
of epoxy thermoset are evident by blending with polyurethane resin. The damp-
ing characteristics of epoxy polymer composites can be tailored by optimization of
polyurethane composition [18, 19].
This research work reports the optimum composition of polyurethane blend and its
effect on the dynamic mechanical properties of flax fiber-reinforced epoxy polymer
composite.
Dynamic Mechanical Analysis of Flax … 477

2 Materials and Methods

2.1 Materials

Flax fiber was purchased from the Go Green Fiber Ltd., Chennai. The properties
of flax fiber are follows, fiber strength (88–1500 MPa), density (1.4 g/cm3 ), specific
modulus (60–80 GPa), and elongation at failure 1.2–1.6% [20]. The used epoxy resin
grade was LY556 and of density 1.25 (g/cm3 ), Young’s modulus (3.5 GPa), tensile
strength (50 MPa), and fracture toughness (0.5 MPa) [21]. Polyurethane properties
are followed as density (1.2 g/cm3 ), Young’s modulus (0.01 GPa), tensile strength
(20 MPa), and fracture toughness (0.1 MPa) [15]. Diethylenetriamine (DETA) was a
curing agent of C4H13N3 purchased from the Vasavibala Chemical, Chennai. Methyl
ethyl ketone (MEK), a grade of C2 H5 ·CO·CH3 as a compatizer was purchased from
the Chen chem chemical, Chennai, India.

2.2 Fabrication of Blend Composites

The composite laminates were made by hand layup technique that followed a grav-
ity compression loading. Flax woven was treated with 5% NaOH solution for three
hours and then washed by running tap water. Treated woven fibers were dried under
direct sunlight to remove moisture. The releasing agent (Mansion Wax) was properly
spread inside of mold for easy separation. Polymer blend laminate has a dimension
of 300*300*5 mm3 and is fabricated. To prepare epoxy-polyurethane blend compos-
ite, epoxy resin and polyurethane were dissolved separately in methyl ethyl ketone
(MEK). The polyurethane pre-polymer was acting as a slender and its concentra-
tions (5, 10, 15, and 20%) in the epoxy resin as shown in Table 1. The polymer blend
resin mixture was prepared by mixing epoxy-polyurethane resins at room temper-
ature. The ultrasonic stirring process was employed to mix the polyurethane resin
in the epoxy resin. The chosen ratio of 10% of diethylenetriamine (DETA) and 2%
of methylethylketone (MEK) has been added. Then, a woven flax layer was an ini-
tial layer, above which flax layer by layer were stacked. Finally, blend mixture was
applied between the flax layers for effective binding. The composite laminates were

Table 1 Polyurethane
Blend composites % of polyurethane blend (by weight)
content of composite samples
BC0 0
BC5 5
BC10 10
BC15 15
BC20 20
478 K. Rajkumar et al.

Fig. 1 Typical blend composite laminate

cured under dead weight loading for 24 h. The typical blend laminate is shown in
Fig. 1.

2.3 Characterization of Blend Composites

The mechanical and dynamic mechanical properties were evaluated for the fabricated
blend composites as given below.
Tensile and Flexural Strength
The specimens were cut by abrasive water jet machine for both tensile and flexural
tests. The tensile property was recorded from the universal testing machine (UTM).
The tensile specimen was confirming the ASTM-D638 standard and is shown in
Fig. 2a. Loading on the sample was continued until the specimen breaks. The flex-
ural property was recorded using the universal testing machine with a facility of
three-point bending. The flexural specimen was prepared according to ASTM-D790
standard, as shown in Fig. 2b. Five specimens of tensile and flexural strength were
tested for each blend compositions.
Dynamic mechanical analysis
DMA measurements were carried out in Du Pont 983 analyzer. Dynamic mechanical
properties were measured in tensile mode at a fixed frequency of 1 Hz under nitrogen
gas purging. The specimen was heated from 34 to 185 °C with a heating rate of
5 °C/min. The quantities of the storage modulus (E  ), loss modulus (E  ), loss factor
(tan δ), and temperature were recorded in the data logger. These data were further
used to construct the damping characteristics curve.
Dynamic Mechanical Analysis of Flax … 479

Fig. 2 Fabricated samples a Tensile and b flexural samples

3 Results and Discussion

3.1 Tensile and Flexural Strength

The tensile strength and modulus of blend composites are shown in Fig. 3a, b. This
figure shows that tensile strength and modulus of blend composites increased with
increase in polyurethane composition. An addition of polyurethane blend provides
significant changes in the network structure. It reduces the tensile strength when
polyurethane composition was exceeding 10%. Increases in polyurethane resin con-
tent above the 15 and 20% (BC15 and BC20) the failure of composite occurs under
lower tensile stress. It can be seen that composite BC10 blend composition shows a
higher tensile strength. The improvement of tensile strength was 25.5% which is more
than that of base composite (BC0). This is due to saturation of polyurethane inter-
penetrating network with epoxy rings [16]. Similarly, BC10 blend composite shows
a maximum tensile modulus. It was almost 44.4% greater than BC0 blend composite.

Fig. 3 a Tensile strength and b tensile modulus of blend composites

480 K. Rajkumar et al.

Fig. 4 a Flexural strength and b flexural modulus of blend composites

While, BC5 and BC15 blend composites observed higher tensile modulus by 22.2
and 31.1% than the BC0 blend composite.
The flexural strength and modulus of blend composites are shown in Fig. 4a, b.
Figure 4a shows a flexural strength which increases with increasing polyurethane
blend composition. This could be an addition of polyurethane blend which signifi-
cantly increases interlinking networks of the epoxy matrix. IPNs are improving the
strength and flexi of the composite. Additionally, flax fiber resisting bending stress
induced upon it. Higher flexural strength was obtained from the BC10 blend com-
posite. It is observed that BC10 was 40.4% higher than the BC0 blend composite.
Moreover, BC5, BC15, and BC20 blend composites are showing 12.3, 25.8, and
9% higher than BC0 blend composite. Similarly, BC10 blend composite reaches a
maximum flexural modulus and of 21.5% more than the BC0 blend composite.

3.2 Effect of Composition on Dynamic Mechanical Properties

Storage modulus (E )

The variation of storage modulus of blend composites as a function of temperature at
1 Hz frequency is shown in Fig. 5. It is observed that the storage modulus increases
with an increase in polyurethane blend composition. The blend composites of BC15
and BC20 were gradually decreasing its storage modulus value when compared to
BC10. By comparing the different polymer blend composites, it is found that the
value of E  increases up to BC10 and then starts to decrease with polyurethane
composition. In the glassy region, the maximum and lowest value of E  was found to
be 1.30 E+010 Pa and 0.80 E Pa for blend composites BC0. The storage modulus of
the blend composites decreases as temperature increases due to the loss in stiffness
of fibers [17]. In the rubbery region, it is observed that the composite BC10 has the
highest value of E  whereas composite BC20 has the lowest value. The lowest value
of E  indicates an increase in molecular mobility in the blend composites matrix.
Dynamic Mechanical Analysis of Flax … 481

Fig. 5 Variation of storage modulus of blend composites at 1 Hz

The high value of E  , specifically BC10, indicates a good adhesion between polymer
blend composites. This fact shows that blend composite BC10 has a better interface
bonding as compared to other blend composites. This figure also shows a shifting
of glass transition temperature toward the left side as increasing in polyurethane
composition in the matrix.
Loss modulus (E )
The variation of loss modulus of blend composites as a function of temperature at 1 Hz
frequency is shown in Fig. 6. Similar to storage modulus, the value of loss modulus
of blend composite was also found to be increasing with increase in polyurethane
composition up to 10%. It can be observed that the value of E  increases up to the
glass transition temperature (T g ) and then decreasing with temperature. In glassy
region, the maximum values of loss modulus of blend composites BC0, BC5, BC10,
BC15, and BC20 are found as 5.05E+7 Pa, 5.96E+7 Pa, 7.38E+7 Pa, 6.4E+8 Pa, and
1.87E+7 Pa, respectively. The peak height (maximum value) of loss modulus curve
shows the T g of corresponding composites. The blend composite BC10 shows the
highest value of T g (93 °C) which indicates better thermal stability than the other
blend composites.
Loss factor (tan δ)
The variation of loss factor of blend composite as a function of temperature is shown
in Fig. 7. The temperature corresponding to the tan δ peak is known as the glass
transition temperature of the composite. The order of maximum value of tan δ for
blend composites of BC10, BC5, BC0, BC15, and BC20 is found that 0.452, 0.448,
0.428, 0.417, and 0.378, respectively. The highest peak of tan δ value was found for
BC10 composite which means that better damping property reached as compared
482 K. Rajkumar et al.

Fig. 6 Variation of loss modulus of blend composites at 1 Hz

Fig. 7 Shows the variation of loss factor blend composites at 1 Hz

to other blend composites [19]. The lower value of tan δ is found for the blend
composite BC20. Damping properties of different blend composite are shown in
Table 2. The capability of damping was improving up to BC10 and begins to decrease
as polyurethane composition increases. This clearly observed from Table 2. The
similar trend is observed for the T g of composites. It can be noted that the temperature
Dynamic Mechanical Analysis of Flax … 483

Table 2 Damping properties of different blend composite

Types of blend composite (tan δ)max T g (°C) Temperature range (tan δ > 0.3)
BC0 0.428 86 78–94 (16 °C)
BC5 0.448 87 78–95 (17 °C)
BC10 0.452 93 83–102 (19 °C)
BC15 0.417 89 81–97 (16 °C)
BC20 0.378 90 82–98 (16 °C)

range of BC10 composite was 19 °C. This is indicative of damping characteristics

which exist for a longer temperature range when compared to other composites.

4 Conclusions

Dynamic mechanical properties of polymer blend reinforced with flax fiber compos-
ite were studied and the following conclusions are drawn.
1. The tensile strength and modulus of BC10 composition reaches a maximum
value and then decreases with further increase in blend composition. The tensile
strength and its modulus improvement were 25.5 and 40.4% higher than that of
base composite (BC0). This is attributed to the interpenetrating networks (IPNs)
of polyurethane with epoxy cross-link networks.
2. The higher flexural strength and flexural modulus were obtained for the BC10
blend composite. The flexural strength and modulus were increased by 40.4 and
21.5% when compared to the BC0 composite.
3. It is found that the value of storage modulus increases with increase in
polyurethane composition up to 10% in the matrix and then decreases with further
increase in polyurethane composition. Due to the reduction in stiffness of matrix
and flexi or rubbery nature, the loss modulus (E  ) of blend composite BC10 shows
the highest value of T g (93 °C) which shows better thermal stability. The blend
composite BC10 reveals a higher value of tan δ, which is indicative of higher
dampness. Among the dynamic mechanical tested composites, the dampness of
BC10 exists for a longer temperature range.

References

1. Ali M, Liu A, Sou H, Chouw N (2012) Mechanical and dynamic properties of coconut fiber
reinforced concrete. Constr Build Mater 30:814–825
2. Yan LB, Chouw N, Yuan XW (2012) Improving the mechanical properties of natural fiber fabric
reinforced epoxy composites by alkali treatment. J Reinforced Plastics Compos 31(6):425–437
484 K. Rajkumar et al.

3. Liang S, Gning PB, Guillaumat L (2012) A comparative study of fatigue behaviour of flax/epoxy
and glass/epoxy composites. Compos Sci Technol 72(5):535–543
4. Ku H, Wang H, Pattarachaiyakoop N, Trada M (2011) A review on the tensile properties of
natural fiber reinforced polymer composites. Compos Part B: Eng 42(4):856–873
5. Sgriccia N, Hawley MC, Misra M (2008) Characterization of natural fiber surfaces and natural
fiber composites. Compos Part A: Appl Sci Manuf 39(10):1632–1637
6. Genc G, El Hafidi A, Gning PB (2012) Comparison of the mechanical properties of flax and
glass fiber composite materials. J VibroEng 14(2):783
7. Ramraji K, Rajkumar K, Sabarinathan P (2019) Tailoring of tensile and dynamic thermo-
mechanical properties of interleaved chemical-treated fine almond shell particulate flax fiber
stacked vinyl ester polymeric composites. In: Proceedings of the Institution of Mechanical
Engineers, Part L: Journal of Materials: Design and Applications, p 1464420719849616
8. Li X, Tabil LG, Panigrahi S (2007) Chemical treatments of natural fiber for use in natural
fiber-reinforced composites: a review. J Polym Environ 15(1):25–33
9. May CA (1998) Epoxy resins. Marcel Dekker, New York
10. Lee L, Nevile K (1986) Handbooks of epoxy resins. McGraw-Hill, New York
11. Maxwell D, Young RJ, Kinloch AJ (1984) Hybrid particulate-filled epoxy-polymers. J Mater
Sci Lett 3(1):9–12
12. Kinloch AJ, Shaw SJ, Hunston DL (1983) Deformation and fracture behaviour of a rubber-
toughened epoxy: 2. Failure criteria. Polymer 24(10):1355–1363
13. Chen S, Tian Y, Chen L, Hu T (2006) Epoxy resin/polyurethane hybrid networks synthesized
by frontal polymerization. Chem Mater 18(8):2159–2163
14. Kostrzewa M, Hausnerová B, Bakar M, Dalka M (2011) Property evaluation and structure
analysis of polyurethane/epoxy graft interpenetrating polymer networks. J Appl Polym Sci
122(3):1722–1730
15. Chen CH, Chen MH (2006) Synthesis, thermal properties, and morphology of blocked
polyurethane/epoxy full-interpenetrating polymer network. J Appl Polym Sci 100(1):323–328
16. Jin H, Zhang Y, Wang C, Sun Y, Yuan Z, Pan Y, Xie H, Cheng R (2014) Thermal, mechanical,
and morphological properties of soybean oil-based polyurethane/epoxy resin interpenetrating
polymer networks (IPNs). J Thermal Anal Calorimetry 117(2):773–781
17. Chern YC, Tseng SM, Hsieh KH (1999) Damping properties of interpenetrating polymer
networks of polyurethane-modified epoxy and polyurethanes. J Appl Polym Sci 74(2):328–335
18. Jawaid M, Khalil HA, Hassan A, Dungani R, Hadiyane A (2013) Effect of jute fiber loading
on tensile and dynamic mechanical properties of oil palm epoxy composites. Compos Part B:
Eng 45(1):619–624
19. Wang C, Jia J (2014) Damping and mechanical properties of polyol cross-linked
polyurethane/epoxy interpenetrating polymer networks. High Perform Polym 26(2):240–244
20. Mohammed L, Ansari MN, Pua G, Jawaid M, Islam MS (2015) A review on natural fiber
reinforced polymer composite and its applications. Int J Polym Sci
21. Chawla KK (2012) Composite materials: science and engineering. Springer Science & Business
Media
Material Processing Technology
Comparative Study of Ball Nose and Flat
End Milling on A356 Alloy/SiCp Metal
Matrix Composite

K. Jayakumar

Abstract The end milling process is mainly used to quickly remove large amounts
of material during heavier machining operations. It is most suitable for making slots,
profiles and plunge cutting. The ball nose end milling is used to produce complex
3D sculptured surfaces for making moulds and dies. But, a variation of cutting force
and surface roughness in both end milling processes is not well understood due
to the constant change in cutting tool–workpiece engagement during these milling
processes.
Aluminium alloy–SiC particle-reinforced metal matrix composites (MMCs) are
used in many industries in different shapes through different machining processes.
Therefore, the present study focuses on the comparative performance study of ball
end and flat end milling on A356/SiC MMC.
Composite was prepared from A356 alloy powders (avg. size-50 µm) and SiC
powders (avg. size-1 µm). MMC with 90 vol.% A356 alloy powder and 10
vol.% SiC powders were synthesized using vacuum hot pressing (VHP) process
at 600 °C temperature with 25 MPa pressure. VHP is mainly used to achieve high
densification with uniform distribution of reinforcement.
Both ball and flat end mill machining experiments were conducted as per Taguchi
L9 array. Experiments were carried out by varying cutting speed (100, 150 and
200 m/min), feed (0.1, 0.2 and 0.3 mm/rev) and depth of cut (0.4, 0.7 and 1 mm)
for both machining processes. From the experimental results, it is observed that
the flat end mill gave more material removal rate (MRR) in cm3 /min with low
surface roughness (Ra) value than ball nose milling.

Keywords A356/SiCp MMC · Vacuum hot pressing technique · Flat end milling ·
Ball nose end milling · Material removal rate · Surface roughness

K. Jayakumar (B)
Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Kalavakkam, Chennai 603110, India
e-mail: kjayakumar@ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 487

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_43
488 K. Jayakumar

1 Introduction

Metal matrix composites (MMCs) are materials which combine metallic matrix
and hard ceramic reinforcement to manufacture composite materials with supe-
rior properties to conventional metallic alloys. Among modern composite materi-
als, particulate-reinforced metal matrix composites (PMMCs) are finding increased
applications due to their favourable properties, including high mechanical properties
and good wear and corrosion resistance. Silicon carbide particle (SiC)-reinforced
aluminium-based MMCs are among the most common MMCs and commercially
available due to their economic production. Increasing quantities of metal–matrix
composites are being used to substitute conventional materials in many applica-
tions, particularly in the automobile, aerospace, electronics, armour and recreational
industries, owing to increasing performance requirements [8]. In the modern manu-
facturing process, the objective is to produce low-cost, high-quality products in tiny
time. Automated and flexible manufacturing systems are working for that purpose
along with CNC machines that are capable of achieving high accuracy with low
processing time. End milling is the second most common process (after turning)
for metal cutting and especially for the finishing of machined flat parts, in many
applications. Also, ball nose end milling is widely used for generating 3D sculptured
surfaces for making moulds and dies.
Although SiC particle-reinforced aluminium matrix composites are regarded as
one of the new engineering materials that have many applications, the abrasive parti-
cles reinforcement used in these materials makes them difficult to machine especially
if the reinforcement is more than 10 vol. % of base matrix and also brittleness started
increasing. If there is no solution to the machining problem, it is most likely that such
composites would not find wide application. Many works have been carried about
the machining of this material with turning, drilling, grinding and electric discharge
grinding, etc. Kuram et al. [1] studied the effect of spindle speed, feed and depth of cut
on tool wear, cutting force and surface roughness during microball nose end milling
of Al 7075 material. They found the optimum process parameter for achieving higher
machinability of the selected alloy. End milling study was carried out on Al 2124
alloy with 20% B4 C MMC to analyse the effect of machining parameter on surface
roughness [2]. Conclusions were medium cutting speed, high feed and low depth
of cut are suitable for machining the composite with low surface roughness value.
Al 2014-T6 alloy was machined using high-speed ball nose end milling process by
Dikshit et al. [3]. They identified surface roughness value which was influenced by
cutting speed and axial depth of cut then the radial depth of cut and feed. Jayakumar
[4] did an experimental investigation on ball nose end milling study on A356-based
MMC and identified the best machining parameter for getting low cutting force and
surface roughness value. Liu et al. [5] investigated microslot milling of 45% SiCp/Al
composite using a single-edge PCD end mill of ∅ 1 mm. They got optimum condition
to get minimum Ra of 0.4 µm.
Optimization of parameters in the end milling of Al Al5059/SiC/MoS2 aluminium
hybrid metal matrix composites was carried out by Daniel et al. [6]. Major findings
Comparative Study of Ball Nose and Flat … 489

from the study increased in particle size of the composites tend to decrease in surface
roughness, temperature, cutting forces and increased the MRR. Recently, ball nose
end milling on Inconel 718 was conducted to study the effect of radial and axial depth
of cut, feed and cutting speed on surface finish using minimum quantity lubrication
[7]. It was identified that feed effects on the quality of surface finish whilst the radial
depth of cut controls Ra variation.
From the available literature, it was found that ball nose end milling studies of
A356 Al-based MMCs are in the initial stage and it needs further detailed analysis.
Hence, the present study focuses on ball nose end milling of A356 alloy with 10
vol.% SiC MMC and its comparison with flat end milling machining process using
the same process parameter. The output performance measures considered for both
processes are surface roughness and MRR.

2 Experimental Details

2.1 Synthesis of Composite

A356 alloy powder (average size of 50 µm) and SiC particles (average size of 1 µm)
of 10 vol. % were used as raw materials for processing of MMC. The major challenge
in synthesizing the particulate MMCs is uniform mixing of finer reinforcement in the
matrix. Vacuum hot pressing (VHP) shared with powder metallurgy (P/M) process
was used in the present study to consolidate and process the composite with good
densification (relative density—nearly 95%) and more isotropic in nature. Hence,
the P/M process was used rather than stir casting of this alloy. The vacuum hot press
used in this work is available at VSSC, Trivandrum, Kerala.
In sintering or VHP, at first, mixed powders were cold compacted at a pressure
of 15 MPa. Then, the hot compaction was carried out at 600 °C temperature with
25 MPa pressure in a vacuum (10−5 mbar) to avoid oxidation of A356 alloy. The load
and temperature were maintained to a dwell time of 30 min for effective densification
of the powder compacts. The above-mentioned pressure and temperature were taken
from the literature [4]. The hot pressed blank is shown in Fig. 1 which has nearly
125 mm in diameter and 20 mm thickness. Figure 2 shows the microstructure images
of the composite taken from scanning electron microscope (SEM) at 1000 × . From
the image, it was observed that the SiC particles are fairly well distributed on the
matrix.

2.2 Milling Study

Figure 3a shows the experimental set-up. Rectangular piece of size 72 × 62 ×

20 mm (Fig. 3b) was cut from the processed composite and end milling studies on
490 K. Jayakumar

Fig. 1 Processed composite

blank

Fig. 2 Microstructure of
MMC with 10% SiC

the composite material were carried in dry condition. HSS ball nose cutter of Ø
16 mm with two uncoated carbide inserts (R 390-11 T3 04E-NL H13A) and HSS
end mill cutter of Ø 16 mm with two uncoated cemented carbide inserts (WIDIA
XDHT 09 03 08) were used in this experimentation. The factors contributing to
surface roughness and MRR during end milling of MMCs are work material (vol. %
of reinforcement, particle size, type of ceramic particle and the base alloy chemical
composition, etc.), machining parameters (up and down milling, cutting speed, feed
and depth of cut), tool material and geometry, etc. Among these, cutting speed (100,
150, 200 m/min), feed (0.1, 0.2, 0.3 mm/rev) and axial depth of cut (0.4, 0.7, 1 mm)
were considered as the process parameter for the present work.
Comparative Study of Ball Nose and Flat … 491

Fig. 3 a Ball nose end milling b machined workpiece after ball nose milling c machined workpiece
after flat end milling

A good experimental design can decrease the number of experiments at the same
time the error of the model does not change intensely. In the present work, Taguchi’s
L9 orthogonal array was used for three-level experiments with nine different condi-
tions for flat end milling and ball nose end milling process for designing the exper-
iments and is given in Table 1. During the experiments, inserts were indexed and
changed at the end of five experiments to reduce the wear in the inserts. Slots were
made for 16 mm width during both milling processes for each experiment.
The roughness (Ra) of the machined samples was measured using roughness tester
(Model: Form Talysurf 50 Infra). Measurements were taken using the cut-off length
(λc ) of 0.8 mm and no. of samples of 5 using stylus with diamond tip of Ø 2 µm and
an average of 3 readings taken at different locations in each slot was reported.
The weight of the workpieces before and after the experiments has been measured
using a digital weighing balance (accuracy 1 mg). From the weight loss, actual density
and machining time, MRR was calculated.
492 K. Jayakumar

Table 1 Experimental results

Exp. V c -m/min f-mm/rev d of MRR MRR Surface Surface
No. cut-mm (cm3 /min)-ball (cm3 /min)-flat roughness roughness
nose milling end mill (µm)-ball (µm)-flat
nose end mill
milling
1 100 0.1 0.4 0.30 1.19 0.564 0.41
2 100 0.2 0.7 1.58 7.18 2.491 1.75
3 100 0.3 1 3.59 8.82 1.797 1.58
4 150 0.1 0.7 0.96 2.91 1.202 2.44
5 150 0.2 1 3.62 7.23 2.032 1.63
6 150 0.3 0.4 1.74 5.51 2.804 2.51
7 200 0.1 1 2.11 5.82 1.328 1.71
8 200 0.2 0.4 1.20 10.86 1.223 1.18
9 200 0.3 0.7 4.57 13.43 1.331 1.17

3 Results and Discussions

3.1 Comparison of MRR Between Ball Nose and Flat End

Milling Processes

Material removal rate (MRR) is crucial in manufacturing industries for higher pro-
duction and it can be attained by selecting proper machining parameters like cutting
speed, feed and depth of cut. From Fig. 4, it was found that the MRR is increased
with an increase in cutting speed, feed and depth of cut for both ball nose end milling
and flat end milling processes. Increase in cutting speed caused the material to be
soft due to the increase in temperature at the shear zone which reduced the strength
of the composite workpiece that caused the effortless removal of chips. Increase in
feed and depth of cut (DOC) increased the normal pressure, size of chip cut per tooth
and chip load which improved the MRR. MRR is low for ball nose milling than flat
end milling. The reason is that more compressive force is given to workpiece during
ball nose end milling than flat end milling due to inadequate space available for the
chip flow.

3.2 Comparison of Surface Roughness Between Ball Nose

and Flat End Milling Processes

From Fig. 5a, it was found that surface roughness increased initially and then later
it decreased for both milling processes. The reason is the development of built-up
edge (BUE) whilst machining Al-based MMCs at lower cutting speed. BUE is break
off when it becomes large at medium cutting speed. When this happens, broken
Comparative Study of Ball Nose and Flat … 493

10 (a) 10
(b)
MRR (cm3/min)

MRR (cm3/min)
5 5

0 0
100 150 200 0.1 0.2 0.3
Cutting speed (m/min) feed (mm/rev)
Ball nose Flat end mill Ball nose Flat end mill

10
(c)
MRR (cm3/min)

0
0.4 0.7 1
depth of cut (mm)
Ball nose Flat end mill

Fig. 4 a Cutting speed versus MRR (cm3 /min) b Feed versus MRR (cm3 /min) c Depth of cut
versus MRR (cm3 /min)

BUE may pass off the underside of the chip and it randomly scratches the already
machined surface and causes saw-toothed surface which finally makes the work
surface as rough. As the cutting speed increases, the BUE vanishes, chip fracture
decreases, thus the surface finish is improved. From Fig. 5b, the surface roughness of
the composite is increased with a rise in feed for both milling cutters. The explanation
is that, the feed rate has increased the distance between the two successive tool paths,
which leads to increase in Ra value. As the depth of cut increased, the cutting process
vibrations increases [8] and the development of high normal pressure on the rake
faces of the tool resulted in the formation of BUE that worsened the surface finish
(Fig. 5c). Surface roughness value was low for flat end mill cutting process than
ball nose end milling process. Compared with the flat end mill profile, the surface
produced by ball end mill is complex and a new cut is overlapped which generated a
new surface roughness by removing the earlier cut thus the non-uniform feed marks
being formed on the machined surface which increased surface roughness value.
Using design expert software’s optimization option, cutting speed of 186 m/min,
feed 0.29 mm/rev and 0.9 mm depth of cut were recommended as an optimum
condition for ball nose end milling process to get high MRR and low surface value.
Similarly, cutting speed of 200 m/min, feed 0.3 mm/rev and 0.83 mm depth of cut
was recommended as an optimum condition for flat end milling process.
494 K. Jayakumar

(a) (b) 2
2.5

Surface roughness (μm)

Surface roughness (μm)

2 1.5

1.5
1
1
0.5 0.5

0
0
100
150 0.1
200 0.2
0.3
Cutting speed (m/min) feed (mm/rev)
Ball nose Flat end mill Ball nose Flat end mill
(c)
2
Surface roughness (μm)

1.5

0.5
0
0.4
0.7
1
depth of cut (mm)
Ball nose Flat end mill

Fig. 5 a Cutting speed versus surface roughness (µm) b Feed versus surface roughness (µm)
c Depth of cut versus surface roughness (µm)

4 Conclusions

In the present work, fairly homogeneous MMC of A356-10% SiC powders was
synthesized using vacuum hot pressing technique.
The effects of cutting speed, feed and depth of cut on MRR and surface roughness
during flat end milling and ball nose milling of Al-based MMC were investigated
using Taguchi’s L9 experimental design method.
Flat end milling process gave better responses than ball nose mill in terms of high
MRR with low surface roughness value.
The optimum machining parameter was found to improve the machinability of
the MMC for both milling processes.
Comparative Study of Ball Nose and Flat … 495

References

1. Kuram E, Ozcelik B (2013) Multi-objective optimization using Taguchi based grey relational
analysis for micro-milling of Al7075 material with ball nose end mill. Measurement 46(6):1849–
1864
2. Boswell B, Islam MN, Davies IJ, Pramanik A (2017) Effect of machining parameters on the
surface finish of a metal matrix composite under dry cutting conditions. Proc Inst Mech Eng
Part B: J Eng Manuf 231(6):913–923
3. Dikshit MK, Puri AB, Maity A (2017) Optimization of surface roughness in ball-end milling
using teaching-learning-based optimization and response surface methodology. Proc Inst Mech
Eng Part B: J Eng Manuf 231(14):2596–2607
4. Jayakumar K (2018) Study of cutting force and surface roughness in Ball nose end milling of
vacuum hot pressed A356 alloy/SiCp metal matrix composite. Mater Today: Proc 5(2):6526–
6533
5. Liu J, Cheng K, Ding H, Chen S, Zhao L (2018) An investigation of the influence of phases’
removal ways on surface quality in micro milling SiCp/Al composites. Proc CIRP 71:59–64
6. Ajith Arul Daniel S, Pugazhenthi R, Kumar R, Vijayananth S (2019) Multi objective predic-
tion and optimization of control parameters in the milling of aluminium hybrid metal matrix
composites using ANN and Taguchi-grey relational analysis. In: Defence technology
7. Kasima MS, Hafiza MSA, Ghanib JA, Haronb CHC, Izamshah R, Sundic SA, Mohamed SB,
Othman IS (2019) Investigation of surface topology in ball nose end milling process of Inconel
718. Wear 426:1318–1326
8. Niu Z, Cheng K (2019) An experimental investigation on surface generation in ultraprecision
machining of particle reinforced metal matrix composites. Int J Adv Manuf Technol 103:1–9
Numerical Modeling of Orthogonal
Machining Process Using Smoothed
Particle Hydrodynamics—A Parametric
Study

Sarath Babu Thekkoot Surendran , C. S. Sumesh , and Ajith Ramesh

Abstract Some of the challenges faced during numerical modeling of machining

processes using finite element method are: extreme deformations, complicated and
discontinuous contact conditions between the workpiece and cutting tool, and the
chances of self-contact due to chip curling. Another difficulty encountered in FEM
is in the form of mass and energy losses. Lately, the smoothed particle hydrodynam-
ics (SPH) method has developed as a potential alternative for modeling machining
processes due to its ability to handle severe deformations while avoiding energy
losses and mass losses encountered by traditional FE Model. This method has been
implemented in commercial finite element package Abaqus, for solving problems
involving localized severe deformations. Several control parameters are used in a typ-
ical SPH formulation. The purpose of this research work is to investigate the influence
of the important factors such as the type of SPH formulation, mass scaling, particle
density, artificial bulk viscosity, and the smoothing length in the numerical modeling
of orthogonal machining of AISI 1045 steel. The challenges involved in accurately
modeling this highly nonlinear problem is handled using the Abaqus/Explicit inte-
gration scheme along with the Johnson–Cook material model. Results from this
parametric study are validated with the results from previously published literatures.

Keywords Finite element method (FEM) · Smoothed particle hydrodynamics

(SPH) · Abaqus/explicit · Orthogonal machining · Johnson–cook material model ·
Smoothing length

1 Introduction

Numerical modeling normally gives close to accurate results for problems including
nonlinear ones which analytical methods find difficult to handle. They are neither
costly nor time-consuming as compared to experimental analysis. Machining is a

S. B. T. Surendran · C. S. Sumesh (B) · A. Ramesh

Department of Mechanical Engineering, Amrita School of Engineering, Coimbatore, Amrita
Vishwa Vidyapeetham, India
e-mail: cs_sumesh@cb.amrita.edu

© Springer Nature Singapore Pte Ltd. 2021 497

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_44
498 S. B. T. Surendran et al.

controlled material removal process in which a raw material is converted into desired
shape and size. Orthogonal machining is a machining process in which the cutting
edge of the tool is perpendicular to the direction of motion of the cutting tool. AISI
1045 steel used in this study is a medium carbon steel that has good machinability,
good weldability, and high strength and impact properties both for hot rolled and
normalized conditions. It is widely used in gears, shafts, connecting rods, crankshafts,
hydraulic clamps, etc.
Orthogonal machining uses a wedge-shaped tool in which the cutting edge is
perpendicular to the cutting direction. The chip is formed from the shear deforma-
tion along the shear plane. The important tool geometry angles are rake angle and
clearance angle. The direction of chip flow is determined by the rake angle, whereas
clearance between tool and workpiece is endowed by the clearance angle. This paper
presents numerical modeling of orthogonal machining using SPH. Numerical mod-
eling has many advantages over experimental and analytical methods. Numerical
methods can give close to accurate results for problems even when analytical methods
fail to solve the problem. It is cheaper compared to experimental analysis.
Smoothed particle hydrodynamics, abbreviated as SPH, is one among the most
popular particle methods available. The SPH particles move in space following the
conservation laws of continuum fluid dynamics. It has been proved stable to solve
engineering problems. Hence, commercial FE packages like Abaqus/Explicit and LS-
Dyna incorporated this method in their codes. In this research work, Abaqus/Explicit
package is used for modeling.
In the literature, there are many researchers who have done research in the field
of numerical modeling, orthogonal machining, smoothed particle hydrodynamics
(SPH) and combination of these fields. Given below are details of some of the previous
research work related to the machining area.
Niu et al. [1] introduced an improved SPH method and used it for modeling
orthogonal cutting and predicted the cutting force and chip morphology under dif-
ferent cutting conditions. This modification is achieved by approximating density
and kernel gradient. Olleak and El-Hofy [2] used SPH method in the modeling
of high-speed machining. The paper also predicted the chip morphology and cut-
ting forces during high-speed orthogonal machining of Ti6Al4V. Ramesh et al. [3]
predicted the deformation and machining characteristics for high-speed orthogonal
machining of Ti6Al4V using a detailed FE model. The influence of feed rate, spindle
speed, and rake angle on output parameters such as cutting force and surface finish
are studied. The paper also tackled the issue of self-contact using a unique model.
Avachat and Cherukuri [4] studied the SPH control parameters in detail in modeling
orthogonal machining. The paper evaluated the effects of various parameters like
smoothing length, type of SPH formulation, and particle density. The influence of
the above-mentioned parameters on stress distribution as well as chip morphology
of orthogonal machining of AISI 1045 steel is also studied. Bagci et al. [5] presented
the simulation and analysis of 3D machining using SPH. AISI H13 Steel was used in
the model. The model accurately predicted the cutting forces. M.B. Liu and G.R. Liu
[6] were the people who wrote the first book on SPH. In the year 2010, the authors
Numerical Modeling of Orthogonal Machining … 499

published a paper discussing the SPH method and its latest development. This paper
gives a very good background on SPH.
Kershah et al. [7] estimated cutting coefficients of AISI 1045 steel using finite ele-
ment modeling. The paper also contains a parametric study which correlates cutting
parameters to cutting coefficients. Madaja and Píškaa [8] presented a comparison of
SPH results with experimental results and FE model result for orthogonal cutting
simulation of Al 2024-T351. The study found that as particle spacing reduces, the
particles tend to form segmented chips. Sumesh and Ramesh [9] studied the effect of
process parameters like depth of cut, cutting speed, and feed rate on surface integrity
of turned Al 6061–T6 alloys using RSM. Zetterberg [10] discussed the important
parameters and different possibilities to form a chip in metal cutting simulation in
Abaqus/Explicit. It emphasizes on how the flank wear and hardening implemented in
the material model affects the formation of the chip. Villumsen et al. [11] predicted
the cutting forces of orthogonal machining using the SPH method and compared with
the existing literature to validate the model. The authors also carried out a sensitivity
analysis of machining parameters. Haddag et al. [12] analyzed the effect of orthog-
onal machining parameters on chip segmentation of Al2024-T351 alloy. The paper
also considered the effect of feed and rake angle on chip morphology and cutting
force.
Barge et al. [13] developed a numerical model and studied the physical and pro-
cess parameters dependence using finite element modeling. The paper discusses the
numerical modeling of orthogonal machining. Zhang et al. [14] used digital image
correlation technique to obtain the deformation field containing elastic as well as
plastic parts. A new methodology has been proposed to obtain the cutting stress
field. Limido et al. [15] evaluated SPH method for high-speed cutting modeling. The
developed approach in the paper is used for finding the cutting forces in ball end
milling. The paper used both a 2D model and a 3D model using the software LS-
Dyna. Chen et al. [16] developed FE model for high-speed machining of Ti6Al4V
alloy. Cleary et al. [17] used SPH for modeling thin-walled high-pressure die casting
using SPH. The paper compared the results with the experimental data and found that
the results were accurate. The study also performed flow analysis using the validated
model. Hu et al. [18] used SPH for modeling high-pressure die casting. The paper
compared SPH results with the finite element method and found that they are com-
parable and thus validated the model. The paper showed that the method is able to
capture the fine details of the free surface motion, including plume shape, frequency,
and phase of oscillation.
The above-mentioned research works have helped extensively in the presented
research. In this paper, the influence of the important factors such as type of SPH
formulation, particle density, smoothing length, mass scaling, and artificial bulk
viscosity has been studied. The challenges involved in accurately modeling this
highly nonlinear problem is handled using the Abaqus/Explicit integration scheme
along with the Johnson–Cook material model.
500 S. B. T. Surendran et al.

2 Smoothed Particle Hydrodynamics

2.1 History

Smoothed particle hydrodynamics is a mesh-free Lagrangian particle method. This

method mainly uses Lagrangian formulation along with particle approximation. A
group of particles having material properties assigned individually represents the
state of the system. The particles in this system behave depending upon the gov-
erning equations. A pre-defined mesh is not required to connect particles in SPH
method, like in conventional FEM. In the year 1977, Lucy, Gingold, and Monaghan
invented the method to solve astrophysical problems. The first book on SPH was
published by Liu GR and Liu MB in the year 2003 [19]. The authors were working
on high explosives and found that the existing grid-based methods were incapable of
handling the phenomena due to excessive mesh distortion. The search for advanced
methods ended up in SPH which later became popular. It was further extended to solid
mechanics problems involving large deformations which lead to the incorporation
of this method in the commercial FE packages like Abaqus/Explicit and LS-Dyna.
SPH can naturally handle problems with extremely large deformations. Unlike other
mesh-free methods which use nodes as only interpolation points, SPH particles carry
material properties.

2.2 Mathematical Formulation

Function f (y) is approximated as


< f (y) >= w(y − y, h) f (y  )d (1)
Ω

where
< > —the approximation
w(y − y ,h)—the kernel function
h—the smoothing length
y —the particle within the smoothing length

The conditions satisfied by the kernel function are:

w(y − y ,h) = 0 for | y − y | ≥ 2 h
The normalization condition:

w(y − y  , h)dy  = 1 (2)
Ω
Numerical Modeling of Orthogonal Machining … 501

As h tends to zero, the kernel function becomes the Dirac delta function,
i.e., limh→0 w(y − y , h) = δ(y − y , h)
Thus, limh→0 <f (y) > = f (y)
So, for N discretized points, the function approximation becomes:


N
mj
< f (y ) >=
i
f (y j )w(y i − y j , h) (3)
j=1
j

where
mj
j
—Volume of the particle j with mass m j

The value of a variable is computed by the contributions from the neighboring

particles j for any particle i, where j = 1, 2…N. Here, the kernel function is nonzero.
Cubic B-spline is the commonly used kernel function. It was proposed by Monaghan
given by,
⎧ 
⎨ 1 − 2 r 2 + 4 r 3 , 0 ≤ r ≤ 1;
3 3
c
W (r, h) = D × 41 (2 − r )3 , 1 ≤ r ≤ 2; (4)
h ⎩
0, Otherwise.

where

y − y
r=
h
D = the number of spatial dimensions and C = scale factor
C = 2/3 for D = 1,
C = 10/7π for D = 2,
C = 1/π for D = 3.

3 Numerical Modeling of Orthogonal Machining Using

SPH

The numerical model uses a tungsten carbide rigid tool. It has a rake angle of 8°
and a clearance angle of 10°. The workpiece material used is AISI 1045 steel with
dimensions 3.6 mm × 0.3 mm. The out-of-plane thickness is 0.04 mm. The material
properties of workpiece are displayed in Table 1. Equation of state is used along with
the material properties with US-UP formulation. The remaining parameters used in
the model are as follows [4]:
(a) Speed of sound in the workpiece
(b) c = 6.35 × 106 mm/s
 502

Table 1 Material properties of AISI 1045 steel [4]

A (MPa) B (MPa) n C m Room temperature Melting Reference strain Density (kg/m3 ) Shear modulus
(°C) temperature (°C) rate (GPa)
553 600.8 0.234 0.013 1 27 1460 1 7800 76.923
S. B. T. Surendran et al.
Numerical Modeling of Orthogonal Machining … 503

(c) slope s = 0.0

(d) Gamma = 0.0
The stress–strain relationship for the material is defined using Johnson–Cook (J-
C) model. J-C model is found to provide a proper description of the material behavior
which is unique to nonlinear large deformation processes by considering the high
strain rates, temperature-dependent visco-plasticity, and large strains [20].

 

ε  
T − Tr

σ = A + Bεn 1 + C ln 1− (5)
ε0 Tm − Tr

where
εn —equivalent plastic strain

A, B, C, m and n—material constants

ε
ε0
—plasticstrain rate
T m—Reference melt temperature
T r —Reference temperature
The model uses static-kinetic exponential decay as the friction formulation. The
static coefficient of friction is taken as 0.5, dynamic coefficient value as 0.3, and
decay coefficient value as 0.1. The default values of artificial bulk viscosity factor
Q1 is changed from 0.06 to 0.5 and Q2 is changed from 1.2 to 1.5 [4].
Direct SPH continuum particle analyses technique has been used for validation
instead of FE conversion to SPH. In the former technique, a node set was created and
inertia mass was assigned to create mass elements in the .inp file. After completing the
CAE model, the mass elements were renamed as PC3D to convert mass elements into
SPH particles, in the .inp file. The solid elements of the workpiece were deleted. Then
*mass definition was removed and the element set name of the continuum particle
elements (PC3D) was copied to the *Solid Section element set name. Characteristic
length was also defined in the *Solid Section. In the assembly module, a node-based
surface for the workpiece was created to define the contact. *Section Controls was
used to change the SPH to NSPH and also to define the smoothing length. Finally, the
contact between the workpiece and the tool was defined in the .inp file. An additional
damage model is not required while modeling using SPH method, which is another
added advantage.

4 Results and Discussions

Following numerical controls were used for SPH modeling of the machining pro-
cess. The influence of each parameter on the optimal modeling of the problem is
investigated.
504 S. B. T. Surendran et al.

Fig. 1 Tool passing without

interaction

4.1 Influence of Mass Scaling

Mass scaling is a numerical technique that is used to increase the time step by scaling
up the density in specific elements. Initially, a mass scaling factor of 10,000 was used
in the study. Initial results indicated the absence of any material removal, although
the tool had contact with the workpiece (see Fig. 1). Hence, a reduced mass scaling
factor of 100 was used and was found to initiate the material removal. However,
the computational time increased to a significantly large value. Therefore, no mass
scaling was used in this study for better results. It has been understood that an
increased mass scaling can lead to high inertial effects that can corrupt the results.

4.2 Need of Direct SPH

In this work, initially, only the bottom and left faces of the workpiece were encastered.
However, the results indicated that boundary conditions were not transferred to the
SPH particles. Figure 2 shows the end results of the above-mentioned condition. This
problem was resolved by using smoothed particle hydrodynamics instead of finite
element conversion to SPH particles [21]. In the SPH analyses, a node set had to be

Fig. 2 Encastered
workpiece moving along
with the tool
Numerical Modeling of Orthogonal Machining … 505

Fig. 3 Particles spreading

created and inertia mass had to be assigned to the nodes to create mass elements in
the .inp file. Later, these mass elements had to be replaced with continuum particle
elements (PC3D).

4.3 Contact Issues

Another major problem that was encountered was the inability to use general contact
algorithm. The reason being due to the extensive nodal penetration of tool nodes
to the workpiece surface. This was tackled by using the traditional master–slave,
surface-to-surface contact algorithm.

4.4 Necessity of Additional Non-default Constraints

The initial simulation results indicated the movement of SPH particles in the Z-
direction instead of flowing over the tool rake (see Fig. 3). This issue was resolved by
assigning encaster boundary conditions to the bottom most and left most nodes, and
a displacement/rotation boundary condition to the rest of the nodes of the workpiece
where the displacement along the Z-direction and rotations about Y and X axes were
constrained.

4.5 Effect of Smoothing Length

Smoothing length is the radius of the sphere of influence within which any parti-
cle may affect the neighboring particle. Initially a value of 0.005 was assumed for
smoothing length based on the calculation of half of the characteristic length [4].
This, however, leads to non-physical results (see Fig. 4). Then, a higher value of
0.02 was assumed which resulted in the termination of the simulation with an error.
506 S. B. T. Surendran et al.

Fig. 4 Non-physical result due to low value of smoothing value

Fig. 5 Non-physical result due to constant smoothing length

The error indicated that the smoothing length was too large as the number of parti-
cles exceeded the maximum specified. Between these two extreme values, a suitable
trade-off value was reached as 0.01, which was used as a constant smoothing length.
Figure 5 shows the results of constant smoothing length that lead to breaking up of
workpiece or flow of chips over the tool. This issue was resolved by using a variable
smoothing length option in the .inp file. Maximum number of particles within the
smoothing length was maintained at a default value of 140, since the computational
memory required to run the simulation was too high even for a slight increase from
the default value.
Influence of Artificial Bulk Viscosity and SPH formulation
The von Mises stress contour plot was found to be uniformly distributed for Q1
= 0.5 than the default value of Q1 = 0.06. Another important parameter was the
type of SPH formulation. The default SPH formulation was found to be inadequate
and inconsistent for modeling the process, and it leads to non-physical results (See
Fig. 6). Thus, NSPH and XSPH were used simultaneously instead of default SPH
formulation.

5 Validation

Although the numerical model was developed, the initial results were not found to
match with the literature since the simulation had many issues which are mentioned
in the previous section. After making necessary modifications based on the trial-and-
error simulations, the model was finally developed successfully. This work was able
Numerical Modeling of Orthogonal Machining … 507

Fig. 6 Non-physical result due to default SPH formulation

Fig. 7 Validated model

to reproduce the results obtained from the literature, Chinmay et al. [4] and Kershah
et al. [7]. The average stress in the shear zone was found to be matching with the
literature. The von Mises stress contours of the validated model are shown in Fig. 7.

6 Conclusions

This paper discusses the development of a numerical model for simulating orthogonal
turning of AISI 1045 steel using SPH method. The FE package Abaqus/Explicit was
used. The results obtained are found to be in good agreement with the results found
in the literature.
By varying SPH control parameters and boundary conditions, parametric studies
were performed. All the simulations were conducted by modifying the .inp file,
except for FE conversion to SPH results.
508 S. B. T. Surendran et al.

Influence of mass scaling factor was studied and as expected computational time
decreased with increasing the factor. However, for better modeling of orthogonal
machining, it was found that mass scaling need not be used as it changes the whole
dynamics.
Initially, the encaster boundary conditions given to the workpiece were not trans-
ferred during the conversion of elements to SPH particles. This issue was solved
using the direct SPH method instead of FE conversion to SPH method.
The significance of using variable smoothing length was also investigated. When
low smoothing length was used, the particles were found to fly out of the domain,
since there were a less number of particles in the sphere of influence. On the other
hand, the maximum number of particles in the sphere of influence was maintained
at a default value of 140. Even a slight increase from the default value was found to
increase the computational memory required to run the simulation.
It was found that when the correct SPH parameters were used, the computational
results obtained were in good agreement with the results from conventional FEM.
It was also found that SPH can naturally handle self-contact, mesh distortion, and
material separation. However, the SPH method has some limitations. The values of
some SPH parameters used in the model are obtained by trial-and-error methods.
Also, computational time was found to increase significantly due to particle sorting
and neighbor search.

References

1. Niu W, Mo R, Liu GR, Sun H, Dong X, Wang G (2017) Modeling of orthogonal cutting process
of A2024-T351 with an improved Sph method. Int J Adv Manuf Technol 95:905–919
2. Olleak AA, El-Hofy HA (2015) Prediction of cutting forces in high speed machining of
Ti6Al4 V using Sph method. In: Proceedings of the ASME 2015 International Manufacturing
Science And Engineering Conference MSEC 2015 June 812, 2015, Charlotte, North Carolina,
USA
3. Ramesh A, Sumesh CS, Abhilash PM, Rakesh S (2015) Finite element modelling of orthogonal
machining of hard to machine materials. Int J Mach Machinability Mater 17(6):543–568
4. Avachat CS, Cherukuri HP (2015) A parametric study of the modeling of orthogonal machining
using the smoothed particle hydrodynamics method. In: Proceedings of the ASME 2015 Inter-
national Mechanical Engineering Congress and Exposition IMECE 2015, 13–19 Nov 2015,
Houston, Texas
5. Bagci E (2011) 3-D Numerical analysis of orthogonal cutting process via Meshfree method.
Int J Phys Sci 6(6):1267–1282
6. Liu MB, Liu GR (2010) Smoothed particle hydrodynamics (Sph): an overview and recent
developments. Arch Comput Methods Eng 17:25–76
7. Kershah T (2000) Prediction of cutting coefficients during orthogonal metal cutting process
using FEA approach
8. Madaja M, Píškaa M (2013) On the Sph orthogonal cutting simulation of A2024-T351 alloy.
In: 14th CIRP Conference on Modeling of Machining Operations, Procedia Cirp, vol 8, pp
152–157
9. Sumesh CS, Ramesh A (2018) Numerical modelling and optimization of dry orthogonal turning
of Al 6061 T6 alloy. Periodica Polytechnica Mech Eng 62(3):196–202
Numerical Modeling of Orthogonal Machining … 509

10. Zetterberg M (2014) A critical overview of machining simulations in Abaqus. Degree Project.
In: Solid mechanics, Second Level Stockholm, Sweden
11. Villumsen MF, Fauerholdt TG (2008) Simulation of metal cutting using smooth particle
hydrodynamics. Ls-Dyna Anwenderforum, Bamberg
12. Haddag B, Atlati S, Nouari M, Barlier C, Zenasni M (2012) Analysis of the cutting parameters
influence during machining Aluminium alloy A2024-T351 with uncoated carbide inserts. Eng
Trans 60(1):31–39
13. Barge M, Rech J, Hamdi H, Bergheau J-M (2005) Numerical modelling of orthogonal
cutting numerical, physical and process parameters dependence. In: Proceedings of the
2nd International Conference on Manufacturing Engineering (ICMEN), 5–7 Oct 2005,
Kallithea-Chalkidiki, Greece
14. Zhang D, Zhang X-M, Xu W-J, Ding H (2016) Stress field analysis in orthogonal cutting process
using digital image correlation technique. J Manuf Sci Eng 139:031001-1 (March 2017)
15. Limido J, Espinosa C, Salaun M, Mabru C, Chieragatti R, Lacome JL (2011) Metal cutting
modelling Sph approach. Int J Mach Machinability Mater 9(¾)
16. Chen G, Ren C, Yang X, Jin X, Guo T (2011) Finite element simulation of high-speed machin-
ing of Titanium alloy (Ti-6Al-4V) based on ductile failure model. Int J Adv Manuf Technol
56:1027–1038
17. Cleary PW, Savage G, Ha J, Prakash M (2014) Flow analysis and validation of numerical
modelling for a thin walled high pressure die casting using Sph. Comp Part Mech. https://doi.
org/10.1007/s40962-017-0144-9
18. Hu MY, Cai JJ, Li N, Yu HL, Zhang Y, Sun B, Sun WL (2017) Flow modeling in high-pressure
die casting processes using Sph model. Am Foundry Soc. https://doi.org/10.1007/s40962-017-
0144-9
19. Liu GR, Liu MB (2003) Smoothed particle hydrodynamics: a Meshfree particle method. World
Scientic, Singapore
20. Dassault Systmes Simulia Corporation. ABAQUS Analysis user’s guide (2013)
21. Dassault Systmes Simulia Corporation. ABAQUS Analysis user’s guide (2016), Section 15.2.2
Acoustic Emission-Based Grinding
Wheel Sharpness Monitoring Using
Machine Learning Classifier

J. Revant, Rahul Sree Kumar, K. Rameshkumar, and D. S. B. Mouli

Abstract Grinding is one of the important secondary manufacturing processes used

to improve the dimensional accuracy, surface finish and geometric form of the compo-
nent. Grinding wheel consists of abrasive particles, which perform the metal removal
function. The sharpness of the grinding wheel is one of the important factors for
achieving the required surface geometry in the component. In this study, a simple
device used to measure the sharpness of the abrasive particles of the grinding wheel is
designed and fabricated. Aluminium oxide grinding wheel conditions are established
using the sharpness of the abrasive grinding wheel. Grinding process is monitored
using acoustic emission (AE) Sensor. AE features are extracted in time domain and
dominated features which contain useful information about the grinding wheel that
are identified. A correlation between grinding wheel condition and AE feature is
established using ANN-based machine learning classifier.

Keywords Grinding · Acoustic emission (AE) · Machine learning classifiers ·

ANN

1 Introduction

Grinding is an abrasive machining process carried out using grinding wheel. The
grinding wheel is embedded with randomly distributed abrasive particles like alu-
minium oxide, silicon carbide, diamond or cubic boron nitride. Abrasive particles of
different grit sizes are used in the machining process based on metal removal rate
and surface quality requirements. The grinding process starts with sharp abrasive
particles, and the abrasive particles will become dull and lose its cutting ability due
to wear. This phenomenon is called as the glazed condition of the grinding wheel.
The wear is attributed due to attritious wear or progressive wear [1]. Also, during
grinding of soft materials, workpiece chip will clog into the grinding wheel and

J. Revant · Rahul Sree Kumar · K. Rameshkumar (B) · D. S. B. Mouli

Department of Mechanical Engineering, Amrita School of Engineering, Amrita Vishwa
Vidyapeetham, Coimbatore 641112, India
e-mail: k_rameshkumar@cb.amrita.edu

© Springer Nature Singapore Pte Ltd. 2021 511

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_45
512 J. Revant et al.

deteriorates the cutting ability of the abrasive wheel and called as clogging of the
grinding wheel. Once the grinding wheel becomes glazed or clogged, it is necessary
to subject the grinding wheel for dressing operation. The dressing is a grinding wheel
conditioning operation, which will expose the abrasive grits and make the grinding
wheel free from glazing and clogging. Truing operation is performed for new grind-
ing wheels especially for a precision grinding process for ensuring concentricity of
grinding wheel with the work surface. Dressing and truing operations are integrated
as a single process for the conventional grinding wheels.

1.1 Sharpness Evaluation of the Grinding Wheel

The sharpness of the abrasive particles in a grinding wheel plays a prominent role in
determining the capacity of the abrasive wheel to remove the material and achieving
the required finish in the work surface. The sharpness of the wheel face is also directly
related to heat generation and grinding forces developed during grinding. Nakayama
et al. [2] proposed a procedure for evaluating the abrasive wheel sharpness. Alexandre
et al. [3] proposed a test bench for measuring the sharpness of the abrasive wheel to
study the dresser condition. In this study, the methodology proposed in [1, 3] is used
to design and fabricate the sharpness measurement device which is used to establish
the conditions of the grinding wheel.

1.2 Grinding Process Monitoring

Sensor-based machine condition monitoring is increasingly used in industries to pre-

dict the fault condition in advance to prevent the catastrophic failure of the machine
tool. Vibration, motor current, optical, force, temperature, and acoustic emission
sensors are commonly used for monitoring the conditions of the machines or pro-
cesses. Among these sensors, the AE sensor is found suitable for precision machining
processes such as grinding. AE is happening at higher frequencies in the range of
100 kHz to 1 MHz. It is possible to eliminate the external disturbances with much ease
and only the signals relevant to the faulty conditions are utilized using the AE-based
condition monitoring techniques.
AE-based condition monitoring of the grinding process has been investigated by
many researchers. Neuro-fuzzy classification algorithms were proposed by Lezan-
ski [4] using AE, force and vibration sensors. Liao et al. [5, 6] monitored grinding
wheel conditions using hidden Markov models, genetic clustering algorithm, boosted
classifiers and neural networks. Support vector machines (SVM) are used by Sub-
rahmanya and Shin [7] for grinding condition monitoring. Yang and Yu [8] proposed
AE-based grinding wheel wear monitoring system using SVM algorithms with AE
signature. Advancements in sensor technology, hardware development, signal pro-
cessing methods and feature extraction techniques were discussed in detail by Roth
Acoustic Emission-Based Grinding Wheel … 513

et al. [9]. ANN models were developed by Moia et al. [10] for monitoring dress-
ing operation using AE signals. Dias et al. [11] identified a range of frequency for
the occurrence of roundness, cylindricity and surface roughness. In a cylindrical
grinding process, Arun et al. [12] predicted grinding wheel conditions in a cylin-
drical grinding process using classifiers such as SVM, decision trees and artificial
neural network (ANN). Off late, decision tree, Naive Bayes, SVM and artificial neu-
ral network models are used to predict the tool conditions in high-speed precision
machining process [13–15]. In a surface grinding process, dresser conditions of alu-
minium oxide (Al2 O3 ) grinding wheel was studied by Alexandre et al. [3] with AE
and fuzzy methods.
In this paper, the grinding wheel sharpness measurement device has been designed
and fabricated. This device is used to identify the conditions of the grinding wheel. A
grinding wear plot was established for the complete grinding cycle starting from sharp
wheel condition to the wornout condition. Grinding experiments were carried out
using an Al2 O3 grinding wheel and a correlation is established between the features
extracted from AE and grinding wheel sharpness condition. A statistical prediction
model is established using feed-forward back-propagation neural network.

2 Grinding Wheel Sharpness Measurement Test Bench

The principle involved in measuring the sharpness is shown in Fig. 1. Cylindrical

steel disc is subjected to ground with the surface of the abrasive wheel with the
normal force, F N . During the grinding test, the cylindrical disc is pushed by the
normal force (F N ) generated due to the cutting action. The normal force intensity
reduces when the contact area is increased. The displacement of the cylindrical disc
is measured with the help of transducer. The amount of displacement depends on

Fig. 1 Principle of grinding wheel sharpness measurement [2]

514 J. Revant et al.

the sharpness of the grinding wheel face. A relationship between ‘displacement’ and
‘time’ is established to arrive at the sharpness of the grinding wheel. The sharpness
of the grinding wheel is estimated using Eq. (1).
Abrasive grinding wheel sharpness,
√
2b 8r 2/3
K = a (1)
3F N

where ‘b’ is the width of the disc, ‘r’ is the radius of the disc, ‘FN ’ is the normal force
applied to the disc over the grinding wheel, ‘a’ is the gradient of the characteristic
curve of grinding wheel, i.e. regression line obtained from the plot of ‘displacement’
versus ‘t(2/3) ’. A device used for measuring the grinding wheel sharpness has been
designed and fabricated based on the principle of sharpness measurement as shown
in Fig. 1. The photograph of the device is shown in Fig. 2. The specification of the
device is as shown in Table 1. The grinding wheel grade used in this study is ‘AA60-
K5-V8’. The details of the grinding wheel are shown in Table 2. The sharpness
measurement device is fixed over the surface grinding machine table using a fixture.

Fig. 2 Photograph of the grinding wheel sharpness measurement device

Table 1 Specification of the grinding wheel sharpness measurement device

S. No. Parameters Value
1 Overall dimension 100 mm × 60 mm × 80 mm
2 Disc diameter 25 mm
3 Distance between disc centre and pivot point 55 mm
4 Counterweight added to keep the disc in contact 105 g
Acoustic Emission-Based Grinding Wheel … 515

Table 2 Specification of the grinding wheel (AA60-K5-V8)

S. No Parameters Value
1 Grinding wheel size OD—200 mm; ID–—31.75 mm; Thickness—20 mm
2 Grinding wheel material Aluminium oxide (White)
3 Grain size/Grit Medium
4 Grade Medium
5 Type of bond Vitrified

2.1 Grinding Wheel Sharpness Evaluation

Grinding tests were conducted in a surface grinding machine by fixing the sharp-
ness measurement device over the grinding table. The grinding wheel sharpness is
evaluated for all the states of the abrasive wheel using the sharpness measurement
device fabricated in this study. The grinding wheel sharpness is quantified for sharp
and dull-wornout conditions separately. Time–displacement plots are drawn, and the
gradient of the regression lines are computed. Time–displacement plots are shown
in Figs. 3 and 4. Figure 3 shows the regression plot of the sharp condition, and Fig. 4
shows the dull-wornout condition. Sharpness parameters of grinding wheel of three
different conditions are shown in Table 3.
Grinding experiments were repeated to validate the sharpness of the grinding
wheel. From the results, it is observed that the sharpness is lower in the dull or
wornout condition due to clogging or natural wear of the abrasive particles in the
grinding wheel. Lower sharpness value indicates that the grinding wheel has to be
dressed for improving its cutting ability. Sharpness values of the grinding wheel are
indicated in Table 3 for sharp and dull conditions of the grinding wheel.
Grinding wheel wear plot is drawn as a function of metal removal rates of the grind-
ing wheel and work material. Experiments were conducted starting with a dressed
grinding wheel with sharp abrasive edges. Experiments were continued until the

Fig. 3 Time–displacement plot of sharp grinding wheel condition

516 J. Revant et al.

Fig. 4 Time–displacement plot of wornout grinding wheel condition

Table 3 Sharpness parameters

Parameters Sharp condition Dull condition
Disc width ‘b’ 4 mm 4 mm
Disc radius ‘r’ 12.5 mm 12.5 mm
Applied force ‘FN ’ 1.02 N 1.02 N
Regression gradient ‘a’ 0.4596 0.0145
Sharpness ‘K’ 4.505 mm3 /N s 1.063 mm3 /N s

wheel became a dull and glazed condition. During the grinding process at differ-
ent intervals of time, the metal removal rates in workpiece and grinding wheels are
recorded and grinding wheel wear plot is established and shown in Fig. 5. Three

Fig. 5 Grinding wheel wear plot indicating grinding wheel conditions

Acoustic Emission-Based Grinding Wheel … 517

grinding wheel conditions, viz. Sharp-initial, sharp-intermediate and dull-wornout

are also indicated in Fig. 5. Once the grinding wheel becomes dull, it is subjected to
dressing operation.

3 AE Monitoring of Grinding Process

The main objective of AE monitoring of the grinding process is to predict the grinding
wheel condition based on the sharpness parameter using the features extracted from
the AE signature. The experimental set-up used in this study is shown in Fig. 6. The
experimental set-up for AE measurement consists of (a) surface grinding machine,
(b) AE sensor, (c) AE sensor attachment, (d) AE sensor data acquisition system, (e)
AE win software, (d) machine learning classifier and (e) signal processing hardware.
AE sensor used in this study is supplied by M/s Physical Acoustic Corporation,
USA. Micro 30 D sensor used in this study is capable of capturing AE signature in
the frequency range 100–350 kHz. An AE attachment is used to hold the AE sen-
sor to capture the acoustic emissions from the grinding wheel. AE sensor output is
connected to a pre-amplifier and to the data acquisition system. The signal is con-
verted into digital form and stored in the computer memory for further processing.
AE features are extracted using AE-Win software supplied by M/s Mitras Corpora-
tion, USA. After extracting the features, those having wealthy information about the
grinding wheel conditions are selected. The selected features are used for developing
a statistical model using artificial neural network (ANN) classifier for predicting the
grinding wheel condition.

Fig. 6 Experimental set-up for AE measurement

518 J. Revant et al.

3.1 Feature Selection

Grinding experiments were carried out from the sharp-initial condition to the dull-
wornout condition. AE signatures were acquired for the complete grinding cycle. AE
signals were captured at a sampling rate of 1 MSPS with a threshold of 40 dB. AE fea-
tures, viz. Count, RMS, amplitude, rise time, energy and duration are extracted from
the raw signal. The features having useful information about the grinding wheel con-
ditions are selected by plotting the feature value versus grinding time. It is observed
that the AE features RMS and Count have a strong correlation with the grinding wheel
conditions as shown in Fig. 7. RMS and Count feature data are used to train the feed-
forward back-propagation neural network classifier for predicting the grinding wheel
condition using a sharpness parameter.

4 Neural Network Modelling

Neural networks or multi-layer perceptrons have the ability to extract useful infor-
mation from the AE sensor feature data and have the capability to extract patterns and
detect the conditions of the grinding wheel. RMS and Count data extracted from the
AE signature are provided as an input to neural network with the known condition
of the surface grinding wheel. Training of the neural network is carried out using
back-propagation algorithms. After training, the neural network model is tested with
AE signal feature data for which the condition of the grinding wheel is not known.
Training and testing procedure of neural network classifiers is shown in Fig. 8.
Multi-layer perceptron used in this study consists of three layers namely (a) input
layer, (b) hidden layer and (c) output layer. A supervised learning technique, back-
propagation algorithm, is used for training the network. Input layer consisting of
neurons is based on the input data. In this case, 1363 AE data (RMS and Count) are
used for training and testing the network. Among 1363 AE data, 320 are belonging
to the sharp-initial state of the abrasive wheel. 636 samples are belonging to the
sharp-intermediate state of the abrasive wheel. The balance 407 samples belong

Fig. 7 a Grinding time versus RMS. b Grinding time versus Count

Acoustic Emission-Based Grinding Wheel … 519

Fig. 8 Training and testing procedure in neural net classifier

to the dull-wornout state of the abrasive wheel. The samples were extracted by
conducting the experiment in appropriate grinding wheel conditions established in
this study. Experiments were conducted to optimize the number of hidden layers. It
is found that the no. of hidden layers as ‘3’ that yields good results. The sigmoidal
transfer function is used. Learning rate and momentum are taken to be 0.3 and 0.2,
respectively. A ‘10’ fold cross-validation is used for training and testing of the neural
network classifiers. In a ‘10’-fold cross-validation process, the data sets are divided
into ‘10’ groups and perform the training on nine groups and validate the results using
the remaining group. To reduce the variability, 10 times of validation experiments
are performed using different groups of training data leaving out the validation (test)
data. The results are averaged over 10 times of validation. Neural network classifier
is implemented using the Weka platform [16]. Performance of the classifier is studied
using classification accuracy, misclassification rate, Kappa and ROC statistics.

4.1 Performance Measures

The performance of the classifier is studied using a confusion matrix. Performance

measures, viz. (a) classification accuracy, (b) misclassification rate, (c) Kappa statis-
tics and (d) receiver operating characteristic (ROC) are extracted from the confusion
matrix. These matrices are used for evaluating the ANN classifier considered in this
study. Classification accuracy of a classifier is computed using Eq. (2).

Total No. of correctly classified instances

Classification accuracy, % = × 100 (2)
Total No. of instances
Landis and Koch [17] proposed statistics called ‘Kappa’ for evaluating the per-
formance of the classifiers. A ‘Kappa’ value of 0.81–1 is referred to as a ‘perfect
classifier’. A ‘Kappa’ value of 0.61–0.80 is a ‘substantial classifier’, 0.41–0.60 is
a ‘moderate’ one, 0.21–0.40 is termed as ‘fair’ and 0–0.20 is called as a ‘slight’
classifier. The ‘Kappa’ statistic is computed using the expression shown in Eq. (3).
Receiver operating characteristic (ROC) of the classifier is found using true positive
and false positive rates obtained from the confusion matrix.
520 J. Revant et al.

po − pe (Observed accuracy − Expected accuracy)

Kappa = = (3)
1 − pe (1 − Expected accuracy)

where ‘po ’ is the observed accuracy and ‘pe ’ is the expected accuracy.

5 Results and Discussions

AE features RMS and Count were given as an input to ANN classifiers for training.
Training and testing data are acquired from the grinding wheel conditions arrived
using the grinding wheel sharpness data. The confusion matrix obtained after train-
ing and testing the AE features and the ANN classifier performance is shown in
Table 4. The ANN classifier is able to predict the grinding wheel conditions with
an overall accuracy of 90.201%. Among 1363 instances, 1230 instances were clas-
sified correctly according to its conditions. Misclassifications of AE features are
more in the features which were acquired in the boundary regions of ‘sharp-initial’
and ‘sharp-intermediate’. ANN is unable to differentiate the very minute variations
in the sharpness of the abrasive particles in the boundary regions. There are 21
instances of ‘sharp-initial’ misclassified as ‘sharp-intermediate’ and 50 instances of
‘dull-wornout’ are misclassified as ‘sharp-intermediate’ condition.
Kappa statistics value of 0.8468 justifies the good performance of the ANN classi-
fiers. Receiver operating characteristic (ROC) curve statistics provided by the ANN
classifiers also demonstrate the diagnostic ability of the classifier and show that the
ANN is able to distinguish the grinding wheel conditions with good accuracy. ROC
statistics were arrived by using the true positive rate and false positive rates of the
classifier. It is observed that ANN is able to classify the grinding wheel condition
with good accuracy. The time taken for building the ANN model for grinding wheel

Table 4 Confusion matrix: Performance of ANN classifier

Grinding wheel conditions
Predicted condition
Sharp-Initial Sharp-Intermediate Dull-Wornout
Actual condition Sharp-Initial 298 21 1
Sharp-Intermediate 35 578 23
Dull-Wornout 3 50 354
Correctly classified instances 1230
Overall accuracy, Avg., % 90.2421
Incorrectly classified instances 133
Misclassification rate, Avg., % 9.7579
Kappa statistics 0.8468
ROC area 0.973
Acoustic Emission-Based Grinding Wheel … 521

condition prediction is 0.53 s using ‘Intel i5 processor’ with 4 GB internal memory.

ANN model proposed in this study can be extended for grinding wheel condition
prediction in real time.

6 Conclusions

A grinding wheel sharpness measurement device has been designed and fabricated
in this study. This device will be useful in automated manufacturing environments
for detecting the grinding wheel conditions. AE sensor is used for capturing the
signals of the grinding process and grinding wheel conditions are established using
the grinding wheel wear plot. Time domain AE features such as RMS and Count were
found to have a good association with the abrasive wheel conditions. A statistical
model using ANN is developed for estimating the grinding wheel condition. AE
sensor with ANN-based machine learning classifiers is able to predict the grinding
wheel conditions with an accuracy of 90.2%. Kappa and ROC statistics are also
justifying the effectiveness of ANN classifier in predicting the conditions of the
grinding wheel. Further studies can be taken up to develop an AE-based real-time
grinding wheel condition monitoring system using the data and knowledge generated
in this study.

Acknowledgements This research is supported by Directorate of Extramural Research and Intel-

lectual Property Rights (ER & IPR), Defense Research and Development Organization (DRDO),
ERIP/ER/0803740/M/01/1194, 13 January 2010.

References

1. Jackson MJ, Davim JP (2011) Machining with abrasives. Springer, New York, pp 4–5
2. Nakayama K, Takagi JI, Irie E, Okuno K (1980) Sharpness evaluation of grinding wheel face
by the grinding of steel ball. CIRP Ann 29(1):227–231
3. Alexandre FA, Lopes WN, Dotto FRL, Ferreira FI, Aguiar PR, Bianchi EC, Lopes JC (2018)
Tool condition monitoring of Aluminum oxide grinding wheel using AE and fuzzy model. Int
J Adv Manuf Technol 96(1–4):67–79
4. Lezanski P (2001) An intelligent system for grinding wheel condition monitoring. J Mater
Process Technol 109(3):258–263
5. Liao TW, Ting CF, Qu J, Blau PJ (2007) A wavelet-based methodology for grinding wheel
condition monitoring. Int J Mach Tools Manuf 47(3):580–592
6. Liao TW (2010) Feature extraction and selection from acoustic emission signals with an
application in grinding wheel condition monitoring. Eng Appl Artif Intell 23(1):74–84
7. Subrahmanya N, Shin YC (2008) Automated sensor selection and fusion for monitoring and
diagnostics of plunge grinding. J Manuf Sci Eng 130(3):031014
8. Yang Z, Yu Z (2012) Grinding wheel wear monitoring based on wavelet analysis and support
vector machine. Int J Adv Manuf Technol 62(1–4):107–121
9. Roth JT, Djurdjanovic D, Yang X, Mears L, Kurfess T (2010) Quality and inspection of
machining operations: tool condition monitoring. J Manuf Sci Eng 132(4):041015
522 J. Revant et al.

10. Moia DFG, Thomazella IH, Aguiar PR, Bianchi EC, Martins CHR, Marchi M (2015) Tool
condition monitoring of aluminum oxide grinding wheel in dressing operation using acoustic
emission and neural networks. J Brazilian Soc Mech Sci Eng 37(2):627–640
11. Dias EA, Pereira FB, Ribeiro Filho SLM, Brandão LC (2016) Monitoring of through-feed
centerless grinding processes with acoustic emission signals. Measurement 94:71–79
12. Arun A, Rameshkumar K, Unnikrishnan D, Sumesh A (2018) Tool condition monitoring of
cylindrical grinding process using acoustic emission sensor. Mater Today: Proc 5(5):11888–
11899
13. Krishnakumar P, Rameshkumar K, Ramachandran KI (2018) Acoustic emission-based tool
condition classification in a precision high-speed machining of titanium alloy: a machine
learning approach. Int J Comput Intell Appl 17(03):1850017
14. Krishnakumar P, Rameshkumar K, Ramachandran KI (2018) Feature level fusion of vibration
and acoustic emission signals in tool condition monitoring using machine learning classifiers.
Int J Progn Health Manag 9:1–15
15. Krishnakumar P, Rameshkumar K, Ramachandran KI (2018) Machine learning based tool
condition monitoring using acoustic and vibration data in high speed milling. Int J Intell Decis
Technol 1:1–18
16. https://www.cs.waikato.ac.nz/~ml/weka/
17. Landis JR, Koch GG (1977) The measurement of observer agreement for categorical data.
Biometrics 159–174
Turning Process Characteristics
of Aluminium Matrix Hybrid Composite
Using Grey Relational Surface
Methodology

A. Gnanavelbabu , V. Arunachalam , K. T. Sunu Surendran ,

R. Saranraj , and K. Rajkumar

Abstract Aluminium metal matrix composites have generated huge attention in

modern industries because of its high strength-to-weight ratio, low density and cor-
rosion resistance. Turning is a widely used material removal process to produce
cylindrical products. Presence of solid lubricant reinforcements such as graphite
(Gr), carbon nanotube (CNT) and hexagonal boron nitride (hBN) could increase
the feasibility of turning on aluminium metal matrix composites. Aluminium alloy
6061 reinforced with B4 C (5–15 vol.%) and hBN (10 vol.%) particulates were fab-
ricated using stir casting process. The effect of process parameters such as feed,
speed, depth of cut and volume percentage of boron carbide has been investigated
to explore the impact on tangential force, cutting power and tool wear. Experiments
were conducted using Taguchi’s L27 orthogonal array (OA). The factors affecting
the responses were optimized by applying grey-response surface methodology (G-
RSM) using grey relational grade (GRG) calculated by normalized S/N ratio and
grey relational coefficient (GRC).

Keywords Aluminium hybrid composites · Turning · Grey relational grade ·

Response surface methodology

1 Introduction

Metal matrix composite (MMC) has been considered as an important class of mate-
rial, as it progressively replaces conventional materials utilized in structural engineer-
ing, aeronautics, mechanical engineering and in many other sectors [1]. MMCs are
considered for engineering applications where high specific strength, stiffness and
lightweight are important. Common matrix metals include aluminium, magnesium,

A. Gnanavelbabu (B) · V. Arunachalam · K. T. Sunu Surendran · R. Saranraj

Department of Industrial Engineering, CEG, Anna University, Chennai 600025, India
e-mail: agbabu@annauniv.edu
K. Rajkumar
Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai 603110, India

© Springer Nature Singapore Pte Ltd. 2021 523

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_46
524 A. Gnanavelbabu et al.

copper and titanium in its pure or alloy forms. Aluminium metal matrix composites
(AMMCs) are suitable material for aerospace, defence and automotive applications.
Generally, aluminium (6 series) alloys have been considered broadly because of its
merits such as strength, castability, formability, corrosion resistance and low cost
comparing to other aluminium alloys.
Ceramic reinforcements such as silicon carbide (SiC), boron carbide (B4 C) and
titanium carbide (TiC) are common reinforcements in AMMC [2, 3]. Incorporation of
such reinforcements improved the specific properties of the metal matrix phase [4, 5].
In addition, graphite, graphene, carbon nanotube (CNT) and hexagonal boron nitride
(hBN) are commonly used solid lubricants in developing AMMCs. A. Gnanavelbabu
et al. investigated the cutting quality characteristics of AA6061-B4 C-hBN hybrid
composites by abrasive water jet machining and reported favourable results [6]. C.
Muthazhagan et al. found a positive influence of solid lubricant on the mechanical
properties of aluminium-B4 C composites. Also, they found that the addition of B4 C
conversely increased the hardness of the composites [7].
Design of experiment (DOE) plays a vital role in reducing the number, time
and cost of the experiments. Optimizations of process parameters are important for
improving the quality characteristics of the responses. W.H. Yang et al. optimized
the cutting parameters for turning operations for steel using Taguchi’s orthogonal
array, S/N ratio ANOVA [8]. A. J. Makadia et al. investigated the influence of tool
geometry (nose radius) and cutting parameters (cutting speed, feed and depth of cut)
on the surface finish obtained in turning of mild steel by applying Taguchi-response
surface methodology (T-RSM) [9]. R. Adalarasan et al. investigated the quality char-
acteristics of laser cutting parameters using grey-response surface methodology (G-
RSM) and found the optimum level of cutting parameters by applying grey relational
analysis and G-RSM [10]. A. Dey et al. predicted the optimal combination of pro-
cessing parameters by hybrid approach of grey-based response surface methodology
(G-RSM) and found the significant model using grey-based approach [11].
It is observed from the literature that optimization of multiple responses based on
orthogonal array with GRA tends to take advantage of both the Taguchi method and
the GRA. Hence, an approach named GRSM which combines with GRA is presented
to find out the optimal parameter setting. This paper presents a detailed experimental
analysis on the turning process of AA6061-B4 C-hBN hybrid composites by applying
G-RSM for optimizing the machining quality characteristics such as tangential force,
cutting power and tool wear.

2 Materials and Methods

Aluminium alloy 6061 is the most commonly used alloy which has a better combina-
tion of mechanical properties and it was chosen as the matrix material.B4 C and hBN
were the reinforcement particulates of size 10–25 µm and 5 µm, respectively. The
composition of the fabricated hybrid composites is shown in Table 1. Aluminium
hybrid composites were fabricated using two-stage stir casting process.
Turning Process Characteristics of Aluminium … 525

Table 1 Composition of fabricated composites

Composite hBN (vol.%) B4 C (vol.%) AA6061 (vol.%)
Comp A 10 5 85
Comp B 10 10 80
Comp C 10 15 75

Table 2 Factors and levels of the input parameters

Factors Speed (RPM) Feed (mm/rev) Depth of cut (mm) Reinforcement (vol.%)
[A] [B] [C] [D]
Low 60 0.05 0.2 5
Medium 120 0.10 0.4 10
High 180 0.15 0.6 15

The fabricated composites comp A, comp B and comp C were cut into required
dimensions for performing turning operation. Taguchi’s L27 orthogonal array was
used to collect the experimental data and to analyse the characteristics of turning
parameters of AMMC. Turning parameters namely speed, feed, depth of cut and solid
lubricant are chosen for controlling the responses and each parameter is designed
with three levels, denoted by, low, medium and high as shown in Table 2.
CNC turning centre with 360–480 V and 40/60 Hz was used to perform the
experiments. Polycrystalline diamond (PCD) was used as the cutting tool. PCD tools
are preferred instead of polycrystalline cubic boron nitride (PCBN) tools as they
exhibit high fracture resistance and low adhesion property. Also, it improves the
machining productivity and tool life characteristics. Each experiment was carried
out using a new cutting edge every time to obtain accurate response values which
are tangential force, cutting power and tool wear. The values of tangential force (F z )
and tool wear (T w ) are obtained directly from CNC turning machine, and the cutting
power (Pc ) is calculated using Eq. 1.

D ∗ F ∗ S ∗ Kc
Pc = (1)
60 ∗ 103 ∗ η

where Pc –cutting power, F—feed, S—speed, D—depth of cut, K c —Specific cutting

force and ï—machine coefficient. The results of CNC turning process are shown in
Table 3.
The grey relational analysis was performed for the multi-performance optimiza-
tion, in which the optimization processes are gone through, until the grey relational
grade (GRG) obtained. Higher the value of GRG, closer the parameter to its optimum
value. The resultant GRG value reflects the single optimization value of the multiple
experimental performances. Within the range of sequences to acquire exact solu-
tions, one has to pre-process the data which are related to group of sequences called
as grey relational generation by Deng, J.L. (1989). Converting an original sequence
526 A. Gnanavelbabu et al.

Table 3 Experimental data

Trail A (RPM) B (mm/rev) C (mm) D (vol.%) F z (N) Pc (kW) T w (mm)
1 60 0.05 0.2 5 5 1.571 0.071
2 60 0.05 0.2 10 10 3.143 0.045
3 60 0.05 0.2 15 8 2.514 0.1
4 60 0.1 0.4 5 9 2.829 0.152
5 60 0.1 0.4 10 11 3.457 0.123
6 60 0.1 0.4 15 12 3.771 0.121
7 60 0.15 0.6 5 11 3.457 0.179
8 60 0.15 0.6 10 15 4.714 0.145
9 60 0.15 0.6 15 13 4.086 0.162
10 120 0.05 0.4 5 13 8.171 0.209
11 120 0.05 0.4 10 13 8.171 0.175
12 120 0.05 0.4 15 7 4.4 0.185
13 120 0.1 0.6 5 12 7.543 0.222
14 120 0.1 0.6 10 14 8.8 0.2
15 120 0.1 0.6 15 12 7.543 0.201
16 120 0.15 0.2 5 16 10.057 0.256
17 120 0.15 0.2 10 15 9.429 0.223
18 120 0.15 0.2 15 13 8.171 0.225
19 180 0.05 0.6 5 11 10.371 0.289
20 180 0.05 0.6 10 18 16.971 0.245
21 180 0.05 0.6 15 11 10.371 0.262
22 180 0.1 0.2 5 14 13.2 0.302
23 180 0.1 0.2 10 19 17.914 0.269
24 180 0.1 0.2 15 12 11.314 0.292
25 180 0.15 0.4 5 15 14.143 0.322
26 180 0.15 0.4 10 20 18.857 0.28
27 180 0.15 0.4 15 14 13.2 0.315

into a decimal sequence between 0.00 and 1.00 is known as pre-processing raw data.
The pre-processing starts with calculating signal-to-noise ratio (S/N). S/N ratio was
calculated to reduce the variability in the process. The objective of this study is to
minimize the tangential force, cutting power and tool wear for better characteristic
performance. Hence, the S/N ratio was calculated by considering “smaller the bet-
ter” condition for all three responses. The obtained sequence normalized with the
“smaller is better” condition.
Further, GRC was calculated to express the relationship between the best and
actualized S/N ratio. Finally, the average of grey relational coefficients was taken to
get grey relational grade (GRG). Further, response surface methodology (RSM) was
Turning Process Characteristics of Aluminium … 527

applied to optimize and generate ANOVA and regression equations. Surface plots and
contour plots were generated using G-RSM. Statistical analysis of the experimental
data was performed using design expert software.

3 Results and Discussions

The current optimization is focused on the improvement of minimum tangential

force, cutting power and tool wear in turning process. From GRC values of F z , Pc and
T w , grey relational grade (GRG) was calculated. The experiments of grey relational
analysis were ranked based on the higher order of their GRG values. The GRG values
for all experiments are shown in Table 4. Experiment no. 26 has the highest GRG
(0.9611). Therefore, experiment no. 26 is the optimal process parameters setting
for minimum tangential force, cutting power and tool wear simultaneously (i.e. the
best multi-performance characteristics) among the twenty-seven experiments. Since
there is a possibility for a certain degree of uncertainties on GRG, RSM model is
employed on GRG for further optimizing the responses.
The GRG representing the multiple responses was taken as a single performance
and used in the generation of response surfaces. The RSM method involves develop-
ing equation of quadratic model using the decision variables. The generated regres-
sion equation creates the way to clear about the behaviour of system by quantifying
the association and thereby exploring the individual and interaction effects of turn-
ing parameters on the responses. The GRG values are entered in the design expert
software and the mathematical model are obtained as shown in Eq. 2.

GRG = +0.66 + 0.15 ∗ A + 0.048 ∗ B + 0.011 ∗ C−0.014 ∗ D

+ 0.015 ∗ A ∗ B−0.032 ∗ A ∗ C−0.20 ∗ A ∗ D
− 0.27 ∗ B ∗ C − 1.025E − 003 ∗ B ∗ D + 6.007E − 003 ∗ C ∗ D
+ 0.023 ∗ A2 + 0.013 ∗ B2 −0.069 ∗ D (2)

The coefficient of the regression equation is determined and contour and surface
plots are plotted for the p values lower than the significance level (α) equals to
0.05. The results of response surface regression and analysis of variance for the grey
relational grades are shown in Table 5.
Based on the results from ANOVA, the response surfaced models generated are
significant as shown in Table 6.
Referring to the ANOVA table, the regression model p value showed 0.0001, which
indicated that the model was more significant. Thus, the model is more suitable to
predict the values. The significance of the regression fitted model is determined by
incorporating the coefficient correlation. The obtained R-sq. and R-sq. (adj) for the
model are 94.24 and 88.48, respectively. Thus, the statistical measure of value shows
that the model is fitted closely. Response surface optimizer in design is used to find
the parameters which provide the best results. The goal is to maximize the GRG
528 A. Gnanavelbabu et al.

Table 4 GRG calculation

Trail S/N ratio (normalized) GRC GRG
Fz Pc Tw Fz Pc Tw
1 0 0 0.231 0.333 0.333 0.394 0.3523
2 0.5 0.279 0 0.5 0.409 0.333 0.4158
3 0.339 0.189 0.405 0.430 0.381 0.456 0.4214
4 0.423 0.236 0.618 0.464 0.395 0.567 0.4723
5 0.568 0.317 0.510 0.536 0.422 0.505 0.4866
6 0.631 0.352 0.502 0.755 0.435 0.501 0.5028
7 0.568 0.317 0.701 0.536 0.422 0.626 0.5243
8 0.792 0.442 0.594 0.706 0.472 0.552 0.5754
9 0.689 0.384 0.653 0.616 0.484 0.585 0.5483
10 0.689 0.663 0.780 0.616 0.597 0.694 0.6343
11 0.689 0.663 0.690 0.616 0.597 0.617 0.6101
12 0.242 0.414 0.718 0.397 0.460 0.639 0.4956
13 0.631 0.631 0.811 0.575 0.575 0.725 0.6225
14 0.742 0.693 0.757 0.660 0.619 0.673 0.6501
15 0.631 0.631 0.760 0.575 0.575 0.676 0.6070
16 0.839 0.747 0.883 0.756 0.664 0.810 0.7407
17 0.792 0.721 0.813 0.706 0.641 0.728 0.6904
18 0.689 0.663 0.817 0.616 0.597 0.732 0.6463
19 0.568 0.759 0.955 0.536 0.675 0.909 0.6996
20 0.923 0.957 0.861 0.868 0.921 0.782 0.8604
21 0.568 0.759 0.895 0.536 0.675 0.826 0.6764
22 0.742 0.856 0.967 0.660 0.776 0.938 0.7886
23 0.963 0.979 0.908 0.931 0.960 0.845 0.9147
24 0.631 0.794 0.950 0.575 0.708 0.909 0.7271
25 0.792 0.884 1 0.706 0.812 1 0.8356
26 1 1 0.928 1 1 0.875 0.9611
27 0.742 0.856 0.988 0.660 0.776 0.978 0.8009

values so that the response values can be optimized. The optimal parameter settings
are shown in Table 7 at which our objective function can be satisfied.
Figure 1 illustrates the plot of actual versus predicted values of responses and it
is observed that all the points of the experimental runs are spotted randomly inside
the reliable range of residuals.
Surface plots and contour plots for GRGs are given in the Figs. 2 and 3, respec-
tively. The response surface plots are drawn based on the generated regression equa-
tion associating the dependent variable GRGs with the independent variables such
as tangential force, cutting power and tool wear. Increased level of speed and feed
Turning Process Characteristics of Aluminium … 529

Table 5 Response surface regression

Term Coefficient SE Coefficient T value P value
Constant 0.6883 0.0340 20.26 0.000
Speed 0.1647 0.0125 13.13 0.000
Feed 0.0619 0.0177 3.49 0.004
DOC −0.002 0.0177 −0.13 0.898
% OF hBN −0.013 0.0125 −1.08 0.298
Speed*Speed 0.0094 0.0217 0.43 0.672
Feed*Feed −0.006 0.0251 −0.03 0.980
DOC*DOC −0.013 0.0251 −0.54 0.595
% of hBN* % of hBN 0.068 0.0217 −3.15 0.008
Speed*Feed −0.012 0.0251 −0.48 0.638
Speed*DOC −0.004 0.0251 −0.19 0.852
Speed*% of hBN −0.020 0.0154 −1.32 0.210
Feed*% of hBN −0.001 0.0154 −0.07 0.948
DOC*% of hBN 0.0060 0.0154 0.39 0.702

improved the GRG values and responses as shown in Fig. 2. It is observed that speed
and feed are directly proportional to the responses, while the volume percentage of
hBN is inversely proportional to the responses. From the surface plot, the maximum
value of GRG is obtained for Comp B with high cutting speed, high feed and medium
depth of cut.
The contour plots (Fig. 3) shows the graph of GRG and process parameters,
where the dark green colour indicates higher the GRG value and light green colour
indicates lower the GRG value. They hold values in the contour plots which are
120 rpm, 0.1 mm/rev, 0.4 mm and 10% for speed, feed and depth of cut and volume
percentage of hBN, respectively. Lower the cutting speed lead to striations and cause
rough surface finish while higher the cutting speed caused thermal damage to the
material. Hence, a moderate level (120 rpm) of cutting speed is desired for turning
operation.

4 Conclusion

The G-RSM optimization tool was applied for the optimization of CNC turning
process parameters on AA6061-B4C-hBN hybrid composites and the following
conclusions were drawn.
1. The derived second-order polynomial equation has shown good correlation
between the predicted and experimental values.
530 A. Gnanavelbabu et al.

Table 6 ANOVA for GRG

Source Sum of DOF Mean F value P value Remarks
squares square
Model 0.60 13 0.046 16.35 <0.0001 Significant
A Speed 0.082 1 0.082 29.00 0.0001
B Feed 8.732E − 1 8.732E − 2.96 0.1093
003 003
C DOC 5.787E − 1 5.787E − 0.20 0.6587
004 004
D % hBN 3.324E − 1 3.324E − 1.17 0.2984
003 003
AB 3.491E − 1 3.491E − 0.12 0.7312
004 004
AC 1.160E − 1 1.160E − 0.41 0.5333
003 003
AD 4.918E − 1 4.918E − 1.74 0.2104
003 003
BC 8.403E − 1 8.403E − 0.30 0.5952
004 004
BD 1.260E − 1 1.260E − 4.450E − 0.9478
005 005 003
CD 4.332E − 1 4.332E − 0.15 0.7021
004 004
A2 1.371E − 1 1.371E − 0.48 0.4989
003 003
B2 4.368E − 1 4.368E − 0.15 0.7009
004 004
C2 0.00 0
D2 0.028 1 0.028 0.0076
Residual 0.037 13 2.833E − 9.94
003
Cor total 0.64 26
Std.dev.—0.053 R2 —0.9424 Adj. R2 —0.8848 Adq. precision—14.078

Table 7 GRSM results at

Input Optimum level
10% hBN
Speed 180
Feed 0.15
Depth of cut 0.3374
% hBN 8.6364
Turning Process Characteristics of Aluminium … 531

Fig. 1 Predicted versus actual response plot

Fig. 2 Surface plot for GRG

2. Experiment no. 26 has the highest GRG value (0.9611) and it is the optimal
process parameters setting to obtain the best multiperformance characteristics.
3. Effect of speed is much more prominent than the effects of feed rate and depth
of cut, on the tangential force, cutting force and tool wear.
532 A. Gnanavelbabu et al.

Fig. 3 Contour plot for GR

4. The optimal process parameters from the grey relational analysis for turning pro-
cess were speed—180, feed—0.15, depth of cut—0.3374 and %B4 C—8.6364.

Acknowledgements The authors gratefully acknowledge the financial support from “DST-PURSE
Phase II” for research consumables.

References

1. Nicholas JC, Boswell B, Davies J, Islam MN (2017) Review of machining metal matrix
composites. Int J Adv Manuf Technol 90:2429–2440
2. Rai RN (2006) A study on the machinability behavior of Al-TiC composite prepared by in situ
technique. Mater Sci Eng 428:34–40
3. Torralba JM, Da Costa CE, Velasco F (2003) P/M aluminum matrix composites: an overview.
J Mater Process Technol 133:203–206
4. Chawla N, Chawla KK (2013) Metal matrix composites, 2nd edn. Springer, New York
5. Lin JT, Bhattacharyya D, Ferguson B (1998) Chip formation in the machining of SiC-particle-
reinforced aluminium-matrix composites. Compos Sci Technol 58(2):285–291
6. Gnanavelbabu A, Rajkumar K, Saravanan P (2018) Investigation on the cutting quality char-
acteristics of abrasive water jet machining of AA6061-B4C-Hbn hybrid MMC. Mater Manuf
Process (45):1313–1323
7. Muthazhagan C, Gnanavelbabu A, Bhaskar GB, Rajkumar K (2014) Influence of graphite
reinforcement on mechanical properties of Al-boron carbide composites. Adv Mater Res
495:398–402
8. Makadia AJ, Nanavati JI (2013) Optimization of tool geometry and cutting parameters for
turning operations based on response surface methodology. In: Proceedings of AMPC, Chennai,
pp 6–8
9. Yang WH, Tarng YS (1998) Design optimization of cutting parameters for turning operations
based on the Taguchi method. J Mater Process Technol 84:122–129
Turning Process Characteristics of Aluminium … 533

10. Adalarasan R, Santhanakumar R, Rajmohan M (2015) Optimization of laser cutting parameters

for Al6061/SiCp/Al2 O3 composite using grey based response surface methodology (GRSM).
Measurement 73:596–606
11. Dey A, Debnath S, Pandey KM (2017) Optimization of electrical discharge machining process
parameters forAl6061/cenosphere composite using grey based hybrid approach. Transform
Nonferrous Metals Soc China 27:998–1010
Corrosion Behaviour of Hot Rolled
AA8015 in Natural Seawater at 1.37 µm
Surface Roughness

O. Olaogun and Esther Titilayo Akinlabi

Abstract Aluminium alloy 8015 (AA8015) is a general-purpose alloy having a wide

range of applications that include household appliances and marine ship structures.
Electrochemical testing of hot rolled AA8015 with a particular roughness immersed
in natural seawater solution was investigated. The AA8015 under investigation in this
research is manufactured by a continuous casting method and subjected to mechanical
hot rolling process at a temperature above its recrystallization temperature into 7 mm
thickness plate. A roughness average of 1.37 µm with three cold-mounted samples
was accomplished on an automatic dual multi-purpose grinding-polishing machine
with the aid of 320 grit silicon carbide waterproof paper sheet. Corrosion experiments
via electrochemical reactions were conducted on the corrosion samples immersed in
natural seawater with the aid of a computerized controlled Ivium potentiostat in an
open electrochemical polarization cell arrangement at ambient temperature condi-
tion. Observation of the surface morphologies of the corroded samples was achieved
by a high megapixel camera and scanning electron microscope (SEM). Findings
show asymmetric polarization curves for all the investigated samples, showing an
actively corroding region. Consequently, the SEM examination confirms corrosion
occurring at localized areas in the form of pitting.

Keywords AA8015 · Electrochemical corrosion · Mechanical hot rolling ·

Surface roughness

O. Olaogun (B) · E. T. Akinlabi

Department of Mechanical Engineering Science, University of Johannesburg, Johannesburg,
South Africa
e-mail: yinka.olaoluwa@live.com
E. T. Akinlabi
Department of Mechanical Engineering, Covenant University, Ota, Nigeria
O. Olaogun
Department of Mechanical Engineering, Kwara State University, Malete, Nigeria

© Springer Nature Singapore Pte Ltd. 2021 535

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_47
536 O. Olaogun and E. T. Akinlabi

1 Introduction

Aluminium alloy 8015, among other aluminium alloys, is equally known for its
excellent corrosion resistant property owing to the very thin oxide film layer formed
on the metallic surface. Hot rolled aluminium alloy 8015 is distinctly based on the
fact that it has undergone mechanical hot rolling process achieved with heat. The hot
rolled aluminium alloy 8015 product has the best applications where dimensional
tolerances are not as important as the overall material strength and likewise, surface
finish is not a major concern [1].
Corrosion of aluminium alloys has significant records by various scholars [2–12].
The presence of intermetallic particles in aluminium alloys, usually on the surface,
causes a weak thin oxide film layer [13]. Research shows that these intermetallic
particles act as a cathodic role to the aluminium matrix in corrosive environments
containing chloride ions [14]. These intermetallic alloying element particles have
a varying degree of solid solubility in the aluminium matrix. Polmear et al. [15]
reveal the solid solubility of individual alloying elements in aluminium. Findings
show that zinc, silver, magnesium and copper have high solid solubility compared
to iron, nickel, manganese and silicon, thereby suggesting that intermetallic alloying
elements of the latter are the main causes of localized corrosion in aluminium alloys.
Electrochemical corrosion studies of aluminium alloys in natural seawater have
been reported in the literature [16–18], and all reveal the mode of corrosion by pitting.
Information on electrochemical corrosion of hot rolled aluminium alloy 8015 is yet
to be reported in the literature.
This present work was motivated to investigate the corrosion behaviour of hot
rolled aluminium alloy 8015 in natural seawater at 1.37 µm surface roughness aver-
age. Electrochemical corrosion test using Ivium potentiostat was used to analyse
the open-circuit potential (OCP), polarization measurement and corrosion rate of
the aluminium alloy 8015. The aluminium alloy 8015 specimen samples, before and
after electrochemical corrosion, were analysed with TESCAN VEGA scanning elec-
tron microscope having energy dispersive spectrometry to observe the microstructure
and morphologies of the corroded surface. The details of the corrosion behaviour are
discussed in the present study.

2 Methods

2.1 Material and Corrosion Sample Preparation

The electrochemical experiments were performed with three specimen samples of

aluminium alloy 8015 prepared by a continuous casting method and mechanical
hot rolling process. This alloy under study was obtained industrially from Tower
Corrosion Behaviour of Hot Rolled AA8015 … 537

Aluminium Rolling Mill Nigeria Plc, Sango-Ota, Nigeria. The elemental composition
as determined by Bruker Elemental Optical Emission Spectrometry is shown in
Table 1.
Prior to the hot rolling process, the aluminium alloy 8015 slab formed by solidifi-
cation of metallic solution of pure aluminium and the alloying elements were fed into
the furnace and heated to above recrystallization temperature up to 55 °C. The molten
alloy was continuously casted and solidified using roll casters into a continuous strip
of approximately 20 mm thickness. Further mechanical deformation process in a hot
rolling mill in tandem reduced the thickness to 7 mm plate.
The hot rolled aluminium alloy 8015 specimen samples having thickness of 7 mm
were prepared for electrochemical corrosion testing according to the ASTM stan-
dard G1-03 [19]. The specimen samples were square-cut into dimensions 10 mm
by 10 mm. Insulated copper wire with SANS 60227-5 specification, cut into short
lengths, was adhered to the cut specimen samples using epoxy glue for connec-
tivity. Subsequent cold mounting was achieved and further grinding with 320-grit
SiC paper was done. Surface roughness measurement using HOMMEL-ETAMIC
TURBO WAVE V7.53 roughness and contour metrology before and after sample
preparation reads 2.50 µm and 1.37 µm, respectively.

2.2 Electrochemical Corrosion Test Procedure

Potentiodynamic electrochemical experiments were conducted in a conventional

three-electrode open glass-cell configuration containing 200 ml natural seawater
solution using a CompactStat Potentiostat, an Ivium Technologies product, with cor-
rosion analysis software. This instrument controls the driving mechanism required
for electrochemical reactions that takes place on the working electrode. The hot
rolled aluminium 8015-alloy with 1.37 µm surface roughness serves as working
electrode, a cylindrical coiled platinum wire as counter-electrode, and Ag/AgCl as
a reference electrode. The three-electrode cell was utilized to allow the potential
taking place between working and reference electrodes to be controlled. The electro-
chemical experiments were performed at room temperature of 20 ± 2 °C within the
succeeding order: Open-circuit potential (OCP) versus time for an hour duration and
polarization measurements over a potential range starting from 250 mV lower than
the OCP value and ending at 250 mV directly above the OCP value at a scan rate of
0.67 mV/s. These experiments procedures were followed according to ASTM stan-
dard G69 and G5 for corrosion potential and polarization measurements, respectively
[20, 21]. The hot rolled metal alloy surface before and after the electrochemical test
was subjected to macro- and micro-observation by a high pixel camera and scanning
electron microscope (SEM), respectively.
 538

Table 1 Measured element composition of aluminium alloy 8015

Si Fe Cu Mn Mg Cr Zn Ti Al
Weight % 0.4780 1.3380 0.1460 0.0740 0.0350 0.00670 0.0720 0.0110 Bal.
O. Olaogun and E. T. Akinlabi
Corrosion Behaviour of Hot Rolled AA8015 … 539

3 Results and Discussion

3.1 Open-Circuit Potential and Potentiodynamic Polarization

The progression of the open-circuit potential (OCP) as a function of time for the
hot rolled aluminium alloy 8015 samples (at 1.37 µm surface roughness) exposed to
natural seawater electrolyte solution is presented in Fig. 1.

Fig. 1 Display of
open-circuit potential (OCP)
versus time for hot rolled
aluminium alloy 8015 (Ra ≈
1.37 µm) in natural seawater
540 O. Olaogun and E. T. Akinlabi

Fig. 2 Polarization curves

of hot rolled aluminium alloy
8015 samples immersed in
natural seawater at Ra ≈
1.37 µm

Figure 1 illustrates that the OCP variation for the three hot rolled aluminium alloy
8015 samples with 1.37 µm surface roughness slightly shifted to the lower negative
values. This observation shows that the inert thin-film oxide layers of the samples
exhibit similar resistance to seawater electrolyte attack. The recorded OCP values
for samples 1, 2 and 3 against Ag/AgCl are −0.693 V, −0.696 V and −0.714 V,
respectively.
The polarization curves generated from the potentiodynamic experiments are
illustrated in Fig. 2. The polarization curves show the anodic scan (progresses in
the positive direction) and cathodic scan (progresses in the negative direction). Point
O denotes the open circuit or rest potential. At this point, the measured current is
relatively zero, confirming that the sum of the anodic and cathodic reaction rates on
the hot rolled AA8015 electrode is zero. In the anodic scan, the potential increases
up to region B. At region A, the steeply rise of the anodic current confirms dominant
metal oxidation reaction taking place. Likewise, in the cathodic scan, a fast reduc-
tion reaction is observed. Rapid decrease in the applied potential in region C shows
evidence of no change in the reaction rate.

3.2 Corrosion Rate

Corrosion rate of the hot rolled aluminium alloy 8015 at 1.37 µm was analysed from
the polarization curve using Tafel analysis. The Tafel analysis for the three samples
is given in Fig. 3. The straight lines that best fit the data at high potentials were
achieved at selected potential range markers on both the anodic and cathodic curves.
The straight lines were extrapolated until it intersects. Table 1 shows the computed
corrosion potential (E corr ), corrosion current density (I corr ) polarization resistance
(Rp ) and corrosion rate (Table 2).
Corrosion Behaviour of Hot Rolled AA8015 … 541

Fig. 3 Tafel analysis of hot

rolled aluminium alloy 8015
at Ra ≈ 1.37 µm. a Sample
A1; b Sample A2; c Sample
A3

Samples 2 and 3 computed corrosion rate is selected to determine the average

corrosion rate of 0.029385 mm/y. This confirms that the hot rolled aluminium alloy-
8015 of 7 mm plate will have a service life of 238 years in natural seawater before
it is dissolute completely.
542 O. Olaogun and E. T. Akinlabi

Table 2 Computed corrosion

Sample A1 Sample A2 Sample A3
rate of hot rolled aluminium
alloy 8015 at 1.37 µm Tafel data Tafel data Tafel data
surface-roughness E corr (V) −0.7281 −0.7264 −0.715
I corr (A/cm2 ) 1.12E − 06 2.43E − 06 2.97E − 07
Rp () 9879 4968 3520
Corrosion rate 0.01217* 0.02646 0.03231
(mm/y)
* indicates the corrosion rate value asterik in the table was not use
to calculate the mean corrosion rate

(a) (b)

Black
spots
showing
localised

Fig. 4 Macrographs showing photographic observation for hot rolled aluminium alloy 8015 sample
at 1.37 µm surface roughness. a Image before corrosion of hot rolled sample; b images after
corrosion for hot rolled sample

3.3 Visual and Microstructural Observation

Figure 4 shows the visual camera images of the prepared hot rolled alloy surface
before and after electrochemical corrosion. A larger surface area of the aluminium
alloy surface is attacked in the form of localized corrosion with the existence of black
spots observed on the surface of the alloy. Further microstructural observation using
TESCAN VEGA scanning electron microscope revealed pitting corrosion, as shown
in Fig. 5.

4 Conclusions

This research has provided an electrochemical study on corrosion behaviour of alu-

minium alloy 8015 at 1.37 µm in natural seawater solution. From the experimental
investigation, it is shown that:
Corrosion Behaviour of Hot Rolled AA8015 … 543

Fig. 5 SEM images of corroded hot rolled aluminium alloy 8015 at 1.37 µm surface rough-
ness. a Image before corrosion at 271- × magnification; b–d images after corrosion at amplified
magnifications

• The potentiodynamic polarization curves for the three samples are similar, show-
ing an actively corroding region. This confirms precision of the experimental
investigation.
• Mean corrosion rate of the hot rolled aluminium alloy 8015 is at 1.37 µm, and
surface roughness under natural seawater condition at room temperature of 20 ±
2 is given as 0.029385 mm per year.
• Microstructural analysis of the hot rolled aluminium alloy 8015 surface after
electrochemical corrosion reveals localized corrosion in the form of pitting.

Acknowledgements The Council for Scientific and Industrial Research (CSIR), South Africa, is
acknowledged for provision of financial support.
544 O. Olaogun and E. T. Akinlabi

References

1. Dieter GE, Kuhn HA, Semiatin SL (2003) Handbook of workability and process design. ASM
International, USA
2. Gundersen JTB, Aytaç A, Ono S, Nordlien JH, Nişancıoğlu K (2004) Effect of trace elements on
electrochemical properties and corrosion of aluminium alloy AA3102. Corros Sci 46(2):265–
283
3. Laferrere A, Parson N, Zhou X, Thompson G (2013) Effect of microstructure on the corrosion
behaviour of extruded heat exchanger aluminium alloys. Surf Interface Anal 45(10):1597–1603
4. Shahidi M, Gholamhosseinzadeh MR (2015) Electrochemical evaluation of AA6061 aluminum
alloy corrosion in citric acid solution without and with chloride ions. J Electroanal Chem
757:8–17
5. Krüger L, Schwarz F, Mandel M, Hockauf M (2015) Electrochemical corrosion studies of
ultrafine-grained aluminium alloy EN AW-6063. Mater Corros 66(3):226–232
6. Dai X, Wang H, Ju L, Cheng G, Cong H, Newby BZ (2016) Corrosion of aluminium alloy
2024 caused by Aspergillus Niger. Int Biodeterior Biodegrad 115:1–10
7. Aribo S, Fakorede A, Ige O, Olubambi P (2017) Erosion-corrosion behaviour of aluminum
alloy 6063 hybrid composite. Wear 376–377:608–614
8. Tian G, Jin G, Zhang W, Huang Z, Yang Z (2017) Investigation on electrochemical corrosion
characteristic of 2a14 aluminum alloy in nitric acid. Surf Rev Lett 24(Supp 01):1850016
9. Deepa P, Padmalatha R (2017) Corrosion behaviour of 6063 aluminium alloy in acidic and in
alkaline media. Arab J Chem 10:S2244
10. Donatus U, Thompson GE, Omotoyinbo JA, Alaneme KK, Aribo S, Agbabiaka OG (2017)
Corrosion pathways in aluminium alloys. Trans Nonferrous Metals Soc China 27(1):55–62
11. Zhang X, Zhou X, Hashimoto T, Liu B (2017) Localized corrosion in AA2024-T351 alu-
minium alloy: transition from intergranular corrosion to crystallographic pitting. Mater Charact
130:230–236
12. Liu B, Zhang X, Zhou X, Hashimoto T, Wang J (2017) The corrosion behaviour of machined
AA7150-T651 aluminium alloy. Corros Sci 126:265–271
13. Seri O (1994) The effect of NaCl concentration on the corrosion behavior of aluminum
containing iron. Corros Sci 36(10):1789–1803
14. Verma AS, Sumankant S, Yashpal NM (2015) Corrosion behavior of aluminum base particulate
metal matrix composites: a review. Mater Today: Proc 2(4–5):2840–2851
15. Polmear I, John D, Nie J-F, Qian M (2017) 2-physical metallurgy of aluminium alloys. Light
alloys. Elsevier Ltd, pp 31–107
16. Al-Moubaraki AH, Al-Rushud HH (2018) The red sea as a corrosive environment: corrosion
rates and corrosion mechanism of aluminum alloys 7075, 2024, and 6061. Int J Corrosion
2018:1–15
17. Lameche-Djeghaba S, Benchettara A, Kellou F, Ji V (2014) Electrochemical behaviour of pure
aluminium and Al–5%Zn alloy in 3% NaCl solution. Arab J Sci Eng 39(1):113–122
18. Kim S, Jang S, Han M, Park J, Jeong J, Chong S (2013) Mechanical and electrochemical
characteristics in sea water of 5052-O aluminum alloy for ship. Trans Nonferrous Metals Soc
China 23(3):636–641
19. ASTM International. (a). ASTM G1-03 standard practice for preparing, cleaning, and
evaluating corrosion test specimens. ASM International
20. ASTM G5-14 Standard Reference test method for making potentiodynamic anodic polarization
measurements. ASTM G5-14 Standard Reference test method for making potentiodynamic
anodic polarization measurements
21. ASTM International. (c). ASTM G69-97 (2009) Standard test method for measurement of
corrosion potentials of aluminum alloys. ASM International
Influence of Tool Profiles on Heat
Transfer Analysis in Al-6061 Alloy Using
Friction Stir Welding

M. Bala Chennaiah, A. Sreenivasulu, and K. Ravi Kumar

Abstract The following paper describes the thermal analysis on friction stir welded
aluminum alloy AA 6061. The tool for this investigation is made up of high-speed
steel material. We considered tool profile as tapered and the pin is with threaded
and without threaded and also discusses the process parameters and their effects
of thermal analysis of similar welded aluminum alloys. The joint is fabricated at a
tool rotational speed of 825 rpm, feed rate of 32 mm/min and angle of 3 degree for
the joining process. There is no surface-level cracks in a welded region by using
liquid penetrant testing (NDT) method. The heat transfer model and evolution of
temperature during the process of FSW are calculated theoretically, and the values
of heat rate and peak temps are evaluated. These are compared with practical analysis
values based on ANSYS values.

Keywords Aluminum alloys · Tapered tool · Heat transfer · Temperature

distribution · Tensile · Hardness

1 Introduction

Friction stir welding (FSW) is a process of joining of metals in solid state. This
process is efficient in energy, eco-friendly and comfortable to do. It is considered
that the most significant development in the metal joining process in the twenty-first
decade. 6061 Aluminum alloys are widely used for structural and marine applica-
tions. They have the advantage of being lightweight at the same time exhibiting
good tensile strength. Non-consumable tool is used to join two faces of workpieces
without melting the workpiece material. In Friction stir welding (FSW) major role
play the tool material, tool profile, rotational speed, axial force and also temperature
distribution perpendicular to the tool moment. Heat is generated by friction between
the rotating tool and the workpiece and plastic deformation of workpiece, which
leads to a softened region near the FSW tool. While the tool is traversed along the

M. Bala Chennaiah (B) · A. Sreenivasulu · K. Ravi Kumar

ME Department, V.R. Siddhartha Engineering College, Vijayawada, India
e-mail: chennai303.mech@vrsiddhartha.ac.in

© Springer Nature Singapore Pte Ltd. 2021 545

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_48
546 M. Bala Chennaiah et al.

joint line, it mechanically intermixes the two pieces of metal, and forges the hot and
softened metal by the mechanical pressure, which is applied by the tool, much like
joining clay. Thus, a joint is produced in solid state. It is primarily used on wrought
or extrude and particularly for structures which need very high weld strength.

2 Literature Review

A Daneji et al. [1] studied about “ Influence of tool profile and processing parameters
on welding defects and mechanical properties of friction welding of AA 6061.”
In this study, he took rotational speed, welding speed, pin height and pin thread
pitch as input parameters. However, the weld quality-related defects and mechanical
properties were treated as output parameters. A cylindrical tool of external thread
is taken as tool geometry, rotational speed of 1750 rpm and weld speeds of 6, 9,
12 mm/min. From the experimental results, he concludes that welding speed plays
a key role toward the heat generation and lead to the generation of wormhole in
the welding nugget in case of wrong selection. The thread pitch has a controlling
influence on the stirring action and it can contribute toward the material accumulation
toward the retrieving side of the welding.
M.S. Srinivasa Rao et al. [2] studied about “Study of weld characteristics during
friction stir welding of aluminum alloy 6061.” They selected AA6061 of dimensions
200*60*6 mm and two types of tools such as taper threaded tool and half grooved
tool. They used 900 rpm, feed 16 mm/min, and 2 degrees of tilt angle as process
parameters for the welding process. They observed that there is no influence on tool
profile on hardness value. Hardness is mainly influenced by rotational speeds. The
maximum Brinell’s hardness numbered 69 BHN occurred at 900 rpm welded using
the half grooved tool.
Ch Mohana Rao et al. [3] studied on “Friction Stir Welding of Aluminium Alloys
6061-To-6061 Similar Metals.” Friction stir welding is performed to join 4 mm thick
plates of 6061 aluminum to aluminum with varied parameters (like tool rotation
speed (rpm), welding speed (mm/min), and the joining conditions are characterized.
All welds were defect-free. Microstructure of weld and microhardness were shown
at center of the weld. Tensile strength was good. As tensile strength increases from
710 to 1120 rpm, hardness decreases from 710 to 1120 rpm.
Abdul Arif et al. [4] studied on “Temperature Validation for Friction Stir Welding
of Dissimilar Aluminum Alloys.” The results of thermal analysis which are simulated
using ANSYS® are used to evaluate the temperature distribution in the workpiece
geometry. From ANSYS software, evaluated temperature distribution in dissimilar
plates and compared actual models. Values of temperatures received by the finite
element method are close to the real temperature distribution in the welded construc-
tion. % of error between experimental and simulation results do not exceed 5%. The
analysis results are closely matched with experimental results.
CM Chen et al. [5] studied the FEM of friction stir welding—thermal and
thermo-chemical analysis. They took the AA6061-T6 as base metal of dimensions
Influence of Tool Profiles on Heat Transfer … 547

240*50*6 mm, AIRKOOL STEEL of shoulder and diameter of 24 mm, 6 mm is

used as a tool material. They took the rotational speed of 500 rpm and weld speeds
of 140 mm/min, 350 mm/min for welding. In the experimental results, they found
that if there is an increase in speed from 140 mm/min to 350 mm/min, there will
be a decrease in temperature of the weld plate, especially in the weld zone. They
calculated the heat generation during welding and used it as an input for ANSYS
analysis of temperature, and they got the highest temperature 425 degrees at weld
speed of 140 mm/min and rotation of 500 rpm (Figs. 1 and 2).
W. Tang et al. [6] “Studied about heat input and temperature distribution in friction
stir welding”. They took AA6061 as a base material of dimensions 300*100*6.5 mm,
rotational speeds of 300 rpm,1200 rpm and travel speed was 2 mm/s. In the studied
they observed that grain size of the weld zone is notably smaller and more equiaxed
than that of parent metal as a result of recrystallization. They used thermocouples for
measuring the temperature along the weld zone. The maximum temperature obtained
is 450°. Perumalla Janki Ramulu et al. [7], “ Experimental Study On Temperature

Fig. 1 Before the FSW

process

Fig. 2 After the FSW

process
548 M. Bala Chennaiah et al.

Evolution During Frication Welding Of 6061-T6 Aluminum Alloy.” they were con-
sidered AA 6061 as a base metal of thickness 2.1 mm, three types tools of shoulder
diameters 12 mm, 15 mm, 18 mm and pin with equal probe lengths are used for entire
process. They used thermocouples to measure the temperature distribution. They con-
ducted the experimental work taking rotational speed of 1300, 1400, 1500 rpm and
weld speeds of 90,100,120 mm/min and tool tilt angle of 2.5°. They concluded that
increase in shoulder diameter and rotations speeds increase the temperature.
Summary of Literature Survey: Work done has been done on FSW of Al6061
plates of different thickness, different researches concluded that tool rotational speed,
welding speed, axial force, pin profile, pin diameter, shoulder profile, shoulder diam-
eter, tool tilt angle are the important process parameters which determine the quality
of weld. Before the start of the experiment this values has to be chosen carefully.
Improperly selection this process parameters leads to weaker joints containing several
defects like tunnel defect voids, cracks etc.

3 Experimental Procedure

Aluminum plates of 6 mm thick, Al 6061 alloy base metal were cut to the required
size (150 m × 50 mm) by power hacksaw cutting and milling. Square butt joint
configuration (300 mm × 300 mm) was prepared, and single-pass welding proce-
dure was followed to fabricate the joints. The chemical composition and mechanical
properties of base metals are presented in Tables 1 and 2. The non-consumable tool
is made of high-speed steel (HSS) material. The shoulder and pin of the tool are
designed and machined by using the CNC machine. The shoulder diameter profile
of the tool (20 mm) is flat and pin length (5.6 mm) is tapered with thread, and the
other one is tapered without thread. Then, both tools are made by using heat treat-
ment process. Machining process is done by the universal milling machine and the
plates are tightened by clamping force of 2 kN by varying process parameters that
are tool rotational speed, feed rate, and tilt angle. The process parameters are shown
in Table 3. The plates were successfully welded using friction stir welding tech-
nique. The welded samples were first visually inspected to element any defective
workpieces. All the samples passed this test. Next, the samples were subjected to
non-destructive tests. Only one sample (Al6061 and Al6061 without threaded tool)

Table 1 Mechanical properties of Al 6061

Property Yield strength Tensile strength Elasticity modulus Brinell hardness
Value 276 MPa 310 MPa 68.9 GPa 95 HB

Table 2 Heat value at tool

S. No. Tool rotational speed Amount of heat (kW)
rotational speed
1. 700 5.18
Influence of Tool Profiles on Heat Transfer … 549

Table 3 Heat and

Plate Heat value (kW) Temperature (°C)
temperature distribution on
plates 6061-6061 3.626e3 399.714
6061-6061 2.59e3 362.285

Table 4 Theoretical and experimental temperature values

Plate Theoretical temperature (°C) Practical temperature (°C) % error
6061-6061 385.65 399.714 3.51
6061-6061 385.65 362.285 6.05

was found, and all others were non-defective. In the present work, 6 mm thick com-
mercial grade aluminum plates have been welded by using different types of tool
(tapered and tapered with thread). From the numerical analysis, it is observed that
the tool with tapered threaded pin profile generates less temperature than the tool
with tapered pin profile because for the same base diameter the cylindrical pin has
more contact area with the workpiece than the contact area of pin with conical pin
profile (Table 4).

3.1 Heat Transfer Analysis

A 3D finite element transient thermal model was developed in the present work to
analyze the heat transfer and temperature distribution in the FSW process. The actual
conditions were accommodated as far as possible in the present thermal model. The
following assumptions were made in developing the thermal model of a FSW process
(Figs. 3 and 4).
The temperature calculation is based on Fourlier’s equation:

dT
ρc = div(k · gradT ) + qin Ω (1)
dt
where q is the power generated by friction between the tool and the top of the
workpiece and by the plastic deformation work of the central weld zone, T is the
temperature, n is the conductivity, r is the material density, and c is the heat capacity.
The main heat source in FSW is generally considered to be the friction between the
rotating tool and the welded plates, and the “cold work” in the plastic deformation
of material in the vicinity of the tool. Considering an element at the contact surface
between the tool shoulder and the top surface of workpiece, the rate of heat generation
derived from the friction in the element at radius r is:

d q̇ = 2π ω.r 2 μ(T ) p(T )dr (2)

550 M. Bala Chennaiah et al.

Fig. 3 6061 with thread tool—heat distribution

Fig. 4 6061 without threaded tool—heat distribution

Influence of Tool Profiles on Heat Transfer … 551

The rate of heat generation (caused by the friction) over the entire interface of the
contact will be:

R0
2  
q̇ = 2π ω.r 2 μ(T ) p(T )dr = π ωμ(T ) p(T ) R03 − r03 (3)
3
r0

The rate of heat generation at the interface between the shoulder and the top of
the workpiece surface is a function of the coefficient of friction m, angular velocity
w, and radius r. As the m (T ) and p (T ) are dependent on the local temperature and
the radius r, Eq. (3) is difficult to evaluate. As the temperature increases, the friction
coefficient is expected to decrease, and the work of the plastic formation increases.
In this model, the p is from the experimental measurement. A constant value of the
friction coefficient is used to approximate the comprehensive effect of both factors
of thermal and plastic effects during FSW, and the predicted temperature history is
verified by the measured ones. The heat generation in the vicinity of the probe is
considered the same as that at the shoulder periphery with a radius equal to that of
the probe.

3.2 Heat Analysis

From the considering, the Fourier heat equation

dT
ρc = div(k · gradT ) + qinΩ
dt
2  
q= π ωμ(T ) p(T ) R03 − r03
3
We assume that
Shoulder diameter R0 = 20 mm
Probe diameter r 0 = 7 mm
We taken μ = 0.5

2π N
ω=
60
At 700 rpm,

2 2π 700 5000  
q= π 0.5  2 (T ) p(T ) 0.01 − 0.0035
3 3
3 60 π 3.5∗10−3
= 5185.7 W
= 5.18 kW
552 M. Bala Chennaiah et al.

For 6061-6061(P = 5.18 kW),

π 2
5.18 × 103 W = 5.67 × 10−8 m2 K ×
W
2.0 × 10−3 × T 4
4
5.18 × 103 × 4
= T4
5.67 × 10−8 × π × 202 × 10−6

Tthe = 385.65 ◦ C

4 Conclusions

In this experiment, the two process parameters of FSW, i.e., feed rate and tool rota-
tional speed are varied to find the effect on the thermal properties of friction stir
welded Al 6061 alloy plates of 6.0 mm thickness of those parameters. A combina-
tion of tool rotational speeds and welding speeds were employed in this experiment.
As a result, two joints were fabricated. The following are the results that are derived
from these experimentations.
• We chosen two types of tools profiles: one of them is tapered with thread and
another type is tapered without thread tool.
• Then, the highest temperature is generated at taper with thread tool, the amount
of increase temperature in %, i.e., 6.05%

Acknowledgements We sincerely thanks to VR Siddhartha Engineering College management pro-

viding necessary resources and Dr. N. Vijaya Sai, Head of the Department, Mechanical Engineering,
VRSEC for his continued encouragement to completion of this work.

References

1. Daneji A, Ali M, Pervaiz A (2018) Influence of tool geometry and processing parameters on
welding defects and mechanical properties for friction stir welding of 6061 alloy. Mater Sci Eng
346
2. Srinivasa Rao MS, Ravi Kumar BVR, Manzoor Hussain M. Experimental study of weld char-
acteristics during friction stir welding (FSW) of Aluminum Alloy (Aa6061-T6). Int J Res Eng
Technol. ISSN: 2319-1163
3. Mohana Rao C, Mallikarjuna Rao K (2017) Friction stir welding of Aluminium Alloys 6061-
To- 6061 similar metals. Int J Mech Eng Technol (IJMET) 8(1):264–269
4. Arif A et al (2013) Temperature validation for friction stir welding of dissimilar Aluminum
Alloys. In: International Conference on Advances in Mechanical Engineering, ICAME-2013
5. Chen CM, Kovacevic R (2003) Finite element modeling of friction stir welding thermal and
thermo mechanical analysis, by Received 12 May 2003; Accepted 10 June 2003
Influence of Tool Profiles on Heat Transfer … 553

6. Tang W, Guo X, Mcclure J. Heat input and temperature distribution in friction stri welding.
University of Texas at El pasa
7. Ramulu PJ, Ganesh Narayana R (2014) Experimental study on temperature evolution during
frication welding of 6061-T6 Aluminum Alloy. (AIMTDR 2014) 12th–14th Dec 2014
Tensile, Hardness, and Impact Properties
of Amorphous Al–Si–Mg Cast Alloys

Olayide R. Adetunji, Adeniran S. Afolalu, Mufutau A. Mustapha,

Oluwasegun J. Adelakun, Samson O. Ongbali, and Abiodun A. Abioye

Abstract This research work focused on tensile, hardness, and influence properties
of amorphous Al–Si-Mg cast alloys. Enhancement of aluminum alloy with Mg and
Si was performed after which heat treatments were carried out on the produced
samples. The result showed the percentage of Mg as 0.14, 4.14, 0.14, 0.32, and
0.20%, for as-cast aluminum, annealed, age-hardened, hardened, and glassy-phase
samples, respectively. The percentages of Si were 2.01%, range of 1.94–4.10% and
3.37% for as-cast, heat-treated, and glassy-phase aluminum alloys, respectively. The
impact tests showed an increase in impact value from 5.423 J of as-cast sample to
8.135 J of age-hardened sample. Hardened sample has the lowest impact value of
2.712 J while glassy-phase sample has the highest impact value of 10.847 J. The
ultimate tensile strength of annealed sample is higher compared to as-cast while
glassy-phase has the highest value. The SEM of the annealed alloy sample contains
bright and dark regions indicating α and β phases, respectively.

Keywords Mechanical property · Amorphous al alloy · SEM characterization

1 Introduction

Aluminum is a substance having a unique significant as a result of the proportion of

its capacity to weight [1], that is, simplicity in recovering the substance through a
process of recycling, opposing deterioration of materials, high tendency of transmit-
ting heat, and electricity. Utilization of alloys of aluminum, composite, and titanium
was adopted in reducing aircraft weight. Wide variety of solutions were assumed
in the design, operation, and manufacturing of automobiles in which the alloy of
aluminum constitutes a part of it [2]. In most instances, alloys of wrought and cast

A. S. Afolalu (B) · O. J. Adelakun · S. O. Ongbali · A. A. Abioye

Mechanical Engineering Department, College of Engineering, Covenant University, Ota, Nigeria
e-mail: Sunday.afolalu@covenantuniversity.edu.ng
O. R. Adetunji · M. A. Mustapha
Mechanical Engineering Department, Federal University of Agriculture, Abeokuta, Nigeria

© Springer Nature Singapore Pte Ltd. 2021 555

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_49
556 O. R. Adetunji et al.

aluminum are applicable in joining the roof of automobile, engine, and side panels to
the underbody, [3]. Steel or iron was previously used for automobile; but discovery
has tested alloys of aluminum proven to result into lighter part which might occur to
be complicated most times [4]. As a result of the increase in part design or a material
property, it is most likely to generate a specific part that will perform the activities
of other parts, thereby making it easy to minimize the parts to be produced [5].
The particular combination of metals and another element type leading to Al–
7Si–0.4 Mg is normally utilized in automobile part, aircraft; as a result of the Mg
in the combination, deterioration of materials is opposed [6]. Cast Al–Si–Mg alloys
are applicable in automobile and aviation industry as a result of their excellent char-
acteristics and the ratio of its capacity to weight [7]. Mechanical properties required
for the aforementioned application are excellent opposition to weakening of mate-
rials, high proof strength, and high resistant to fracture [8]. Heat treatment of cast
Al–Si–Mg alloys is normally examined in the aspect of homogeneous solid solution,
undesirable low-temperature changes been hindered, and accelerated changes occur-
ring in the properties [9], and this method for strengthening materials is utilized in
other to achieve a preferred combination of its mechanical properties [10, 11]. This
research work therefore aimed to produce amorphous Al–Si–Mg cast alloys with an
improved mechanical property.

2 Materials and Methodology

The material, tools, and equipment required for the production of the aluminum alloys
and heat treatments included aluminum ingot, silicon ingot, magnesium ingot, alloy
additives, heating furnace, vernier calipers, and hacksaw.

2.1 Procedure for Producing Aluminum Alloy Casting

The aluminum (Al), magnesium (Mg), and silicon (Si) ingots were properly dried
with acetone. In other to easily detach the mold from the casting, graphite coatings
are made available within the mold. The casting of aluminum alloy was carried out
at research center in the country. The molds were prepared toward a particular high
temperature. Electric induction furnace was used for the melting of metal, and the
alloy was inverted twice and was allowed to cool in sand. The alloys produced were
categorized into as-cast sample and heat-treated samples. The cast samples were
analyzed to determine the percentage of elemental composition in the alloys.
Tensile, Hardness, and Impact Properties … 557

2.2 Procedure for Heat Treatment

In other to enhance the mechanical properties of cast alloy, heat treatment process
was done on mechanical test samples. Heat treatment process comprises mainly of
age-hardening, annealing, hardening, and glassy phase which were carried out at the
research center.

2.3 Chemical Composition Analysis of the Test Samples

Percentage weight elemental composition of the alloys produced was carried out
in a local aluminum rolling mill. The equipment used for the analysis was optical
emission spectrometer. The test was carried out following the procedure below:
The samples were prepared by forming into the shape accommodated by the
machine. This involved cutting and filling to shape, and prepared samples as described
above were then sparked using argon gas after which the sample was inserted into
metal analyzer for proper analysis.

2.4 Materials and Equipment for Mechanical Testing

of the Samples

Various mechanical tests were carried out on the samples; hardness and impact
tests were carried out at the university laboratory. Tensile test was carried out in
another university. The equipment used included meter rule, testing equipment appa-
ratus, hardness testing equipment, tensile strength testing equipment, and inspection
equipment apparatus.
Hardness Impact and Tensile Tests. Rockwell hardness testing machine was
used for Brinell hardness test using heat-treated samples and control samples. The
surface was prepared by grinding and polishing. 20 mm length of no specific diameter
was sectioned out from the sample and fixed on the machine using a dwell time of
15 s.

2.5 Procedure for Microstructure Examination

The SEM is majorly applicable in the evaluation of samples that are thick by putting
together electrons which are emitted from the sample under experiment so as to
obtain secondary electrons. In addition, other signals including electron beam were
also collected. This was done using Phenom ProX scanning electron microscope,
model number-800-07334.
558 O. R. Adetunji et al.

3 Results and Discussion

The findings of this research are contained in Tables and Figures.

3.1 Results of Aluminum Casting

The casting produced was grouped into two for the purpose of comparison on the test
to be carried out: The first group is as-cast (i.e., non-heat-treated samples), and the
second is heat-treated sample. Both groups of samples were subjected to chemical
analysis to determine the percentage weight composition of elements in the aluminum
alloy produced (Fig. 1).

3.2 Results of Chemical Analysis

The result of chemical analysis is shown in Table 1.

3.3 Results of Scanning Electron Microscopy Test

The scanning process as earlier explained was followed by the EDX analysis with
the acquired microstructure using the Phenom identification software. The scanning
electron micrograph in Fig. 2 shows the microstructure of annealed alloy sample.
The bright region indicates the primary solid solution, and the dark region shows the
non-equilibrium eutectic solid solution between the grains. The energy dispersive
X-ray as shown in Fig. 3 reveals the chemical composition. Figure 4 shows the SEM
of age-hardening test carried out on the alloy samples. The emergence of a scattered
solid form from a solution in combining two or more metals is the aim of the solid-
hardening process. The fine solid in the combination of the two or more metals as
seen in the micrograph prevents abnormality separation in mobility by forcing the
abnormality to pass through the solid particles or encompasses round it. Preventing
abnormality in mobility makes combination of two or more metals to be strengthened
(Figs. 5 and 6).
Tensile, Hardness, and Impact Properties … 559

Fig. 1 a SEM image of annealed test sample b Energy dispersive spectrometer

3.4 Results of Mechanical Tests

Mechanical tests carried out on the samples were impact test, hardness test, and
tensile test, and the results are as given below. These results are given in Tables 2
and 3.
 560

Table 1 Percentage chemical composition of element in the cast sample

Heat-treat % Elemental Composition
Mg Si Mn Cu Zn Ti Fe Na B Al Sn Pb
As-cast 0.14 2.01 0.36 0.42 0.30 0.01 0.59 0.00 0.00 96.31 0.01 0.02
Annealing 4.14 3.94 0.90 0.24 2.48 0.12 3.00 0.12 0.059 81.51 2.81 0.68
Age-Hardening 0.14 1.94 0.21 0.40 0.31 0.01 0.72 0.01 0.00 96.24 0.01 0.02
Hardening 0.32 4.10 0.12 0.51 0.41 0.01 2.81 0.00 0.00 91.68 0.02 0.03
Glassy-Phase 0.20 3.37 0.42 0.37 0.01 2.36 0.00 0.00 93.12 0.01 0.03
0.12
O. R. Adetunji et al.
Tensile, Hardness, and Impact Properties … 561

Fig. 2 a SEM image of the As-cast sample b Energy dispersive spectrometer (EDS)

3.5 Discussions of Results

The results show the percentage weight of elements in the as-cast, heat-treated, and
glassy-phase aluminum alloys. The as-cast aluminum alloy shows a percentage of
Mg 0.14%, heat-treated samples show annealed sample to be 4.14%, age-hardened
sample 0.14%, hardened sample 0.32%, and glassy phase shows 0.20%. The per-
centage of Si in the as-cast Al alloy is 2.01 by wt% as the range is between 1.94%
and 4.10% in the heat-treated aluminum alloy while glassy phase is 3.37%.
The impact value of as-cast sample is 5.423 J just as it is in annealed sample; age-
hardened sample experiences an increase in the impact value to 8.135 J, and hardened
sample has aÀ reduced impact value of 2.712 J. Glassy phase has the highest impact
value of 10.847 J.
Comparatively, the results of hardness (HRB) value of as-cast sample is 68, and
this value remained in hardened sample and increased to 70 in the annealed sample
while the value reduced to 66 in age-hardened sample. Glassy phase only shows the
lowest rate of reduction to 62. This result conformed to those of other researchers
like Adeyemi et al. [10].
562 O. R. Adetunji et al.

Fig. 3 a SEM image of age-hardened sample b Energy dispersive spectrometer

The tensile yield strength of as-cast is 82 MPa, and the ultimate tensile strength
is 94.22 MPa while the percentage elongation is 3.62. The test also shows tensile
yield strength in annealed sample to be 98 MPa, ultimate tensile strength increased
to 108.24 MPa, and the percentage elongation is 2.81. In age-hardened sample shows
tensile yield strength increased to 114 MPa, ultimate tensile strength is 91.81 MPa,
and percentage elongation reduced to 1.60. The results show the tensile strength to be
116 MPa, ultimate tensile strength as 87.98 MPa, and percentage elongation is 1.62.
Glassy-phase sample shows tensile yield strength to be 116 MPa, ultimate tensile
strength to be 115.87 MPa, and percentage elongation to be 2.68.
The bright region in the SEM of the annealed alloy sample indicates the primary
solid solution which accounted for solid solution hardness got in the result, and the
dark region shows the non-equilibrium eutectic solid solution between the grains
which accounted for the higher tensile and impact toughness values. This is in line
with the findings of Chen et al. [11, 12]. The energy dispersive X-ray reveals the
Tensile, Hardness, and Impact Properties … 563

Fig. 4 a Micrograph of hardened test b Energy dispersive spectrometer (EDS)

traces of the chemical composition while Al (65.8%), O (11.6%), and Br (14.9%)

were predominant; traces of Mg, C, and Si were seen.

4 Conclusions

The results of analysis show that the produced Al–Si–Mg alloys exhibit excellent
casting characteristics evident from the results of microstructural test. Mechanical
properties including impact toughness, hardness, and tensile were of good values.
Glassy-phase alloy had the best impact toughness value followed by age-hardened
one. The best hardness value was obtained for hardened alloy and followed by as-cast,
age-hardened, and lastly glassy phase. The best yield and ultimate tensile strength
564 O. R. Adetunji et al.

Fig. 5 a SEM image of glassy-phase test sample b Energy dispersive spectrometer

were obtained for glassy-phase alloy, followed by age-hardened, annealed, hardened,

and as-cast. As-cast alloy had the best ductility followed by annealed one.

Acknowledgements We acknowledge the financial support offered by Covenant University in the

actualization of this research work for publication.
Tensile, Hardness, and Impact Properties … 565

Fig. 6 Comparison of the results of tensile test samples

Table 2 Summary of results for impact tests

S/N Sample Impact value (J) Hardness test (HRB)
1 As-Cast 5.423 68
2 Annealed 5.423 70
3 Age-Hardened 8.135 66
4 Hardened 2.712 68
5 Glassy-Phase 10.847 62

References

1. Adetunji OR, Olukuade ML, Simka W, Sowa M, Adesusi OM, Okediran IK (2017) Production
and characterization of amorphous Aluminium-Copper alloy for aerospace applications. Eur J
Eng Res Sci 2(2):1–5
2. Afolalu SA, Abioye AA, Dirisu JO, Okokpujie IP, Ajayi OO, Adetunji OR (2018) Investigation
of wear land and rate of locally made HSS cutting tool. In: AIP conference proceedings, vol
1957, no 1. AIP Publishing, p 050002
3. Afolalu SA, Abioye OP, Salawu EY, Okokpujie IP, Abioye AA, Omotosho OA, Ajayi OO
(2018) Impact of heat treatment on HSS cutting tool (ASTM A600) and its behavior during
machining of mild steel (ASTM A36). In: AIP conference proceedings, vol 1957, no 1. AIP
Publishing, p 050003
4. Afolalu SA, Okokpujie IP, Salawu EY, Abioye AA, Abioye OP, Ikumapayi OM (2018) Study
of the performances of nano-case treatment cutting tools on carbon steel work material during
turning operation. In: AIP conference proceedings, vol 1957, no 1. AIP Publishing, p 050001
566 O. R. Adetunji et al.

5. Al-Khazraji KK, Hanna WA, Muhammed OS (2011) Study the effect of polymeric quenching of
wear resistance of eutetic modified Aluminum-Silicon alloy. J Minerals Mater Characterization
Eng 10(10):941–958
6. Abioye A, Abioye OP, Ajayi OO, Afolalu SA, Fajobi MA, Atanda PO (2018) Mechanical and
microstructural characterization of ductile iron produced from fuel- fired rotary furnace. Int J
Mech Eng Technol 9(1):694–704
7. Raj KS, Amit T, Das S (2016) Microstructure and mechanical properties of Al-Si alloy in
As-cast and heat-treated condition. Am J Eng Res (AJER) 5(8):133–137
8. Abdulwahab M, Madugu IA, Asuke F, Fayomi OSI, Ayeni FA (2013) Effect of thermal ageing
treatment on the mechanical properties of antimony-modified A356.0-type Al-Si-Mg alloy. J
Mater Environ Sci 4(1):87–92
9. Umaru OB, Abdulwahab M, Hafsat NM, Maidawa H, Murtada AO, Suleiman TM (2016)
Influence of a low temperature ageing on the properties of Al-6.5%Si-0.35%Mg alloy. Nigerian
J Technol (NIJOTECH) 35(3):550–554
10. Adeyemi DI, Bolaji A, Mosobalaje OA, Oluyemi JO, Moshood DS (2013) Effect of heat
treatment on some mechanical properties of 7075 Aluminum alloy. Int J Mater Res 16(1):190–
194
11. Chen R, Wang L, Apelian D (2000) fludized bed reactor heat treatment of A356 alloy:
microstructure analysis and mechanical properties. Trans Am Foundry Soc 2(4):289–304
Experimental Investigation
on the Influence of Tool Geometry
in Minimum Chip Thickness
of Microendmilling Using Cutting Force
Analysis

M. Prakash, M. Kanthababu, A. Arul Jeya Kumar,

and V. Prasanna Venkadesan

Abstract Minimum uncut (t m ) chip thickness is the promising factor of the

microendmilling process, which is affected by many factors. The tool geometry is
one among them. In this work, the effect of different tool geometries, two flutes and
four flutes of microendmill (500 µm diameter), are investigated, during slot milling
of copper alloy (BSS 249) by considering the surface roughness (Ra ) and cutting
forces. The results show that the feed rate has more impact on the Ra . A higher
order of Ra value is observed with the two flutes microendmill than that of with
four flutes microendmill, due to the closer tool path generation of the subsequent
passes of the four flutes. It is also observed that the transition of chip formation from
interrupted chips to continuous chips has occurred with two flutes and four flutes.
From the cutting forces analysis, it is found that the feed force (F y ) is greater than
that of the transverse forces (F x ) irrespective of the tool geometry. The behaviour
of cutting forces also helps to identify the minimum uncut chip thickness (t m ). The
angular shifting of the cutting forces is also observed concerning the tool geometry.
The material removal mechanism reveals that the ploughing effect is more dominant
than that of the shearing effect with two flutes microendmill. The present investi-
gations have shown that four flutes microendmill is more suitable for tool-based
microendmilling in order to achieve better Ra with minimum cutting forces.

Keywords Microendmill · Tool geometry · Surface roughness · Cutting forces ·

Uncut chip thickness

M. Prakash (B) · A. Arul Jeya Kumar

SRM Institute of Science and Technology, Kattankulathur, India
e-mail: prakashmuniyandi@gmail.com
M. Kanthababu
CEG, Anna University, Chennai, India
V. Prasanna Venkadesan
Indian Institute of Information Technology Design & Manufacturing, Kancheepuram, India

© Springer Nature Singapore Pte Ltd. 2021 567

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_50
568 M. Prakash et al.

1 Introduction

Recently, with more demand for accurate and precise micro-components, the impor-
tance of tool-based micromachining processes has been increased in the num-
ber of fields, such as medical, aerospace and automobile [1]. Among the tool-
based micromachining processes, microendmilling is used for manufacturing the
micro-components like microsurgical equipment, micro-impellor and micro-nozzle.
Microendmilling uses microendmills with diameter in the sub-millimetre range. The
edge radius of a typical microendmill is in the order of a few microns, and also,
the uncut chip thickness (t) is usually lesser than the edge radius. Due to the size
of the microendmill, it is very difficult to achieve the quality in the finished com-
ponents. Previous investigations have indicated that the tool life in the tool-based
micromachining is unpredictable, and also it fails prematurely [2–4].
Microendmilling distinguished from macro-regime machining in terms of size
effect and t [5]. The t is greatly influenced by the chip formation process in the
microendmilling. If the t is less than the minimum uncut chip thickness (t m ) (i.e. 0.33
of nose radius (r e )), ploughing is found to be dominating than that of the shearing [4],
which resulted in the increase of cutting forces. New by et al. [5] made an attempt
to calculate the uncut chip thickness (t) based on the feed per flute and cutting edge
radius, which is shown in Eqs. 1 and 2.

n 2 1 2
Uncut chip thickness (t) = tx sin α − tx sin α cos α + t cos2 α (1)
2πr e 2re x

where
F × 1000
tx = (2)
S×n

where t—uncut chip thickness (µm), t x —feed per flute (µm), F—feed rate
(mm/min), S—rotational speed (rpm), n—number of flute, α—tool rotational angle,
r e —tool nose radius.
Generally, in tool-based micromachining processes the quality of the machined
surface is mostly affected by the characteristics of cutting forces, which are mostly
affected by the tool geometry [6]. Therefore, there is a need for more investigation on
the effect of cutting forces generated during microendmilling. However, few attempts
were made by researchers to study the effect of cutting forces in conventional/macro-
regime endmilling [6–11].
Literature review related to the studies on the effect of tool geometry on the
cutting forces is briefly presented here. Fang et al. [6] carried out the experimental
investigations to study the effect of different tool geometries such as two flutes,
triangle type and semicircular type and its failure modes using cutting force signals
in the microendmilling of brass material. They have observed that the tool rigidity
is much higher with the triangular-based microendmill, than that of the two flutes
Experimental Investigation on the Influence … 569

microendmill. However, the machining quality is found to be poor with triangular-

based microendmill than that of two flutes types of microendmill. Filiz et al. [7]
investigated the effect of two flutes of the microendmill with cutting forces during
microendmilling of copper. They have found that the t m and associated ploughing
effects induce variations in the cutting forces. Li et al. [8] carried out experimental
analysis on the tool wear and surface roughness using two flutes 0.1 mm diameter
microendmill in microendmilling of OFHC Copper. They have observed that, the
effect of the t m on the surface roughness is found to be lower, when the feed per tooth
is lower than that of the t m . Lai et al. [9] developed an analytical model based on the
finite element simulations for microendmilling. It is resulted that the edge radius is
the major factor for the t m . Bissacco et al. [10] found that during microendmilling the
effect of machining process parameters is not similar to that of macro-regime milling
due to the size effect and the reduced stiffness of the microendmill. These differences
are mainly due to the changes in physical phenomena, t m and microstructure of the
work material.
From the literature, it is observed that during microendmilling, the tool geometry
plays a significant role in terms of generation of the t m subsequently cutting forces and
surface roughness. It is also necessary to understand the microendmilling mechanism
with respect to the tool geometry [11–14]. From the literature review, it is also
observed that the effect of cutting forces in microendmilling is not yet fully explored.
In addition to the cutting forces, the other factors that are affected by the tool geometry
in microendmilling are entry and exit of the flutes, which occurs repeatedly during the
cutting process, run out of the tool tip, t m , ploughing, indentation, elastic recovery,
etc. [15–30].
Therefore, in this work investigations are carried out to identify the effect of
different tool geometries such as two flutes and four flutes of the microendmill by
analysing the behaviour of uncut chip thickness (t) along with the cutting forces
(transverse force (F x ) and feed force (F y )).

2 Experimental Set-up

Microendmilling was performed using MIKROTOOLS made multi-process machine

tool driven by a 100 W AC servo drive motor that can provide a rotational speed
up to 5000 rpm. The X-Y table has optical linear scale with resolution of 0.1 µm.
Two flutes carbide microendmill (Model: E9661; Make: AXIS, India) and four flutes
carbide microendmill (Model: E9671; Make: AXIS, India) of 500 µm diameter each
coated with aluminium titanium nitride (AlTiN) of 400 nm thickness are used for
the experimentation (Fig. 2). The workpiece material used in the work is copper
alloy (BSS 249), and its chemical composition is copper 55.70%, lead 3.30% and
zinc 38.58%. Vickers’s hardness of the workpiece is 146.10 HV. The workpiece is
mounted on the worktable of the machine, and the microendmill is clamped to the
spindle unit of the machine (Fig. 1).
570 M. Prakash et al.

Fig. 1 Experimental set-up closer view

Fig. 2 Photograph of the microendmills

Experiments are conducted as per the design of experiments approach (DoE),

using the L16 orthogonal array, and optimal parameters based on the surface rough-
ness such as rotational speed of 2800 rpm, feed rate of 10 mm/min and depth of cut
(DOC) of 50 µm are obtained [27]. The above-optimized cutting parameters are used
to analyse the effect of different tool geometry, namely two flutes and four flutes.
The surface roughness (Ra ) is measured on the machined surfaces at three different
locations (entry, middle and exit along the length of cut) in the feed direction using
non-contact instrument Talaysurf CCI Lite, with 0.983 µm cut of length and 333 µm
traverse length. The average value is recorded and given in Table 1.
Experimental Investigation on the Influence … 571

Table 1 Allocation of input parameters in L16 OA and experimental results

Ex. No. Speed [S] Feed [F] Surface roughness [Ra ] Cutting Forces (N)
(rpm) (mm/min) (µm) Two flutes Four flutes
Two flutes Four [F x ] [F y ] [F x ] [F y ]
flutes
1 700 5 0.194 0.145 0.037 0.771 0.400 0.470
2 700 10 0.163 0.126 0.583 0.771 0.819 0.444
3 700 15 0.169 0.159 1.050 0.875 1.500 0.695
4 700 20 0.219 0.204 2.376 2.498 2.910 1.320
5 1400 5 0.192 0.106 0.057 0.955 0.545 0.389
6 1400 10 0.122 0.101 0.117 0.300 0.491 0.529
7 1400 15 0.108 0.093 0.844 0.910 1.140 0.623
8 1400 20 0.174 0.108 0.759 0.774 1.107 0.292
9 2100 5 0.163 0.079 0.121 0.903 0.538 0.011
10 2100 10 0.088 0.071 0.390 0.827 0.809 0.381
11 2100 15 0.104 0.063 0.084 0.449 0.611 0.351
12 2100 20 0.114 0.086 0.229 0.690 0.617 0.354
13 2800 5 0.125 0.062 0.137 0.938 0.336 0.293
14 2800 10 0.091 0.075 0.178 0.675 0.468 0.388
15 2800 15 0.076 0.066 0.134 0.487 0.425 0.375
16 2800 20 0.101 0.067 0.001 0.419 0.382 0.257

The cutting forces in transverse (F x ), feed (F y ) and axial (F z ) direction were

measured using piezoelectric cutting force dynamometer (Fig. 1). The dynamometer
(Make: Kistler, Model: 9256C2) is mounted on the fixture unit of the machine and
then connected to a charge amplifier, which in turn connected to a PC for acquiring the
data using data acquisition system with DynoWare software (Fig. 1). The directions of
cutting forces (F x , F y and F z ) are measured using the dynamometer which is shown
in Fig. 3. F x , F y and F z are measured along the X-, Y- and Z-axes, respectively.
The cutting force signals are measured at the sampling rate of 1000 Hz with the
regular intervals of 120 s. During microendmilling (slot milling), it is observed that
both downmilling and upmilling occurs, which is also similarly observed by Kang
et al. [20]. The schematic indicates the direction of cutting forces measured during
downmilling and upmilling is shown in Fig. 4a, b, respectively. Figure 4 indicates
that ploughing occurs during the rotation of the microendmill from A to B and E to
D regions during upmilling and downmilling, respectively, whereas shearing occurs
during the rotation of the microendmill, from B to C and C to D regions during
upmilling and downmilling, respectively.
From the preliminary studies, it is observed that the axial force is found to be
non-significant in all the machining conditions due to the lower order of depth of cut
(i.e. 50 µm), which is also similarly observed by Bissacco et al. [10]. Therefore, the
572 M. Prakash et al.

Fig. 3 Schematic diagram of the cutting force components in microendmilling

Fig. 4 Schematic of the direction of cutting forces in the workpiece in microendmilling

axial force is not considered for further analysis. The resultant force is also calculated
by considering transverse and feed force using Eq. 3, which is given as

Resultant force (Fr ) = Fx2 + Fy2 (3)
Experimental Investigation on the Influence … 573

3 Results and Discussion

The effect of two different microendmill geometries such as two flutes and four flutes
during microendmilling on the surface roughness (Ra ), the uncut chip thickness (t),
and cutting forces (F x and F y ) is analysed. They are briefly presented below.

3.1 Surface Roughness (Ra )

Experiments are carried out by varying the rotational speed and feed rate with constant
depth of cut of 50 µm. The effect of two different tool geometries such as two flutes
and four flutes of the microendmill at various cutting parameters on the Ra is shown
in Fig. 5. Typical surface topography of the machined surface with two flutes and
four flutes of microendmill is shown in Fig. 6. From Figs. 5 and 6, it is observed that
Ra values are found to be lower in the four flutes microendmill than that of two flutes
microendmill. This may be due to the formation of continuous chips with four flutes
than that of the two flutes (Fig. 7). This may also be due to the formation of closer tool
path (i.e. due to the increase in number of flutes in the four flutes microendmill) for
subsequent passing of microendmill in the feed direction as illustrated by Bao et al.
[16]. In the case of microendmilling with two flutes, higher Ra is observed due to
interrupted chip formation (Fig. 7). It is also observed that Ra is found to be decreased
with the increase in the rotational speeds up to the feed rate of 12 mm/min both with
the two flutes and four flutes microendmill. However, Ra is found to be increased
as the feed rate increases beyond 12 mm/min. There is no significant variation in
Ra at the higher rotational speed above 1400 rpm with increase in feed rate beyond
12 mm/min. Therefore, in order to achieve a better Ra , four flutes microendmill

0.25 0.25

700 rpm
700 rpm
0.20 0.20
Surface roughness [Ra] (µm)
Surface roughness [Ra] (µm)

1400 rpm

0.15 0.15
2100 rpm 1400 rpm
2400 rpm
0.10 2100 rpm
0.10
2400 rpm

0.05 0.05

0.00 0.00
5 10 15 20 25 5 10 15 20 25
Feed (mm/min) Feed (mm/min)

(a) (b)

Fig. 5 Tool geometry effect on surface roughness in microendmilling. a Two flutes microendmill.
b Four flutes microendmill
574 M. Prakash et al.

(a) (b)

Fig. 6 Typical surface topography of the machined surface in microendmilling. a Two flutes
microendmill. b Four flutes microendmill

100µm 1 0 0 µm

(a) (b)

Fig. 7 Typical chips obtained in microendmilling. a Two flutes microendmill. b Four flutes
microendmill

with higher rotational speed along with moderate feed rate are found to be highly
desirable.

3.2 Uncut Chip Thickness (t)

The behaviour of the uncut chip thickness (t) is analysed during the half rotation
for both two flutes and four flutes microendmill by varying the rotational speed.
Figure 8 indicates the two regions of material removal mechanism in microendmill
based on the uncut chip thickness (t) such as ploughing (regions A to B and D to E)
Experimental Investigation on the Influence … 575

15 15
Feed - 10 mm/min A - B: Ploughing Feed - 10 mm/min A - B: Ploughing
DOC - 50 µm B - C: Upmilling
DOC - 50 µm B - C: Upmilling
10 C - D: Downmilling C - D: Downmilling
10
700 rpm D - E: Ploughing

Shearing
C D - E: Ploughing

Shearing
1400 rpm 700 rpm

Uncut chip thickness (t) [ m]

Uncut chip thickness (t) [ m]

2100 rpm 5 1400 rpm

5 C C
2800 rpm C C
BB C DD A D E
0 A E 0 C
B B D D B B D
A 2800 rpm 2100 rpm
A E
E
A E -5 A

Ploughing
-5

Ploughing
-10 -10

A
E -15
-15 0 30 60 90 120 150 180
0 30 60 90 120 150 180
Tool rotational angle ( [degree] Tool rotational angle ( [degree]

(a) (b)

Fig. 8 Behaviour of uncut chip thickness of different tool geometry with varying speed in
microendmilling. a Two flutes microendmill. b Four flutes microendmill

and shearing (regions B to C and C to D). During shearing (regions B to C and C to

D), both upmilling and downmilling occurred.
Figure 8a indicates that in the case of two flutes microendmill, with the increase
in the rotational speed and there is a decrease in the thickness value of uncut chip
thickness (t). However, ploughing (regions A to B and D to E) is found to be more
dominant than that of the shearing (regions B to C and C to D). This is also reflected
in the formation of interrupted chip (Fig. 6a), in which the chips are not formed until
the uncut chip thickness (t) reaches the minimum uncut chip thickness (t m ) during
the subsequent passes of microendmill along the feed direction. This indicates that
the chip formation did not occur during the single pass of the microendmill, and
therefore, subsequent passes are required for the chip formation process.
Figure 8b indicates the behaviour of uncut chip thickness (t) with four flutes
microendmill with the increase in the rotational speed. From the Fig. 8b, both plough-
ing (regions A to B and D to E) and shearing (regions B to C and C to D) are observed
during the half tool rotation of the microendmill, irrespective of the rotational speed
(Fig. 8b). However, the size of uncut chip thickness (t) is found to be lesser than that
of the two flutes microendmilling. It is also observed that in the four flutes microend-
mill, the shearing (regions B to C and C to D) is found to be more dominating than
that of the ploughing (regions A to B and D to E), which is also reflected in the
formation of continuous chip (Fig. 7b).
Figure 9a, b shows the effect of feed rate with respect to two flutes and four flutes
microendmill on the uncut chip thickness (t) during the half rotation of the microend-
mill. Both ploughing (regions A to B and D to E) and shearing (regions B to C and
C to D) are observed during the half tool rotation of the microendmill, irrespective
of the rotational feed rate (Fig. 9). From Fig. 9a, it is observed that the uncut chip
thickness (t) is found to increase with the increase in the feed rate. It is also observed
that with the increase in the feed rate, ploughing effect dominates the shearing effect.
However, in the case of microendmilling with four flutes microendmill (Fig. 9b), it
is found that the size of uncut chip thickness (t) and the effect of ploughing are
576 M. Prakash et al.

Speed - 2800 rpm A - B: Ploughing Speed - 2800 rpm A - B: Ploughing

4 DOC - 50 µm B - C: Upmilling 4 DOC - 50 µm B - C: Upmilling
C - D: Downmilling C - D: Downmilling
C D - E: Ploughing D - E: Ploughing

Shearing
20 mm/min

Shearing
Uncut chip thickness (t) [ m]

Uncut chip thickness (t) [ m]

15 mm/min C 20 mm/min
2 10 mm/min 2
C 15 mm/min C 10 mm/min
C
C C
0 A B D DE 0 A B C D DE
B B B 5 mm/min D D B B D D
A 5 mm/min E
A E A E

Ploughing

Ploughing
-2 A E -2

A E
-4 -4

0 30 60 90 120 150 180 0 30 60 90 120 150 180

Tool rotational angle ( [degree] Tool rotational angle ( [degree]
(a) (b)

Fig. 9 Behaviour of uncut chip thickness of different tool geometry with varying feed in
microendmilling. a Two flutes microendmill. b Four flutes microendmill

lesser than that of with the two flutes microendmill. This may be due to the closer
trochoidal tool path generation of the four flutes microendmill than that of two flutes
microendmill, which is also similarly observed by Newby et al. [5]. These results
indicate that the tool path generation in microendmilling significantly depends on
the number of flutes and the feed rate, which is also similarly observed by Tansel
et al. and Bao et al. [14, 16]. Higher feed rate with more number of flutes results
in continuous chip formation, which is also similarly observed in the conventional
endmilling [11–13].

3.3 Cutting Forces

Figures 10 and 11 show the typical cutting forces in the transverse and feed directions,

2
Speed - 2800 rpm
Un chip Chip Un chip Chip
Transverse force [Fx] (N)

Four flute microendmill Feed - 10 mm/min

forming forming forming forming Two flute microendmill
zone zone DOC - 50 µm
zone zone
1 B
B
O E
A C
0 O C
A E O - A : Un chip forming zone
A - E : Chip forming zone
A - B: Ploughing
-1 B - C: Upmilling
D C - D: Downmilling
D D - E: Ploughing
Two flute microendmill Four flute microendmill
-2
0 180 360 540 720 900 1080 1260 1440 1620 1800
Tool rotational angle (degree)

Fig. 10 Typical transverse force acquired during microendmilling

Experimental Investigation on the Influence … 577

2 C Four flute microendmill Speed - 2800 rpm

C Feed - 10 mm/min
DOC - 50 µm
1
Feed force [FY ] (N)

O
B D O
A
0
E A B D E O - A : Un chip forming zone
Un chip Un chip Chip
Chip A - E : Chip forming zone
forming forming forming Two flute microendmill
forming
zone zone zone A - B: Ploughing
-1 zone
B - C: Upmilling
C - D: Downmilling
Two flute microendmill Four flute microendmill D - E: Ploughing
-2
0 180 360 540 720 900 1080 1260 1440 1620 1800
Tool rotational angle (degree)

Fig. 11 Typical feed force acquired during microendmilling

respectively, for two flutes and four flutes microendmill. From Figs. 10 and 11, it
is observed that the signature of the transverse force and feed force signals for both
two flutes and four flutes microendmill is found to be in increasing and decreasing
trend both in positive and negative directions with respect to the tool rotational angle.
However, from Figs. 10 to 11, it is observed that, at the earlier stages with four flutes
microendmill there is the significant angular shift in cutting forces at the region of
unchip forming zone (O to A), where no chips are formed, than that of with the two
flutes microendmill. This may be due to the closer trochoidal tool path of the four
flutes microendmill than that of two flutes microendmill. This is well agreed with the
result of Bao et al. [16]. It is also observed from Figs. 10 and 11, that the signature
of the cutting force signal during unchip forming zone (O to A) and chip forming
zone (A to E) is found to be different from the four flutes microendmill than that
of two flutes microendmill, due to the change in frequency of contact of two flutes
microendmill than that of four flutes microendmill.
From Fig. 10, it is observed that the direction changes in the transverse force
from upmilling (B to C) to downmilling (C to D) with both two flutes and four flutes
microendmill as shown in Fig. 4. However, the magnitude of the transverse force
during the upmilling is lower in the case of four flutes microendmill. But in the case
of downmilling, there is no significant variation is observed. Transverse force due
to ploughing (regions A to B and D to E) is more dominant in the case of two flutes
microendmill. From the above, it is observed that, the microendmilling with two
flutes is mostly affected by ploughing than that of shearing.
From Fig. 11, it is observed that increasing and decreasing trends of feed force
from upmilling (B to C) and downmilling (C to D) in both microendmills as shown
in Fig. 4. However, the magnitude of feed force during upmilling is lesser in the case
of four flutes microendmill than that of two flutes microendmill.
Figure 12a, b shows the cutting forces behaviour with respect to the uncut chip
thickness (t) and tool geometry. Feed force is found to be the maximum when the
uncut chip thickness (t) reaches the maximum limit. From Fig. 12, it is also observed
that transverse force approaches nearly zero, when the uncut chip thickness (t) is
578 M. Prakash et al.

4 4 4 4
Speed - 2800 rpm A - B: Ploughing Speed - 2800 rpm A - B: Ploughing
3 Feed rate - 10 mm/min B - C: Upmilling
3 Feed rate - 10 mm/min
3 B - C: Upmilling 3
DOC - 50 µm C - D: Downmilling
DOC - 50 µm C - D: Downmilling
D - E: Ploughing

Uncut chip thickness ( m)

D - E: Ploughing
Uncut chip thickness ( m)

2 C 2 2 2
Feed force (Fy)
1 1 C

Cutting force (N)

1 1

Cutting force (N)

Feed force (Fy) D
0 B 0 0 0
D A B
Transverse force (Fx) E
-1 A E -1 -1 -1
Uncut chip thickness Uncut chip thickness
Transverse force (Fx ) -2 -2
-2 -2
Ploughing Ploughing P Shearing P
Shearing -3 -3
-3 -3
Upmilling Downmilling
Upmilling Downmilling
-4 -4 -4 -4
0 30 60 90 120 150 180 0 30 60 90 120 150 180

Tool rotational angle [ (degree) Tool rotational angle [ (degree)

(a) (b)

Fig. 12 Typical behaviour of cutting forces with the uncut chip thickness of different tool geometry
in microendmilling. a Two flutes microendmill. b Four flutes microendmill

found to be the maximum level. The transverse force is found to change the direction
from positive direction to negative direction during downmilling (regions C to D)
while using two flutes and four flutes microendmill (Fig. 12a, b). By comparing
Fig. 12a, b, it is observed that, shearing mode of material removal mechanism is
dominant with four flutes than the two flutes microendmill. This may be due to the
increase in contact area between the tool and workpiece. Figure 12 also indicates that
the transverse force and feed force are found to be lesser in the case of four flutes
microendmill due to the reduced uncut chip thickness (t).
Figure 13 shows the tool geometry effect on the resultant forces (F r ) with various
rotational speed and feed rate with the constant depth of cut (50 µm). It is observed
that, the rotational speed and feed rate are compounded irrespective of the tool
geometry at the lowest speed of 700 rpm, as similar to the conventional endmilling.
The same trend is observed by Lai et al. [9]. In conventional endmilling, increased
rotational speed and feed rate resulted in the increasing trend of the cutting force

4 4
700 rpm

3 3 700 rpm
Resultant force (N)

Resultant force (N)

2 2

1400 rpm 1400 rpm

1 2100 rpm 1 2100 rpm

2400 rpm 2400 rpm

0 0
5 10 15 20 5 10 15 20
Feed (mm/min) Feed (mm/min)
(a) (b)

Fig. 13 Tool geometry effect on cutting forces in microendmilling. a Two flutes microendmill.
b Four flutes microendmill
Experimental Investigation on the Influence … 579

values [11]. At lower speed of 700 rpm along with increased feed rate, higher cutting
forces are observed due to the dominance of rubbing and ploughing. However, when
the rotational speed increases above 2100 rpm the resultant forces tends to decrease.
In the case of two flutes microendmill, the cutting forces are found to be higher
than that of four flutes microendmill due to the effect of ploughing than that of the
shearing [28].

4 Conclusions

This work presented the investigations on the effect of tool geometries of microend-
mill with two flutes and four flutes during microendmilling of copper alloy (BSS
249) on the surface roughness (Ra ), the uncut chip thickness (t) and cutting forces
(feed and transverse forces). The following conclusions are drawn from the present
investigations.
• Four flutes microendmill resulted in lesser Ra than that of the two flutes microend-
mill, due to the closer tool path generation of the flutes in the microendmill in
the feed direction and also due to the formation of continuous chip. Therefore,
to achieve the better Ra during microendmilling, four flutes microendmill with
higher speed and with moderate feed rate is recommended.
• Based on the behaviour of uncut chip thickness (t), it is found that both plough-
ing and shearing are observed during the half tool rotation of the microendmill,
irrespective of the rotational speed and feed rate, in both two flutes and four flutes
microendmills. However, the size of uncut chip thickness (t) and the ploughing
effect is found to be lesser while using four flutes. Shearing is found to be domi-
nant during microendmilling with four flutes microendmill due to the increase in
contact area of tool–workpiece interface.
• From the cutting forces analyses, it is found that the signature of the cutting force
signals for both two flutes and four flutes microendmill are similar with respect
to the tool rotational angle. However, significant angular shift is observed in the
case of four flutes microendmill due to the closer trochoidal tool path generation
of successive passing of microendmill. Transverse and feed forces are found to be
low in the case of four flutes microendmill due to the reduced uncut chip thickness
(t) in the same parametric level of microendmilling. Cutting forces are found to
be higher with two flutes microendmill due to the dominance of ploughing than
that of the shearing.
580 M. Prakash et al.

References

1. Masuzawa T (2000) State of the art of micromachining. Ann CIRP 49:473–488

2. Alting L, Kimura F, Hansen HN, Bissacco G (2003) Micro engineering. CIRP Ann Manuf
Technol 52:635–657
3. Dornfeld D, Min S, Takeuchi Y (2006) Recent advances in mechanical micro machining. CIRP
Ann Manuf Technol 55:745–768
4. Chae J, Park S, Freiheit T (2006) Investigation of micro cutting. Int J Mach Tools Manuf
46:313–332
5. Newby G, Venkatachalam S, Liang SY (2007) Empirical analysis of cutting force constants in
micro-end-milling operations. J Mater Process Technol 41–47:192–193
6. Fang FZ, Wu H, Liu XD, Liu YC, Ng ST (2003) Tool geometry study in micromachining. J
Micromech Microeng 13:726–731
7. Filiz S, Conley CM, Wasserman MB, Ozdoganlar OB (2007) An experimental investigation of
micro-machinability of copper 101 using tungsten carbide micro-endmills. Int J Mach Tools
Manuf 47:1088–1100
8. Li H, Lai X, Li C, Feng J, Ni J (2008) Modelling and experimental analysis of the effects of tool
wear, minimum chip thickness and micro tool geometry on the surface roughness in micro-
end-milling. J Micromech Microeng 18. https://doi.org/10.1088/0960-1317/18/2/025006
9. Lai X, Li H, Li C, Lin Z, Ni J (2008) Modelling and analysis of micro scale milling considering
size effect, micro cutter edge radius and minimum chip thickness. Int J Mach Tools Manuf
48:1–14
10. Bissacco G, Hansen HN, De Chiffre L (2005) Micromilling of hardened tool steel for mould
making applications. J Mater Process Technol 167:201–207
11. Koshy P, Dewes RC, Aspinwall DK (2002) High speed end milling of hardened AISI D2 tool
steel (~ 58 HRC). J Mater Process Technol 127:266–273
12. Matsumura T, Hiramatsu T, Shirakashi T (2005) A study on cutting force in the milling process
of glass. J Manuf Process 7(2):102–108
13. Li HZ, Zhang WB, Li XP (2001) Modelling of cutting forces in helical endmilling using
predictive machining theory. Int J Mech Sci 43:1711–1730
14. Tansel I, Rodriguez O, Trujillo M, Paz E, Li W (1998) Micro-end-milling: I. Wear and breakage.
Int J Mach Tools Manuf 38:1419
15. Friedrich CR, Coane PJ, Vasile MJ (1997) Micromilling development and applications for
micro-fabrication. J Microelectron Eng 35:367
16. Bao WY, Tansel IN (2000) Modeling micro-end-milling operations. Part II: tool run-out. Int J
Mach Tools Manuf 40:2175–2192
17. Fang FZ (2002) Cutting edge effect in micromachining. Presentation to the Cutting Committee
Meeting (Part II) in 52nd CIRP General Assembly (Spain)
18. Mian AJ, Driver N, Mativenga PT (2011) Identification of factor that dominate the size effect
in micro-machining, Int J Mach Tools Manuf 51:383–394
19. Yun HT, Heo S, Lee MK, Min BK, Lee SJ (2011) Ploughing detection in micromilling process
using the cutting force signal. Int J Mach Tools Manuf 51:377–382
20. Srinivasa YV, Shunmugam MS (2013) Mechanistic model for prediction of cutting forces in
microendmilling and experimental comparison. Int J Mach Tools Manufa 67:18–27
21. Sooraj VS, Mathew J (2011) An experimental investigation on the machining characteristics
of microscale endmilling. Int J Adv Manufa Technol 56:951–958
22. Kang IS, Kim JS, Seo YW (2010) Investigation of cutting force behaviour considering the effect
of cutting edge radius in the micro-scale milling of AISI 1045 steel. Proc IMechE 225:163–171
23. Malekian M, Park SS, Jun MBG (2009) Modeling of dynamic micro-milling cutting forces.
Int J Mach Tools Manuf 49:586–598
24. Park SS, Malekian M (2009) Mechanistic modeling and accurate measurement of microend-
milling forces. CIRP Ann Manuf Technol 58:49–52
Experimental Investigation on the Influence … 581

25. Cuba Ramos A, Autenrieth H, Straub T, Deuchert M, Hoffmeister J, Schulze V (2012) Charac-
terization of transition from ploughing to cutting in micromachining and evaluation of minimum
chip thickness of cut. J Mater Process Technol 212:594–600
26. Afazov SM, Zdebski D, Ratchev SM, Segal J, Liu S (2013) Effect of micro-milling conditions
on the cutting forces and process stability. J Mater Process Technol 213:671–684
27. Prakash M, Kanthababu M (2013) In-process tool condition monitoring using acoustic emission
sensor in microendmilling. Int J Mach Sci Technol 17(2):209–227
28. Chen W, Teng X, Huo D, Wang Q (2017) An improved cutting force model for micro milling
considering machining dynamics. Int J Adv Manuf Technol 93(9–12):3005–3016
29. Sahoo P, Patra K (2019) Mechanistic modeling of cutting forces in micro-end-milling consid-
ering tool run out, minimum chip thickness and tooth overlapping effects. Mach Sci Technol
4;23(3):407–430
30. Chen W, Huo D, Teng X, Sun Y (2017) Surface generation modelling for micro end milling
considering the minimum chip thickness and tool runout. Proc CIRP 1(58):364–369
Temperature and RF Power Effect
on the Morphology and Structural
Properties of TiC Thin Film Grown
by RF Magnetron Sputtering

Olayinka Oluwatosin Abegunde, Esther Titilayo Akinlabi,

and Oluseyi Philip Oladijo

Abstract In this research study, the effect of temperature and radio frequency (RF)
power on the surface morphology and structural properties of titanium carbide (TiC)
thin film was studied. TiC thin film was deposited on the surface of commercially
pure titanium (CpTi) alloys to enhance its surface properties. During the deposition
process, the sputtering temperature and RF power were varied while other sputtering
parameters were kept constant. Field emission scanning electron microscope was
used to analyze the surface profiling of the film, and atomic force microscope was
used to study the surface roughness. Raman spectroscopy was used to determine the
film composition, and grazing incidence X-ray diffractometer was used to analyze
the phase composition. Nanoindentation was carried out to understand the hardness
and Young’s modulus properties of the coating. The results reveal the dependency
of the thin film properties on the independent process parameters.

Keywords RF power · Temperature · Thin film · TiC

1 Introduction

TiC coatings are extensively applied for modifying and enhancing the surface prop-
erties of materials. This is due to the superior quality of their properties such as high
hardness, low coefficient of friction, low wear rate, good electrical conductivity,
excellent thermal resistance and stability, high modulus, excellent biocompatibility
adhesion and chemical inertness [1–6].

O. O. Abegunde (B) · E. T. Akinlabi · O. P. Oladijo

University of Johannesburg, Johannesburg 2006, South Africa
e-mail: oabegunde@uj.ac.za
E. T. Akinlabi
Covenant University, Ota, Nigeria
O. P. Oladijo
Botswana International University of Science and Technology, Palapye, Botswana

© Springer Nature Singapore Pte Ltd. 2021 583

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_51
584 O. O. Abegunde et al.

Numerous techniques like laser deposition, electrochemical deposition, chemical

vapor deposition, and physical vapor deposition are available for making surface coat-
ing depending on the area of application and coating thickness [7–9]. Among these
several techniques, RF magnetron sputtering, a variant of physical vapor deposition,
has proven to be one the most promising coating techniques due to advantageous
factors like deposition at low temperature, dense and uniform coating, easy to con-
trol the process parameters and ability to be used on both conductive and insulating
materials [10, 11]. During RF magnetron sputtering process, several process param-
eters like sample holder configuration, sputtering power, vacuum temperature and
pressure, sputtering bias potential and current, deposition time and substrate–target
distance determine the behavior of the resulting thin film properties [12–17]. The
distribution of the temperature and RF power is two crucial parameters focused on
in this research work. Distribution of the temperature within the vacuum chamber
is important for the growth of thin film and affects the coalescence and deposition
rate. Thinner film tends to form a region with high-temperature profile compared to
region with low-temperature distribution [18, 19]. RF magnetron power is another
vital process parameter in growing thin films. RF power determines the intensity of
the plasma formation, which influences the bombardment of the target and coales-
cence of the adatoms on the surface of the substrate [13, 20]. Therefore, optimizing
and controlling the interaction between these process parameters can help to achieve
film of desired and specific properties.
In this research study, the effect of the RF power and temperature was studied
on the evolving properties (microstructure, roughness, hardness, Young’s modulus,
structure, and Raman) of TiC thin film deposited on commercially pure titanium
substrate.

2 Experimental Setup

2.1 RF Magnetron Sputtering

TiC thin films were grown on commercially pure titanium (CpTi) using HHV TF500
versatile RF magnetron sputtering coater. Before deposition of the film, the substrate
(CpTi) was ground using silica carbide paper and then polished to remove any surface
impurities. Further, cleansing of the surface was done by using acetone to remove
any form of lubricant or contaminant on the surface. The purity of the TiC target used
is 99.99%. During the sputtering process, the RF power and temperature were varied
while other sputtering parameters were kept constant. Three runs of experiments
were conducted, and the experimental matrix is shown in Table 1. The sputtering
apparatus was first evacuated to a base pressure of 1.13 × 10−5 mbar. The reason for
the low pressure is to provide a long mean free path for a collision between the target
and substrate and allows for the control and minimization of contaminants. After
evacuation, the system is refilled with argon to a partial pressure of 2.5 × 10−3 mbar.
Temperature and RF Power Effect … 585

Table 1 Experimental matrix of the process parameters for thin film deposition
S/N RF Power (W) Temperature (°C)
A 150 80
B 200 90
C 250 100

The flow rate of argon was kept constant at 12 sccm while the sputtering time was
maintained at 2.5 h. The working pressure was constant throughout the experiments.
Maintaining constant pressure helps to place many argon atoms in the path of the
ions and ejected coatings.

2.2 Characterization Techniques

A veeco Di2100 atomic force microscope was used to evaluate the surface topogra-
phies of the samples. Image scan size at 5 × 5 µm2 was obtained in tapping mode.
All analysis was performed in ambient temperature. Nanoscope software was used
for capturing and analyzing the images from the surface of the samples. Field emis-
sion scanning electron microscope (ZEISS Gemini*2, Germany) (FESEM) capable
of capturing nanoscale images effortlessly at very high magnification was used to
observe the surface morphology evolution. The FESEM can take images at very
high magnification and images were taken at 50,000× magnification. Raman anal-
yses were performed on TiC thin films using an alpha300R (WITec) confocal laser
Raman microscopy coupled with frequency-doubled Nd-YAG laser of 532 nm wave-
length. A magnification of ×50 on Nikon objective lens was used in capturing the
Raman spectra intensity. A laser power of 2 mW at room temperature was used to
prevent the burning of the film surface. Before the acquisition of the Raman spectra,
beam centering and Raman spectra calibration were done using standard Si(111)
samples. The Raman spectrum of the substrate was obtained and used to compare
with the TiC thin films deposited. Grazing incidence X-ray diffractometer (PAN-
alyticals’s Xpert Pro with Cu K-alpha and wavelength 1.540598 A) at a very low
angle of incidence that allows penetration into the thin film was used to study the
structural properties of the thin film. Young’s modulus and hardness thin film of the
nanomechanical properties were obtained using Hysitron TI 950 Triboindenter. The
control mode used was load controlled, and the load pattern was modeled to follow a
trapezoidal geometry of loading dwelling and unloading profile with dwelling time
of 15 s at peak load.
Berkovich tip was used as the diamond indenter with a tip radius of curvature
of 100 nm. From the analyzed load-displacement curves, Oliver and Pharr analysis
method was used to calculate Young’s modulus of the thin film [21–23]. The depth of
penetration for all the samples was limited to 10% of the coating thickness to eliminate
the effect of the substrate on the thin film properties [24, 25]. The hardness and
586 O. O. Abegunde et al.

Young’s modulus reported corresponding to average of 10 measurements performed

on the thin film coating.

3 Result and Discussion

Figure 1 depicts the microstructural evolution of the surface morphology of the coat-
ing at three different process parameters. Three different distinct features and growth
modes are noticeable at each process parameter. For sample A, the microstructural
growth pattern is similar to island growth which shows an early stage of film growth
and nucleation on the surface and the adherence bond between the adatoms is stronger
than the bonding between the adatoms and substrate. The coatings are sparingly dis-
tributed on the surface with granular grain shape. The coating becomes dense and
homogenous distributed as the power and temperature increased. This might be asso-
ciated with the fact that at sufficient temperature and power, the rate of nucleation
and growth of the film is rapid. In sample B, the images reveal a columnar structure
as expected. It is of no doubt that these columnar structures usually associated with
ceramic coating Amina et al. [2]. For sample C, two layers of coating are noticeable.
The topmost layer is coarse with high porosity while the latter layer coating is evenly
distributed. This implies that further increased in RF power and temperature leads to
thicker coating surface and formation of a new layer with the nucleation mechanism
pattern of the Stranski–Krastanov mechanism.
The 3D morphology of the surface roughness is shown in Fig. 2, and the statistical

Fig. 1 FESEM images taken at a 50,000× magnification of the surface morphology

Fig. 2 3D images from the AFM surface profiling at scan size of 5 × 5 µm2
Temperature and RF Power Effect … 587

Table 2 Statistical details of the AFM results

Experimental run Z-range (nm) RMS (nm) Mean roughness Skewness Kurtosis
(nm)
A 397.25 51.93 39.68 0.263 3.898
B 268.35 30.58 20.48 1.646 8.316
C 440.66 53.467 40.330 0.451 4.223

details are summarized in Table 2. The 3D AFM structure for sample 1 exhibits
extended columnar grain growth which agrees with the result of the FESEM images.
The AFM morphology of samples 2 shows evenly distributed coating, and for sample
3, the surface becomes uneven.
The surface roughness reduced from 39.68 nm for sample A to 20.48 nm in sample
B. This phenomenon might due to the homogenous morphology of sample B. The
roughness increases for sample C which has the highest roughness and most uneven
surface due to incomplete formed multilayer surface. The hardness and Young’s
modulus are presented in Table 3. Similarly, the graph is shown in Fig. 3. The highest
hardness value is associated with sample B that exhibits uniform coating of the film.
Sample C has the least hardness value. Although for thin film, the hardness is not

Table 3 Hardness and Young’s modulus results from nanoindentation

Experimental Hardness (GPa) Young’s Wear resistance Plasticity Index
run Modulus (GPa) H 3 /E 2 (GPa) H/E
A 5.00 ± 2.04 134.19 ± 21.39 0.007 0.037
B 6.08 ± 2.90 139.03 ± 25.67 0.012 0.044
C 4.09 ± 1.99 113.21 ± 34.44 0.005 0.036

Fig. 3 Force versus

displacement curve of the
nanoindentation plot
588 O. O. Abegunde et al.

Fig. 4 XRD spectra of TiC

thin film coating on CpTi

enough to conclude the resistance of the surface to deformation or failure, the wear
resistance which is the ratio of H3 /E2 and plasticity index (H/E) is used to study the
film resistance. The wear resistance and plasticity index were highest at B. The wear
resistance and plasticity increase with an increase in RF power and temperature and
then decrease as the RF power and temperature were raised from 200 W and 90 °C
to 250 W and 100 °C, respectively.
The crystallographic structures from the XRD are shown in Fig. 4. The patterns
of the TiC peaks obtained exhibit a preferential orientation of 110, and other planes
found are 200, 220, and 311. TiC films have a rock salt-type (NaCl type) face-centered
cubic (FCC) with diffraction peaks intensity at 110, 200, 220, and 311 plane reflection
[26]. This agrees with the result, and different peak intensities were noticed due to
the RF power and temperature of the coating. The highest peak intensity was found
at 110 for all the samples. 110 spectra are known to result in high hardness value [2].
Sample A produces the highest peak value at 110 dues to the geometrical factor on
the slip system [14].
Raman spectroscopy is responsible for determining the crystal structure quality,
and the response graph is shown in Fig. 5. Both the D and G bands for TiC are
noticeable on the spectra graph. The D band for TiC was found around 500–700
(1/cm). Sample B tends to have smaller Raman intensity peaks compared to A and C.
Raman peaks are caused by lattice mismatch, chemical composition, defects, short-
range order, crystalline structure, and internal stresses. Since sample B possesses
better homogenous and uniformly distributed coating, it can be concluded that the
lattice is better matched compared to A and C with less defect.
Temperature and RF Power Effect … 589

Fig. 5 Raman spectra of

TiC thin film coating on
CpTi substrate

4 Conclusion

The observation from the characterization shows how RF power and temperature
contribute to the properties of TiC thin film coating. Three distinct growth types
were noticed depending on the process parameters, and the lowest surface roughness
was noticed at the film with the dense and uniformly distributed coating. The XRD
revealed the preferential plane of 111 has the highest peak, and the hardness properties
show substantial dependence on the process parameters.

Acknowledgements The authors will like to acknowledge CSIR South African and the University
of Johannesburg for funding this project.

References

1. Kulkarni M et al (2014) Biomaterial surface modification of titanium and titanium alloys for
medical applications. Nanomedicine 111:111
2. Djafer ZA et al (2014) Deposition and characterization of titanium carbide thin films by
magnetron sputtering using Ti and TiC targets. Appl Surf Sci 312:57–62
3. Longo G et al (2010) Effect of titanium carbide coating by ion plating plasma-assisted depo-
sition on osteoblast response: a chemical, morphological and gene expression investigation.
Surf Coat Technol 204(16–17):2605–2612
4. Brama M et al (2007) Effect of titanium carbide coating on the osseointegration response
in vitro and in vivo. Biomaterials 28(4):595–608
5. Galevsky G et al (2015) Titanium carbide: nanotechnology, properties, application. In: IOP
conference series: materials science and engineering, p 012017
6. Bugaev SP et al (2003) Amorphous hydrogenated carbon films deposited by a closed-drift ion
source. Laser Part Beams 21(2):285–289
7. Abegunde OO et al (2019) Overview of thin film deposition techniques. AIMS Mater Sci
(2):174
590 O. O. Abegunde et al.

8. Zhang S et al (2003) Recent advances of superhard nanocomposite coatings: a review. Surf

Coat Technol 167(2–3):113–119
9. Guo HC et al (2017) Recent progress of atomic layer deposition on polymeric materials. Mater
Sci Eng C 70:1182–1191. https://0-doi-org.ujlink.uj.ac.za/10.1016/j.msec.2016.01.093
10. Alexeeva OK (2016) Application of the magnetron sputtering for nanostructured electrocata-
lysts synthesis. Int J Hydrogen Energy 41(5):3373–3386
11. Wu F et al (2016) Characteristics of the structure and properties of ZnSnO3 films by varying the
magnetron sputtering parameters. Acta Metallurgica Sinica (English Letters) 29(9):827–833
12. Toku H et al (2010) Influence of process parameters on the growth of pure-phase anatase and
rutile TiO2 thin films deposited by low temperature reactive magnetron sputtering. Braz J Phys
40(3):340–343
13. Ghoranneviss ZG (2016) Effects of various deposition times and RF powers on CdTe thin film
growth using magnetron sputtering. J Theor Appl Phys 10(3):225–231
14. Zhang X et al (2013) Effects of substrate temperature on the growth orientation and optical
properties of ZnO: Fe films synthesized via magnetron sputtering. J Alloys Compd 574:149–154
15. Fancey K (1995) A coating thickness uniformity model for physical vapour deposition systems:
overview. Surf Coat Technol 71(1):16–29
16. Kim K, Kim HW, Lee CM (2003) Effect of growth temperature on ZnO thin film deposited on
SiO2 substrate. Mater Sci Eng B 98(2):135–139
17. Gennardo D (2010) Design, construction, and optimization of a magnetron sputtering system
for urania deposition
18. Dong H et al (2006) Study and analysis on the film uniformity. Piezoelectr Acoustoopt 28:578–
580
19. Tian M, Liu D (1991) Handbook of thin film science and technology
20. Li FWJ, Qi Y, Guo WDY (2016) Characteristics of the structure and properties of ZnSnO3
films by varying the magnetron sputtering parameters. 29(9):827–833
21. Oliver WC, Pharr GM (1992) An improved technique for determining hardness and elastic
modulus using load and displacement sensing indentation experiments. J Mater Res 7(6):1564–
1583
22. Oliver WC, Pharr GM (2004) Measurement of hardness and elastic modulus by instrumented
indentation: advances in understanding and refinements to methodology. J Mater Res 19(1):3–
20
23. Fang T, Jian S, Chuu D (2004) Nanomechanical properties of TiC, TiN and TiCN thin films
using scanning probe microscopy and nanoindentation. Appl Surf Sci 228(1–4):365–372
24. Bec S et al (1996) Improvements in the indentation method with a surface force apparatus.
Philos Mag A 74(5):1061–1072
25. Mani et al (2005) Effects of residual stress on the mechanical and structural properties of TiC
thin films grown by RF sputtering. Surf Coatings Technol 194(2–3):190–195
26. Chun S (2010) Bias voltage effect on the properties of TiN films by reactive magnetron
sputtering. J Korean Phys. Soc 56(4):1134–1139
Effect of Temperature on the Surface
Characteristics of Anodized Aluminium
Tubes

A. L. Sriram Sudhan and A. Brusly Solomon

Abstract Thin porous coating was found to be vital for boiling heat transfer
enhancement in various heat transfer devices that use a phase change mechanism.
However, the formation of a stable thin porous surface is a challenging process. It is
known that the aluminium tubes were used to fabricate different kinds of heat pipes
which are used in space applications. Therefore, developing nanoporous coating in
an aluminium tube is essential to improve the performance of heat pipes. Anodiza-
tion is an electrochemical process that applied mostly for aluminium materials to
protect the surface from corrosion. This anodized aluminium surface can also be
utilized for heat transfer applications by controlling the morphology of the surface
through adjusting the process parameters of the anodization. One of the most critical
parameter that controls the morphology of coating surface is the temperature of the
anodizing cell. Therefore, in this study, the effect of cell temperature on the coating
thickness, pore size and contact angle is analysed while maintaining the electrolyte
concentration, cell voltage, electrolyte flow rate and anodizing time as constant. An
aluminium tube of the outer and inner diameter of 19 mm and 16.8 mm, respectively,
with a length of 350 mm was used for anodization. The anodized aluminium tubes
were characterized by analysing the surface morphology, coating thickness, pore size
and contact angle. It was found that the formation of micro/nanoporous structure in
the inner layer of aluminium tube with the pore size ranging from 2.05 µm to 3.6 µm
and contact angle less than 10° between the temperature range of 20–35 °C when
10% H2 SO4 solution is used as the electrolyte.

Keywords Anodization · Temperature · Nanoporous surface · Coating thickness ·

Capillary rise · Heat transfer

A. L. Sriram Sudhan · A. Brusly Solomon (B)

Micro and Nano Heat Transfer Lab, Centre for Research in Material Science and Thermal
Management, Department of Mechanical Engineering, Karunya Institute of Technology and
Sciences, Coimbatore, India
e-mail: abruslysolomon@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 591

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_52
592 A. L. Sriram Sudhan and A. Brusly Solomon

1 Introduction

The capillary rise is a critical parameter for an efficient operation of heat transfer
device such as heat pipe. To enhance the capillary rise, anodization (electrochemical
process) was performed on the inner surface of the aluminium tube. The anodization
process prevents corrosion and also increases the number of nucleation sites by
forming micro/nanopores [1, 2]. By controlling the anodizing parameters, surface
morphology can be modified for efficient heat transfer applications [3, 4]. Singh
et al. [5] studied the thermal performance of anodized flat thermosyphon with a heat
input range of 50–300 W, various fill ratio (40%, 60% and 100%) and different
inclination angles (0°, 45°, 90°). It was found that the thermal resistance was reduced
by more than 20% in the anodized surface when compared to the non-anodized
surface. Also, it was noticed that the pore density was increased up to 90% compared
to the non-anodized surface due to the anodization. Solomon et al. [6] investigated
the thermal performance of anodized two-phase thermosyphon by varying the heat
input in the range of 50–300 W with the combination of R600a and R290 as working
fluid. It was found that the reduction in thermal resistance of 17% to 23% while
varying the inclination angle from 0° to 90°. Further, the heat transfer coefficient
was increased up to 33% for anodized thermosyphon. Solomon et al. [7] studied
the thermal performance of anodized grooved heat pipe with various parameters
such as heat input in the range of 25–250 W, fill ratio of 10–70% and inclination
angle of 0°–90° using R600a as working fluid. It was found that the formation of
micro/nanostructures enhances the heat transfer performance by 39% while reducing
the thermal resistance of 30% compared to the non-anodized surface. Due to the
anodization process, the generation of nucleation site is more; thus, it makes the
surface more wettable and leads to capillary rise and heat transfer. Weng et al. [8]
studied the transfer performance of the anodized heat pipe by developing anodic
aluminium oxide nanotubes in the inner surface of the heat pipe. It was found that heat
transfer was enhanced by 4.12% and the reduction in thermal resistance was 58.68%
by anodization. From the above literature studies, it is evident that the anodized
coating formed in the inner surface of the heat pipe enhances the thermal performance
of the heat pipe. This paper aimed at modifying the surface morphology such as
pore size, coating thickness and surface wettability, which are key for heat transfer
enhancement. Also, the effect of temperature on the pore size, coating thickness and
contact angle of anodized surface is studied and discussed.

2 Experimental Analysis

2.1 Anodization of Aluminium Tube

The anodizing facility consists of three systems, namely the electrolyte flow system,
chilling unit and DC power supply. The aluminium tube of OD 19 mm, ID 16.8 mm
Effect of Temperature on the Surface … 593

and length of 350 mm is used as an anode, and stainless-steel rod of 10 mm OD

is used as a cathode. The cathode is placed inside the anode, and it is supported
by Teflon sleeves at both ends, as shown in Fig. 1. Initially, etching is done by
circulating NaOH (Merck) solution of 5% for about 1 min and then washed with
de-ionized water. Secondly, desmutting is done by circulating 70% of nitric acid
solution (Merck) for 2 min, and the excess material is removed from the surface
of aluminium and then cleaning is done with DI water. Finally, the aluminium tube
is cleaned by passing de-ionized water for 5 min. After the cleaning process, the
electrolyte solution is prepared with 10vol.% H2 SO4 solution (Merck) and another
solution with 7% H2 SO4 + 3% oxalic acid. For anodization, initially 1 V was set,
and then the voltage was raised 1 V per minute up to 15 V, and then the voltage
was kept constant. The anodization is performed at different temperatures ranging
from 20 to 35 °C with H2 SO4 solution and with a temperature range from 10 to
30 °C for a hybrid solution having 7% H2 SO4 and 3% oxalic acid solution. More
detailed procedure for anodizing the inner tube of aluminium tube is presented in
our previous study [4]. After the anodization process, the anodized aluminium tubes
were characterized using SEM micrograph, and from the SEM image, the pore size
was calculated by using a reference scale provided in the same image.

Fig. 1 Anodization cell set-up

594 A. L. Sriram Sudhan and A. Brusly Solomon

3 Results and Discussions

3.1 Anodization of Aluminium Tube

Variations of anodizing current with respect to the anodizing time at different tem-
peratures while using H2 SO4 as electrolyte solution during the anodization process is
shown in Fig. 2. During the voltage increment, a thick anodized layer is formed, and
when the voltage is maintained at 15 V, the pore formation starts. After the pore
formation started, the current consumption increases slightly, as shown in Fig. 2.
Therefore, as the temperature is maintained, the increase in current leads to more
prominent pore formation. Also it is noticed that the pore size is increasing with the
increase in temperature of the anodizing cell. Therefore, it was decided to decrease
the temperature further to minimize the pore size. Moreover, the literature suggest-
ing that adding oxalic acid reduces the pore size formation. Therefore, oxalic acid
of 3% is combined with the 7% H2 SO4 solution is considered while decreasing the
temperature. The variation of current at different temperature with respect to the
anodizing time using 7% H2 SO4 + 3% oxalic acid as electrolyte solution is shown
in Fig. 3. It was observed that the current increases with the increase in time for all
anodizing temperatures as similar to H2 SO4 as an electrolyte solution. The thick-
ness of the coating for this combination of electrolyte is skinny, and it is around 3
to 5 µm. Though the thickness of the anodized layer is thin, there were no pores
found on the anodized surface. From these experiments, we can conclude that using
H2 SO4 solution as an electrolyte at an anodizing temperature of 20 °C, the pore size
(2.053 µm) and coating thickness (6.446 µm) are found to be less which is suitable
for enhanced heat transfer applications.

9
8
7
Current (Amps)

6
5 Temp at 35°C
4 Temp at 29°C
3 Temp at 25°C
2 Temp at 20°C
1
0
0 20 40 60 80
Time (Min)

Fig. 2 Variation of current with respect to the anodizing time at different temperatures using H2 SO4
as an electrolyte
Effect of Temperature on the Surface … 595

2.5
Current (Amps)
2
Temp at 10°C
1.5 Temp at 15°C
Temp at 20°C
1
Temp at 25°C
0.5 Temp at 29°C

0
0 20 40 60 80
Time (Min)

Fig. 3 Variation of current with respect to anodizing time at different temperature using 3% oxalic
acid + 7% H2 SO4 as electrolyte

3.2 Characterization of Anodized Aluminium Tube

In our previous study [4, 7], anodization results for the electrolyte flow rate of
380 ml/min, the anodizing voltage of 1 to 15 V and the concentration of 10 vol%
of H2 SO4 were discussed. In the present study, the surface morphology of anodized
aluminium tube using H2 SO4 as electrolyte solution with an increase in temperature
is shown in Fig. 4a–d and Table 1. It was found that the pore size is increasing with
the increase in anodizing temperature. The pore size is found to increase suddenly
from 2 to 3.6 µm when the temperature is increased from 29 to 35 °C. The maxi-
mum experimental temperature was limited at 35 °C due to the formation of larger
pore size, which is against our requirement of smaller pore size. To find the thick-
ness of coating formed at the surface of the inner layer, SEM photograph for 10%
of H2 SO4 as electrolyte solution is shown in Fig. 4 (e, f, g and h). A sponge-like
substance was formed in the inner layer of the aluminium tube, which represents the
coating thickness. This sudden decrease in coating thickness may happen due to the
dissolution of the aluminium oxide layer in the inner layer of the tube. A similar
dissolution pattern is common in anodization practices [9]. This dissolution occurs
with consuming more current of 9 A while forming coating at the inner side of the
aluminium tube and if it is continuing the dissolution may take place. With these
above discussions, anodization concerning increasing in temperature may result in
more dissolution. The anodizing temperature of 20 °C with a minimum pore size
of 2.053 µm and minimum coating thickness of 6.44 µm is found to be optimum,
which will be suitable for enhanced heat transfer applications.
Surface morphology of anodized aluminium tube while using 7% H2 SO4 + 3%
oxalic acid as electrolyte solution is given in Fig. 5a–e. It was observed that the
596 A. L. Sriram Sudhan and A. Brusly Solomon

Fig. 4 SEM image of anodized aluminium tube a at 20 ° C b at 25 °C c at 30 °C d at 35 °C, coating

thickness of anodized aluminium tube e at 20 °C f at 25 °C g at 30 °C h at 35 °C

pore size is 4.54 nm for the anodizing temperature of 29 °C. When the anodization
temperature decreases from 29 to 10 °C, there was no pore size noticed. This combi-
nation of 7% H2 SO4 + 3% oxalic acid electrolyte consumes less current (about two
amps) for anodizing compared to 10% H2 SO4 electrolyte. Due to less consumption
of current, the coating layer formed on the inner surface of the aluminium tube is
found to be very thin with cracks and breaks. Though the coating layer is very thin,
it is necessary to estimate the coating thickness from the SEM photographs, and it is
shown in Fig. 5 f–j. With the decrease in anodizing temperature, the coating thickness
also decreases up to a temperature of 15 °C and then starts increasing for 10 °C and
it is around 4.7 µm. As a result, this combination of 7% H2 SO4 + 3% oxalic acid of
electrolyte will not be suitable for anodization since there was no pore formation.
To find the wettability of the anodized surface, the contact angle of the surface was
measured using Drop Shape Analyzer (DSA 25, Kruss GmBH) with a needle diam-
eter of 0.652 mm. The contact angle for the tube anodized using 7% H2 SO4 + 3%
oxalic acid as electrolyte solution is given in Fig. 6a–e. With the increase in tem-
perature, the contact angle was found to be decreasing, and it is in the range of
98.4–135.6°. Further, it was found that the minimum contact angle obtained at a
Table 1 Anodization conditions
Experiment Electrolyte Electrolyte Anodizing Final cell Acid Anodizing Average Average Average
number flow rate Solution Temperature voltage concentration time coating pore size contact
(ml min−1 ) (°C) (Vfinal ) (V/V %) (minutes) thickness (µm) angle (θ)
(µm)
Uncoated – – – – – – – – 110
1 380 H2 SO4 20 15 10 60 6.446 2.053 <10
2 380 H2 SO4 25 15 10 60 7.465 2.167 <10
3 380 H2 SO4 29 15 10 60 10.97 2.333 <10
Effect of Temperature on the Surface …

4 380 H2 SO4 35 15 10 60 4.84 3.61 <10

5 380 3% oxalic 29 °C 15 10 60 5.425 0.00453 98.4
acid + 7%
H2 SO4
6 380 3% oxalic 25 15 10 60 4.038 No pore 104.8
acid + 7% size sited
H2 SO4
7 80 3% oxalic 20 15 10 60 3.851 No pore 110.5
acid + 7% size sited
H2 SO4
8 380 3% oxalic 15 15 10 60 3.604 No pore 113.8
acid + 7% size sited
H2 SO4
9 380 3% oxalic 10 15 10 60 47.032 No pore 135.5
acid + 7% size sited
H2 SO4
597
598 A. L. Sriram Sudhan and A. Brusly Solomon

Fig. 5 SEM image of anodized aluminium tube a at 10 °C b at 15 °C c at 20 °C d at 25 °C e at

29 °C, Coating thickness of anodized aluminium tube f at 10 °C g at 15 °C h at 20 °C i at 25 °C
j at 29 °C

higher temperature is 29 °C. The minimum contact angle implies that the anodized
surface is hydrophilic, where the solid–liquid interaction is more, which leads the
surface to produce capillary rise. The contact angle for the inner surface of aluminium
tube coated using 10% of H2 SO4 as an electrolyte solution is much less than 10°,
which cannot be measured as the liquid drop is spreading quickly.
Effect of Temperature on the Surface … 599

Fig. 6 Contact angle measurement of anodized aluminium tube a at 10 °C b at 15 °C c at 20 °C

d at 25 °C e at 29 °C

4 Conclusion

The effects of anodization on the aluminium tube with varying temperature for two
different electrolytes were analysed and discussed. From this study, the following
conclusions arrived.
• The pore size and coating thickness were found to be increasing with an increase
in temperature for anodization using 10% of H2 SO4 as an electrolyte solution
• When the cell temperature was varied from 10 to 30 °C, there were no pores found
in the anodization process with 7% H2 SO4 + 3% oxalic acid as an electrolyte
solution
• H2 SO4 of 10% electrolyte solution with an anodizing temperature of 20 °C is found
to be the optimum conditions due to the minimum the pore size and minimum
coating thickness.
It was noticed that the anodized surface with the above conditions seems to be
more hydrophilic, which will be suitable for enhanced heat transfer applications.
600 A. L. Sriram Sudhan and A. Brusly Solomon

Acknowledgements The authors would like to thank DST–SERB

(DST/SERB/YSS/2015/001084) for the financial support for this research work.

References

1. Lee W, Kim J, Splitting PW, Template AO (2009) Properties of nanostructures obtained by

anodization of aluminum in phosphoric acid at moderate potentials properties of nanostructures
obtained by anodization of aluminum in phosphoric acid at moderate potentials
2. Thamida SK, Chang H, Thamida SK, Chang H (2002) Nanoscale pore formation dynamics
during aluminum anodization nanoscale pore formation dynamics during aluminum anodization,
vol 240
3. Renjith Singh R, Selladurai V, Solomon AB, Balakrishnan T (2015) Effect of nucleation’s site on
the performance of anodized thermosyphon. In: Proceedings of the 23rd National Heat and Mass
Transfer Conference and 1st International ISHMT-ASTFE Heat and Mass Transfer Conference,
IHMTC2015, 17–20 December, 2015, Thiruvananthapuram, India
4. Solomon AB, Noel M (2015) IHMTC2015-30 anodization and evaluation of an aluminium
thermosyphon with, pp 1–9
5. Renjith Singh R, Selladurai V, Ponkarthik PK, Solomon AB (2015) Effect of anodization on the
heat transfer performance of flat thermosyphon. Exp Therm Fluid Sci 68:574–581
6. Solomon AB, Roshan R, Vincent W, Karthikeyan VK, Asirvatham LG (2015) Heat transfer
performance of an anodized two-phase closed thermosyphon with refrigerant as working fluid.
Int J Heat Mass Transf 82:521–529
7. Solomon AB et al (2017) Characterisation of a grooved heat pipe with an anodised surface. Heat
Mass Transf und Stoffuebertragung 53(3):753–763
8. Weng H, Yang M-H (2018) Heat transfer performance enhancement of gravity heat pipes by
growing AAO nanotubes on inner wall surface. Inventions 3(3):42
9. Abd-elnaiem AM, Mebed AM, Gaber A (2014) Effect of the anodization parameters on the
volume expansion of anodized aluminum films, vol 8, no 2013, pp 10515–10525
Experimental Investigation on Heat
Pipe-Assisted Cooling During Milling
Process of AISI 1040
I. Kantharaj, D. J. Hiran Gabriel, Julius Benedict Prakash,
and S. Mohanasundaram

Abstract The cutting temperatures during milling operation are high owing to the
hardness of work specimens and due to friction at the tool–work and tool–chip inter-
faces. This elevated temperature results in the deformation of the cutting edges during
dry milling. The most common method employed to reduce the heat that is generated
is the application of cutting fluids in the cutting zone. The tool life during milling is
prominently enhanced by providing a higher quantity of cutting fluid at the cutting
zone. Though the cutting fluid effectively removes heat from the tool, its sudden
application causes thermal shocks in the tool. This thermal shock makes the tool
brittle specifically at the cutting edges resulting in tool wear. Also, coolant appli-
cation has a negative impact on the environment as well as the machine operator’s
health. Moreover, tool cost is higher than the cost of safekeeping and clearing of cut-
ting fluids. Machining with heat pipe is a newly established technique to lower the
problem involved with cutting fluid. Thus, in this experiment, a customized milling
tool is fabricated such that it can house a ring-shaped heat pipe. The experiments
were conducted in conditions of dry milling with no coolant and heat pipe-assisted
cooling. The results obtained in the experiments confirm that milling with heat pipe-
assisted cooling exhibited reduced cutting force and cutting temperature than dry
machining. Thus, heat pipe-assisted cooling can be suggested as an alternative for
conventional methods.

Keywords Dry milling · Cutting fluid · Heat pipe

I. Kantharaj (B) · J. B. Prakash · S. Mohanasundaram

Department of Mechanical Engineering, Karunya Institute of Technology and Sciences,
Coimbatore, Tamil Nadu 641114, India
e-mail: ikantharaj@gmail.com
D. J. Hiran Gabriel
Department of Mechanical Engineering, PSG College of Technology, Coimbatore, Tamil Nadu
641004, India

© Springer Nature Singapore Pte Ltd. 2021 601

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_53
602 I. Kantharaj et al.

1 Introduction

Metal machining plays a significant part in all manufacturing industries. Machining

industry has incorporated continuous changes in technology and processes to increase
the effectiveness of manufacturing [1]. Metal machining is a cutting process that
removes the unwanted material from the work material to produce the desired shape
and size [2]. Machining includes turning, drilling, boring, milling processes with
single or multi-point cutting tools to remove excess material from the work to attain
a required geometry during the metal removal operation [3]. Temperature rise and
friction at the tool–work interface generate difficulties concerning the quality of
finished surface and tool life [4]. According to several authors, the contact between
workpiece and tool is the source friction produced during the metal removal process.
The tool wear is due to thermal stresses produced during metal cutting and affects
the surface finish [5]. During metal removal, heat energy produced at the tool–chip
interface and it is carried away by coolants. To overcome these issues, cutting fluids
can be used. Cutting fluid is a lubricant as well as a coolant for metal machining
processes. A typical cutting fluid contains oil, oil with water emulsion, gel, aerosols
or mist with air. It is widely used to increase the level of productivity and surface
quality of the specimen during machining [6]. In machining, the cutting fluid, along
with cutting parameters like depth of cut, cutting speed and feed rate, plays a vital
role. Though the principal role of the cutting fluids is to cool the primary cutting
zone and lubricate the interfaces, it also reduces the consumption of power, prevents
corrosion on the surface of the machined component and also prolongs the tool life
[7]. The global demand for lubricant was high at 39.4 million tons a year in 2015. It is
predicted to be in the range of 43.9 million tons in 2022. Most of the lubricants used
are engine oils, hydraulic lubricants and gear oils. A metal cutting fluid represents 5%
of the global market of lubrication. Though the cutting fluid has several advantages,
it also suffers from certain shortcomings such as cost, environmental issues, operator
safety and health [8]. Around 17% of the total cost of machining is ascribed to use of
coolants in the machining industry [9]. In recent years, alternatives for cutting fluids
have been established to overcome the downsides of cutting fluids. The alternatives
include cryogenic cooling, dry machining, minimum quantity lubrication (MQL),
solid lubrication, gaseous cooling and nanofluids [10]. Dry machining and MQL are
being widely used among these alternatives. The tool wear was found to be higher at
elevated temperatures of dry machining [11]. During dry machining, achieving the
required geometrical accuracy was difficult due to the above-mentioned shortcomings
[12]. MQL applied in the form of mist during machining, can be inhaled by the
metalworking operator, which can be carcinogenic [13]. Embedding the heat pipe
with the tool during machining is the better solution to avoid above-said problems and
to remove the heat produced in the primary cutting zone without applying coolant
[14–16]. In order to evaluate the efficiency of heat pipe-assisted milling over dry
milling, experiments were conducted and the results are discussed below.
Experimental Investigation on Heat Pipe-Assisted Cooling … 603

2 Materials and Methods

2.1 Investigation Parameters

Milling is the most commonly employed metal removal process in the industry. The
main advantage of the process is that it can create a wide variety of features on
the component by removing the material from the component. In milling, a large
amount of heat is produced at cutting zone due to the shearing of the workpiece by
the milling cutter. In dry machining, the elevated temperature results in high tool wear
and surface finish of the workpiece being poor. Thus, to remove the heat generated,
a large amount of cutting fluid is employed in the cutting zone. Though cutting fluid
aids in better surface finish and increased tool life, it adds up to the indirect cost of
the product. Also, cutting fluid makes the work floor unclean and hazardous to the
operator. Thus, it is suggested to implement heat pipe in the process for the effective
transfer of heat from the tool.
The main objective of the experimental work is to study dry machining and heat
pipe-assisted cooling in the milling process. The parameters considered for the com-
parison are temperature, cutting force and surface roughness. For a proper investi-
gation, a nine-level experimental run was designed with variation in metal cutting
parameters such as spindle speed, feed and depth of cut. The preset values of the
nine-level experimental are given in Table 1.

2.2 Tool and Heat Pipe Fabrication

For the experimental purpose, a new tool was designed and fabricated. The material
used for the fabrication of the tool is HSS. To embed the ring-shaped heat pipe, a
housing position is made using EDM. The metal removal process using EDM was

Table 1 Parameters used in

Run Speed (m/min) Feed (mm/rev) Depth of cut
the investigation for each run
(mm)
of the experiment
S1 250 16 0.25
S2 250 20 0.5
S3 250 25 0.75
S4 500 20 0.25
S5 500 25 0.5
S6 500 16 0.75
S7 710 25 0.25
S8 710 16 0.5
S9 710 20 0.75
604 I. Kantharaj et al.

Cavity for
Heat pipe

Fig. 1 Fabricated milling cutter with cavity

Table 2 Dimensions of the

Parameter Dimension (mm)
heat pipe
Outer diameter 70
Inner diameter 54
Wall thickness 0.5
Height 35

carried out in a way that does not weaken the tool as well as enabling the heat
pipe to reach the maximum possible distance near to the cutting zone. A pre-shaped
electrode in the shape of the required cavity is used in the process. The fabricated
tool is shown in Fig. 1.
The heat pipe was made using thin copper plates of 0.5 mm thickness. Special
care was taken in order to maintain the ring shape since any deformation in the
shape would lead to mismatch with the housing slot on the tool. Distilled water is
the working fluid used in the heat pipe. The heat pipe filling is done using a vacuum
pump and is sealed properly. The dimensions of the heat pipe are given in Table 2.
The fabricated heat pipe shown in Fig. 2 is fitted in the housing cavity made on
the tool. The fit was made in such a way that the heat pipe does not come off the tool.
The assembly of milling tool with the heat pipe is shown in Fig. 3.

2.3 Experimental Set-up

AISI 1040 steel workpiece was used to carry out experiments with heat pipe-assisted
cooling method. The parameters measured during the experiments were temperature,
surface roughness and cutting force. The cutting force during milling was measured
using a Kistler 9257B dynamometer. Mahr TR100 surface roughness tester of type
Mahr Surf GD 25 was used in measuring the surface roughness of the workpiece.
Experimental Investigation on Heat Pipe-Assisted Cooling … 605

Fig. 2 Fabricated heat pipe

Fig. 3 Milling cutter

embedded with heat pipe

Among the various parameters measured using the surface roughness tester, rough-
ness average (Ra) is considered in the study. The temperature during the milling
operation is measured using Amprobe IR-750 an infrared thermometer.
The fabricated heat pipe is housed on the custom-made milling tool. A heat pipe is
a tool that transfers heat using the ideas of phase transition and thermal conductivity
to transfer heat efficiently. A heat pipe is always fabricated using a high thermally
conductive material like copper. The part of the heat pipe which is exposed to the
higher temperature is the hot interface and part that is exposed to the atmosphere is
the cold interface. At hot interfaces, the liquid which is in contact with the highly
conductive material is converted to vapour. This vapour reaches the cold interface of
the heat pipe where the temperature is low. The vapour at the cold interface due to the
low temperature condenses to become liquid again giving away the latent heat to the
atmosphere. This condensed liquid is then returned to the hot interface by capillary
action, centrifugal force, or gravity and the cycle restarts.
606 I. Kantharaj et al.

Infrared thermometer

Milling Cutter

Work Piece

Vice connected to
Dynamometer

Fig. 4 Experimental set-up

The set-up of the experiment consists of a milling machine to which an arbour

is fixed. The tool is mounted in the machine using the arbour. The dynamometer is
fixed with the work table along with the vice of the machine for the measurement of
the cutting forces. The workpiece is mounted on the vice and checked for levelling.
The infrared thermometer is used for the measurement of the cutting temperature.
Surface roughness tester measures the value of surface roughness. The experiments
were conducted for both the conditions of dry and heat pipe-assisted cooling with
cutting parameters remaining the same. The experimental set-up is shown in Fig. 4.

3 Results and Discussion

The effect of the milling process was studied under the conditions of dry machining
and heat pipe-assisted machining. The cooling performance of the heat pipe-assisted
cooling in milling is analysed by comparing the values of surface roughness, cutting
temperature and cutting forces.
It is clear from Table 3 that irrespective of speed of spindle, depth of cut and feed
values of surface roughness, cutting force and cutting temperature obtained during
milling with heat pipe cooling has an improvised result in comparison with that of
dry machining. The cutting temperature of dry machining is so high since there is no
external agent such as coolant for heat removal. Due to an increase in temperature and
cutting forces, dry machining results in the improper surface finish of the component.
This elevated temperature of the tool leaves burnt marks on the machined surface of
the component.
Figure 5 shows a comparison of cutting temperatures obtained during each run,
from the graph it is clear that in heat pipe-assisted milling machining happens at a
lower temperature. The heat pipe housed on the milling cutter effectively removes
heat from the cutting zone due to its thermal conductivity and phase transition prop-
erties. This reduced temperature during the milling process aids in the production of
Experimental Investigation on Heat Pipe-Assisted Cooling … 607

Table 3 Observation during the experiments

Run Cutting temperature (°C) Cutting force (N) Surface roughness (µm)
Dry milling Heat pipe Dry milling Heat pipe Dry milling Heat pipe
assisted assisted assisted
milling milling milling
S1 37.4 31.5 90.6 77.6 1.143 1.234
S2 42 35.6 109.4 103.3 1.101 1.201
S3 58.9 42.3 215.7 135.6 1.144 1.164
S4 38.2 34.7 223.8 130.3 0.954 1.098
S5 43.2 38.6 196.95 181.9 0.871 0.997
S6 66.3 49.1 170.27 146.1 0.852 0.894
S7 39.8 35.2 219.5 142.6 0.857 0.942
S8 52.2 39.1 274.9 189.8 0.582 0.64

Fig. 5 Comparison of cutting temperature

good quality of surface finish on the component along with a considerable reduction
in cutting force which is shown in Fig. 6. The surface finish of the obtained compo-
nent is measured for each run of the experiment, and a comparison is made between
dry and heat pipe-assisted milling plotted in the form of a graph as shown in Fig. 7.
Since there is no usage of cutting fluid in the heat pipe-assisted milling process, it
causes no hazard to the environment as well as the operator.
608 I. Kantharaj et al.

Fig. 6 Comparison of cutting forces

Fig. 7 Comparison of surface roughness

4 Conclusion

For the milling process under study, for each experimental run, parameters such as
surface roughness, cutting force and temperature are considered and a comparison
is made between both dry machining and heat pipe-assisted machining. From the
results obtained, it is observed that the heat pipe-assisted milling process exhibited
improved surface finish in comparison with dry machining, due to the lower cutting
temperature observed during the machining process. Also, there is a considerable
reduction in cutting force, thus conserving energy. Since heat pipe-assisted cooling
does not use cutting fluid, it eliminates the process of safekeeping and disposal of
cutting fluid, thereby reducing the cost of the final product. Thus, heat pipe-assisted
milling will serve as a better alternative for conventional dry milling process. This
Experimental Investigation on Heat Pipe-Assisted Cooling … 609

process also aids in the promotion of greener manufacturing environment since it

does not consume any coolant for the process.

References

1. Shokoohi Y, Paiva JM, Fox-Rabinovich G, Bork CAS, Veldhuis SC (2019) Evaluation of the
superabsorbent coolant as a new approach to semi-dry machining. Int J Adv Manuf Technol
101:2037–2050
2. Kalpakjian S (2001) Manufacturing engineering and technology. Pearson Education India
3. Liew PJ, Shaaroni A, Sidik NAC, Yan J (2017) An overview of current status of cutting fluids
and cooling techniques of turning hard steel. Int J Heat Mass Transf 114:380–394
4. Sharma VS, Dogra M, Suri N (2009) Cooling techniques for improved productivity in turning.
Int J Mach Tools Manuf 49:435–453
5. Chan C, Lee W, Wang H (2013) Enhancement of surface finish using water-miscible nano-
cutting fluid in ultra-precision turning. Int J Mach Tools Manuf 73:62–70
6. Arulraj GA, Wins KLD, Raj A (2014) Artificial neural network assisted sensor fusion model
for predicting surface roughness during hard turning of H13 steel with minimal cutting fluid
application. Proc Mater Sci 5:2338–2346
7. Liang L, Liu X, Li X-Q, Li Y-Y (2015) Wear mechanisms of WC–10Ni3Al carbide tool in dry
turning of Ti6Al4V. Int J Refract Metal Hard Mater 48:272–285
8. Adler D, Hii W-S, Michalek D, Sutherland J (2006) Examining the role of cutting fluids in
machining and efforts to address associated environmental/health concerns. Mach Sci Technol
10:23–58
9. Brinksmeier E, Meyer D, Huesmann-Cordes AG, Herrmann C (2015) Metalworking fluids—
mechanisms and performance. CIRP Ann 64(2):605–628
10. Anton S, Andreas S, Friedrich B (2015) Heat dissipation in turning operations by means of
internal cooling. Proc Eng 100:1116–1123
11. Das SR, Dhupal D, Kumar A (2015) Experimental investigation into machinability of hardened
AISI 4140 steel using TiN coated ceramic tool. Measurement 62:108–126
12. Debnath S, Reddy MM, Yi QS (2016) Influence of cutting fluid conditions and cutting parame-
ters on surface roughness and tool wear in turning process using Taguchi method. Measurement
78:111–119
13. Shokrani A, Dhokia V, Newman ST (2012) Environmentally conscious machining of difficult-
to-machine materials with regard to cutting fluids. Int J Mach Tools Manuf 57:83–101
14. Ghosh S, Rao PV (2015) Application of sustainable techniques in metal cutting for enhanced
machinability: a review. J Clean Prod 100:17–34
15. Robinson Gnanadurai R, Varadarajan AS (2016) Investigation on the effect of cooling of the
tool using heat pipe during hard turning with minimal fluid application. Eng Sci Technol Int J
19(3):1190–1198
16. Stephenson DA, Agapio JS (2016) Metal cutting theory and practice, 3rd edn. CRC Press, Boca
Raton
Development of a Modified Magnetic
Moulding Set up for Improved Heat
Transfer

B. Anand Ronald, M. Harshal, K. Barath Varadaraj, and G. Gopinath

Abstract Magnetic moulding is a relatively lesser-known process used for casting

of metallic materials. The conventional magnetic moulding set up has certain short-
comings, due to heat transfer from the molten metal to the container wall through
the steel balls; simultaneously heat is generated from the container wall due to Joule
Thomson heating from the current-carrying copper coils, leading to the possibility
of stagnation in heat transfer and thereby retarding the heat carried away to the steel
balls. In order to overcome this, the current work deals with, the conventional mag-
netic moulding set up modified with water circulating jacket, so as to enhance the
heat transfer, which is expected to improve the solidification rate and hence have
control over the grain size of the cast products.

Keywords Magnetic moulding · Heat transfer · Water jacket

1 Introduction

Metal casting is one of the widely used primary manufacturing processes. Currently,
there is a greater need for manufacturing processes that can reduce lead time, cost of
manufacturing without compromising on the quality of products. Also, many efforts
are made towards production strategies/processes that reduce damage to the envi-
ronment. Magnetic mould casting is an innovative process having a great potential
to replace conventional casting methods due to various advantages associated with
it. In magnetic moulding steel balls are used instead of moulding sand and an elec-
tromagnetic field, surrounding the moulding flask acts as a binding agent [1]. This
process is similar to the lost foam process because the pattern used is an expend-
able polystyrene pattern. Once the cast metal is solidified the electromagnetic field is
interrupted so that the steel balls would collapse by itself; the part is recovered and the
steel balls could be reused in the subsequent castings [2]. Compared to the moulding

B. Anand Ronald (B) · M. Harshal · K. Barath Varadaraj · G. Gopinath

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, Tamil Nadu 603110, India
e-mail: anandronaldb@ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 611

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_54
612 B. Anand Ronald et al.

sand (mixture of sand, clay and water) used in sand casting, this mould has no mois-
ture content to generate steam and vapour, which generally results in casting defects
like blow holes and pores when they are trapped in the mould itself [3]. Therefore,
replacing the sand with alternate materials with better thermal conductivity, like steel
balls improves the heat transfer rate and reduces the solidification time of casting.
This paper deals with the development of a modified magnetic moulding set up to
overcome the heat transfer limitations in the conventional magnetic moulding set up.

2 Modified Magnetic Moulding Set up

The type of process employed for casting has an effect over properties like surface
finish, microstructure, hardness, toughness, etc. of the resultant product. Magnetic
mould casting is an application of electromagnetism to the process of casting. Steel
shots constitute the mould that is formed by application of magnetic field on them.
The application of magnetic field induces temporary magnetic force between steel
shots giving strength to mould [4]. This reduces the time elapsed in the ramming
process. Also, breaking of mould becomes easier by using magnetic field, as it is
required to switch off the supply to turn down the magnetic field and mould breaks.
The conventional magnetic moulding set up (Fig. 1) consists of a mild steel con-
tainer inside which the steel balls are filled with the pattern at the center. The pattern
is made using Expendable Polystyrene pattern (EPS) and it is cut according to the
cast product size and shape required. Polystyrene is in solid-state (glassy) at room

Fig. 1 Schematic diagram of the conventional magnetic moulding setup. 1. Steel shots. 2.
Expendable polystyrene pattern (EPS). 3. Container. 4. Copper coil
Development of a Modified Magnetic Moulding … 613

(a) Top View (b) Front View (c) 3D model

Fig. 2 Schematic diagram of the modified magnetic moulding set up

temperature, but when heated above 100 °C it evaporates. Copper coils of 22 gauge
were wound surrounding the mild steel container. The number of turns of the copper
wire was 500 with 8 A DC power supply given.
When the molten metal is poured over the polystyrene pattern, the molten metal
evaporates the polystyrene, and the molten metal takes the shape of the pattern, trans-
fers the heat to the surrounding steel balls, which act as mould wall. This contributes
to nucleation and thereby encourages directional solidification of the mould mate-
rial. In the conventional magnetic moulding set up, the heat that is dissipated from
the molten metal moves towards the steel balls by conduction and then to the mild
steel container wall. Simultaneously, certain amount of heat is generated due to Joule
heating/Ohmic heating, from the current-carrying copper coils that are wound around
the container. This can lead to stagnation of heat in the steel balls and hence lead to
reduction in solidification time for the cast material. To overcome this stagnation of
heat, a modified magnetic moulding set up was developed with a water jacket in the
container wall.
The water jacket was developed such that room temperature water enters inlet
port available at the bottom, the heat is transferred to the water during circulation
and the hot water exits through the top port. The schematic diagram of the modified
magnetic moulding set up is shown in Fig. 2.

3 Calculations

Calculation of amount of heat from aluminium

The amount of heat from the molten aluminium is calculated as follows:
Mould Dimension:

Diameter = 30 mm = 0.03 m
Radius = 15 mm = 0.015 m
614 B. Anand Ronald et al.

Volume of Mould = πr 2 h
= π × (0.015)2 × (0.06)
= 4.24 × 10−5 m3
Mass of Aluminium = Volume × Density
= 4.24 × 10−5 × 2700
= 0.11448 kg
Temperature Difference T = 700 ◦ C − 30 ◦ C
Required = 670 ◦ C

Specific Heat of Aluminium Cp = 0.91 kJ/kg K

Amount of heat from the mould
Q 1 = mCp T
= 0.1148 × 0.91 × 670
= 69.99 kJ
Therefore, Q1 = 69.99 kJ

Calculation of amount of heat produced in the coil

According to Joule’s law,

Heat produced in the coil H = I 2 RT

To find Current (I ), Selecting the Output Current (I ) from the transformer as
I = 3A
To find Time (T ), Time of current supply to the coil T = 5 min
5 × 60 = 300 s
ρl
To find Resistance (R), Resistance of the coil R = 
A

Resistivity of copper coil = 1.72 × 10−8 m

Length of the wire = No of turns × Circumference of the outer cylinder
= 500 × π × 144 = 226.080 m
Area of the coil = πr 2 = 6.28 × 10−6 m2
Resistance of the copper coil R = 0.6192 
Heat produced in the coil H = I 2 RT
H = 32 × 0.6192 × 300
= 1671.83 J
= 1.67183 kJ
Q2 = H = 1.67183 kJ
Development of a Modified Magnetic Moulding … 615

Amount of magnetic induction produced in current-carrying coil

μI
Magnetic Induction B = T
2πr
To find Magnetic Permeability (μ),
Magnetic Permeability for materials

Magnetic Permeability of MS cylinder = 1.262344 × 10−6 H/m

Magnetic Permeability of Water = 1.256627 × 10−6 H/m
Magnetic Permeability of Steel shots = 5.044567 × 10−6 H/m
Average Permeability of materials = 2.521179 × 10−6 H/m

To find Current (I),

Selecting the Output Current (I) from the transformer as

Current I = 3A

To find Radius of the coil (r),

Selecting copper coil of grade (SWG 19)

Radius of the copper coil r = 1 mm = 0.001 m

μI 2.52117 × 10−6 × 3
Magnetic Induction B = =
2πr 2 × π × 0.001
= 1.20377 × 10−3 T
Actual Magnetic Induction B = N × B
= 500 × 1.20377 × 10−3
= 0.60188 T

Mass flow rate of water at time ‘t’ second

Total Amount of Heat from the setup Q total = Q 1 + Q 2

= 69.99 + 1.67183
= 71.66183 kJ
Qtotal = 71.66183 kJ

Mass flow rate of water to decrease 5 °C at 30 s

616 B. Anand Ronald et al.

Fig. 3 Actual photograph of

the inner and outer
concentric cylinders during
fabrication

Total Amount of Heat from the setup Q total = ṁCp T

Q total
ṁ =
T × Cp × 30
= 0.119 kg/s
= 0.119 l/s

Thus the Mass Flow rate determined is equal to 0.119 l/s.

4 Fabrication of the Mild Steel Container

Two concentric cylinders are used so that there can be a passage of fluid in between
the walls of the closed cylinder. A cylindrical mild steel pipe is chosen for this
purpose. As per the design calculation, we require a 2 mm thick-walled cylinder so
that magnetic field can penetrate. From the determined flow rate for proper removal
of heat the diameters of the cylinders are selected as 80 and 140 mm (outer diameter).
So in order to achieve the required dimensions 88 and 148 mm mild steel pipe (Outer
diameter) of thickness 4 mm is purchased and is turned using four jaw lathe to 80
and 140 mm diameter with 2 mm thickness (Fig. 3).

5 Conclusions

A Modified Magnetic Mould casting setup was successfully fabricated. Since the
modified set up has a water jacket, it is expected that the modified set up will give bet-
ter heat transfer, which can be ascertained by evaluating (at the outlet port) the amount
of heat carried away. Also after solidification, the mechanical (tensile strength) and
Development of a Modified Magnetic Moulding … 617

metallurgical properties (fine grain size) of the cast material will give insight into the
improvement due to modification of the existing magnetic moulding set up.

References

1. Wittmoser A (1975) The new third generation of molding processes. AFS Trans 63–72
2. Suganth Kumar KJ, Senthilkumaran S, Rajendra Boopathy S, Ramesh A (2008) Theoretical and
experimental investigation of mold strength in magnetic molding compacts. J Mater Process
Technol 205:235–242
3. Rajkolhe R, Khan JG, Gajanan SS (2014) Defects, causes and their remedies in casting process:
a review. Int J Res Advent Technol 2(3). E-ISSN: 2321-9637
4. Anand Ronald B, Arun Prakash C, Suba Karthik M (2016) Influence of steel shots size on tensile
properties of magnetic moulded MMC. Appl Mech Mater 852:118–122
Quality Assessment Studies on AA7075
Plate in Hot Rolling Process

K. Jayakumar

Abstract Rolling is an important material-forming process and more than 90% of

ferrous and nonferrous metals and alloys are processed to their usable shapes by
rolling. In the rolling process, permanent deformation is achieved by subjecting the
material to high compressive stress by allowing the material to pass through the gap
between two rotating cylindrical rolls. In 7075 aluminum alloy, zinc as the primary
alloying element. It is strong, with a strength comparable to many steels, and has
good fatigue strength and average machinability. In order to improve further the
strength of the AA7075 alloy, the hot rolling study was carried out on a 6.62 mm
thickness plate. Hot rolling was carried out as per L9 orthogonal array by varying
roller speed (750 and 850 rpm), the number of passes (10 and 20) and preheating the
workpiece at 350 and 500 °C. From the experimental results, it was observed that
the hardness value was increased from 42.6 to 48.3 HRB with an increase of all the
selected process parameters. The thickness of the rolled plate was reduced from 6.62
to 5.53 mm at a higher number of passes. Dye penetrant inspection was carried out
to see the surface-level defects or microstructural changes. The rolled plate with low
roller speed, less number of passes, and higher preheating temperature showed more
surface-level defects or microstructural deformation.

Keywords AA7075 alloy · Hot rolling study · Hardness · Thickness variation ·

Dye penetrant inspection

1 Introduction

AA7075 alloy is mainly used in Aircraft fittings, Gears and shafts, Fuse parts, Missile
parts, Worm gears, keys, aircraft, aerospace and defence applications, bike frames,
all-terrain vehicle (ATV) sprockets, etc. It has lower resistance to corrosion than

K. Jayakumar (B)
Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Kalavakkam,
Chennai 603110, India
e-mail: kjayakumar@ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 619

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_55
620 K. Jayakumar

many other aluminum alloys but has significantly better corrosion resistance than
the 2000 series aluminum alloys. Its relatively high cost limits its use. Rolling is the
most important material-forming process by which properties of the materials can be
improved further by grain refinement. In hot rolling, the workpiece is heated nearer to
its recrystallization temperature and hence the compressive force required for defor-
mation is comparatively less and porosity of the rolled metal can be minimized. Also,
the smaller capacity rolling stand can be used for rolling operation. Much research
work has been conducted on the rolling of different aluminum alloys. Panigrahi et al.
[1] did rolling studies on AA7075 alloy at cryogenic and room temperature and they
compared its mechanical properties from tensile tests and hardness testes results.
The conclusion from the study was rolling with cryogenic temperature is good. The
effects of rolling parameters on the mechanical behavior of AA6061 alloy was carried
out by Rajabi et al. [2]. They concluded that higher rolling temperature and strain rate
were improved the tensile strength of the material. Guo et al. [3] suggested that the
true stress-strain curves at 300–500 °C temperature and strain rates of 0.01–10 s−1
were obtained through isothermal compression tests, providing a data basis for the
hot processing map of AA7075 alloy. From 2015 onwards, the accumulative roll
bonding process (ARB) came to replace the conventional rolling to further improve
the properties of the AA alloys. Alvandi et al. [4] studied the mechanical proper-
ties of fine structured AA7075 alloy by ARB process at room temperature. Results
showed that the tensile and hardness of the sample were improved by ARB compared
with alloy with a coarse grain sample. Guo et al. [5] evaluated the microstructure of
hot-rolled AA7075 alloy through Friction Stir Processing (FSP). Major conclusions
were the maximum superplastic elongation of 776.4% was achieved at 500 °C and
strain rate 10−3 s−1 for the hot-rolled AA7075 subjected to one pass FSP. Microstruc-
ture and deformation behavior of Al–Fe–Mn alloys with different Fe contents during
Cold Rolling was tried by Pan et al. [6].
From the available literature, it was found that different researchers have tried to
improve the mechanical properties of the AA7075 alloy by different rolling processes.
But, they evaluated the improvement in properties by destructive testing. However,
the present work focused on the improvement of mechanical properties of AA7075
by hot rolling process and assessment of the mechanical properties by destructive
and nondestructive testing through dye penetrant testing.

2 Experimental Details

The AA7075-T651 alloy with the chemical composition in maximum weight % of

6.1 Zn, 2.9 Mg, 2 Cu, 0.5 Fe, 0.4 Si, 0.28 Cr, 0.2 Ti, and Al balance (91.4%) was taken
as raw material. The initial size of the plate is 300 mm × 300 mm × 6.62 mm. To
improve further the mechanical properties of the alloy, eight sample pieces of 20 ×
70 × 6.62 mm were cut from the plate and used for the further rolling study. Before
the rolling process, the thickness and hardness of a parent sample were measured
Quality Assessment Studies on AA7075 Plate in Hot Rolling Process 621

Roller

AA7075 sample

Fig. 1 Rolling of AA7075 sample

using digital vernier caliper and Rockwell hardness tester with B scale respectively.
Initial hardness value obtained as 42.6 HRB and thickness as 6.62 mm.
The hot rolling mill (Maximum Temperature = 1200 °C) facility which is available
in the Department of Metallurgical and Materials Engineering at IIT Madras was
used in the present work. Figure 1 shows the experimental setup with the rolling
of aluminum plate. The hot rolling study was conducted by varying roller rotational
speed, the number of passes and preheating temperature of the workpiece. The rolling
parameter was selected based on trial and error as well as machine constraints. Rolling
pressure or load was kept as constant with 0.5 ton. Eight experiments were conducted
as per three factors with two level-full factor experimental designs due to facility
constrain.
After rolling, the hardness test was conducted for all eight pieces to check the
improvement of hardness. Rockwell B scale hardness test conditions are Load-
100 kg, Indenter-1/16 in. or 1.6 mm diameter ball indenter. Hardness was measured
in each plate at three different locations and the average is considered. Dwell time
of 10–15 s was maintained during the hardness test to allow for elastic recovery.
Thickness variation of the rolled samples was also measured and Table 1 lists the
experimental design with results. During thickness measurement, thickness variation
was observed in all samples from middle to edges. Average of thickness variation
from all samples is 0.03 mm.
Figure 2 shows the samples after the rolling process in which the first sample
which is present in the right end is a not rolled one and the remaining eight pieces
are rolled one as per eight experimental conditions.
To check the internal defect of the rolled samples, Dye penetrant (DP) based
nondestructive testing (NDT) was carried. DP test is sensitive to small surface dis-
continuities/defects, gives real-world results and affordable one than other NDTs.
There are three types of liquids that were used in the DP test and they are cleaner,
penetrant, and developer.
622 K. Jayakumar

Table 1 Rolling experimental condition and results

S. No. Roller speed No. of passes Temperature Hardness Thickness
(rpm) (°C) (HRB) (mm)
1 750 10 350 45.6 6.11
2 850 10 350 44.5 6.34
3 750 20 350 46.6 5.98
4 850 20 350 45.3 5.53
5 750 10 500 44.3 6.31
6 850 10 500 46.3 6.22
7 750 20 500 47 5.8
8 850 20 500 48.3 6.23

Fig. 2 Samples after rolling process

The cleaner is used to remove the dust particles on the workpiece surface and the
penetrant is used to fill the surface cracks. Dwell time of 15 min was maintained after
applying penetrant and Fig. 3 shows the rolled plate after applying the developer.
A developer is a nonaqueous wet type one and they are acetone, isopropyl alcohol,
or a propellant that is a combination of these two.

8 7 6 5 4 3 2 1 Base material

Fig. 3 Rolled specimens after applying developer

Quality Assessment Studies on AA7075 Plate in Hot Rolling Process 623

3 Results and Discussions

3.1 Effect of the Rolling Parameter on Hardness

and Thickness Variation

The effect of a rolling parameter on hardness variation is shown in Fig. 4 in which

the number of passes in hot rolling is the most sensitive factor than the other two
factors.
Main effect graphs were obtained from Minitab software based on the experi-
mental results. The hardness of the sample was increased for an increase in all the
selected process parameters. As the number of passes and temperature during rolling
is increased, the work hardening and dislocation density are increased which caused
the grain refinement and fragmented precipitates which improved the hardness.
As the rolling speed increased, the hardness of the strip was also increased which
is due to the decrease of time taken for temperature drop which caused the grain
refinement during the application of compressive load in the rolling process.
Similar to hardness, the thickness of the rolled plates was increased owing to
the increase of roller speed and preheating temperature of a workpiece (Fig. 5). The
thickness of the rolled strip is increased for an increase in rolling speed since the time
for temperature drop was less which caused more deformation. The thickness of the
AA7075 strip was decreased because of the increase in the number of passes. The
reason is that as the no. of passes increased, the edges got tore instead of deformation
which finally decreased the thickness.
An increase in the thickness of the sample as the increase in temperature was
observed due to more deformation during rolling instead of compression.

Fig. 4 Main effect plot for hardness variation

624 K. Jayakumar

Fig. 5 Main effect plot for thickness variation

3.2 Effect of the Rolling Parameter on Surface Defects

From the DP test after applying developer, sample number 3 and 4 are shown in
Fig. 6 has more surface-level microcracks. The reason is that the rolling process was
carried out with more number of passes (20 numbers), low temperature (350 °C)
with high speed (850 rpm) in those experiments. Due to these effects, the material is
not able to deform and got slightly torn on the edges which caused microcracks.
Overall, from these rolling studies, it was observed that the experiment number 7
and 8 can be recommended for getting good mechanical properties of the AA7075
strip.

Micro level surface cracks

Sample 2 Sample 3 Sample 4 Sample 5

Fig. 6 Rolled samples with surface-level defects

Quality Assessment Studies on AA7075 Plate in Hot Rolling Process 625

4 Conclusions

• In this study, the hardness, thickness variation, and surface-level defects of

AA7075 aluminum alloy plate through the hot rolling process were investigated.
• The hardness of the AA7075 sample is increased from 42.6 to 48.3 HRB by hot
rolling process.
• The thickness of the sample is reduced from 6.62 to 5.53 mm through the hot
rolling process. The number is passes of the sample during the rolling process
was identified as the dominant factor for both responses.
• Dye penetrant inspection was carried out to see the surface-level defects or
microstructural changes from that microcracks or defects were observed for the
third and fourth experiments.

References

1. Panigrahi SK, Jayaganthan R (2008) A comparative study on mechanical properties of Al 7075

alloy processed by rolling at cryogenic temperature and room temperature. Mater Sci Forum
584:734–740
2. Rajabi F, Zarei-Hanzaki A, Eskandari M, Khoddam S (2013) The effects of rolling parameters
on the mechanical behavior of 6061 aluminum alloy. Mater Sci Eng A 578:90–95
3. Guo L, Yang S, Yang H, Zhang J (2015) Processing map of as-cast 7075 aluminum alloy for hot
working. Chin J Aeronaut 28(6):1774–1783
4. Alvandi H, Farmanesh K (2015) Microstructural and mechanical properties of nano/ultra-fine
structured 7075 aluminum alloy by accumulative roll-bonding process. Procedia Mater Sci
11:17–23
5. Guo ML, Tan MJ, Liu FC, Song X, Chua BW (2016) Microstructure evolution in hot rolled
7075 Al via friction stir processing. In: AIP conference proceeding, vol 1769, no 1, pp 1–6
6. Pan Y, Shen Y, Zhao P (2018) The microstructure and deformation behavior of Al-Fe-Mn alloys
with different Fe contents during cold rolling. Metals 753(8):1–6
Artificial Neural Network and Genetic
Algorithm-Based Models for Predicting
Cutting Force in Turning of Hardened
H13 Steel
K. Leo Dev Wins, B. Anuja Beatrice, D. S. Ebenezer Jacob Dhas,
and V. S. Anita Sofia

Abstract The manufacturing sector in the modern era is striving hard to reduce the
cost of production by employing innovative techniques. One such technique is hard
turning where the workpiece is heat treated to the requisite hardness, and the final
size and shape of the component are obtained directly through hard turning process.
Hard turning is generally carried out with a huge amount of cutting fluid to enhance
the output performance. Since petroleum-based emulsions are easily available in
the market at reasonable price, they are widely used in industries. Petroleum-based
cutting fluids create a number of environmental and health issues. In this perspective,
pure dry turning is a logical substitute as it does not possess the harmful effects
connected with the cutting fluids. The feasible tool life and surface quality are often
disturbed while carrying out the machining operation under pure dry condition. Under
such circumstances, the concept called minimal cutting fluid application (MCFA)
performs itself as a possible solution. This paper investigates the effect of applying
cutting fluid using MCFA technique at the critical contact zones while hard turning of
H13 steel. An artificial neural network (ANN) model was developed for the prediction
of the main cutting force, and its ability to predict cutting force (F z ) was analyzed.
An effort is made to optimize the cutting parameters to accomplish minimum cutting
force using genetic algorithm.

Keywords Hard turning · Artificial neural network · Minimal cutting fluid

application · Genetic algorithm

K. Leo Dev Wins (B) · D. S. Ebenezer Jacob Dhas

Karunya Institute of Technology and Sciences, Coimbatore, Tamil Nadu, India
e-mail: leo@karunya.edu
B. Anuja Beatrice · V. S. Anita Sofia
Sri Krishna Arts and Science College, Coimbatore, Tamil Nadu, India

© Springer Nature Singapore Pte Ltd. 2021 627

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_56
628 K. Leo Dev Wins et al.

1 Introduction

The manufacturing sector is increasingly adopting newer methods and techniques

to reduce the cost of production and manufacturing lead time to ensure productivity
growth and delivery of top quality products in the most effective and economi-
cal manner [1]. In this endeavor, lot of researchers took it as a promising area of
research to device new prospects in manufacturing processes. By virtue of rigidity,
good compressive strength, and high wear resistance, hardened steels meet industry
demands in the fabrication of high-performance components. They are highly used
in the field of automotive and die tool industry [2]. In the recent years, the concept of
hard machining has gained great importance in metal cutting. The increasing need
for more efficient machining techniques and the timely emergence of high-quality
cutting tool materials such as CBN have led to the wide usage of hard turning pro-
cess. Reduced equipment costs, few process steps, and flexibility are the influencing
factors for the selection of hard turning compared to conventional practices [3].
Hard turning process necessitates large quantities of cutting fluids. Cooling and
lubrication of workpiece and cutting tool during hard turning operation are normally
executed by free flow of cutting fluids to reduce friction and heat developed during
machining. The high temperature at the cutting zone during hard turning produces
harmful fumes which not only create adverse effects to the workers but also causes
environmental pollution [4]. With the advent of new standards regarding the environ-
ment, health and safety, some of the ingredients in many of the cutting fluids were
identified as problematic. Because of such adverse effects created by cutting fluids,
a number of government agencies have established regulations and standards for the
need of reducing particulate exposure [5].
Cutting fluid usage, storage, and disposal affect the cost-effectiveness of the
machining processes. Industries spend 7–17% of the manufacturing cost on cutting
fluids [2]. Under such situation, dry machining is a viable alternative but it requires
machine tools and cutting tools that can take enormous cutting forces. Hence, such
a scheme cannot be executed as such in the existing shop floor. Enormous amount
of vibration is produced which may also lead to premature tool failure [6].
For reducing harmful effects linked with the use of cutting fluids, minimal quantity
lubrication (MQL) technique was developed. MQL method also improves surface
finish and dimensional accuracy [7] but adds the risk of mist in the shop floor [8].
Alternative to this problem, Varadarajan et al. developed a new fluid application
system called minimal cutting fluid application system (MCFA) in 2002 [9], by
excluding the role of compressed air from the system. In this method, cutting fluid
is delivered at high velocity and is pumped directly to the cutting zone with the help
of an injector in a pulsating manner, and this process created very less oil mist [10].
The results proved that the MCFA reduces cutting force and vibration and improves
surface quality and tool life when compared to wet cooling and dry machining [11].
Analytical models often use unfavorable assumptions and simplifications to get
superior and proper results. In real-time applications, artificial intelligence-based
modeling approaches are preferred, in that ANN is found to be attractive, reliable,
Artificial Neural Network and Genetic Algorithm … 629

Table 1 Chemical composition of H13 steel (weight %)

C Mn Si P S Cr Mo Fe
0.430 0.214 1.08 0.0330 0.0290 5.02 1.13 Balance

and viable [11]. Özel and Karpat used ANN model for predicting flank wear and
surface finish during hard turning [12]. An ANN model was developed by Leo et al.
for predicting surface finish based on ANN during surface milling of AISI 4340 steel
of hardness 43 HRC with minimal fluid application [13].
Genetic algorithm with the basic principle of genetics and evolution is one of the
optimization methods. It mimics biological evolution for solving search and opti-
mization problems [14]. Computer programs based on genetic algorithm approach
helps to produce optimal solutions [15].
The literatures review revealed that, lot of research works have been reported on
the optimization of surface finish and cutting force during hard turning of different
steels but no work is reported on the prediction of cutting force using ANN during
hard turning of H13 steel with minimal cutting fluid application. In this research,
hard turning process was assisted with minimal cutting fluid application technique,
which reduced the usage of cutting fluid to a minimum amount of 8 ml/min. Based on
the earlier researches in the optimization while machining of hardened steels [16], an
attempt was also made to achieve minimum cutting force by optimizing the cutting
parameters using genetic algorithm. The prediction model derived out of GA was
found to be highly accurate with more than 99% accuracy.

2 Selection of Workpiece and Cutting Tool

H13 tool steel of 45 HRC was used as the workpiece material in this research. It
is one of the toughest tool steels. It has high toughness and red hardness. Its good
hardenability makes it useful in making extrusion tools and dies.
Accordingly, a cylindrical rod of 65 mm diameter with 355 mm length was used
for the experimentation. Table 1 shows the percentage composition of elements in
the workpiece material. The tool used for hard turning was SNMG120408 (MT TT
5100) hard metal inserts and tool holder used was PSBNR2525M12.

3 Experimental Setup

Experimental setup used an all-geared Kirloskar Turnmaster-35 lathe with variable

feed and speed drive. Figure 1 shows the experimental setup. Variable DC controllers
were used to control the feed rate and speed independently.
630 K. Leo Dev Wins et al.

Fig. 1 Experimental setup

Kistler multi-component piezoelectric dynamometer of type 9257B was used

for the measurement of cutting force. Screening experiments were conducted to
identify the main factors that have major effect on the output parameter, and thorough
analysis was done to fix the upper and lower limits of the factors. Accordingly, trial
runs were conducted by changing one of the input parameters while keeping other
parameters constant. Accordingly, spindle speed of the lathe was fixed between 77
and 115 m/min. In line with this factor, feed rate of the lathe was fixed between 0.05
and 0.1 mm/rev and the depth of cut was fixed between 0.5 and 1 mm. The fluid
application parameters were kept constant. Accordingly, pressure at the injector was
kept at 100 bar, frequency of pulsing at 500 pulses/min, quantity of application at
8 ml/min, and the cutting fluid concentration as 20% of oil in water [17].
Five-level factorial central composite rotatable design with full replications was
used for the design of experiment because these designs can include information from
a correctly planned factorial experiment. Table 2 presents five levels of parameters.
Table 3 shows the design matrix with the experimental values and results.

Table 2 Five levels of cutting parameters

Cutting parameters Levels
Cutting speed (mm/min) 77 85 96 107 115
Feed rate (mm/rev) 0.05 0.06 0.07 0.09 0.1
Depth of cut (mm) 0.5 0.6 0.75 0.88 1
Artificial Neural Network and Genetic Algorithm … 631

Table 3 Experimental data

Run order Type of data Cutting speed Feed rate (f ) Depth of cut Cutting force
(v) (m/min) (mm/rev) (d) (mm) (F z ) (N)
1 Training 85 0.06 0.6 211.8
2 Training 107 0.06 0.6 248.2
3 Training 85 0.09 0.6 250
4 Training 107 0.09 0.6 260.7
5 Training 85 0.06 0.88 228
6 Training 107 0.06 0.88 240
7 Testing 85 0.09 0.88 318
8 Training 107 0.09 0.88 321
9 Training 77 0.07 0.75 245.3
10 Training 115 0.07 0.75 254
11 Training 96 0.05 0.75 239.3
12 Training 96 0.1 0.75 345
13 Training 96 0.07 0.5 202.7
14 Testing 96 0.07 1 281.5
15 Training 96 0.07 0.75 276.5
16 Training 96 0.07 0.75 277.3
17 Training 96 0.07 0.75 276.5
18 Training 96 0.07 0.75 277.1
19 Training 96 0.07 0.75 277.5
20 Testing 96 0.07 0.75 278.1

4 Network Algorithm for ANN

It has been proved by many researchers that the artificial neural network (ANN)
model has better accuracy and yields more admissible results when compared with
the simple tools such as Taguchi method and regression analysis. Artificial neural
network seems to be the most preferable algorithm for developing a predictive model
for complex nature of the hard turning of H13 steel. Artificial neural network was
developed and trained using quickprop (QP) algorithm, and cutting force values
were predicted. The neural network contains one input layer and an output layer.
The hidden layer was changed into single and multi-layer types with the changes in
number of neurons, and cutting force was predicted with the help of best configuration
which gave better coefficient of determination with least MSE error through trial-
and-error method [18]. “Tansig” was considered as the transfer function. Equation (1)
shows the sigmoid function used in this experimentation.

1
F(x) = (1)
1 + e−x
632 K. Leo Dev Wins et al.

Fig. 2 ANN layer

configurations with limiting
RMSE values

In order to get the best possible network structure, ANN model was trained with
80 different network configurations. Some of the network configurations which gave
lower values of RMSEs are presented in Fig. 2. Based on the results, 3-4-1 network
structure was selected with multilayer normal feedforward algorithm.
This ANN model consists of one input layer with three neurons corresponding to
cutting speed, feed rate, and depth of cut, one hidden layer with four neurons, and
an output layer corresponding to cutting force as shown in Fig. 3.
The regression coefficient (R) value of 0.99996 was obtained for the 3-4-1 network
configuration, which is very close to unity, and it shows a close relationship between
observed and predicted data. Table 4 shows the comparison between test patterns with
experimental results for validating the ANN model. The comparison results revealed

Fig. 3 The 3-4-1 network of

the ANN model

Table 4 Comparison between testing data and experimental results

Run Testing data Cutting force (F z ) (N)
order v (mm/min) f (mm/rev) d (mm) Experimental Prediction % Error
results by ANN
1 85 0.09 0.88 318 312.5 1.73
2 96 0.07 1 281.5 276.3 1.85
3 96 0.07 0.75 278.1 273.6 1.62
Artificial Neural Network and Genetic Algorithm … 633

Table 5 Optimum
Machining parameter Optimum value
machining parameters using
hybrid ANN-GA approach Cutting speed (m/min) 77.00109
Feed rate (mm/rev) 0.0600000325
Depth of cut (mm) 0.610000123

Table 6 Confirmatory test results

Optimum cutting force predicted through Confirmatory experimental result (F z ) % Error
GA approach (F z ) (N)
195.4946 198.076 1.303

the good concurrence between the predicted values and experimental results. No
significant interaction effect between the variables was found.

5 Optimization Using Hybrid ANN-GA

In this present investigation, an effort is made to find the optimum values of cutting
parameters that will lead to minimum cutting force. To solve the optimization problem
for this study, a hybrid ANN-GA approach was used. Predictive model was trained
and developed [19]. The developed artificial neural network (ANN) was used as the
fitness function for genetic algorithm (GA).
The machining parameters leading to minimum cutting force are the optimization
problem in this case. The constraints in this case are cutting speed (77–115 m/min),
feed (0.05–0.1 mm/rev), and depth of cut (0.5–1 mm). The GA parameters used were
population size of 30, mutation rate of 0.1, crossover rate of 0.8, and of uniform
crossover type. The optimum machining parameters which led to minimum cutting
force using hybrid ANN-GA approach are given in Table 5. The optimum cutting
force value obtained was 195.4946 N.
Confirmatory tests were conducted to check the adequacy of the optimum param-
eter combinations, and the experimental results showed good agreement with the
predictions. The confirmatory test results are shown in Table 6.

6 Conclusions

• The developed ANN model can be very much helpful in setting up the machining
parameters to reduce the force needed to achieve the best results.
• The new scheme is well suitable for the industry as it can be directly practiced in
the shop floor with no modifications in the existing setup.
634 K. Leo Dev Wins et al.

• Hybrid ANN-GA approach can be effectively used for optimizing cutting

parameters to achieve minimum cutting force.
• MCFA technique can be used in promoting a green environment in the metal
cutting industries, by minimizing risks involved in the use of huge quantity of
cutting fluid.

References

1. Enabling manufacturing competitiveness and economic sustainability. In: Proceedings of

the 4th international conference on changeable, agile, reconfigurable and virtual production
(CARV2011), Montreal, Canada (2011)
2. Attanasio A, Gelfi M, Giardini C, Remino C (2006) Minimal quantity lubrication in turning:
effect on tool wear. Wear 260(3):333–338
3. Aronson RB (1995) Why dry machining. Manuf Eng 33–36
4. NIOSH (2001) Metal working fluids—recommendation for chronic inhalation studies. National
Institute for Occupational Safety and Health, Cincinnati, OH USA
5. Marano RS, Smolinnski JM, Esingulari CWM (1997) Polymer additives as mist suppressant
in metal cutting fluids. J Soc Tribol Lubr Eng 25–32
6. Sam Paul P, Varadarajan AS (2013) Performance evaluation of hard turning of AISI 4340 steel
with minimal fluid application in the presence of semi-solid lubricants. Proc Inst Mech Eng
Part J J Eng Tribol 227(7):738–748
7. Itoigawa F, Childs THC, Nakamura T, Belluco W (2006) Effects and mechanisms in minimal
quantity lubrication machining of aluminium alloy. Wear 260:339–344
8. Dhar NR, Kamruzzaman M, Ahmed M (2006) Effect of minimum quantity lubrication (MQL)
on tool wear and surface roughness in turning AISI-4340 steel. J Mater Process Technol
172:299–304
9. Varadarajan AS, Philip PK, Ramamoorthy B (2002) Investigations on hard turning with minimal
cutting fluid application (HTMF) and its comparison with dry and wet turning. Int J Mach Tools
Manuf 42:193–200
10. Robinson GR, Varadarajan AS (2012) Investigation on the effect of an auxiliary pulsing jet of
cutting fluid on the top side of the chip during hard turning with minimal fluid application. Int
J Mach Mach Mater 12:321–336
11. Chryssolouris G, Guillot M (1990) A comparison of statistical and AI approaches to the
selection of process parameters in intelligent machining. ASME J Eng Ind 112:122–131
12. Özel T, Karpat Y (2005) Predictive modeling of surface roughness and tool wear in hard turning
using regression and neural networks. Int J Mach Tools Manuf 45:467–479
13. Leo Dev Wins K, Varadarajan AS, Ramamoorthy B (2007) Artificial neural network assisted
sensor fusion model for predicting tool wear online during hard turning. In: 9th international
symposium on measurement and quality control, IIT Madras, pp 267–270
14. Bhoskar T, Kulkarni O, Kulkarni N (2015) Genetic algorithm and its applications to mechanical
engineering: a review. Mater Today 2(4–5):2624–2630
15. Goldberg DE (2003) Genetic algorithm in search. Optimization and machine learning. Pearson
Education Asia, New Delhi
16. Kaur M, Bhattacharya A, Singh M, Batish A (2014) Hard turning: parametric optimization
using genetic algorithm for rough/finish machining and study of surface morphology. J Mech
Sci Technol 28(5):1629–1640
17. Raj A, Leo Dev Wins K, Varadarajan AS (2016) Optimization of fluid application parameters
during hard turning of AISI H13 tool steel using minimal cutting fluid application. Int J Res
Mech Eng 4(3):190–196
Artificial Neural Network and Genetic Algorithm … 635

18. Zhang G, Patuwo BE, Hu MY (1998) Forecasting with artificial neural network; The state of
the art. Int J Forecast 14:35–62
19. Mia M, Dhar NR (2017) Prediction and optimization by using SVR, RSM and GA in hard
turning of tempered AISI 1060 steel under effective cooling condition. Neural Comput Appl
1–22
Experimental Study of the Effect
of TiN–Zn Coated High-Speed Steel
Cutting Tool on Surface Morphology
of AL1060 Alloy During Machining
Operation

I. P. Okokpujie, C. A. Bolu, O. S. Ohunakin, and Esther Titilayo Akinlabi

Abstract In machining operations, surface morphology is a significant parameter

in developing mechanical components, as the level of the surface roughness gives the
consumer the notion of whether the product is well developed or not. Also, the surface
roughness of the mechanical component will affect the component during machining
operations. The application of coatings in cutting tools is very significant, as coatings
will help in improving the cutting tool life as well as protect it from rapid wear and
corrosion rate. This research work focuses on the investigation of the significance
of TiN–Zn coatings on high-speed steel (HSS) when compared with non-coated
HSS cutting tool on the surface morphology, machining parameters and machining
time on 1060 aluminum alloy during the machining operation. Computer numerical
control (CNC) program and excel software was employed to investigate the effects
the Coated cutting tool, various depth of cut and such as 1, 1.5, 2, 2.5, and 3 mm
and at a constant cutting speed of 1250 rpm, constant feed rate of 10 mm/rev, with
machining time varies from 15 to 500 s on the machined surface morphology of the
1060 aluminum alloy. After each machining time, the average surface roughness was
determined, from three-point of the work piece, that is the starting point, middle point,
and the end point. The result shows that the TiN–Zn coated HSS cutting tool on the
surface morphology of 1060 aluminum alloy was found to reduce surface roughness
with about 15.5% and both minimum surface roughness of 0.96 and 1.06 µm was
achieved for the TiN–Zn coated and uncoated HSS cutting tool, respectively, with
cutting time of 15 s at 1 mm depth of cut, which is viable for machining operations.

I. P. Okokpujie (B) · C. A. Bolu · O. S. Ohunakin · E. T. Akinlabi

Department of Mechanical Engineering, Covenant University, Ota, Ogun State, Nigeria
e-mail: imhade.okokpujie@covenantuniversity.edu.ng
O. S. Ohunakin
Faculty of Engineering and the Built Environment, University of Johannesburg, Johannesburg,
South Africa
E. T. Akinlabi
Department of Mechanical Engineering Science, University of Johannesburg, Auckland Park,
Kingsway Campus, Johannesburg 2006, South Africa

© Springer Nature Singapore Pte Ltd. 2021 637

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_57
638 I. P. Okokpujie et al.

Keywords Coatings · Machining · Surface morphology · Aluminum 1060 alloy ·

HSS cutting tool

1 Introduction

Coating and its application in cutting tool used for metal to metal machining are
very significant, as the coatings help to improve the wear rate of the cutting tool life,
prevent the cutting tool from corrosion attack, increases the electrical, thermal, and
mechanical characteristic of the cutting tool material, which enables the cutting tool
to withstand high temperature at the cutting zone and vibration during machining
operations [1–3]. Machining operations are greatly affected by the cutting tools.
The continuous failures of cutting tools during machining operations can give rise
to high cost of production of mechanical parts [4, 5]. Coating is the application of
additives on the surface of working materials for improved mechanical properties.
The state-of-the-art technology during the machining process is the use of coated
cutting tool. Improvements in covering hardware and procedures currently empower
us to deliver a wide scope of various hard nitride and oxidic films and to store them
on different apparatus substrates as monolayer, multilayer, or composite coatings [4,
6–13]. Ghani et al. [14] examined the wear rate of carbide cutting tool coated with
titanium nitride (TiN) and the uncoated carbide cutting tool, using different values
of machining speed, feed rate, and depth of cut. From the result, the authors confirm
that the uncoated cutting tool, crack with little or no time and the wear rate with
the machining parameters were very high, when compared with the carbide cutting
tool coated with TiN. Huang [15] also carried out tool evaluation in associations of
tool life centered on flank wear at standard machining conditions, such as machining
speed, depth of cut, and feed rate. They discovered that machining speed plays a
great role in analyzing the tool life evaluation, due to the fact that the increase of the
cutting speed led to gradual reduction of the tool wear, followed by depth of cut and
feed rate.
According to Schulz et al. [16] who worked on the chemical vapor deposition
(CVD) and physical vapor deposition (PVD) of TiN, TiC, and TiAl on cutting tool,
the result shows that the cutting tool life was improved significantly when compared
with the uncoated cutting tool. Fahad [17] confirmed that the application TiN coating
on the cutting tool leads to a great reduction of heat from getting to the cutting tool,
the heat rate was analyzed and the coated cutting decrease the heat by 17% for
traditional cutting speed and 60% for high-speed machining region.
The comparison of different coatings on cutting tools and uncoated cutting tools
were also studied by Astrand et al. [18] on heat distribution in the cutting tool.
This proves that coating has great influence in tool flank face and racks on the tool
wear. Grzesik [19] and Corduan et al. [20] also confirm that the multilayer coatings on
cutting tool done by physical vapor deposition performed better when compared with
the chemical vapor deposition method. Bayraktar et al. [21] produced a permanent
Al-35Zn alloy using mold casting method and carried out cutting evaluations on
Experimental Study of the Effect of TiN–Zn Coated … 639

titanium aluminum nitride (TiAlN) coated carbide tool with ordinary carbide tool
without coatings on A-35Zn alloy during end milling operation. The study employs
CNC vertical milling machining center, using a constant feed rate, depth of cut and
machining speed. From their result it is clearly seen that the TiAlN increases the
cutting force and the surface roughness when compared with the uncoated cutting
tool during the machining operation, the authors clamped that the coatings applied
reduce the contact area between the cutting tool and the working material, which also
reduces the sharpness of the cutting teeth. Quadrini [22], Qin [23] and Heaney [24]
contradict this study, saying surface roughness and cutting force can be optimized
with the application of coated cutting tool with smaller grain size.
According to Tamerabet et al. [25] the elimination of the lubrication process
from machining operation, which is dry machining process, gives a clear picture to
study and know all the factors that can lead to cutting tool damages. To optimize
the cutting condition, there is need to know the wear pattern, contact angles, and the
factors in the operating environment, this idea is to select the important or the most
influential factor values that can help to eliminate the wear behavior. The work of
the authors focusses on the application of multilayer coating on the cutting tool and
the optimization of the cutting parameters on tool wear and temperature at the tool-
chip cutting region for two different cutting tools of KC810 and KC910 during dry
turning of SAE 1030 carbon steel. The study is to achieve the optimized cutting tool
conditions using Taguchi techniques. The result shows that machining speed is the
most significant parameter on the wear rate study and the temperature distribution.
From the observations of the authors, the KC810 applied to perform better during the
turning of carbon steel with a longer lifetime of 228 min during operation. Cutting
tool also has a great impact on surface roughness of the mechanical component if not
properly optimized and controlled. From the study carried out from [26–32] shows
that vibration and surface roughness can also be affected by coated cutting tool and
non-coated cutting tool. Therefore, the need to study cutting tool performance on
surface roughness is very significant.
However, this experimental study is aimed at analyzing the performance of coated
TiN–Zn HSS cutting tool on surface morphology during the machining operation of
Aluminum 1060 alloy.

2 Materials and Methods

The material used for this research work are coated HSS cutting tool, non-coated
HSS and AL1060 alloy rectangular bar size of length of 2000 mm, width of 50 mm
and thickness of 5 mm. The machining experiment was carried out using WARCO
X6323A vertical milling machine, Vernier caliper, Steel rule, Mitutoyo surface
roughness tester. After each experiment, the surface morphology is studied by mea-
suring the surface with the Mitutoyo surface roughness tester, in order to determine
the effects of the coated cutting tool on the surface morphology of AL1060 alloy.
Tables 1 and 2 shows the chemical composition, mechanical, thermal and physical
properties of AL1060 alloy.
 640

Table 1 Chemical composition of 1060 aluminum alloy [20]

Element Fe Si Cu V Zn Mg Mn Ti AL
Weight % 0.4 0.25 0.05 0.05 0.1 0.03 0.03 0.03 Reminder
I. P. Okokpujie et al.
Experimental Study of the Effect of TiN–Zn Coated … 641

Table 2 Physical,
Properties Units
mechanical, and thermal
properties of AL1060 alloy Density 2700 kg/m3
Melting start (solidus) 649 °C
Tensile strength 130 MPa
Yield strength 94 MPa
Elastic modulus 80 GPa
Poisson’s ratio 0.33
Thermal expansion 23.6 (10−6 /°C)
Thermal conductivity 234 (mK)

Table 3 The parameter

Cutting speed Feed rate Axial depth of cut
employed for this
(m/min) (mm/rev) (mm)
investigation
1250 10 1, 1.5, 2, 2.5 and 3
1250 10 1, 1.5, 2, 2.5 and 3
1250 10 1, 1.5, 2, 2.5 and 3
1250 10 1, 1.5, 2, 2.5 and 3

2.1 Experimental Set-up for the Turning Operation

of AL1060 Alloy

Table 3, present the parameters employed for the machining analysis, of different
axial depth of cut, while the cutting speed and feed rate were kept constant. Figure 1,
shows the experimental set up of the SIEG 3/10/0016 milling machine employed
for this research work. After each machining process, the surface morphology is
measured with the surface roughness tester and the machining process was carried
out between 15 and 500 s, under the dry machining conditions.

3 Result and Discussion

The result obtained in this experimental study is presented in Figs. 1, 2, 3, 4 and

5. Figure 1 presents the effects of the coated HSS cutting tool and the uncoated
HSS cutting tool on the surface morphology of Aluminum 1060 alloy, with different
cutting time ranging from 15 to 480 s with 1 mm depth of cut during the machining
operations. From Fig. 1, it is clearly seen that there was continues increase of the
surface roughness with an increase in cutting time, for both the TiN–Zn coated
cutting tool and the uncoated cutting tool. This trend is being seen as the same in
the experimental result of varying the depth of cut from 1 to 3 mm in this study.
However, due to the significant mechanical properties of the coatings applied, there
642 I. P. Okokpujie et al.

DOC = 1 mm (HSS) DOC = 1 mm (TiN-Zn HSS)

4.06
3.96
SURFACE MORPHOLOGY μM

3.45
3.35
3.06
2.96
1.875

2.01
1.775

1.91
1.84
1.76

1.74
1.66

1.66
1.56
1.355
1.255
1.31
1.21
1.06
0.96

15 30 60 90 180 210 260 310 360 410 480

DOC = 1 mm (HSS) 1.06 1.31 1.355 1.66 1.76 1.84 1.875 2.01 3.06 3.45 4.06
DOC = 1 mm (TiN-Zn
0.96 1.21 1.255 1.56 1.66 1.74 1.775 1.91 2.96 3.35 3.96
HSS)
CUTTING TIME (SECS)

Fig. 1 Surface morphology variation versus cutting time with depth of cut of 1 mm, machining
speed 1250 rpm and feed rate 10 mm/rev for TiN–Zn coated HSS and non-coated HSS

DOC = 1.5mm (HSS) DOC = 1.5 mm (TiN-Zn HSS)

4.79
4.7
SURFACE MORPHOLOGY (μM)

3.68
3.59
3.26
3.17
2.082
1.992
1.972
1.882
1.86
1.79

1.77
1.73
1.64
1.372

1.7
1.318

1.282
1.228
1.083
0.993

15 30 60 90 180 210 260 310 360 410 480

DOC = 1.5mm (HSS) 1.083 1.318 1.372 1.73 1.79 1.86 1.972 2.082 3.26 3.68 4.79
DOC = 1.5 mm (TiN-Zn
0.993 1.228 1.282 1.64 1.7 1.77 1.882 1.992 3.17 3.59 4.7
HSS)

CUTTING TIME (SECS)

Fig. 2 Surface morphology variation versus cutting time with depth of cut of 1.5 mm, machining
speed 1250 rpm and feed rate 10 mm/rev for TiN–Zn coated HSS and non-coated HSS
Experimental Study of the Effect of TiN–Zn Coated … 643

4.81
DOC = 2 mm (HSS) DOC = 2 mm (TiN-Zn HSS)

4.73
SURFACE MORPHOLOGY (μM)

4.01
3.93
3.56
3.48
2.66
2.58
2.04
1.96
1.91
1.85

1.83
1.77
1.76
1.68
1.485
1.405
1.36
1.28
1.092
1.012

15 30 60 90 180 210 260 310 360 410 480

DOC = 2 mm (HSS) 1.092 1.36 1.485 1.76 1.85 1.91 2.04 2.66 3.56 4.01 4.81
DOC = 2 mm (TiN-Zn HSS) 1.012 1.28 1.405 1.68 1.77 1.83 1.96 2.58 3.48 3.93 4.73

CUTTING TIME (SECS)

Fig. 3 Surface morphology variation versus cutting time with depth of cut of 2 mm, machining
speed 1250 rpm and feed rate 10 mm/rev for TiN–Zn coated HSS and uncoated HSS

4.86
DOC= 2.5 mm (HSS) DOC = 2.5 mm (TiN-Zn HSS)

4.9
4.35
4.29
SURFACE MORPHOLOGY (μM)

3.76
3.7
3.06
3
2.06
1.94
1.88
1.86
1.81
1.75
1.555
1.495

1.8

2
1.43
1.37
1.101
1.041

15 30 60 90 180 210 260 310 360 410 480

DOC= 2.5 mm (HSS) 1.101 1.43 1.555 1.81 1.86 1.94 2.06 3.06 3.76 4.35 4.9
DOC = 2.5 mm (TiN-Zn
1.041 1.37 1.495 1.75 1.8 1.88 2 3 3.7 4.29 4.86
HSS)
CUTTING TIME (SECS)

Fig. 4 Surface morphology variation versus cutting time with depth of cut of 2.5 mm, machining
speed 1250 rpm and feed rate 10 mm/rev for TiN–Zn coated HSS and non-coated HSS
644 I. P. Okokpujie et al.

DOC = 3 mm (HSS) DOC = 3 mm (TiN-Zn HSS)

4.98
4.87
4.66
4.59
SURFACE MORPHOLOGY (μM)

4.06
3.99
3.43
3.36
2.26
2.19
1.96
1.91

1.89
1.85

1.84
1.78
1.558
1.488
1.472
1.402
1.115
1.045

15 30 60 90 180 210 260 310 360 410 480

DOC = 3 mm (HSS) 1.1151.4721.558 1.85 1.91 1.96 2.26 3.43 4.06 4.66 4.98
DOC = 3 mm (TiN-Zn HSS) 1.0451.4021.488 1.78 1.84 1.89 2.19 3.36 3.99 4.59 4.87

CUTTING TIME (SECS)

Fig. 5 Surface morphology variation versus cutting time with depth of cut of 3 mm, machining
speed 1250 rpm and feed rate 10 mm/rev for TiN–Zn coated HSS and non-coated HSS

were significant reductions in the surface roughness across the various depths of cut
applied in this study, from Figs. 1, 2, 3, 4, and 5. This result is in line with Fahad [16]
but it contradicted the study done by Bayraktar et al. [20] applying TiAlN coatings on
carbide cutting tool during machining Al-35Zn alloy. From the study it was observed
during the development of the said coatings, there was no critical consideration of
the coating thickness put in place so when it was deposited on the cutting tool it
caused the coated cutting tool to have a blunt surface, which definitely increases the
cutting force and the surface roughness during operations.

4 Conclusion

The study of TiN–Zn coated HSS cutting tool on surface morphology of aluminum
1060 alloys during machining operation has been done experimentally. From the
study, it has been confirmed that TiN–Zn coated HSS cutting tool when compared
with the uncoated HSS cutting tool during machining operations has significant
effects on the surface morphology of the aluminum 1060 alloy during the machining
operation. The following finding was observed
Experimental Study of the Effect of TiN–Zn Coated … 645

• The experimental result proves that cutting time has a great effect on surface
morphology, as the cutting time and the depth of cut increases there was continuous
increase in the surface morphology (roughness).
• The minimum surface morphology of 0.96 and 1.06 µm was achieved for the
TiN–Zn coated and uncoated HSS cutting tool at a cutting time of 15 s at 1 mm
depth of cut.
• From the observation, during the machining operation, the TiN–Zn coated HSS
cutting tool was able to perform at high temperature at the cutting zone without
loosening its hardness.

Acknowledgements The author wishes to appreciate the covenant university management for their
sponsorship.

References

1. Ucun İ, Aslantas K, Bedir F (2013) An experimental investigation of the effect of coating
material on tool wear in micro milling of Inconel 718 super alloy. Wear 300(1–2):8–19
2. Okokpujie IP, Ohunakin OS, Bolu CA, Okokpujie KO (2018) Experimental data-set for pre-
diction of tool wear during turning of Al-1061 alloy by high speed steel cutting tools. Data
Brief 1(18):1196–1203
3. Ezugwu CA, Okonkwo UC, Sinebe JE, Okokpujie IP (2016) Stability analysis of model regen-
erative chatter of milling process using first order least square full discretization method. Int J
Mech Appl 6(3):49–62
4. Nwoke ON, Okonkwo UC, Okafor CE, Okokpujie IP (2017) Evaluation of chatter vibration
frequency in CNC turning of 4340 alloy steel material. Int J Sci Eng Res 8(2):487–495
5. Afolalu SA, Salawu EY, Okokpujie IP, Abioye AA, Abioye OP, Udo M, Adetunji OR, Ikuma-
payi OM (2017) Experimental analysis of the wear properties of carburized HSS (ASTM A600)
cutting tool. Int J Appl Eng Res 12(19):8995–9003
6. Ogundimu O, Lawal SA, Okokpujie IP (2018) Experimental study and analysis of variance of
material removal rate in high speed turning of AISI 304L alloy steel. In: IOP conference series:
materials science and engineering, vol 413, no 1, p 012030. IOP Publishing, UK
7. Afolalu SA, Okokpujie IP, Salawu EY, Abioye AA, Abioye OP, Ikumapayi OM (2018) Study
of the performances of nano-case treatment cutting tools on carbon steel work material during
turning operation. In: AIP conference proceedings, vol 1957, no 1, p 050001. AIP Publishing,
New York
8. Afolalu SA, Abioye AA, Idirisu J, Okokpujie IP, Ikumapayi OM (2018) Abrasion wear of
cutting tool developed from recycled steel using palm kernel shell as carbon additive. Prog Ind
Ecol Int J 12(1/2):206–218
9. Okokpujie IP, Ikumapayi OM, Okonkwo UC, Salawu EY, Afolalu SA, Dirisu JO, Nwoke ON,
Ajayi OO (2017) Experimental and mathematical modeling for prediction of tool wear on the
machining of aluminium 6061 alloy by high speed steel tools. Open Eng 7(1):461–469
10. Okokpujie IP, Salawu EY, Nwoke ON, Okonkwo UC, Ohijeagbon IO, Okokpujie KO (2018)
Effects of process parameters on vibration frequency in turning operations of perspex material,
pp 700–707
11. Afolalu SA, Abioye OP, Salawu EY, Okokpujie IP, Abioye AA, Omotosho OA, Ajayi OO
(2018) Impact of heat treatment on HSS cutting tool (ASTM A600) and its behaviour during
machining of mild steel (ASTM A36). In: AIP conference proceedings, vol 1957, no 1, p
050003. AIP Publishing, New York
646 I. P. Okokpujie et al.

12. Afolalu SA, Abioye AA, Dirisu JO, Okokpujie IP, Ajayi OO, Adetunji OR (2018) Investigation
of wear land and rate of locally made HSS cutting tool. In: AIP conference proceedings, vol
1957, no 1, p 050002. AIP Publishing, New York
13. Afolalu SA, Oluwaseyifunmi O, Ongbali SO, Abioye AA, Okokpujie IP, Oloyede OR (2018)
The study of the impact of nano carbon additives on ASTM A53 mild steel during machining.
In: IOP conference series: materials science and engineering, vol 413, no 1, p 012028. IOP
Publishing, UK
14. Ghani JA, Choudhury IA, Masjuki HH (2004) Wear mechanism of TiN coated carbide
and uncoated cermets tools at high cutting speed applications. J Mater Process Technol
10(153):1067–1073
15. Huang Y, Liang SY (2005) Effect of cutting conditions on tool performance in CBN hard
turning. J Manuf Process 7(1):10–16
16. Schulz U, Peters M, Bach FW, Tegeder G (2003) Graded coatings for thermal, wear and
corrosion barriers. Mater Sci Eng A 362(1–2):61–80
17. Fahad M, Mativenga PT, Sheikh MA (2012) A comparative study of multilayer and functionally
graded coated tools in high-speed machining. Int J Adv Manuf Technol 62(1–4):43–57
18. Åstrand M, Selinder TI, Fietzke F, Klostermann H (2004) PVD-Al2 O3 -coated cemented carbide
cutting tools. Surf Coat Technol 188:186–192
19. Grzesik W, Van Luttervelt CA (2005) Analytical models based on composite layer for com-
putation of tool-chip interface temperatures in machining steels with multilayer coated cutting
tools. CIRP Ann 54(1):91–94
20. Corduan N, Himbart T, Poulachon G, Dessoly M, Lambertin M, Vigneau J, Payoux B (2003)
Wear mechanisms of new tool materials for Ti-6AI-4V high performance machining. CIRP
Ann 52(1):73–76
21. Bayraktar S, Hekimoglu AP, Turgut Y, Haciosmanoglu M (2018) A performance comparison
study of uncoated and TiAlN coated carbide end mill on machining of the Al-35Zn alloy.
In: IOP conference series: materials science and engineering, vol 295, no 1, p 012013. IOP
Publishing, UK
22. Quadrini F, Tagliaferri V, Polini R (2008) Effect of coating roughness on performance of small
CVD diamond coated tools. Mach Sci Technol 12(3):390–404
23. Qin F, Chou YK, Nolen D, Thompson RG (2009) Coating thickness effects on diamond coated
cutting tools. Surf Coat Technol 204(6–7):1056–1060
24. Heaney PJ, Sumant AV, Torres CD, Carpick RW, Pfefferkorn FE (2008) Diamond coatings
for micro end mills: enabling the dry machining of aluminum at the micro-scale. Diam Relat
Mater 17(3):223–233
25. Tamerabet Y, Brioua M, Tamerabet M, Khoualdi S (2018) Experimental investigation on tool
wear behavior and cutting temperature during dry machining of carbon steel SAE 1030 using
KC810 and KC910 coated inserts. Tribol Ind 40(1)
26. Okonkwo UC, Okokpujie IP, Sinebe JE, Ezugwu CA (2015) Comparative analysis of alu-
minium surface roughness in end-milling under dry and minimum quantity lubrication (MQL)
conditions. Manuf Rev 2(30):1–11
27. Okokpujie IP, Okonkwo UC (2015) Effects of cutting parameters on surface roughness during
end milling of aluminium under minimum quantity lubrication (MQL). Int J Sci Res 4(5):2937–
2942
28. Nwoke ON, Okokpujie IP, Ekenyem SC (2017) Investigation of creep responses of selected
engineering materials. J Sci Eng Dev Environ Technol (JOSEDET) 7(1):1–5
29. Okokpujie IP, Ajayi OO, Afolalu SA, Abioye AA, Salawu EY, Udo M (2018) Modeling and
optimization of surface roughness in end milling of aluminium using least square approximation
method and response surface methodology. Int J Mech Eng Technol (IJMET) 9(1):587–600
30. Dunmade I, Udo M, Akintayo T, Oyedepo S, Okokpujie IP (2018) Lifecycle impact assessment
of an engineering project management process—a SLCA approach. In: IOP conference series:
materials science and engineering, vol 413, no 1, p 012061. IOP Publishing, UK
31. Okokpujie I, Okonkwo U, Okwudibe C (2015) Cutting parameters effects on surface roughness
during end milling of aluminium 6061 alloy under dry machining operation. Int J Sci Res
4(7):2030–2036
Experimental Study of the Effect of TiN–Zn Coated … 647

32. Chuan SP, Ghani JA, Tomadi SH, Hassan CH (2012) Analysis of Ti-base hard coating
performance in machining process: a review. J Appl Sci (Faisalabad) 12(18):1882–1890
Experimental Investigation of Surface
Roughness in End Milling of AA6061
Alloy with Flooded Cooling
and Minimum Quantity Lubrication
(MQL) Technique

D. Nathan, D. Elilraja, T. Prabhuram, and S. Prathap Singh

Abstract This article presents an experimental investigation on the effect of surface

roughness during end milling of aluminium alloy 6061 grade in minimum quantity
lubrication (MQL) technique. The MQL technique is becoming increasingly more
popular due to the safety of the environment. The cutting speed, depth of cut, feed rate
and MQL flow rate are selected input parameters in this study. This experiment was
conducted based on the Orthogonal Array (L27 ). The comparative effectiveness of
surface roughness was investigated between MQL technique end milled components
and flooded lubrication end milled components. The findings of this study show that
surface roughness values have a significant difference upon application of the MQL
technique.

Keywords Minimum quantity lubrication · High-speed machining · Surface

roughness · Orthogonal array

1 Introduction

Nowadays, modern industries are looking for ways to reduce the use of lubricants in
metal cutting operations due to ecological and economical limitations. Lubrication
in the machining process becomes one of the most important challenges. Almost
conventional flooded lubrication with water-based lubricants in high-speed machin-
ing has a negative effect on tool life [1]. Although various lubricants are used in the
industries, unfortunately, these cutting fluids have many negative characteristics like
costly; need more quantity; they need a periodical replacement; and, lead to health
and environmental issues [2, 3]. Minimum Quantity Lubrication otherwise called
“micro lubrication” is the latest technique used for delivering coolant at the tool–
workpiece interface. In order to improve the effectiveness of high-speed machining,

D. Nathan (B)
Dr. Mahalingam College of Engineering and Technology, Pollachi, India
e-mail: nathan14576@gmail.com
D. Elilraja · T. Prabhuram · S. Prathap Singh
St. Joseph’s Institute of Technology, Chennai, India
© Springer Nature Singapore Pte Ltd. 2021 649
S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_58
650 D. Nathan et al.

MQL came into existence because it assists in lowering temperature in chip tool inter-
face as well as tool–workpiece interface and maintains the sharpness of the cutting
edges. Vegetable oil or synthetic easter oil are used instead of mineral oil which can
improve lubrication properties in MQL [4]. The exposure to flooded metal cutting
fluid may lead to adverse health effects and safety issues, including toxicity, cancer,
dermatitis, and respiratory disorders [5]. Comparative experiments were performed
to investigate machining of titanium alloy (Ti–6Al–4V) and difficult-to-cut alloys
under different lubrication conditions such as dry, MQL, and high-pressure air [6].
Although dry machining without using any cutting fluids may result in tremendous
environmental and economic benefits, it suffers a rise in temperature at workpiece tool
interface and hence reduced tool life. The goal of MQL is to use the minimum amount
of cutting fluids during machining, typically of a flow rate of 50–500 ml/h. Out of
total machining costs, 7–17% of the cost is spent for cutting fluids while 15–20%
of total machining cost is used for coolants compared to the tool cost which ranges
from 2 to 4% [7–9]. Thus minimum use of lubricants in machining can be a prime
indicator of sustainable manufacturing. End milling is one of the most widely used
machining operations in the industry because of its high MRR and good surface qual-
ity [10]. Usually, in the end milling process, abundant amounts of liquid coolant are
applied. Since liquid coolant provides intermittent cooling in end milling causes an
increase in temperature variations at the tool–workpiece interface and hence buildup
edge [11]. MQL has been effectively applied in increasing tool life while machining
components [12]. Also, the MQL technique possesses many advantages over dry,
air and flooded lubrication besides requiring some accessories and attachment like
chip removal system in machine tool [13]. MQL parameters such as lubricant flow
rate, nozzle angle, and distance between the nozzle tip and the tool cutting edge have
remarkable effects on the machining performance also in their work optimal condi-
tion of nozzle distance, MQL flow rate and nozzle angle using Taguchi Method [14].
The influence of dry, flood and minimum quantity lubrication on diameter error and
circularity was studied in aluminium, mild steel 1030 and alloy steel 4340 and found
that circularity has a significant effect [15]. Significant parameters for end milling of
AA6061T6 alloy end milling using an HSS tool under flood cooling were studied in
the statistical method and Genetic Algorithm [16]. Due to high adhesive properties
and smearing effects on the cutting tool edges, effective lubrication is required for
aluminium alloys [17]. However, very few investigations found employing MQL
into the end milling of aluminium alloys [18]. Although most of the researches have
been concerned mainly with the turning, drilling and grinding processes involved
with MQL. The role of MQL is still to be explored in end milling [19]. This leads to
further exploration of the effect of MQL in end milling of aluminium alloys. Hence
in this study, end milling of aluminium alloy is investigated by comparing the surface
roughness of the components machined under flooded cooling and MQL cooling.
Experimental Investigation of Surface Roughness in End Milling … 651

2 Materials and Methods

2.1 Materials

Aluminium alloy (AA6061) workpiece of dimensions 50 mm × 50 mm × 25 mm

is machined using “STM Manford VL 610” CNC vertical milling machine. Table 1
shows the chemical composition of AA6061 aluminium alloy. Table 2 describes the
properties of AA6061 aluminium alloy. A twist fluted PVD carbide coated HSS end
milling cutter of diameter 12.5 mm has been used as a cutting tool. End milling of
the AA6061 aluminium workpiece has been carried out at various spindle speed,
table feed, and depth of cut values. Same machining parameter combinations were
used in end milling under the flooded cooling condition and MQL condition. “ST
KOOL 74AL” Water-soluble oil was used as the flood coolant with a concentration
of 10%, emulsion (pH-9). Flood coolant was used at 2 bar pressure and 0.5 LPM
flow rate during cutting. A Portable surface tester “Mitutoyo-Surf test SJ-210” is
used for surface roughness measurement. According to ISO 2488:1996 procedure,
surface roughness measurement was carried out along the feed direction for both
flooded cooling condition and MQL condition end milling operation.

2.2 Selection of Design Matrix

In order to analyze the machining process, a large number of random experiments

have to be done. The Taguchi Method can be used to study the whole parameter
space with a minimum number of experiments. Depending on the number of control
parameters and their levels the suitable Orthogonal Array can be selected. L27 orthog-
onal array was prepared for machining parameters combinations of speed, feed and
depth of cut to study the effect of surface roughness on end milling operation. Table 3
depicts the set of cutting conditions based on L27 Orthogonal Array.

2.3 MQL Setup

A custom-designed MQL set up was fabricated for this study. Compressed air at 2 bar
pressure is passed through the FRL unit. Then at the mixing chamber, compressed
air is mixed with coolant at 6.588 l/h. A nozzle of diameter 2 mm sprayed out the
coolant air mixture at the tool–workpiece zone as shown in Fig. 1. Figure 2 describes
the arrangement of MQL setup.
 652

Table 1 Chemical composition of AA6061 by weight %

Al Si Fe Cu Mn Cr Zn Ti Each others Other total
Available 0.4–0.8 0.7 max 0.15 0.8–1.2 0.04–0.35 0.25 max 0.15 max 0.05 max 0.15 max
D. Nathan et al.
Experimental Investigation of Surface Roughness in End Milling … 653

Table 2 Properties of
Properties Value
AA6061 aluminium alloy
Density (ρ) 2.70 g/cm3
Young’s modulus (E) 68.9 GPa
Tensile strength (σ t ) 124–290 MPa
Elongation (ε) at break 12–25%
Poisson’s ratio (v) 0.33
Melting temperature (T m ) 585 °C
Thermal conductivity (K) 151–202 W/m-K
Specific heat capacity (c) 897 J/kg-K

Table 3 Machining
Exp. No. Speed (rpm) Feed (mm/min) Depth of cut
parameters
(mm)
1 700 200 0.4
2 700 200 0.8
3 700 200 1.0
4 700 500 0.4
5 700 500 0.8
6 700 500 1.0
7 700 1000 0.4
8 700 1000 0.8
9 700 1000 1.0
10 1500 200 0.4
11 1500 200 0.8
12 1500 200 1.0
13 1500 500 0.4
14 1500 500 0.8
15 1500 500 1.0
16 1500 1000 0.4
17 1500 1000 0.8
18 1500 1000 1.0
19 2000 200 0.4
20 2000 200 0.8
21 2000 200 1.0
22 2000 500 0.4
23 2000 500 0.8
24 2000 500 1.0
25 2000 1000 0.4
26 2000 1000 0.8
27 2000 1000 1.0
654 D. Nathan et al.

Fig. 1 Photograph of minimum quantity lubrication arrangement

Air FRL cool-

Compressor Unit anttank

Mixing Nozzle
Chamber

Fig. 2 Block diagram of MQL setup

3 Results and Discussions

3.1 Effect of Depth of Cut and Feed at 700 RPM

Figure 3 describes that, at 700 rpm, the highest surface roughness values were
observed at 1000 mm/min feed with flood cooling. Lowest surface roughness val-
ues were given by 200 mm/min with flood cooling. Comparing the various depth
Experimental Investigation of Surface Roughness in End Milling … 655

Fig. 3 Influence of depth of cut and feed with surface roughness at 700 rpm

of cut, the lowest depth of cut of 0.4 mm produces low surface roughness. Irre-
spective of depth of cut, almost minimum variation in surface roughness observed
in 500 mm/min feed with flood cooling. Even under flood cooling higher feed rate
results in lesser heat dissipation at the workpiece and cutting edge interface subse-
quently increases the surface roughness values. At higher feed rate with MQL, it is
evident that heat dissipation is still lower as the depth cut increases which results in
a higher gradient of surface roughness. Hence, the strong influence of flood cooling
at this speed, feed, and depth of cut combinations were observed.

3.2 Effect of Depth of Cut and Feed at 1500 RPM

Figure 4 reveals that, at 1500 rpm, lower feed of 200 mm/min MQL end milling
operation produces lowest surface roughness because of sufficient heat dissipation
with MQL was achieved at the tool chip interface. On the other hand, the higher
feed rate of 1000 mm/min with MQL gives the highest surface roughness by sticking
phenomena of aluminium chips due to insufficient exposure of lubricant at the tool–
workpiece interface. At lower feed rate, irrespective of the depth of cut it is evident
that flood cooling produces almost constant surface roughness values as the heat
dissipation and consequently low friction at tool chip interface.

3.3 Effect of Depth of Cut and Feed at 2000 RPM

From Fig. 5, at highest cutter speed and feed rate, as the temperature at the tool–
workpiece interface became high, it is evident that MQL produces the highest surface
656 D. Nathan et al.

Fig. 4 Influence of depth of cut and feed with surface roughness at 1500 rpm

Fig. 5 Graph showing influence of depth of cut and feed with surface roughness at 2000 rpm

roughness values due to insufficient flow of lubricant at the tool–workpiece interface.

Alternately, no distinct change in the surface roughness values with the influence of
MQL and flooded cooling has been observed in respect of feed rate and depth of cut.

3.4 Analysis of Variance (ANOVA)

Analysis of Variance is performed to investigate the influence and contribution of each

process parameter under MQL conditions and flooded cooling for surface roughness
in end milling of AA6061 alloy. Table 4 indicates the significant parameters and
Experimental Investigation of Surface Roughness in End Milling … 657

Table 4 ANOVA for surface roughness under MQL

Source DF Adj SS Adj MS F-value P-value % Con
Speed 2 2.1023 1.05117 33.85 0 40.70
Feed rate 2 2.8406 1.42032 45.74 0 55.09
DOC 2 0.2129 0.10645 3.43 0.052 4.13
Error 20 0.621 0.03105
Total 26 5.7769

Table 5 ANOVA for surface roughness under flood cooling

Source DF Adj SS Adj MS F-value P-value % Con
Speed 2 2.49834 1.24917 14.35 0 53.41
Feed rate 2 2.16466 1.08233 12.44 0 46.30
DOC 2 0.01307 0.00654 0.08 0.928
Error 20 1.74067 0.08703
Total 26 6.41675

contribution of each parameter under the MQL condition. From Table 4 it is evident
that feed rate is the most significant parameter with 55.09% contribution and followed
by the speed with 40.7% contribution of the total effect. The linear model expresses
that 89.25% variability in surface roughness.
Table 5 describes the result of ANOVA under flood cooling for surface roughness
with linear model prediction. From Table 5 it is observed that speed gives a major
contribution of 53.41% followed by 46.3% of contribution by feed rate in minimizing
surface roughness. The model has 72.87% variability in surface roughness.
Figure 6 describes the comparison of surface roughness values of the end milled
specimens both under minimum quantity lubrication and flood cooling for various
experiments. From the comparison graph, it is found that a considerable amount of
deviation in the surface roughness values has been noted. The trend line deviates
much in MQL cutting conditions compared with flood cooling conditions.
From literature, MQL increases tool life and reduces surface roughness in machin-
ing, but this work gives contradictory results. However, other researchers claimed
that MQL application has given poor tool life compared to flood cooling conditions
in some work tool material combinations. Many researchers argued that cooling
method or machining performance depends on workpiece tool material. This study
supports the argument and hence AA6061 aluminium alloy and carbide coated tool
worsen surface roughness in end milling operation.
658 D. Nathan et al.

Fig. 6 Comparison of surface roughness values between MQL cooling and flood cooling

4 Conclusions

End milling of AA6061 aluminium alloy with a carbide-coated end mill tool under
MQL and flooded cooling conditions were performed successfully by varying cut-
ter speed, feed, and depth of cut. Within the limitations of current research, it was
concluded that the application of MQL does not give favorable results in reduc-
ing surface roughness compared with the flooded cooling condition. The surface
roughness of AA6061 machined with MQL was higher than that obtained with flood
cooling method irrespective of the other cutting conditions. ANOVA was employed
to investigate the significant parameters for surface roughness under both MQL and
flood cooling conditions. As future work, other MQL settings such as distance from
the nozzle tip to the tool–workpiece interface, angle of impinging spray, coolant flow
rate and compressed air pressure are to be optimized.

References

1. Dhar NR, Islam S, Kamruzzaman M (2007) Effect of minimum quantity lubrication (MQL)
on tool wear, surface roughness and dimensional deviation in turning AISI-4340 steel. G.U. J
Sci 20(2):23–32
2. Rozzi JZ, Sanders JK, Chen W (2010) The experimental and theoretical evaluation of an indirect
cooling system for machining. J Heat Transf 133(3):1–10
Experimental Investigation of Surface Roughness in End Milling … 659

3. Wichmann H, Stache H, Schmidt C, Winter M, Bock R, Herrmann C, Bahadir M (2013)

Ecological and economic evaluation of a novel glycerol based biocide-free metalworking fluid.
J Cleaner Prod 43:12–19
4. Woods S (2005) Going green. Cutting Tool Eng 57(2)
5. Abdalla HS, Baines W, McIntyre G, Slade C (2007) Development of novel sustainable neat-
oil metalworking fluids for stainless steel and titanium alloy machining. Part 1—formulation
development. J Adv Manuf Technol 34:21–33
6. Liu ZQ, Chen M, An QL (2015) Investigation of friction in end-milling of Ti-6Al-4V under
different green cutting conditions. Int J Adv Manuf Technol 78:1181–1192
7. Rahman MM, Yusoff AR, Kadirgama K (2012) Investigation of flow behavior in minimum
quantity lubrication nozzle for end milling processes. Int J Autom Mech Eng 6:768–776
8. Puvanesan M, Rahman MM, Najiha MS, Kadirgama K (2014) Experimental investigation of
minimum quantity lubrication on tool wear in aluminum alloy 6061-T6 using different cutting
tools. Int J Autom Mech Eng 9:1538–1549
9. Najiha MS, Rahman MM (2015) Experimental study on minimum quantity lubrication in end
milling of AA6061-T6 using TiAlN coated carbide tools. Int J Autom Mech Eng 11:2771–2785
10. Lakshmi VVK, Subbaiah KV (2012) Modelling and optimization of process parameters during
end milling of hardened steel. Int J Eng Res Appl 2:674–679
11. Boswell B, Voges E (2012) The effect of combined cold air and minimum liquid cooling on
end milling. Adv Mech Eng 2(1):1–7
12. Boubekri N, Shaikh V, Foster PR (2010) A technology enabler for green machining: minimum
quantity lubrication (MQL). J Manuf Technol Manage 21:556–566
13. Kadu MB, Hambire U (2017) Review on MQL in green environments used for machining
process. Int J Eng Res Appl 7(10:3):23–26
14. Safiei W, Rahman MM, Rusdan SA (2018) Experimental investigation of MQL optimum
parameters in end milling of AA6061-T6 using Taguchi method. Int J Eng Technol 7(4.35):186–
189
15. Islam MN (2013) Effect of additional factors on dimensional accuracy and surface finish of
turned parts. Int J Mach Sci Technol 17(1). https://doi.org/10.1080/10910344.2012.747936
16. Najiha MS, Rahman MM, Yusoff AR (2013) Modeling of the end milling process for aluminum
alloy AA6061T6 using HSS tool. Int J Autom Mech Eng 8:1140–1150
17. Tosun N, Huseyinoglu M (2010) Effect of MQL on surface roughness in milling of AA7075-T6.
Mater Manuf Process 25:793–798
18. Boswell B, Islam MN (2013) The challenge of adopting minimal quantities of lubrication for
end milling aluminium. In: Yea G-C (ed) Lecture notes in electrical engineering, IAENG
transactions on engineering technologies. Springer Science + Business Media Dordrecht,
Australia
19. Boswell B, Voges E (2011) The effect of combined cold air and minimum liquid cooling on end
milling. In: Ma M (ed) International conference on mechanical, industrial, and manufacturing
engineering. Information Engineering Research Institute, Melbourne, Australia, pp 54–57
Material Joining Technology
Performance Study of Dissimilar Alloy
Joints of SS321 and SS347 Under MIG
Welding Process

S. Nishanth, S. Prem, T. Prakash, J. Siva, M. Sathish Kumar,

and N. Siva Shanmugam

Abstract An attempt is made in this research work to study the influence of filler
wire materials on the dissimilar alloys between SS321 and SS347 are made using
Robotic MIG welding process Preliminary experimental welding trials are conducted
by varying the MIG welding process parameters like current, welding speed and stick
out the distance. The prepared metal coupons of two different stabilized stainless
steels are welded using ER321 as well as ER347 filler wires. The performance of
the joint is studied by mechanical testing and microstructural characterization. The
results of mechanical testing showed that the welded joint with ER347 filler wire
exhibited improved tensile strength and hardness value than the joint made by ER321
filler wire.

Keywords SS321 · SS347 · Dissimilar metal welding · MIG

1 Introduction

The highest demand for enhanced material performance in both military and civilian
sectors is the main driving force for advances in new alloys and technologies for
newer designs. In this work type 321 stainless steel, an austenitic alloy is stabilized
with titanium, making it resistant to intergranular attack caused by chromium car-
bide precipitation is joined using gas metal arc welding with dissimilar 347 which
resembles the same chemical composition using two different filler wires ER321 and
ER347 the specimens are prepared and welded in specific conditions and tested for
mechanical properties along with microstructure.

S. Nishanth · S. Prem · T. Prakash · J. Siva · M. Sathish Kumar (B)

Department of Mechanical Engineering, SNS College of Engineering, Coimbatore, Tamil Nadu,
India
e-mail: Sathishmfg14@gmail.com
N. Siva Shanmugam
Department of Mechanical Engineering, National Institute of Technology Tiruchirappalli,
Tiruchirappalli, Tamil Nadu, India

© Springer Nature Singapore Pte Ltd. 2021 663

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_59
664 S. Nishanth et al.

It is identified that stabilized austenitic stainless steel such as 347 and 321 are
widely used in components designed for high-temperature applications like nuclear
reactors, boilers, superheaters and chemical reactors [1]. This material can be selected
for high temperatures services due to its high creep and intergranular corrosion
resistance. However, for services into the sensitization range of temperatures (450–
850 °C) [2]. Austenitic stainless steel type AISI 321, has an incredible resistance to
corrosion and oxidation [3].

2 Experimental Procedure

2.1 Material Selection

The grade SS347 is chosen as one of the dissimilar parent alloys which are stabilized
austenitic stainless steel with the addition of niobium. The intermittent temperature
is about 800–1650 °F (427–897 °C). It is nonmagnetic. The grade SS321 is another
alloy to be welded with 347 which is a stabilized austenitic stainless steel which is
similar to SS304 but with a titanium addition of at least five times the carbon content.
The intermittent temperature is about 800–1500 °F (427–816 °C). It is normally
magnetic. This grade is equivalent to SS304 in the annealed condition and stronger
if weldments in these grades have not subjected to annealing or if this grade is put
into general service in the 420–900 °C range. Table 1 gives the material composition
of the alloys.

2.2 Applications of These Dissimilar Alloys

This alloy has been extensively used in Expansion joints Spiral Welded tube for burner
pipes and flues, Bellows, Woven or welded screens for high-temperature mineral
processing Furnace parts, Heating element tubing, Heat Exchangers. Due to the
above-said reasons, their weld characteristics have been studied. Stabilized austenitic
stainless steel (SS) such as type 347 and 321 are widely used in components designed
for high-temperature applications like nuclear reactors, boilers, superheaters and
chemical reactors [1].

2.2.1 Experiment

The experiment is carried out in robotic gas metal arc welding machine shown in
Figs. 1 and 2. The specifications of the GMA welding setup is DM 350 MIG power
source of OTC Daihen. GMA welding torch was attached with a 6 axis OTC Daihen
robot. The experimental study denotes that MIG weld process provides higher quality
Table 1 Composition of 321/347 grades
Grade C Mn Si P S Cr Mo Ni N Other
321 Min – 2.00 0.75 0.045 0.030 17.0 – 9.0 0.10 T=5(C+N)
Max 0.08 – – – – 19.0 – 12.0 – 0.70
347 Min 0.08 2.00 0.75 0.045 0.030 17.0 – 9.0 0.10 Nb=10(C+N)
Performance Study of Dissimilar Alloy Joints of SS321 …

Max – – – – – 19.0 – 13.0 – 1.0

665
666 S. Nishanth et al.

Fig. 1 Robot setup

Fig. 2 Welding process

Performance Study of Dissimilar Alloy Joints of SS321 … 667

welds in a wide variety of metal and alloys. Therefore, it is most commonly used to
join stainless steel and other metals. The helium and argon are shielding gases that
are used to protect the welding zone from the atmosphere [4].
The specimen preparation is carried out to get sound welding joints.
The specimen size is 100 × 50 × 6 mm (two pieces). The butt weld joint is
preferred. The specimen is set to resist the degrees of freedom in a welding fixture
available. The welding is carried out with parameters arrived from the previous
literature. The welding current is chosen as 140 A for welding. The filler wire of two
types one is ER321 and ER347 is taken. The specimens were welded using ER321
and ER347 separately and subject to mechanical test.

3 Results and Discussion

3.1 Tensile and Hardness Properties

The tensile tests were carried out at room temperature, specifically, the methods of
determination of yield strength, ultimate tensile strength, elongation, and reduction
of area. The specimens were prepared according to ASTM E8. The tests were carried
out in x-direction and with Tinius Olsen H50KL tensile testing machine. The Fig. 3
shows the specimen size as per the standards.
Some of the literature findings reveal that the tensile test sample fractures occurred
at the base metal of SS321 austenitic stainless steel, which shows that the strength
of the welded joint was better than that of base metal [5].
The tensile test provides data on the strength and ductility of materials under
uniaxial tensile stresses. The results of the tensile test, which was conducted for
dissimilar welds of low alloy steel to austenitic stainless steel, showed that fracture
happened in the HAZ of low alloy steel [6].
In this work, it is studied that the tensile strength of the base metals SS321 and
347 is about 515 MPa. After welding the dissimilar combination, the specimens were
subjected to tensile test. The results of the tensile tests were shown in Table 2. The
comparison is to find the best suitable filler wire for the dissimilar alloy combinations

Fig. 3 Test specimen for tensile test

668 S. Nishanth et al.

Table 2 Tensile test results

S. No. Filler/welding Ultimate stress Yield stress (MPa) Total elongation
current (MPa) (%)
1 ER321/140 A 330 255 17.6
2 ER347/160 A 363 343 20.9

Fig. 4 Stress–strain curve for specimen welded with ER347

Fig. 5 Tensile test

having high tensile strength. It is found that the weld specimen of the combination
of filler wire 347 and welding current 160 A has high tensile strength.
The stress–strain curve is shown in Fig. 4 for the weld specimen welded with filler
ER347. The ultimate tensile strength is obtained at 18% strain and the upper yield
point occurs at 13% of strain. The specimen before testing and after testing is shown
in Fig. 5.

3.1.1 Hardness Test

The hardness measurements were performed to determine the strength along with
hardness. Vickers hardness provides medium indentations. The specimen is subjected
Performance Study of Dissimilar Alloy Joints of SS321 … 669

Fig. 6 Vickers hardness test 300 289

250 228 204.9
200
150
SS347
100
50
0
SS321 SS347 WELD

to Vickers hardness test. Hardness is lower in the HAZ region compared to the weld
metal and base metal regions from welding technique [7].
Figure 6 shows the values obtained in the hardness test. Vickers scale is considered.
The specimen welded with ER347 has hardness value of 289 HV and the specimen
welded with ER321 has 257 HV in the welded region. The base metal always gave
the highest hardness value for all the welding parameters [8].

3.2 Microstructure Properties

The microstructure properties were studied using SEM analysis. The Fig. 7 shows a

Weld zone

Fig. 7 SEM images

670 S. Nishanth et al.

uniform flow of filler material and good weld joint. The scanning electron microscopy
was used for the examination of weld microstructure. In the interdendritic region
where the small white precipitates which were sometimes segregated clearly from
the field were observed [9].

4 Conclusion

The present investigation on dissimilar alloy deduces the following conclusions.

The butt welding joint is best suitable for testing tensile strength. The filler wire
ER321 and ER347 are used to weld the joint which improves strength and hardness
to a possible extent. The tensile strength of a specimen with ER347 is higher than
the specimen welded with ER321. The hardness tested with specimen (welded with
ER347) is more when compared to (ER321 filler wire) weld specimen. The neck
formation and failure of specimen occur at the ultimate strength of 363 MPa in the
specimen welded with ER347. The neck formation and failure of the specimen occur
at ultimate strength of 330 MPa in the specimen welded with ER321. The hardness
of the weld joint is more than the base metal signifies the lesser tendency to abrasion.
The results of mechanical testing showed that the welded joint with ER347 filler wire
exhibited improved tensile strength and hardness value than the joint made by ER321
filler wire. This is due to the fact that the presence of intermetallic compounds, i.e.
NbC (revealed during metallographic study) in the weld metal is observed to be more
in joints produced with ER347 filler wire. In the SEM analysis, the specimen welded
with ER347 resembles a more uniform grain structure than the specimen welded
with ER321. In the future, the welding current and speed may be optimized along
with filler wire to improve joint properties.

References

1. Chandra K, Kain V, Tewari R (2012) Microstructural and electrochemical characterisation of

heat-treated 347 stainless steel with different phases. Corros Sci 67:118–129
2. Moura V, Kina AY, Tavares SS, Lima LD, Mainier FB (2007) Influence of stabilization heat
treatments on microstructure, hardness and intergranular corrosion resistance of the AISI 321
stainless steel. J Mater Sci 43:536–540
3. Nakhodchi S, Shokuhfar A, Iraj SA, Thomas BG (2015) Evolution of temperature distribu-
tion and microstructure in multipass welded AISI 321 stainless steel plates with different
thicknesses. J Pressure Vessel Technol 137
4. Kang BY, Prasad YK, Kang MJ, Kim HJ, Kim IS (2009) The effect of alternate supply of
shielding gases in austenite stainless steel GTA welding. J Mater Process Technol 209:4722–
4727
5. Mohan Kumar S, Siva Shanmugam N (2018) Studies on the weldability, mechanical properties
and microstructural characterization of activated flux TIG welding of AISI 321 austenitic
stainless steel, vol 5, no 10. IOP Publishing Ltd Materials Research Express
Performance Study of Dissimilar Alloy Joints of SS321 … 671

6. Arivazhagan N, Surendra Singh S, Prakash S, Reddy GM (2011) Investigation on AISI 304

austenitic stainless steel to AISI 4140 low alloy steel dissimilar joints by gas tungsten arc,
electron beam and friction welding. Mater Des 32:3036–3050
7. Lakshminarayanan K, Shanmugam K, Balasubramanian V (2009) Effect of welding processes
on tensile and impact properties, hardness and microstructure of AISI 409M ferritic stainless
joints fabricated by duplex stainless steel filler metal. J Iron Steel Res Int 16:66–72
8. Gulenc B, Develi K, Kahraman N, Durgutlu A (2005) Experimental study of the effect of
hydrogen in argon as a shielding gas in MIG welding of austenitic stainless steel. Int J Hydrogen
Energy 30(13–14):1475–1481
9. Hajiannia I, Shamanian M, Kasiri M (2013) Microstructure and mechanical properties of AISI
347 stainless steel/A335 low alloy steel dissimilar joint produced by gas tungsten arc welding.
Mater Des 50:566–573
10. Green G, Higginson R, Hogg S, Spindler S, Hamm C, Najorka J (2015) Analysis of ferrite
formed in 321 grade austenitic stainless steel. Mater Sci Technol 31:418–425
Effect of Friction Stir Welding Tool
on Al–SiC Composites by Varying Tool
Pin Profile and Tool Material

P. Jayaseelan , T. V. Christy , and S. J. Vijay

Abstract Efficient joints can be accomplished through the friction stir welding
(FSW) process with the selection of appropriate tool profile, tool material and D/d
ratio. This experiment to weld Aluminium 6061 with 10% silicon carbide reinforce-
ment was carried out in a CNC FSW machining set-up, the tool materials being H13
and OHNS, the tool profiles plain pentagon, plain taper and threaded cylinder and the
D/d ratio, a common 2. The different welding parameters chosen were 1000 rpm tool
rotation speed and 25 mm/min travel speed with an axial load varying between 3.8
and 4.3 kN. The joint produced with the threaded cylinder tool, compared with the
rest of the tools, exhibited the maximum tensile strength of 150 MPa and hardness
of 41 HRC and was found to be defect-free with a uniform grain size distribution.

Keywords FSW · Profile · Material · D/d ratio · Al–SiC · Tensile strength ·

Hardness

1 Introduction

Friction stir welding was first developed in December 1991 by “the Welding Institute”
(TWI). Friction stir welding procedure is a process of solid-state connection without
any external filler. This entire process is carried out in the CNC machine setup.
Large alloys, like highly alloyed 2xxx and 7xxx series, are not weldable because
their microstructure and porosity in the area of melting are poorly solidified. This
entire process is carried out in the CNC machining setup [1]. When there are high

P. Jayaseelan · S. J. Vijay
Department of Mechanical Engineering, School of Engineering and Technology, Karunya
Institute of Technology and Sciences, Coimbatore, India
e-mail: jeevaglory@gmail.com
S. J. Vijay
e-mail: vijayjoseph.2001@gmail.com
T. V. Christy (B)
Professor of Mechanical Engineering, PRIST, Vallam, India
e-mail: drchristytv@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 673

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_60
674 P. Jayaseelan et al.

levels of thermal stress and solidification of shrinkage during the welding process,
Aluminium is susceptible to Solidification cracking or hot cracking can be found
out. FSW has several benefits in the welding process which can be listed: FSW
does not need a joint preparation for plates, the result is high-quality welding with
increased tensile strength, expectation fatigue and corrosion resistance [2]. In order to
improve material mechanical properties, combining with various alloying elements
shows high performance, and heat treatable alloys. In this experimental process, we
used Aluminium Silicon Carbide alloy as a weldment. The four major parameters
determine the quality and strength of joint are mainly classified into (i) Speed of the
Rotation, (ii) Traverse speed, (iii) Plunging depth, (iv) Tilt angle. Then, the selection
of appropriate tool profile, tool material, tool geometry is risky and challenging one
[3]. By determining the proper ratio for the shoulder and pin, the tool was selected
[4]. Palin pentagon, Plain Tapering and Threaded cylinder with various instruments
are the tool profiles used in this experimental work. After the welding process once
completed, the scans of the weldment are important to check the different zones and
its welded quality. The areas discussed here are (i) Base Material (BM) or Parent
Zone (PZ) (ii) Heat affected zone (HAZ) (iii) Thermo-Mechanically affected zone
(TMAZ) (iv) Weld zone (WZ) or Nugget Zone (NZ). In this experimental work, we
finally discussed the various Tensile strength and hardness for the welded specimen
in different tool profiles and the values are tabulated, and the micrographs for the
weldment zones are shown.

2 Experimental Procedure

The experiment was initially started by selecting the proper tool profiles which
namely Plain Pentagon, Plain tapered, threaded cylinder tool pin and the material
used for the tools are H13, OHNS are selected for use. Hence the D/d ratios for the
tool profiles are listed in Table 1. Hence, the listed tool profiles are allowed to weld
in the Al–SiC weldment are measured the tensile strength and hardness values are
compared. The tools used in this experiment are shown in Fig. 1. The experiment was
therefore performed with a composite aluminium metal matrix. Al-10% SiC with the
dimensions 100 mm × 50 mm × 6 mm is produced under the in situ casting process
with the composition of the metal plates. As we discussed before, the composition
of the Aluminium alloy with other alloys possess greater mechanical strength [5]. If
the material composition is balanced, then the defect-free weld can be obtained [6].
The Al-10% SiC MMC microstructure is displayed in Fig. 2.

Table 1 FSW tools

Tool No. Tool profile Tool material D/d
descriptions
1 Plain pentagon H13 2
2 Plain tapered OHNS 2
3 Threaded cylinder H13 2
Effect of Friction Stir Welding Tool on Al–SiC Composites … 675

Fig. 1 Al-10% SiC MMC

optical micrograph

Fig. 2 Fabricated FSW tools

Then, the proper selection of the tool material is important that the tool should
withstand such mechanical properties [7]. The selection of the material is important
so that the good quality of the weld can be achieved.
H13 and OHNS with proper tool geometry are the tools used in this experiment
which have a strong temperature resistance to thermal fatigue and resistance to wear.
The parameters used in this experiment are (1) 1000 rpm tool rotation speed; (2)
25 mm/min traversing speed; (3) 3.8–4.3 kN axial load. The parameters are fixed after
we did some trial experiments which yielded defect-free weld [8]. The parameter used
676 P. Jayaseelan et al.

Table 2 Welding parameters

Tool rotational Traverse speed Axial load (kN)
speed (RPM) (mm/min)
1000 25 3.8–4.3

in this experiment is tabulated as shown in Table 2. Then the experiment was followed
with this fixed parameter for all the tool profiles, then the results are discussed. The
connection between the number of revolutions and the speed of soldering directly
affects the toughness of the fracture and the energy needed to initiate and propagate
the crack [9].

3 Results and Discussion

Specimens are welded with above-mentioned parameters and it was prepared for
testing the hardness and tensile strength. The test was conducted as per ASTM E08
standard. Table 3 shows the tensile strength of the welded Aluminium Silicon car-
bide (Al–SiC) is welded with different tool profiles (Plain Pentagon, Plain Tapered,
Threaded Cylindrical). And it was shown that Tool number 3 (Threaded Cylinder—
H13) yields high tensile strength than the other tool profiles. And it is clearly stated
that it yields 150 MPa, 118 MPa, 127 MPa with respect to Threaded cylinder—H13,
Plain Tapered—OHNS, Plain Pentagon—H13 respectively. As we know that it yields
high tensile strength only if the grains are evenly distributed (uniformly). That the
condition is satisfied in the case of Threaded Cylinder—H13 and its microstructure
are given in Fig. 3. The welding temperature and welding speed are mainly deter-
mined for the UTS of the friction stir welds in EN AW-2219-T87 [10]. Other than the
H13, Threaded cylinder, tools profile, and materials made like H13—plain pentagon
and OHNS—plain tapered couldn’t even uniformly distribute than the H13 Threaded
cylinder profile. So, the other two tools yield slightly lower than the H13—Threaded
Cylinder profile. The optimal combination of process parameters for the high-speed
steel (HSS) tool [11].
After the test, the maximum hardness for the welding specimens (Al–SiC) was
determined by a hardness test. The hardness value for the welded specimen is shown
in Table 4. It is clearly stated that again H13—Threaded Cylinder shows the maxi-
mum hardness value of 41 HRC compared to OHNS—Plain tapered and H13—Plain

Table 3 Tensile strength of

Tool No. Tool profile Tool material Tensile strength
weldments
(in MPa)
1 Plain pentagon H13 127
2 Plain tapered OHNS 118
3 Threaded H13 150
cylinder
Effect of Friction Stir Welding Tool on Al–SiC Composites … 677

a) BM b) HAZ c) TMAZ

10μm 10μm 10μm

Considerable d) WNZ
grain
refinement
with the
precipitates 10μm

e) TMAZ f) HAZ g) BM

10μm 10μm 10μm

Fig. 3 Composite FSW micrographs using the threaded cylinders tool H13

Table 4 Hardness of
Tool No. Tool profile Tool material Hardness (in
weldments
HRC)
1 Plain pentagon H13 40
2 Plain tapered OHNS 26
3 Threaded H13 41
cylinder

Pentagon of 26 HRC and 40 HRC, respectively. Hence, the Mechanical characteri-

zation results that H13—Threaded cylinder profile yielded high strength. Hardness
increases during welding at a high tool rotation and reduces wear [12]. The tool
material does contribute significantly to joint characteristics [13].
In current work, Al 10% SiC using the H13—Threaded cylinders show, according
to the metallurgical characterization, that in the Weld zone microstructure in Fig. 3(d)
the uniform distribution of particles is clearly indicated. Through examining of the
weld zone, the uniform distribution of particles with a defect-free weld is obtained
for H13—Threaded cylinder than the other two tool profiles.
As we know, metallurgical characterization is divided into four major categories.
This includes the Weld Zone (WZ), THM Zone (TMAZ), the Thermal Area (HAZ),
the THM Zone (NZ), or BA (BM) Zones. Therefore Al–SiC composite welded
678 P. Jayaseelan et al.

Table 5 Macrostructure of weldments

Tool profile Macrograph of FSW joint

Plain pentagonal 4.2 kN, 25 mm/min, 1000 rpm

H13
Tool No. 1

Cylindrical tapered 3.5 kN, 25 mm/min, 1000 rpm

OHNS
Tool No. 2

Threaded cylinder 4.3 kN, 25 mm/min, 1000 rpm

H13
Tool No. 3

micrographs are shown in Fig. 3. The grain structure of the HAZ is well known to be
almost the same as base metal (BM). Since, when compared to Thermo-Mechanically
Affected zone (TMAZ) grain size are smaller than the other grains due to its plastic
deformation near the weld zone (WZ). The heat is allowed to pass through the TMAZ.
The welding zone with a maximum heat generation is thus smaller than TMAZ and
BM in grain size. In different areas of the welded joint, the asymmetry of the sold
joints and the changes in metallurgical transformations that occur around the pine
and below the tool during their joint movement affect the effect strength value [14].
Also, the macrostructure of the Welded zone of Al–SiC is shown along with the
Tensile Strength (in MPa) as shown in Table 5. As we see that, weldment made
using the Tool No 3 has exhibits defect-free joint and yield high tensile strength with
desired hardness.

4 Conclusion

Significant difficulties in achieving high-quality welds are the right choice for the
FSW tool. As already stated, the material, pin profile and ratio of D/d affect soldering
quality significantly and therefore require a thorough investigation. Detailed investi-
gations were carried out using tools of various profiles, such as friction stir welding
aluminium metal matrix composites (Al–SiC). Simple flap, cylinders, materials such
as simple tapering and threading. FSW, OHNS and H13 and D/d ratio 2 are allowed
on commercial CNCs. The results show that Al-MMC joints can produce superior
Effect of Friction Stir Welding Tool on Al–SiC Composites … 679

tensile strength and durability in a threaded cylinders H13 tool with a ratio of D/d
2. The ongoing work in the production of quality welds using the FSW process is
considered an important contribution.

References

1. Barath VR, Vignesh RV, Padmanaban R (2018) Analysing the strength of friction stir welded
dissimilar aluminium alloys using Sugeno fuzzy model. IOP Publishing, Coimbatore
2. Bosneag A, Constantin MA, Nitu E, Iordache M (2018) Friction stir welding of three dissimilar
aluminium alloy: AA2024, AA6061 and AA7075. IOP Publishing, Romania
3. Palani K, Elanchezhian C, Avinash K, Karthik C, Chaitanya K, Sivanur K, Reddy KY (2018)
Influence of friction stir processing parameters on tensile properties and microstructure of dis-
similar AA8011-H24 and AA6061-T6 aluminium alloy joints n nugget zone. IOP Publishing,
Kanchipuram
4. Jayaseelan P, Christy TV (2019) Behaviour of threaded cylinder tool in the friction stir welding
of Al-ZrB2 and Al-SiC composite. Int J Mech Eng Technol 10(1):1751–1761
5. Daneji A, Ali M, Pervaiz S (2018) Influence of tool geometry and processing parameters on
welding defects and mechanical properties for friction stir welding of 6061 aluminium alloy.
IOP Publishing, Dubai
6. Chand S, Vineetha S, Madhusudhan D, Krishna CS, Devi GK, Hemarao K, Naidu GG (2018)
Effect of FSW welding speed on microstructure and microhardness of Al-0.84Mg-0.69Si-
0.76Fe alloy at moderate rotational tool velocity. IOP Publishing, Andhra Pradesh
7. Jayaseelan P, Christy TV, Rubesh, Prabhu G, Srinivasan R (2019) Behaviour of threaded
cylinder tool in the friction stir welding of Al–ZrB2 composite. J Comput Theor Nanosci
16(2):719–724
8. Vijayan D, Abhishek P (2018) Multi objective process parameter optimization of friction stir
welding using NSGA-II. IOP Publishing, Kanchipuram
9. Jayaseelan P, Christy TV (2019) Behavior of threaded cylinder tool in the friction stir welding
of Al-SiC composite. In Materials today: Proceedings (Vol. 16)
10. Zhang H, Zhao C, Cui B, Yan X, Dou C, Shao T, Pan R (2018) The corrosion behaviour of cold
spray coating on 2219 aluminium alloy joints prepared by friction stir welding. IOP Publishing,
Beijing
11. Milcic M, Vuherer T, Radisavljevic I, Milcic D, Kramberger J, Andjelkovic B (2018)
Mechanical behavior of Al 2024 alloy welded by friction stir welding. IOP Publishing, Beograd
12. Chernykh IK, Vasil’ev EV, Matuzko EN, Krivonos EV (2018) Upgrading weld quality of a
friction stir welded aluminium alloys AMG6. IOP Publishing, Omsk
13. Constantin MA, Bosneag A, Nitu E, Iordache M (2018) Orientation of process parameter values
of TIG assisted FSW of copper to obtain improved mechanical properties. IOP Publishing,
Romania
14. Jayaseelan P, Christy TV, Vijay SJ, Nelson R (2019) Effect of tool material, profile and D/d ratio
in friction stir welding of aluminium metal matrix composites. Materials Research Express,
IOP Science Publishing
Fabrication of AISI 304 Austenitic
Stainless Steels with TiN Addition Using
Spark Plasma Sintering Method

Babatunde Abiodun Obadele, Ramokone Marcia Mafafo,

Walter Roberto Tuckart, and Peter Apata Olubambi

Abstract The spark plasma sintering (SPS) is a type of electric discharge sintering
technique used to consolidate metallic/ceramic powders. In this study, spark plasma
sintering was used to fabricate 304 stainless steel with TiN addition. The best param-
eters that influence the sintering process such as sintering temperature, heating time
and pressure were investigated. Results show that by varying the amount of TiN,
the sintered properties of the composites could be significantly affected. The relative
density decreased with the increment of TiN. Microhardness values vary from 270
to 350 HV0.1.

Keywords Spark plasma sintering · 304 stainless steel · Density · Hardness

1 Introduction

In modern industrial applications, engineering materials are being subjected to harsh

demanding environments that require materials to possess good mechanical proper-
ties, good oxidation and corrosion resistance. Engineering materials such as titanium,
aluminium and stainless steels are finding wider application in such environment
[1]. Austenitic stainless Steels (ASS) type 304 is widely used in these engineering
environments. 304 ASS has excellent corrosion resistance properties in various envi-
ronments; as a result, austenitic stainless steels find application in various industrial
application which include thermal power plants, power generation plants, marine,

B. A. Obadele (B) · R. M. Mafafo · P. A. Olubambi

Department of Chemical Engineering Technology, University of Johannesburg,
Doornfontein Campus, Johannesburg, South Africa
e-mail: obadele4@gmail.com
B. A. Obadele
Department of Chemical, Materials and Metallurgical Engineering, Botswana International
University of Science and Technology, Palapye, Botswana
W. R. Tuckart
Tribology Group, Universidad Nacional del Sur/CONICET, Av. Alem 1253,
8000 Bahía Blanca, Argentina

© Springer Nature Singapore Pte Ltd. 2021 681

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_61
682 B. A. Obadele et al.

urban water, chemical, aircraft gas turbine engines and food industries [2]. Apart from
corrosive environment application, high impact velocity of particles is frequently
encountered in components used in these industries. Examples are coal-fired boilers
used in the power generation industries [3]. Although austenitic stainless steels are
well known for their good corrosion properties, they are famous for poor mechanical
properties such as friction and wear. According to Sun and Bell [4], when ASSs are
in contact with each other or other materials, they experience severe metallic wear
which results from the adhesion junction and severe plastic deformation occurring
between the surfaces of the materials in contact with each other. Furthermore, ASS
fails mainly because of wear rather than corrosion-related material loss problems.
According to More et al. [5], 50% or more investigators focuses on steel as their
target material because steel is used in many industrial applications, however, it
has been noticed that steel material can be effectively replaced or alloyed with some
other materials such as titanium. Mainly because titanium is the second-largest target
material and they have good wear properties. Titanium nitride (TiN), which has
relatively good wear properties, is intentionally alloyed to 304 ASS to improve the
hardness and wear properties of the resulting material.
There are several methods that can be used to consolidate 304L ASS with TiN.
Such methods include hot isostatic pressing (HIP), hot pressing (HP), microwave
sintering and spark plasma sintering (SPS) [6, 7]. This study focuses on ASS grade
304 reinforced with TiN under high-temperature conditions.

2 Experimental Procedures

Commercial as-received 304-austenitic stainless steel and TiN were chosen as the
starting powders. The two feedstock powders were characterized for particle size
distribution (PSD) using Microtrac S3500 analyzer. It uses the three red laser diodes
(tri-laser) that provides accurate and repeatable particles sizes [8]. Powders were
weighed according to the proportions as shown in Table 1. The weighed powders
were then mixed (under dry environments) using Turbula Shaker Mixer T2F for 8 h at
a speed of 72 rpm. The function of mixing is to achieve uniformly blended particles
without altering the properties required for the next steps [9]. Admixed powders
were then characterized using SEM-EDX and the phases formed were analyzed
using XRD. The admixed powders were also characterized by PSD.

Table 1 Chemical
Number Powder Weight ratio
compositions of admixed
powders 1 304-ASS 100
2 304-ASS/2TiN 98:02
3 304-ASS/4TiN 96:04
4 304-ASS/6TiN 94:06
Fabrication of AISI 304 Austenitic Stainless Steels … 683

Table 2 Spark plasma

Sintering conditions Weight ratio
sintering conditions
Sintering temperature (°C) 1100
Holding time (min) 10
Heating rate (°C/min) 100
Vacuum pressure (MPa) 50

The admixed powders were then consolidated using spark plasma sintering tech-
nique in which specimens with dimensions of 35 mm × 1.5 mm were produced.
The substrates were machined using the robofil 240-SL electrical discharge machine
(EDM) wire cut. Scanning electron microscopy (SEM, Tescan Vega 3) which has the
energy-dispersive X-ray spectrometer (EDS) attached to it was used to analyze the
morphology of the feedstock and admixed powders to confirm the element present
in the powders. Phase identification of the powders was also analyzed with X-ray
diffraction (XRD Rigaku Ultima IV). Admixed powders were sintered using spark
plasma sintering technology (H-HPD25-FCT Systeme GmbH Germany) in a 40 mm
graphite die and the thickness between the graphite foil (used to separate the graphite
die and the punches) and the powder was 6.5 mm. The densification process took
place at varying operating parameters under a heating rate of 100 °C/min from room
temperature up to the desired temperature (1100 °C) and holding time of 10 min. The
temperature was monitored using the optical pyrometer throughout the SPS process.
A constant vacuum pressure of 50 MPa was applied in each test. The sintering param-
eters used are presented in Table 2. After sintering, the samples were sand-blasted
to remove any excess graphite material on them. Furthermore, the densities of the
sintered samples were accurately weighed using the Archimedes for five times using
Mittler electronic balance and the average densities were recorded.
The sintered samples were ground and polished to 3 μm diamond finish. The sam-
ples were further polished with fumed silica using a polishing suspension of 0.2 μm to
achieve a mirror-like surface. The morphology of the sintered specimen was investi-
gated using scanning electron microscope model TESCAN Vega 3 which is equipped
with Energy-dispersive spectroscope. The Rigaku Ultima IV X-ray diffractometer
XRD equipped with a graphite-monochromatic Cu Kα radiation source at 40 kV and
30 mA was used to perform the XRD analysis. Collection of diffractograms took
place in the 2θ range between 3° and 90° with a step size of 0.01 [8]. The purpose
of the XRD analysis is to identify the phases present/formed in the composites. The
effect of adding TiN on the hardness of the resulting composite was carried using
the INNOVATEST Falcon 500 Vickers microhardness tester. The specimens were
indented under a load of 100 g and dwell time of 10 s on 10 different positions. The
average values were calculated and recorded.
684 B. A. Obadele et al.

3 Results and Discussion

The SEM morphologies and EDS analysis of the starting 304 ASS and TiN powders
used in this study are given in Fig. 1. From Fig. 1a, it could be seen that 304 ASS
particles are spherical in shape, smooth and non-porous with a few satellites. This
is typical for powders that were produced by gas atomization [10]. From Fig. 1b, it
could be seen that the TiN powder show an irregular shape and it is porous. The TiN
powder could have been produced by the sol-gel method. This method is used to
produce metal-oxides or metal chlorides (usually known as precursors), especially
that of silicon and titanium. The morphology of particles produced by this technique
is porous. EDS reveals Fe, Cr and Ni in the 304 ASS powder as the major elements
and Ti and N as the main elements in the TiN starting powder.
The XRD spectra of as-received 304 ASS and TiN powders are given in Fig. 2. In
Fig. 2a, the XRD pattern shows the phases present in the starting 304 ASS powder.
The phases that have emerged contain elements such as Cr, Ni, and Fe. Similarly, in
Fig. 2b, the diffractometer reveals that there is only TiN present in the powder. The
XRD results of the starting powders obtained in this study are in line with the EDS
analysis.
The particle size distribution curves are presented in Fig. 3. The PSD graph repre-
sents the amount of materials in percentage passing through a specific screen size in
microns. From the graph, it could be observed that for pure 304 SS, 304 SS/2TiN, 304

a b

Fig. 1 SEM morphologies of a as-received 304 ASS and b TiN particles with EDS analyses
Fabrication of AISI 304 Austenitic Stainless Steels … 685

Fig. 2 XRD pattern of the feedstock a 304 ASS and b TiN powders

100
90
80
70 304 ASS
TiN
% Passing

60 304 ASS + 2TiN

304 ASS + 4TiN
50 304 ASS + 6TiN
40
30
20
10
0
10 100 1000
Size (microns)

Fig. 3 PSD of 304 stainless steel matrix composite

SS/4TiN and 304 SS/6TiN, 50% (D50) of the particles pass through approximately
15 μm sieves. In the case of TiN, 50% (D50) of the particles pass through 5.5 μm.
From SEM analysis of the powders (see Fig. 1), TiN particles were identified by
smaller particle size whereas 304 ASS particles were classified by larger particles.
The PSD curves confirm that the particles sizes of TiN powder to be smaller than
that of 304 ASS powder.
Relative densities for the SPS composites used in the study are shown in Fig. 4.
From the graph, it could be observed that the densities of the composites decrease

Fig. 4 Relative density of

composites after SPS process
686 B. A. Obadele et al.

with an increasing amount of TiN reinforcing material in the steel matrix. The reason
for this trend may be due to the lower density of TiN (5.43 g/cm3 ) than that of 304 SS
(8.03 g/cm3 ). Relative densities of the composites were 97.65%, 96.64%, 96.56% and
96.23% for compositions without and with TiN (0 wt%, 2 wt%, 4 wt% and 6 wt%)
respectively. Sulima et al. [11] reported that the apparent densities of the composites
decrease with increasing content of the reinforcing material in the matrix. This is
so because of the lower densities the reinforcing material usually possess. On their
studies on the effect of SPS parameters on densification and properties of steel matrix
composites, they indicated that the apparent densities of the steel matrix decrease as
the content of TiB2 reinforcing material increases in the steel matrix. On the other
hand, a decrease in relative density could result from the formation of pores around
the TiN phase as shown in Fig. 5.
Figure 5 presents the SEM micrographs of spark plasma sintered composites of
different content of TiN. It could be observed that most of the TiN are distributed
within the grain boundaries of the matrix material. As the amount of TiN increases
from 5 to 6%, more of TiN (white phase) segregate along the grain boundaries. SEM
images (Fig. 5a–d) reveal the presence of micropores within the grain and along

Fig. 5 SEM micrograph of SPS 304 ASS without and with 2%, 4% and 6% TiN
Fabrication of AISI 304 Austenitic Stainless Steels … 687

♣ -Austenite ♦ -FeN 0.0939 ♥ -TiN

♣♦

♣
♣
304ASS-6TiN ♥ ♥

304 ASS-4TiN

304 ASS-2TiN

304 ASS

10 20 30 40 50 60 70 80 90
2-theta (deg)

Fig. 6 X-ray diffractograms of 304 SS composite after SPS process

grain boundaries of 304 ASS. The decrease in sintered relative density (see Fig. 4)
correlates with SEM images presented in Fig. 5.
The XRD analysis of the SPS stainless steel composites with various amount
of reinforcement (TiN) in wt% sintered at a sintering temperature (1100 °C), heat-
ing rate (100 °C/min) and holding time (10 min) is given in Fig. 6. The peaks were
scanned at room temperature under argon atmosphere. The XRD reveals the presence
of the following phases; TiN with a chemical formula TiN0.90, phases containing
iron, chromium and nickel such as FeCr0.29Ni0.16C0.06, austenite. The main ele-
ments of 304 ASS which are Fe, Ni and Cr, are revealed, these elements lie close to
each other on the periodic table. Therefore, they possess similar properties and lattice
parameters. In situations like this, it is practically impossible to accurately determine
the phase composition of the spark plasma sintered 304 ASS composites because
the XRD analysis of these composites can only prove the solid solution of iron [11,
12]. There has been no peak shift during sintering. However, there has been a slight
change in peak intensities. The change in peak intensities could result from the sin-
tering temperature, time and heating rate. The XRD patterns before and after spark
plasma sintering process are very similar. There has been no significant transforma-
tion or phase change during spark plasma sintering. However, cubic nitride (chemical
formula, FeN0.0939) has been detected in the major peak of ASS 304 with 2, 4 and
6 wt% TiN at peak position 2θ = 43.884°. This finding could be due to the intense
phenomena that take place during the SPS process, which include surface activation,
diffusion, surface melting as well as necking between sintered powder particles and
plastic flow. An increase in mass transportation is obtained during sintering and this
improves the bonding between powder particles. The iron nitride that has emerged in
the studied material after sintering is formed as a result of iron and nitrogen diffusing
in the grain boundary of the steel matrix. It is very crucial to mention that the SPS
process involves a stage where the material undergoes non-equilibrium conditions
688 B. A. Obadele et al.

Fig. 7 Microhardness of
304 SS composites before
erosion studies

and this has a great impact on the reactions as well as the diffusions which take place
during SPS [12].
Microhardness profiles of ASS 304 without and with TiN additions are given in
Fig. 7. As expected, the microhardness values increase as the amount of TiN in the
specimen increases. This could be due to the hard phase TiN presence in the matrix
material. ASS 304 with 6% TiN shows the highest microhardness value of about 350
HV0.1 while ASS 304 without and with 2% and 4% TiN microhardness values of
275, 330, 338 HV0.1 respectively.

4 Conclusion

In this study, the effect of TiN on the microstructure and hardness behaviour of
304 austenitic stainless steel has been evaluated. Based on the results obtained, the
following conclusion could be drawn; SPS results given in terms of relative densities
show a decrease in relative density with increasing content of TiN. Also, an increase
in TiN value in the matrix increased the hardness values.

Acknowledgements The authors are grateful to Tribology Group, Universidad Nacional del Sur,
Bahía Blanca, Argentina for the use of equipment and facilities and The Institute of Nanoengi-
neering Research (INER), Tshwane University of Technology, Pretoria for the use of sintering and
characterization facilities.

References

1. Obadele BA, Lepule ML, Andrews A, Olubambi PA (2014) Tribocorrosion characteristics

of laser deposited Ti–Ni–ZrO2 composite coatings on AISI 316 stainless steel. Tribol Int
78:160–167
Fabrication of AISI 304 Austenitic Stainless Steels … 689

2. Bregliozzi G, Di Schino A, Ahmed S-U, Kenny J, Haefke H (2005) Cavitation wear behaviour
of austenitic stainless steels with different grain sizes. Wear 258:503–510
3. Samandi M, Shedden B, Smith D, Collins G, Hutchings R, Tendys J (1993) Microstructure,
corrosion and tribological behaviour of plasma immersion ion-implanted austenitic stainless
steel. Surf Coat Technol 59:261–266
4. Sun Y, Bell T (2002) Dry sliding wear resistance of low temperature plasma carburised
austenitic stainless steel. Wear 253:689–693
5. More SR, Bhatt DV, Menghani JV (2017) Recent research status on erosion wear—an overview.
Mater Today Proc 4:257–266
6. Smuk O (2004) Microstructure and properties of modern P/M super duplex stainless steels
(Doctoral dissertation, Materialvetenskap).
7. Brytan Z, Dobrzański L, Grande MA, Rosso M (2009) Characteristics of vacuum sintered
stainless steels. J Achievements Mater Manuf Eng 33:126–134
8. Dada OR (2015) Characterization and beneficiation of foundry greensand waste from ferrous
foundries. University of Johannesburg
9. Obadele BA, Masuku ZH, Olubambi PA (2012) Turbula mixing characteristics of carbide
powders and its influence on laser processing of stainless steel composite coatings. Powder
Technol 230:169–182
10. Obadele BA, Ige OO, Olubambi PA (2017) Fabrication and characterization of titanium-nickel-
zirconia matrix composites prepared by spark plasma sintering. J Alloy Compd 710:825–830
11. Sulima I, Putyra P, Hyjek P, Tokarski T (2015) Effect of SPS parameters on densification and
properties of steel matrix composites. Adv Powder Technol 26:1152–1161
12. Oke SR, Ige OO, Falodun OE, Obadele BA., Shongwe MB, Olubambi, PA (2018) Optimization
of process parameters for spark plasma sintering of nano structured SAF 2205 composite.
Journal of materials research and technology, 7(2), 126–134.
Effect of Process Parameters on Bead
Width of 202 Grade Stainless Steel Gas
Tungsten Arc Welded Plates Using
Response Surface Methodology

R. Sudhakaran, P. S. Sivasakthivel, M. Subramanian, S. Mahendran,

M. Sathish Kumar, and R. Vijayakumar

Abstract Welding quality can be accomplished by meeting quality prerequisites,

for example, bead geometry by detecting either directly or indirectly the different
procedure parameters engaged with the procedure. Lacking weld bead measure-
ments may add to the disappointment of a welded structure. This paper presents
an improvement of a scientific model for foreseeing bead width for gas tungsten
arc welding of 202-grade stainless steel plates. This model depends on the connec-
tion between procedure parameters and bead width and can be utilized to foresee
the impact of procedure parameters on bead width. To assemble the required infor-
mation for demonstrating, real tests were done dependent on a central composite
rotatable plan with 31 test runs. The procedure factors considered here incorporate
welding current (I), welding speed (V ), gas flow rate (Q) and welding gun angle (θ ).
The sufficiency of the model was tried utilizing ANOVA procedure. The outcomes
demonstrate that the proposed model can foresee bead width with sensible precision.
Primary and cooperation impacts of the control factors are exhibited in a graphical
structure that helps in choosing rapidly the procedure parameters to accomplish the
ideal outcomes.

Keywords Gas tungsten arc welding · Central composite rotatable design ·

ANOVA · Bead width

R. Sudhakaran (B) · S. Mahendran · M. Sathish Kumar

Department of Mechanical Engineering, SNS College of Engineering, Coimbatore, Tamil Nadu,
India
e-mail: absudha@yahoo.com
P. S. Sivasakthivel
School of Mechanical Sciences, SASTRA University, Thanjavur, India
M. Subramanian · R. Vijayakumar
Department of Mechanical Engineering, SNS College of Technology, Coimbatore, Tamil Nadu,
India

© Springer Nature Singapore Pte Ltd. 2021 691

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_62
692 R. Sudhakaran et al.

1 Introduction

With the expanded utilization of mechanized welding, the determination of welding

process parameters and welding methodology must be increasingly explicit to guar-
antee that the weld bead parameters of good quality are gotten at least expense with
high repeatability. GTAW quality is emphatically portrayed by the weld bead geom-
etry. This is on the grounds that weld bead geometry assumes an imperative job in
deciding the mechanical properties of the weld [1]. Subsequently, the input welding
factors which impact the bead width should consequently be appropriately chosen
to acquire a satisfactory bead width and henceforth a great joint. EL Rayes et al. [2]
considered the impact of different crossbreed welding parameters on bead geometry.
They conducted experiments on 316 L austenitic stainless steel work piece. They
changed arc and laser capacity to contemplate their effect on different bead mea-
surements. They found that arc control impacts bead width. Esme et al. [3] improved
weld bead geometry in the TIG welding process utilizing Gray relational investi-
gation and Taguchi technique. They researched the multi-response optimization of
TIG welding process for an ideal parametric blend to yield positive bead geometry
of welded joints utilizing the Taguchi technique. They directed sixteen exploratory
runs dependent on an orthogonal array of Taguchi strategy to advance the target
work for example bead width. Thao and Kim [4] built up a connection demonstrate
for foreseeing weld bead geometry for lab joint in gas metal arc welding process.
They led tests dependent on full factorial plan with two dimensions of five procedure
parameters to get bead geometry utilizing a GMA procedure. They created linear,
curvilinear and interaction models for the procedure and found that forecast capac-
ities of interaction models are solid than the linear and curvilinear models. From
the investigation, it was induced that welding voltage, arc current, welding veloc-
ity, and welding gun angle have vast noteworthy impacts on bead width. Siva et al.
[5] advanced the weld bead parameters of nickel-based overlay saved by plasma
exchanged arc surfacing. They led tests dependent on a five-factor, five-level cen-
tral composite rotatable design and created numerical models utilizing the regression
method. They found that bead width expanded when the welding current is expanded
and diminished when the welding speed is expanded. Mostafa and Khajavi [6] opti-
mized welding process parameters for friction controlled arc welding process. They
built up a scientific model for foreseeing weld penetration as an element of welding
process parameters. The outcomes demonstrated that weld penetration achieves its
most extreme esteem when welding current, arc voltage, nozzle to plate distance
and electrode to work angle are greatest and welding speed is least. Raja Dhas and
Kumanan [7] proposed a clever strategy, versatile neuro fluffy derivation framework
(ANFIS) to anticipate the weld bead width in the submerged arc welding process for
a given arrangement of welding parameters. They structured the analyses utilizing
Taguchi’s standards and utilized the outcomes to build up various multiple regression
models. Information got from the model was utilized to prepare the smart network.
They found that the proposed ANFIS is adaptable, precise in anticipating bead width.
Menaka et al. [8] assessed bead width amid welding by infrared thermal imaging. A
Effect of Process Parameters on Bead Width … 693

ton of work has been completed to foresee weld bead width for gas metal arc weld-
ing, GTAW, Friction stir welding process for various evaluations of stainless steels
yet no work has been accounted for anticipating bead width for 202-grade stainless
steel which can be utilized as a substitute for 304-grade stainless steel in many appli-
cations like food processing equipment, kitchen hardware, sink and apparatus and
medicinal instruments and so on. Increasingly over the collaboration impacts of pro-
cedure parameters on bead width are not intricately talked about. Thus an endeavor
has been made to connect GTAW process parameters, for example, welding current
(I), welding speed (V ), gas flow rate (Q) and welding gun angle (θ ) with bead width.
A factually planned examination dependent on central composite rotatable design
was utilized for the improvement of numerical model [9]. The direct and intuitive
impacts of procedure parameters on bead width were examined. The created model
was helpful to decide quantitatively bead width of 202 stainless steel GTAW plates.

2 Experimental Procedure

The trials were structured and dependent on four-factor five-level central composite
rotatable plans with full replication method [10]. These trials were directed according
to the design matrix utilizing Lincoln V 350 PRO electric computerized welding
machine. A servo motor driven controller was utilized to keep up uniform welding
speed. The fundamental trial set up utilized comprised of a travelling carriage with a
table for supporting the specimens. A power source was kept prepared. A welding gun
was held stationary in an edge over the table and was given a connection for setting
the expected nozzle to plate distance and welding gun angle individually. The nozzle
to plate distance was kept consistent at 2.5 mm all through the experimentation
procedure. A high recurrence connection was utilized to produce the arc at this
separation. Test plates of size (100 mm × 30 mm × 5 mm) were cut from 202
stainless steel plates and one surface was cleaned to evacuate oxide scale and earth
before welding. The concoction synthesis of AISI stainless steel plate is given in
Table 1.
Argon gas with flow rates between 5 and 25 l/min was utilized for protecting the
weld bead. The reason for utilizing the protecting gas was to shield the weld region
from climatic gases, for example, oxygen, nitrogen, carbon dioxide and water vapor.
The welding machine and controller utilized for leading the trials is shown in Figs. 1
and 2.
 694

Table 1 Chemical composition of stainless steel 202 grade

SAE designation %C %Mn %Si %Cr %Ni %P %S %N %Fe
202 0.15 9.25 0.49 17.1 4.1 0.06 0.03 0.25 70.01
R. Sudhakaran et al.
Effect of Process Parameters on Bead Width … 695

Fig. 1 Gas tungsten arc

welding machine

Fig. 2 Manipulator used for

welding

3 Plan of Investigation

3.1 Identification of Procedure Parameters

The autonomously controllable procedure parameters impacting profundity of bead

width were recognized to empower the completing of exploratory work and building
up the scientific model. These are welding current (I), welding speed (V ), shielding
gas flow rate (Q), and welding gun angle (θ ).
696 R. Sudhakaran et al.

Table 2 Welding parameters and their levels

Parameter Unit and notation Levels
−2 −1 0 1 2
Welding current A (I) 70 80 90 100 110
Welding speed mm/min (V ) 170 180 190 200 210
Shielding gas flow rate l/min (Q) 5 10 15 20 25
Welding gun angle Degrees (θ) 50 60 70 80 90

3.2 Finding the Limits of the Procedure Parameters

The working extents of each picked factor were settled by coordinating fundamental
runs. This was finished by contrasting one of the components while keeping whatever
is left of them as relentless characteristics. The working extent of every method
parameter was constrained by exploring the dot for a smooth appearance with no
conspicuous distortions, for instance, surface porosity, undercut, etc. The maximum
furthest reaches of a given factor were coded as (+2) and lower limit was coded as
(−2). The coded qualities for centre qualities were resolved using the Eq. (1) [11]

2(2X − (X max − X min ))

Xi = (1)
X max − X min

where X i is the required coded estimation of a variable X and is any estimation of

the variable from X min to X max . The chosen procedure parameters with their points
of confinement and documentation are given in Table 2.

3.3 Development of the Design Matrix

The plan framework picked to lead the investigations was a five-level, four-factor
central composite rotatable design comprising of 31 sets of coded conditions and
involving a half replication 24 = 16 factorial structure in addition to eight star focuses
and seven centre focuses. All welding factors at the middle level (0) comprise the
inside focuses while the mix of each welding factors at either its lower level (−2)
or its more elevated amount (+2) with the other two factors at the moderate level
establish the star focuses. Subsequently, the 31 test runs permit the estimation of
straight, quadratic and two way intuitive impacts of the procedure factors on bead
width [12]. Examinations were led at random to maintain a strategic distance from
schematic errors crawling into the exploratory methodology.
Effect of Process Parameters on Bead Width … 697

Fig. 3 Bead profile of the

welded specimen

3.4 Recording the Responses

To measure the bead width the following processes were carried out on the specimens
(1) Segmenting, (2) Granulating, (3) Polishing, (4) Etching and (5) Profile tracing.
Segmenting: The transverse areas of each weld were cut utilizing band saw and
care was taken to keep away from twisting of the delicate austenitic stainless steel.
Granulating: It was performed so as to expel the cold work from cutting. The
granulating was done at rates of roughly 300 RPM.
Polishing: The specimens were then harsh cleaned by hand. So as to acquire
better edge evenness, the specimens were polished utilizing silicon carbide grating
papers of evaluations 100, 220, 400, 600 and 800 individually. The specimens were
then cleaned utilizing a rough—slurry of alumina (Al2 O3 ) and water (H2 O) on a
polishing machine.
Etching: After cleaning, etching was done on the specimens. Etching was impor-
tant for analyzing the microstructure of the weld bead. The etchant utilized was
Marble’s reagent which is a blend of HCl (50 ml), CuSO4 and H2 O (50 ml). The
polished countenances of every specimen were swabbed utilizing the etchant for
around 50–60 s to uncover the weld bead.
Profile Tracing: The bead profiles of the specimens were followed utilizing a
reflective type optical profile projector.
The followed bead profiles were examined so as to decide the bead width. The
bead width was estimated with the assistance of AUTOCAD. The followed bead
profile is shown in Fig. 3. The welded specimens are shown in Fig. 4.
The design matrix and measured value of bead width are shown in Table 3.
The welding conditions continue as before for trial keeps running from 25 to
31 yet the reaction shifts marginally. This is because of the impact of obscure and
eccentric factors called noise factors that creep into the tests. To represent the effect
of these obscure factors on the reaction rehashed runs were incorporated into the
design.

3.5 Recording the Responses

The development of a numerical model was done dependent on the regression tech-
nique to foresee the bead width. The response surface capacity speaking to bead
width can be communicated as α = f (I, V, Q, θ ) and the relationship chose is a
698 R. Sudhakaran et al.

Fig. 4 Welded specimens

second request reaction surface for k factors is given by Eq. (2) [9]


k 
k 
k
Y = b0 + bi X i + bi j X i X j + bii X i2 (2)
i=1 i, j=1 i=1
i= j

Factual programming bundle (Minitab Release 14) [11] was utilized to figure
the estimations of these coefficients. An underlying numerical model was created
utilizing the coefficients got from the above conditions.

3.6 Testing the Coefficients for Significance

The estimation of the regression coefficients gives a thought about what degree the
control parameters influence the response quantitatively. The less critical coefficients
are dispensed with alongside the responses with which they are related without giving
up a significant part of the exactness. This is finished by utilizing a student’s t-test
[13]. As per this test, when the determined value of ‘t’ comparing to the coefficient
surpasses the standard classified an incentive for the likelihood rule kept at 0.75, the
coefficient ends up noteworthy. The final numerical model was created utilizing just
the noteworthy coefficients. The final scientific model as controlled by the above
examination is given by Eq. (3).

W = 3.612 + 0.384I − 0.150V − 0.211Q + 0.416θ − 0.191Q 2

+ 0.098I V − 0.157I Q − 0.056V Q + 0.182V θ (3)

The square numerous estimations of ‘R’ of the full model and the diminished
model are exhibited in Table 4. It is apparent from the table that the full model has
the higher estimation of balanced square multiple ‘R’ than the diminished model and
Effect of Process Parameters on Bead Width … 699

Table 3 Design matrix and responses

S. No. Welding process parameters in coded values Bead width (mm)
I V Q θ°
01 −1 −1 −1 −1 3.0268
02 1 −1 −1 −1 3.7861
03 −1 1 −1 −1 2.3194
04 1 1 −1 −1 3.5200
05 −1 −1 1 −1 3.0144
06 1 −1 1 −1 3.3449
07 −1 1 1 −1 2.0593
08 1 1 1 −1 2.9645
09 −1 −1 −1 1 3.5187
10 1 −1 −1 1 4.3151
11 −1 1 −1 1 3.3384
12 1 1 −1 1 4.7879
13 −1 −1 1 1 3.4434
14 1 −1 1 1 3.7218
15 −1 1 1 1 3.3701
16 1 1 1 1 3.8312
17 −2 0 0 0 2.7917
18 2 0 0 0 4.4517
19 0 −2 0 0 4.0100
20 0 2 0 0 3.3366
21 0 0 −2 0 3.4124
22 0 0 2 0 2.4511
23 0 0 0 −2 2.9074
24 0 0 0 2 4.6083
25 0 0 0 0 3.6754
26 0 0 0 0 3.7935
27 0 0 0 0 3.4325
28 0 0 0 0 3.4425
29 0 0 0 0 3.4788
30 0 0 0 0 3.7886
31 0 0 0 0 3.7345

Table 4 Comparison of square multiple ‘R’ values and standard error of estimate for full and
reduced models
Response Adjusted square multiple R Standard error estimate
Full model Reduced model Full model Reduced model
Bead width 0.950 0.959 0.140 0.126
700 R. Sudhakaran et al.

the diminished model has lesser estimations of the standard error of estimate than
that of the individual full model. Thus the diminished model is superior to the full
model.

3.7 Checking the Adequacy of the Model

The adequacy of the model was tried utilizing the Analysis of Variance Technique
(ANOVA). According to this procedure [14], the determined estimation of the F—
the proportion of the model created ought not to surpass the standard estimation of
F-proportion for an ideal dimension of certainty, i.e., 95% and the determined esti-
mation of R-proportion of the model created ought to surpass the standard organized
estimation of the R-proportion for a similar certainty level. In the event that these
conditions are satisfied, the model is viewed as sufficient. The aftereffects of ANOVA
are exhibited in Table 5. It is obvious from the table that the model is sufficient.

4 Results and Discussion

The scientific model created can be utilized to foresee bead width by substituting
the estimations of the individual procedure parameters. The impact of the procedure
parameters on the bead width was concentrated utilizing the created model. The direct
impact of procedure parameters was contemplated utilizing the created model. The
direct impact was concentrated by keeping all the procedure parameters at the mid-
dle level with the exception of the parameter whose direct impact was examined.
The interaction of the parameters is considered by keeping every one of the param-
eters at the middle level aside from the parameters whose interaction impacts are
contemplated. The direct impact of the considerable number of parameters and the
interaction impacts of welding process parameters which have solid interaction on
bead width are examined underneath.

4.1 Direct Effect of Welding Current and Welding Speed

on Bead Width

Figure 5 demonstrates the direct impact of welding current on bead width. It is

obvious from the figure that bead width increments essentially as the welding current
increments. This is because of the way that heat input increments as the welding
current increments. The expansion in heat input outcomes in preheating of the work
piece amid forwarding welding. Subsequently, additionally liquefying of the base
metal happens. This builds the spreading of warmth which results in an increment of
Table 5 Results of ANOVA analysis.doc
Parameter First-order term Second-order term Lack of fit Error term F ratio R ratio Remarks
SS DOF SS DOF SS DOF SS DOF
Bead width 9.298 4 2.201 10 0.175 15 0.161 16 0.433 47.525 Model is adequate
Effect of Process Parameters on Bead Width …

SS Sum of Squares, DOF Degrees of Freedom

Mean Sum of Squares = Sum of Square Terms/DOF
F ratio = Ms of Lack of Fit/Ms of Error Terms
R ratio = Ms of First-Order Term and Second-Order Term/Ms of Error Term
F ratio (15, 6, 0.05) = 0.433
R ratio (9, 6, 0.05) = 4.10
701
702 R. Sudhakaran et al.

Fig. 5 Direct effect of welding current on bead width

Fig. 6 Direct effect of welding speed on bead width

bead width. Figure 6 demonstrates the direct impact of welding speed on bead width.
Welding speed is one of the primary factors that control heat info and bead width.
Bead width and measurements of the warmth influenced zone diminish with the
expansion in welding speed. This is because of the way that heat input is contrarily
corresponding to welding speed. Henceforth, there is an abatement in bead width as
the welding speed increments.

4.2 Direct Effect of Shielding Gas Flow Rate and Welding

Gun Angle on Bead Width

Figure 7 demonstrates the direct impact of shielding gas flow rate on bead width.
At the point when the gas flow rate is changed from lower level to more elevated
amount there is a peripheral increment in bead width up to (−1) level and afterward
Effect of Process Parameters on Bead Width … 703

Fig. 7 Direct effect of shielding gas flow rate on bead width

there is a diminishing pattern up to (+2) level. This is because of the way that at
higher gas flow rates more warmth is diverted by the gas. These outcomes in less
bead width and diminishing in measurements of the warmth influenced zone. Thus
there is an abatement in bead width for comparing increment in shielding gas flow
rate. Figure 8 demonstrates the direct impact of welding gun angle on bead width.
From the figure, bead width increments as the welding gun angle increments. This is
because of the way that at higher gun angles the preheating of the base metal is high
and the base metal gets more introduction to the arc. Thus there is an expansion in
bead width for comparing increment in gun angle.

Fig. 8 Direct effect of welding gun angle on bead width

704 R. Sudhakaran et al.

4.3 Interactive Effect of Welding Speed and Welding Current,

Welding Gun Angle on Bead Width

Figure 9 speaks to the contour diagram for interactive effect of welding speed and
welding current on bead width. From the contour plot, when the welding speed is
at its lower limit for example (−2) level, bead width is in the scope of 3–3.5 mm
for increment in welding current from (−2) to (0) level. It increments and is in the
scope of 3.5–4 mm for the welding current from (0) to (+2) level. At the point when
the welding speed is at (−1) level, bead width is in the scope of 2.5–3 mm for the
welding current up to (−1) level. It increments and is in the scope of 3–3.5 mm for the
welding current up to (+1) level. It is in the scope of 3.5–4 mm for the welding current
at (+2) level. For the following dimension of welding speed for example (0) level,
bead width is in the scope of 2–2.5 mm for the welding current up to (0) level. It is in
the scope of 3–3.5 mm for the welding current at (+1) level and increments to 4 mm
for the welding current at (+2) level. At the point when the welding speed is at (+1)
level, bead width is under 2 mm for the welding current at (−2) level. The bead width
builds like (0) level for different dimensions of welding current. A similar pattern is
watched for welding speed at (+2) level. The expansion in bead width for increment
in welding current for all dimensions of welding speed is because of the beneficial
outcome of welding current which is more noteworthy than the negative impact of
welding speed. Figure 10 demonstrates the contour plot for interactive of welding
speed and welding gun angle on bead width. From the contour plot, when the welding
gun angle is at (−2) level, bead width is in the scope of 4–4.5 mm for all dimensions
of welding speed. At the point when the welding gun angle is at (−1) level, bead
width is in the scope of 3.5–4 mm for the welding speed up to (−1) level. It at that
point diminishes and is in the scope of 3–3.5 mm for the welding speed at (0) level.
It further reduces and is in the scope of 2.5–3 mm for the welding speed at (+1) level.
It tumbles to under 2.5 mm for the welding speed at (+2) level. At the point when
the welding gun angle is at (0) level, a similar pattern is watched for bead width for

Fig. 9 Interactive effect of

welding speed and welding
current for bead width
Effect of Process Parameters on Bead Width … 705

Fig. 10 Interactive effect of

welding speed and welding
gun angle for bead width

all dimensions of welding speed. For the following dimension of welding gun angle
for example (+1) level, bead width is in the scope of 4–4.5 mm for all dimensions of
welding speed. At the point when the welding gun angle is at its maximum farthest
point for example (+2) level, bead width is in the scope of 4–4.5 mm for the welding
speed up to (0) level. It at that point insignificantly increments to 4.5 mm for the
other two dimensions of welding speed. From the above investigation, it tends to
be derived that bead width increments for increment in welding gun angle for all
dimensions of welding speed. This is because of the beneficial outcome of welding
gun angle which is more critical than the negative effect of welding speed.

4.4 Interactive Effect of Welding Current, Welding Speed

and Shielding Gas Flow Rate on Bead Width

Figure 11 speaks to the shape outline for the interactive effect of welding current and
shielding gas flow rate on bead width. From the contour plot, when the shielding gas
flow rate is at (−2) level and welding current is differed from (−2) to (+2) level bead
width increments from 1.5 to 5 mm. At the point when the shielding gas flow rate is at
(−1) level, bead width increments from 2 to 5 mm for increment in welding current
from (−2) to (+2) level. A similar pattern is watched for gas flow rate at (0) and (+1)
level. At the point when the shielding gas flow rate is expanded to its furthest restrict
for example (+2) level, bead width is in the scope of 2–3 mm for increment in welding
current from (−2) to (+2) level. From the above investigations, it may be surmised
that bead width increments when the welding current increments for all dimensions
of gas flow rate. This is because of the way that the beneficial outcome of welding
current is more critical than the negative effect of shielding gas flow rate on bead
width. Figure 12 demonstrates the contour plot for interactive effect of shielding gas
flow rate and welding speed on bead width. As appeared in the contour plot, when
the shielding gas flow rate is at (−2) level bead width is in the scope of 2–2.5 mm for
706 R. Sudhakaran et al.

Fig. 11 Interactive effect of

welding current and
shielding gas flow rate for
bead width

Fig. 12 Interactive effect of

welding speed and shielding
gas flow rate for bead width

increment in welding speed from (−2) to (+2) level. At the point when the shielding
gas flow rate is at (−1) level, bead width is in the scope of 3.5–4 mm for increment
in welding speed from (−2) to (+2) level. At the point when the shielding gas flow
rate is at (0) level bead width is more noteworthy than 4 mm for welding speed at
(−2) level. It at that point diminishes and is in the scope of 3.5–4 mm for different
dimensions of welding speed. For the following dimension of shielding gas flow rate
for example (+1) level, bead width is in the scope of 3.5–4 mm for the welding speed
up to (+1) level. It at that point diminishes and is in the scope of 3–3.5 mm for the
welding speed up to (0) level. It further declines and is in the scope of 2.5–3 mm for
the welding speed up to (+1) level. It is in the scope of 2–2.5 mm for the welding
speed at (+2) level. At the point when the shielding gas flow rate is at its maximum
breaking point for example (+2) level, bead width is in the scope of 3–3.5 mm for
the welding speed at (−2) level. It is in the scope of 2.5–3 mm for the welding speed
at (−1) level. It further declines and is in the scope of 2–2.5 mm for the welding
Effect of Process Parameters on Bead Width … 707

speed at (0) and (+1) levels. It goes underneath 2 mm for the welding speed at (+2)
level. From the above investigation, it very well may be induced that bead width
diminishes when the welding speed increments for all dimensions of gas flow rate.
This is because of the way that both welding speed and shielding gas flow rate have
a negative effect on bead width.

5 Conclusions

The accompanying ends were touched base at from the present examination.
1. The regression display created in this work from the experimental information
can be utilized to control the procedure so as to accomplish the ideal weld quality
in butt-welded plates.
2. The percentage of error got for profundity of entrance among predicted and
experimental esteem falls inside the breaking point of 95% certainty level. Sub-
sequently, the created regression display is fit for making the forecast of bead
width with sensible exactness.
3. Central composite design can be advantageously used to examine the immediate
and intuitive impacts of various blends of procedure parameters inside the scope
of the examination.
4. Out of four procedure parameters chose for examination, welding current and
welding gun angle has constructive outcome on bead width where as welding
speed and shielding gas flow rate has negative impact on bead width.
5. Out of the distinctive blends of procedure parameters the following combinations,
i.e., welding current and welding speed, welding current and shielding gas flow
rate, welding speed and shielding gas flow rate and welding speed and welding
gun angle have solid interactions on bead width.
6. The greatest bead width acquired from experimental investigations was 4.78 mm
when the procedure parameters, for example, welding current, welding speed,
and welding gun angle were at 100 A, 200 mm/min and 80° individually and
shielding gas flow rate was at 10 l/min.
7. The least bead width acquired from experimental investigations was 2.06 mm
when welding current and welding gun angle were at 80 A and 60° separately
whereas welding speed and shielding gas flow rate were at 200 mm/min and
20 l/min individually.

References

1. Giridharan PK, Murugan N (2009) Optimization of pulsed GTA welding process parameters
for the welding of AISI 304L stainless steel sheets. Int J Adv Manuf Technol 40(5–6):478–489
2. EL Rayes M, Walz C, Sepold G (2004) The influence of various hybrid welding parameters on
bead geometry. Weld J 147–153
708 R. Sudhakaran et al.

3. Greyjevo OGTZ, Esme U, Bayramoglu M, Kazancoglu Y, Ozgun S (2009) Optimization of

weld bead geometry in TIG welding process using grey relation analysis and Taguchi method.
Mater Technol 43(3):143–149
4. Thao DT, Kim IS (2009) Interaction model for predicting bead geometry for lab joint in GMA
welding process. Comput Mater Sci Surf Eng 1(4):237–244
5. Siva K, Murugan N, Raghupathy V (2009) Modelling, analysis and optimisation of weld bead
parameters of nickel based overlay deposited by plasma transferred arc surfacing. Surf Eng
1:174–182
6. Mostafa NB, Khajavi MN (2006) Optimisation of welding parameters for weld penetration in
FCAW. J Achievements Mater Manuf Eng 16(1):13–17
7. Raja Dhas JE, Kumanan S (2007) ANFIS for prediction of weld bead width in a submerged
arc welding process. J Sci Ind Res (India) 66(4):335–338
8. Menaka M, Vasudevan M, Venkatraman B, Raj B (2005) Estimating bead width and depth of
penetration during welding by infrared thermal imaging. Insight Non-Destr Test Cond Monit
47(9):564–568
9. Cochran WG, Cox GM (1963) Experimental designs. Asia Publishing House, India
10. Vel Murugan V, Gunaraj V (2005) Effects of process parameters on angular distortion of gas
metal arc welded structural steel plates. Weld J 84(11):165–171
11. Mini tab release 14, Minitab Inc (2004)
12. Montgomery DC (2001) Design and analysis of experiments. Wiley, New York
13. Yang LJ, Bibby MJ, Chandel RS (1993) Linear regression equations for modeling the
submerged-arc welding process. J Mater Process Tech 39(1–2):33–42
14. Murugan N, Gunaraj V (2000) Prediction and optimization of weld bead volume for the
submerged arc process—part 1. Weld Res 331–338
Feasibility Study of TIG Welding
of AA6063-AA7075 Alloys

D. Nathan, S. Ashwin Kannan, and P. Krishna Kumar

Abstract The aim of this paper is to study the feasibility of welding dissimilar
aluminum alloys AA6063 and AA7075 using destructive and nondestructive testing.
Dissimilar Aluminum alloys AA6063 and AA7075 plates were butt-jointed using the
tungsten inert gas (TIG) welding method. The feasibility of welding of dissimilar
aluminum alloy can be confirmed by both strength calculations as well as identifica-
tion of welding defects. The feasibility of welding of dissimilar aluminum alloy joint
was studied through visual appearance, microstructures, tensile strength, hardness,
liquid penetrant test and ultrasound test. Microstructures at different zones of dis-
similar TIG joints were identified. The average tensile strength and hardness values
were calculated. Using the liquid penetrant test and ultrasound test, welding defects
were identified. It can be concluded that TIG welding of a dissimilar aluminum alloy
is feasible.

Keywords Aluminum alloys · TIG welding · Nondestructive testing · Liquid

penetrant test

1 Introduction

Automotive manufacturing environments have been focusing on light metal such as

alloy without compromising its functionality. Aluminum alloys are widely used in
the industry due to their promising properties like high strength to weight ratio, high
thermal conductivity, low density and high corrosion resistance. Alloys have differ-
ent varieties. Out of these varieties, the 6xxx series aluminum alloys are heat treat-
able, have high strength and excellent in corrosion resistance. Structural members
and architectural elements where strength and stiffness are significant employ 6xxx
series alloys due to its high extrudability. Other applications of 6xxx series include

D. Nathan (B)
Dr. Mahalingam College of Engineering and Technology, Pollachi, India
e-mail: nathan14576@gmail.com
S. Ashwin Kannan · P. Krishna Kumar
St. Joseph’s Institute of Technology, Chennai 600119, India
© Springer Nature Singapore Pte Ltd. 2021 709
S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_63
710 D. Nathan et al.

weld structural members such as truck and marine frames, pipelines and railroad cars.
Besides, the 7xxx series alloys are also heat treatable with the highest strength of
all aluminum alloys. AA7075 is also wide applications in aircraft as well as in auto-
motive industries [1]. The joining of dissimilar alloys becomes a prime demand in
the automotive and aircraft industries for products with improved quality in terms of
strength and ductility but light in weight. In order to produce vehicles with economical
fuel consumption and manage pollution, many parts are manufactured by combin-
ing alloys that have lightweight and good corrosion resistance such as aluminum
alloys [2].
Tungsten Inert Gas (TIG), also known as Gas Tungsten Arc Welding (GTAW) is
the preferred welding process of alloys because of the comparatively easier applica-
bility and superior economy [3]. The TIG welding process is also suitable to weld
thin plates, resulting in high-quality and defect-free welded joints [4]. The arc is
established between the tip of a non-consumable tungsten electrode and the work
piece with a shielding gas applied to protect oxidation of the weld pool area [5].
In the TIG welding, problems may occur during the welding process where the
welds may fail in the soft and over age heat-affected zone [6]. The limitations exist
to join dissimilar aluminum alloys that are mainly related to the presence of an oxide
layer. This oxide layer weakens the grain boundary cohesion and consequently the
weld becomes prone to inter-granular cracking experienced during welding-pool
solidification. Also, thermal conductivity and composition with alloying elements are
one of the major causes these materials are difficult to join [7, 8]. Hence, feasibility of
welding dissimilar alloys needs to be studied deeply before applied to the fabrication
of engineering products.
In this study, the weldability of dissimilar aluminum alloys 6061 and 7075
was studied using both destructive and nondestructive testing. Visual inspec-
tions, microstructure identification, tension test and hardness test are conducted to
investigate the effect of the welding process.

2 Materials and Methods

2.1 Materials

AA6063 and AA7075 plates of dimensions 200 mm × 50 mm × 10 mm were used

in this study. The chemical compositions and mechanical properties are tabulated
in Tables 1 and 2, respectively. Alloy plates are prepared by polishing with 180
grit emery paper at the faying surface and cleaned with ethanol. Lincoln ER5356
aluminum welding wire is used as filler material. After surface preparation, manual
TIG welding of the dissimilar alloy was done. Welding parameters which are tungsten
electrode diameter of 3.2 mm, 60° electrode tip angle and argon shielding gas were
Table 1 Chemical compositions of AA6063 and AA7075 by weight%
Grade Al Cr Cu Fe Mg Mn Si Ti Zn Other, each
AA6063 Max 97.5% Max 0.1% Max 0.1% Max 0.35% 0.45–0.9% Max 0.1% 0.2–0.6% Max 0.1% Max 0.1% Max 0.05%
AA7075 87.1–91.4% 0.18–0.28% 1.2–2% Max 0.5% 2.1–2.9% Max 0.3% Max 0.4% Max 0.2% 5.1–6.1% Max 0.05%
Feasibility Study of TIG Welding of AA6063-AA7075 Alloys
711
712 D. Nathan et al.

Table 2 Properties of AA6063 and AA7075

Grade AA6063 AA7075
Physical properties
Density (×1000 kg/m3 ) 2.69 2.8
Mechanical properties
Young’s modulus (GPa) 68.9 70–80
Tensile strength (MPa) 241 230
Yield strength (MPa) 214 105
Elongation (%) 12% 17%
Poisson’s ratio 0.3 0.33
Shear strength (MPa) 152 150
Fatigue strength (MPa) 68.9 160
Thermal properties
Coefficient of thermal expansion (×10−6 /°C) 25.6 23.2
Thermal conductivity (W/mK) 201–218 130

Table 3 Welding parameters

Parameters Level 1 Level 2 Level 3
and its level
Welding current (A) 150 175 200
Welding voltage (V) 18 20 22
Gas flow rate (L/min) 5 7 9

fixed. The other parameters such as welding current welding voltage and gas flow
rate are varied. The various welding parameters and its level are tabulated in Table 1
(Table 3).

2.2 Experimentation

Tungsten Inert Gas Welding is a multi-factor metal joining technique. The various
process parameters include welding current, welding voltage, gas flow rate, elec-
trode diameter, nozzle gap, etc. influence on weld bead geometry, weldment quality
and also mechanical-metallurgical properties of the weldment. A limited number
of experimental runs have been done in this study to use three levels of welding
process parameters namely welding current, welding voltage, gas flow rate. Single
side 60° V groove has been made before welding. Experiments have been carried
out to create a single pass butt joint of dissimilar aluminum plates by varying the
process parameters in the TIG welding. The tensile test specimens were prepared as
per ASTM standards.
Feasibility Study of TIG Welding of AA6063-AA7075 Alloys 713

Table 4 Experimentation
Exp. No. Welding Welding Gas flow rate
matrix
current (A) voltage (V) (L/min)
1 150 18 5
2 150 20 7
3 150 22 9
4 175 18 7
5 175 20 9
6 175 22 5
7 200 18 9
8 200 20 5
9 200 22 7

2.3 Orthogonal Array

Taguchi’s orthogonal arrays are standard experimental design tables in order to per-
form only a small number of experimental trials to find the influence of the main
factors on the response. The minimum number of experiments to be conducted is
calculated from the formula given below

N = 1 + V (L − 1) (1)

N Number of experiments needed

V Number of input parameters
L Number of levels of each parameter.

In this study,
V = 3 and L = 3, Hence
N = 1 + 4(3 − 1) = 9.
Taguchi experimental design of experiments suggests L9 orthogonal array, where
nine experiments are sufficient to optimize the parameters. Hence at least nine exper-
iments are to be conducted. Based on this L9 orthogonal array (OA) is to be selected
which has at least nine rows, i.e., nine experimental runs (Table 4).

3 Results and Discussion

3.1 Tension Test Effect of Gas Flow Rate

Destructive testing was done on three specimens (specimen 2, specimen 4 and spec-
imen 9). The results are as shown in Fig. 1. The result shows the specimen 2, welded
714 D. Nathan et al.

Fig. 1 Tensile test results

Table 5 Results of tensile

Sample Id 2 4 9
test
Tensile strength in MPa 34.86 68.83 36.99
Yield stress in MPa 29.95 61.03 27.49
Elongation in % 1.50 0.90 0.24
Fracture location At weld At weld At weld

at low welding current, low welding voltage and medium gas flow rate has a low
tensile strength.
The result of the testing done on specimen 4 which is welded under medium
welding current, medium welding voltage and low gas flow rate gives out high tensile
strength. The result depicts the specimen nine welded low welding current, low
welding voltage and medium gas flow rate brings very low tensile strength (Table 5).

3.2 Hardness Test

The Vickers Hardness test was done on all specimens under 10 kgf load with a dwell
time of 10 s. Before doing hardness test the specimens were prepared by polishing
with help of different grades of emery paper in order to get clear indentation marks.
Hardness measurement was carried from either side of base metal towards the center
of the weld. The results of specimens 2, 4, and 9 were plotted in a graph as shown in
Fig. 2. The result shows the specimen 2, welded at low pulsed current, low pulsed
voltage and medium gas flow rate has a high hardness structure in HV.
The result of the testing done on specimen 4, which is welded under medium
pulsed current, medium pulsed voltage and low gas flow rate gives out medium
Hardness structure in HV. The result depicts the specimen 9, welded high pulsed
current, high pulsed voltage and high gas flow rate bring very low Hardness structure
in HV.
Distribution of microhardness values was analyzed and depicted in Fig. 3. The
graph shows that hardness value decreases at HAZ on both sides of welding due to
the coarsening of grain structure. Hardness value has an increasing trend near the
Feasibility Study of TIG Welding of AA6063-AA7075 Alloys 715

Fig. 2 Hardness test results

Fig. 3 Nugget profile of different experiments

center of the weld bead on both sides of the welded region. It is found to be AA 7075
side has higher hardness distribution compared to the AA 6063 side.

3.3 Microscopic Examinations

Microscopic examinations of TIG-welded plates with different welding parameters

were studied. It is found the microstructure of the test specimen 9 shows a uniform
grain structure with no distinct heat-affected zone. In specimen 4, the structure of
the HAZ is coarse-grained. The presence of large grains due to the overheating
zone causes weakening the strength properties of the joint. The microstructure of
the test specimen 4 reveals distinct grain coarseness and heat-affected zone. In the
716 D. Nathan et al.

a) Specimen 2 b) Specimen 4 c) Specimen 9

Fig. 4 Microscopic view of a specimen 2, b specimen 4, c specimen 9

Fig. 5 a Liquid penetrant test, b crack detection

test specimen, 2 grain distortion was found clearly. A small amount of intermetallic
phases in the form of fine precipitates is present in the grains and grain structure
found to be coarse (Fig. 4).

3.4 Liquid Penetrant Test

Liquid penetrant test results are depicted in Fig. 6. A maximum flaw length of 10 mm
was found in the test number 1 is an unacceptable value. Test number 3, 4, 5, 6, 7
and 8 have an average flaw length of 2.91 mm. Test number 2 and 9 have no defects
(Fig. 5).
Feasibility Study of TIG Welding of AA6063-AA7075 Alloys 717

Fig. 6 Liquid penetrant test result

3.5 Ultrasonic Test

The calibrated ultrasound flaw detector can calculate the depth of the discontinuity
and the thickness of the component accurately. Figure 7 shows the ultrasonic testing
on TIG-welded specimen.
From Fig. 8, it is evident that the maximum flaw length of 3.8 mm occurs in test
number 5 at the gas flow rate of 9 L/min and welding of 175 A. Test numbers 2, 8,

Fig. 7 Ultrsonic testing of

welded specimen
718 D. Nathan et al.

Fig. 8 Ultrsonic test result

and 9 show minimum flaw length of 3.4 mm. Average flaw length of 3.5 mm found
in test 3 and 7. Hence, it is evident that all the test specimens have flaws in the weld.

3.6 Effect of Gas Flow Rate

From literature, gas flow rate in the TIG welding process of Aluminum alloy has
a significant effect on weld quality. Shielding gas has a greater influence on weld
strength, porosity, ductility, impact strength and resistance to corrosion. Argon is used
as shielding gas in this work. Welding defects such as porosity, lack of penetration and
fusion are taken care of by suitable shielding gas at a certain flow rate. Hence the gas
flow rate is considered as one of the welding parameters in this study. Observations
show a significant variation in tensile strength with respect to gas flow rate.

4 Conclusion

This investigation found the feasibility of various TIG welding by estimating mechan-
ical properties and microscopic study of the weld. Taguchi’s L9 Orthogonal Array
has been implemented in the experimentation. Tensile testing, hardness testing and
microscopic examinations have been done. Liquid penetrant test and ultrasonic
inspection were successfully conducted on welds produced at various parameter
combinations. It can be concluded that based on the various parameters of welding
Feasibility Study of TIG Welding of AA6063-AA7075 Alloys 719

the liquid penetrant testing and ultrasonic technique successfully detected the defects
present in the welds and can be said to be appropriate in this regard. This study shows
the idea to choose suitable welding parameters for the better welding of dissimilar
alloy.

References

1. Fukuda T (2012) Weldability of 7000 series alloy materials. Weld Int 26:256–269
2. Hayat F (2012) Effect of aging treatment on the microstructure and mechanical properties of
the similar and dissimilar 6061-T6/7075-T651 RSW joints. Mater Sci Eng A 556:834–843
3. Kumar TS, Balasubramanian V, Sanavullah MY (2007) Influences of pulsed current tungsten
inert gas welding parameters on the tensile properties of AA 6061 alloy. Mater Des 28:2080–2092
4. Nascimento MP, Voorwald HJC, PayãoFilho JDC (2011) Fatigue strength of tungsten inert
gas-repaired weld joints in airplane critical structures. J Mater Process Technol 211:1126–1135
5. American welding. Society 261–269 (2012)
6. Mutombo K, Du Toit M (2011) Corrosion fatigue behaviour of alloy 6061-T651 welded using
fully automatic gas metal arc welding and ER5183 filler alloy. Int J Fatigue 33:1539–1547
7. Lakshminarayanan AK, Balasubramanian V, Elangovan K (2007) Effect of welding processes
on tensile properties of AA6061 alloy joints. Int J Adv Manuf Technol 40:286–296
8. Song JL, Lin SB, Yang CL, Ma GC, Liu H (2009) Spreading behavior and microstructure
characteristics of dissimilar metals TIG welding–brazing of aluminum alloy to stainless steel.
Mater Sci Eng A 509:31–40
Effects of Processing Parameters
on Temperature Distributions, Tensile
Behaviour and Microstructure
of Friction Stir Welding of Dissimilar
Aluminium Alloys

Olatunji P. Abolusoro and Esther Titilayo Akinlabi

Abstract Research has shown that there is a correlation between process parameters,
temperature variations, tensile strength, microstructures and the durability of the
welds obtained in friction stir welding (FSW). The need to institute this correlation
is crucial in order to achieve a weld free of defects and having sound mechanical
behaviours for industrial applications. This work examines process parameters effects
on variations of weld temperature, tensile strength and microstructures in dissimilar
FSW of 6101-T6 and 7075-T651 aluminium alloys with a plate thickness of 6 mm.
The welding was done with rotational speeds of 1250, 1550 and 1850 rpm and
traverse speeds of 50 and 110 mm/min. The results obtained indicate that processing
parameters significantly affect the temperature distributions in the weld. Increase in
rotational speed increases the temperature but cause a decrease in tensile strength.
While the increase in travel speed cause a reduction in temperature which results to
increase in the tensile strength. Highest tensile strength of 143 MPa was obtained
at 1250 rpm and 50 mm/min but better mixing of both materials was achieved at
1550 rpm and 50 mm/min.

Keywords Aluminium · Friction stir welding · Temperature · Processing

parameters

1 Introduction

One of the metals that are broadly used in many engineering applications is alu-
minium. The extensive use of this alloy emanates from its properties. One of which
is its high strength to weight ratio. FSW is a technology that has been in use for
joining both similar and dissimilar aluminium alloys. The unique advantages of this

O. P. Abolusoro (B) · E. T. Akinlabi

Department of Mechanical Engineering Science, University of Johannesburg, Johannesburg 2006,
South Africa
e-mail: abolusoroolatunji@yahoo.com
E. T. Akinlabi
Department of Mechanical Engineering, Covenant University, Ota, Ogun State, Nigeria

© Springer Nature Singapore Pte Ltd. 2021 721

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_64
722 O. P. Abolusoro and E. T. Akinlabi

welding technology which include non-requirements of filler metals as the case in

non-solid state welding technology and environmentally friendly have endeared it to
its wide usage in industries for joining aluminium alloys [1–3]. There are parameters
that govern the friction stir welding process. Two most prominent of these factors are
the rotation and travel speed of the tool pin on the workpiece. These two parameters
affect largely the heat generated during the welding and consequently affect temper-
ature distributions, tensile and microstructures of the weld [4–7]. Aissani et al. [8]
reported the influence of process parameters on tensile behaviour. Lombard et al. [9]
optimized the rate of feed and rotation of tool effects on tensile strength and concluded
that tool rotational speed impact more on tensile behaviour. Many other researchers
have reported the relationship between these factors, tensile and microstructures [10–
14] Also, the influence of these processing parameters on temperatures distributions
have been examined by some researchers. Umasanka and Vijay studied the relation-
ship between tool rotational speed, mechanical impact strength and temperature on
AA6101-T6 and AA6351-T6 alloys welded using FSW. Their result indicates that
temperature goes higher with an increase in rotational speed and that the impact
strength depends on the pressure and the duration of impact [15]. Yoshima et al.
welded dissimilar ADC12 and AA6061 alloy and examined the temperature distri-
butions during the welding. The temperature measurements were done at various
offset values. Their results indicate that temperature goes higher with an increase
in tool offset and the temperature at any point depends on its distance from the
welding tool which generates the heat [16]. Modelling of temperature distributions
in FSW of one of the chosen alloys for this work (AA7075) has been attempted.
Kandasamy et al. analyzed temperature changes in FSW of AA7075 aluminium
alloy using finite element approach. The temperature measurement was done using
k-type thermocouple on the workpiece and hand infrared thermometer on the tool.
Their result demonstrated close agreement with the experimental values obtained
[17]. Many researchers have utilized various approaches to investigate temperature
changes during FSW. Some of these methods include thermal cameras, thermocou-
ple inserted in the welding tool or in the workpiece [18–22]. Other methods like
ultrasound and neutron-based techniques have also been used [23, 24]. However,
the most commonly used method observed in literature is that of a thermocouple
embedded in workpiece [15, 16, 25, 26]. This particular method was employed in
this study. Thermocouple positions have also been reported to affect the accuracy of
temperatures measured. While close proximity of the thermocouple to the rotating
tool pin or the stir zone gives a better estimate of the weld temperature, the likely
destruction of the thermocouple wires as the pin progresses limits the use of this
method [15]. Hence in this study, the thermocouple was embedded in the workpiece
in an area close to the heat-affected zone. Although numerous researches on FSW
of dissimilar aluminium alloys have been carried out, however, welding parameter’s
relationship with microstructures, tensile and temperature variations especially on
FSW of 7075-T651 and 6101-T6 have not been fully reported. This paper, therefore,
presents this relationship for more understanding.
Effects of Processing Parameters on Temperature Distributions … 723

2 Experimental Procedures and Processing Parameters

2.1 Elemental Composition and Properties of the Sample

Dissimilar aluminium alloys 6101-T6 and 7075-T651 were used for the experiment.
The elemental constituents of the two materials are indicated in Table 1 while Table 2
shows the mechanical properties.

2.2 Process Parameters

Three tool rotational speeds and two feed rates were engaged in this study as presented
in Table 3.

2.3 Experimental Procedures

The tool for the experiment was fabricated using high carbon steel H13. It has shoul-
der diameter of 22 mm with a tapered pin profile of big diameter 7.3 mm, small
diameter 5.5 mm and pin length of 5.6 mm as illustrated in Fig. 1. The two alloy
plates of 7075-T651 and AA6101-T6 were cleaned and arranged in butt configura-
tion with AA6101-T6 on the advancing side and 7075-T651 on the retreating side
as illustrated in Fig. 2. The FSW was performed along the rolling direction of the
two alloy samples with the tilt angle at 20 . Two thermocouple wires were embedded
in the workpiece 2 mm away from the welding zone on both sides of the plates
and are placed 40 mm apart. The thermocouples were connected to a temperature
data capturing machine for in-process temperature measurements. The thermocouple
arrangement is shown in Fig. 3. Four different experiments were carried out using
the parameters shown in Table 3. Three tensile samples cut according to ASTM E8
standard were tested for each weld and the average for each weld taken for evalua-
tion. The microstuctural analysis was done on the nugget zone of each weld using
an optical microscope.
 724

Table 1 Elemental constituents of the two materials

Alloy Si Fe Cu Mn Mg Cr Zn Ti Al
6101-T6 0.53 0.14 0.01 0.002 0.600 0.001 0.003 0.008 Others
7075-T651 0.40 0.50 1.70 0.300 2.40 0.22 5.50 0.20 Others
O. P. Abolusoro and E. T. Akinlabi
Effects of Processing Parameters on Temperature Distributions … 725

Table 2 Mechanical properties of the alloy

Alloy Tensile strength Ultimate tensile Elongation (%) Thermal
(Mpa) strength (Mpa) conductivity
(W/m °C)
6101-T6 172 180 15 167
7075-T651 462 575 11 130

Table 3 Processing parameters

Rotational speed (rpm) Travel speed (mm/min)
1250 50
1550 50
1850 50
1850 110

Fig. 1 Schematic diagram of the welding tool

726 O. P. Abolusoro and E. T. Akinlabi

Fig. 2 Workpiece retreating side

Fig. 3 Schematic diagram of the thermocouple arrangements T0, T1, T2, T3 on the alloy plates

3 Results and Discussion

3.1 Weld Appearance

The weld joints were carefully observed after the welding and were found to be
externally free form defects although little flashes are noticed in some of the welds
as shown in the photograph in Fig. 4.
Effects of Processing Parameters on Temperature Distributions … 727

Fig. 4 Welded plates with embedded thermocouple

Fig. 5 The highest

temperature profile of the 600
processing parameters used
500
Temperature (oC)

400

300
T1 1250/50
T2 1550/50
200
T3 1850/50
T4 1850/110
100

0
0 2 4 6 8 10 12 14
Time (Min)

3.2 Temperature Profile of the Weld

Heat is produced as a consequence of the friction between the workpiece and the
rotating tool. This generated heat influenced the thermo-mechanical welding process
and impact on the mechanical behaviours of the joints. The highest heat takes place
at the workpiece and rotating tool interface. The temperature profile of the welding
process with time is presented in Fig. 5. The temperatures recorded at the 6101-T6
side of the weld which is the advancing side were lower than those at the 7075-T651
side which is the retreating side. This is as a result of different thermal conductivity
influence of the welded plates. The 6101-T6 alloy on the advancing side of the welds
has a higher thermal conductivity than 7075-T651 placed on the retrieving side.
This implies that the rate of heat transfer and heat losses is higher in 6101-T6 than in
7075-T651, hence the aluminium alloy 7075-T651 on the retreating side retains more
heat than 6101-T6 and therefore gave higher temperature values than the 6101-T6
alloys during the welding. The temperature profile also indicates that the temperature
728 O. P. Abolusoro and E. T. Akinlabi

generally increases as the welding progresses and get to the peak in the middle of
the weld.

3.3 Effects of Tool Rotational Speed on the Temperature

Distribution

The highest temperature of about 560 °C was recorded at a rotational speed of

1850 rpm while the least was about 365 °C at a rotational speed of 1250 rpm. The
temperature generally increases across the studied samples as the tool rotational
speed increases. This is due to higher heat generation as the rotating tool speed goes
higher own to increase in frictional movement between the rotating tool and the
workpiece.

3.4 Effects of Welding Speed on the Temperature Distribution

Varying the travel speed also influenced the temperature distributions. At 1850 rpm
and 50 mm/min, the highest temperature recorded was 560 °C while at 110 mm/min
travel speed with the same rotational speed of 1850 rpm, the highest temperature
was 480 °C. The lower temperature recorded at the higher travel speed could be as
a result of less heat generated own to less dwelling time of the rotating tool as it
moves faster on the workpiece. While the higher temperature obtained at the lower
travel speed was due to more dwelling time of the rotating tool as it moves slower
on the workpiece. Also, the ratio of the rotational speed to the welding speed of
the tool decreases from 37 to 17 rpm min/mm as the speed rises from 50 mm/min
to 110 mm/min. This decrease in the ratio leads to a reduction in heat input and
consequently lowers the temperature of the weld.

3.5 Tensile Behaviour of the Welds at Constant Travel Speed

At a steady travel speed of 50 mm/min, the ultimate tensile strength reduces as

the rotational speed rises. This is shown in Fig. 6. This emanates from more heat
generated as the rotational speed increases as shown in the temperature profile in
Fig. 5. The excessive heat cause deterioration of the precipitates at the welding
zone which leads to poor grain bonding and consequently decreases the mechanical
strength.
Effects of Processing Parameters on Temperature Distributions … 729

Fig. 6 Effect of rotational 160 Ultimate Tensile Strength

speed on ultimate tensile
strength 140

Ultimate tensile Strength (MPa)

120

100

80

60

40

20

0
1200 1400 1600 1800 2000
Rotational Speed (rpm)

3.6 Tensile Behaviour of the Welds at Constant Rotational

Speed

At a steady rotational speed of 1850 rpm, the rising of feed rate from 50 to
110 mm/min causes the ultimate tensile strength to increase by 4 MPa. The increase
follows a drop in the temperature as shown in Fig. 5. This implies that there was
an improvement in grain bonding as the temperature drops and consequently an
improvement in the ultimate tensile strength of the joint. This is shown in Fig. 7.

Ultimate tensile Strength

160

140
Ultimate Tensile Strength (MPa)

120

100

80

60

40

20

0
50 100
Travel Speed (mm/min)

Fig. 7 Effect of travel speed on the ultimate tensile strength

730 O. P. Abolusoro and E. T. Akinlabi

Fig. 8 Macrographs of the welded zone. a 1250 rpm/50 mm/min, b 1550 rpm/50 mm/min,
c 1850 rpm/50 mm/min, d 1850 rpm/110 mm/min

3.7 Macrostructure and Microstructure of the Welds

The macrographs of all the welds (Fig. 8a–d) show that there was sufficient heat for
plasticization for all the welds. The welds exhibited onion ring structures at all the
rotational speed. Lamellae structure pattern of material flow was observed in all the
parameters. Larger material movement is observed at the 1850 rpm and 110 mm/min.
However, full penetration of the alloys into one another occurred more at 1550 rpm
and 50 mm/min rotational speed and feed rate respectively.
The microstructure varies across the length and the breath of the weld and can
be divided into unmixed, mechanically mixed and mixed regions as reported by
some authors [27–29]. The unmixed region is observed close to the tool shoulder
areas. Fine-grain structure is observed here as shown in Figs. 9a, 10a, 11a and 12a
as a result of dynamic recrystallization during the welding. The mechanically mixed
region (Figs. 9b, 10b, 11b and 12b) is the interface region of both alloys at the fusion
zone. It contains the microstructures of both alloys. Weck’s reagent etchant used
revealed 7075-T651 as a lighter colour and the 6101-T6 as brown colour. At the
mixed region, the tool pin stirring mechanism moved plastically deformed materials
in alternate layers of 7075-T651 and 6101-T6 to form lamellae structural pattern
(Figs. 9c, 10c, 11c and 12c.
Material penetration of the aluminium alloy into each other was highest at
1550 rpm and 50 mm/min rotational speed and travel speed, respectively.
Effects of Processing Parameters on Temperature Distributions … 731

Fig. 9 Microstructures of weld nugget zone at 100× for 1250 rpm/50 mm/min, a unmixed,
b mechanically mixed, c mixed regions

Fig. 10 Microstructures of weld nugget zone at 100× for 1550 rpm/50 mm/min, a unmixed,
b mechanically mixed, c mixed regions

Fig. 11 Microstructures of weld nugget zone at 100× for 1850 rpm/50 mm/min, a unmixed,
b mechanically mixed, c mixed regions

Fig. 12 Microstructures of weld nugget zone at 100× for 1850 rpm/110 mm/min, a unmixed,
b mechanically mixed, c mixed regions
732 O. P. Abolusoro and E. T. Akinlabi

4 Conclusions

Conclusions from the results obtained are as follows:

1. Increase in the tool rotational speed result in an increase in the temperature
generated during the welding process.
2. Increase in travel speed decreases the temperature.
3. Temperature profile varies relative to welding time
4. Tensile strength decreases as the tool rotational speed increases.
5. There is an increase in tensile strength as the welding speed increases.
6. Material distributions in the welding zone are not even. The volume of material
swept into one another was more in 1850 rpm at a welding speed of 110 mm/min.
However, the highest penetration of the alloy into one another occurred at the
1550 rpm rotational speed and at the welding speed of 50 mm/min.

Acknowledgements The authors appreciate the head and all the scholars of the Friction stir lab-
oratory of the India Institute of Technology Kharagpur where this experiment was carried out and
to the University of Johannesburg South Africa for sponsorship.

References

1. Padhy GK, Wu CS, Gao S (2017) Friction stir based welding and processing technologies -
processes, parameters, microstructures, and applications: a review. J Mater Sci Technol 34:1–38
2. Sidhu MS, Chatha SS (2012) Friction stir welding—process and its variables: a review. Int J
Emerg Technol Adv Eng 2(12):275–279
3. Abolusoro OP, Akinlabi ET (2019) Wear and corrosion behaviour of friction stir welded
aluminium alloys—an overview. Int J Mech Prod Eng Res Dev 9(3):967–982
4. Mishra RS, Mahoney MW, Sato Y, Hovanski Y (2016) Friction stir welding and processing
VIII
5. Silva ACF, De Backer J, Bolmsjö G (2017) Temperature measurements during friction stir
welding. Int J Adv Manuf Technol 88(9–12):2899–2908
6. Khandkar MZH, Khan JA, Reynolds AP (2003) Prediction of temperature distribution and
thermal history during friction stir welding: input torque based model. Sci Technol Weld Join
8(3):165–174
7. Yan F, Zhang Y, Fu X, Li Q, Gao J (2019) A new calculating method of frictional heat and its
application during friction stir welding. Appl Therm Eng 153(January):250–263
8. Givi MKB, Asadi P (2014) Advances in friction-stir welding and processing
9. Lombard H, Hattingh DG, Steuwer A, James MN (2008) Optimising FSW process parameters
to minimize defects and maximize fatigue life in 5083-H321 aluminium alloy. Eng Fract Mech
75(3–4):341–354
10. Prabha KA, Putha PK, Prasad BS (2018) Effect of tool rotational speed on mechanical properties
of aluminium alloy 5083 weldments in friction stir welding. Mater Today Proc 5(9):18535–
18543
11. Aydin H, Bayram A, Esme U, Kazancoglu Y, Guven O (2010) Application of grey relation
analysis (Gra) and Taguchi method for the parametric optimization of friction stir welding
(FSW) process. Appl Grey Relat Anal 44(4):205–211
Effects of Processing Parameters on Temperature Distributions … 733

12. Abd Elnabi MM, Elshalakany AB, Abdel-Mottaleb MM, Osman TA, El Mokadem AA (2019)
Influence of friction stir welding parameters on metallurgical and mechanical properties of
dissimilar AA5454-AA7075 aluminum alloys. J Mater Res Technol 8(2):1684–1693
13. Rajakumar S, Balasubramanian V (2012) Establishing relationships between mechanical prop-
erties of aluminium alloys and optimized friction stir welding process parameters. Mater Des
40:17–35
14. Kumar SR, Rao VS, Pranesh RV (2014) Effect of welding parameters on macro and microstruc-
ture of friction stir welded dissimilar butt joints between AA7075-T651 and AA6061-T651
alloys. Procedia Mater Sci 5:1726–1735
15. Das U, Toppo V (2018) Effect of tool rotational speed on temperature and impact strength of
friction stir welded joint of two dissimilar aluminum alloys. Mater Today: Proc 5:6170–6175
16. Takayama Y, Akutsu Y, Choshiro N, Kato H, Watanabe H (2010) Temperature measurement
during friction stir welding of dissimilar aluminum alloys, pp 0–5
17. Kandasamy J, Sairam J, Hussain MM (2018) ScienceDirect ICMPC_2018 comparative analysis
of temperature variation in friction stir welding. Mater Today Proc 5:18798–18804
18. Sun YF, Fujii H, Sato Y, Morisada Y (2019) Friction stir spot welding of SPCC low carbon
steel plates at extremely low welding temperature. J Mater Sci Technol 35:733–741
19. Silva-Magalhães A, De Backer J, Martin J, Bolmsjö G (2019) In-situ temperature measurement
in friction stir welding of thick section aluminium alloys. J Manuf Process 39:12–17
20. Hamilton C, Dymek S, Sommers A (2010) Characteristic temperature curves for aluminum
alloys during friction stir. Weld J
21. Covington JL, Robison W, Webb BW (2006) Experimental characterization of tool heating
during friction stir welding. Mech Eng 179–184
22. Buffa G, Donati L, Fratini L, Tomesani L (2006) Solid state bonding in extrusion and FSW:
Process mechanics and analogies. J Mater Process Technol 177(1–3):344–347
23. Feng Z et al (2007) In situ neutron diffraction measurements of temperature and stresses during
friction stir welding of 6061-T6 aluminium alloy. Sci Technol Weld Join 12(4):298–303
24. Fehrenbacher A, Duffie NA, Ferrier NJ, Pfefferkorn FE, Zinn MR (2014) Effects of
tool-workpiece interface temperature on weld quality and quality improvements through
temperature control in friction stir welding. Int J Adv Manuf Technol 71(1–4):165–179
25. Upadhyay P, Reynolds AP (2010) Effects of thermal boundary conditions in friction stir welded
AA7050-T7 sheets. Mater Sci Eng A 527(6):1537–1543
26. Miles M, Karki U, Hovanski Y (2014) Temperature and material flow prediction in friction-stir
spot welding of advanced high-strength steel. JOM 66(10):2130–2136
27. Palanivel R, Koshy Mathews P, Murugan N, Dinaharan I (2012) Effect of tool rotational speed
and pin profile on microstructure and tensile strength of dissimilar friction stir welded AA5083-
H111 and AA6351-T6 aluminum alloys. Mater Des 40:7–16
28. Ouyang JH, Kovacevic R (2002) Material flow and microstructure in the friction stir butt welds
of the same and dissimilar aluminum alloys. J Mater Eng Perform
29. Shigematsu I, Kwon YJ, Suzuki K, Imai T, Saito N (2003) Joining of 5083 and 6061 aluminum
alloys by friction stir welding. J Mater Sci Lett
Optimizing the Parameters for Friction
Stir Welding of an Aluminium Alloy

M. Selvaraj

Abstract A novel slip factor accounted for thermal model is being used to predict the
temperature with time and the effect of welding and rotations on heat input for weld
length peak-temperature of the friction stir welding process. Material used in this
study is aluminium alloy of AA6061 grade which is used in aerospace applications.
For various rotations and weld velocities, friction stir welding experiments were
conducted on aluminium alloy plates. The plates were fixed with thermocouples
at different locations from weld centerline to measure the temperatures during the
welding process. The temperature with time and locations predicted by the model
along the transverse direction are closely matching with the experimental results.
The heat input for unit length of weld and peak-temperature increases as rotation
increases and decreases as weld velocity increases. The energy required for welding
for length is minimum at the combination of lower rotation and higher weld velocity
in the defect-free zone.

Keywords Friction stir welding · Aluminium alloy · Thermal model · Energy

required · Temperature

1 Introduction

Heat flow modeling during the FSW process was done by Rajamanickam et al. [1]
and Riahi and Nazari [2] using a heat flow model to obtain the thermal history.
Song and Kovacevic [3] presented the isotherms and temperature with time of FSW
process by simulating the welding process. Woo et al. [4] and Chao and Tang [5]
developed a heat flow model for the FSW process, numerically predicted tempera-
ture with time. Prasanna et al. [6] modeled the FSW process of steel, predicted the
temperature with time. Nandan et al. [7] presented a 3D thermal and material flow
model and predicted the thermal history, velocity and strain rate during the FSW
process. Hamilton et al. [8, 9] done simulation of FS welding of aluminium alloy

M. Selvaraj (B)
Sri Sivasubramaniya Nadar College of Engineering, Chennai, Tamil Nadu, India
e-mail: selvarajm71@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 735

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_65
736 M. Selvaraj

to predict the thermal cycle. 3-dimensional model for FSW considering energy due
to friction and plastic deformation is developed by Emam and Domiaty [10] and
predicted change of peak normalized temperature with the effective energy. Selvaraj
et al. [11, 12] proposed a thermal model for FS welding of steels assuming slip factor
as a function of temperature to predict the temperature with time and energy required
per unit weld length during the welding process.
The literature survey indicates that most of the thermal studies concentrate on
predicting temperature distribution and temperature with time during FSW process.
Most of the researchers have not predicted the energy required for different com-
binations of weld velocity and rotation. In some combinations of weld, parameters
require to weld with minimum energy. It will save a lot of energy. In this paper, atten-
tion has been given to predicting the energy-efficient rotational and weld velocity in
the FS welding process using a mechanics-based heat flow model.

2 Experiment Work

This study uses AA6061 aluminium alloy which is used in applications like aerospace
and marine. The material was cut into a plate of 150 × 75 × 6 mm size. Two plates are
aligned and fixed by clamps. Aligned plates were welded using an M2 steel tool as
shown in Fig. 1. The tool and welding parameters are given in Table 1. The welding
experiments were done by an FSW machine with 12 kW motor as shown in Fig. 2.

Fig. 1 FSW tool dimensions

Optimizing the Parameters for Friction Stir Welding … 737

Table 1 FSW parameters

Parameters Value
Radius of shoulder, r s 9 mm
Max. radius of pin, r 1 4 mm
Min. radius of pin, r 2 2 mm
Length of pin, h 5.5 mm
Axial load, F n 15.5 kN
Rotations, N 300–2000 rpm (7 speeds)
Translational velocity, V 300, 600, 900, 1400 and
1800 mm/min

Fig. 2 Experimental setup

One mm diameter four K type thermocouples were fixed at different locations.

They are at 4, 10, 15 and 25 mm from the center of weld on the advancing side. The
temperature during the process was measured using data logger.

3 Thermal Model

3.1 Mathematical Model

Shoulder heat
Friction heat at shoulder
2π
Q sh−fr = μ pn ω (rs3 − r13 ) (1)
3
738 M. Selvaraj

Pressure in axial direction,

Fn
pn = (2)
πrs2

Tool angular velocity,

2π N
ω= (3)
60
The friction coefficient (μ) is linearly varying from 0.5 to 0.4 between the room
and solidus temperature [13].
Shear deformation heat at shoulder
2π
Psh−pl = τ ω(rs3 − r13 ) (4)
3
Shear stress,
σy
τ=√ (5)
3

where σ y is the material yield stress which is shown in Table 2.

The expression for heat generated at the shoulder is obtained by merging the heat
generation due to friction and shear deformation with stick factor.
Shoulder heat

Q sh = δ.Q sh−pl + (1 − δ)Q sh−fr (6)

Total heat
Calculated pin heat at side Q ps and bottom Q pb are
Pin heat,

Q pin = Q ps + Q pb (7)

Total tool heat,

Q T = Q sh + Q ps + Q pb (8)

Shoulder and pin heat flux

The heat over weld length

QT
Ql = × 60 (9)
V
Table 2 Properties of 6061-T6 aluminium alloy [2]
Temperature (°C) 26 38 93 147 205 260 316 372 425
Yield strength (MPa) 276 275 265 249 219 159 66 34 17
Density (kg/m3 ) 2700 2686 2686 2666 2656 2656 2631 2621 2601
Optimizing the Parameters for Friction Stir Welding …

Thermal con. (W/m °C) 167 171 178 185 191 202 205 215 225
Heat capacity (J/kg °C) 896 921 979 1005 1029 1051 1079 1105 1134
739
740 M. Selvaraj

ht = 15 W/m2 oC
ht= 15 W/m2 oC

Weld center
line
Area heat flux
ht = 15 W/m2 oC Plate

Volume heat flux

ht = 15 W/m2 oC
ht = 0 W/m2oK

ht = 300 W/m2 oC

Fig. 3 FE 3D model with boundary conditions and heat flux

By assuming 90% of heat transferred to workpiece [14], shoulder heat flux

expression arrives linear with radius of the tool [13].

3Q shr
qsh = 0.9 × for rp ≤ r ≤ rs (10)
2π(rs3 − rp3 )

Heat around the pin is taken as volume heat flux in the shape of hollow cylinder
based on the shear layer thickness of 2 mm as stated by Chen et al. [15].

Q pin
qpin = 0.9 × (11)
π [(rp + 2)2 − rp2 ]h

3.2 Finite Element Model

See Fig. 3.

4 Results and Discussion

4.1 Thermal History

The variation of temperature with respect to time during FS welding of aluminium

alloy for different rotations and weld velocities at four different locations from the
weld centerline is shown in Fig. 4. For all input parameters, the thermal history
predicted by the model matches closely with that of experimental history. It shows
that this model is reliable to predict the temperature distribution during FS welding.
Optimizing the Parameters for Friction Stir Welding … 741

(a) Weld velocity 300 mm/min & rotation 300 rpm. (b) Weld velocity 300 mm/min & rotation 1200
rpm.

(c) Weld velocity 300 mm/min & rotation 2000 rpm. (d) Weld velocity 900 mm/min & rotation 300 rpm.

(e) Weld velocity 900 mm/min & rotation 1200 rpm. (f) Weld velocity 900 mm/min & rotation 2000
rpm.

Fig. 4 Temperature with time at various locations from the weld centerline for different Weld
velocity and rotations
742 M. Selvaraj

(g) Weld velocity 1800 mm/min & rotation 300 rpm. (h) Weld velocity 1800 mm/min & rotation 1200 rpm
.

(i) Weld velocity 1800 mm/min & rotation 2000 rpm.

Fig. 4 (continued)

Analyzing the temperature with time of 300 mm/min weld velocity and different
rotations, the temperature increases and drops sharply for 300 rpm than 1200 rpm.
It also shows that the peak-temperature is high and wider cycle is obtained for 1200
as compared to 300 rpm. But the variation of temperature with time and peak-
temperature between 1200 and 2000 rpm is not significant. For a constant weld
velocity (300 mm/min), when speed is increased from 300 to 1200 rpm, heat gen-
eration increases and leads to variation in temperature with time. The variation in
temperature with time between 1200 and 2000 rpm is not significant because vari-
ation in heat generation is not significant. The same kind of pattern is visible in the
other weld velocities also.
Analyzing the effect of various weld velocity for a constant rotation (300 rpm), the
rise and dip in temperature are very sharp and peak-temperature decreases as weld
velocity increases. As weld velocity increases, the heat source quickly approaches
and passes away the measuring location and it leads to a sharp rise and drops in
temperature.
Optimizing the Parameters for Friction Stir Welding … 743

4.2 Effect of Weld Parameters on Energy Required

and Peak-Temperature

The energy required for length of weld and peak-temperature for various weld veloc-
ity and rotations are illustrated in Figs. 5 and 6. It is noted that the heat input for
length of weld and peak-temperature increases as rotation increases but decreases as
weld velocity increases. It is also noted that at lower rotations slope of the curve is
more and it reduces as rotation increases. This is due to reduction in heat generation
rate as rotation increases. The heat input for length of weld is inversely propor-
tional to the weld velocity. The heat input for length of weld and peak-temperature

Fig. 5 Energy required per

mm length of weld for
various weld velocity and
rotations

Fig. 6 Simulated
peak-temperature for various
weld velocity and rotations
744 M. Selvaraj

Energy required per mm legth of the weld in J/mm

800

600

400

200

0
2000
1500 2000
1000 1500
1000
500 500
Translational velocity in mm/min 0 0
Rotational speed in rpm

Fig. 7 Energy required for length of weld for various weld velocity and rotations

follow the same trend. It is clear that heat input for length of weld influences the
peak-temperature (Fig. 7).
In the defect-free zone, the energy required per unit weld length is 188.2 J/mm
at the combination of lower rotation (1200 rpm) and higher weld velocity
(1800 mm/min). The energy required per unit weld length is 549.6 at the combi-
nation of higher rotation (1200 rpm) and lower weld velocity (300 mm/min). The
energy required at the combination of higher rotation and lower weld velocity is 2.9
times higher as compared with the energy required per unit weld length the combi-
nation of lower rotation and higher weld velocity. It shows that the energy required
for welding for length is minimum at the combination of lower rotation and higher
weld velocity in the defect-free weld zone. Hence it is vital to see the combination
of lower rotation and higher weld velocity in order to weld with minimum energy.
This will save a lot of energy required for welding.

5 Conclusions

1. The predicted temperature with time and location in transverse direction has
closely merged with that of experimental values.
2. The energy required per weld length is minimum at the combination of lower
rotation and higher weld velocity.
Optimizing the Parameters for Friction Stir Welding … 745

References

1. Rajamanickam N, Balusamy V, Reddy GM, Natarajan K (2009) Effect of process parameters

on thermal history and mechanical properties of friction stir welds. Mater Des 30:2726–2731
2. Riahi M, Nazari H (2010) Analysis of transient temperature and residual thermal stresses in
FS welding of aluminum alloy 6061-T6 via numerical simulation. Int J Adv Manuf Technol
55:143–152
3. Song M, Kovacevic R (2003) Thermal modeling of FS welding in a moving coordinate system
and its validation. Int J Mach Tools Manuf 43:605–615
4. Woo W, Choo H, Withers PJ, Feng Z (2009) Prediction of hardness minimum locations during
natural aging in an aluminum alloy 6061-T6 friction stir weld. J Mater Sci 44:6302–6309
5. Chao YJ, Tang W (2003) Heat transfer in FS welding—experimental and numerical studies.
Trans ASME 125:138–145
6. Prasanna P, Rao BS, Mohana Rao GK (2010) Finite element modeling for maximum
temperature in FS welding and its validation. Int J Adv Manuf Technol 51:925–933
7. Nandan R, Roy GG, Lienert TJ, Debroy T (2007) Three-dimensional heat and material flow
during FS welding of mild steel. Acta Mater 55:883–895
8. Hamilton C, Dymek S, Sommers A (2008) A thermal model of FS welding in aluminum alloys.
Int J Mach Tools Manuf 48:1120–1130
9. Hamilton C, Sommers A, Dymek S (2009) A thermal model of FS welding applied to Sc-
modified Al–Zn–Mg–Cu alloy extrusions. Int J Mach Tools Manuf 49:230–238
10. Emam SA, Domiaty AE (2009) A refined energy-based model for friction-stir welding. World
Acad Sci Eng Technol 53:1016–1022
11. Selvaraj M, Murali V, Rao SK (2013) Thermal model for FS welding of mild steel. Multidiscip
Model Mater Struct 9:49–61
12. Selvaraj M (2014) A temperature dependent slip factor based thermal model for FS welding
of stainless steel. Sadhana-Acad Proc Eng Sci 38:1393–1405
13. Soundararajan V, Zekovic S, Kovacevic R (2005) Thermomechanical model with adaptive
boundary conditions for FS welding of Al 6061. Int J Mach Tools Manuf 45:1577–1587
14. Dickerson T, Shi Q, Shercliff, HR (2003) Heat flow into FS welding tools. In: 4th international
symposium on FS welding, Park City Utah, USA
15. Chen Z, Pasang T, Qi Y (2008) Shear flow and formation of Nugget zone during FS welding
of aluminium alloy 5083-O. Mater Sci Eng, A 474:312–316
Machining of ZE41 Magnesium Alloy
in WEDM Using Taguchi Approach

G. Selvakumar , S. Ram Prakash , and E. Caleb Kovilpillai

Abstract This study explains the effects of process parameters of Wire Electrical
Discharge Machining (WEDM) on the machinability and surface characteristics of
Magnesium ZE41 alloy. ZE41 alloy has a plethora of applications in the structural
industry, aerospace industry, military equipment, video cameras, and vibration test-
ing equipment. In this study, Pulse on time (Ton), Pulse off time (Toff), Current
(I) and Wire Tension (WT) were considered as the input parameters to conduct the
experiments for the performance measures Material removal rate (MRR) and Sur-
face roughness (Ra). Experimentation works were conducted by using Taguchi’s L9
orthogonal array. Analysis of variance (ANOVA) was performed to identify the sig-
nificant parameters. XRD test was carried out to identify the phase constituents of
the material. Further, Surface morphology of the machined surface is analyzed using
SEM micrographs.

Keywords ZE41 magnesium alloy · WEDM · Material removal rate · Surface

roughness · Process parameters

1 Introduction

The demand for alloy materials is rising abound in recent years due to its appealing
properties like high strength to weight ratio, ductility, formability, and high resistance
to chemical attack. Alloying is an effective technique to enhance the mechanical prop-
erties of the materials by combining the beneficial properties of two or more metals
[1]. It is the most predominantly used material in all engineering applications since
pure metals have limitations in its properties. Magnesium is the 6th widely available
lightweight material in the earthly resources [2]. Magnesium alloys have renowned
properties like good castability, formability, high sound dampening capacity, and
low density [3]. Among aluminum, steels, and copper-base alloys, magnesium alloy

G. Selvakumar · S. Ram Prakash (B) · E. Caleb Kovilpillai

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Kalavakkam, Chennai 603110, India
e-mail: ramprakashnec@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 747

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_66
748 G. Selvakumar et al.

is the supreme material for structural applications due to its high strength to weight
ratio. However, it has poor hardness, wear, and corrosion resistance. Pure magne-
sium is highly susceptible to corrosion and weakens the material behavior which
can be enhanced by adding alloying elements [3]. Magnesium and its alloys have a
plethora of applications in biomedical engineering like screws and plates for bone
fixation, tissue engineering scaffolds, as well as on cardiovascular stents [4], automo-
bile, aerospace, electronics, and structural industries [5]. Machining of magnesium
alloy at the required shape with the superior surface finish is necessary to perform its
intended functions. So, it is indeed crucial to machine the material with minimal sur-
face roughness and higher material removal rate. It is identified that nontraditional
machining is the best way out to cut intricate shapes with a better surface finish.
Amongst all other nontraditional machining processes, WEDM is the apt machining
process to process magnesium alloys. The material removal mechanism for WEDM
is the melting and vaporization of the material due to the electrical discharges. It has
the potential to cut any intricate shapes with great accuracy of electrically conduc-
tive material [6]. A material with high electrical and thermal conductivity would be
preferred as wire electrodes. Brass wire and zinc-coated brass wire are commonly
used wire electrodes in the recent era due to its utmost performance [7].
Janardhan and Samuel [8] developed a pulse discrimination algorithm to study the
effect of machining parameters at spark gap and revealed that the surface roughness
(Ra) and roundness error are majorly affected by the occurrence of arc regions, width
of arc and average ignition delay time. Weingartner et al. [9] investigated the shape
and size of craters produced during single discharge experiments and developed
an electro-thermal model. Further, the experimental results are then compared with
the simulation results. Mohammadi et al. [10] discussed the effects of machining
parameters for roundness error and surface quality of a cylindrical component and
further presented the mathematical model equation for predicting the roughness using
regression analysis.
Few researchers have attempted on processing ZE41 magnesium alloy in WEDM
but there is a gap in investigating the surface integrity and morphology of the
machined surface. The objective of the present work is to study the effect of process
parameters of WEDM on machining of ZE41 alloy for the responses MRR and sur-
face roughness. Surface morphology of the machined surface was examined through
SEM micrographs. Besides, the phase constituents of the alloy were investigated
using XRD pattern.

2 Materials and Methods

The experimental work is performed in Ratnaparki CNC WEDM machine, model:

smart cut 2530. The machine setup contains a machine tool, control system, and
circulation type dielectric unit with a capacity of 250 L. The chemical composition
of the ZE41 magnesium alloy used in this experimentation work is shown in Table 1.
The phase constituents of the ZE41 magnesium alloy have been identified by XRD
Machining of ZE41 Magnesium Alloy in WEDM Using Taguchi Approach 749

Table 1 Chemical compositions of ZE41 Magnesium alloy

Material Zinc Rare earth Zirconium Magnesium
Magnesium ZE41 alloy 3.5–5 0.8–1.7 0.4–1 Balance

Table 2 Parameters and their levels

S. No. Process parameters Level 1 Level 2 Level 3
1 Pulse on time—Ton (µs) 0.1 0.2 0.3
2 Pulse off time—Toff (µs) 30 40 50
3 Peak current—I (A) 1 3 5
4 Wire tension—WT (N) 18 20 22

spectrum. Copper is used as target material, generator voltage and tube current were
set as 45 kV and 40 mA, respectively. In Fig. 1, from the obtained XRD peaks, it
is evident that magnesium occupies the dominant phases and few traces of zinc are
also detected when compared with JCPDS data.
Zinc-coated brass wire is used as wire material with the diameter 250 µm and
deionized water is used as dielectric medium. Taguchi’s L9 orthogonal array was
employed to conduct the experiments. The controllable factors considered for this
experiment were based on previous literature and trial experiments. Table 2 dis-
cusses the controllable parameters and their levels considered for conducting the
experiments. The constant parameters used for these experiments were wire feed:
3 m/min, Gap voltage:30 V, Water Pressure: 9 bar.
MRR is calculated by the following expression:

MRR (mm3 /min) = cutting speed (mm/min) × kerf width (mm) × thickness (mm)
(1)

Surface roughness is measured by MitutoyoSurftest, Model: SJ 210 with the cut-

off length as 0.25 mm. While Taguchi’s parametric analysis approach, the better per-
formance of WEDM for MRR is indicated by choosing the problem type as “larger is
better” and “smaller is better” is chosen for surface roughness. Table 3 represents the
experimental results for the L9 orthogonal array. In Taguchi’s approach, the WEDM
performance analysis for MRR is marked as “Larger is better”. Besides, “smaller is
better” is chosen for surface roughness. The analysis of variance (ANOVA) method
has been deployed to figure out the contribution of individual factors for the responses
MRR and Ra (Fig. 1).
750 G. Selvakumar et al.

Table 3 Experimental results

S. No. Ton (µs) Toff (µs) I (A) WT (N) MRR (mm3 /min) Ra (µm)
1 0.1 30 1 18 2.822 2.83
2 0.1 40 3 20 2.405 2.64
3 0.1 50 5 22 2.202 2.63
4 0.5 30 3 22 10.797 3.56
5 0.5 40 5 18 9.867 2.95
6 0.5 50 1 20 9.602 3.51
7 1 30 5 20 11.020 3.70
8 1 40 1 22 10.606 3.32
9 1 50 3 18 11.557 3.12

Fig. 1 XRD pattern of ZE41 magnesium alloy

3 Results and Discussion

From Fig. 2 and Table 4, it is revealed that the pulse on time is the most predominant
factor which majorly influences the MRR. Next to pulse on time, current plays a
major role in deciding MRR. The rest of the factors have minor influences over
MRR. The increase in pulse on time and current increases the material removal rate,
this is due to the fact that the amount of discharge energy rises up drastically which
in turn melt and vaporize large lump of material and produces huge crater. Therefore,
an increasing trend is observed in Fig. 3 for both pulse on time and current.
Machining of ZE41 Magnesium Alloy in WEDM Using Taguchi Approach 751

Fig. 2 Effect of parameters on MRR

Table 4 ANOVA result

Factor DF MRR Ra
AdjSS % of contribution AdjSS % of contribution
Ton 2 77.67 69.36 6.3157 53.80
Toff 2 12.256 10.94 0.777 6.62
I 2 14.13 12.62 2.6744 22.78
WT 2 1.042 0.93 0.1478 1.26
Error 0 6.883 1.8333
Total 8 111.981 11.7481

From Fig. 3 and Table 4, it is revealed that the pulse on time plays the major role
followed by the current on determining the surface roughness of the machined sur-
face. The amount of discharge energy applied is directly proportional to the longevity
of the pulse and the magnitude of discharge energy per pulse decides the crater size.
An increase in pulse on time increases the discharge energy which tends to produce
large craters and creates an uneven surface. Hence, increase in pulse on time and
current diminish the surface quality of the material [11].
Figure 4 explains the surface morphology of the machined surface. In WEDM,
material removal takes place in the form of craters. The size of the crater depends on
the intensity of the electrical discharge. As the discharge energy increases, the more
will be the crater size. It is observed from the SEM images that the machined surface
752 G. Selvakumar et al.

Fig. 3 Effect of parameters on Ra

Fig. 4 a Microcracks on the machined surface, b presence of molten globules, c crater and
microcrack formation of the machined surface

contains microcracks, microvoids, craters, and molten globules. In some regions,

it shows the accumulation of fragmented debris on the machined surface. Further,
it witnesses the random presence of molten globules due to the rapid cooling effect
of the deionized water. During spark emission, gas bubbles are formed which later
explodes and produces microcracks on the surface. The presence of microcracks
and voids deteriorates the mechanical strength of the material. Proper flushing along
with appropriate pulse off time should be provided to reduce the recast layer on the
machined surface.
Machining of ZE41 Magnesium Alloy in WEDM Using Taguchi Approach 753

4 Conclusion

In this work, WEDM machining of ZE41 alloy was performed and the effect of
process parameters was studied for the responses MRR and Ra. The XRD analysis
has been conducted to identify the phase constituents of the material. Pulse on time
and current were identified as the prime factors which influence the material removal
rate and surface roughness. The surface morphology of the machined surface has
been studied with the SEM micrographs. Molten globules, microcracks, pockmarks,
and microvoids were observed in the machined region.

References

1. Chen J, Tan L, Yu X, Etim IP, Ibrahim M, Yang K (2018) Mechanical properties of magnesium
alloys for medical application: a review. J Mech Behav Biomed Mater 87:68–79
2. Singh K, Singh G, Singh H (2018) Review on friction stir welding of magnesium alloys. J
Magn Alloys 6:399–416
3. Dinodi N, Nityananda Shetty A (2013) Electrochemical investigations on the corrosion
behaviour of magnesium alloy ZE41 in a combined medium of chloride and sulphate. J Magn
Alloys 1:201–209
4. Hohna S, Virtanena S, Boccaccini AR (2019) Protein adsorption on magnesium and its alloys:
a review. Appl Surf Sci 464:212–219
5. Luo AA (2013) Magnesium casting technology for structural applications. J Magn Alloys
1:2–22
6. Selvakumar G, Thiruppathi Kuttalingam KG, Ram Prakash S (2018) Investigation on machining
and surface characteristics of AA5083 for cryogenic applications by adopting trim cut in
WEDM. J Braz Soc Mech Sci Eng 40:267
7. Gnanavelbabu A, Saravanan P, Rajkumar K, Karthikeyan S, Baskaran R (2018) Optimization
of WEDM process parameters on multiple responses in cutting of Ti-6Al-4V. Mater Today:
Proc 5:27072–27080
8. Janardhan V, Samuel GL (2010) Pulse train data analysis to investigate the effect of machining
parameters on the performance of wire electro discharge turning (WEDT) process. Int J Mach
Tools Manuf 50:775–788
9. Weingartner E, Wegener K, Kuster F (2012) Wire electrical discharge machining applied to
high-speed rotating workpieces. J Mater Proc Technol 212:1298–1304
10. Mohammadi A, Tehrani AF, Emanian E, Karimi D (2008) A new approach to surface roughness
and roundness improvement in wire electrical discharge turning based on statistical analyses.
Int J Adv Manuf Technol 39:64–73
11. Selvakumar G, Sarkar S, Mitra S (2013) An experimental analysis of single pass cutting of
aluminium 5083 alloy in different corner angles through WEDM. Int J Mach Machinabil Mater
13(2/3)
Comparative Study of Friction Stir
Welding and Underwater Friction Stir
Welding on Magnesium ZE41 Alloy

S. Cyril Joseph Daniel and A. K. Lakshminarayanan

Abstract 5-mm-thick ZE41 magnesium alloy plates were subjected to normal fric-
tion stir welding (in air) and under water friction stir welding (UWFSW). A compar-
ative study was made to investigate the process parameter and resultant properties of
microstructure and tensile behaviour. The friction stir welding (FSW) was carried out
at two different speeds 660 and 1220 rpm in order to obtain the influence of rotation
speed on the performance of underwater joints. And the tool transverse speed and
the tool tilt angle were fixed at 40 mm/min and 2.5°, respectively. From this investi-
gation, it can be found that the UWFSW joint made using the higher tool rotational
speed of 1220 rpm exhibited good tensile properties when compared to FSW joints
without any defects like tunnel-type defects or hot cracking. The higher cooling rate
in the UWFSW caused reduced thermomechanically affected zone (TMAZ) in the
weld. The joint welded at UWFSW made the grain size fine when compared to the
grain size of joints at the air weld condition. It is also noted that the hardness was
increased in the UWFSW specimen due to the fine grains present in the stir zone.
For air weld, the fracture occurred in the weld region and for UWFSW the fracture
moved away from the stir zone.

Keywords Friction stir welding · Magnesium ZE41 · Underwater ·

Microstructure · Mechanical properties

1 Introduction

Magnesium alloys have brought more and more attention lately in response to the
rising eco-awareness. Due to the fact that magnesium possesses the lowest density
among all structural metals (such as Al, Mg and Ti), the purpose of magnesium alloys
will reduce the energy consumption and has very high vibration damping capacity.

S. Cyril Joseph Daniel (B) · A. K. Lakshminarayanan

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai,
Tamil Nadu 603110, India
e-mail: cjdaniel1994@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 755

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_67
756 S. Cyril Joseph Daniel and A. K. Lakshminarayanan

In addition, magnesium also has the advantage of recyclability, further decreasing

the squandering of natural resources. Based on the lower density and recyclability,
magnesium alloys have gradually become the highly potential metallic material. The
light alloys with lower density, higher strength and good corrosion resistant properties
magnesium are widely used. Magnesium alloys are the lightweight material among
the structural metals, which have two-thirds density (1.74 g/cm3 ) of the aluminium
alloys (2.7 g/cm3 ) and also four times lighter the steels which have the density of
7.86 g/cm3 . It has a good ductility and excellent casting ability.
Friction stir welding is the solid-state welding process which is recently adopted
in all possible processes. In FSW process, the non-consumable tool moves between
two plates which are to be welded together by the effect of friction between tool
and the workpiece. Moreover, the stirring action and pressure applied play the major
role in FSW process. The lower heat generation in FSW stands unique and avoids
many defects due to higher temperature on fusion welding. Expanded details about
the process of FSW were discussed in many previous publications [1–5].
Nowadays, there are many methods like GTAW, AGTAW, GMAW, SMAW and
laser beam welding followed for welding of magnesium ZE41 alloy [1, 6]. Some
researchers had tried different materials with FSW in underwater condition and got
better mechanical properties [7]. But mechanical properties and microstructural dis-
tribution for joining of similar magnesium ZE41 underwater condition have not
been studied fully. In this paper, similar magnesium ZE41 alloy was welded together
using both FSW at underwater condition and FSW at air condition. The microstruc-
ture distribution, tensile strength and fracture region were investigated to understand
a comparative study of welding of underwater FSW and FSW of magnesium ZE41
alloy.

2 Experimental Procedure

As per the requirement, the magnesium ZE41 alloy (200 mm × 40 mm × 5 mm)

was prepared and checked for defect-free condition. Their chemical composition and
mechanical properties are shown in Table 1. The material was completely cleaned
to make sure the absence of the foreign particles which can cause defects during
the welding. The as-received alloys have coarse grains, so the faces of the plates
were machined and finished to get the flat surface. The plates were placed together
on the welding direction, and tightly clamped to the table to avoid distortion during
the welding. The process was done at a XL6026AV machine, which is a modified form

Table 1 Chemical properties of magnesium ZE41

Grade Zn Ce Zr Mg
% Min 3.5 0.8 0.4 Remaining
Max 5 1 1
Comparative Study of Friction Stir Welding and Underwater … 757

of vertical friction stir welding machine with the 5HP capability shown as in Fig. 5.
The machine has 360 cm of table width and 1370 cm of table length (Fig. 5), and
the chamber for underwater welding was designed and fabricated with dimensions
of 400 mm × 280 mm × 75 mm length, width and height, respectively. This set-up
is designed to hold the material to be welded and capable of carrying the water to
make the material immersed condition (Figs. 1, 2, 3 and Table 2).

Fig. 1 Schematic of weld specimen

Fig. 2 Joint to be fabricated (Mg ZE41)

758 S. Cyril Joseph Daniel and A. K. Lakshminarayanan

Fig. 3 Chamber for underwater FSW

Table 2 Mechanical properties of magnesium ZE41

Tensile Modulus of Poisson’s Elongation Brinell’s Melting point
strength elasticity ratio hardness
218 MPa 44.12 GPa 0.35 4.50% 55–70 532–638 °C

Figure 4 shows that the non-consumable tool was made up of high carbon steel
with dimension of shoulder diameter (D) 15 mm, insert pin diameter (d) 5 mm and
pin length (L) of 4 mm, respectively. Two plates were welded at rotational speed of
1220 rpm and travel speed of 40 mm/min. The tool’s tilt angle used is 2.5° to get
effective material flow during tool transverse. Various parameters used in this study
were indicated in Table 3.
After the FSW process at both conditions, the samples were cut perpendicular to
the direction of welding using electrical discharge machine. The cross section was
observed by the optical microscopy (OM). Afterwards, the prepared samples were
mounted and polished using SiC papers with different grit sizes. Polished surfaces
were swabbed with the etching solution (picric acid, ethanol, acetic acid and water of
4.2 g, 70 ml, 40 ml and 10 ml, respectively) until the macrostructure was revealed. The
etched samples were cleaned with ethanol for few seconds. Microstructure images
were obtained by the optical microscope (Figs. 6, 8, 9 and 10).
Comparative Study of Friction Stir Welding and Underwater … 759

Fig. 4 FSW tool

Table 3 FSW process

Parameter Value
parameter and tool
nomenclature Tool rotation speed (rpm) 660 1220
Tool transverse speed (mm/min) 40
Axial force (KN) 5
Tool tilt angle (°) 2.5
Shoulder diameter (mm) 15
Shoulder length (mm) 20
Pin diameter (mm) 5
Pin length (mm) 4
D/d ratio of tool 3
Tool profile Taper cylindrical

3 Results and Discussions

3.1 Appearances and Microstructures

Figure 6 shows the optical micrograph of ZE41 base alloy. The α-Mg matrix has
T-Phase: Mg7 Zn3 -RE precipitates which are distributed over the matrix eutectically.
There are some more finely distributed particles in the matrix; they are Zr4 Zn3 or
Zn2 Zr3 zirconium-rich particles. These zirconium-rich particles have grain size of
760 S. Cyril Joseph Daniel and A. K. Lakshminarayanan

Fig. 5 XL6026AV semi-automatic milling machine

T - Phase

Intermetallic particles

α - Mg

Fig. 6 Optical micrograph of ZE41 alloy

Comparative Study of Friction Stir Welding and Underwater … 761

Fig. 7 Welded joint of ZE41 alloy

around 4 μm [8, 9]. From the Mg–Zr phase diagram, below 653 °C, the α-Zr particles
and α–Mg particles are present in the magnesium matrix and they are not dissolved
in the base material [10, 11]. The brighter contrast of the grain boundaries shows
that the presence of higher atomic number of the secondary-phase particles in the
grain boundary. The T-phases in the Mg matrix show a white- and light-coloured
zone in the SEM micrograph of ZE41 alloy, and similar observations are reported
by Neil et al. [12].
Figure 7 shows the FSW joint of ZE41. It also showed no defects on surface when
the rotational speed of tool was 1220 rpm and transverse speed was 40 mm/min. And
the spiral formation on upper surface of the joints can be seen. It shows that material
flow during the welding process and the spiral obtained due to the relation between
tool rotation and the substrate movement. Figures 8 and 9 represent the macrograph
and its microfeatures of underwater FSW and FSW, respectively. The boundary
between stir zone (SZ) and the thermomechanically affected zone (TMAZ) is called
delineation. From Fig. 8, it can be seen that the delineation in the advancing side is
sharper and the delineation in the retreating side is gradually changed. It evidences
the state of plastic flow during welding process. In the stir zone, the friction effect is
produced by the welding tool and the material on the advancing side is squeezed to
the back of the pin [6].
Heat dissipation rate was high on the underwater friction stir welding since water
at room was in contact with magnesium plate; thus, the thermal gradient developed
during welding was dissipated from plate. So the frictional heat between tool and
the specimen was absorbed by the surrounding water. So that there was no enough
time given for growth of recrystallized grains than the air weld. On the other side,
the heat dissipation rate in the normal air atmosphere was low where sufficient time
was given for grain refinement than the underwater condition. So it had enough time
762 S. Cyril Joseph Daniel and A. K. Lakshminarayanan

Fig. 8 Micrograph and macrograph of underwater welded sample: a advancing side, b retreating
side, c SZ and base interface, d stir zone and e macroimage

for grain growth near the interface. So the geometry of stir zone in the air weld
condition was quite higher than the underwater condition.
The macrostructure of friction stir welded ZE41 alloy is shown in the figure in
which grain refinement occurred heavily in stir zone. The FSW weld joint reveals
the bathtub-like curve which denoted the stir zone, a narrow region of interface like
structure that divides the stir zone from HAZ in TMAZ (Figs. 8 and 9). This TMAZ
appeared as a resultant of mechanical stirring effect occurring between tool and
workpiece. The retreating and advancing side of the weld metal can be identified by
grain refinement pattern. The sudden grain refinement was observed in advancing
side due to the initial area of contact was at advancing side during the welding. Then
the plasticized material was wiped towards the rotating direction of the tool. It is
called as retreating side where the grain growth took place eventually.
Figure 8d shows that the stir zone of the underwater welding has smaller cross
section when compared to the stir zone of the air weld FSW, because the higher heat
dissipation in underwater limited the grain growth at the interface region.
Comparative Study of Friction Stir Welding and Underwater … 763

Fig. 9 Micrograph and macrograph of air welded sample: a advancing side, b retreating side, c SZ
and base interface, d stir zone and e macroimage

Figures 10a and b show that the average grain size measured for underwater
welded specimen is 25 μm and for the air welded specimen is 40 μm. This is due to
the cooling rate was higher in the underwater FSW when compared to air FSW.

Fig. 10 Heat-affected zone: a underwater condition and b air welded condition

764 S. Cyril Joseph Daniel and A. K. Lakshminarayanan

Table 4 Micro- and macrograph of tensile samples: (a) air welded conditions and (b) underwater
welded conditions
S. Macrograph Micrograph Observation
No.
1. Failure
occurs in
the TMAZ

2. Failure
occurs in
the base

3.2 Tensile Properties

The tensile strength of underwater friction stir welded ZE41 alloys is up to 189 MPa
and for air weld condition is 168 MPa. Maximum elongation of air welded sample
is 20% and for underwater weld is 8%. The normal elongation for magnesium ZE41
is 4%. In air weld condition, the heat input at the stir zone is higher and it causes
excessive grain growth. Therefore, the sample exhibits lower tensile strength when
compared to the weld done at underwater condition.
From Table 4, we can see that the failure occurs in the stir zone for air weld speci-
men and the fracture is located at the base metal for underwater welded specimen. It
denotes that the weld strength of the underwater specimen is higher than the strength
of air weld specimen. The fracture location of friction stir weld underwater has less
elongation than the air weld FSW. This is because of the formation of precipitation
on the welding region decelerates the elongation of grains in interface.

4 Conclusions

In this study, the investigation of UWFSW and FSW on MgZE41 alloy was
experimentally conducted in which the process parameter, resultant properties,
microstructure and tensile behaviour were studied.
Comparative Study of Friction Stir Welding and Underwater … 765

• Heat dissipation rate was high on the underwater friction stir welding since water
at room was in contact with magnesium plate; thus, thermal gradient developed
during welding was dissipated from plate. So the frictional heat between tool and
the specimen was absorbed by the surrounding water. So that there was no enough
time given for the growth of recrystallized grains than the air weld.
• On the other side, the heat dissipation rate in the normal air atmosphere was low.
And so sufficient time was given for grain refinement than the underwater condi-
tion. So it had enough time for grain growth near the interface. So the stir zone in
the air weld condition was quite higher than the underwater condition.
• The tensile strength of underwater friction stir welded ZE41 alloys was up to
189 MPa and for air weld condition was 168 MPa. Maximum elongation of air
welded sample was 20%, and for under water weld was 8%. The normal elongation
for magnesium ZE41 is 4%.
• The failure occured in the stir zone for air weld specimen, and it was located at
the base metal in underwater welded specimen.
• The heat input at the stir zone was higher, and it caused excessive grain growth.
Therefore, the sample exhibited lower tensile strength when compared to the weld
done at underwater condition.

References

1. Rajasekaran R, Lakshminarayanan AK, Vasudevan M, Vasantharaja P, Role of welding pro-

cesses on microstructure and mechanical properties of nuclear grade stainless steel joints. Proc
IMechE Part L: J Mater: Design Appl. https://doi.org/10.1177/1464420719849448
2. Esparza JA, Davis WC, Murr LE (2003) Microstructure-property studies in friction-stir welded,
thixomolded magnesium alloy AM60. J Mater Sci: Mater Electron 38(5):941–952
3. Zhang HJ, Liu HJ, Yu L (2011) Microstructure and mechanical properties as a function of
rotation speed in underwater friction stir welded aluminum alloy joints. Mater Des 32:4402–
4407
4. Esparza JA, Davis WC, Trillo EA, Murr LE (2002) Friction-stir welding of magnesium alloy
AZ31B. J Mater Sci Lett 21(12):917–920
5. Zhao Y, Wang Q, Chen H, Yan K (2014) Microstructure and mechanical properties of spray
formed 7055 aluminum alloy by underwater friction stir welding. Mater Des 56:725–30
6. Zhao Yong, Zhengping Lu, Yan Keng, Huang Linzhao (2015) Microstructural characterizations
and mechanical properties in underwater friction stir welding of aluminum and magnesium
dissimilar alloys. Mater Des 65:675–681
7. Huda Z, Edi P (2013) Materials selection in design of structures and engines of supersonic
aircrafts: a review. Mater Des 46:552–60
8. Mofid MA, Abdollah-Zadeh A, Malek Ghaini F (2012) The effect of water cooling during
dissimilar friction stir welding of Al alloy to Mg alloy. Mater Des 36:161–167
9. Koganti R, Lakshminarayanan AK, Ramprabhu T (2019) Understanding the effect of tool
rotational speed on microstructure and mechanical properties of friction stir processed ZE41
grade magnesium alloy. In: Advances in materials and metallurgy, Lecture notes in Mechanical
engineering. Springer Nature Singapore Pte Ltd. https://doi.org/10.1007/978-981-13-1780-
4_41
10. Sanchez C, Nussbaum G, Azavant P, Octor H (1998) Structural material properties, microstruc-
ture and processing. Mater Sci Eng A 252:269
766 S. Cyril Joseph Daniel and A. K. Lakshminarayanan

11. Neil WC, Forsyth M, Howlett PC, Hucthinson CR, Hinton BRW (2009) Corrosion of
magnesium alloy ZE41—the role of microstructural features. Corros Sci 51:387–394
12. Genghua C, Datong Z, Wen Z, Cheng Q (2015) Microstructure evolution and mechanical
properties of Mg–Nd–Y alloy in different friction stir processing conditions. J Alloy Compd
636:12–19
Characterizing the Tensile Deformation
Behavior of Friction Stir Welded
Dissimilar Joints Using Acoustic
Emission Technique

A. Venkatakrishna, A. K. Lakshminarayanan, K. Radhika,

and R. Rajasekaran

Abstract In the present investigation, P91 and 316LN dissimilar weld joints were
fabricated at the tool speeds of 600 and 900 rpm with 0.5 and 1 mm offset condi-
tions. For all mentioned welding conditions, the tensile testing was carried out at the
constant strain rate of 3.3 × 10−4 S−1 . A non-destructive acoustic emission (AE)
technique was used to analyze the tensile deformation behavior at elastic and plastic
deformation stages. The AE signals were recorded until the crack initiation and the
deformation behavior is compared with base metals. The obtained results were useful
to identify the deformation characteristics and joint integrity in different tool speed
and offset conditions.

Keywords Dissimilar weld joints · Tensile deformation · Acoustic emission ·

Strain rate · AE signals

1 Introduction

The dissimilar weld joints of P91 steel and low carbon and nitrogen enhanced stain-
less steel 316LN were widely used in pressure vessels, piping materials, and steam
generator applications in fast-breeder nuclear reactors [1, 2]. During sudden start-ups
and shutdowns of the nuclear reactor, high cyclic thermal stresses were generated due
to the mismatches in thermal expansion coefficients of the dissimilar weld metals.
This cyclic stresses are superimposed on the residual welding stresses and leads to
service failure/premature failure during long time service of the welded structure [3].
Therefore, by online monitoring, the material behavior through a wide range of non-
destructive testing techniques, the real-time information on material integrity can be
predicted and the failures of the welded structures can be controlled in advance.
Acoustic emission technique (AE) is a non-destructive technique that used to
monitor the structural integrity, dislocation kinetics, crack initiation and inclusions

A. Venkatakrishna (B) · A. K. Lakshminarayanan · K. Radhika · R. Rajasekaran

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, Tamil Nadu 603110, India
e-mail: venkatakrishnaa@ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 767

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_68
768 A. Venkatakrishna et al.

in various types of materials [4]. Due to its simplicity in testing, the deformation
characteristics have been analyzed in different kinds of materials such as brass [5],
copper [6], aluminum [7], and magnesium [8]. But only a few researchers [9–15]
have shown interest in characterizing the tensile behavior in a wide variety of steels
and stainless steels. During tensile deformation, the variation in AE signals will
depend on the physical, chemical, and the structural properties of the base materials.
The acoustic signals will generate high activity due to high precipitate content, low
dislocation densities, different grain sizes, and variation in carbon content in different
weld thermal cycles [9]. Khamedi et al. [10] investigated the acoustic behavior of
DP steel under tensile loading. The rate of deformation and modes of failure were
comparatively correlated with the use of AE frequency signals and the counts. Barat
et al. [11] studied the tensile deformation behavior of AISI-type 304LN SS in low
temperature, it is depicted that the AE counts were gradually increased due to the
evolution of martensite during low-temperature deformation. Han et al. [12] studied
tensile deformation behavior of Q345 steel weld and base metals. It is reported that
the crack initiation in weld was comparably higher than the base metal. It is due to
the presence of oxide inclusions and coarse grain structures in the weld samples.
Raj et al. [13] studied the systematic variation and a high source of AE signals
that generated during the prior stage of deformation in austenitic stainless steels;
it is due to the material dislocations and multiplication of existing dislocation, i.e.,
Frank-Read grain boundary during the tensile straining. Venkataraman et al. [14]
monitored the AE activity of 316 austenitic stainless steel, and the AE count rate
is varied with respect to different strain rates during tensile testing. In maximum
straining condition, the generated AE counts and signal frequency were high due
to high deformation rate and the source activation of acoustic signals. Haneef et al.
[15] investigated the tensile behavior of grade 316 austenitic stainless steel. The
AE parameters such as RMS voltage and dominant frequency were monitored in
different strain rates from 1.4 × 10−3 S−1 to 1.4 × 10−2 S−1 . The RMS voltage was
increased with the increase in the strain rate. The dominant frequency was decreased
during necking, but it is increased due to the increased dislocation velocity at high
strain rates.
From the above literature, it is identified that by using AE technique, the tensile
deformation behavior was characterized in various kinds of metallic and non-metallic
materials. But only a few studies were available on characterizing the weld defor-
mation behavior using the AE technique. Moreover, no work has been carried out on
monitoring the tensile deformation behavior of ferritic–martensitic–austenitic dis-
similar weld joints using a friction stir welding process. In this present study, an
attempt was made to study the acoustic emission characteristics during the tensile
deformation of P91-316LN dissimilar joints. The weld integrity in different tool
speed and offset conditions was characterized, and the differences in AE parameters
were compared with base metal studies.
Characterizing the Tensile Deformation Behavior of Friction … 769

Table 1 Physical properties of the base metals

Base metals Density (kg/m3 ) Specific heat Melting point (°C) Thermal
(J/kg K) conductivity
(W/mK)
P91 7.77 × 10−9 620 1420 26–28
316LN 7.99 × 10−9 470 1450 16.2–16.5

2 Experimental Details

2.1 Materials and Weld Specimens

The materials used in the present study were P91, a nuclear grade ferritic/martensitic
steel, and 316LN, a corrosion-resistant austenitic stainless steel. The physi-
cal properties of the base metals and its chemical compositions were listed
in Tables 1 and 2.
The normalized and tempered condition of P91 steel and solution annealed 316LN
was received in the form of 10-mm-thick plates. With the use of an electrical discharge
machine (EDM), the plates were sliced to the thickness of 3 mm. The sliced plates
were fabricated to the length of 125 mm and a breadth of 60 mm. Before conducting
the FSW trials, the machined plates were cleaned with methyl alcohol (CH3 OH)
solution, and the centerline of the weld plates was tightly clamped to each other.
The P91 steel plate was arranged in the retreating side, and the 316LN SS plate
was arranged in the advancing side. By using tungsten lanthanum oxide tool, FSW
process was carried out. The weld trials were carried out at the rotational speeds (ω)
of 600 and 900 rpm with 0.5 and 1 mm offset towards the P91 side. For all FSW trials,
constant welding speed (v) of 50 mm/min was maintained. The process parameters
used for the FSW process were shown in Table 3. The representation of the FSW
tool and its schematic view of the FSW process were shown in Fig. 1.

2.2 Acoustic Emission Monitoring and Testing Procedure

The weld joints and the base metals were machined according to ASTM E8M-09
standards for testing the tensile properties of the weld joint. The graphical represen-
tation of the tensile sample is shown in Fig. 2a. The tensile deformation behavior was
characterized by using 100 kN hydraulic tensile testing machine (Make: INSTRON,
Model: 8801), and the testing was accomplished at the constant strain rate of 3.3 ×
10−4 S−1 .
Two piezoelectric AE sensors (Make: Physical Acoustics, Model: micro 30) with
a frequency range of 150–400 kHz was fixed at the shoulder regions of the tensile
samples. A very small quantity of silicon grease gel was used as a sensor cou-
plant in gauge transition regions. To obtain proper signals without any disturbances,
 770

Table 2 Chemical compositions of the base metals measured in (wt%)

C Cr Mo Ni Si Mn N Cu P V Nb Fe (bal)
P91 0.09 9.04 0.95 0.42 0.28 0.37 0.04 0.15 0.01 0.23 0.09 88.3
316LN 0.02 18.0 2.74 12.3 0.54 1.97 0.08 0.19 0.04 – 0.005 63.9
A. Venkatakrishna et al.
Characterizing the Tensile Deformation Behavior of Friction … 771

Table 3 FSW process

Parameters Notations Units Process variables
parameters
Tool and pin Taper cylindrical
profile
Tool pin offset x mm 0.5, 1.0
Average axial F kN 32 kN
load
Rotational speed N rpm 600, 900
Welding speed v mm/min 50

Fig. 1 FSW tool, dimensions, and the welding process

petroleum-based lubricant gel was applied between the sensor interface and around
the specimen surface. During tensile loading, the AE sensor starts sensing the defor-
mation levels and generates the acoustic signals in the shoulder region. These signals
were captured and processed by pre-amplifier and 100–300 kHz range bandpass fil-
ter. An average threshold frequency was set to eliminate the external noises that
occurred during tensile testing. The AE signals produced during tensile deformation
were simultaneously recorded and processed by using an AE workstation (Make:
Physical electronics), and the tensile deformation results were evaluated by using
AEwin disp software v2.19. Finally, the fractured surfaces of the deformed tensile
samples were studied by using scanning electron microscopy (Make: Hitachi, Model:
3400 N). The graphical and pictorial views of the AE experimental setup were shown
in Fig. 2b.
772 A. Venkatakrishna et al.

Fig. 2 Schematic of tensile sample and AE experimental setup with block diagram

3 Results and Discussion

3.1 Tensile Properties and Joints Characteristics

The tensile properties of the base metals were illustrated along with the acoustic
emission characteristics as shown in Fig. 3. The tensile and yield strength of P91 steel
were high around 36 and 150 Mpa. But the rate of elongation was low around 32%
as compared to 316LN SS. In P91 steel (Fig. 3a), high tensile strength was observed
due to co-precipitation of MX and M23 C6 precipitates in ferrite and coarse martensite
grain boundaries [16]. In 316LN (Fig. 3d), the elongation percentage was increased
due to the effect of work hardening. Dissimilar joints between these two materials
exhibited different tensile properties in different tool speed and offset conditions
as shown in Figs. 4 and 5. In both 600 and 900 rpm weld conditions, the tool was
offset toward the P91 side. The tensile strength of 600 rpm 0.5 condition (Fig. 4a)
was reduced due to inadequate heat input and insufficient mixing of 316LN material
toward the P91 area. Therefore, the tensile strength of 600 rpm and 0.5 mm weld
condition was reduced around 110 Mpa than the base metal. Similarly, by increasing
the offset up to 1 mm toward P91 side (Fig. 4d), the weld region was not softened
Characterizing the Tensile Deformation Behavior of Friction … 773

Fig. 3 Tensile properties of the base metals and its AE characteristics

Fig. 4 Tensile properties of 600 rpm joints and its AE characteristics

properly. So, the defects such as micropores and tunnel defect were developed around
the 316LN HAZ side. Therefore, the tensile strength was drastically decreased around
315 Mpa. Moreover, in both offset conditions, the fracture was observed in the weld
region.
In 900 rpm weld with 0.5 offset condition (Fig. 5a), material mixing proportions
were good due to increased heat input. The obtained tensile strength was almost
equal to 316LN base metal properties. When compared to both base materials, a
slight drop in strength around 25 Mpa was noticed. It is due to the coarsening effect
774 A. Venkatakrishna et al.

Fig. 5 Tensile properties of the 900 rpm joints and its AE characteristics

of grains during welding. The yield strength was increased up to 66% than 316LN,
and the elongation percentage was increased up to 21% than the P91 base metal.
But in 1 mm offset condition (Fig. 5d), the mixing ratio of 316LN was reduced;
therefore, the tensile properties are comparatively reduced than 0.5 offset condition.
Moreover, in both offset conditions, ductile mode of fracture was observed in P91
HAZ region. The variation in tensile properties for both base metals and weld joints
was illustrated in Table 4.

Table 4 Tensile test results of the base metals and the weld joints
Strain rate 3.3 × 10−4 S−1
P91 316LN 600 rpm and 600 rpm and 900 rpm and 900 rpm and
0.5 mm 1 mm offset 0.5 mm 1 mm offset
offset offset
Tensile 618 582 479 303 569 554
strength
(Mpa)
Elongation 54 82 31 12 75 60
(%)
Yield 431 290 262 212 356 304
strength
(Mpa)
Characterizing the Tensile Deformation Behavior of Friction … 775

3.2 Acoustic Emission Characteristics During Tensile

Deformation

AE behavior during tensile testing was characterized in four regions; they are: (1) 0–
3% elastic deformation region, (2) 3–10% yield region, (3) UTS and necking region,
and (4) final fracture after UTS [15]. During tensile testing, the transient elastic AE
waves were generated from the shoulder region of the tensile sample and the rate of
deformation is identified in the form of counts. The emission characteristics during
AE analysis differ in three stages—they are (1) crack initiation, (2) crack propagation,
and (3) the final stage of failure. In the crack initiation stage, the emission character-
istics will be low, but in crack propagation stage, the emission characteristics will be
increased, and during the fracture stage, the AE signal and emission characteristics
will be in maximum peak level [17]. Similar to these emission characteristics, the
variation of AE counts with respect to time was identified in P91 and 316LN base
metals and its weldments.
In P91 base metal (Fig. 3b), the AE counts have been increased at the beginning
stage of plastic deformation, but sudden rise and fall of signals were observed till
the yield region; it is due to the movement of dislocations during straining. In crack
initiation region, the dislocation densities were increased and necking formation
was observed, and the amplitude signals were slightly increased with respect to the
acoustic signals. But at the stage of crack propagation, the dislocation motion and its
velocity were increased and the rise in AE signals was observed. With the increase of
AE counts at maximum straining condition, the duration of acoustic signals (Fig. 3c)
has increased to 70,000 μS. Generally, the AE signals generated during deformation
were mainly depended upon the microstructure of the metal, strain rate, and presence
of any inclusions [18]. Due to single-phase austenite microstructure in 316LN base
metal, AE signals are observed at the constant strain levels corresponding to the
phenomenon of macroyielding (Fig. 3e). But compared to P91 steel, the AE counts
were less during crack initiation. In the crack propagation region, AE signals were
generated due to the effect of dislocation motion and twinning. At the final stage of
fracture, the burst AE signals were observed due to the increase in source activation
signals at the constant strain rate. Therefore, the duration of acoustic with respect to
amplitude (Fig. 3f) was increased up to 140,000 μS.
In 600 rpm weld joint with 0.5 offset condition (Fig. 4b), the AE counts were
slightly increased until the homogenous plastic deformation stage. But at the begin-
ning stage of crack propagation, the AE counts were gradually increased until the
end of the tensile test. Due to high dislocation movements in the weld region, the
stress concentration was increased and the material was failed in an earlier stage as
compared to base metals. Therefore, the amplitude of AE signals was increased up
to 100 dB and the AE burst signals were increased until the minimum duration of
25,000 μS (Fig. 4c).
In 1 mm offset condition (Fig. 4d), the AE counts were increased at the beginning
stage of plastic deformation. Sudden rise and fall of AE peaks were continuously
monitored in the crack initiation region. But at the crack propagation region, the AE
776 A. Venkatakrishna et al.

counts have increased again till the occurrence of a failure in the tensile specimen.
The sample was failed without necking, so it is concluded that the sample is failed due
to the presence of increased dislocations in the weld region. Therefore, the material
was failed in an earlier stage as compared to 600 rpm 0.5 offset weld condition.
Similarly, the AE signals were constantly increased up to 100 dB, and the AE burst
signals were increased until the minimum duration of 25,000 μS (Fig. 4e).
In 900 rpm weld joint with 0.5 offset condition (Fig. 5b), the AE counts were very
low until the final stage of elastic deformation, but a sudden increase of AE signals
was noticed due to the decohesion of inclusions during crack initiation. The recorded
AE counts were too weak for a long time up to 800 s of testing time (i.e., up to neck-
ing), because the rate of deformation of this weld joint was high as compared to other
welds. Due to improved tensile properties, AE counts was increased upto 160,000 μs
during crack propagation stage of the tensile sample. Similarly, in 1 mm offset con-
dition (Fig. 5f), a peak of AE emission was noticed due to the presence of coarse
grains in the plastic deformation region. But other than the obtained AE signals,
the counts were almost the same as the 0.5 mm offset condition. But the amplitude
frequency signals were reduced up to 40 dB than the 0.5 mm offset condition; this
change is attributed with the effect of moving dislocations associated with the tensile
deformation.

4 Conclusion

Using the AE technique, tensile deformation behavior has been successfully moni-
tored for P91-316LN FSW joints that welded at the tool speed of 600 and 900 rpm
with 0.5 and 1 mm offset conditions. From the experimental investigations, some
important results were observed and it is listed below as follows:
1. In 600 rpm welds, AE signals were generated in yielding stage of deformation—
it is mainly associated due to the increased fraction of dislocation density in the
weld region.
2. Due to the improper coalescence of dissimilar materials in 600 rpm weld con-
dition, the ductility of the joint has been comparably reduced in 1 mm offset
condition. Hence, AE counts and signal duration were found to be increased at
the beginning stage of plastic deformation.
3. In 600 rpm weld joints, crack initiated in weld region, and for 900 rpm joints,
the fracture occurred in P91-HAZ region for both offset conditions. The 900 rpm
joints exhibited good tensile properties as compared to 600 rpm joints. But com-
pared to base metals, the strength of 900 rpm joints was reduced due to the
presence of inclusions and coarse grains in the weld joints.
4. During necking, frequencies and amplitude signals were found to be increased
in 900 rpm weldments; it is attributed to the increase of the material dislocations
during deformation at the constant strain level.
Characterizing the Tensile Deformation Behavior of Friction … 777

5. The amplitude signals and AE waves generated during tensile deformation were
varied in both 600 and 900 rpm weld joints. In 600 rpm condition, the peaks
were totally varied due to the presence of defects in 1 mm offset condition. But
in 900 rpm, slight changes in AE counts and amplitude were observed during
deformation in both offset conditions.
Therefore, it is concluded that by increasing the tool speed with a low offset value,
the defect formation in dissimilar metals has been reduced and the joint integrity
has been improved. Moreover, by using the AE technique, overall material life can
be predicted in earlier stages of deformation; therefore, the joint failures can be
controlled in advance.

Acknowledgements Authors are thankful to Mr.T Haneef, Scientific Officer D, and Dr. C.K
Mukhopadhyay Head, NDTD/MMG, IGCAR, Kalpakkam, India, for providing the acoustic emis-
sion test facility in their laboratory. The authors are thankful to Mr. Thirukumaran, Research fellow,
NDTD/MMG, IGCAR, for providing constant encouragement and support during the experimental
analysis.

References

1. Zinkle SJ, Ghoniem NM (2011) Prospects for accelerated development of high-performance

structural materials. J Nucl Mater 417:2–8
2. Chai M, Duan Q, Hou X, Zhang Z, Li L (2016) Fracture toughness evaluation of 316LN
stainless steel and weld using acoustic emission technique. ISIJ Int 56(5):875–882
3. Bhaduri AK, Venkadesan S, Rodriguez P (1994) Transition metal joints for a steam
generators—an overview. Int J Press Vessels Pip 58:251–265
4. Ríos-Soberanis CR (2011) Acoustic emission technique, an overview as a characterization tool
in materials science. J Appl Res Technol 9(3):367–379
5. Pattnaik A.B. Jha B.B and Sahoo R.: Effect of strain rate on acoustic emission during tensile
deformation of α-brass. Materials Science and Technology 29(3),294–299 (2013)
6. Hunaidi O, Chu WT (1999) Acoustical characteristics of leak signals in plastic water
distribution pipes. Appl Acoust 58:235–254
7. Skalski VR, Lyasota IM (2014) Features of acoustic-emission signals during the initiation of a
fatigue failure in a welded joint of an aluminum alloy of the Al Cu-Mn system. Russ J Nondestr
Test 50(2):120–126
8. Máthis K, Chmelik F, Trojanová Z, Lukác P, Lendvai J (2004) Investigation of some magnesium
alloys by use of the acoustic emission technique. Mater Sci Eng A 387–389:331–335
9. Wadley HNG, Scruby CB (1991) Cooling rate effects on acoustic emission-microstructure
relationships in ferritic steels. J Mater Sci 26(21):5777–5792
10. Khamedi R, Fallahi A, Refahi Oskouei A (2010) Effect of martensite phase volume fraction on
acoustic emission signals using wavelet packet analysis during tensile loading of dual phase
steels. Mater Design 31(6):2752–2759
11. Barat K, Bar HN, Mandal D, Roy H, Sivaprasad S, Tarafder S (2014) Low-temperature tensile
deformation and acoustic emission signal characteristics of AISI 304LN stainless steel. Mater
Sci Eng A 597:37–45
12. Han Z, Luo H, Cao J, Wang H (2011) Acoustic emission during fatigue crack propagation in
a micro-alloyed steel and welds. Mater Sci Eng A 528(25–26):7751–7756
13. Raj B, Jayakumar T (1991) Acoustic emission during tensile deformation and fracture in
austenitic alloys. ASTM-STP-1077, p 218
778 A. Venkatakrishna et al.

14. Venkataraman B, Mukhopadhyay CK, Raj B (2004) Effect of variation of strain rate on thermal
and acoustic emission during tensile deformation of nuclear grade AISI type 316 stainless steel.
Mater Sci Technol 20:1310–1316
15. Haneef T, Lahiri BB, Bagavathiappan S, Mukhopadhyay CK, Philip J, Rao BPC, Jayakumar
T (2015) Study of the tensile behavior of AISI type 316 stainless steel using acoustic emission
and infrared thermography techniques. J Mater Res Technol 4(3):241–253
16. Zhang S, Melfi T, Narayanan BK (2016) Effects of precipitates on mechanical properties of
P91 submerged arc welds. Sci Technol Weld Joining 21(2):147–156
17. Yu J, Ziehl P, Zárate B, Caicedo J (2011) Prediction of fatigue crack growth in steel bridge
components using acoustic emission. J Constr Steel Res 67(8):1254–1260
18. Ono K (2005) Current understanding of mechanisms of acoustic emission. J Strain Anal Eng
Design 40(1):1–15
Study of Infrared Thermography
on Tensile Behavior of Laser Beam
Welded 316LN Austenitic Stainless Steel
R. Rajasekaran, A. K. Lakshminarayanan, A. Venkatakrishna,
and K. Radhika

Abstract In this investigation, the tensile behavior of laser beam welded 316LN
austenitic stainless steel was studied by infrared thermography (IRT) and compared
with the base metal tensile behavior. Initially, microstructural characterization of
base metal and weld bead was carried out by optical microscopy (OM). Base metal
comprises an average grain size of 60 μm and LBW exhibited refined grains at
the fusion zone. Very narrow HAZ around 15 μm was recorded at LBW interface.
Tensile samples were prepared as per ASTM standard E8, and the test was conducted
at room temperature of 26.5 °C. A strain rate of 4.4 × 10−4 s−1 was used during
the tensile test of the base metal and LBW samples. Temperature variations of the
base metal and weld bead were recorded by an infrared camera at different stages
of deformation. Compared to LBW sample base metal sample deformed more and
displayed the percentage of elongation as 64.14%, whereas LBW sample displayed
less percentage of elongation around 53.2% due to lesser deformation as compared
to the base metal. This is attributed to grain refinement during rapid solidification
of LBW process at fusion zone. High temperatures were recorded on the base metal
(39.2 °C) and center of the LBW fusion zone (30.8 °C) before just the time of fracture.
This high temperature of the base metal over LBW sample indicates that the base
metal deformed more compared to laser beam welded sample.

Keywords Infrared thermography technique (IRT) · Laser beam welding (LBW) ·

316LN austenitic stainless steel · Stain rate · Deformation

1 Introduction

Austenitic stainless steels are used widely in nuclear power plants for their excellent
mechanical and corrosion resistance properties. Many structural components, except
the nuclear core structures, are built by 316LN grade austenitic stainless steel for

R. Rajasekaran (B) · A. K. Lakshminarayanan · A. Venkatakrishna · K. Radhika

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, Tamil Nadu 603110, India
e-mail: rajasekaranr@ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 779

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_69
780 R. Rajasekaran et al.

its superior mechanical and corrosion resistance property at high temperatures [1].
During the structural component fabrication process the joining of 316LN together is
done using conventional GTAW process. Many research works are going on joining
of 316LN using high energy density and solid-state welding techniques [2, 3]. Joining
of 316LN using the LBW process could improve mechanical properties in the fusion
zone. Improvement in toughness property by LBW process on 409 M ferritic stainless
steel was reported in earlier research work [4]. During the service, 316LN joints
may experience different types of load/stress conditions. This phenomenon leads
to deformation in that particular material as well as joints. Hence, the study of the
deformation behavior of 316LN base and its joints is essentially one.
Infrared thermography is a contactless nondestructive technique used in many
engineering applications due to the accurate monitoring of temperature variations in
fast inspection rates [5]. Assessment of material deformation is successfully carried
out on metals, polymers, and composite materials using infrared thermography tech-
nique [6–8]. This is followed by the concept of infrared radiations from a deforming
object detected by IR camera, and the radiations are converted in terms of temper-
ature patterns. These temperature patterns are stored and analyzed using computer
workstations. Deformation distribution in the material and its severity is evaluated by
the intensity of the temperature patterns. In this experimental investigation, IRT has
been used for studying tensile deformation of laser beam welded 316LN austenitic
stainless steel.

2 Experiment

2.1 Material

Austenitic stainless steel 316LN was taken to study about IR thermography during
the tensile test. The details of the chemical composition of 316LN austenitic stainless
steel are given in Table 1. Plate form of the base material with a length of 300 mm,
the width of 75 mm, and a thickness of 3 mm was prepared to make weld joint.
Edge preparation was made by square butt configuration. Argon with a purity level
of 99.995% was used as a purging gas during the welding process. The following
process parameters were used for fabricating LBW joint: (i) laser power: 2.5 KW
and (ii) welding speed: 1500 mm/min.

2.2 Microstructural Evolution

Metallographic samples of base metal and weld joints (transverse side) were polished
up to 2500 girt emery sheet. Further samples were polished by using 0.5-μm diamond
paste to eliminate micro-sized impurities. After polishing, etching process was done
Table 1 Chemical composition of 316LN austenitic stainless steel
Weight % of elements
C Cr Ni Mo Mn Si S P N Fe
316LN 0.026 17.8 11.7 2.4 1.6 0.45 0.009 0.026 0.095 Balance
Study of Infrared Thermography on Tensile Behavior of Laser …
781
782 R. Rajasekaran et al.

by a solution which contains 1:3 ratios of HNO3 to HCl as per ASTM standard
designation E407. Microstructures of base metal and the LBW fusion zone were
captured by olymphus BX 51 M optical microscope.

2.3 Infrared Thermography

Figure 1 depicts experimental setup of IRT during the tensile test. Figure 2 depicts
samples of base metal and LBW after the tensile test. Tensile test of the base metal
and LBW samples were performed by Dartec, UK, 100KN hydraulic servo motor
computer-controlled universal testing machine (UTM). 2 mm/min crosshead speed

Fig. 1 Experimental setup

of IRT during tensile test,
a universal testing machine
and IR camera and b tensile
sample of LBW

Fig. 2 Samples after the

tensile test
Study of Infrared Thermography on Tensile Behavior of Laser … 783

Fig. 3 Microstructures, a 316LN base, b LBW: fusion zone–base metal interface

was maintained during the tensile test. Thermographic images were recorded by the
IR camera model: CEDIP silver 420 with 25 m/C temp sensitivity.

3 Result and Discussion

3.1 Microstructural Characterization

Figure 3 depicts the microstructural morphologies of 316LN base metal and LBW
joint. The base metal consists complete austenite phase with grain size around 60 μm.
Annealing twins were clearly observed in the base metal microstructure. LBW fusion
zone exhibited the primary austenite mode of solidification. Following morphologies
were recorded at the fusion zone of the LBW process viz: (i) epitaxial grain growth,
(ii) equiaxed dendrite, and (iii) columnar structure. Higher welding speed of LBW
process resulted in the formation of very narrow weld bead and HAZ around 15 μm
[9].

3.2 Deformation Study by IR Thermography

Thermographic images of the base metal at different stages during tensile deformation
are shown in Fig. 4a. The highest temperature around 39 °C was recorded at stage
III (before fracture) in the base metal. Temperature curve and its different stages are
explained in Sect. 3.3. Temperature evolution curves along the axial length of the
base metal sample at different stages are shown in Fig. 4b. Necking was observed
almost center of the gauge portion in the base metal. This is clearly correlated with
the peak point of the temperature evolution curve.
Thermographic images of the LBW sample at different stages during tensile defor-
mation are shown in Fig. 5a. The highest temperature around 32 °C was recorded
784 R. Rajasekaran et al.

Fig. 4 a Thermographic images during tensile deformation of base metal and b evolution of
temperatures in axial length of the base metal

Fig. 5 a Thermographic images during tensile deformation of LBW joint and b evolution of
temperatures in axial length of LBW sample

at stage III (before fracture) in LBW sample. Temperature evolution curves along
the axial length of the LBW sample (Fig. 5b) revealed that the position of the peak
temperature is away from the center of the gauge portion (LBW fusion zone).

3.3 Tensile Behavior Versus Peak Temperature

Transverse tensile properties of the base metal and the LBW joint are given in Table 2.
Base metal showed more percentage of elongation as compared to the LBW tensile
sample. This phenomenon is mainly attributed to the grain refinement at the fusion
zone during the solidification process. Failure of LBW sample occurred at the base
metal side around 8 mm from the fusion zone. This failure location indicating that the
LBW joint contains more tensile strength compared to the base metal. Further micro-
tensile samples [10] were extracted from LBW fusion zone along weld direction to
asses UTS of LBW joint. From UTS evaluation, it was seen that the LBW joint
exhibited 11.7% more strength (688 MPa) compared to the base metal UTS value.
Study of Infrared Thermography on Tensile Behavior of Laser … 785

Table 2 Tensile properties of base and LBW joint

S. No. Sample Crosshead Strain Yield UTS % of
specification speed rate (s−1 ) (MPa) (MPa) Elongation
(mm/min)
1. 316LN Base 2 4.4 × 319 614 64.14
10−4
2. LBW 2 4.4 × 320 619 53.2
10−4

Evolution of peak temperature with the stress plot of the base metal and the LBW
samples is shown in Figs. 6 and 7, respectively. This temperature curve is divided into
four main stages based on the deformation behavior of the material. They are (i) linear
temperature drop stage, (ii) temperature rise stage, (iii) rapid temperature rise stage,
and (iv) final temperature decline stage [11, 12]. The considerable temperature drop

Fig. 6 a Evolution of peak temperature of the base metal during deformation with respect to time
and b stress versus strain curve of base metal

Fig. 7 a Evolution of peak temperature in LBW sample during deformation with respect to time
and b stress versus strain curve for LBW sample
786 R. Rajasekaran et al.

was observed in stage I due to the thermoelastic effect of the material. This effect
occurred up to yield point of the base metal and LBW sample. When the plastic
deformation started, the continuous rise in temperature was noticed on both base
metal and LBW sample. This is attributed to the generation of heat due to plastic
deformation effect [13]. Sudden rise in temperature occurred during the necking and
cracking stage of the samples. This is attributed to the local plastic deformation of
the initiated crack tip during necking. On stage III, the base metal sample exhibited
39.2 °C, whereas the LBW sample exhibited a peak temperature of 32.4 °C. But the
fusion zone of the LBW sample experienced peak temperature of 30.8 °C before
fracture (stage III). Lesser deformation in the gauge portion of the LBW sample
compared to the base metal sample is the reason for this temperature difference.

4 Conclusions

• Higher welding speed and solidification behavior of LBW process influenced the
formation of refined grains at fusion zone compared to the base metal grains. This
grain refinement increased strength in the fusion zone of the LBW process over
the base metal.
• During tensile deformation study, peak temperatures were recorded just before to
the time of fracture for both LBW and base metal samples. This phenomenon is
attributed to the increase of temperature with respect to increasing deformation in
metals. Moreover, failure locations were revealed in terms of temperature patterns
in advance before the occurrence of fracture during the tensile test on different
samples.
• From IR thermography images, peak temperatures of 39.2 and 30.8 °C were
observed at base metal and LBW fusion zone, respectively. This increased tem-
perature at the base metal indicates that the base metal has undergone more
deformation compared to LBW sample.

References

1. Mannan SL, Chetal S et al (2003) Selection of materials for prototype fast breeder reactor.
Trans Indian Inst Met 56(2):155–178
2. Vignesh A, Vijay Prakaash VG, Lakshminarayanan AK (2015) Improving wear resistance of
AISI 316LN austenitic stainless steel using friction stir processing. Appl Mech Mater 787:421–
425
3. Rajasekaran R, Lakshminarayanan AK, Vasudevan M, Vasantharaja P, Role of welding pro-
cesses on microstructure and mechanical properties of nuclear-grade stainless steel joints. J
Mater: Design Appl. In: Proceedings of the institution of mechanical engineers, part L. SAGE.
https://doi.org/10.1177/1464420719849448
4. Lakshminarayanan AK, Balasubramanian V (2012) Characteristics of laser beam and friction
stir welded AISI 409M ferritic stainless steel joints. J Mater Eng Perform 21:530–539
Study of Infrared Thermography on Tensile Behavior of Laser … 787

5. Bagavathiappan S, Lahiri BB, Saravanan T, Philip J, Jayakumar T (2013) Infrared thermography

for condition monitoring—a review. Infrared Phys Technol 60:35–55
6. Punnose S, Mukhopadhyay A, Sarkar R, Alam Z, Das D, Kumar V (2013) Determination of
critical strain for rapid crack growth during tensile deformation in aluminide coated near-α
titanium alloy using infrared thermography. Mater Sci Eng, A 576:217–221
7. Górka Jacek, Janicki Damian, Fidali Marek (2017) Thermographic assessment of the HAZ
properties and structure of thermo mechanically treated steel. Int J Thermophys 38(183):1–21
8. Kumar Jalaj, Baby Sony, Kumar Vikas (2008) Thermographic studies on IMI-834 titanium
alloy during tensile loading. Mater Sci Eng, A 496:303–307
9. Lakshminarayanan AK, Balasubramanian V (2012) Influences of welding processes on
microstructure and mechanical properties of modified 12 wt% Cr ferritic stainless steel. Int
J Manuf Res 7(4):331–352
10. Lakshminarayanan AK, Balasubramanian V (2012) Evaluation of microstructure and mechan-
ical properties of laser beam welded AISI 409M grade ferritic stainless steel. J Iron Steel Res
Int 19(1):72–78
11. Haneef Thodamrakandy et al (2015) Study of the tensile behavior of AISI type 316 stainless
steel using acoustic emission and infrared thermography techniques. J Mater Res Technol
4(3):241–253
12. Yan Z, Zhang H, Chen P, Wang W (2017) Anisotropy of fatigue behavior and tensile behavior
of 5A06 aluminum alloy based on infrared thermography. J Wuhan Univ Technol Mater Sci
Ed 32(1):155–161
13. Kutin Marina, Ristic Slavica, Burzic Zijah, Puharic Mirjana (2010) Testing the tensile features
of steel specimens by thermography and conventional methods. Sci Tech Rev 60(1):66–70
Production and Operation Management
Evaluation of Ergonomics Issues
in Repetitive Scrap Handling Work
in Automobile Industries

Aravind Babu Dasari and Dhinesh Balasubramanian

Abstract Ever since the origin of the evolution of human being either named as
ape or man, striving to survive with safety, by becoming wise on live hazards, he
has been innovating as many ways and means for the safe living/working. To this
day, the struggle for safety is existing but in a sophisticated manner, as he grew from
being wise to being an intellectual. This paper discusses the ergonomics issues in an
automobile industry handling the hard substance, among the engineering industries,
as the manual scrap handling system is one of the most critical safety concern aspects
which includes more manual effort, due to its impact on the manpower expenses,
medical expenses, personal protective equipment expenses, and time consumption.
Ergonomic assessment, as a tool and method for analyzing human activities and
their interactions with the surrounding environment, is thus crucial for designing
operations and workplaces that achieve high safety. In the engineering industry, how-
ever, the constant repetitive work environment and laborious tasks cause traditional
approaches to ergonomic analysis, such as manual observations and measurements
to require substantial time and effort to yield reliable results. This study mainly con-
centrates on scrap handling because of its repetitive actions with heavy loads and
also explores the adaptation and integration of various existing methods for data
collection, analysis, and output representation potentially available for comprehen-
sive ergonomic analysis. The proposed framework integrates the 3DSSP’s (Static
Strength Prediction) localized fatigue report for the inputs such as measurement
of body angles using goniometer and NIOSH (National Institute of Occupational
Safety and Health) weight lifting equation. The proposed framework is demonstrated
through a case study using data from on-site scrap handling system through 3DSSP
software.

A. B. Dasari (B) · D. Balasubramanian

Mepco Schlenk Engineering College, Sivakasi, India
e-mail: daravindb6@gmail.com
D. Balasubramanian
e-mail: dhineshbala91@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 791

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_70
792 A. B. Dasari and D. Balasubramanian

Keywords 3DSSP software · NIOSH weight lift equation · AS-IS model ·

Goniometer · Localized fatigue report · Manual observation-based ergonomics
analysis

1 Introduction

The engineering industry is labor- and machine-intensive, and worker activities can
significantly affect the success of the engineering industry and scrap handling opera-
tions. Labor is one of the most crucial resources and has the highest direct impact on
the outcomes in the industry, including time and cost. Furthermore, labor operations
in scrap handling involve physically demanding motions and tasks that frequently
expose workers to risk in their working environments, leading to a rate of injuries
and chronic injuries that are among the highest of any engineering industry’s scrap
handling system.
The risk involved in the repetitive manual scrap handling system gets its peak with
respect to the duration of the work, work atmosphere, distance between the loading
and unloading region, and loading to unloading pathway of the scrap movement.
The repetitive manual scrap handling system leads to the considerable effects on the
expenses such as personal protective equipment (PPE) and medical expenses.
Golabchi et al. [1] state that the constantly changing work environments and
laborious tasks subject traditional approaches to ergonomic analysis such as man-
ual observations and measurements and his proposed framework used sensing for
data collection, action recognition, and simulation modeling, by using different tech-
niques that could facilitate the assessment of manual operations and could enhance
the implementation of ergonomic practices during the work by reducing time, com-
plexity and effort required, improvement achieved by proposing automated, simple
and reliable analysis of manual operation.
Munck-Ulfsfält et al. [2] inferred that the ergonomics work is not a separate
entity, but it is based on a strategy and much easier to get the management and other
employees to realize, accept, and understand the risk involved in ergonomics work.
Training on the load ergonomics concerned will be the key to success.
Dul et al. [3] argued that by contributing the shared goals of business performance,
ergonomists would also be better able to reach their traditional health and safety
objectives and also discussed the organizational strategies and business goals to
which ergonomics contribute.
Leung et al. [4] studied the impact of many organizational stressors and stress on
construction work safety behaviors and injury incidents. To achieve this, they studied
a number of workers, also by using factor analysis, identified the major organizational
stressors (inappropriate safety equipment, unfair reward and treatment, lack of goal
setting and poor physical treatment, lack of provision of training), two types of
stressors (physical and emotional), and safety behaviors and concluded that setting
personal, long-term, and specific work goals with the workers would reduce the
injury incident rate directly.
Evaluation of Ergonomics Issues in Repetitive Scrap Handling … 793

Golabchi et al. [5] discussed the approach that enabled modeling of manual activ-
ities at human motion level in order to provide standard task durations in addition to
initial insight into the level of ergonomic risks associated with a planned production
scenario.
Golabchi et al. [6] proposed the concept by making the approach with motion
data from the 3D model of the jobsite to evaluate the risk factors that probably could
produce more physical loads on the human body through biomechanical analysis.
The framework was proposed after analyzing the jobsite ergonomic safety.
Tran et al. [7] introduced the metric-based learning approach for human activity
identification, with the abilities to reject invisible actions, and discussed the vital
role of invisible actions when applying human activity recognition to real-world
applications.
Wei et al. [8] stated that their proposed model could generate physically realistic
animations that react to external forces or changes in physical quantities of human
bodies and interaction environments and evaluated their system by comparing against
ground-truth motion data and alternative methods.
Case et al. [9] discussed the digital human modeling practices for the last four
decades which is the first usable tool and still existing for the designing of the
ergonomics-related issues.
This current study deals with the procedure of analyzing and evaluating the issues
in the ergonomics in the automobile industry, analyzing the number of machines,
plant layout, and the frequency of the scrap produced from different operations,
gathering the anthropometric details through personal interview, video recording the
manual scrap handling process for analyzing AS-IS model through MOST method,
measuring the body angles using goniometer as input to the 3DSSP software for
analyzing the load on the body parts and their ability of bearing the load, evaluating
the RWL (Recommended Weight Limit) and LI (Lifting Index) for safe lifting weight,
and giving the remedial action according to the NIOSH evaluation technique.

2 Case Study and Analysis

The main activities carried out to complete the task include picking up s45c grade
steel scrap from the trolley and throw it into the scrap dump yard. As the first step, the
work was video graphed to extract time lapses and types of activities and then angles
of the worker’s body joints were measured during his bending, picking, twisting, and
throwing work postures. By using goniometer, the body angles were measured.

Details of the Machinery in the Automobile Industry

Figure 1 the graph gives information about the number of machines present in
the industry for a particular operation which produces the scrap and also gives
the information about the frequency of the scrap unloading per shift for different
operations.
794 A. B. Dasari and D. Balasubramanian

Fig. 1 Number of machinery and frequency of scrap with respect to the operation

2.1 Safety Assessment Module

The biomechanical analysis component of the framework enables the evaluation

of an operation by examining the loads and their exertion on the human body and
comparing them to safe limits. The results can be used along with productivity
analysis output to improve the operation and select the optimal remedial action.
Table 1 gives the details about the size of the different types of scrap produced from
a different type of operations in the automobile industry after its particular operation.
S45C grade low-carbon steel is the scrap produced from different machines.
Table 2 gives the details about the number of exertions taken to unload the scrap
from the trolley with respect to the scrap produced from their particular operation.
Figure 2 shows the exact replica of the manual scrap unloading procedure followed
in the automobile industry. The practice involves bending, holding, picking, and
throwing operations. The whole scrap is unloaded from the trolley of the above
shown size, which is made of MS material weighing about 100 kg (Fig. 3).

2.2 Activities Sequence and Duration

Table 3 gives the details about the time lapse used for the MOST method of analysis,
which represents the duration in seconds taken for the single exertion to get completed
Evaluation of Ergonomics Issues in Repetitive Scrap Handling … 795

Table 1 Different types of scrap from different operations for manual scrap handling
S. No. Operation Scrap sample
1 Facing

2 Turning

3 Turning and ID boring

4 VMC-multi-angle cut

5 Boring

Table 2 Number of exertions

S. No. Operation name No. of exertions for
for unloading different types
unloading
of scrap
1 Facing 550
2 Turning 460
3 Turning and ID boring 225
4 VMC-multi-angle cut 220
5 Boring 420

in the sequence, the actions for the scrap handling being identical for the all scraps
produced from different operations.

AS-IS Model
The AS-IS model of the plant layout gives the detailed description of the unloading
work pathway from its source to destination, besides the details like distance. As per
the above plant layout, the sources of the unloading trolley are not the same, but the
796 A. B. Dasari and D. Balasubramanian

Fig. 2 Manual operation done by the worker

Fig. 3 AS-IS model representation of the factory layout, for scrap collection

destination of unloading the trolley is the same. The pathway is drawn connecting
the individual source to the destination.

Analysis of the Direct and Indirect Expenses Due to the Inappropriate

Ergonomics
Figure 4 gives the details about the average monthly usage of the expensive gloves for
3 scrap unloading workers used for handling the scrap unloading process; because of
the variation in the method of unloading (crowbar and hand unloading), irrespective
of unloading time, the unloading system depends on the shape of the scrap. Around
Evaluation of Ergonomics Issues in Repetitive Scrap Handling … 797

Table 3 Details of the action sequence and duration of the manual scrap lifting operation
S. No. Activity sequence Duration
1 @00 h:00 m:01 s

Bend
2 @00 h:00 m:04 s

Hold
3 @00 h:00 m:05 s

Pickup
4 @00 h:00 m:06 s

Throw

45–50 expensive gloves are used on an average for the work of scrap unloading. The
average life span of the gloves used is hardly 2 days.
Figure 5 gives the monthly medical expenses in a year for 3 scrap unloading
workers; the inference from the graph is that on an average the monthly medical

Fig. 4 Number of gloves used by the scrap handling worker during a month
798 A. B. Dasari and D. Balasubramanian

expenses which is been referred from the monthly medical expenses bill of the
particular workers are approximately 1500 Indian rupees.
Figure 6 gives the information about the average of different types of medical
expenses incurred in a month; the inference from the pie chart is that the medical
expenses are major for the hip and torso pains.

Fig. 5 Medical expenses of scrap handling workers during a year

Fig. 6 Pie chart for the different types of medical expenses for a month
Evaluation of Ergonomics Issues in Repetitive Scrap Handling … 799

3 Methodology

This study proposes and tests an integrated framework that couples data acquisition
and 3D static simulation program with an analysis of manual operations to enable
effective evaluation of those manual operations for a comprehensive ergonomic anal-
ysis. The three objectives are (1) exploring the data associated with the different
angles of the body, (2) examining possibility and applicability of sharing data among
those different methods, and (3) testing, feasibility, and effectiveness of integrat-
ing the various methods. The framework is composed of the following modules:
simulation, as-is modeling, and safety assessment. This analysis was held by gath-
ering information about the conditions in the work environment through manual
measurement.
The workers’ normal activities at the work site were video graphed, and then
extracts were obtained of the type of work, the sequence of operation, and duration
of tasks and were used to build a simulation and get analysis report of the particu-
lar operation. The simulation made through 3DSSP software served to evaluate the
static strength of the operation and gave the information about the strength or pressure
applied on the particular parts of the body because of the particular operation and
helped to evaluate the duty cycle based on the number of and duration of exertions.
This 3DSSP 7.0 simulation program has the advantage of built-in range manipula-
tor for duty cycle limit, exertion limits, etc., according to American Conference of
Governmental Industrial Hygienists (ACGIH) 2016 (Fig. 7).

3.1 Manual Observation-Based Ergonomic Analysis

To carry out a thorough ergonomic analysis, information about the effects of physical
activities on a worker’s body needs to be available. Main contributors to those effects
are the type, duration, and sequence of manual tasks. A complete ergonomic analysis
involves evaluating the motions and postures of working people and the physical
attributes of job site to assess present work conditions and propose the new designs
for manual operations (safe conditions) and working spot (work spot dimensions).
To carry out assessments, the ergonomists need to go through the three stages: (i)
data collection, (ii) data analysis, and (iii) interpretation and presentation of results.
Ahead of data collection, the ergonomist has to plan the analysis process and
define the strategy based on certain conditions of work being analyzed. The required
inputs for each relevant data are collected, traditionally by observing the subjects
(posture, anthropometry), their motions during work (leaning, bending), and the
work environment particulars (workbench, tools, equipment).
After the completion of data collection, the ergonomist uses the gathered data to
conduct an ergonomic evaluation using tools such as ergonomics assessment check-
list (RULA, ROSA) and time and motion studies (MTM, MOST), giving the details
800 A. B. Dasari and D. Balasubramanian

Sensing
• Video Camera • Visual Checking
• Direct Interview • Goniometer measurement

Individual personal Study on the plant and

Interview for machine layout their
Anthropometry working

Video recording (time Scrap Frequency

lapse for work) analysis Report
and
Body angles Analysis on Expenses analyzing
measurement at work spent on Medical and
with Goniometer PPE. correcƟve
acƟon
Body angles AS-IS modeling for
measurement at work the factory layout for
with scrap handling
Goniometer operation.

Body angles as input to NIOSH weight lifting

3DSSP equation for studying
simulationsoftware. the safe working load

Analysis Report for 3D Low Back

Compression, Percentage of population
capable

Fig. 7 Methodology for the evaluation and analysis of ergonomics in the automobile industry

like worker’s posture (body angles/joint angles) along with the frequency and dura-
tion of exposure allowing posture-based tools to report on the level of ergonomic
risk associated with the task.
This work involves three stages of the ergonomics analysis. They are (i) data acqui-
sition, (ii) analysis (simulation modeling), and (iii) representation (result/corrective
action).
Evaluation of Ergonomics Issues in Repetitive Scrap Handling … 801

3.2 MTM Method Study

Methods Time Measurement (MTM) is the motion time system that is primarily used
in industrial settings to analyze the methods to perform any manual task or operation
and as a product of analysis, set the standard time that a worker should complete that
particular task.

3.3 MOST Method Study

MOST systems are designed to provide the optimal combination of speed, detail,
and accuracy of the work/operation. It was designed to be much faster than the
conventional work measurement techniques such as time study [10].

Integrated Ergonomic Analysis

To carryout thorough ergonomic analysis, the availability of information regarding
the effects of physical activity is imperative. Physical activity type, duration, and
sequence of the manual task are the main contributors to the physical effect.
Ergonomists rely on human judgment and estimates in acquiring data when
designing the new operations for new or prospective workplaces, which can lead
to uncertain information. This issue can be addressed by linking simulation model.
Predetermined motion time systems (PMTSs), which enable calculating the stan-
dard duration of the physical/manual task based on the involved movements, as this
simulation environment allows not-yet-existing scenarios to be conveniently modeled
and explored.

Simulation Model
To create a simulation model of manual operation and analyze its operational activ-
ities, either human observation or sensing methods have to be used to gather the
required inputs. Human observation typically requires time, effort, and expertise,
but adding the human observation along with simulation software enables more
facts regarding the ergonomics issues of the particular work.
The developed simulation model using 3DSSP software represents the current
status of an ongoing operation which is also used to serve as a base model to evaluate
other scenarios for operation in terms of safety to find the most reliable.

NIOSH Weight Lifting Equation

The NIOSH weight lifting equation is an important tool mostly used by occupational
health and safety experts to assess the manual material handling risks associated with
the workplace while performing lowering tasks and lifting.
It offers the recommendation regarding the weight limit (RWL); the recommended
weight lifting equation is as follows [11]:

RWL = LC(51) × HM × VM × DM × AM × FM × CM
802 A. B. Dasari and D. Balasubramanian

• H = Horizontal position of an object relative to the body.

• V = Vertical position of an object relative to the floor.
• D = Distance of the object moved vertically.
• A = Asymmetry angle/twisting requirement.
• F = Duration of lifting and frequency.
• C = Coupling/quality of worker grip on an object (M stands for multiplier).

As-Is Modeling Module

The as-is modeling module that includes main components such as the current condi-
tions of the existing workplace (building and objects) has to be modeled, and the path
that each worker’s 3D animation will use in the virtual representation is identified
through path planning.
The above table gives detailed information about the different types of operations
and the number of exertions took to unload the scrap totally from the trolley. This
exertion count for unloading each trolley made a path for analyzing the ergonomics
in the automobile industry, which includes the anthropometry data of the individual
and manual body angle measurement using the goniometer.
The manual measurements taken from goniometer are to be taken as the input to
the most useful ergonomic analysis software (3DSSP), which eventually gives the
resultant details such as analysis summary report and localized fatigue report.

Details of Weight and Number of Exertions for Unloading Different Operations

Figure 8: the graph gives details about the weight of the scrap lifted manually and
the number of exertions taken for the employee to unload the scrap present in the
trolley. The scrap produced from facing operation is less in weight compared to the
weights of other produced scraps, due to the low weight and the irregular size, the
holding capacity of this #05 operation scrap is considerably low.

4 Analysis Using 3DSSP Software for All the Scrap

Handling Operations

Table 4 gives the details regarding the anthropometry details of the worker performing
manual scrap handling taken prior to the starting of the 3DSSP analysis (Fig. 9).
The simulation of the images is done using the 3DSSP software from the input
parameters taken through goniometer.
Table 5 gives the details regarding the input parameters for the 3DSSP software
being taken manually using the goniometer, though it is a time-consuming process
but the most effective and accurate process of taking the body angles of the worker
during the working posture (Fig. 10).
The simulation of the images is done using the 3DSSP software from the input
parameters taken through goniometer (Table 6).
Evaluation of Ergonomics Issues in Repetitive Scrap Handling … 803

Fig. 8 Details of weight & number of Exertions for unloading different operations

Table 4 Anthropometry data of the employees working in manual scrap handling

S. No. Details Person 1 Person 2 Person 3
1 Gender Male Male Male
2 Height (cm) 165 163 167
3 Weight (kg) 54 52 54
4 Shoe height (cm) 2.5 2.5 2.5
5 Age 48 46 48

By using the proposed body angles measuring algorithm on the human body with
a goniometer, recognized scrap handling procedure which is induced from different
operations is obtained.

Inference from the Analysis Summary Report for the 3D Low Back Compression
Figure 11 gives details obtained from the analysis summary report which discusses
the 3D low back compression (Newton) with respect to the scrap produced from
different operations.
Figure 12 gives the details of the low back compression for the different steps
followed in handling (unloading) the scrap produced from #05 operation.

Inference of the Analysis Summary Report for the Percent of the Population
Capable
Figure 13 gives details in percentage about the capability of the population to execute
the particular operation; the analysis summary gives the report for the following
body parts such as wrist, elbow, shoulder, torso, hip, knee, and ankle. The inference
804 A. B. Dasari and D. Balasubramanian

Step-1 Step-2 Step-3 Step-4

Step-5 Step-6 Step-7

Fig. 9 The above figures are the steps an unloading person follows from pick to drop

that can be derived from the analysis summary report obtained from the 3DSSP
software is that while manually handling this type of scrap nearly 96.5% of people
can effectively and capably handle the wrist load, almost 100% of people can capably
handle the elbow load, on an average 98.5% population can capably handle the
shoulder load, around 93.75% of the population can manage the torso load, only
90.25% of population can capably manage the hip load, on an average 95.5% of
population can handle the knee loads, and 96.25% of population can capably handle
the ankle load while handling the scrap. As a result, the load on the hips while
handling (unloading) the scrap is comparatively more than other body parts loads.
Table 7 gives information about the percent of the population capable of bearing
the loads on the individual body part. The inference from the table is that only 90.25
and 93.75% of the population are capable of handling the load on the hip region and
the torso region, respectively.
The results of the localized fatigue analysis report from 3DSSP software are used
to give awareness about the work duration limits and also threshold duty cycle limits
of the particular unloading work.
Evaluation of Ergonomics Issues in Repetitive Scrap Handling … 805

Table 5 Input parameters for the steps involved in the manual scrap handling of the facing operation
scrap
Parameters Step-1 Step-2
Left Right Left Right
Horizontal Vertical Horizontal Vertical Horizontal Vertical Horizontal Vertical
Hand (°) 106 −55 106 −55 106 −53 106 −56
Fore arm 106 −55 106 −55 106 −53 106 −56
(°)
Lifting 0 0
weight (N)
Trunk Flexion Axial Lateral bending Flexion Axial Lateral bending
angles rotation rotation
Head 44 0 0 12 0 −1
Angles 45 0 0 12 0 0
Parameters Step-3 Step-4
Left Right Left Right
Horizontal Vertical Horizontal Vertical Horizontal Vertical Horizontal Vertical
Hand (°) 110 −53 110 −54 106 −52 110 −50
Fore arm 110 −53 110 −54 106 −52 110 −50
(°)
Lifting 0 27.5
weight (N)
Trunk Flexion Axial Lateral bending Flexion Axial Lateral bending
angles rotation rotation
Head 12 1 −1 6 −1 −1
Angles 12 0 0 7 0 0
Parameters Step-5 Step-6
Left Right Left Right
Horizontal Vertical Horizontal Vertical Horizontal Vertical Horizontal Vertical
Hand (°) 106 −52 110 −50 118 −66 120 −41
Fore arm 106 −52 110 −50 118 −66 120 −41
(°)
Lifting 27.5 27.5
weight (N)
Trunk Flexion Axial Lateral bending Flexion Axial Lateral bending
angles rotation rotation
Head 50 0 0 43 25 −4
Angles 33 0 0 26 0 0
Parameters Step-7
Left Right
Horizontal Vertical Horizontal Vertical
Hand (°) 43 −12 162 13
(continued)
806 A. B. Dasari and D. Balasubramanian

Table 5 (continued)
Parameters Step-7
Left Right
Horizontal Vertical Horizontal Vertical
Fore arm (°) 43 −12 162 13
Lifting weight (N) 0
Trunk angles Flexion Axial rotation Lateral bending
Head 60 40 −6
Angles 75 10 −4

#10 scrap Handling #20 Scrap Handling #50 Scrap Handling #60 Scrap Handling

Fig. 10 The figures are the unloading postures of the scrap produced from #10, #20, #50, #60
operations

4.1 NIOSH Weight Lifting Equation Analysis

Table 8 gives the total information regarding the maximum and minimum weights
that the scrap unloading worker handles for different scrap operations; minimum
average load and maximum average load of the handling material are used for the
calculation of the NIOSH weight lifting equation (Fig. 14).
From Table 9, it is clear that the Lifting Index (LI) of this scrap unloading work
is in the highest risk factor, it badly needs the corrective remedial action to reduce
the Lifting Index value for the safe lifting work, and it should be less than 1.0.

5 Results

5.1 3DSSP Analysis Output

Table 10 gives the details about the center of gravity of the particular position of the
scrap unloading work for each step; this is obtained based on the weight of the load
that the scrap handling person picks and the position of the torso; in the above figures,
Evaluation of Ergonomics Issues in Repetitive Scrap Handling … 807

Table 6 Input parameters for the steps involved in the manual scrap handling of the other operation
scrap
Parameters Turning Turning and ID Boring
Left Right Left Right
Horizontal Vertical Horizontal Vertical Horizontal Vertical Horizontal Vertical
Hand (°) 117 −54 115 −51 115 −51 109 −48
Fore arm 117 −54 115 −51 115 −51 109 −48
(°)
Lifting 46.1 46.1
weight (N)
Trunk Flexion Axial Lateral bending Flexion Axial Lateral bending
angles rotation rotation
Head 19 0 0 18 −1 1
Angles 18 0 0 16 0 0
Parameters Multi-angle cut Boring
Left Right Left Right
Horizontal Vertical Horizontal Vertical Horizontal Vertical Horizontal Vertical
Hand (°) 115 −51 109 −48 115 −51 109 −48
Fore arm 115 −51 109 −48 115 −51 109 −48
(°)
Lifting 52 54.9
weight (N)
Trunk Flexion Axial Lateral bending Flexion Axial Lateral bending
angles rotation rotation
Head 18 −1 0 18 −1 3
Angles 16 0 0 16 0 0

the red dot represents the center of gravity; if the red dot is out of the green region,
the particular position is not acceptable to work and that at any time the worker may
fall down or cause to himself heavy strain to the hip and the torso region.
Table 11 gives the details about the center of gravity for the other scrap handling
operations; this kind of analysis is made according to the ACGIH 2016 standards of
the ergonomic analysis.

The 3D Analysis Summary Report

Figures 15 and 16 gives the result about the body parts which are subjected to high
strain for the population of 25%, The entire analysis is done in the software with
25% strength population; among the 25% of the population, the hip and the torso
are the major parts that are affected badly and among that population 90.25% and
93.75%, respectively, are capable of handling the loads on the particular operation.

NIOSH Weight Lifting Equation

The NIOSH weight lifting equation confirms that (ref Table 9), the Lifting Index (LI)
value is out of the limit, i.e., exceeding 1 by 0.11, thereby indicating that it critically
needs engineering and administrative control for remedial action.
808 A. B. Dasari and D. Balasubramanian

Fig. 11 3D low back compressions for different scrap handling operations

Fig. 12 3D low back compressions for handling facing operation scrap

Evaluation of Ergonomics Issues in Repetitive Scrap Handling … 809

Fig. 13 Percentage of population capable to handle the loads

Table 7 Percent of
S. No. Load on body parts for Percent of population
population capable of
different operation capable (%)
handling the scrap with
respect to body parts 1 Wrist 96.5
2 Elbow 100
3 Shoulder 98.5
4 Torso 93.75
5 Hip 90.25
6 Knee 95.5
7 Ankle 96.25

Table 8 Calculation of the

S. No. Operation Min weight Max weight
NIOSH lifting equation
name lift/exertion lift/exertion
(kg) (kg)
1 Facing 2.0 2.8
2 Turning 3.0 4.7
3 Turning and ID 4.0 5.11
boring
4 Multi-angle 4.0 5.3
cutting
5 Boring 5.0 5.59
810 A. B. Dasari and D. Balasubramanian

Fig. 14 a Represents NIOSH weight lifting equation details, b represents the dimensions of the
trolley, c, d represent the real-time action with NIOSH weight lifting standards

6 Conclusions

This evaluation and analysis study on the ergonomics in manual scrap handling
system of a particular automobile industry has enabled to draw the following
conclusions

3DSSP Software Analysis

• Confirms that the load is heavy on the hip and torso of the scrap handling person.
• The unacceptable level of the center of gravity of the scrap handling person even-
tually leads to fall of person and strongly supports the heavy load on the hip and
torso region of the worker.

NIOSH Lifting Equation

• The NIOSH weight lifting equation indicates that the Lifting Index (LI) is above
the acceptable value of 1, i.e., (1.11). So the scrap handling operation critically
needs the engineering and administrative controls as corrective action.
Evaluation of Ergonomics Issues in Repetitive Scrap Handling … 811

Table 9 Calculation of NIOSH lifting equation

Lifting task Unloading Multiplier Unloading Multiplier
stock origin origin stock destination
destination
Horizontal location (25.34–63.5) 58.42 HM = 50.8 HM = 0.50
cm H 0.43
Vertical location (0–177.8) cm V 9.9 VM = 49.76 VM = 0.92
0.80
Travel distance (25.4–177.8) cm D 45.11 DM = 45.11 DM = 0.92
0.92
Angle assymetry (0–342.3) A 0 AM = 0 AM = 1.00
1.00
Coupling (1 = good, 2 = fair, 3 = 3 CM = 0.90 3 CM = 0.90
poor) C
Frequency (0.2–15 lifts/min) F 7.5 FM = 0.22 7.5 FM = 0.22
Average weight lifted in kg 3.6 3.6
Maximum weight lifted in kg 4.6 4.6
Duration (1, 2, 8 h) 8 8h 8 8h
Recommended Weight limit RWL 3.3 4.3
= HM * VM * DM * AM * CM *
FM
Lifting index (LI) = load/RWL 1.11 0.84

MOST Method

• The MOST method of analysis for the scrap lifting operation produced from all
the operations has the quick and shorter time duration of handling the scrap from
its origin to the destination.
• It also serves the same to the plant layout; the traveling distance of unloading
trolley to bring from origin to destination is the shortest and quickest as per the
time duration and the plant layout.
 812

Table 10 Details on the center of gravity for the steps involved in facing operation scrap unloading
S. No. Step posture Center of gravity %MVC (maximum voluntary %DC (maximum exertion
contraction) duty cycle)
1 Critical Are within the limits Neck Flexion and torso
flexion are out of limits

2 Are within the limits Wrist flexion, elbow flexion,

shoulder abduction, and torso
Unacceptable flexion are out of limits

3 Are within the limits Wrist ulnar/radius, elbow

flexion, shoulder abduction,
neck flexion, lateral bending,
and torso are out of limits
Critical

4 Are within the limits Wrist flexion, Ulnar/radius,

elbow flexion, shoulder
rotation backward/forward,
abduction, neck lateral
bending, Torso flexion and
rotation are out of limits
Acceptable
A. B. Dasari and D. Balasubramanian
Table 11 Details on the center of gravity for the steps involved in boring operation scrap unloading
S. No. Step posture Center of gravity %MVC %DC
1 Are within the limits Wrist flexion and ulnar/radius, elbow flexion, shoulder
rotation backward, frontward, neck flexion, torso flexion
are out of limits

Acceptable
2 Are within the limits Wrist flexion and Ulnar/radius, elbow flexion, shoulder
rotation backward, frontward, neck flexion, torso flexion
are out of limits

Acceptable
3 Are within the limits Wrist flexion and ulnar/radius, elbow flexion, shoulder
rotation backward, frontward, neck flexion, torso flexion
are out of limits
Evaluation of Ergonomics Issues in Repetitive Scrap Handling …

Acceptable

(continued)
813
Table 11 (continued)
814

S. No. Step posture Center of gravity %MVC %DC

4 Are within the limits Wrist flexion and ulnar/radius, elbow flexion, shoulder
rotation backward, frontward, neck flexion, torso flexion
are out of limits

Acceptable
A. B. Dasari and D. Balasubramanian
Evaluation of Ergonomics Issues in Repetitive Scrap Handling … 815

Fig. 15 Average percent of the population capable

Fig. 16 Indication for the percentage of population considered for the analysis

References

1. Golabchi A et al (2018) An integrated ergonomics framework for evaluation and design of

construction operations. Autom Constr 95:72–85
2. Munck-Ulfsfält U (2003) Corporate ergonomics programme at Volvo car corporation. Appl
Ergon 34:17–22
816 A. B. Dasari and D. Balasubramanian

3. Dul J, Neumann WP (2009) Ergonomics contributions to company strategies. Appl Ergon

40(4):745–752
4. Leung MY, Chan IYS, Yu J (2012) Preventing construction worker injury incidents through
the management of personal stress and organizational stressors. Accid Anal Prev 48:156–166
5. Golabchi A, Han S, Seo J, Han S, Lee S, Al-Hussein M (2015) An automated biomechanical
simulation approach to ergonomic job analysis for workplace design. J Constr Eng Manage
141(8)
6. Golabchi A (2015) Integration of ergonomic analysis into simulation modelling of manual
operations. In: Simulation in production and logistics, pp 491–500
7. Tran D, Sorokin A (2008) Human activity recognition with metric learning. ECCV, Part I,
LNCS 5302:548–561
8. Wei X (2011) Physically valid statistical models for human motion generation. ACM Trans
Graph 30(3): Article 19
9. Case K, Marshall R, Summerskill S (2016) Digital human modeling over four decades. Int J
Digit Hum 1(2):112–131
10. Senthil J, Haripriya G (2016) Time analysis with most technique. Int J Chem Sci 14(S2):519–
526
11. https://ergo-plus.com/niosh-lifting-equation-single-task/
Reduction of Terminal Rejections
and Failure Cost through Kaizen at Shop
Floor of a Wire Harness Manufacturing
Company

Pramod Kumar and Jaiprakash Bhamu

Abstract In the cutthroat market, demand for higher-quality products is currently

the most important task in front of organization and management looks for better
effectiveness. In manufacturing units, high cost, declining profit boundaries, dis-
crepancy of quality of products, and delivery of such products in the competitive
market are the factors those have encouraged many organizations to focus on ways
to increase the involvement of employees to improve quality and productivity. Kaizen
is one of the employee contribution methods intended to bring together all level of
workforce in an organization for setting principles of excellence and achieving better
consequences. In this paper, it is tried to analyze the performance of Kaizen in a wire
harness manufacturing unit which is considered for the study, and the paper presents
a case study of the selected company to reduce rejections due to high contact gap in
terminal that further confides decrement in failure cost.

Keywords Kaizen, root cause analysis · Cost-effectiveness · Process

improvement · Productivity · Rejection level

Abbreviations

OEE Original equipment efficiency

PDCA Plan-do-check-act
AWP Advance wire process
NG Not good
BOP Begin of process
DOL Direct on line
WIP Wire in process

P. Kumar (B) · J. Bhamu

Government Engineering College, Bikaner, India
e-mail: pramodfaujdar1984@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 817

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_71
818 P. Kumar and J. Bhamu

1 Introduction

In this study, effective lean tool such as Kaizen has been adopted for process improve-
ment in a wire harness manufacturing company and Kaizen is used for betterment
of cost-effectiveness at the shop floor in the company through teamwork. The study
deals with various aspects of Kaizen as a lean manufacturing tool and how improve-
ments can be made by adopting practices of Kaizen in wire harness manufacturing
companies. OEE method was applied to reduce the cost by using Kaizen technique to
a semiconductor industry which results in cost reduction of 7% annually for the use of
gases and chemicals [1]. Kaizen was adopted to reduce operational cost and improve
efficiency [2]. Kaizen-based case study was performed in Romania for improvement
through PDCA cycle which results in reduction in service interventions, labor costs,
and cost of replaced parts [3]. Kaizen signifies small improvements that have been
made in the quality as a result of continuing efforts. Kaizen method concerned to
work out the part mismatch problems in automobile assembly production line [4].

2 Methodology of the Study

Step-by-step Kaizen procedure has been attempted to solve the problem by data col-
lection, root cause analysis, selection of the best solution method, corrective actions,
and documentation. Identifying the problems, analyzing them, signifying solutions,
implementing them, and getting feedback are all sequential procedures that help to
resolve work-related problems properly and efficiently. The sequential method of
Kaizen is followed to solve work-related problems. The study is carried out over a
period of three months, during which team of Kaizen worked together to identify
the major problem and found a suitable solution at the shop floor of the company.
This study was performed on the basis of different Kaizen tools and problem-solving
techniques for collecting and analyzing the data. The sequential procedure that we
followed is shown by a flowchart in Fig. 1.

3 Case Study

3.1 Company Background

The ABC Company is a professionally managed company known for its quality and
attitude toward perfection in the field of auto-electrical and electronic components.
Appreciating the continually evolving needs of the automotive industry, the company
has constantly developed new products for its customers and has endeavored to
remain at par with the leading technology available worldwide. The selected com-
pany is among the largest manufactures of wiring harnesses in India. A well-equipped
Reduction of Terminal Rejections and Failure Cost through … 819

Identification of Problem

1. ABC Analysis
Selection of Problem 2. Ranking Method
3. Pareto Analysis

Theme selection

Define Problem

Regular imple-
Data Collection
mentation status

Fix the Target

Follow-up Review
Cause & Effect
Diagram, Why-Why
Root Cause Analysis Analysis, Pareto
Achive-
Analysis
ments/Tangible/Intangible
Benifits
Developing Solution

Counter Measures Counter Measures & Actions

Fig. 1 Flowchart for problem solving case study

design and manufacturing facility combined with strong engineering capability

ensures production of world-class products.

3.2 Problem Recognition

Team members identified problems by visiting the shop floor of production depart-
ment of the company, and after discussion with shop floor workers, 52 problems
(shown in Table 1) have been recognized in crimping section of WIP area of the
company.
820 P. Kumar and J. Bhamu

Table 1 Problem recognition

S. No. Problem
1 Improper overlapping in taping
2 Improper storage system on shop floor
3 Tape wastage
4 Coupler interchange at coupler assembly stage
5 Improper storage of tape
6 Plating problem during storage
7 Improper issuing system of processed wires
8 Short molding in coupler
9 Cutting variation in terminal during crimping
10 Mixing in coupler
11 Improper fitment gap during crimping
12 Mixing of processed wires
13 Improper identification of job card
14 Improper storage of processed wires
15 Improper fitment of fuse case
16 Cutting variation in terminals during crimping
17 Terminal damaged during handling
18 Lock broken in 2p male coupler
19 Band harness missing
20 Improper pulling of band harness
21 Improper branching of taping
22 Time loss in opening of tape
23 Missing of protector cord
24 Wire interchange at coupler assembly stage
25 Terminal damage at wire assembly stage
26 Excess tool setup time at coupler insertion fixture
27 Excess tool setup time in continuity testing stage
28 Testing pins wear out
29 Terminal damaged at coupler insertion stage
30 Excess changeover time of taping fixture
31 Roller bearing of conveyor broken
32 Improper height of hangers
33 Improper identification of wire spool at AWP
34 Improper record storage of used job cards
35 Improper storage of raw material at coupler insertion
stage
36 Improper storage of fixtures
(continued)
Reduction of Terminal Rejections and Failure Cost through … 821

Table 1 (continued)
S. No. Problem
37 Improper display of work instruction at shop floor
38 Improper storage of master sample
39 Time loss in changeover of terminal
40 Terminal damage during taping
41 Break down of 2.8 m terminal
42 Time loss in opening the tape
43 Tape wastage during taping
44 Tongue revert back
45 Contact gap in terminal
46 Improper knurling in terminal
47 Terminal face damaged
48 Leg height variation in terminal
49 Improper identification system of sleeves
50 Terminal blackout
51 Excess machine setup time of continuity tester
52 Conveyor hanger broken

3.3 Problem Selection

It is required to select the problem which affects the most and due to which more
rejections are occurring. So to identify this problem in this study, three basic methods
are followed, i.e., ABC analysis, Pareto analysis and rating method.

ABC Analysis
After ABC analysis, it is observed that out of 52 problems, 43(shown serial no. 1–43
in Table 1) problems are in category of A, which have minimum involvement of other
department and only 5 problems (serial no. 44–48 in Table 1) are under B category in
which involvement of other department is necessary and 4 problems (serial no. 49–
52 in Table 1) are signified in C category which is needed management support for
implementing solutions. The problems which have minimum involvement of other
department are considered in ‘A’ category and which have involvement of other
department is necessary are in ‘B’ category. In ‘C’ category problems management
support may be needed in implementing solutions.

Pareto Analysis
Pareto analysis is basically a statistical technique used for selection of the problem
which has significant overall affect on operations. Pareto analysis not only shows
you the most important problem to solve, also gives you a score showing how severe
the problem is. Out of 52 problems, it is decided to select a major problem that
is responsible for increasing rejection level of terminal. After discussion and deep
822 P. Kumar and J. Bhamu

inspection, the team members decided to take one problem out of 5 problems from B
category problems. With the help all members’ opinion, all these five problems are
assigned score of severity of rejections. After plotting the score on chart, it is found
that highest rejections terminal are occurring due to improper gap between leg of
terminal so this problem is first selected to solve. Pareto chart is shown in Fig. 2.

Rating Method
Rating method is also an important technique which is also utilized for selection of
the major problem for responsible for most of rejections of terminal. In this method,
rating from 1 to 5 is decided based on three factors, i.e., occurrence, detection and
extent of impact. 1 rating is for LOW, and 5 is for HIGH. Problem selected on the

100.00
88.61
74.68

100 100
55.69
Mark Obtained

80 80
35.44

60 60

Cum %
40 28 40
16 15
20 11 9 20
0 0
CONTACT TERM. IMPROPER LEG TONGUE
GAP FACE KNURLING HEIGHT REVERT
DAMAGE VARIATION BACK

PROBLEMS

Fig. 2 Pareto chart analysis

Photo 1 Contact gap NG

Reduction of Terminal Rejections and Failure Cost through … 823

basis of highest priority numbers is calculated by multiplying the rating of given

three factors.
For the occurrence factor, 5 rating is given for most frequently occurring problem
and 1 for the problem which have lowest frequency of occurring. In detection, 1
rating is allotted for the problem that can be easily detected at the stage of generation
time and 5 for the problem which cannot be detected. In extent of impact, 5 rating is
allotted for the problem having impact on safety and 1 rating given to the problem
which is having no significant impact (Table 2).

Table 2 Severity score of impact

S. No. Problem Severity
1 Contact gap in terminal 28
2 Terminal face damage 16
3 Improper knurling in terminal 15
4 Excessive tool setup time in crimping section 11
5 Tongue revert back 9

Table 3 Rating for three

Rating Occurrence
factors
5 Every hour
4 Every day
3 Every week
2 Every month
1 6 or more month
Rating Detection
1 Can be detected at the stage of generation (fool
proofing)
2 Can be detected at the next stage using gauge/jig
3 Can be detected by sample inspection
4 Can be detected by 100% inspection
5 Can’t be detected
Rating Extent of impact
5 Impact on safety
4 Functionality is affected
3 Sec. functionality is affected
2 Inconvenience to customer/user
1 Insignificant impact
824 P. Kumar and J. Bhamu

4 Defining the Problem

The problem that creates major rejections is found the improper contact gap or height
in 1.6F female type of terminal designed to slip inside a socket contact. The gap should
be as per standards decided by the company, but at most of working station, this was
not found as per the standards. As per the quality standards, this gap should not be
greater than 2 mm (shown in Photo 2), but after measuring this gap, it was found
5–6 mm or up to 7 mm (shown in Photo 1). Due to this improper gap, the rejections
levels were very high. As per data collected for 6 days in month of April 2018, the
rejection was 66% because of this improper contact gap.

4.1 Impact of the Problem

The problem that was occurring before the improvement affected so many things in
the company, i.e., increase in-house rejections, decrease in vendor rating, decrease
in customer satisfaction, loss in productivity, increase in rework cost, and operator
fatigue.

Photo 2 Contact gap ok

Reduction of Terminal Rejections and Failure Cost through … 825

Table 4 Priority score calculation

S. No. Problem Occurrence Detection Extent of Priority No.
impact
1 Low, 5 high 5 Low, 1 high 1 Low, 5 high O*D*E
O D E
1 Tongue revert 2 4 4 32
back
2 Contact gap in 3 4 4 48
terminal
3 Improper 1 2 3 6
knurling in
terminal
4 Terminal face 2 4 2 16
damage
5 Leg height 2 5 1 10
variation in
terminal

4.2 Objectives

The objectives of our team are to ensure 100% customer satisfaction, increase vendor
rating, reduce in-house rejections, increase productivity, increase confidence level of
operator, and reduce the cost of failure. As per (Table 4) rating method, the priority
number of ‘contact gap in terminal’ is high so it is decided to solve this problem on
priority basis.

5 Process Flow

The flow process of wire harness manufacturing section of the company is shown
in Fig. 3. The raw material is received at raw material store and then goes through
BOP; after that, the material is transferred to main store of the company. After issuing
from main store, it reaches to wire assembly section and then transfers to coupler
assembly. In coupler assembly, coupler is fixed to wires and this semi-finished product
is transferred for board tapping process, and after that, this is electrically tested and
cover connector is inserted. After complete assembly, the harness is sent to final
inspection stage, and after inspection, it is ready for dispatch (Table 3).
826 P. Kumar and J. Bhamu

Problem
2 Area
1. Receiving store
2. BOP
Problem 3. Main Store
3 10 Area 4. Wire Assembly
5. Coupler Assem-
bly
Problem 6. Board Tapping
9 Area
4 7. Electrical Testing
8. Cover connector
insert
5 8 9. Final Inspection
10. DOL
Problem
6 7 Area

Fig. 3 Process flow diagram for indicating the problem

Leg Height
Variation Tongue revert
6% back
Improper 6%
Knurling 8%

Contact Gap
66%
Terminal face
damage
14%

Fig. 4 Data representation on pie chart (refer Table 5)

6 Data Collection

For analysis of the impact of the five B category problems, data are collected for
non-conforming pieces of terminal. After the analysis of six-day observations, it was
found that 10 pieces are non-conforming due to leg height variation, 13 are due to
improper knurling, 22 are due to face damaged, 9 are due to tongue revert back, and
highest non-conforming pieces are due to improper contact gap of the terminal. This
analysis is also shown in pie chart; according to that, 66% pieces of terminal are
non-conforming due to contact gap problem. Total rejection level at different stages
of production process is shown in Fig. 5. Out of total received quantity 1,120,000,
Reduction of Terminal Rejections and Failure Cost through … 827

1120000

. Rej.Qty

72500
8000 2000 150

Total BOI Stage Process Stage DOL Stage Customer

Recieved Qty End
Different Stages

Fig. 5 Rejections faced at different stages during April 2018

90000 82650
80000
70000
60000
50000
Rejection from
40000 35000
Dec17 to April18
30000
20000
20000
10000 1500
2000
0
Dec Jan Feb March April

Fig. 6 Month-wise rejection from December 2017 to April 2018

72,500 pieces are rejected at BOI, 8000 pieces are rejected during process stage,
2000 at DOL, and 150 are rejected at customer end. Five month rejections status is
shown in Fig. 6. It is observed that the rejection is continuously increasing due to the
present problem as in the month of April 2018 this is found 82,650 pieces (Fig. 4).

7 Root Cause Finding

Why-Why analysis method is used to move past symptoms and understand the true
root cause of a problem. It is said that only by asking ‘Why?’ Successively, it can
investigate into a problem deeply enough to understand the ultimate root cause. With
the help of this technique, it is observed that first tool setup rolled is due to insufficient
828 P. Kumar and J. Bhamu

training given to operator, latch/lever stopper wear problem is due to not analyzing
tool life properly, the problem of improper setting of lever stroke length occurs due
to inadequate tool design, and gap induced during rework of wire interchange is the
problem which is due to inadequate rework method (Tables 4 and 5).

Table 5 Data collection for problem analysis

Category of Data collection of 2nd week of March 2018 Total NC pcs Priority
problem Day-1 Day-2 Day-3 Day-4 Day-5 Day-6
Leg height 1 2 2 1 3 1 10 4
variation
Contact gap 15 16 12 27 19 24 103 1
of term
Improper 2 1 3 2 2 3 13 3
knurling
Terminal 3 2 3 4 5 5 22 2
face
damaged
Tongue 1 2 1 2 1 2 9 5
revert back

Table 6 Why-Why Analysis for root cause finding

Problems WHY1 WHY2 WHY3 WHY4
First tool setup First setup Operator is not Operator is not Training is not
piece rolled non-conforming aware of the aware of the provided to
pieces not disposition of impact of the operator about
separated by first setup pieces first setup pieces the impact of
operator during job setup first setup pieces
pieces
Latch/lever Stopper wear Improper Preventive Actual tool life
stopper wear out out before its preventive maintenance is not analyzed
life maintenance schedule is not properly
adequate
Improper setting Die setup length Stopper screw is Screw locking is Tool design is
of lever stroke is more than loose during not adequate for not adequate
length required production high-speed press
Gap induced Contact gap Pin use for There is no other Rework method
during rework of increased during terminal option to release is not adequate
wire interchange terminal releasing from the terminal
releasing coupler
Reduction of Terminal Rejections and Failure Cost through … 829

Table 7 Data analysis

S. No. Causes Observations No. of failures
Nos. In (%)
1 First tool setup piece rolled 15 times 9 53
2 Latch/lever stopper wear out 15 times 5 29
3 Improper setting of lever stroke length 15 times 2 12
4 Gap induced during rework of wire interchange 15 times 1 6

8 Data Analysis

After the root analysis of causes of failure, it is found that major causes are there due
to which rejections level is high that are shown in Table 7. Observations are taken 15
times, and out of that, 9 times failure occurred due to tool setup problem and 5 times
due to wear out lever stopper, only 2 failure are because of improper setting of lever
stroke length and 1 time due to gap induced during rework of wire interchange. So
it is observed from data that 53% failures are due to tool setup problem (Table 6).

9 Developing Solution

Theme that got maximum percentage in ABC analysis, rating method, and Pareto
analysis is directly related to company’s esteemed customer and fit/function at cus-
tomer end, internal customer harassment, overall quality rating, time loss in pro-
ductivity and rejection increase/rejection cost increase, so it is must to eliminate
the root causes of this problem that are identified during root cause finding stage.
Training is provided to the operator about the impact of first set pieces, and customer
complaints were displayed at the notice board. Fresh pieces were used while rework-
ing the wire interchange, and training is provided to operator for change of rework
method that concludes to adequate method of reworking. Tool life is analyzed again
and displayed, and fixed punch is used at the place of lever/latch. To implement the
improvements, following are the resistances that are responsible for further step to
overcome the causes of failures.
Before the implementation of the counter measures, the rejection level found
82,650 pieces at the end of April 2018, it is observed that after implementation of
countermeasure 1, rejection level reduced to 55,600 pieces, after countermeasure 2,
it reduced to 45,600 pieces, after countermeasure 3, it is found 40,000 pieces and
after implementation of countermeasure 4, the rejections level found at zero level.
This is great achievement of our objectives (Table 8).
830 P. Kumar and J. Bhamu

Table 8 Countermeasures adopted for elimination of root causes

S. No. Root causes Countermeasures
1 Training is not provided to operator about Training is provided to operator about the
the impact of first setup pieces impact of first setup pieces
Customer complaint is displayed at shop
floor
2 Rework method is not adequate Fresh pieces are used while reworking
the wire interchange
Training is provided to operator for
change of rework method
3 Actual tool life is not analyzed properly Tool life is analyzed again and displayed
4 Tool design is not adequate Fixed punch is used in place of lever/latch

10 Calculation for Cost-Effectiveness

Monthly rejection qty. of terminal = 10,150 pcs.

Cost of reject w/p term = 10,150 * 0.90 paisa = Rs. 9135.
Internal failure cost
Operator cost for processing = 2 min.* 8000 = 267 h.*20 = Rs. 5340.
Operator cost (DOL) = 2000 * 2 min. = 67 h * 20 = Rs. 1340.
Total = Rs. 6680.
External failure cost
Customer visit cost (5-times) = 200 km * 5 = 1000 = Rs. 1000.
Operator cost for rework at customer end = 6 days * 160 = Rs. 960
Staff cost = 16 h * 72 = Rs. 1152
Total cost save = 9135 + 5340 + 1340 + 1000 + 960 + 1152 = Rs. 18,927/month
Total cost saving per year = Rs. 227,124 (Table 9).

Table 9 Probable resistance

S. No. Problem resistance Step to overcome
1 Supplier may not accept this as they Counseling the supplier and to convince
might think that this will be an extra them that this will help them for a smooth
burden to them and enjoyable work
2 Operator will not agree to implement the Counseling the operator about the impact
change method of rework of the problem and convince them that
this will help to avoid the occurrence of
the problem at customer end
3 Shutdown of the machine is required for A system is developed to overcome the
tool modification, likely to lead problem by planning the tool
production loss modification on turn basis
Reduction of Terminal Rejections and Failure Cost through … 831

82650

90000 55600
80000
70000 45600
40000
60000
50000
40000
30000 0
20000
10000
0
Before CM's After CM1 After CM2 After CM3 AFTER
CM4

Fig. 7 Effectiveness of countermeasures

11 Conclusion

The theme of this was to reduce material cost and failure cost by solving the contact
gap problem of the terminal in WIP area by adopting Kaizen lean tool. In this regard,
total 52 problems are identified, out of those 43 in A, 5 in B, and 4 in C category.
This problem affects that total 82,650 pieces were rejected maximum in month of
April 2018. In this case, to remove the contact gap, a modified mechanism with fixed
punch installed in place of existing mechanism. After improvement in mechanism,
the rejection level reduced to zero and the total cost saved Rs. 227,124/year. At the
closing stages of this work, it concludes that Kaizen technique is very effective to
solve the work-related problems in the organizations. This way of improving the
production processes and getting cost-effectiveness is very precious for building far
above the ground productivity of the company (Fig. 7).

References

1. Kikuchi K, Kikuchi T, Takai T (2007) Method of overall consumable effectiveness. In: IEEE
international symposium on semiconductor manufacturing, Santa Clara, USA, pp 1–4
2. Modarress B et al (2005) Kaizen costing for lean manufacturing: a case study. Int J Prod Res
43(9):1751–60
832 P. Kumar and J. Bhamu

3. Veres C, Marian L, Moica S (2017) Case study concerning effects of Japanese Management
Model. Procedia Eng 181:1013–1020
4. Chandrasekaran M, Kannan S, Pandiaraj P (2008) Quality improvement in automobile assembly
production line by using kaizen. Manuf Technol Today 7(3):33–38
Solving the Flexible Job Shop Scheduling
Problem Using a Hybrid Artificial Bee
Colony Algorithm

Rylan H Caldeira , A. Gnanavelbabu , and J. Joseph Solomon

Abstract In this work, a hybrid artificial bee colony algorithm is proposed for solv-
ing the flexible job shop scheduling problem (FJSP) which is a classification of
the classical job shop scheduling problem (JSP) considered to NP-hard in nature.
In FJSP, an operation can be processed on a set of capable machines with different
processing times, thereby dealing with a routing and sequencing problem. The objec-
tive considered is to minimize the makespan. The basic artificial bee colony (ABC)
algorithm stresses on the balance between global exploration and local exploitation.
However, the drawback of the basic ABC algorithm is that it converges prematurely
and may get trapped in the local optima. Hence to improve its exploration capability
in local space, it is hybridized using a Tabu search (TS) algorithm. At first, initial
solutions are generated with certain quality and diversity as food sources using mul-
tiple strategies in combination. Crossover and mutation operations are carried out for
machine assignment and operation sequencing separately generating new neighbor-
ing solutions. Lastly, a local search strategy based on TS is proposed to enhance the
local search capability. Kacem’s and Brandimarte’s benchmark instances are used to
compare the performance of the proposed approach to five other well-known algo-
rithms in the literature. Experimental results revealed the superiority of the proposed
approach in solving FJSP.

Keywords Flexible job shop scheduling · Artificial bee colony algorithm · Tabu
search · Makespan

R. H. Caldeira · A. Gnanavelbabu (B) · J. Joseph Solomon

Department of Industrial Engineering, CEG Campus, Anna University, Chennai,
Tamil Nadu 600025, India
e-mail: agbabu@annauniv.edu

© Springer Nature Singapore Pte Ltd. 2021 833

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_72
834 R. Caldeira et al.

1 Introduction

The process of converting the resources into a finished or semi-finished product by

adding value to it is called manufacturing. There are various types of manufacturing
systems based on their size and activity. They are job shop, batch production, mass
production, flow production, and assembly lines. This work deals with job shop since
it is an extensively adopted type of production system. A job shop is a type of man-
ufacturing process in which small batches of a variety of customized products are
produced. JSP is the process of scheduling these jobs to machines. FJSP which is an
extension of the JSP is an NP-hard combinatorial optimization problem. These prob-
lems are complex as each operation can be processed on more than one machine with
different processing times. It consists of two sub-problems, routing, and scheduling.
FJSP is usually solved by adopting a hierarchical or integrated approach. Brandi-
marte [1] proposed a hierarchical strategy for solving the scheduling problem by
using TS. In recent times, hybrid meta-heuristics have become a popular tool for
solving the FJSP as the hybridization of two or more meta-heuristics have proven
to improve the performance of each meta-heuristic while eliminating the drawbacks
of the other. ABC is one of the popular meta-heuristics. ABC provides a structured
search framework that is characterized by individual improvement, cooperation, and
competition of the population, and thus, it can perform global exploration efficiently
(large-scale search). But the original ABC algorithm falls back in its ability to per-
form exploitation effectively by converging to a solution in earlier iterations, thus
being incapable of exploring the search space effectively. In order to improve the
local search capability of the ABC algorithm, a local search technique based on TS
is introduced which greatly enhances the exploitation capability of the algorithm.
A beer froth ABC was proposed by Sharma et al. [2] in which to maintain proper
harmony amid exploration and exploitation capabilities a beer froth phenomenon
inspired position update was incorporated. A novel ABC based on the cosine simi-
larity is proposed by Xiang et al. [3], a method based on cosine similarity for choosing
a better neighbor individual to overcome the insufficiency of slow convergence in
ABC. Levy flight ABC proposed by Sharma et al. [4] illustrated that there is a higher
probability of skipping the optimal solution owing to large step sizes and a search
strategy based on Levy flight was integrated with ABC which balances the exploita-
tion and exploration capability. A hybrid ABC for FJSP proposed by Thammano
and Phu-ang [5] where dispatching rules and the harmony search algorithm were
used to create the initial solutions after which one of the two search techniques was
randomly selected with a probability that is proportional to their fitness values. Sim-
ulated annealing algorithm or filter and fan algorithm was employed to escape from
the local optimum. A hybrid ABC for the JSP proposed by Zhang et al. [6] aimed at
minimizing the total weighted tardiness in JSP. Considering the high complexity, a
novel ABC algorithm was proposed for solving the problem. A neighborhood prop-
erty of the problem was discovered, and then, a tree search algorithm was devised
to enhance the exploitation capability of ABC. An effective ABC for the FJSP pro-
posed by Wang et al. [7] stresses the balance between global exploration and local
Solving the Flexible Job Shop Scheduling Problem Using … 835

exploitation. A local search strategy based on critical path theory was proposed to
enhance the local intensification capability of the onlooker bees.
A hybrid ABC for solving the FJSP proposed by Li et al. [8] considered the total
flow time criterion, where TS heuristic is introduced to perform a local search for
employee, onlooker, and scout bee phases. A best-so-far ABC proposed by Banharn-
sakun et al. [9] biases the solution direction toward the best-so-far solution rather
than a neighboring solution as proposed in original ABC method and also uses the
set theory to describe a mapping of their proposed method to the problem in the
combinatorial optimization domain. Among different approaches, ABC algorithm
proposed by Karaboga and Gorkemli [10] is a widely employed swarm intelligence
algorithm for scheduling FJSP problems. ABC algorithm was first proposed to solve
the multi-variable and multi-modal continuous functions. Localization and multi-
objective evolutionary optimization for FJSP proposed by Kacem et al. [11] present
two new approaches to jointly solve routing and scheduling by localization and then
by evolutionary approach. Mehrabad and Fattahi [12] proposed a TS algorithm to
solve the FJSP to minimize makespan. Hurink et al. [13] performed TS for solving
the FJSP.
In this work, we propose a hybrid ABC with TS to improve the exploration
capability in local space (exploitation). The proposed algorithm consists of the basic
ABC phases, initialization, employee bee phase, and the scout bee phase. Here, solu-
tions are initialized as two separate chromosomes to facilitate individual crossover.
The two chromosomes are machine assignment and operation sequencing. Machine
assignment follows the random rule, maximum time remaining rule, and most num-
ber of remaining operations rule, whereas operation sequencing follows the random
rule, global minimum processing time rule, and work-time considered rule. These
rules are adopted so that the population generated is diverse and feasible.
Employee bee phase initiates with the modified precedence operation crossover
operator (MPOX) for crossover in operation sequencing. A uniform crossover is
used for machine assignment. For half of the population chosen at random maximum
workload reduction, mutation operator is used for modifying machine assignment.
The mutation operator is used to enhance the exploration capability of the algorithm.
The solution obtained is now introduced into a TS local search. This search performs
an iterative search on the entire population to get new and better solutions in each
step. TS is chosen as it prevents the local search from converging on a solution
prematurely, thereby improving the local exploration capability of the algorithm.
When the solution does not improve, a random solution is generated on behalf of the
scout bee phase which is compared with the TS solution. A new better population is
obtained from both of these solutions and taken for the next ABC iteration until the
termination criteria are satisfied.
836 R. Caldeira et al.

2 Problem Definition

The formulation of FJSSP is as follows: There is a set of ‘m’ machines c = 1, 2, …,

m and a set of ‘n’ jobs a = 1, 2, …, n. Each job ‘a’ consists of a sequence of ‘k’
operations. Processing of each operation Oab (b = 1, 2, …, k) of job ‘a’ has to be
done on single machine ‘M c ’ out of a set of given available machines M ab (for M c e
M ab , M ab e M). Each operation needs a single machine picked from a set of capable
machines. Further, FJSP requires to set starting and ending times of each operation.
The FJSP needs to find both allotment and an operations sequence on the machines
so that given criteria are satisfied. If there is at least one operation M ab e M, it is
partial flexibility FJSP (P-FJSP), whereas if M ab = M for every operation, it is total
flexibility FJSP (T-FJSP). The objective is to minimize the makespan.

3 Framework of the Proposed Approach

3.1 Artificial Bee Colony Algorithm

The ABC algorithm is a meta-heuristic designed by observing the foraging behavior

of honey bee, introduced by Karaboga [1] in 2005. Here, all possible solutions to
the optimization problem are considered as food sources, and the honey bees search
to find the best solution. There are three types of bees in the population which are
involved in finding the food source. They are
• Employer bees. For all employer bees, generate new candidate solutions and
retain the candidate solutions possessing better objective function value.
• Onlooker bee phase. For all onlooker bees, select an employed bee randomly.
Produce a new candidate solution. Compute fitness of individual, and if the fitness
of the new solution is better than the existing solution, replace the older solution.
• Scout bees. If any source is exhausted, replace it by randomly generated solution
by scout and memorize the best solution until stopping criteria are met. Old is
substituted by an arbitrarily picked source within a specified region.

3.2 Solution Representation

The problem is solved hierarchically, and thus, solutions are represented as separate
vectors each for operation sequencing and machine assignment which is termed as
the two-vector solution representation (see Fig. 1). To guarantee an initial population
with a certain quality, different heuristics are adopted to generate a population.
Solving the Flexible Job Shop Scheduling Problem Using … 837

Fig. 1 Two-vector solution representation

For machine assignment, the following heuristics are employed:

• Global minimum processing time rule (20%);
• Workload considered rule (20%);
• Random rule (60%).
For operation sequencing, the following heuristics are employed:
• Random rule (80%);
• Most time remaining rule (10%);
• Most no. of operations remaining rule (10%).

3.3 Crossover and Mutation

In this phase, the two vectors are modified separately by crossover and mutation
operators. Machine assignment vector undergoes a uniform crossover (see Fig. 3),
where the new vector unew is derived from the initial vector ui and the vector with
the best solution u’ by generating a binary vector and then a maximum workload
reduction mutation operator is used to reduce the assignment of the machine with
maximum workload.
Operation sequencing vector undergoes a modified precedence operation
crossover (MPOX) operation (see Fig. 2), where a random subset of half the jobs is
generated, and they are given priority from one vector, whereas the other jobs are

Fig. 2 Modified precedence

operation crossover
838 R. Caldeira et al.

Fig. 3 Uniform crossover

assigned from another vector from left to right. Then, a maximum workload reduc-
tion operator is applied which identifies the machine with the maximum workload
and re-assigns possible operations to other machines with lesser workloads. This acts
as the mutation operator which is performed on only half of the population (Fig. 3).

3.4 Tabu Search

The TS is a local search heuristic that uses a neighborhood search procedure to

iteratively move from one possible solution to x to another improved solution x  in
the neighborhood of x until a terminating criterion is fulfilled. Local search techniques
usually get stuck in less efficient areas and do not explore the search space completely.
TS carefully explores the neighborhood of each solution as the search progresses.
So to prevent the heuristic from returning to a recently visited solution, a short-term
memory of recent moves within the search space is recorded, and this prevents future
moves from repeating these recent moves. That is if a new potential solution has been
previously visited within a short time period, it is marked as Tabu (forbidden) and
the algorithm does not consider that solution. Thus, the solutions admitted to the new
neighborhood, N * (x), are determined through the use of memory structures. Using
these memory structures, the search progresses iteratively moves from the current
solution x to an improved solution x  in N * (x).
These memory structures form a list known as the Tabu list which will contain a
set of rules and banned solutions used to filter solutions that will be admitted to the
neighborhood N * (x) to be explored by the search. So a Tabu list is a short-term set
of the solutions that have been visited in the recent past (less than n iterations ago,
where n is the number of previous solutions to be stored which is also called the
Tabu tenure). So a Tabu list consists of solutions that have changed by the process
of moving from one solution to another.
Solving the Flexible Job Shop Scheduling Problem Using … 839

3.5 Proposed Hybrid ABC

The parameters for the algorithm like population size, number of iterations, and bee
colony size are set. The initial population solution representation is done by using
two-vector notations employing population initialization rules.
In the employee bee phase, the machine assignment vector undergoes uniform
crossover and a maximum workload reduction mutation operation where the muta-
tion is performed on only half of the population, whereas the operation sequencing
vector undergoes a modified operation crossover operation. Then, for the whole mod-
ified population, a TS local search is performed which provides the best solution in
every iteration without converging predominantly, thus improving the exploration
capability of the ABC algorithm in local search space. The flowchart of the proposed
hybrid ABC is illustrated (see Fig. 4).

4 Experimental Results

The proposed approach was coded in MATLAB software on an i3 processor with a

speed of 1.8 GHz on a Windows 10 operating system. Computational experiments
are conducted on all ten instances of Brandimarte [1] and five instances of Kacem
et al. [11] since these instances are widely solved in literature and hence extensive
comparison of solution methodologies can be carried out. Kacem’s instances range
from sizes of 4 jobs and 5 machines to 15 jobs and 10 machines with a single
instance of 8 jobs and 8 machines with partial flexibility, while the remaining have
total flexibility. Brandimarte’s instances consist of problems ranging from sizes of
10 jobs and 6 machines to 20 jobs and 15 machines with medium flexibility.
From Table 1, it can be seen that the proposed hybrid approach obtains best solu-
tions in ten out of fifteen benchmark instances. In comparison with other algorithms,
it can be seen that the proposed algorithm outperformed the algorithm proposed by
Xing et al. [14] in seven problems. The performance of the hybrid ABC was superior
to TLBO proposed by Buddala et al. [15] in six problem instances. The proposed
approach gave better results to two problems each compared to AIA proposed by
Bagheri et al. [16] and HTSA proposed by Li et al. [17]. The proposed ABC outper-
formed BBO proposed by Rahmati et al. [18] in eight problem instances. The conver-
gence of Brandimarte’s MK09 problem is shown (see Fig. 5). Gantt chart obtained
using the proposed approach is shown for Kacem’s 10 * 15 problem instance (see
Fig. 6).
840 R. Caldeira et al.

Fig. 4 Flowchart of the proposed hybrid ABC

Solving the Flexible Job Shop Scheduling Problem Using … 841

Table 1 Results of benchmark problems

S. No. Problem Xing’s HTSA AIA BBO TLBO Proposed
size algorithm [17] [16] [18] [15] ABC
[14]
1 4×5 11 11 NA 11 11 11
2 8×8 14 14 14 14 14 14
3 10 × 7 11 11 NA NA 11 11
4 10 × 10 7 7 7 7 7 7
5 15 × 10 11 11 11 12 12 11
6 (MK01) 10 × 6 42 40 40 40 40 40
7 (MK02) 10 × 6 28 26 26 28 28 27
8 (MK03) 15 × 8 204 204 204 204 204 204
9 (MK04) 15 × 8 63 61 60 64 63 60
10 15 × 4 181 172 173 173 172 172
(MK05)
11 10 × 15 60 65 63 66 65 63
(MK06)
12 20 × 5 148 140 140 144 144 140
(MK07)
13 20 × 10 523 523 523 523 523 523
(MK08)
14 20 × 10 308 310 312 310 311 307
(MK09)
15 20 × 15 212 214 214 230 214 214
(MK10)
Note The significance of the “Bold” is to indicate that the proposed problem solution is better than
benchmark problem solutions

5 Conclusion and Future Scope

This paper considers the NP-hard FJSP with the makespan criteria, and the same
is solved by a proposed hybrid algorithm where a TS is incorporated in the ABC
algorithm. The effectiveness of using TS as a local search is highlighted by the
improved exploration and convergence capability of the ABC algorithm in the local
search space. The proposed hybrid ABC algorithm showed superior performance in
comparison with other well-known meta-heuristics on Kacem’s and Brandimarte’s
instances.
This work can be extended to solve other scheduling problems with constraints.
842 R. Caldeira et al.

Fig. 5 Convergence of MK09 instance

Fig. 6 Gantt chart of

Kacem’s 10 × 15 instance

References

1. Brandimarte P (1993) Routing and scheduling in a flexible job shop by tabu search. Ann Oper
Res 41(3):157–183
2. Sharma N, Sharma H, Sharma A (2018) Beer froth artificial bee colony algorithm for job-shop
scheduling problem. Appl Soft Comput 68:507–524
3. Xiang W et al (2018) A novel artificial bee colony algorithm based on the cosine similarity.
Comput Ind Eng 115:54–68
Solving the Flexible Job Shop Scheduling Problem Using … 843

4. Sharma H, Bansal JC, Arya KV, Yang XS (2016) Lévy flight artificial bee colony algorithm.
Int J Syst Sci 47(11):2652–2670
5. Thammano A, Phu-ang A (2013) A hybrid artificial bee colony algorithm with local search for
flexible job-shop scheduling problem. Procedia Comput Sci 20:96–101
6. Zhang R, Shiji S, Cheng W (2013) A hybrid artificial bee colony algorithm for the job shop
scheduling problem. Int J Prod Econ 141(1):167–178
7. Wang L, Zhou G, Xu Y, Wang S, Liu M (2012) An effective artificial bee colony algorithm for
the flexible job-shop scheduling problem. Int J Adv Manuf Technol 60(1–4):303–315
8. Li JQ, Xie SX, Pan QK, Wang S (2011) A hybrid artificial bee colony algorithm for flexible
job shop scheduling problems. Int J Comput Commun Control 6(2):286–296
9. Banharnsakun A, Sirinaovakul B, Achalakul T (2012) Job shop scheduling with the best-so-far
ABC. Eng Appl Artif Intell 25(3):583–593
10. Karaboga D, Gorkemli B (2012) A quick artificial bee colony-qABC-algorithm for optimization
problems. In: International symposium on innovations in intelligent systems and applications.
IEEE, pp 1–5
11. Kacem I, Hammadi S, Borne P (2002) Approach by localization and multiobjective evolutionary
optimization for flexible job-shop scheduling problems. IEEE Trans Syst Man Cybern Part C
(Appl Rev) 32(1):1–13
12. Saidi M, Fattahi P (2007) Flexible job shop scheduling with tabu search algorithms. Int J Adv
Manuf Technol 32(5–6):563–570
13. Hurink J, Jurisch B, Thole M (1994) Tabu search for the job-shop scheduling problem with
multi-purpose machines. Oper Res Spektrum 15(4):205–215
14. Xing LN, Chen YW, Yang KW (2009) An efficient search method for multi-objective flexible
job shop scheduling problems. J Intell Manuf 20(3):283–293
15. Buddala R, Mahapatra SS (2018) An integrated approach for scheduling flexible job-shop using
teaching–learning-based optimization method. J Ind Eng Int 1–12
16. Bagheri A, Zandieh M, Mahdavi I, Yazdani M (2010) An artificial immune algorithm for the
flexible job-shop scheduling problem. Futur Gener Comput Syst 26(4):533–541
17. Li JQ, Pan QK, Liang YC (2010) An effective hybrid tabu search algorithm for multi-objective
flexible job-shop scheduling problems. Comput Ind Eng 59(4):647–662
18. Rahmati SHA, Zandieh M (2012) A new biogeography-based optimization (BBO) algorithm
for the flexible job shop scheduling problem. Int J Adv Manuf Technol 58(9–12):1115–1129
Linear Programming in Market
Management Using Artificial Intelligence

Shetty Vignesh Uday, Hamritha, and Gaurav Chaudhary

Abstract The intelligence exhibited by machines in decision making using mathe-

matical algorithms in contrast to the human intelligence is known as artificial intel-
ligence (AI). It is an area of engineering science that focuses on the modeling of
intelligent systems that work and act in response like humans. The tasks that AI
includes are face recognition, voice recognition, planning, learning, and predicting.
Linear programming problem (LPP) is a tool to solve the real-life problems using
the standard mathematical approach that gives us the optimal solution. This paper
aims at providing the investors an aid in decision making while advertising their
business with combination of the tools that are available. Neural network model will
be formulated and trained on the basis of dataset that includes past data of variables
such as amount of investment, type of business, demographical area, and consumer
attributes. The output nodes include the variables based on the type of advertisement.
After training and testing are finished, user provides the model with the inputs and
the output predicts the optimal percentage of marketing of product in different fields.
This way the model will help us to make sound decisions in allocating the total
investment among the variables.

Keywords Linear programming problem · Marketing · Artificial intelligence ·

Productivity

1 Introduction

Artificial intelligence is widely used in decision making and prediction using math-
ematical algorithms. Artificial neural network (ANN)-based system works similar

S. V. Uday (B) · Hamritha

Department of Industrial Engineering and Management, Ramaiah Institute of Technology,
Bangalore 560054, India
e-mail: shettyvignesh01@gmail.com
G. Chaudhary
Department of Electronics and Communication Engineering, Ramaiah Institute of Technology,
Bangalore 560054, India

© Springer Nature Singapore Pte Ltd. 2021 845

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_73
846 S. V. Uday et al.

to a human brain in an individual’s body. This is an easier method and guarantees

consistency across the marketing chain [1]. An artificial neural network is a network
of many interconnected nodes connected by unidirectional communication chan-
nels, called weights that carry numeric data which adjusts itself as per the dataset
that contains past data of variables such as amount of investment, type of business,
demographical area, and consumer attributes.
Linear programming enables us to make sound decisions in several fields based
on the given constraints. There are several ways to solve a linear programming itself
depending upon the type of the problem [2]. This is one of the oldest and most
dependable methods which have already proved giving the most optimized results
in several fields like scheduling, food and agriculture, energy and transportation
problems. It is fundamentally used for allocating the limited resources since they are
not abundantly available in the nature.
One of the most upcoming technologies of this era is the concept of artificial
intelligence, where the system is trained and tested for predicting, categorizing, and
collecting the data using facial recognition, theoretical data, speech recognition,
virtual reality, image processing, visual audio manipulation, etc [3].
The computer-based system simulates the problems on par with human processing
system. There are several ways where the system could be classified depending on
the processing technique and applications.
A product to be sold needs to have a solid marketing strategy, existing marketing
scenario has been completely evolved, and compared to the past years, there are
several types of marketing methods like digital marketing, direct marketing, inbound
marketing, cause marketing, guerrilla marketing, relationship marketing, multi-level
marketing, personalization marketing, database marketing, etc. But the confidence
interval of the product promising the returns plays a vital role in the type of marketing
strategy that is adopted.
So, this paper in the whole discusses the algorithm and crucial part of the choosing
appropriate marketing techniques using artificial-based system. The basic objective
function of linear programming is to maximize the reach, frequency, and quality of
exposure within the constraints of cost and availability of variables. Each of these
depends on the product, demography, investment, customer behavior, and economy
[4].
Henceforth, as the constraints increase, the factors on which it depends will contain
huge set of data. As the data size increases, it is difficult to study and analyze the
data manually. Therefore, linear programming problem can be framed based on type
of the problem.
The real-world linear programming problem has a large set of variables and is quite
difficult to solve manually which can be resolved with the help of programs and solve
the problem based on the input set of constraints and obtain feasible solutions. This
can be illustrated using limited set of constraints but when making use of software
which are available in the market for solving the problem constraints considered
are many. Algorithms are written for solving problem in the area of marketing, and
also it would help in the prediction of the market returns of the product.
Linear Programming in Market Management … 847

2 Methodology

The formulation of a linear programming problem considers only few variables and
can be solved manually, but on real time applications, the variables considered are to
be many. The main focus is not to solve a linear programming problem, but utilizing
the linear programming system by a computer-embedded program that can solve the
figures of large variables and generate the best suitable answer based on the past data
and also configures the future prediction of the product chosen by the people.
In this system, input is taken from the end user like minimum investment, maxi-
mum investment, area, etc., and these inputs are fed to the system with ‘n’ variable,
and from the past records, it gives the output how the amount should be invested in
each format of marketing. So, data in the system enables the customer to decide on
the investment and thus helps to make a smart business.

2.1 Common Terminologies Used in Linear Programming

Decision Variables The decision variables are those which will decide the output of
a problem statement. Decision variables need to be identified to solve any problem.
For the defined problem, decision variables are “add input constraint names the ones
that are visible” investment, type of business, city, time (in months), and assets.
Objective Function It is the prime function which is to be worked upon depending
upon the problem. In this problem, the company wishes to know the better option
in investing money in advertisement (“out of 4 options, add name of output nodes
here”) digital marketing, newspapers, pamphlets, and banners.
Following is the algorithm can be considered for obtaining the desired output:
The model comes under multiple class neural network classification that involves
building neural networks that map the input to the network and output containing
more than two classes. In this case, dependency on the input and output is not known.
Therefore, three hidden layers are used with reLu as activation function for hidden
layer and Softmax for output layer (Fig. 1).
Supervised Learning During the training of ANN under supervised learning, the
input vector is presented to the network, which will give an output vector. The
obtained result vector is compared with the desired result vector. An objection signal
is generated if there is a difference between the actual result and the desired result
vector.
The weights are adjusted (added or subtracted with value) until the actual output
is matched with the desired output. It is a non-recurrent network that has processing
units or nodes in layers, and all the nodes in a layer are connected to the nodes
of the previous layer and also to node of next layer. The connection has different
weights upon them (chosen randomly before training). These weights change as per
848 S. V. Uday et al.

Fig. 1 Basic neural network

model

the dataset using back-propagation algorithm to give us an output with acceptable

error.
Activation functions are used to map the output of a layer between 0 to 1 and −
1 to 1 depending on the type of function used. Rectified linear activation function—
reLu—is the mostly used activation function (Fig. 2).
f (z) is zero, if z is less than zero. f (z) is z, if z is greater than zero. The function
and its derivative both are monotonic.
Softmax activation function is used for final layer, where the Softmax calculates
a probability for every possible class. By applying the standard exponential function
to each element ‘z’ of the input vector, normalize these values by dividing the sum
of all these exponentials, as shown in the formula below.

Fig. 2 Graphical
representation of input
output of reLu
Linear Programming in Market Management … 849

eβz j
σ (z) =  K (1)
βz k
k=1 e

Training Algorithm For training, use reLu activation function for hidden layers and
Softmax for the output layer. The training of neural network will have the following
four phases.
Step 1: Feed-forward phase:
• Initialize to start the training with following inputs
– Weights (choose random weights)
– Learning rate α (should be small)
• Calculate activation function (a) of unit i in layer j using below relation
 n 
j
 j
ai =g xi Hi (2)
i=1

where i is layer of which the activation function is being calculated.

Step 2: Calculate the error and adjust weights as follows:
• Cost function in neural network for multiple output will be:

 m 
1  (i)
J (θ ) = − t log(yθ (xi ))k + (1 − tk ) log(1 − yθ (xi ))k
i
(3)
m i=1

where m is total number of nodes in a layer, t is desired output, and (yθ )i is ith output.
Step 3: Back propagation:
Intuition: δ ij = error of node j in layers i.
Intuition for output layer is:

δ (m−1)
j = y (m)
j − t (m)
j

where δ (2)
j is intuition of the output layer
**In most cases, back propagation is calculated by the backend library itself**
Step 4: Update the weights in the network until dataset classified correctly or stopping
criterion satisfied.
∂J
H d+1 = H d − ∝ (4)
∂h
where ∝ = learning rate (should be low), d is number of iterations, and J is loss
function
850 S. V. Uday et al.

Table 1 Results obtained for

Output class Probability (%)
the input variables
Digital marketing 0.84
Newspaper 0.10
Pamphlets 0.05
Banners 0.01

Testing Stage Simulate the network with new dataset. By using various test cases,
neural network will be trained and that network will be used to solve real cases by
linear programming problem.
Outcome Data is captured to get the best possible way of advertising, calculated by
the neural network

3 Results and Discussion

Considering the marketing issues and availability, it is difficult to obtain the solutions
considering all the variables. Hence, by using a formulated mathematical model, it
is easier to obtain the probability of variables. Hence, the objectives can be met and
profits can be maximized in real-life applications.
As marketing strategy is being one of the most important criteria of a product, it is
appreciated if it reaches the end user with necessary requirements in short duration.
The above model will help the marketing team to meet these demands by training
the algorithm. The probability of each variable will be obtained, and with the help of
these values, suitable constraints and variables can be identified. This model helps
the user decide on expenditures and types of marketing.
To demonstrate the working of the trained neural network model, the input is
taken as [200,000, ‘non-profit’, ‘Mumbai’, 15, 3,500,000].
Passing the above as input, the probability obtained is as follows (Table 1).
The above output obtained will help marketing team on understanding the types
of marketing and the priority.

4 Conclusion

This paper gives a complete set of algorithms for the system to train and learn the
marketing environment and predict the most accurate outputs using the system of
linear programming. As the time passes, large datasets will be collected and the
system would have more variables to match and work upon; this will keep system
improving.
In the future, the scope of obtaining the feedback in the system can be incorporated
which will help the system constantly to update and develop. The developed system
Linear Programming in Market Management … 851

will help both the marketing firms to develop individually and also reach to the
customers based on their preferences. Each of the products marketed varies from
each other, based on the type of advertisement which varies from the demographical
area. Based on place, people, and economy, a customized marketing plan can be
formulated considering all possible variables.
Manually, only few variables can be formulated, as the complexity increases, and
when the real-life solutions are needed to meet the customer satisfaction, ANN would
give the best solutions to the problem within least time which gives a better impact
on the marketing firm.

References

1. Kennedy MP, Chua LO (1988) Neural networks for nonlinear programming. IEEE Trans Circuits
Syst CAS 35:554–562
2. Ghasabi-Oskoei H, Mahdavi-Amiri N (2006) An efficient simplified neural network for solving
linear and quadratic programming problems. Appl Math Comput J 452–464
3. Wen U-P, Lan K-M, Shi H-S (2009) A review of Hopfield neural networks for solving
mathematical programming problems. Eur J Oper Res 675–687
4. Gao X, Liao L-Z (2010) A new one-layer, neural network for linear quadratic programming.
IEEE Trans Neural Netw 21(6)
Queueing Theory an Index Factor
for Production Inventory Control
in Automotive Industry—A Review

Sunday A. Afolalu, Segun Oladipupo, Samson O. Ongbali,

Abiodun A. Abioye, Ademola Abdulkareem, Mfon O. Udo,
and Oluseyi O. Ajayi

Abstract In this paper, various approaches to inventory control within the automo-
tive industry were reviewed using queueing theory. Different models used in past
literature were stated and the model considered to be most effective in this review
is dock management modeling. This model was used to analyze inventory control
and how effectively automotive industries can minimize inventory by getting the
component needed in the assembly line just in time, this helps to reduce additional
costs for warehouse maintenance and capital tied down in form of excess inventory.
Analytical and simulation models are the mathematical models that are considered
in this review as they are used in several papers by different authors.

Keywords Queueing theory · Index factor · Inventory · Automotive · Model

1 Introduction

The resultant effect of a queue could be a result of server factors of which the inability
of servers not to meet up with the demand of the customers is one of them, this
increases waiting time and length of the queue [1]. Other factors that affect a queueing
system are service time, arrival pattern, number of servers, etc. Queues can be found in
various settings which could either be in the manufacturing or service organization;
in this case, the major concern in the manufacturing sector taking the automotive
industry as a case study. In recent times, there has been an increase in variety of
products and mix models in order for the manufacturers to effectively compete in
the global market. For manufacturers to stay relevant, flexibility in manufacturing
is an important concept that must be embraced [2]. However, flexibility has its own
cons which are high initial investment which requires large inventory and this is a
non-negligible cost [3].

S. A. Afolalu (B) · S. Oladipupo · S. O. Ongbali · A. A. Abioye · M. O. Udo · O. O. Ajayi

Mechanical Engineering Department, College of Engineering, Covenant University, Ota, Nigeria
e-mail: sunday.afolalu@covenantuniversity.edu.ng
A. Abdulkareem
Department of Electrical and Information Engineering, Covenant University, Ota, Nigeria
© Springer Nature Singapore Pte Ltd. 2021 853
S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_74
854 S. A. Afolalu et al.

Manufacturing car, parts are stamped from steel rolls and the entire body is then
welded to produce a shell, and then each composition is finally welded in the body
on the moving assembly line [4]. It was observed in the automotive industry that
inventory management is a key factor in maximizing productivity and minimizing
waste [5]. Inventory can be in different forms like raw material, work in progress,
finished products, and goods in the warehouse between customers and suppliers. It
is important for all industries to manage inventory properly as it is determinant on
how productive and profitable the business can be one of the most successful models
used to tackle inventory problems in this industry is just-in-time manufacturing that
makes use of the supermarket rule of taking exactly what is needed in the assembly
line from the warehouse [6].
The efficiency of the production system can be determined using the batch size
and throughput. The batch size and the throughput must be aligned to optimize the
system in order to make sure that the system is efficient [7]. Queueing network
was used in this paper to study how throughput and batch size affects the optimal
utilization of resources, especially machine resources in manufacturing systems. This
study is focused on plants in the manufacturing industry that carry out assembly
operations. The result of this study shows that utilization is a function of batch size
and throughput, which means an increase or decrease in batch size or throughput
directly affects the increase and decrease in utilization. If the demand requirement
is higher than the capacity bottlenecks will definitely occur. For the manufacturing
industry to remain relevant, sustainable, and compete strongly they must always
find a way of improving efficiency in the production process. The production line
where production is carried out is divided into three parts, namely single, batch,
and mixed-model line [8]. The machines in the production lines have a cycle time
in which they operate. In a manufacturing system, the efficiency of the production
operation is based on how the machines utilize the resource in particular cycle time.
Some operations did not make full use of the resources, this resulted in increased
cycle time which also affects the throughput rate at which a product enters and
leaves the production line. However, the ideal batch size and throughput for a given
production can be defined with the help of the queueing theory in order to optimally
use resources. In the manufacturing system, modeling can be done by using either an
analytical model or a simulation model. Queueing network is one of the analytical
models used in evaluating manufacturing system [9].

2 Reviewed Literature

In the automotive assembly, the techniques and emerging technologies used in vehicle
assembly plants are discussed in this paper. Assembly processes such as handling,
joining, etc., that involves the use of technology, as well as supporting systems were
considered in this paper. A summary of existing technologies is provided. Mass
customization requirements have been a challenge for the assembly systems as they
are unable to keep up, this means that upgrading technology is necessary to achieve
Queueing Theory an Index Factor for Production Inventory … 855

high flexibility. For operation and system, design flexibility is a key factor that should
be integrated into future assembly plants [10].
The automotive industry faces great challenges such as advancement in technol-
ogy, new competitors, and industry restructuring, and sourcing globally. The quick
change from mass production to mass customization is dependent on the fact that
customers wants certain features incorporated into their products and manufactur-
ers have to provide many variants using fewer resources in a relatively short period
[11]. The ability of the manufacturer to provide several variants per model and bring
onboard new models at a fast rate is restricted by the current equipment used in mass
production and also current technologies that are not capable of supporting product
variability. In automotive assembly increased complexity requires an encompass-
ing view of the main attributes of manufacturing that should be considered when
manufacturing decisions relating to cost, quality, flexibility, and time are taken [12].
In the automotive industry, the typical framework of a plant comprises four phases
- stamping, body shop, painting, and final assembly. Assembly is carried out mainly
in the body shop or final assembly. The assembly of the body in white (BIW) uses high
level of automation, while the final assembly stage introduces a combination of hybrid
human/machine systems. Four approaches can thus be identified when designing the
assembly system; (1) flexible assembly (2) manual assembly, (3) semi-automated
assembly, (4) fixed assembly.
Generally, different vehicles in an automotive assembly make use of the same
assembly line known as mixed-model assembly line (MMAL). The MMAL are
distinguished because of their ability not to hold large inventory when assembling
different models of a given product [13]. Demand for more secured and faster ramp-
up processes which are cost-intensive is one of the major problems modern assembly
system has to tackle. Steady increase in number of products demanded underpins
the goal, which is a result of improved innovation and increased market launches
of product variant and new product. Groover [9] further expounded that the present
trend that Original Equipment Manufacturers (OEMs) in automotive industry follow
is incorporating equipment, product and process standardization.
One of the probabilistic models used for predicting and making a decision under
uncertainty are Bayesian networks. The inventory, production quantity and deliver
quantity of any manufacturing organization such as the automotive industry varies
according to the several unexpected events [14]. Thus, predicting inventory in pro-
duction is sacrosanct in order to cope with the variations in the market. This paper
considers the way to adjust production for the car parts process through the use of
a dynamic Bayesian network. Material in a production system (known as factory
automation) is delivered to clients and stored in inventory for subsequent shipment
through each manufacturing, sub-assembly and final product process [15].
To utilize a system like factory automation, attributes of production such as pro-
duction rate of the company varies because of the irregular changes found generally
in the operating capabilities and also the strength of the labor force. There are random
changes to product order which means managers of each process have to decide on
the output of products by using their know-how and experience to tackle this issue
[16]. Taking a look at the inventory in a Kanban way, the inventory is considered
856 S. A. Afolalu et al.

to be a cost. Kanban is used to schedule for just-in-time manufacturing, it oper-

ates on the supermarket principle of picking the things necessary for production at
the moment from the store. From a production point of view, Kanban controls the
logistical and supply chain [17]. In a manufacturing system, putting uncertainties
into consideration is of upmost importance because it helps proper planning against
unforeseen events. Production planning models that put uncertainties into consider-
ation are better off as they generate excellent planning than the models that don’t
account for uncertainties [18]. This paper reviews existing production planned liter-
ature under uncertainty. According to Pearl [22], uncertainty can be defined as the
distinction between the information already gained and the amount of information
actually needed to carry out the task. There are many ways in which uncertainty have
an impact on production processes. It is classified into two groups, uncertainty about
the environment and system uncertainty. Environmental uncertainty is those beyond
uncertainties in demand and supply; this type of uncertainty is not influenced by the
production process itself. System uncertainty refers to uncertainty that is influenced
directly by the production process, such as failure of the production system, quality
uncertainty, lead time uncertainty, operation yield uncertainty and changes to the
structure of the product. The typology of uncertainty is used in this paper [19].
The supply chain of automotive industry is characterized for a properly organized
structure, where the partial exchange of forecasting amongst manufacturers and their
first-tier suppliers are requirement demand; this means that suppliers in most cases
must be ready to handle a variety of potential requirements. In order to handle such
varieties, it important to consider technological tools for efficiently managing the
decision-making in this area [20].
The task of logistics planning in the automotive industry was studied in this
paper using the macroscopic simulation model [21]. In specific terms, macroscopic
simulation model lies between simulation models that are object-based discrete-
event or flow-based continuous simulation models. On a broader view, processes
of logistic flow are represented using macroscopic model via a piecewise flow rate
that is constant rather than modeling flow objects individually. The manufacture of
assembled products with various components such as cars, engines, etc. involves
the collaborative effort of several factories. The individual components are to be
designed and organized properly in order to meet customers’ expectations [22]. The
main concern here is the decision problems (medium and long-term), which contains
supply chain macroscopic models. As stated by Lee et al., continuous worldview is
used at this level, as representing the individual flows of numerous parts becomes
quite tedious. The macroscopic models are based on the paradigm of Forrester system
dynamics that are currently used by large French automotive companies [23]. In this
article, the supply chain presented is made of two production areas with a total
of five existing plants. The results show the real advantages that can be achieved.
Through ferrous metal separation, 75% of the material weights are successfully
recovered by the automotive recycling infrastructure once the lifecycle of a vehicle
is complete [24]. Nevertheless, the rate at which automotive manufacturers make use
of nonmetallic and nonferrous material has increased and this a major challenge for
the automotive recycling industry [25]. Recycling planning models were proposed in
Queueing Theory an Index Factor for Production Inventory … 857

this paper for automotive shredders in order to make a short-term strategic decision
with regards to the level at which materials should be processed and reprocessed
through multiple passes. A model known as mixed-integer linear programming is
used to determine whether to separate or combine materials for shipment [26].

3 The Problem

3.1 Material Management

Generally, in supply chain materials move through the distribution center from the
supplier and finally arrive at the moor of the assembly plant. The arriving components
or materials are then taken immediately to the assembly line or the inventory. Prompt
arrival of the components determines the correct feeding of the assembly line because
of any mistake in leaving a part out result to heavy cost for production re-scheduling
[27].

3.2 Feasible Alternatives

There are two ways of addressing any of the delays from the supplier, through means
of transport or dock, they discussed as follows: Transport management: this is taking
into consideration when the supplier doesn’t deliver on time or problems arise while
transporting the goods. In this type of scenario is an alternative which is to take
advantage of faster means of transportation in order to reduce the lead time of goods
arriving at the plant. It should be noted that change to a faster means of transportation
will lead to additional cost for faster delivery [28].
Dock management takes care of any form of disruptive event that can occur during
delivery of goods.

4 Solution

4.1 Dock Management Model

Askin et al. [28] stated that incoming disruptive events managed by this model led
to a reduction in the total cost of updating the dock work, which helps to identify the
most appropriate dock management solution. The main objective here is to remember
the actual production time that displays the time when the assembly line requires an
incoming order. Based on this time, calculating when the order at the dock should be
ready is possible at maximum. The time difference between production time and time
858 S. A. Afolalu et al.

required to transfer orders for the assembly line from the warehouse is known as due
time [29]. This transfer can be managed in two ways, which are the Standard method
and strategies that are faster which include advantageous transport and an increase in
price which is used to make up for the delay experienced that arrives after due time.
Two moments name acceptable due time_S1 and _S2 are defined by the two strategies
mentioned above, these are calculated as the difference between respective transfer
time and actual production time. Any delayed order arriving after these moments are
transferred for production re-scheduling [30]. Also, the time between the specific
window time when the production re-scheduling is required and due time is known
as buffer time. Arrivals after the buffer time are placed within the re-scheduling time.
Orders that takes place after the estimated arrival time is due to disruption during
order delivery [31].
A specific amount of time known as the unloading duration is used to unload and
transfer goods to the warehouse after the order arrives at the dock. After unloading,
the order is then placed on the assembly line. If the estimated readiness time is before
the due time, then there is no additional cost. If the time of readiness after the due
date is estimated, then the buffer time or cost of re-scheduling will be additional as
a matter of time [32].

5 Conclusions

A number of manufacturing industries like the automotive industry took the use of
queueing theory in inventory management, leading to reductions in lead times to
maximize profit, mathematical models that include analytical models and simulation
models were evaluated in various literature, meanwhile in this paper, the model
considered to be effective is dock management, the use of the model result in proper
inventory keeping of materials necessary for the assembly line in order to avoid
stoppage during production [33–35].

Acknowledgements We acknowledge the financial support offered by Covenant University in

actualization of this research work for publication.

References

1. Balaji K, Kumar VSS (2014) Multicriteria inventory ABC classification in an automobile

rubber components manufacturing industry. Procedia CIRP 17:463–468
2. Başak Ö, Albayrak YE (2015) Petri net based decision system modeling in real-time scheduling
and control of flexible automotive manufacturing systems. Comput Ind Eng 86:116–126
3. Bhatnagar R, Lin B (2018) The joint transshipment and production control policies for multi-
location production/inventory systems. Eur J Oper Res
Queueing Theory an Index Factor for Production Inventory … 859

4. Bruniaux R, Pierreval H, Caux C (1999) Simulating supply chains with system dynamics. In:
Proceedings of the ESS’99 European simulation symposium, Nuremberg, Germany, 26–28
Oct, pp 209–213
5. Bruniaux R (2000) Simulation continue de réseaux de sites industriels: application aux chaines
logistiques dans le secteur automobile. Ph.D. Dissertation, in French, Université Blaise Pascal,
Clermont-Ferrand, France
6. Bruniaux R, Caux C (2007) A continuous simulation approach for supply chains in the
automotive industry. Simul Model Pract Theory 15(2):185–198
7. Boysen N, Emde S, Hoeck M, Kauderer M (2015) Part logistics in the automotive industry:
decision problems, literature review and research agenda. Eur J Oper Res 242(1):107–120
8. Dalmijn WL (1990) Practical applications of eddy current separators in the scrap recy-
cling industry. In: van Linden JHI, Stewart DL Jr, Sahai Y (eds) Proceedings of the second
international symposium on recycling of metals and engineered materials, pp 303–314
9. Groover MP (1987) Automation, production systems, and computer integrated manufacturing.
Prentice-Hall, Englewood Cliffs, pp 36–40, 404–407
10. Ho C (1989) Evaluating the impact of operating environments on MRP system nervousness.
Int J Prod Res 27:1115–1135
11. Michalos G, Makris S, Papakostas N, Mourtzis D, Chryssolouris G (2010) Automotive assem-
bly technologies review: challenges and outlook for a flexible and adaptive approach. CIRP J
Manufact Sci Technol 2(2):81–91
12. Hu SJ, Zhu X, Wang H, Koren Y (2008) Product variety and manufacturing complexity in
assembly systems and supply chains. CIRP Ann Manuf Technol 57(1):45–48
13. Lang S, Reggelin T, Wunder T (2017) Mesoscopic simulation models for logistics planning
tasks in the automotive industry. Procedia Eng 178:298–307
14. Marsudi M, Shafeek H (2013) Production line performance by using queuing model. IFAC
Proc Vol 46(9):1152–1157
15. Marsudi M, Dzuraidah AW, Che Hassan CH (2009) Application of spreadsheet and queuing
network model to capacity optimization in product development. World Acad Sci Eng Technol
34:1068–1071
16. Salawu EY, Okokpujie IP, Afolalu SA, Ajayi OO, Azeta J (2018) Investigation of production
output for improvement. Int J Mech Prod Eng Res Dev 8(1):915–922
17. Afolalu AS, Enesi YS, Kehinde O, Samuel UA, Ikechi VI, Remilekun RE (2018) Failure mode
and effect analysis a tool for reliability evaluation. Eur J Eng Res Sci 3(4):65–68
18. Salawu EY, Ajayi OO, Inegbenebor AO, Afolalu SA, Ongbali SO (2018) Pareto analysis of
product quality failures and cost effects in bottling machines—a lean thinking solution for
alcohol industry. Int J Mech Eng Technol 9(11):2380–2388
19. Afolalu SA, Ayuba SU, Ihebom IV, Elewa RR, Oluyemi K (2018) Role of reliability man-
agement tools and documentations—a review. Glob J Eng Sci Res Manage 5(3):39. ISSN
2349-4506
20. Meyer A, Sejdovic S, Glock K, Bender M, Kleiner N, Riemer D (2017) A disruption man-
agement system for automotive inbound networks: concepts and challenges. EURO J Transp
Logist 1–32
21. Mula J, Poler R, García-Sabater JP, Lario FC (2006) Models for production planning under
uncertainty: a review. Int J Prod Econ 103(1):271–285
22. Pearl J (1986) Fusion, propagation and structuring in belief networks. Artif Intell 29(3):241–288
23. Lee YH, Cho MK, Kim SJ, Kim YB (2002) Supply chain simulation with discrete continuous
combined modeling. Comput Ind Eng 43(1–2):375–392
24. Rabelo L, Magdy H, Lertpattarapong C (2004) Analysis of supply chains using system dynam-
ics, neural networks and eigenvalues. In: Proceedings of the 2004 winter simulation conference,
Dec, Washington, DC, USA
25. Abioye AA, Atanda PO, Abioye OP, Akinlabi SA, Akinlabi ET, Bolu CA, Afolalu SA, Ajayi
OO, Ohijeagbon IO (2018). A review on automotive industries and foundries in Nigeria. In:
IOP conference series: materials science and engineering, vol 413, no 1. IOP Publishing, p
012003
860 S. A. Afolalu et al.

26. Ramírez-Granados M, Hernàndez JE, Lyons AC (2014) A discrete-event simulation model for
supporting the first-tier supplier decision-making in a UK’s automotive industry. J Appl Res
Technol 12(5):860–870
27. Remiel FM, Cauvin A, Ferrarini A (2014) Conceptual design and simulation of an automotive
body shop assembly line. IFAC Proc Vol 47(3):760–765
28. Askin RG, Mitwasi MG, Goldberg JB (1993) Determining the number of kanbans in multi-item
just-in- time systems. IIE Trans 25(1):89–98
29. Sethi SP, Yan H, Zang Q (2002) Optimal and hierarchical controls in dynamic stochastic
manufacturing systems: A survey. Manuf Serv Oper Manage 4:133–170
30. Shin J, Kim S, Lee JM (2015) Production and inventory control of auto parts based on predicted
probabilistic distribution of inventory. Dig Commun Netw 1(4):292–301
31. Shiina T, Birge JR (2003) Multistage stochastic programming model for electric power capacity
expansion problem. Jpn J Ind Appl Math 20:379–397
32. Spengler T, Ploog M, Schroter M (2003) Integrated planning of acquisition, disassembly and
bulk recycling: a case study on electronic scrap recovery. OR Spectrum 25:413–442
33. Ajayi OO, Agarana MC, Akinwumi II, Okokpujie PI, Salawu EY, Abioye AA, Afolalu AS,
Leramo RO (2019) Modelling rate of traffic-induced building vibrations in Sango-Ota, Nigeria:
an assumption based analysis. WIT Trans Built Environ 182:303–313
34. Weiss EN, Tucker C (2018) Queue management: elimination, expectation, and enhancement.
Bus Horiz 61(5):671–678
35. Williams JAS, Wongweragiat S, Qu X, McGlinch JB, Bonawi-tan W, Choi JK, Schiff J (2007)
An automotive bulk recycling planning model. Eur J Oper Res 177(2):969–981
Design and Analysis of ASRS Using AGV
for Rapid Inventory Storage System

Pranav Santhosh Nair, Sourabh Nair, Soukhin Sarkar,

Arockia Selvakumar Arockia Doss, and D. Dinakaran

Abstract Non-competitive market pricing is easy to use and maintain but it is a major
issue in automated guided vehicles (AGVs) that are available in the current market.
Space is a very valuable commodity in urban areas. The current AGV available in
the market needs optimization and is specifically designed for one particular form of
storage system. In order to reduce the complexity of the system, hard locating method
approach is used. Hard locating is a process that assists the operator in avoiding
mistakes. In this work, an AGV for rapid inventory storage system is proposed. The
AGV consists of a 5-degrees of freedom (DOF) robotic arm with a two-fingered
gripper is used for picking and placing the boxes. The gripper is equipped with
a piezoelectric sensor to sense the size of the object. By using this approach, the
existing rack can be used without much alteration. This system has the capability to
increase the utilization of the warehouse. This paper presents an optimized automated
storage and retrieval system (ASRS) with the help of AGV with reduced cost and
complexity.

Keywords Automated guided vehicle · Automated storage and retrieval system ·

Warehouse optimization · Piezoelectric sensor

1 Introduction

The design and implementation of a 7-channel line sensor for line following robot
were having advanced functionalities such as software-based automatic sensor cal-
ibration and individual threshold per sensor channel. The system was implemented
using light-emitting diodes, phototransistor and an on-board microcontroller Arduino
which communicates with any navigation control system using serial communication
[1]. The optimization problem of AGV in smart warehouse dispatching mechanism

P. S. Nair · S. Nair · S. Sarkar · A. S. Arockia Doss (B)

School of Mechanical Engineering, Vellore Institute of Technology, Chennai, India
e-mail: arockia.selvakumar@vit.ac.in
D. Dinakaran
Centre for Robotics and Automation, HITS, Chennai, India
© Springer Nature Singapore Pte Ltd. 2021 861
S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_75
862 P. S. Nair et al.

is studied with genetic algorithm [2]. The AGV logistics and distribution compo-
nents which are used for AGV in discrete manufacturing process are explained [3].
AGV is a kind of simple mobile robot, which integrates light, sound, electricity
and computer, and it is applied in the fields of manufacturing industry and logis-
tics, such as automobile industries [4]. An AGV with low-cost camera was used to
capture images to figure out the surrounded conditions and decide the next move.
Region-based convolutional neural network-based deep learning algorithm was used
to analyse the images [5]. In smart factory, as one of key features for our industry, it
requires automation within a manufacturing plant such as a shop floor. Automated
guided vehicle (AGV) systems may be the best idea, whose accuracy depends on
some factors. An AGV with radio frequency identification (RFID)-enabled position-
ing system is used in for smart factory [6]. The robotic gripper was designed and
fabricated taking into account the economics, simplicity and reliability. Suitable-end
effectors can be installed for particular application [7]. The portable robotic arm with
4-degrees of freedom working model was able to imitate the movement of a human
arm with imperceptible delay and had a payload capacity of 150 gm that is achieved
[8]. The ASVS/RS rack design is focused on avoiding collision and deadlock in the
highroad traffic network, by restricting the layout and traffic regulations of trunk
rails, storage rails, lifters, etc. [9]. The ASRS is basically used to store or retrieve the
loads in the warehouse or out of it. The computer will calculate wherein the storage
area the item can be retrieved from and does the retrieval operation. The components
are stored into or retrieved from the racks, and the computer updates the inventory
system [10].
Automated storage and retrieval systems have many potential applications. The
most common ASRS application is the storage and retrieval of inventory component,
but other ASRS applications can greatly increase the efficiency of an operation as
well [11]. An automated guided vehicle (AGV) is a robot that follows marker (wires)
on the floor, magnets or lasers for navigation. With the rise of online shopping, the
requirement of large warehouses has increased exponentially. To manage these large
warehouses, an army of workers is needed, which is not very economical. A better
option would be to implement an automated storage and retrieval system by using
automated guided vehicle which will greatly help the institution in economic terms
and increase the storage density of the warehouse. The objective of the proposed
system is to find the perfect balance between the two systems such that the overall
efficiency of the warehouse storage can be optimized. Implementing human work-
force, wherein large ever-growing demand for storage places, human efficiency may
not update with the technology. Hence to increase the efficiency and storage den-
sity of the packages, implementation of ASRS becomes vital. The ASRS reduces
the probability of human-caused errors, and the AGV increases the efficiency and
the work duration. This is cheaper to handle, and very less maintenance is required.
The integration and optimization of these two complex systems are presented and
detailed in this proposed work.
Design and Analysis of ASRS Using AGV for Rapid Inventory … 863

Fig. 1 Warehouse layout

2 Proposed System

Figure 1 shows the warehouse layout. In front of each box levels, a line or a junction
is created to position the bot in front of the box. The boxes will be kept in the rack
using locating strips so that the position of the box will be known and fixed; hence,
the arm could be hardcoded to reach the designated position of the boxes. Based
on the proposed fixing position of the boxes, the complexity, both mechanically
and electrically, is reduced. Base of the AGV is kept as three-fourth of the track
width (distance between two columns) and not too short of the wheelbase that will
misbalance the centre of gravity of the bot while handling the load. The standby
dimensions of the bot are 800 * 630 * 900. These two package sizes (12.5 × 9.75
× 3.25 and 18 21.5 × 5.25 ) are the most common size used in the shipping
industry. The rack is a simple 3-level rack and used to store objects of various sizes
in inventory systems. The rack system is designed to help with the hardcoded gripper
mechanism and ensure zero error and arrangement of boxes. The arena is designed
such that the bot will have to go through minimum number of iterations. It will
receive a set of location addresses via Bluetooth, using the location instructions; it
will take the necessary turns, will grab the package with the help of gripper and get
back to its original position.

3 Design and Analysis

The 5-DOF manipulator is mounted on the chassis which is having four wheels
driven by DC motors. The end effector is driven by a DC motor for open and close
operation of the jaw, and it is designed to have a piezoelectric sensor to adapt to the
size of the object. The line following guidance system is designed for free movement
of the bot without any interference. The precise measurement of the centre of the
bot to the corner of the columns is kept in mind while marking the path. The gripper
is flexible with its gripping length; i.e. it can lift packages of various dimensions.
864 P. S. Nair et al.

Fig. 2 Arm assembly ideal

condition

Linear ball screw is used for moving the movable arm of the gripper. Figures 2 and
3 show the standby and fully extended position of the arm, respectively. The robotic
arm consists of 5 DOF, which is controlled using electric linear actuators and high
torque DC motors. The arm along with gripper has a lifting capacity of 100 kgs; i.e.
it can lift package weighing up to 100 kg. The AGV components details are tabulated
with specifications in Table 1.
The electric linear actuator proposed in this work is a simple gear and screw
mechanism. The electric motor rotates the main gear connected to the motor shaft
which is connected to the gear which is in turn connected to the screw. The screw
moves up or down depending on the direction of motion of the motor. The robotic
arm consists of three links as shown in Figs. 4 and 5. Link 2 slides over link 1 in order
to extend the overall length of primary link from 500 to 750 mm. Because of this
extendable link, the arm is able to lift objects placed at a height of 1.7 m. The sliding
motion is controlled by a square gear DC motor of torque 30 kgcm and 500 RPM.
The inclination of all the links is controlled by linear actuators of sizes 4 and 8 in.

4 Simulation

Structural analysis of the chassis of the bot is done to ensure that it can handle
the stress while handling maximum load and has enough lifetime before service is
needed. Kinematics is the field of focus while designing the arm. The load acting on
the joints, torque acting on the juncture and the material required to suit the conditions
Design and Analysis of ASRS Using AGV for Rapid Inventory … 865

Fig. 3 Arm assembly

extended condition

Table 1 Details of AGV

S. No. Part name Dimensions (mm)
components
1 Base chassis 800 * 550 * 210
2 Link 1 500 * 70 * 96
3 Link 2 340 * 125 * 70
4 Link 3 550 * 70 * 2.5
5 Linear actuator 4 cm 205 (min) 305 (max)
6 Linear actuator 8 cm 320 (min) 520 (max)
7 Wheel diameter 100

are calculated. The gripper force directly depends upon the power of the motor used
to move the jaws; keeping this in considerations, the gripper is designed. In order to
check the safety of the components, the components are simulated in the ANSYS
16.0 and the displacements of base and the links are calculated and tabulated in Table
2.
Figure 6 shows the total deformation of base chassis for the payload of 2500 N; the
deformation is found to be 0.000511 m; and it is found to have FOS of 4.8927. Figure 7
shows the total deformation of Link 1 for the payload of 2500 N; the deformation is
found to be 0.000144 m; and it is found to have FOS of 1.9249. Figure 8 shows
866 P. S. Nair et al.

Fig. 4 Prototype made for

static analysis

Fig. 5 Line following robot

using IR sensors

Table 2 Details of AGV components

S. No. Part name Force applied (N) Displacement (m) Factor of safety
1 Base 2500 0.000511 4.8927
2 Arm Link 1 2500 0.000144 1.9249
3 Arm Link 2 2500 0.000144 4.3403
4 Arm Link 3 1000 0.00028158 1.9169

the total deformation of Link 2 for the payload of 2500 N; the deformation is found
to be 0.000144 m; and it is found to have FOS of 4.3403. Figure 9 shows the total
deformation of Link 3 for the payload of 1000 N; the deformation is found to be
0.00028158 m; and it is found to have FOS of 1.9169.
Design and Analysis of ASRS Using AGV for Rapid Inventory … 867

Fig. 6 Deformation results

of base chassis

Fig. 7 Deformation results

of Link 1

Fig. 8 Deformation results

of Link 2

5 Control System Design

Figure 10 shows the line following sensors implemented in the AGV. The basic
manoeuvrability is found to be perfect, and further, it is preceded with the line
following mechanism. Two IR sensor arrays are placed one behind the other to
follow a straight line. Black tapes are used to mark the line. The operator will be
using a Bluetooth transmitter to send the coordinates to the Bluetooth receiving
868 P. S. Nair et al.

Fig. 9 Deformation results of Link 3

Fig. 10 Line detection

sensors

module situated on the AGV. The code will be received by the master Arduino.
The code consists of the coordinates of the rack where the package is placed or
is to be placed. The master Arduino will in turn control 2 other slave Arduinos
which controls both arm and base, respectively. The base controller Arduino will
manage the motor RPM, and the line following mechanism is to guide the bot to
the designated position according to the coordinates received from the operator. The
arm controller Arduino will be hardcoded to reach the hardcoded position of the
package. In Fig. 11, the control system of the AGV is shown. The Bluetooth module
will receive the command from the user which will in turn send the received command
to the master Arduino. According to the system layout, the master Arduino will give
commands to the slave Arduino 1 and slave Arduino 2 which controls the arm and
base of the AGV, respectively. The command received by the slave Arduino 1 will
Design and Analysis of ASRS Using AGV for Rapid Inventory … 869

Fig. 11 Control system unit

make the arm position itself in the desired position. The slave Arduino 2 will make
the base of the bot manoeuvre and position itself using the line following algorithm.
The code sent by operator consists of three basic parts. This helps in keeping the
code simple and prevents data skipping by Arduino due to excessive data values. An
example of the code that will be sent to the master Arduino situated on the bot by
the means of Bluetooth is given in Table 3.
The Bluetooth receiver is embedded on the Arduino which is connected to the
transmitter which is in turn controlled by the operator. The code is kept small so
as to prevent data skipping which might occur due to large data values. The master
Arduino is connected to two other slave Arduino which controls the arm and the

Table 3 Code to Arduino

1A 3B 2C
870 P. S. Nair et al.

base, respectively. The first two parts of the code are sent to the base slave Arduino.
The part A of the code is used to identify the rack location. The bot understands
the location of the rack, and the motors start to spin following the line. There are
two-line following sensors placed in line. The first sensor is used to implement the
line following mechanism. The second sensor acts as the counter which counts the
junctions. The value of A is being 1, the bot turns in alley 1 and so on. The part B is
used to identify the location of the rack column. This is used in similar method of line
following and counter like in part A to reach the designated location. After reaching
the location, the arm slave Arduino comes into play. The part C is used to identify
the height where the package is placed in the column. The arm will position itself
according to the location of the package. Because the position of the package is placed
according to the strips, the arm could be hardcoded. This enables swift movement
and less complexity. The arms of the gripper are fitted with a piezoelectric sensor
which helps in identifying whether the package is firmly gripped by the gripper or
not. The bot undergoes minimum number of iterations, simplifying the process. The
bot receives the code, goes to the junction, takes a left turn facing the rack, positions
its gripper in accordance with the height and picks up the package. After the package
is firmly gripped, the arm will retrieve back to its original position and the bot returns
back to the retrieval and distribution centre.
Figure 12 depicts the control layout of the Arduinos (4, 5, 6 from left) and motor
drivers (1, 2, 3, 7 from left) and Bluetooth module connected to the master Arduino
(4 from left).

Fig. 12 Control set-up

Design and Analysis of ASRS Using AGV for Rapid Inventory … 871

6 Summary

An AGV for rapid inventory storage system is fabricated and tested for industrial
application. Based on the design, structural analysis and experimental results the
following points are summarized,
• The line following sensors (line following) that make the AGV follow a single
path which makes the movement of the AGV much faster than the existing AGVs.
• This AGV can perform pick/place operations up to a height of 6 ft from the ground
using the robotic arm without taking out the whole rack for a single object.
• The AGV is very cost-effective. In general, the market value of an AGV is
approximately 5–10 lakhs. But the total cost of this AGV is around 0.30–0.35
lakh.
• Piezoelectric sensor allows AGV to avoid using visual sensor in assisting bot to
pick up the package, which in turn helps us in reducing the cost.
• As the system is hardcoded, the system follows the written code command and
any decision-making cannot be made by the AGV.

References

1. Patil S, Wagh A, Sawant M, Panda S, Bhopale A (2016) Design and implementation of

advanced auto calibrating line following sensor for coloured surfaces with a white line. In:
2016 IEEE 1st international conference on power electronics, intelligent control and energy
systems (ICPEICES), Delhi, pp 1–6
2. Liu S (2008) Optimization problem for AGV in automated warehouse system. In: 2008 IEEE
international conference on service operations and logistics, and informatics, Beijing, pp 1640–
1642
3. Qi BY, Yang QL, Zhou YY (2015) Application of AGV in intelligent logistics system. In: Fifth
Asia international symposium on mechatronics (AISM 2015), Guilin, pp 1–5
4. Shengfang L, Xingzhe H (2006) Research on the AGV based robot system used in substation
inspection. In: 2006 international conference on power system technology, Chongqing, pp 1–4
5. Cheong H-W, Lee H (2018) Concept design of AGV (automated guided vehicle) based on
image detection and positioning. Procedia Comput Sci 139:104–107
6. Lua S, Xub C, Zhongc RY, Wangd L (2017) A RFID-enabled positioning system in automated
guided vehicle forsmart factories. J Manuf Syst 44:179–190
7. Chouhan R, Kanwal F, Ali S, Ali N (2014) Design and development of a prototype robotic
gripper. In: 2014 international conference on robotics and emerging allied technologies in
engineering (iCREATE), Islamabad, pp 317–320
8. Bora P, Nandi V (2015) Low cost shadow function based articulated robotic arm. In: 2015 inter-
national conference on energy, power and environment: towards sustainable growth (ICEPE),
Shillong, pp 1–4
9. Fu X, Zhang B, Yu H (2015) Rack design for an autonomous-steerable-vehicle storage and
retrieval system of a warehouse. In: 2015 7th international conference on intelligent human-
machine systems and cybernetics, Hangzhou, pp 112–116
872 P. S. Nair et al.

10. Farah Hanani MK, Zulkhairi MY, Mohamad Zaihirain MR, MohdAswadi A, Ismail A (2016)
Development of automated storage and retrieval system for flexible manufacturing system. J
Eng Technol 4(1):43–50
11. Cheonga H-W, Leeb H (2018) Requirements of AGV (automated guided vehicle) for SMEs
(small and medium-sized enterprises). Procedia Comput Sci 139:91–94
Optimization Techniques
Multi-response Optimization of Inconel
825 Process Parameters Using LN2
Cooled Zinc-Coated Brass Wire in CNC
Wire-Cut EDM
Midthur A. Salman Khan, C. Nandakumar, B. Mohan,
and R. Senthil Kumar

Abstract Inconel 825 is a nickel-based superalloy which is mostly used in most of

the industries where high corrosion resistance and high temperature withstanding
are needed. Some of these industries are Petrochemical, Chemical, Missiles and
Nuclear power plants. Inconel 825 has possessed high anti-oxidation nature even at
high temperatures, i.e. about 650–760 °C. Inconel 825 cannot be machined in the
traditional machines because of its high strength at elevated temperatures, high strain
hardened and low thermal conductivity. WEDM is an alternative machine which
is used for machining of Inconel 825, because WEDM can machine any material
without considering its hardness if the material is electrically conductive. Liquid
Nitrogen cooled (LN2 ) Zinc-coated brass is taken as a cathode and the Inconel 825 is
taken as an anode during the machining process. Taguchi’s L 27 orthogonally array is
used in the experimentation method. Five parameters of WEDM i.e., Wire Tension,
Wire Feed, Pulse on Time, Pulse off Time, Servo Voltage are taken as inputs and
MRR and Surface finish are taken as the output responses. In this paper, an attempt is
made to study the effect of various parameters and optimize the result. Mathematical
modelling of the output responses is developed.

Keywords Nickel-based superalloy (Inconel 825) · Taguchi method · Material

removal rate · Surface roughness · Liquid cooled nitrogen

M. A. Salman Khan (B)

Department of Mechanical Engineering, KG Reddy College of Engineering and Technology,
Hyderabad 501504, India
e-mail: m.a.salman087@gmail.com
C. Nandakumar
Department of Production Technology, MIT Campus, Anna University, Chennai 600044, India
B. Mohan
Department of Mechanical Engineering, CEG Campus, Anna University, Chennai 600025, India
R. Senthil Kumar
Department of Mechanical Engineering, Mohamed Sathak A.J. of Engineering,
Chennai 603103, India

© Springer Nature Singapore Pte Ltd. 2021 875

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_76
876 M. A. Salman Khan et al.

1 Introduction

Wire Electric Discharge Machining (WEDM) is one of the non-conventional

machines, which is used in machining of the materials without considering the
strength of the material. It can be used to machine components with complex shapes,
profile and sharp edges which cannot be machined in non-traditional machining pro-
cess. Hence it has wide application areas like aerospace industries, medical implants,
electronic industry, automobile industry, etc. [1]. Wire electrode is made of Copper,
Brass, Zinc-coated Brass and Tungsten of about 0.05–0.3 mm diameter. Thermoelec-
tric energy is the working principle of WEDM. Series of spark are generated between
the wire and workpiece whose temperature is around 10,000–20,000 °C [2]. Dielec-
tric fluid is continuously passed between the wire electrode and workpiece. The wire
electrode moves close to the workpiece until the applied voltage between the wire
and the workpiece increases and causes the dielectric to get ionized and breakdown
of dielectric happens. The gap between both the electrodes should be in the range of
10–100 µm. Short-circuiting will happen if the gap reduces less this range. Dielec-
tric fluid is also used to flush these tiny particles evaporated. There will no direct
contact between the tool and the workpiece which reduces the mechanical stress and
vibrations problems during the machining [3]. When cryogenic treatment is made
on the material there will be a change in its crystal structure. Surface finish will
be improved when deep cryogenic treatment is carried out on the material, which
further improves the fatigue strength, corrosion and wear resistance of OHNS Die
Steel [4]. It was observed that the cutting temperature of Ti–6Al–4V has reduced
to 36 and 47% when CO2 and LN2 are used as cryogenic in comparison with wet
machining. Cryogenic machining produces better chip control and minimizes the tool
wear more than dry, wet machining [5]. Navjot Singh et al. [6], investigated effect
on AISI D3 Die Steel using two different Cryogenic Treated Wires in Wire Electri-
cal Discharge Machining by designing the experiments in Taguchi’s L 9 orthogonal
array. Better Material removal rate was obtained in cryogenically treated zinc-coated
diffused brass wire than cryogenic treated plain brass wire. Parveen et al. [7], tried to
optimize the WEDM process parameters when H-13 is machined using Taguchi L 9
orthogonal array method. MRR, Wire wear rate and surface flatness lead to improve-
ments in the machining efficiency. Patil et al. [8], have studied surface finish (Ra) of
Ceramic materials reinforced Al matrix composite and developed the semi-empirical
model. The surface roughness of Al/SiCp composite reduces with an increase in vol-
ume fraction of SiC and in Al/Al2 O3 surface roughness decreases with the increase
in volume fraction of aluminum. A. Thakur et al. have investigated the influence of
cutting speed (V c ) (51, 84, and 124 m/min) on various machining characteristics like
chip morphology, chip thickness ratio, tool wear, surface, and sub-surface integrity
during dry turning of Inconel 825 and have concluded that the surface and sub-surface
integrity obtained with coated tool were superior to that while machining Inconel 825
with uncoated tool. Rajyalakshmi and Venkata Ramaiah [9] have applied an effective
approach, Taguchi grey relational analysis, to experimental results of wire-cut elec-
trical discharge machining (WEDM) on Inconel 825 with consideration of multiple
Multi-response Optimization of Inconel 825 Process Parameters … 877

response measures along with the orthogonal array design of experiment, to find
the best combination of parameters. Kumari et al. [10], have investigated the extent
of machinability in purview of process performance surface integrity on of Inconel
825 Superalloy by using different tool material and flushing conditions. Moreover,
effects of peak discharge current on EDM responses were also discussed. Aspinwall
et al. [11], have employed Roughing and finishing (multiple trim cut) strategies on
two high specification machines with pulse generators designed to provide minimum
workpiece integrity damage. Results include productivity; 3D topographic maps of
workpiece surfaces, microstructural and microhardness depth profile data. Average
recast thickness was less than 11 µm were achieved. Yildiz and Nalbant [12] have
reviewed the application of coolant in a cutting process which can increase tool life,
dimensional accuracy and decrease cutting temperatures, surface roughness, amount
of power consumed in a metal cutting process using nitrogen, as a cryogenic coolant.
Molinari et al. [13], have studied the effect of deep cryogenic treatment (−196 °C)
on the properties of some tool steels by means of both field tests on real tools and
laboratory tests and concluded that the deep cryogenic treatment on quenched and
tempered high-speed steel tools increases hardness and reduces tool consumption.
Neeraj Sharma et al. [14], have used cryogenic treated high carbon high chromium
tool steel (D-2 tool steel) and investigated the effect of parameters on surface rough-
ness for WEDM. They employed Mathematical modelling of the process is carried
with the help of Response surface methodology (RSM) and the central composite
rotatable design (CCRD) for planning the experiments.

2 Research Methodology

A Computer Numerically Controlled (CNC) WEDM is used for the experimentation

whose worktable can withstand a maximum load of 300 Kgs and can accommodate
a workpiece of 200 mm height. The worktable can move in X, Y and Z directions
whereas the wire upper guide can move in U and V directions. A standard wire of
0.25 mm can be passed through upper and lower guides. Deionized water is passed
between the wire and workpiece, which behaves as a dielectric fluid between the wire
electrode and workpiece electrode. In this research to make the Zinc-Coated Brass
wire is used as wire electrode which is cryogenically cooled using Liquid Nitrogen
(LN2 ). Stainless steel box in which a hollow copper tube is at the center through
with the Zinc-coated brass wire is passed. This box is insulated from inside using
thermocol to avoid the dissipation of heat from the box. The cryogenic liquid which
is stored in a separated cylinder is pumped into the stainless steel box continuously
during the machining. 3D model of the cryogenic setup is shown in Fig. 1.
Inconel 825 of 250 × 250 × 5 mm is used as workpiece. Experiments were
designed using Taguchi’s L 27 orthogonally array. The process parameters for machin-
ing are Wire Feed (WF), Wire tension (WT), Pulse on Time (T on ), Pulse off
Time (T off ) and Servo Voltage which are taken at three different levels. The input
parameters and its levels are shown in Table 1.
878 M. A. Salman Khan et al.

Fig. 1 3D modelling of cryogenic setup

Table 1 Machine process parameters and its levels

Symbol Parameter Unit Coding levels
−1 0 1
A Wire tension g 850 1150 1300
B Wire feed mm/min 7 8 9
C Pulse on µs 7 8 9
D Pulse off µs 6 7 8
E Servo voltage V 4 5 6

Gear-shaped machining profile is selected for machining. Using Taguchi 27 exper-

iments were conducted with five parameters and its three different levels. Taguchi
method has reduced the number of test using orthogonal array and effects of the actors
that cannot be controlled can be minimized. It even provides an efficient and sys-
tematic approach to obtain the optimum parameters. Material Removal Rate (MRR)
and Surface Roughness are the responses recorded for 27 experiments. The material
Removal rate is calculated using the formula given below.

MRR = . . . (1)

Volume of Material removed (V m ) = Perimeter × Thickness of workpiece (T m

= 5 mm) × Thickness of Wire (T w = 0.3 mm).

2.1 Calculation of Perimeter

For calculating the perimeter of the Spur gear few assumptions are made to make
the calculation simpler and the Fig. 2 shows the gear profile and Fig. 3 shows
the Dedendum triangle reference, Figs. 4 and 5 gives details about the Dedendum
Multi-response Optimization of Inconel 825 Process Parameters … 879

Fig. 2 Spur gear profile

Fig. 3 Dedendum triangle

reference
880 M. A. Salman Khan et al.

Fig. 4 Gear tooth

dimensions

Fig. 5 Dedendum triangle

triangle.

Pp = Pd + TTSL + TTWA = 132.89 mm.

where Pd = Perimeter Covered at Dedendum

TTSL = Total Tooth Side Length.
TTWA = Tooth Width at Addendum.
where θ = Angle of dedendum width (∠AOB).
O = Centre of Gear.

Pd = Cd −(L twd × N )

Pd =22.21 mm.
where C d = Dedendum Circumference (π × 20.1).
L twd = Arc length from TWD length.
N = No. of tooth.

TTSL = STSL × N × S

TTSL = 91.4832 mm.

where STSL = Single Tooth Side Length (mm).
S = No. of Sides in a Tooth (2).
√ 2
STSL = (h + TSBL2 )

STSL = 3.8118 mm
where h = Gear Tooth Height (3.685 mm)
TSBL = Tooth Side Bottom Length (0.97 mm).

L twd = π d

L twd = 3.411636 mm
Multi-response Optimization of Inconel 825 Process Parameters … 881

where d = Dedendum diameter (20.1 mm)

θ = Angle of arc by TWD
θ = tan − 1 = 19.45°.
Surface roughness of the specimen is calculated using TaylSurf Surface roughness
machine. The surface roughness is calculated at three sides of the specimen and the
average of these values is taken as consideration. Responses for the 27 experiments
are tabulated in Table 2. Machined specimens are shown in Fig. 6.

Table 2 Responses obtained for the experiments

S. No. Wire feed Wire Pulse Pulse Servo Material Surface
WF tension on time on time voltage removal roughness
(mm/min) WT (g) T on T off S v (V) rate MRR (µm)
(µs) (µs) (mm3 /min)
1 7 850 7 6 4 3.24 2.65
2 7 850 7 6 5 3.35 2.80
3 7 850 7 6 6 3.45 2.98
4 7 1150 8 7 4 3.35 2.60
5 7 1150 8 7 5 3.56 2.70
6 7 1150 8 7 6 3.70 2.86
7 7 1300 9 8 4 3.95 2.50
8 7 1300 9 8 5 4.05 2.65
9 7 1300 9 8 6 4.35 2.80
10 8 850 8 8 4 3.25 2.81
11 8 850 8 8 5 3.45 2.95
12 8 850 8 8 6 3.67 3.05
13 8 1150 9 6 4 3.56 2.68
14 8 1150 9 6 5 3.79 2.84
15 8 1150 9 6 6 4.05 3.01
16 8 1300 7 7 4 3.85 2.15
17 8 1300 7 7 5 4.05 2.29
18 8 1300 7 7 6 4.21 2.41
19 9 850 9 7 4 3.65 2.85
20 9 850 9 7 5 3.79 2.98
21 9 850 9 7 6 3.97 3.15
22 9 1150 7 8 4 3.59 2.31
23 9 1150 7 8 5 3.90 2.43
24 9 1150 7 8 6 4.10 2.65
25 9 1300 8 6 4 3.89 2.26
26 9 1300 8 6 5 3.95 2.39
27 9 1300 8 6 6 4.25 2.50
882 M. A. Salman Khan et al.

Fig. 6 Samples specimen

after machining

3 Evaluation of Experiments and Its Analysis

The effect of input parameters on the responses is studied by analyzing Taguchi design
in Minitab software. The optimum solutions of the controlled factors were provided
by Signal to noise (S/N) ratios. The highest value of MRR and lowest value of Surface
Roughness are required. The S/N ratios graphs for MRR and Surface Roughness are
shown in Figs. 7 and 8 which show the optimal machining parameters for obtaining
smaller surface roughness and large MRR.
The parameters best level is one which is having the highest value in S/N response
table. The optimal value of MRR is obtained at Wire feed 9 mm/min, Wire tension
1300 g, Pulse on time 9 µs, Pulse off time 7 µs and Servo voltage 6 V. Similarly for
best surface roughness is obtained Wire feed 9 mm/min, Wire tension 1300 g, Pulse
on time 7 µs, Pulse off time 7 µs and Servo voltage 4 V.
Modelling of variables is the relation between input parameters and the responses
are obtained using regression equations. The linear regression equation of the Mate-
rial Removal Rate and Surface Roughness are obtained in Eqs. 2 and 3 which gives the
linear relation between input parameters with its responses. The response obtained
by Eqs. 2 and 3 will give the liner variation of response with inputs. The R2 value
for the linear regression equation of MRR is found to be 98.9% and that of Surface
roughness is found to be 99.2%

MRR = 0.1854 WF + 0.001009 WT + 0.0198 Ton + 0.0029 Toff + 0.1963 Sv (2)

Ra = 0.0238 WF − 0.000806 WT + 0.2355 Ton + 0.0726 Toff + 0.1947 Sv (3)

The Quadratic regression equation of MRR and Ra is obtained by using Eqs. 4

and 5 which will help to obtain the response for any value of the input instead of
predefined values. Its predicted R2 values are 96.9 and 98.9%.
Multi-response Optimization of Inconel 825 Process Parameters … 883

Fig. 7 Signal to noise ratio for MRR

Fig. 8 Signal to noise ratio for surface roughness

884 M. A. Salman Khan et al.

MRR = 0.805 WF + 0.00728 WT − 1.157 Ton + 0.199 Toff − 0.237 Sv

− 0.001050(WF × WT) + 0.0909(WF × Ton ) − 0.0433(WF × Toff )
+ 0.0250(WF × Sv ) + 0.000236(WT × Ton )
+ 0.000067(WT × Sv ) + 0.0042(Ton × Sv ) + 0.0167(Toff × Sv ) (4)

Ra = −0.276 WF − 0.001193 WT + 1.385 Ton − 0.515 Toff + 0.182 Sv

+ 0.000408(WF × WT) − 0.0902(WF × Ton ) + 0.0639(WF × Toff )
− 0.00083(WF × Sv ) − 0.000387(WT × Ton )
− 0.000017(WT × Sv ) − 0.00167(Toff × Sv ) (5)

3.1 Variations in Material Removal Rate

The Changes which are obtained because of the input responses in the Material
removal rate are studied.

3.1.1 Surface Plot of Wire Tension and Wire Feed

The effect of Wire tension and Wire feed on MRR is shown as a surface plot in
Fig. 9. It is clearly observed from surface plots that there is an increase in MRR
with an increase in Wire tension and Wire feed. As wire tension of LN2 cooled Zinc-
coated brass wire increases, there is an increase in the stiffness of the wire and wire
becomes hardened due to cryogenic liquid which causes removal of more material.
When Wire feed increases the wire moves faster which leads to less time for the
breakdown between the wire and material hence material removal rate decreases.

Fig. 9 Surface plot for

MRR versus WF, WT
Multi-response Optimization of Inconel 825 Process Parameters … 885

Fig. 10 Surface plot for

MRR versus WF, T on

3.1.2 Surface Plot of Wire Feed and Pulse On Time Versus MRR

The effect of Wire feed and Pulse on time are shown on surface plot in Fig. 10, in
which it is observed that when pulse on time through the cryogenic cooled zinc-
coated brass wire was increased, the duration of time increases that leads to evenly
distribution of energy over the wire and removes more amount of material from the
surface. This leads to increase in the amount of energy through the wire. Hence, the
MRR increases.

3.1.3 Surface Plot of Wire Feed and Pulse Off Time Versus MRR

Surface Plot of MRR against Wire feed and Pulse off time is shown in Fig. 11.
The figure clearly depicts that the wire feed is prominent than pulse off time. With

Fig. 11 Surface plot for

MRR versus WF, T off
886 M. A. Salman Khan et al.

Fig. 12 Surface plot for

MRR versus WF, S v

increase in Pulse off time, there is an increase in duration for cooling the wire and
the energy passed through the wire will be reduced. This leads to a decrease in MRR.

3.1.4 Surface Plot of Wire Feed and Servo Voltage Versus MRR

The variation in MRR with Wire Feed and Serve voltage is shown in Fig. 12. If Servo
voltage is increased the dielectric will breakdown easily which in turn causes large
energy between the electrodes, i.e. wire and the material, thus MRR increases.

3.1.5 Surface Plot of Wire Tension and Pulse On Time Versus MRR

The variations in MRR with the Wire tension and Pulse on time are shown on surface
plot in Fig. 13. It is observed that MRR increases with the increase of pulse on time
and Wire tension. Highest MRR is observed when Pulse on time is low and Wire

Fig. 13 Surface plot for

MRR versus WT, T on
Multi-response Optimization of Inconel 825 Process Parameters … 887

Fig. 14 Surface plot for

MRR versus WT, T off

tension is high which indicates that wire tension is more significant compare to Pulse
on time.

3.1.6 Surface Plot of Wire Tension and Pulse Off Time Versus MRR

The variation in MRR with Wire tension and Pulse off time is shown in Fig. 14, which
clearly depicts that the MRR increases with increase in Wire tension and decreases
with an increase in Pulse off time. As wire tension increases the vibrations of the wire
between the upper and lower guide decreases which causes the energy to distribute
evenly, thus MRR increases. When pulse off time increases more delay will be there
in the production of breakdown hence MRR decreases.

3.1.7 Surface Plot of Wire Tension and Servo Voltage Versus MRR

The variation in MRR with wire tension and Servo voltage is shown on surface plot
Fig. 15. It is observed that MRR increases with increase in both wire tension and
Servo voltage. The MRR is highest in Fig. when both servo voltage and wire tension
are highest.

3.1.8 Surface Plot of Pulse On Time and Pulse Off Time Versus MRR

The variation in MRR with Pulse on time and pulse off time is shown in Fig. 16
which clearly depicts that the MRR is high when Pulse on time is Medium and Pulse
off time is less. When Pulse on time increases beyond a certain limit the wire will
break because of short circuit. When Pulse off time is less, the sparks are generated
within short duration of time, and hence MRR is more when Pulse off time is less.
888 M. A. Salman Khan et al.

Fig. 15 Surface plot for

MRR versus WT, S v

Fig. 16 Surface plot for

MRR versus T on , T off

3.1.9 Surface Plot of Pulse On Time and Servo Voltage Versus MRR

MRR variations with Pulse on time and Pulse off time are shown in surface plot
Fig. 17. The figure clears depicts that the MRR increases with an increase in both
Pulse on time and servo voltage. The variation in MRR with Pulse on time and Servo
voltage is linear. Highest MRR is observed when both Pulse on time and Servo
Voltage are high.

3.1.10 Surface Plot of Pulse Off Time and Servo Voltage Versus MRR

The Surface plot of Pulse off time and Servo voltage on MRR is shown in Fig. 18.
The variation in figure clearly depicts that the highest MRR is obtained when Pulse
Multi-response Optimization of Inconel 825 Process Parameters … 889

Fig. 17 Surface plot for

MRR versus WT, S v

Fig. 18 Surface plot for

MRR versus T on , T off

off time is less and Servo voltage is more because the delay in time for the generation
of sparks will be less and breakdown occurs faster.

3.2 Variations in Surface Roughness

The effect of various input parameters on surface roughness is observed by using

surface plots.
890 M. A. Salman Khan et al.

Fig. 19 Surface plot for Ra

versus WF, WT

3.2.1 Surface Plot for Ra Against Wire Tension and Wire Feed

The effect of wire tension and Wire feed on surface roughness is shown on surface
plot in Fig. 19. It is observed from the figure that Surface roughness is least when both
wire tension and wire feed are high. When wire tension of cryogenic cooled brass
wire is increased the stiffness of wire increases and vibrations are reduced which
leads to even spark generation from the wire and material is removed evenly from
the surface. When this Stiff wire is fed continuous, i.e. wire feed increases the newer
wire comes in the spark generation are which further helps to remove the material
evenly. Hence, Surface roughness decreases when both Wire tension and Wire feed
are high.

3.2.2 Surface Plot for Ra with Wire Feed and Pulse On Time

The variation in surface roughness against Wire feed and Pulse on time is shown
in Fig. 20. The figure depicts that Surface roughness increases with increase of
both Wire feed and Pulse on time. As Pulse on time increases, the duration for which
energy distortion between wire and workpiece increases which leads to more amount
of material removal from the same surface hence surface roughness increases. When
Wire Feed is increased the break down between newer wire and work piece takes
place continuously, which leads to more amount of energy gets distorted at once.
Thus there is an increase in surface roughness.

3.2.3 Surface Plot for Ra with Wire Tension and Pulse On Time

A surface variation for Ra with Wire tension and Pulse on time is shown in Fig. 21.
It is observed that Surface roughness increases with increase in pulse on time and
decreases with increase in wire tension. When Pulse on time increases, there is
Multi-response Optimization of Inconel 825 Process Parameters … 891

Fig. 20 Surface plot for Ra

versus WF, T on

Fig. 21 Surface plot of Ra

versus WT, T on

increase in time of spark generation which causes more amount of material getting
eroded from the surface of the work piece. Thus Surface roughness increases with
increase in Pulse on time. When Wire tension is increased there is reduction in
vibrations of the wire which further leads to even distribution of energy between
work piece and the wire, hence surface roughness decreases. It is clearly observed
from the figure the surface roughness is least when Pulse on time is less and wires
tension is high.

3.2.4 Surface Plot for Ra with Wire Tension and Pulse Off Time

The changes in Surface plot for surface roughness with Wire tension and Pulse
off time are shown in Fig. 22. It is observed from the figure that Surface roughness
decreases with increase both Pulse off time and Wire tension, the least surface rough-
ness is obtained when both wire tension and pulse off time are high. When Pulse off
892 M. A. Salman Khan et al.

Fig. 22 Surface plot of Ra

versus WT, T off

time is high the duration for wire to get cool increases and increase in the delay in
formation of spark. Thus Surface roughness will decrease.

3.2.5 Surface Plot for Ra with Wire Tension and Servo Voltage

The effects of Wire tension and Servo voltage on Surface Roughness are shown
in Fig. 23. The surface roughness increases with an increase in Servo voltage and
decreases with increase in Wire tension. As Servo voltage increases the amount of
voltage between the wire and work piece increases the breakdown increases which
leads to an increase in energy transfer between the electrodes, which leads to increase
in surface roughness. The least value of surface roughness is observed in the figure
when wire tension is high and servo voltage is less.

Fig. 23 Surface plot of Ra

versus WT, S v
Multi-response Optimization of Inconel 825 Process Parameters … 893

Fig. 24 Surface plot of Ra

versus T on , T off

3.2.6 Surface Plot of Ra with Pulse On Time and Pulse Off Time

The variations in surface plot of Ra with Pulse on time and Pulse off time is shown
in Fig. 24. The figure depicts that surface roughness decreases with increase in Pulse
off time and increases with increase in Pulse on time. The least surface roughness in
the figure is obtained when Pulse off time is high and Pulse on time is low.

3.2.7 Surface Plot of Ra with Pulse On Time and Servo Voltage

The Surface plot of surface roughness with pulse on time and servo voltage is shown
in Fig. 25. The figures clearly indicate the surface roughness increases with increase

Fig. 25 Surface plot of Ra

versus T on , S v
894 M. A. Salman Khan et al.

Fig. 26 Surface plot of Ra

versus S v , T off

in both servo voltage and pulse on time. The highest value of the surface plot is
obtained when both servo voltage and pulse on time are high.

3.2.8 Surface Plot of Ra with Pulse Off Time and Servo Voltage

The variation in the surface roughness with Pulse off time and Servo voltage is
shown in Fig. 26. When Servo voltage increases there will be an increase in Surface
roughness and when Pulse off time increases the surface roughness will be decreased.
The least value in surface plot is obtained when servo voltage is less and Pulse off
time is high.

4 Surface Analysis

To study the surface Scanning Electron Microscopic images are taken for the highest
MRR and the lowest Surface roughness experimental specimens.
From Fig. 27 it is observed that for high MRR the surface has damaged more.
There are few cracks, debris and globules are observed. The debris is formed when
molten material of the work piece is not flushed properly and gets re-solidifies on the
surface of the workpiece. In Fig. 27, it is observed that more debris is formed because
the molten material during machining gets re-solidifies quickly due to decreases in the
temperature of dielectric as it is in contact with cryogenic cooled wire. Globules are
formed when the material which gets evaporated from the surface of the workpiece
falls back on to the surface during flushing. Due to the difference in the impact of
pulse discharge energy the craters are formed on the work piece.
Multi-response Optimization of Inconel 825 Process Parameters … 895

Fig. 27 SEM image for high MRR specimen

Fig. 28 SEM image for high MRR specimen

From Fig. 28, it is observed that better surface is observed. There are less amount
globules and craters are observed. Surface roughness is improved as the as less energy
is discharged during this experiment which causes less material to discharge and less
molten material gets accumulated and fewer globules are formed.

5 Conclusions

• MRR of Inconel 825 varies from 3.24 to 4.35 mm3 /min.

• Surface Roughness varies from 2.15 to 3.01 µm.
• Highest MRR of 4.35 mm3 /min is obtained when the input parameters are Wire
Feed (WF)–7 mm/min, Wire Tension (WT)–1300 g, Pulse on time (T on )–9 µs,
Pulse off time (T off )–8 µs and Servo Voltage (S v )–6 v.
896 M. A. Salman Khan et al.

• Least Surface Roughness of 2.15 µm is observed when input parameters are Wire
Feed (WF)–8 mm/min, Wire Tension (WT)–1300 g, Pulse on time (T on )–7 µs,
Pulse off time (T off )–7 µs and Servo Voltage (S v )–4 v.
• Wire Tension, Pulse on time and Servo voltage are the most significant factors for
MRR.
• Wire tension, Pulse off time and Wire Feed are the most significant factors for
surface roughness.
• The Mathematical modelling for both MRR and Surface Roughness is found to
be in good agreement with the experimental values.

References

1. Singh T, Misra JP, Singh B (2017) Empirical modelling of wear ratio during WEDM of nimonic
263. Mater Today 5(11):23612–23618
2. Majumder H, Maity K (2018) Prediction and optimization of surface roughness and micro-
hardness using grnn and MOORA-fuzzy-a MCDM approach for nitinol in WEDM. Measure-
ment 118(1):1–13
3. Ajay BT, Mayur Vitthal P, Rajyalakshmi G (2017) WEDM machining on aerospace Materials
for improving material properties. Mater Today: Proc 4(8):9107–9116
4. Singh A, Kumar S (2015) Surface roughness and microhardness evaluation for EDM with
Cu–Mn powder metallurgy tool. Mater Manuf Process 31(4):514–521
5. Dilip Jerold B, Pradeep Kumar M (2013) The Influence of cryogenic coolants in machining of
Ti–6Al–4V. J Manuf Sci Eng 135(3):1–8
6. Singh N, Kumar P, Goyal K (2013) Effect of two different cryogenic treated wires in wire
electrical discharge machining of AISI D3 DIE steel. J Mech Eng 43(2):54–60
7. Kumar P, Bhushan B, Gupta RK (2013) Response of wire electrical discharge machining for
H13 using Taguchi L 9 array. Int J Eng Res Appl 3(5):907–912
8. Patil NG, Brahmankar PK (2009) Some studies into wire electro-discharge machining
of alumina particulate-reinforced aluminum matrix composites. Int J Adv Manuf Technol
48(5–8):537–555
9. Rajyalakshmi G, Venkata Ramaiah P (2013) Multiple process parameter optimization of wire
electrical discharge machining on Inconel 825 using Taguchi grey relational analysis. Int J Adv
Manuf Technol 69(5–8):1249–1262
10. Kumari S, Datta S, Masanta M, Nandi G, Pal PK (2018) Electro-discharge machining of Inconel
825 super alloy: effects of tool material and dielectric flushing. Silicon 10(5):2079–2099
11. Aspinwall DK, Soo SL, Berrisford AE, Walder G (2008) Workpiece surface roughness and
integrity after WEDM of Ti–6Al–4V and Inconel 718 using minimum damage generator
technology. CIRP Ann—Manuf Technol 57(1):187–190
12. Yildiz Y, Nalbant M (2008) A review of cryogenic cooling in machining processes. Int J Mach
Tools Manuf 48(9):947–964
13. Molinari A, Pellizzari M, Gialanella S, Straffelini G, Stiasny KH (2001) Effect of deep
cryogenic treatment on the mechanical properties of tool steels. J Mater Process Technol
118(1–3):350–355
14. Sharma N, Singh A, Sharma R, Deepak (2014) Modelling the WEDM process parameters for
cryogenic treated D-2 tool steel by integrated RSM and GA. Procedia Eng 97(1):1609–1617
Optimization of Process Parameters
During EDM on Inconel Alloy 625

K. Jayakumar

Abstract Inconel 625 is a Nickel–Chromium–Molybdenum based superalloy which

is extensively used in numerous engineering applications like, aerospace, nuclear and
marine fields, aeronautical, chemical, petrochemical industries, oil and gas extrac-
tion industries. It has high strength and outstanding corrosion resistance properties.
The main problem with using this material is their low machinability while machin-
ing with conventional machining processes due to its high mechanical properties.
From different unconventional machining processes (UCMP), the Electrical dis-
charge machining (EDM) process is the most important and versatile UCMP and is
suitable for machining any difficult to machine materials and it is most appropriate
for the manufacturing of dies, molds, and press tools. Normally machining industries
targets the achievement of high-quality products, relating to the surface finish, work-
piece dimensional accuracy, high material removal rate (MRR) at low cost. Hence,
the present research work deals with the optimization of process parameters in EDM
of Inconel 625 alloy for minimizing the surface roughness (Ra) and maximizing
MRR. EDM studies were conducted as per Taguchi’s L 9 orthogonal array by vary-
ing process parameters such as current, voltage, pulse on time (T on ) and electrode
bottom profile. Surface roughness and MRR values were measured. Effects of pro-
cess parameters on the selected machinability responses were analyzed. Optimum
process parameters were found for the optimization of EDM on Inconel alloy 625.

Keywords Inconel alloy 625 · EDM studies · Material removal rate · Surface
roughness

K. Jayakumar (B)
Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Kalavakkam, Chennai 603110, India
e-mail: kjayakumar@ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 897

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_77
898 K. Jayakumar

1 Introduction

Inconel 625 is a nickel-based superalloy that has high strength and resistance to ele-
vated temperatures, corrosion and oxidation. Its capability to survive high stress and
a wide range of temperatures, both in and out of the water, makes it a fitting best for
nuclear and marine uses. Originally Inconel 625 was developed in the 1960s for the
use of making material for steam-line piping. Later, some modifications were made
to be for more creep-resistant and weldable. Because of this, the uses of Inconel
625 has expanded into the chemical processing industry, marine, and nuclear appli-
cations. However, the same properties pose a great challenge during their machin-
ing by conventional machining [1]. Non-conventional machining methods with no
direct contact between tool and workpiece would be a better option to machine such
materials. Electric discharge machining (EDM) is one of the most frequently used
non-conventional machining techniques to machine almost any electrically conduc-
tive material into complex and intrinsic shapes. EDM is an electrothermal process
where material removal takes place through the controlled spark erosion. EDM is the
most widely used machining route for Inconel superalloys because of its less setup
cost and excellent accuracy of the machining performance in terms of quality of the
end product. EDM drilling is most widely performed to machine superalloys in the
aerospace industry, gas turbine, and rocket engine [2].
Due to the technological development, researchers have also worked on EDM
of Inconel alloys (Inconel 625). Dhanabalan et al. [3] conducted EDM studies on
Inconel 625 to analyze the effect of input parameters namely peak current, pulse on
time (T on ), and pulse off time (T off ) on the form tolerances and MRR with circular,
square and hexagonal shaped electrodes.
Torres et al. [4] analyzed the effect of EDM parameters on surface roughness
(Ra), MRR and electrode wear (EW) of an Inconel 600 alloy by varying current,
duty cycle, pulse duration and polarity. They found that +ve positive polarity leads
to higher MRR whereas −ve polarity leads to lower Ra values. The same authors [5]
continued the machining of Inconel 600 alloy using copper infiltrated graphite (C–
Cu). They suggested that the use of negative polarity leads to a higher MRR, higher
EW and higher Ra value. Moreover, they found the optimal condition to obtain a
maximum MRR (Material Removal Rate) of 30.49 mm3 /min was: 8 A, 100 µs and
0.6, respectively, for the current intensity, pulse time and duty cycle. Dikshit et al. [6]
recently used EDM for machining rectangular shape in Inconel 625. Experiments
were conducted by varying peak current, pulse on time and pulse off time as per
central composite design. They found optimum process parameter conditions for
maximizing MRR and minimizing Ra value.
From literature, it was found that the effect of EDM process parameter on different
machinability of Inconel alloys was studied but the effect of electrode bottom profile
on the machinability of Inconel 625 is missing and it needs a detailed study. The
present work focused on the effect of EDM process parameter and electrode bottom
profile on Inconel 625 in terms of MRR and surface roughness.
Optimization of Process Parameters During EDM … 899

2 Experimental Details

Inconel 625 of size 60 × 60 × 6 mm was chosen as workpiece material. Major

constituents in weight % of Inconel 625 are Ni (55%), Cr (23%), Mo (10%) and Fe
(5%). A cylindrical Copper (Cu) electrode of diameter 6 mm was used as electrode
material. Cu can machine all hard materials with low tool wear and it is affordable.
EDM studies were conducted in the Die-Sinking EDM setup (Make: Ratnaparkhi
Electronics EDM 500X300 ZNC; Country: India). Commercial EDM oil was cir-
culated as a dielectric medium. The workpiece is connected to positive polarity and
the electrode is connected to negative polarity during machining. Experiments were
conducted using four major controllable process parameters and individual param-
eters were varied at three levels. The selected process parameters were current (I),
voltage (V), pulse on time (µs) and electrode bottom profile (flat, chamfer and fillet).
Pulse off time was maintained as a constant of 20 µs for all experiments due to the
machine constraint. Fillet profile was created in a lathe machine with 1 mm front
radius and the chamfer profile was machined with 1 mm × 45° dimension.
Taguchi L 9 orthogonal array (OA) design was used to conduct experiments and
Table 1 shows the experimental conditions and results. The EDM hole depth was
kept as constant (5.9 mm) and machining time was noted for each experiment to
calculate MRR. Figure 1 shows the close view of the EDM setup with workpiece
and electrode. Figure 2 shows the machined workpiece.
The machining responses of MRR and average surface roughness (Ra) at the
machined surface were measured for each experimental condition. MRR was calcu-
lated based on weight difference between the workpiece before and after machining
and time taken for each machining. Surface roughness (Ra) of the drilled holes sur-
faces was measured using Mitutoyo SJ210 series surface roughness tester. Cut off
length and number of sample was taken as 0.8 mm and 4, respectively. Roughness
measurements were made at three different areas of each hole surface by keeping the

Table 1 EDM experimental design and experimental results

Sl. No. Current (A) Voltage (V) Ton (µs) Electrode MRR Surface
(g/min) roughness
(µm)
1 15 50 20 Flat 0.092 5.471
2 15 65 30 Chamfer 0.146 7.617
3 15 80 40 Fillet 0.164 9.132
4 20 50 30 Fillet 0.167 9.787
5 20 65 40 Flat 0.178 9.194
6 20 80 20 Chamfer 0.129 6.958
7 25 50 40 Chamfer 0.159 11.910
8 25 65 20 Fillet 0.154 6.779
9 25 80 30 Flat 0.210 11.318
900 K. Jayakumar

Fig. 1 EDM setup for

machining Inconel 625

Fig. 2 Workpiece after

EDM

workpiece in a different orientation and the average of three readings was considered.
Figure 3 shows the arrangement for surface roughness measurement and results of
Ra is listed in Table 1.
Optimization of Process Parameters During EDM … 901

Fig. 3 Arrangement for measurement of surface roughness

3 Results and Discussions

3.1 Effect of EDM Parameter on MRR

Figure 4 shows the main effect plots of MRR. With an increase of current, voltage
and Ton time, the MRR is increased. As current and voltage increased, the power and

Main Effects Plot for MRR (Gram/min)

Current (A) Voltage (V)
0.20
0.18
0.16
0.14
MRR (gram/min)

0.12
0.10
15 20 25 50 65 80
Ton (µs) Electrode
0.20
0.18
0.16
0.14
0.12
0.10
20 30 40 Chamfer Fillet Flat

Fig. 4 Main effect plots for MRR

902 K. Jayakumar

number of sparks produced were increased this improved the MRR. When the Ton
increased, MRR increased at the start but later decreased with further raise in Ton.
It is the result of an additional reduction on the gap between the work and electrode
developing an electrical short circuit instead of spark which reduced the MRR.
MRR obtained from the study for an electrode with chamfer and the flat electrode is
low compared with the fillet profile of the electrode. The reason is that the fillet profile
has more front curvature surface area than the other two profiles which increased
the spark discharges per unit area of the electrode and hence improved the MRR.
Also, for the same EDM process parameters, the density of the electric discharges is
high; faster separation of the metal from the workpiece and effective flushing of the
molten metal happened with fillet profile than chamfer and flat profile [7].

3.2 Effect of EDM Parameter on Surface Roughness

The effect of process parameters on surface roughness is shown in Fig. 5.

The effect of process parameters on surface roughness is shown above plots. From
that, it was observed that the electrode bottom profile has a little effect than the other
three parameters. As power (current and voltage) and Ton of the EDM electric pulse
increased, sparks become more energetic, causes more melting and affect the surface

Main Effects Plot for Surface roughness (µm)

Current (A) Voltage (V)

11
10
9
Surface roughness (µm)

8
7
6

15 20 25 50 65 80
Ton (µs) Electrode

11
10
9
8
7
6

20 30 40 Chamfer Fillet Flat

Fig. 5 Main effect plots for surface roughness

Optimization of Process Parameters During EDM … 903

over a longer period, leading to larger and deeper craters which increased the surface
roughness value [8, 9].
Amount of spark discharged from the electrode bottom is decreased from chamfer
to fillet and flat profile, i.e. amount of spark discharged from fillet bottom profile is
more than chamfer and flat profile, which created the melting boundary as deeper and
wider cavities and increased the roughness from 5.471 to maximum of 11.910 µm.
From the experimental runs, experiment number 8 with fillet bottom profile can be
considered as the optimum condition for better MRR and lower surface roughness.
Design Expert software’s optimization module was used for finding optimum
parameter combination. From that, it was found that the process parameter combina-
tion of current-15 A, voltage-55 V, Ton 40 µs with flat bottom profile electrode was
found as the overall optimum parameter for getting maximum MRR of 0.21 g/min
and minimum surface roughness value of 6.34 µm with the highest desirability of
0.93.

4 Conclusions

In this present study, the influence of EDM parameters on Inconel alloy 625 has been
carried out with a copper electrode using three different bottom profiles. The results
showed that, for MRR and surface roughness, Ton and current were identified as the
most influencing factor.
For getting low surface roughness value, an electrode with a flat profile is recom-
mended than the other two profiles and for high MRR; electrode with fillet profile is
suggested.
Application of Design Expert software has determined the optimal setting of
process parameters for concurrently attaining multi-response features (high MRR
and low Ra value) in relation to the overall EDM performance.

References

1. Gangadharudu Talla S, Gangopadhyay S, Biswas CK (2017) Influence of graphite powder mixed

EDM on the surface integrity characteristics of Inconel 625. Particulate Sci Technol 35(2):219–
226
2. Mishra DK, Rahul, Datta S, Mahapatra SS, Masanta M, Mahapatra SS (2019) Through hole
making by electro-discharge machining on Inconel 625 super alloy using hollow copper tool
electrode. Proc Inst Mech Eng, Part B: J Eng Manuf 233(2):348–370
3. Dhanabalan S, Sivakumar K, Sathiya Narayanan C (2014) Analysis of form tolerances
in electrical discharge machining process for Inconel 718 and 625. Mater Manuf Process
29(3):253–259
4. Torres C, Luis J, Puertas I (2015) Analysis of the influence of EDM parameters on surface finish,
material removal rate, and electrode wear of an INCONEL 600 alloy. Int J Adv Manuf Technol
80(1–4):123–140
904 K. Jayakumar

5. Salcedo AT, Arbizu IP, Pérez CJL (2017) Analytical modelling of energy density and
optimization of the EDM machining parameters of Inconel 600. Metals 166(7):1–21
6. Dikshit MK, Anand J, Narayan D, Jindal S (2019) Machining characteristics and optimization
of process parameters in die-sinking EDM of Inconel 625. J Braz Soc Mech Sci Eng 302:1–14
7. Manohar M, Selvaraj T, Sivakumar D, Gopinath S, George KM (2014) Experimental study
to assess the effect of Electrode bottom profiles while machining Inconel 718 through EDM
process. Procedia Mater Sci 6:92–104
8. S Ahmad, M A Lajis (2013) Electrical discharge machining (EDM) of Inconel 718 by using
copper electrode at higher peak current and pulse duration. IOP Conf Ser: Mater Sci Eng 50:1–7
9. Goyal A, Pandey A, Sharma P (2017) Machinability of Inconel 625 aerospace material using
cryogenically treated WEDM. Solid State Phenom 266:38–42
Optimization of Laser Trepanning
Parameters for Mild Steel by Taguchi
Response Surface Methodology (T-RSM)

A. Gnanavelbabu , V. Arunachalam , K. T. Sunu Surendran ,

V. Dharaniya , and K. Rajkumar

Abstract Laser trepanning plays a vital role in the machining of holes as it needs
to eliminate material only from a thin annular region, not the entire circle. The key
benefits of trepanning are material saving and less energy consumption. In this study,
responses from laser trepanning on mild steel were taken for optimization. Responses
like hole-top circularity (HTC), hole-bottom circularity (HBC) and hole taper (HT)
were recorded according to the variations made in the parameters such as laser power,
gas pressure, speed of spot movement and pulsing frequency as per Taguchi’s L 27
orthogonal array. To optimize the operating parameters, Taguchi–response surface
methodology (T-RSM) was applied. Three-dimensional response surface graphs and
ramp function graph for desirability were plotted for the obtained results. Optimal
level of parameters obtained was laser power-2716.83 W, gas pressure-7 bar, speed of
spot movement-1200 mm/s, pulsing frequency-2000 Hz and trepanning hole-20 mm.

Keywords Laser trepanning · Mild steel · Optimization · Response surface

methodology (RSM)

1 Introduction

Laser-induced material removal is getting wide attention for precision hole drilling
on difficult to cut superior alloys [1]. The laser drilling is categorized into two types:
trepan drilling and percussion drilling. Laser trepan drilling is more popular than
percussion drilling, and it involves trimming around the boundary of the hole to be
generated [2]. Laser trepanning provides better dimensional accuracy [3]. The major

A. Gnanavelbabu (B) · V. Arunachalam · K. T. Sunu Surendran · V. Dharaniya

Department of Industrial Engineering, CEG, Anna University, Chennai 600025, India
e-mail: agbabu@annauniv.edu
K. Rajkumar
Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Chennai, Tamil Nadu 603110, India

© Springer Nature Singapore Pte Ltd. 2021 905

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_78
906 A. Gnanavelbabu et al.

advantage of laser trepan drilling is its less energy consumption for producing hole
with least material wastage [4].
The statistical relationship between the factors and responses in an experiment can
be determined using various optimization tools. The Taguchi DOE is a reliable and
effective method for evaluating several design parameters by reducing the number
of experiments [5]. Avanish Kumar Dubey et al. made robust design using Taguchi
Orthogonal array and performed multi-objective optimization of laser beam cutting
parameters on aluminium alloy sheet using ANOVA [6]. Response surface methodol-
ogy (RSM) is an emerging optimization tool to characterize the process performance.
Murthy B. R. N. et al. integrated Taguchi and response surface methodology to set
optimum combination of process parameters for drilling of GFRP [7]. Sivaraos et al.
by applying Taguchi and RSM methods compared the CO2 laser machining on PVC
foam by considering the parameters such as cutting speed, laser power, frequency,
duty cycle, gas pressure on kerf width. They concluded that both the techniques
predicted near values of average errors and RSM produced promising results over
Taguchi method [8]. R. Adalarasan et al. studied the effect of CO2 laser cutting
parameters on surface roughness and kerf width of AA 6061/Al2 O3 composite using
L 27 orthogonal array. Desirability analysis was generated by RSM and to obtain opti-
mum setting of laser cutting parameters [9]. Rupesh Goyal et al. applied RSM to find
the optimum parameters setting of laser trepan drilling (LTD) to improve the hole
circularity and hole taper on superalloy sheets. They considered pulse width, pulse
frequency, gas pressure and trepanning speed as the input parameters, more circular
hole due to increased peak power for better penetration and higher pulse width and
lower gas pressure reduced the hole taper [10].
From the literature review, not much work has been reported on the optimiza-
tion of laser trepanning parameters on mild steel using Taguchi–response surface
methodology (T-RSM) method. Therefore, in the present work, optimization of
laser trepanning parameters for optimizing the responses such as hole-top circularity,
hole-bottom circularity and hole taper on mild steel is carried out by T-RSM.

2 Materials and Methods

Mild steel is soft, and hence, it can be machined easily. In this work, a mild steel
plate thickens 10 mm was used for laser trepanning. The non-uniformity in melting of
work material, incomplete ejection melt, improper penetration, etc., affect the laser
trepan hole quality characteristics. Hole quality characteristics of laser trepanning
depend on hole circularity on top/bottom and hole taper. The uneven surface in hole
diameter at top and bottom is known as hole taper. The responses such as hole top
circularity (HTC) and hole bottom circularity (HBC) and hole taper (HT) can be
calculated by Eqs. (1–3).

Minimum hole diameter at top

HTC = (1)
Maximum hole diameter at top
Optimization of Laser Trepanning Parameters for Mild Steel … 907

Table 1 Experimental parameters and levels

Symbol Parameters Level 1 Level 2 Level 3
A Laser power (W) 2500 2750 3000
B Gas pressure (bar) 5 6 7
C Speed of spot movement (mm/s) 1000 1100 1200
D Pulsing frequency (Hz) 1000 1500 2000
E Trepanning hole diameter (mm) 10 15 20

Minimum hole diameter at bottom

HBC = (2)
Maximum hole diameter at bottom
    
−1 hole diameter at top-hole diameter at bottom 180
HT = tan × × (3)
2 × plate thickness π

Taguchi DOE as a robust statistical design is used for the optimization of process
parameters. Laser trepanning on mild steel was performed as per the structured
Taguchi’s L 27 orthogonal array with 5 factors and 3 levels as shown in Table 1.
The S/N ratio and ANOVA are applied to study the contribution of input param-
eters in various processes. Taguchi transforms the objective function rate into S/N
ratio of mean standard deviation. It is essential to minimize the responses such as
HBC, HTC, HT always. Therefore, “smaller is better” (Eq. 4) for S/N ratio was con-
sidered for this experiment. P-values of individual input process parameters were
obtained by ANOVA using design expert software.
 
S/N = −10 ∗ log (Y 2 )/n (4)

where Y 2 = average observed data, n = number of observations, y = observed data.

Response surface methodology (RSM) is a popular statistical analysis tool for the
optimization of independent variable and dependent variables. It analyses the rela-
tionships between several explanatory variables and one or more response variables.
The 3D surface graph generated by RSM helps in visualizing the parameter effects
on the responses. The desirability value range between 0 and 1 indicates the nearness
of the outputs towards the target.

3 Result and Discussion

Laser trepanning on mild steel was performed successfully at the defined level of
parameters, and the corresponding responses obtained are shown in Table 2.
908 A. Gnanavelbabu et al.

Table 2 Laser trepanning parameters and responses (L 27 OA)

S. A (W) B C D (Hz) E (mm) HTC HBC HT
No. (bar) (mm/s) (Min/Max) (Min/Max) (°)
1 2500 5 1000 1000 10 0.9977 0.9896 1.3
2 2500 5 1000 1000 15 0.9951 0.9845 1.7
3 2500 5 1000 1000 20 0.989 0.9818 1.2
4 2500 6 1100 1500 10 0.9996 0.9916 1.3
5 2500 6 1100 1500 15 0.9925 0.9914 0.2
6 2500 6 1100 1500 20 0.9915 0.9897 0.3
7 2500 7 1200 2000 10 0.9927 0.9906 0.3
8 2500 7 1200 2000 15 0.9926 0.9848 1.3
9 2500 7 1200 2000 20 0.9918 0.9845 1.2
10 2750 5 1100 2000 10 0.995 0.9921 0.5
11 2750 5 1100 2000 15 0.9926 0.9918 0.1
12 2750 5 1100 2000 20 0.9844 0.9783 1
13 2750 6 1200 1000 10 0.9985 0.9931 0.9
14 2750 6 1200 1000 15 0.9959 0.9881 1.3
15 2750 6 1200 1000 20 0.9948 0.9825 2
16 2750 7 1000 1500 10 0.9934 0.9903 0.5
17 2750 7 1000 1500 15 0.9931 0.9863 1.1
18 2750 7 1000 1500 20 0.9903 0.9768 2.2
19 3000 5 1200 1500 10 0.998 0.9931 0.8
20 3000 5 1200 1500 15 0.9942 0.9901 0.7
21 3000 5 1200 1500 20 0.9904 0.9898 0.1
22 3000 6 1000 2000 10 0.9981 0.9921 1
23 3000 6 1000 2000 15 0.9908 0.9898 0.2
24 3000 6 1000 2000 20 0.9894 0.9866 0.5
25 3000 7 1100 1000 10 0.9957 0.9941 0.3
26 3000 7 1100 1000 15 0.9941 0.9876 1.1
27 3000 7 1100 1000 20 0.9885 0.9847 0.6

3.1 Analysis of Variance (ANOVA)

The ANOVA for hole top circularity is given in Table 3. The model F-value 6.93
implies the model which is significant (S). P-values less than 0.050 indicate signifi-
cant models at 95% of confidence level. In this case, B, D, E, AC, BE are significant
model terms. The obtained values for the HTC from ANOVA were S.D = 0.001715,
mean = 0.993322, R2 = 0.890, Adj. R2 = 0.762, pred. R2 = 0.459.
The ANOVA for hole bottom circularity is given in Table 4. The model F-value
3.354 implies that the model is significant at 95% of confidence level. In this case, E,
Optimization of Laser Trepanning Parameters for Mild Steel … 909

Table 3 ANOVA for HTC

Source SS DOF MS F-value P-value Significant
Model 2.85 × 10−04 14 2.04 × 10−05 6.931993 0.0009 S
A 4.24 × 10−06 1 4.24 × 10−06 1.441922 0.2530
B 1.98 × 10−05 1 1.98 × 10−05 6.748194 0.0233 S
C 5 × 10−07 1 5 × 10−07 0.170023 0.6874
D 4.29 × 10−05 1 4.29 × 10−05 14.57943 0.0024 S
E 1.91 × 10−04 1 1.91 × 10−04 64.87228 0.0001 S
AB 6 × 10−06 1 6 × 10−06 2.04027 0.1787
AC 2.09 × 10−05 1 2.09 × 10−05 7.109208 0.0206 S
AD 0 0
AE 2.8 × 10−06 1 2.8 × 10−06 0.95326 0.3482
BC 3.83 × 10−06 1 3.83 × 10−06 1.301428 0.2762
BD 0 0
BE 2.05 × 10−05 1 2.05 × 10−05 6.984808 0.0215 S
CD 0 0
CE 5.74 × 10−06 1 5.74 × 10−06 1.952142 0.1877
DE 3 × 10−08 1 3 × 10−08 0.010201 0.9212
A2 3.49 × 10−06 1 3.49 × 10−06 1.185381 0.2976
B2 0 0
C2 0 0
D2 0 0
E2 1.67 × 10−07 1 1.67 × 10−07 0.056674 0.8158
Residual 3.53 × 10−05 12 2.94E-06
Corr. total 0.000321 26
where SS sum of squares, DOF degree of freedom, MS mean square, S significant

BC, A2 are significant model terms. The obtained values for the HBC from ANOVA
were S.D = 0.002996, Mean = 0.987989, R2 = 0.796, Adj. R2 = 0.559, pred. R2 =
−0.062.
The ANOVA for hole taper is given in Table 5. The model F-value of 55.57
implies that the model is significant and significant model terms are A, B, C, D, E,
AB, AE, BE. The corresponding values were S.D = 0.001248, Mean = 0.257444,
R2 = 0.984809, Adj. R2 = 0.967087, pred. R2 = 0.905739.

3.2 3D Surface Graphs

To find the effects of interactions of input parameters, 3D surface graphs of HTC,

HBC, and HT were shown in Figs. 1a, b, 2a, b and 3a, b, respectively.
910 A. Gnanavelbabu et al.

Table 4 ANOVA for HBC

Source SS DOF MS F-value P-value Significant
Model 0.00042139 14 30.09 × 10−05 3.354396731 0.0211 S
A 1.068 × 10−05 1 10.68 × 10−05 1.190646047 0.2966
B 3.528 × 10−05 1 35.28 × 10−05 3.9317609 0.0707
C 3.0680 × 1 30.68 × 10−05 3.419178252 0.0892
10−05
D 3.0508 × 1 30.51 × 10−07 0.033999408 0.8568
10−07
E 2.8720 × 1 2.872 × 10−04 32.00691368 0.0001 S
10−04
AB 1.4727 × 1 14.73 × 10−05 1.641205561 0.2244
10−05
AC 2.8167 × 1 28.17 × 10−05 3.139019237 0.1018
10−05
AD 0 0
AE 4.8 × 10−07 1 4.8 × 10−07 0.053493346 0.8210
BC 4.8676 × 1 4.868 × 10−05 5.424621489 0.0381 S
10−05
BD 0 0
BE 1.4008 × 1 1.401 × 10−06 0.156115128 0.6997
10−06
CD 0 0
CE 3.8533 × 1 38.53 × 10−07 0.429432691 0.5246
10−06
DE 4.8 × 10−07 1 4.8 × 10−07 0.053493346 0.8210
A2 7.1501 × 1 71.50 × 10−05 7.968385739 0.0154 S
10−05
B2 0 0
C2 0 0
D2 0 0
E2 1.0417 × 1 1.042 × 10−06 0.11608799 0.7392
10−06
Residual 1.0768 × 12 89.73 × 10−07
10−04
Corr. total 5.2907 × 26
10−04
Optimization of Laser Trepanning Parameters for Mild Steel … 911

Table 5 ANOVA for HT

Source SS DOF MS F-value p-value Significant
Model 1.212 × 10−03 14 8.6569 × 55.5690936 0.0001 S
10−05
A 2.692 × 10−04 1 2.6924 × 172.822494 0.0001 S
10−04
B 5.69 × 10−05 1 5.6889 × 36.5170877 0.0001 S
10−05
C 2.69 × 10−05 1 2.6889 × 17.2600297 0.0013 S
10−05
D 8.643 × 10−06 1 8.6429 × 5.54786669 0.0364 S
10−06
E 2.222 × 10−05 1 2.2222 × 14.2644874 0.0026 S
10−05
AB 2.017 × 10−05 1 2.0167 × 12.9450223 0.0037 S
10−05
AC 6.66 × 10−07 1 6.6667 × 0.42793462 0.5253
10−07
AD 0 0
AE 8.333 × 10−06 1 8.3333 × 5.34918276 0.0393 S
10−06
BC 5 × 10−07 1 5 × 10−07 0.32095097 0.5815
BD 0 0
BE 8.333 × 10−06 1 8.3333 × 5.34918276 0.0393 S
10−06
CD 0 0
CE 3 × 10−06 1 3 × 10−06 1.92570579 0.1905
DE 2.083 × 10−06 1 2.0833 × 1.33729569 0.2700
10−06
A2 1.93 × 10−06 1 1.9286 × 1.23795373 0.2877
10−06
B2 0 0
C2 0 0
D2 0 0
E2 6.667 × 10−07 1 6.6667 × 0.42793462 0.5253
10−07
Residual 1.869 × 10−05 12 1.5579 ×
10−06
Corr. total 1.231 × 10−03 26
912 A. Gnanavelbabu et al.

a b

Fig. 1 a Interaction between A and B, b interaction between A and C

a b

Fig. 2 a Interaction between A and B, b interaction between B and C

a b

Fig. 3 a Interaction between B and A, b interaction between B and E

Optimization of Laser Trepanning Parameters for Mild Steel … 913

Figure 1a shows laser power (X-axis), gas pressure (Y-axis), and HTC (Z-axis).
It inferred from the graph that higher the laser power and higher the gas pressure,
minimizes the HTC value (0.9781). When laser power is maximum and gas pressure
is minimum, and HTC is maximum. Figure 1b shows laser power, speed of spot
movement and HTC are on X-, Y-, Z-axes, respectively. At higher speed on spot
movement and higher laser power, a minimal optimal value 0.9781 is achieved.
Hence, higher the laser power and higher speed on spot movement and minimized
the HTC.
Figure 2a represents laser power, gas pressure and HBC on X-, Y-, Z-axes, respec-
tively. The minimum HBC value 0.9772 was obtained at higher gas pressure and
nominal laser power. The maximum HBC values are obtained at the lower level of
gas pressure and laser power. Figure 2b represents gas pressure at X-axis, speed of
spot movement at Y-axis and HBC at Z-axis. An increase in speed of spot movement
and increase in gas pressure resulted in a minimum HBC.
Figure 3a represents X-axis laser power, Y-axis gas pressure and Z-axis hole taper.
At higher laser power and higher gas pressure, the hole taper was minimum. Maxi-
mum HT obtained at higher laser power and lower gas pressure. Figure 3b represents
gas pressure, trepanning hole and taper hole at X-, Y-, Z-axes, respectively. Hole
taper was reduced significantly at high gas pressure and low trepanning hole. Lower
trepanning value minimizes the hole taper.

3.3 Ramp Function Graph of Desirability

Ramp function graph for desirability is shown in Fig. 4. The desirability value
obtained is 0.992. It proves a higher closeness interval between the target data and
the responses.

Fig. 4 Ramp function graph of desirability

914 A. Gnanavelbabu et al.

4 Conclusion

• Laser trepanning on mild steel plate was successfully carried out as per Taguchi’s
L 27 orthogonal array, and the parameters were optimized using response surface
methodology.
• Optimal combination of parameters obtained through RSM was laser power-
2716.83 W, gas pressure-7 bar, speed of spot movement-1200 mm/s, pulsing
frequency-2000 Hz and trepanning hole 20 mm.
• The ramp graph indicates a desirability value of 0.992 proving a higher degree of
closeness between the target data and the responses.

References

1. Ashkenasi D, Kaszemeikat T, Mueller N, Dietrich R, Eichlera HJ, Illing G (2011) Laser

trepanning for industrial applications. Phys Procedia 12:323–331
2. Choudhury IA, Chong WC, Vahid G (2012) Hole qualities in laser trepanning of polymer
materials. Opt Lasers Eng 50:1297–1305
3. Shi C, Ren N, Wang H, Xia K, Wang L (2018) Effects of ultrasonic assistance on micro hole
drilling based on Nd: YAG laser trepanning. Opt Laser Technol 106:451–460
4. Kumar S, Dubey VK, Pandey AK (2013) Computer-aided genetic algorithm based multi-
objective optimization of laser trepan drilling. Int J Precision Eng Manuf 14:1119–1125
5. Chien W-T, Hou S-C (2006) Investigating the recast layer formed during the laser trepan drilling
of Inconel 718 using the Taguchi method. Int J Adv Manuf Technol 33:308–316
6. Dubey AK, Yadava V (2008) Robustparameter design and multi-objective optimization of laser
cutting for aluminum alloy sheet. Int J Adv Manuf Technol 38:268–277
7. Murthy BRN, Rodrigues LLR, Anjaiah D (2012) Process parameters optimization in GFRP
drilling through integration of Taguchi and response surface methodology. Res J Recent Sci
6:7–15
8. Sivaraos, Milkey KR, Samsudin AR, Dubey AK, Kidd P (2014) Comparison between Taguchi
method and response surface methodology (RSM) in modelling CO2 laser machining. Jordan
J Mech Ind Eng 8:35–42
9. Adalarasan R, Santhanakumar M, Thileepan S (2016) Selection of optimal machining param-
eters in pulsed CO2 laser cutting of Al6061/Al2 O3 composite using Taguchi-based response
surface methodology (T-RSM). Int J Adv Manuf Technol 93:305–317
10. Goyal R, Dubey AK (2014) Quality improvement by parameter optimization in laser trepan
drilling of superalloy sheet. Mater Manuf Process 29:1410–1416
Parametric Optimization of Cracked
Cantilever Beam Using Genetic
Algorithm

Mihir Kumar Sutar, Sarojrani Pattnaik, and Pawan Kumar Modi

Abstract This paper presents a smart crack detection system using genetic algorithm
(GA) for a cantilever beam. The accuracy of prediction is optimized by performing
parametric optimization of the first three natural frequencies. The three modes of
natural frequencies are obtained from the experimental procedures and are optimized
to reduce the percentage of error in predicting the relative crack position and crack
depth from a fixed position. Comparison of analytical method is in agreement with
the experimental values as the error percent between the experimental and predicted
values is below 10%.

Keywords Cracked cantilever beam · Genetic algorithm · Natural frequency ·

Parametric optimization

1 Introduction

Cracks generated in structural and engineering parts lead to their failure, and these
cracks influence the vibration characteristics of the machines. The occurrence of
crack in a machine accounts for decline in natural frequencies, and crack length rises
as the amplitude of vibration increases. This concerned issue has been addressed by
many researchers around the globe, and many methodologies have been suggested
by them.
Mahfouz and Banerjee [1] introduced cracks of diverse lengths at three different
sites along a cantilever steel beam, which was exposed to various amplitudes and
frequencies by the usage of an electromagnetic shaker. The crack length and crack
propagation data have been linked with statistical moments, frequency spectra, and

M. K. Sutar (B) · S. Pattnaik

Mechanical Engineering Department, Veer Surendra Sai University of Technology, Burla, Odisha,
India
e-mail: mihirsutar05@gmail.com
P. K. Modi
Electrical Engineering Department, Veer Surendra Sai University of Technology, Burla, Odisha,
India

© Springer Nature Singapore Pte Ltd. 2021 915

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_79
916 M. K. Sutar et al.

wavelet coefficient data. Khatir et al. [2] used experimentally determined natural
frequencies along with the particle swarm optimization (PSO) technique to detect
an open crack on a cantilever beam made up of steel. Genetic algorithm (GA) is
an optimization method which is bio-inspired, and it supports the ideas of natural
genetics and selection theories as proposed by Charles Darwin. It was developed at
the University of Michigan by John Holland and his co-workers. It provides robust
search in complex spaces [3]. GA is widely used in the field of engineering and
manufacturing. Pathan et al. [4] successfully built up a real-coded constrained GA
along with finite element (FE) method for optimizing the damping action of a unidi-
rectional composite laminate. Pajand and Tavakoli [5] pioneered a proficient process
for detecting cracks in concrete gravity dams using GA and finite element modeling
(FEM). It was found that this hybrid method could predict accurately and quickly
the displacement values in concrete gravity dam analysis.
Pawar and Ganguli [6] used genetic fuzzy approach to determine the location
of damage in cantilever beam and helicopter rotor pads. They used FEM to find
out the variation in beam frequencies due to structural damage. A fuzzy model was
developed and its rule-base, as well as membership functions, was optimized by GA
so that there was an automatic rule generation. The adopted method proved effective
in identifying the structural damage in both the beam and pads. Wang and Chen [7]
applied GA to choose optimal support location of beams with three different boundary
conditions. They considered both elastic and rigid supports, and it was found that
GA’s performance was better as compared to other optimization tools. Cho [8] used
GA to optimize the design of a composite cylindrical shell which is most prone to
buckling and failure on application of axial load, torsional load, or both. It was found
that the GA optimized shell design possessed better mechanical properties as regards
to conventional ones. Lin and Lee [9] applied GA along with regression analysis to
optimize the stacking sequence of laminated composite structures. Khatir et al. [10]
proposed a methodology using GA and firefly algorithm (FA) to identify the location
of damage in simple and continuous beams and two-dimensional frame structure.
The current analysis includes an experimental investigation of obtaining the three
modes of natural frequencies, e.g. first natural frequency (fnf ), second natural fre-
quency (snf ), and third natural frequency (tnf ), at different relative crack length
(RCL) and relative crack depth (RCD) values. These experimental data are used
further to develop a GA controller that can be used to predict the RCL and RCD.
Comparison between the experimental and predicted values indicates a minimal
percentage of error, and thus, the controller is acceptable. This will be helpful to pre-
dict the defects present in beam in the form of crack, without any experimentation,
thus minimizing the cost of experiment. The current analysis focuses on the health
monitoring of beam structural element for their safe functioning.
Parametric Optimization of Cracked Cantilever Beam Using … 917

Fig. 1 Experimental setup

for modal frequency
determination

2 Experimental Procedure

A cantilever beam with single crack was used for the experimentation. The relative
crack positions from the fixed end were varied at an interval of every 100 mm, and
the three modal frequencies were recorded. The results obtained in the experimen-
tation were further used as the input training data for the GA controller. The GA
controller was able to predict the RCL and RCD from the input of the three modal
frequencies. Number of experimental tests has been carried out on an 800 mm long
aluminium cantilever beam having a cross section of 50 mm width and 6 mm thick-
ness. Complete experimental setup is shown in Fig. 1. The three modes of natural
frequencies recorded during the experimentation for different RCL and RCD are
enlisted in Table 1.

3 Genetic Algorithm

GA is an optimization technique based on iteration process, and every iteration is

composed of stages, namely reproduction, crossover, and mutation. Reproduction
is nothing but selection of good parents from initial population who is having the
best fitness value so to produce offspring for the subsequent generation. Crossover
is done by choosing random mating points, and after that the new generation figured
out possess characteristics comparable to that of parents. Mutation is the intermittent
arbitrary alteration of the worth of a string position.
In this study, the fitness of all selected output variables is calculated on the basis of
fitness function based on multi-variable linear regression model. This model consists
of “n” independent input variables (x) whose relationship with the dependent variable
(response y) is given by Eq. 1:
918 M. K. Sutar et al.

Table 1 Three relative natural frequencies at different RCL and RCD from experimentation
Sl. No. Inputs Outputs
Relative 1st Relative 2nd Relative 3rd Relative crack Relative crack
natural natural natural length (RCL) depth (RCD)
frequency Frequency Frequency
“fnf” (A) “snf” (B) “tnf” (C)
1 0.9848 0.9958 0.9974 0.205 0.073
2 0.9673 0.9874 0.9942 0.43 0.084
3 0.9624 0.9948 0.9982 0.568 0.158
4 0.9689 0.9921 0.9948 0.4317 0.2774
5 0.9755 0.9976 0.9972 0.391 0.188
6 0.9655 0.9837 0.9961 0.3952 0.279
7 0.9785 0.9993 0.9987 0.387 0.189
8 0.9732 0.9916 0.9963 0.3772 0.1486
9 0.9852 0.9984 0.9967 0.23 0.24
10 0.9859 0.9989 0.9974 0.236 0.248
11 0.9724 0.9961 0.9817 0.545 0.287
12 0.9755 0.9869 0.9936 0.4827 0.2585
13 0.9824 0.9872 0.9917 0.447 0.406
14 0.981 0.9809 0.9932 0.495 0.424
15 0.9842 0.9926 0.9948 0.2345 0.2017
16 0.9834 0.9685 0.9974 0.535 0.534
17 0.9689 0.9921 0.9948 0.4317 0.2774

y = a + bx1 + cx2 + d x3 (1)

The regression equations for the selected responses, as given by Eqs. (2–3), are
further used as fitness function for process optimization using GA.

y1 = 24.158 − 7.8717x1 − 6.9487x2 − 9.2338x3 (2)

y2 = 13.615 + 4.596x1 − 10.371x2 − 7.6128x3 (3)

where y1 and y2 are responses RCL and RCD; x 1 , x 2, and x 3 are input parameters A,
B, and C, respectively. The flowchart of the GA process is shown in Fig. 2.
Parametric Optimization of Cracked Cantilever Beam Using … 919

Fig. 2 Genetic algorithm

flowchart

4 Confirmatory Experiments

Further confirmatory experiments were conducted at the obtained optimal settings as

predicted by GA. Then, the error percentage between the experimental and predicted
results was computed on the basis of mean absolute percentage deviation (MAPD),
as given by Eq. 4.
n  
100   (E t − P)t 
MAPD = (4)
n t=1  Et 

where E t and Pt are the experimental and predicted values. In Table 2, a comparison
has been made between the experimental and the predicted responses of the cracked
cantilever beam. It is seen that the percentage of error is found to be less than 10% in
all the cases, which shows that GA can be successfully used to forecast the responses
of the cracked cantilever beam.
From the confirmatory experimentation, it was observed that the GA predictions
are well within a limit of 10%, and the model is able to predict the relative crack
length and relative crack depth properly.

5 Conclusions

A genetic algorithm approach for predicting the relative crack position and its depth
is presented in the current analysis. The following conclusions are drawn from the
analysis:
• An aluminium cantilever beam of 800 mm length, 50 mm width, and 6 mm
thickness is taken for the current analysis.
 920

Table 2 Assessment of the experimental and predicted responses of the cracked cantilever beam by GA
S. No. Relative 1st natural Relative 2nd natural Relative 3rd natural Experimental Predicted % error
frequency “fnf” frequency “snf” frequency “tnf” values values
A B C RCL RCD RCL RCD RCL RCD
1 0.9850 0.9993 0.9987 0.2532 0.184 0.238 0.180 5.64 2.43
2 0.9757 0.9992 0.9986 0.2971 0.142 0.313 0.147 −5.38 −4.01
3 0.9834 0.9991 0.9986 0.2752 0.075 0.254 0.072 7.66 3.06
4 0.9828 0.9992 0.9986 0.247 0.169 0.257 0.164 −4.29 3.00
5 0.9640 0.9991 0.9986 0.3894 0.112 0.405 0.107 −4.18 4.44
6 0.9852 0.9993 0.9987 0.2447 0.115 0.237 0.119 2.86 −2.93
7 0.9697 0.9993 0.9986 0.3716 0.075 0.360 0.072 3.06 3.44
8 0.9735 0.9992 0.9986 0.3512 0.106 0.330 0.112 5.80 −5.54
9 0.9821 0.9993 0.9987 0.2791 0.163 0.262 0.170 6.12 −4.28
10 0.9786 0.9992 0.9987 0.2773 0.135 0.289 0.130 −4.47 3.91
11 0.9678 0.9992 0.9986 0.3867 0.117 0.375 0.115 2.81 1.69
12 0.9724 0.9992 0.9986 0.3276 0.118 0.339 0.125 −3.63 −5.47
13 0.9774 0.9992 0.9986 0.2885 0.181 0.300 0.176 −4.05 2.97
14 0.9707 0.9992 0.9986 0.3411 0.142 0.353 0.137 −3.48 3.57
15 0.9662 0.9992 0.9986 0.3672 0.151 0.388 0.158 −5.74 −4.61
16 0.9750 0.9993 0.9985 0.3408 0.167 0.319 0.173 6.30 −3.82
17 0.9640 0.9991 0.9986 0.3899 0.104 0.405 0.100 −4.05 3.45
Average % of error 5.86 7.82
M. K. Sutar et al.
Parametric Optimization of Cracked Cantilever Beam Using … 921

• Relative crack locations are intentionally provided in the beam with a variation
of crack depth to obtain the first three relative natural frequencies.
• The data obtained from the experimental procedure is then used as training data
to prepare the GA model to predict the structural health of the cantilever beam by
giving input of the three relative natural frequencies.
• Prediction of the GA model is found to be satisfactory within a limit of 10%.
• The average prediction error for RCL is found to be 5.86%.
• The average prediction error for RCD is found to be 7.82%.

References

1. Mahfouz IA, Banerjee A (2017) Crack detection and identification using vibration signals and
fuzzy clustering. Procedia Comput Sci 114:266–274
2. Khatir S, Dekemele K, Loccufier M, Khatir T, Wahab MA (2018) Crack identification method
in beam-like structures using changes in experimentally measured frequencies and particle
swarm optimization. CR Mec 346(2):110–120
3. Griesbach JD, Etter DM Fitness-based exponential probabilities for genetic algorithms applied
to adaptive IIR filtering, IEEE transactions. In: Thirty-second Asilomar conference on signals,
systems and computers, 1–4 Nov, 1998
4. Pathan MV, Patsias S, Tagariellia VL (2018) A real-coded genetic algorithm for optimizing the
damping response of composite laminates. Comput Struct 198:51–60
5. Pajand MR, Tavakoli FH (2015) Crack detection in concrete gravity dams using a genetic
algorithm. Proc Inst Civ Eng—Struct Build 168(3):192–209
6. Pawar PM, Ganguli R (2003) Genetic fuzzy system for damage detection in beams and
helicopter rotor blades. Comput Methods Appl Mech Eng 192:2031–2057
7. Wang BP, Chen JL (1996) Application of genetic algorithm for the support location optimization
of beams. Comput Struct 58(4):797–800
8. Cho HK (2018) Optimization of laminated composite cylindrical shells to maximize resistance
to buckling and failure when subjected to axial and torsional loads. Int J Precis Eng Manuf
19(1):85–95
9. Lin CC, Lee YJ (2004) Stacking sequence optimization of laminated composite structures
using genetic algorithm with local improvement. Compos Struct 63(1):339–345
10. Khatir A, Tehami M, Khatir S, Abdel WM (2018) Multiple damage detection and localization
in beam-like and complex structures using co-ordinate modal assurance criterion combined
with firefly and genetic algorithms. J Vibro Eng 20(1):832–842
Study of the Influence of Reinforcement
Parameters on Thermal Conductivity
of Magnesium-Based MMCs Through
Taguchi’s Orthogonal Array Approach

M. R. Shivakumar, N. V. R. Naidu, M. Jai Surya, and D. Indhuja

Abstract Light metal alloy-based composites (MMCs) find applications in auto-

motive and aerospace sectors, mainly due to their high strength-to-weight ratio.
Such composites exhibit enhanced mechanical properties too. However, these MMCs
inherently have inadequate thermal properties. In the present investigation, the
thermal conductivity of pure magnesium/soda–lime glass composite was analyzed
through Taguchi’s orthogonal array approach, and reinforcement parameters were
optimized to realize maximum possible thermal conductivity. Analysis of variance
was applied to find out the contribution of individual reinforcement parameters to
the thermal conductivity of the composite.

Keywords Magnesium composites · Thermal conductivity · Taguchi’s orthogonal

array

1 Introduction

Magnesium and its alloys have found certain specific uses in many industries. The
density of pure magnesium (1.70 g/cm3 ) is just 70% of the density of pure alu-
minum, and this aspect enhances its strength-to-weight ratio. The immediate need
for fuel economy and cost-savings in aerospace and automobile sectors has mandated
increasing uses of magnesium components [1]. Although the mechanical properties
(such as UTS and yield strength) of magnesium alloys are somewhat near to those of
aluminum alloys, their thermal properties (specific heat and thermal conductivity) are
much lower than most other non-ferrous and ferrous alloys. The mechanical proper-
ties of magnesium alloys can be enhanced by incorporating ceramic reinforcements
(to result in MMCs) [2, 3]. But, such reinforcements are likely to lead to adverse

M. R. Shivakumar (B) · N. V. R. Naidu · M. J. Surya · D. Indhuja

Industrial Engineering and Management, M S Ramaiah Institute of Technology, Bangalore,
Karnataka 560054, India
e-mail: mrshivakumar@msrit.edu

© Springer Nature Singapore Pte Ltd. 2021 923

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_80
924 M. R. Shivakumar et al.

effects on the thermal properties. There is an immediate need to analyze such neg-
ative effects in order to make the Mg alloy MMCs more appropriate for specific
end-users.
The following are the different processing routes that are suitable for processing
Mg alloy-based composites [3]:
1. Stir casting,
2. Squeeze casting, and
3. Powder metallurgy methods.
Of the above, the stir casting method appears to be very popular due to its
simplicity, economy of operation, and applicability even to small casting units [4, 5].
The effect of reinforcement on the properties of composites can be analyzed by
varying the reinforcement parameter levels. Optimum parameter levels can be deter-
mined by adopting statistical methods. Design of experiments (DOE) is a method
for scientifically applying statistics in conducting experimental runs. The analysis
involving experimental trials coupled with a methodically planned design procedure
is helpful to lessen the research effort, financial resources, and facilities. Taguchi
simplified the DOE approach with orthogonal array (OA) that significantly lowers
the number of experimental trials needed. Taguchi’s approach to DOE is easy to
adopt, apply, and analyze. Taguchi’s signal-to-noise ratio (S/N ratio) is invariably
used to analyze and optimize the responses [6]. Analysis of variance (ANOVA) is a
tool used to identify the significance of input parameters and to determine the effect
of each input parameter on the properties of the composites [7].

2 Material Selection and Processing

Magnesium and its alloys exhibit good castability, machinability, formability, and
weldability. In view of this and in the light of the ever-increasing applications of
magnesium components in various industries, an experimental investigation on Mg-
based MMCs was planned, with magnesium as the base material (matrix material)
and soda–lime glass powder as the reinforcement (soda–lime glass powder is an
easily available, inexpensive ceramic material) [8].
Magnesium ingots (of German origin) of purity more than 99.50% were procured
from a standard metal supplier. The certified composition (Table 1) is provided by
the manufacturer along with the supply.
The chemical composition of soda–lime glass is given in Table 2. It is noted that
the composition is cited in the literature (8).

Table 1 Chemical composition of magnesium

Elements Mg Fe Si Mn Cu Ni
Percentage by weight 99.92 0.025 0.017 0.026 0.0027 0.0006
Study of the Influence of Reinforcement Parameters on Thermal … 925

Table 2 Composition of
Elements SiO2 Na2 O CaO
soda–lime glass
Percentage by weight 70–73 12–16 10–15

Table 3 Levels of the reinforcement parameters

Parameter Level
1 2 3
A: weight of glass particles added (%) 1.5 3.0 4.5
B: glass particle size (µm) 75 125 210
C: temperature of glass particles (°C) 260 380 500

Magnesium ingots were charged into a graphite crucible covered with Magrex-60
flux and melted in an electric resistance furnace. When the melt reached 700 °C,
which is well above the melting temperature of magnesium, dross was skimmed off
from the surface of the melt after treating with degassing tablet hexachloroethene. The
melt was stirred (using the custom made mechanical stirrer) at a fixed stirring speed
of 400 RPM to form a sufficiently large vortex. Preheated soda–lime glass powder
(varying weight percentages, varying particle sizes and varying preheat temperatures)
was added into the vortex. After adding the particles, the stirring impeller was driven
for some more time till the temperature of the melt dropped to 640 °C. Soon after
this, the melt mixture was poured into the preheated die.
The three levels considered for each reinforcement parameter in the experiments
are shown in Table 3.
For the purpose of experimentation, based on the degrees of freedom for the factor
levels, the minimum number of trials needed (as per total degrees of freedom) is 9.
Taguchi’s L 9 OA was selected suitably. As per the design layout of experimentation,
nine castings were produced.

3 Test Results and Signal-to-Noise Ratio Analysis

The principle of comparative cut bar method (ASTM E1225 test method) was adopted
in the thermal conductivity measurement; this is perhaps the most widely used method
for axial thermal conductivity assessment. Cylindrical test specimen of size 8 mm
diameter and 50 mm length was sandwiched between two reference specimens as
shown in Fig. 1. As the test set-up is heated to predetermined temperatures, the
temperatures at different marked spots on the test specimen were measured accu-
rately using the temperature sensors located between the reference specimens and
the test specimen. These were then used to compute the thermal conductivity K of
the specimen under test.
If K R is the thermal conductivity of the reference specimens, the thermal
conductivity of the unknown sample (K S ) can be derived from the following equation:
926 M. R. Shivakumar et al.

Fig. 1 Test specimen

location (in between the
reference specimens)

 
Ls (T1 − T2 ) + (T3 + T4 )
Ks = K R × × (1)
LR 2(T2 + T3 )

Ls Length of test specimen

LR Length of reference specimens
T 1 , T 2 , T 3 , and T 4 Temperatures as shown in Fig. 1.
The experimental observations were further examined through S/N ratio. These
S/N ratios, which are log functions of desired output, serve as objective functions
for optimization and help in data analysis and prediction of optimum results. The
S/N ratio provides a measure of the impact of noise factors on performance. There
are different types of S/N ratios, viz. smaller-the-better, nominal-the-best and larger-
the-better. Larger the S/N ratio, the more robust the product is against noise. As the
objective of this investigation was to maximize the thermal conductivity, “larger-the-
better” concept was adopted in this work.
For “larger-the-better” characteristics,

S/N ratio = −10 log 10(1/n) (1/yi2) (2)

where “n” is the number of observations and “y” is the observed data.
Thermal conductivity test results and S/N ratios with design layout of experimen-
tation are shown in Table 4. The S/N ratio response values and the range for each
parameter are shown in Table 5.
The plot of the main effect for S/N ratios is shown in Fig. 2.
This plot indicates that among all the factors, weight % (Delta = 1.63) is the most
significant factor affecting the thermal conductivity. The optimal parameter levels
were identified from the main effect plot. The optimum parameter levels are:
Study of the Influence of Reinforcement Parameters on Thermal … 927

Table 4 Experimental results and S/N ratios

Test run A B C Observed values of Mean thermal S/N ratio (dB)
thermal conductivity
conductivity (W/m K)
(W/m K)
1 1.5 75 260 143 150 151 148 43.3975
2 1.5 125 380 145 148 145 146 43.2858
3 1.5 210 500 146 148 144 146 43.2854
4 3.0 75 380 144 135 134 138 42.7631
5 3.0 125 500 137 135 139 137 42.7326
6 3.0 210 260 134 134 137 135 42.6053
7 4.5 75 500 128 120 118 122 41.7113
8 4.5 125 260 121 120 125 122 41.7232
9 4.5 210 380 122 118 123 121 41.6515

Table 5 Main effects

Level A B C
(thermal conductivity) for
S/N ratio 1 43.32 42.62 42.58
2 42.70 42.58 42.57
3 41.70 42.51 42.58
Delta 1.63 0.11 0.01
Rank 1 2 3

Main Effects Plot for SN ratios

Data Means

A B
43.5

43.0

42.5
Mean of SN ratios

42.0

41.5
1.5 3.0 4.5 75 125 210
C
43.5

43.0

42.5

42.0

41.5
260 380 500

Signal-to-noise: Larger is better

Fig. 2 Plot of main effects (data mean) for S/N ratios

928 M. R. Shivakumar et al.

Table 6 Results of confirmation experiment

Factor A B C Thermal Mean thermal conductivity (W/m K)
conductivity
(W/m K)
Value 1.5 75 260 148 150 151 150

Table 7 ANOVA for test results

Source DF Seq SS Adj SS Adj MS F % of contribution
A 2 950.00 950.00 475.00 475.00 99.2
B 2 6.00 6.00 3.00 3.00 0.60
C 2 0.00 0.00 0.00 0.00 0.00
Error 2 2.00 2.00 1.00 0.20
Total 8 958.00

Weight %—1.5%
Particle size—75 µm
Preheat temperature—260 °C.
Confirmation tests were conducted corresponding to the optimum parameter lev-
els, and the results are shown in Table 6. From the results, it is seen that mean thermal
conductivity was maximum at the optimum parameter levels.

4 Analysis of Variance (ANOVA)

The objective of applying ANOVA is to investigate the effect of reinforcement param-

eters on the thermal conductivity. Test results were analyzed using ANOVA, with the
help of MINITAB software. Table 7 shows the contribution of the control parameters.
The results showed that the contribution of weight % of glass reinforcement on the
properties of composites is a high, 99.2%, whereas the contribution of particle size
is negligible (0.60%) and no contribution by glass preheat temperature.

5 Conclusions

Consistent production of magnesium/soda–lime glass powder composites (with vary-

ing levels of weight per cent and particle size) has been established using the stir
casting route, and the following observations are made:
• Thermal conductivity of these Mg alloy composites is affected by the weight
percentage as well as the particle size of the reinforcement.
Study of the Influence of Reinforcement Parameters on Thermal … 929

• Taguchi’s orthogonal array approach has substantially reduced the number of

experiments to be performed to determine the thermal conductivity of the
composites with the entire range of process variables.
• The optimum combination that results in the maximum thermal conductivity is
1.5% weight percentage of the reinforcement (soda–lime glass powder) of particle
size 75 µm.
• Preheating the soda–lime glass powder (prior to adding to the liquid melt) does
not have any effect on the thermal conductivity of the composites.
• ANOVA tests (confirmatory tests) reveal that the weight percentage of soda–lime
glass powder alone contributes 99.2% to the variation of thermal conductivity of
the composites.

Acknowledgements The authors wish to thank Prof. S. Seshan, Indian Institute of Science Banga-
lore, for his involvement in this work. Financial support received from Visvesvaraya Technological
University, Belagavi, is gratefully acknowledged.

References

1. Watarai H (2006) Trend of research and development for magnesium alloy. Sci Technol Trends
Q Rev 8
2. Luo A (1995) Processing microstructure, and mechanical behavior of cast magnesium metal
matrix composites. Metall Mater Trans A 26A:2445–2455
3. Koti V, George R, Shakiba A, Shivananda Murthy KV (2018) Mechanical properties of copper
nanotubes reinforced with uncoated and nickel coated carbon nanotubes. FME Trans 46:623–630
4. Shivakumar MR, Naidu NVR (2019) Characterization of mechanical and thermal properties in
soda–lime glass particulate reinforced LM6 alloy composites ICEMMM 2018. In: Advances in
materials and metallurgy. Springer, pp 149–157
5. Madhoo G, Shilpa M (2017) Optimization of process parameters of stir casting technique using
orthogonal arrays. Int J Adv Res Methodol Eng Technol 1(2)
6. Jayalakshmi, S. Kailas V, Seshan S, Fleury E (2006) Tensile strength and fracture toughness of
two magnesium metal matrix composites. J Ceram Process Res 7(3)
7. Taguchi G, Chowdhury S, Wu Y (2003) Taguchi’s quality engineering handbook. Wiley,
Hoboken, NJ
8. Hasanuzzaman M, Rafferty A, Sajjia M, Olabi AG (2016) Reference module in materials science
and materials. Published by Elsevier
Finite Element Modelling
and Optimisation of Sheet Hydroforming
for Cryo-rolled AA5083 Sheets

Akhil B. Raj, A. Arun , and Ajith Ramesh

Abstract Hydroforming is a manufacturing process that is used to form complex

geometries by applying fluid pressure. The punchless hydroforming process is the
most popular in the race, considering the absence of any punch that helps in reducing
the tooling costs. Based on the part geometries formed, the punchless process can be
classified into three categories, namely sheet hydroforming, shell hydroforming and
tube hydroforming. Sheet metal hydroforming is a hydroforming process that uses
hydrostatic fluid pressure for deforming the blank into a die cavity of the desired
shape. Owing to the inert advantages, like lower tooling costs, reduced processing
steps, remarkable precision, waste reduction and weight reduction, this process finds
extensive use in applications requiring a high strength-to-weight ratio, as in complex
automobile parts. The presented work involves the development of a nonlinear 2D
finite element model for the sheet hydroforming process of AA5182 (aluminium
alloy) using the FE package Abaqus/Explicit and validation of the numerical results
using the available literature (Daryl in Logan: a first course in the finite element
method. Cengage Learning Products, Canada 2007 [2]). The model is first validated
by reproducing the research work carried out by Bharatkumar Modi et al., considering
the input parameters, like blank holding force (BHF), sheet thickness and internal
pressure, and the corresponding output parameters, like material thickness reduction
and die corner radius. The research is further extended by replacing the current
blank material AA5182 using cryo-rolled AA5083. The effect of varying BHF, at
varying annealing temperature of blank considering varying frictional coefficients, is
also studied. This work also investigates the optimisation of the process using a well-
established design of experiments (DOE) technique—response surface methodology
(RSM).

Keywords Sheet hydroforming · Blank holding force (BHF) · Die corner radius ·
Finite element modelling · RSM

A. B. Raj · A. Arun (B) · A. Ramesh

Department of Mechanical Engineering, Amrita School of Engineering, Coimbatore, Amrita
Vishwa Vidyapeetham, India
e-mail: a_arun@cb.amrita.edu

© Springer Nature Singapore Pte Ltd. 2021 931

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_81
932 A. B. Raj et al.

1 Introduction

Forming is one of the most critical manufacturing processes where the material is
deformed and takes the desired shape with improved material strength. Hence, a lot
of research is carried out in this field to improve the process. Deep drawing is one
such forming processes, which uses a die and a punch for forming. The problem with
this kind of conventional forming process is that it deteriorates the surface quality
of the material. Also, the material thickness is non-uniform throughout the formed
surface. To improve this, the punch used in conventional processes is replaced with
hydrostatic fluid pressure. This kind of process is generally known as hydroforming
processes.
Hydroforming can be broadly classified into three: i. shell hydroforming pro-
cesses, ii. tube hydroforming process and iii. sheet hydroforming process [6]. The
current research work focuses on the effect of multiple parameters on a punchless
sheet hydroforming process.
Figure 1 shows a punchless sheet hydroforming set-up. As shown in zone A
(portion under the blank holder) due to tangential compression and radial tensions,
the sheet deforms in deep drawing mode [8]. In zone B, which indicates the die corner
radius, the state of stress is three-dimensional. This is because of the superimposing
of deep drawing and stretch bending components. The central portion of the blank,
i.e. the pole region (zone D), is subjected to uniform hydrostatic pressure due to
which it deforms in stretching mode, which leads to thinning of the blank. Initially,
the region between the die corner radius and the central portion (zone C) is slightly
deformed. During the corner formation, it undergoes additional stretching once the
draw-in is completed.
Blank holding force (BHF) is one of the most critical parameters in the sheet
hydroforming process. A varying BHF can yield better formability in terms of mini-
mum corner radius and lower thinning compared to a constant BHF [8]. Many other
parameters are also studied to improve formability. Varying blank annealing temper-
atures of around 300 °C have given better results for modern lightweight alloys [7,
9, 13]. The annealing temperature is highly dependent on the blank material used for

Fig. 1 Diagram showing

hydroforming of a
cup-shaped part [8]
Finite Element Modeling and Optimization of Sheet Hydroforming … 933

the process. Other vital parameters that have significant effects on formability are
friction and lubrication between the die and the blank surface. Teflon-coated die has
found to be producing less material thinning during the sheet hydroforming process
[3].
In recent years, there has been a tremendous demand for the need for lightweight
and high-strength alloys in the automotive and aerospace industries. Aluminium has
been found as one of the best solutions for this. However, aluminium’s low strength
and formability when compared to carbon steels have restricted its usage. Hence,
cryo-rolled Al alloy sheets that display relatively higher strength due to its ultra-
fine grain structure have occupied the space [4]. Annealing is found to enhance the
formability of cryo-rolled sheets. Cryo-rolling followed by hydroforming has a very
high potential to produce parts having complex shapes with lighter weight and high
strength [9, 11].
In the present work, the experimental results obtained in the literature by Modi
and Ravi Kumar [8] are first replicated using finite element modelling (FEM). The
research is further extended by replacing the current blank material AA5182 using
cryo-rolled AA5083. The effect of varying BHF, at varying annealing temperatures
along with a variable frictional coefficient, is also studied. This work also investigates
the optimisation of the process using a well-established design of experiments (DOE)
technique—response surface methodology (RSM) [14, 15].

2 Methodology

A nonlinear 2D FE model for sheet hydroforming process is developed using

Abaqus/Explicit package [1]. The effects of constant and varying BHF are stud-
ied. The initial model is a replica of experimental work undergone by Modi and Ravi
Kumar [8].
FE model is a powerful tool for analysing the hydroforming process and for saving
cycle time in designing and manufacturing. FEM is one of the powerful tools [2, 5].
The optimisation technology will secure an increased interest in practical application.
Table 1 shows the material property of the blank AA5182.
FE model is a powerful tool for analysing manufacturing processes and for saving
product design cycle time. The optimisation technology will secure an increased
interest in practical applications due to its effectiveness. Table 1 shows the material
property of the blank AA5182.

Table 1 Material properties

Young’s Ultimate Strain Strength
of blank material AA5182
modulus tensile hardening coefficient, K
(MPa) strength index, n (MPa)
(MPa)
152.5 281.2 0.35 594.8
934 A. B. Raj et al.

The arbitrary Lagrangian–Eulerian (ALE) scheme is used to model material flow.

The mesh follows a Lagrangian approach at the region of relatively lower deformation
and solves for displacements. However, in the area of the domain that corresponds
to extensive material flow, the mesh adapts into a Eulerian approach and addresses
for velocities [10, 12].
A surface-to-surface interaction is defined between the die and the bottom portion
of the blank. The die is set as the master and the blank as the slave. The friction for
die–blank interface is taken as semifluid film lubrication, and coefficient of friction
used is 0.3 [8]. In the model, the die is fixed in all degree of freedom and the blank
is allowed to flow in all the three directions.
In problems involving nonlinear geometry, the quality and the level of refinement
of mesh, especially at complicated contours, have a significant role to play in getting
accurate results. A four-node bilinear plane stress quadrilateral with reduced inte-
gration and hourglass control CPS4R is used for the mesh around 18,000 elements
used for the discretisation. It is a planar element with only one integration point.
However, discrete rigid elements: R2D2, are used to discretise the die geometry. It
is a two-dimensional linear rigid element.
Figure 2 shows the assembly model of blank and die that displays the applications
of fluid pressure and BHF. Figure 3 shows the variation of applied fluid pressure
concerning time. The fluid pressure is initially applied at a gradual rate, and once the
time reaches 0.125, there is a steady increase in the applied pressure. The BHF, on
the other hand, is used in two different ways, i.e. constant BHF, as shown in Fig. 4,
and varying BHF, as shown in Fig. 5.
The obtained results are validated with the numerical results available in the
literature [8]. The research work is further extended by replacing AA5182 with
cryo-rolled AA5083 in the validated model. Table 2 displays the material properties
of cryo-rolled AA5083 material.
The variation of input parameters like varying BHF and pressure is plotted in
Figs. 6 and 7. The effect of other parameters, like friction coefficient and blank
annealing temperature, is studied. Friction coefficient ranging from 0.04 to 0.25 is
considered for the simulation [3]. Also, the annealing temperature ranging from

Fig. 2 Meshed 2D assembly model

Finite Element Modeling and Optimization of Sheet Hydroforming … 935

Fig. 3 Plot for fluid pressure versus time [8]

Fig. 4 Constant BHF versus time—AA5182 [8]

936 A. B. Raj et al.

Fig. 5 Variable BHF versus time—AA5182 [8]

Table 2 Material properties

Young’s Ultimate Strain Strength
of AA5083
modulus tensile hardening coefficient, K
strength index (n) (MPa)
(MPa)
186.1 309.8 0.243 533

Fig. 6 Varying BHF versus time—cryo-rolled AA5083 [3]

Finite Element Modeling and Optimization of Sheet Hydroforming … 937

Fig. 7 Fluid pressure versus time—cryo-rolled AA5083 [3]

150 to 300 °C is considered, and their effects on output parameters like thickness
reduction and corner radius are investigated [13].

3 Results and Discussion

The first part of this work is focused on validating the developed model. The influence
of input process parameters like fluid pressure, BHF, friction at the bottom portion
of blank and die, etc., on output parameters like corner radius and material thickness
is investigated. Table 3 shows the comparison of output parameters obtained under
two conditions, i.e. constant BHF and varying BHF. Comparing experimental (from
journal paper) and simulation results, the percentage of difference in the case of
constant BHF for maximum thickness reduction is 3.7 and for variable BHF is 5.95.
For a minimum corner radius, the difference is 4.24 and 0.6 for constant BHF and
varying BHF, respectively.

Table 3 Comparison of experimental (from the literature) and FEM results for sheet hydroforming
for AA5182
Pressure BHF load Max. thickness reduction (mm) Min. corner radius (mm)
path in Experimental [8] FE simulation Experimental [8] FE simulation
hydroforming
Constant 0.8 0.77 37.7 36.1
Variable 0.84 0.79 33.4 33.2
938 A. B. Raj et al.

Fig. 8 Formed diagram

Fig. 9 Von Mises stress measured immediately after deformation in case of varying BHF

Figure 8 shows the forming diagram. Von Mises stress distribution in the material
after the hydroforming process is analysed, and it is found that higher stress con-
centration is at the bottom corners and along with the diagonally opposite faces of
the blank. The instantaneous stress distribution for the blank when the fluid pressure
and BHF (variable and constant BHF) are applied is shown in Fig. 9. Spring-back
behaviour of the deformed blank is analysed, and percentage relaxation in case of
hydroforming using constant BHF technique is 38.3% and that in case of variable
BHF technique is 32.52%. Thus, the spring-back effect is less in case of hydroform-
ing using variable BHF technique. Hence, better results are achieved with variable
BHF technique when compared to the constant blank holding force (BHF) technique.
The deformed shape obtained from the model is almost in agreement with the journal
paper.
Further, the work is extended by replacing the blank material with cryo-rolled
AA5083. Simulations are carried out with varying coefficient friction and varying
blank annealing temperature. Optimisation of output parameters is carried out using
response surface methodology (RSM), and ANOVA table is generated. Figure 10
shows the main effect plot for maximum thickness reduction, and Fig. 11 shows
the main effect plot for a minimum corner radius. In both cases, the coefficient
Finite Element Modeling and Optimization of Sheet Hydroforming … 939

Fig. 10 Main effect plot for maximum thickness reduction

Fig. 11 Main effect plot for minimum corner radius

940 A. B. Raj et al.

of friction and annealing temperature are reflected. The trend displays that, as the
coefficient of friction increases from 0.04 to 0.25, the maximum thickness reduction
is decreasing and the minimum corner radius is increasing until 0.145. Similarly,
annealing temperature increases from 150 to 300 °C, maximum thickness reduction
is falling, and the minimum corner radius is rising.
Three-dimensional surface plot for minimum corner radius versus annealing tem-
perature and coefficient of friction is shown in Fig. 12. Minimum corner radius is
achieved with an increase in the coefficient of friction and with a rise in anneal-
ing temperature. Surface plot for maximum thickness reduction versus annealing
temperature and coefficient of friction is displayed in Fig. 13.
The trend indicates that maximum thickness reduction is obtained by an increase
in both the coefficient of friction and annealing temperature. The optimal value for

Fig. 12 Surface plot for

minimum corner radius vs
annealing temperature and
coefficient of friction

Fig. 13 Surface plot for

maximum radius vs
annealing temperature and
coefficient of friction
Finite Element Modeling and Optimization of Sheet Hydroforming … 941

Table 4 Optimised results vs simulation results

Maximum thickness reduction (mm) Minimum corner radius (mm)
Optimised Simulation Percentage Optimised Simulation Percentage
result result of difference result result of difference
21.15 20.4 3.54 20.59 21.3 3.33

the coefficient of friction is 0.25 and for blank annealing temperature is 155 °C.
Simulations are carried out using the optimum value, which is illustrated in Table 4.
The percentage of difference is around 3.5% only.

4 Conclusion

A detailed finite element model for the punchless hydroforming process is devel-
oped and is validated with the available literature. The work is further extended by
changing the blank material from AA5182 to cryo-rolled AA5083 considering the
extent of applications. Cryo-rolling followed by hydroforming has better potential to
provide complex sheet metal parts using lightweight, high-strength aluminium alloy
sheets. The effect of input parameters like the coefficient of friction and annealing
temperatures for blank is studied, with varying BHF, on output parameters like min-
imum corner radius and maximum thickness reduction. It has been observed that
minimum corner radius and lower thinning are achieved in the case of variable BHF
technique than in the case of constant BHF. It indicates that the variable BHF tech-
nique is superior in forming complex sheet metal parts in sheet hydroforming. The
optimisation studies show an optimum value of 0.25 for the coefficient of friction
and annealing temperature; it is 155 °C.

References

1. ABAQUS (2016) ABAQUS documentation. Dassault Systèmes, Providence, RI, USA

2. Daryl L (2007) Logan: a first course in the finite element method. Cengage Learning Products,
Canada
3. Feyissa FT, Ravi Kumar D (2018) Enhancement of drawability of cryorolled AA5083 alloys
sheets by hydroforming, J Mater Sci Technol 2238–7854 2018. Brazilian Metallurgical,
Materials and Mining Association, pp 7–13, Elsevier Editora Ltd, pp 7–13 (2018)
4. Feyissa F, Ravi Kumar D (2018) Characterization of microstructure, mechanical properties and
formability of cryorolled AA5083 alloy sheets. ASM Int, JMEPEG 27:1614–1627
5. Huges TJR (2000) The finite element method. Dover Publications Inc., Mineola, New York
6. Lang LH, Wang ZR, Kang DC, Yuan SJ, Zhang SH, Danckert J, Nielsen KB (2004) Hydroform-
ing highlights: sheet hydroforming and tube hydroforming, J Mater Process Technol 151(1–3
spec. iss.):165–177
7. Mahabunphachai S, Koç M (2010) Investigations on forming of aluminium 5052 and 6061
sheet alloys at warm temperatures. Mater Des, Elsevier 31:2422–2434
942 A. B. Raj et al.

8. Modi B, Ravi Kumar D (2013) Development of a hydroforming setup for deep drawing of square
cups with variable blank holding force technique. Int J Adv Manuf Technol 66:1159–1169
9. Raja Satish D, Feyissa FT et al (2018) Formability of cryorolled aluminum alloy sheets in
warm forming. Int J Mater, Mech Manuf, 6(2):123–126
10. Ramesh A, Sumesh CS, Abhilash PM, Rakesh S (2015) Finite element modeling of orthogonal
machining of hard to machine materials. Int J Mach Machinability Mater 17:543–568
11. Rao PN, Singh D, Jayaganthan R (2013) Effect of annealing on microstructure and mechanical
properties of Al 6061 alloy processed by cryorolling. Mater Sci Technol 29(1):76–82
12. Sumesh CS, Ramesh A (2018) Numerical modelling and optimization of dry orthogonal turning
of Al6061 T6 alloy. Period Polytech Mech Eng 62(3):196–202
13. Taye F, Das P, Ravi Kumar D, Ravi Sankar B (2014) Characterization of mechanical properties
and formability of cryorolled aluminium alloy sheets. In: 5th International and 26th All India
manufacturing technology, design and research conference (AIMTDR), pp 511.6–511.11
14. Wang L, Chan LC, Lee TC (2007) Process modelling of controlled forming with time variant
blank holder force using RSM method. Int J Mach Tools Manuf 47:1929–1940
15. Zhu X, Liu D, Yang Y, Hu Y, Zheng Y (2016) Optimization on cooperative feed strategy
for radial-axial ring rolling process of Inco718 alloy by RSM and FEM. Chin Soc Aeronaut
Astronaut Beihang Univ 29(3):831–842
Selection of Parameterization Method
for Fitting of Freeform Curves Using
Uniformly Spaced Data

G. Rajamohan

Abstract The term “parameterization” refers to the mapping of measured data

points in Cartesian space into parametric space for the fitting of freeform curves and
surfaces, etc. Among the conventional parameterization methods, the chord length
and centripetal parameterization methods are generally considered adequate for most
of the applications. However, the choice between these two parameterization meth-
ods is not obvious. The present work attempts to analyze the measured data points
in order to select between the chord length and centripetal parameterization meth-
ods for achieving better results while fitting the freeform curves. The chord length
between two adjacent measured data points and the discrete curvature at each point
have been used to develop an approach to select between the two parameteriza-
tion methods considered. For validation purposes, the measured data points with
uniform spacing have been taken from some assumed example freeform curves of
varying complexities. The proposed approach has been found to correctly select the
appropriate parameterization method in all examples considered.

Keywords B-splines · Centripetal parameterization · Chord length

parameterization · Freeform curves and surfaces

1 Introduction

Reverse engineering is carried out to transform the engineering objects, for which the
original part drawings are not available, into computer-aided design (CAD) models
in order to improve their design, manufacture and analysis. In reverse engineering,
the data points from the desired engineering objects are acquired using appropriate
measuring devices, such as 3D scanners. The data points are acquired in Cartesian
space, whereas the CAD models are created in parametric space. The mapping of
data points from Cartesian space to parametric space is known as parameterization.
The accuracy of fitted CAD model depends on adequate parameterization.

G. Rajamohan (B)
National Institute of Foundry and Forge Technology, Ranchi 834003, India
e-mail: grajamohan.nifft@gov.in

© Springer Nature Singapore Pte Ltd. 2021 943

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_82
944 G. Rajamohan

Parameterization also has applications in computer-aided geometric design

(CAGD), such as texture mapping, re-meshing, curve and surface fitting. Among the
available conventional parameterization methods, the chord length and the centripetal
methods are generally used. The present work considers the fitting of parametric
freeform curves to measured data points using the B-splines representation.

2 Literature Review

The problem of fitting of freeform curves (and surfaces) to measured point clouds has
been the subject of research for many years. The existing parameterization methods
can be broadly categorized into conventional (single-step) methods, and evolutionary
and nature-inspired algorithms (optimization-based).
The literature reveals various conventional parameterization methods, such as
uniform, chord length, centripetal, etc. A summary of pros and cons of different
parameterization methods can be found in [1]. These parameterization methods work
fairly well with uniformly spaced data points. The literature also reveals that no
single parameterization method is capable of handling all types of data points (e.g.,
collinear data points). While most of the reported research is focused on open curves
and surfaces, few works based on closed curves and surfaces have been reported [2].
The conventional parameterization methods may not lead to optimal allocation
of parameters for data points with nonuniform spacing. This has led the researchers
to focus on developing algorithms based on some soft-computing techniques, such
as the evolutionary and nature-inspired algorithms. The algorithms reported in this
direction are based on artificial neural networks [3–7], genetic algorithms [8–14],
swarm intelligence [15–18], etc.
The chord length and centripetal parameterization methods are generally adequate
for most practical purposes [19]. The objective of this work is to develop and vali-
date an approach to choose between chord length and centripetal parameterization
methods in order to fit more accurate freeform curves for the given set of measured
data points. The B-splines have been used to represent the freeform curves.

3 Description of the Parameterization Methods

A brief description of the chord length and centripetal parameterization methods is

presented in this section. The parameter u is taken to be in the range [0, 1].
Selection of Parameterization Method for Fitting of Freeform … 945

3.1 Chord Length Parameterization Method [19]

The chord length parameterization method uses the chordal approximation for curves.
Let d be the total chord length, computed using Eq. (1).


N
d= |Q i − Q i−1 | (1)
i=1

Equation (2) can be used to compute the u parameters, i.e., ū i .

|Q i − Q i−1 |
ū 0 = 0; ū N = 1; ū i = ū i−1 + , i = 1, . . . , (N − 1) (2)
d
where, Qi (i = 0,...,N) are the data points. This parameterization method is widely
used and is generally considered adequate.

3.2 Centripetal Parameterization Method [19]

This parameterization method uses a slight variation to the chordal approximation

for curves. Let d be the total chord length, computed using Eq. (3).


N

d= |Q i − Q i−1 | (3)
i=1

Equation (4) can be used now to compute the u parameters, i.e., ū i .

√
|Q i − Q i−1 |
ū 0 = 0; ū N = 1; ū i = ū i−1 + , i = 1, . . . , (N − 1) (4)
d
where, Qi (i = 0,...,N) are the data points. This method gives better results than
chord length parameterization method when the data points take very sharp turns,
i.e., when the curvature changes rapidly.

4 Proposed Approach

The chord lengths and discrete curvatures are used in the proposed approach to
select between the chord length and centripetal parameterization methods. The chord
length is the distance between two adjacent measured points and is computed using
the distance formulae. The discrete curvatures are calculated at all measured points
using the algorithm described in the subsection below. The calculated chord lengths
946 G. Rajamohan

and discrete curvatures are normalized in the range [0, 1], and standard deviations are
calculated. The attribute, i.e., the chord length or discrete curvature, having higher
standard deviation is used to select the most suitable parameterization method. In
case of chord length, the chord length parameterization is selected, and in case of
discrete curvature, the centripetal parameterization method is selected.

4.1 Computation of Discrete Curvature [20]

Let Qi (i = 0, 1,…, N) represent an ordered set of (N + 1) measured data points.

Assuming the sequential connection of these points to form an open polygon, the
curvature at each measured point Qi (denoted as K i ) is the relative curvature of the
curve passing through three adjacent points, viz. Qi−1 , Qi and Qi+1 (Fig. 1a). The
curvature at the interior points (Q1 , Q2 , …, QN −1 ) can be expressed using circular
curvature as K i = (2 * Qi−1 Qi Qi+1 )/(L i L i+1 S i ), where L i = ||Qi − Qi−1 ||, L i+1 =
||Qi+1 − Qi ||, S i = ||Qi+1 − Qi−1 || and Qi−1 Qi Qi+1 is the signed area of triangle. The

Qi αi

Li Li+1
O
R
2π-2αi Qi+1
Si
Qi-1

(a)

QN+
Q1
Q0 Q2
QN-1 QN
QN-2
-
Q0

(b)
Fig. 1 Estimation of discrete curvature; a curvature of a polygon; b additional point supplement
for two end points
Selection of Parameterization Method for Fitting of Freeform … 947

curvature at endpoints Q0 and QN can be computed using the procedure described

below:
Step 1: Compute the tangent vectors at endpoints Q0 and QN by applying the
circular end conditions. The points Q0 , Q1 and Q2 are selected for computing the
tangent vector m̂ 0 at Q0 , and points QN −2 , QN −1 and QN are selected for computing
the tangent vector m̂ N at QN . Let a0 = Q1 − Q0 , b0 = Q2 − Q0 and c0 = a0 ×
b0 . The tangent vector m̂ 0 is then computed as m̂ 0 = (r 0 × c0 )/||r 0 × c0 ||, where
r 0 is computed as r 0 = {||a0 ||2 (b0 × c0 ) + ||b0 ||2 (c0 × a0 )}/(2 ||c0 ||2 ).
Step 2: As shown in Fig. 1b, two additional points Q − +
0 and Q N are added, respec-
tively, along the direction of −m̂ 0 and −m̂ N , satisfying ||Q1 − Q0 || ≥ || Q − 0 − Q0 ||
and ||QN −1 − QN || ≥ || Q +N − QN ||.
Step 3: The curvatures at Q0 and QN are now calculated in the same way as interior
points with Q − +
0 , Q0 , Q1 and QN −1 , QN , Q N , respectively.

5 Results and Discussion

The accuracy of freeform curves (and surfaces) fitted using the measured data points
depends mainly on parameterization. Among the available conventional parameter-
ization methods, the chord length and centripetal methods are mostly used owing
to their simplicity and ease of implementation. The present work is an attempt to
select between these two methods to fit better freeform curves for given measured
data points and to validate the selection approach. The selection is based on the stan-
dard deviation of discrete curvatures and chord lengths, computed using the given
measured data points. The B-splines representation has been used to represent the
freeform curves. The programs were written in C++ language.
In order to make a comparison between the centripetal and chord length parameter-
ization methods, the measured data points have been selected from known freeform
curves. The curves are first discretized in parametric space with fine spacing of 0.005,
giving 201 points in the range assumed, i.e., [0, 1]. Measured data points have been
taken with uniform spacing between them, as the measuring machines typically gen-
erate such data easily. Three runs have been made with the numbers of measured data
points as 11, 21 and 51. The discrete curvature and chord lengths have been computed
using these data points, and standard deviations of the respective measures have been
computed. The chord length parameterization method is chosen when the standard
deviation of chord length is higher and vice versa.
Both the centripetal and the chord length parameterizations have been imple-
mented in order to validate the selection by using the proposed approach. The mea-
sured data points and the respective parameters were used to fit the freeform curves
of degree 3 using interpolation technique [19]. The normal deviations between the
fitted freeform curves and the corresponding point on the known freeform curves
are then calculated. The maximum normal deviation has been used to compare the
948 G. Rajamohan

chord length and centripetal parameterization methods. The selected parameteri-

zation method is supposed to yield lower value of normal deviation. The normal
deviations (mm) obtained using different example freeform curves are shown in
Tables 1, 2 and 3 for the numbers of measurements points of 11, 21 and 51, respec-
tively. The boldfaced entries in column 2 of these tables indicate the mismatch in
selected method using the proposed approach.
As evident from the tables, the prediction accuracy of the proposed approach is
less with fewer data points, and it gets better with increase in the number of data
points. It may be because the discrete curvature calculation algorithm gives better

Table 1 Normal deviations (mm) for measurement data points size of 11

Example No. Selected method Centripetal parameterization Chord length
parameterization
#1 Chord length 0.003163 0.003038
#2 Chord length 0.032463 0.049099
#3 Centripetal 0.008274 0.043202
#4 Centripetal 0.998697 2.588100
#5 Chord length 0.016959 0.026198
#6 Chord length 0.430049 0.635121

Table 2 Normal deviations (mm) for sparse measurement data points size of 21
Example No. Selected method Centripetal parameterization Chord length
parameterization
#1 Chord length 0.000122 0.000087
#2 Centripetal 0.003633 0.008180
#3 Centripetal 0.000687 0.002746
#4 Centripetal 0.039825 0.154892
#5 Centripetal 0.002091 0.003382
#6 Centripetal 0.054421 0.084704

Table 3 Normal deviations (mm) for dense measurement data points size of 51
Example No. Selected method Centripetal parameterization Chord length
parameterization
#1 Chord length 0.000024 0.000022
#2 Centripetal 0.000146 0.000210
#3 Centripetal 0.000033 0.000069
#4 Centripetal 0.000561 0.002775
#5 Centripetal 0.000112 0.000112
#6 Centripetal 0.000587 0.000774
Selection of Parameterization Method for Fitting of Freeform … 949

results with dense points. The examples curves #2, #5 and #6 particularly have
higher curvature variations, while the remaining curves have moderate curvature
variations. Considering this aspect, it appears that the curvature calculation algorithm
yields better values even with smaller numbers of measured points if the variation
in curvature is not severe. With dense measured data points (Tables 2 and 3), the
proposed approach has been able to accurately estimate the suitable parameterization
method in all cases considered. The fitting accuracy also improves significantly. In
most cases, the centripetal method is being suggested as the most suitable method.
This may be because the freeform curves typically have more curvature variations
along their length.

6 Conclusions

Many industrial and technological applications require the fitting of parametric

freeform curves and surfaces to measured points. The parameterization method used
plays crucial role on the accuracy of fitted freeform curves and surfaces. An approach
has been proposed in this paper to select between the chord length and centripetal
methods for fitting of freeform curves using the given sets of measured data points.
The approach has been validated using multiple example freeform curves of varying
complexities. The following conclusions are drawn:
• The accuracy of determination of most suitable parameterization method has been
found to be lower with small number of measured data points, and it improves
with increase in the number of data points.
• The proposed approach works well with uniformly spaced, dense measured data
points. Such a scenario being typically the case in applications such as reverse
engineering, the proposed approach can be of practical use.
The following may constitute the scope for future work:
• In curvature calculation, higher order approximation between the adjacent points
may be tried, instead of straight line approximation, for better curvature estimates.
• The present work uses the interpolation technique to fit freeform curves for
uniformly spaced data points. Use of approximation technique may be attempted.

References

1. Haron H, Rehman A, Adi DIS, Lim SP, Saba T (2012) Parameterization method on B-spline
curve. Math Prob Eng. https://doi.org/10.1155/2012/640472
2. Park H (2001) Choosing nodes and knots in closed B-spline curve interpolation to point data.
Comput Aided Des 33(13):967–974
3. Bishop CM, Roach CM (1992) Fast curve fitting using neural networks. Rev Sci Instrum
63:4450–4456
950 G. Rajamohan

4. Gu P, Yan X (1995) Neural network approach to the reconstruction of freeform surfaces for
reverse engineering. Comput Aided Des 27(1):59–64
5. Hoffmann M, Varady L (1998) Freeform surfaces for scattered data by neural networks. J Geom
Graph 2:1–6
6. Echevarria G, Iglesias A, Galvez A (2002) Extending neural networks for B-spline surface
reconstruction. Lect Notes Comput Sci 2330:305–314
7. Krause FL, Fischer A, Gross N, Barhak J (2003) Reconstruction of freeform objects with
arbitrary topology using neural networks and subdivision techniques. Ann CIRP 52(1):125–128
8. Sarfraz M, Raza SA (2001) Capturing outline of fonts using genetic algorithm and splines. In:
Proceedings of the international conference on information visualization. ISBN: 0-7695-1195-3
9. Renner G, Ekart A (2003) Genetic algorithms in computer aided design. Comput Aided Des
35:709–726
10. Yoshimoto F, Harada T, Yoshimoto Y (2003) Data fitting with a spline using a real coded
genetic algorithm. Comput Aided Des 35:59–64
11. Kumar GS, Kalra PK, Dhande SG (2003) Parameter optimization for B-Spline curve fitting
using genetic algorithms. Proc Congr Evoluti Comput (CEC’03) 3:1871–1878
12. Sun YH, Xiang WJ, Xia DS (2010) Parameter optimization for B-Spline curve fitting based on
adaptive genetic algorithm. J Comput Appl 30:1878–1882
13. Zhao L, Jiang J, Song C, Bao L, Gao J (2013) Parameter optimization for Bezier curve fitting
based on genetic algorithm. Adv Swarm Intell 7928:451–458
14. Hasegawa AY, Tormena C, Parpinelli RS (2014) Bezier curve parameterization using a multi-
objective evolutionary algorithm. Int J Comput Sci Appl 11(2):1–18
15. Galvez A, Cobo A, Pey JP, Iglesias A (2008) Particle swarm optimization for Bezier surface
reconstruction. In: Proceedings of the 8th international conference on computational science
(part II)
16. Galvez A, Iglesias A (2011) Efficient particle swarm optimization approach for data fitting
with free knot B-Splines. Comput Aided Des 43:1683–1692
17. Galvez A, Iglesias A (2012) Particle swarm optimization for non-uniform rational B-Splines
surface reconstruction from clouds of 3D points. Inf Sci 192:174–192
18. Galvez A, Iglesias A (2013) Firefly algorithm for Bezier curve approximation. In: Proceedings
of the 12th international conference on computational science and its applications, pp 81–88
19. Piegl L, Tiller W (1997) The NURBS book. Springer, Berlin
20. Liu GH, Wong YS, Zhang YF, Loh HT (2002) Adaptive fairing of digitized point data with
discrete curvature. Comput Aided Des 34:309–320
Energy Engineering
Experimental Study of an Axial Turbine
for Wave Energy Conversion

Kumud Kumar, Tapas K. Das , R. Srikanth, and Abdus Samad

Abstract With the growing menace of increase in population and pollution, renew-
able sources of energy are being preferred over the fast depleting fossil fuels. Ocean
wave energy is one such form of renewable source with a huge potential all around the
globe. Wave energy conversion (WEC) devices are used to harness the energy from
ocean waves. An oscillating water column (OWC) is a WEC device, which uses an
axial turbine to convert wave energy into useful electrical energy. The experimental
setup designed and assembled at Wave Energy and Fluids Engineering Laboratory,
IIT Madras tested the turbine subjected to bidirectional airflow. The objective of the
preliminary experiment is to measure the rotational speed and pressure drop across
the turbine. The variable parameters are stroke length and time period of the oscil-
lating piston, which simulates different ocean wave conditions. A comparative study
of the parameters is carried out and reported in this article.

Keywords Wells turbine · Wave energy · Experimental study · OWC

Nomenclature

α Angle of attack (degree)

Fax Axial force component (N)
Ftan Tangential force component (N)

K. Kumar · T. K. Das (B) · R. Srikanth · A. Samad

Wave Energy and Fluids Engineering Lab, Ocean Engineering Department,
IIT Madras, Chennai, India
e-mail: mech.tapas@gmail.com
K. Kumar
e-mail: kumud142k@gmail.com
R. Srikanth
e-mail: srikanthrangisetty@gmail.com
A. Samad
e-mail: samad@iitm.ac.in
© Springer Nature Singapore Pte Ltd. 2021 953
S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_83
954 K. Kumar et al.

b Blade width (mm)

c Blade chord length (mm)
D Drag force (N)
L Lift force (N)
Nb Number of blades
V Relative velocity (m/s)
σ Turbine Solidity
h Hub-to-tip ratio
SL Stroke length (m)
CT Cycle time
Rt Rotor tip radius (mm)
Rh Rotor hub radius (mm)
tc Tip clearance (mm)
AR Aspect ratio
N Revolutions per minute
P Differential Pressure drop (kPa)
T Time (s)

Abbreviation

OWC Oscillating water column

WEC Wave energy conversion
CAD Computer aided design
VFD Variable frequency drive
DAQ Data acquisition system

1 Introduction

Wave energy is a form of renewable energy source which is environment-friendly and

has the potential to be economically accessible in the near future. Renewable sources
of energy are unlimited and eco-friendly; they don’t produce harmful pollutant gases
such as carbon dioxide, nitrogen oxides, or sulfur oxides. For countries that are
surrounded by ocean such as India, Japan, Australia, England, to mention a few, and
remotely situated islands, wave energy conversion principle is very promising for
energy needs. Wave energy is the energy harnessed from water waves. When the
wind blows over the water surface, it transfers the energy to waves. Wave power is
directly proportional to wind speed; more the speed of wind more will be the wave
energy. The global wave energy potential is estimated to be 8000–80,000 TW/yr [1].
This energy potential can be harnessed by using wave energy conversion technology
Experimental Study of an Axial Turbine for Wave Energy Conversion 955

to do useful work such as the generation of electricity, desalination of seawater, and

pumping water to reservoirs.
Wave energy is harnessed by using wave energy conversion devices such as point
absorbers, overtopping device, oscillating water columns, and many more. Oscillat-
ing water column (OWC) is a WEC device that is easy to install in both onshore and
offshore conditions [2]. An OWC plant consists of an air chamber with an opening
above the water surface, as shown in Fig. 1. The incident water waves force the
water column to oscillate up and down like a piston-cylinder arrangement inside the
chamber. The air trapped above the water surface is subjected to compression and
decompression simultaneously, and this results in the generation of bidirectional air
flow. The energy from this bidirectional air flow can be harnessed by using an air
turbine designed by professor Alan Arthur Wells at Queens University, Belfast in the
1970s. This air turbine is known as the Wells turbine [3].
Plenty of research has been carried out on both numerical and experimental anal-
ysis of the Wells turbine [4]. While many experimental works on Wells turbine stud-
ied steady-state characteristics of the turbine for different blade geometries [5–7],
researchers also carried out laboratory experiments to study the turbine character-
istics for bidirectional airflow similar to the conditions in OWC [8, 9]. The Wave
Energy and Fluids Engineering Lab (WEFEL) at IIT Madras developed a test rig
with a piston-cylinder arrangement to emulate the bidirectional airflow similar to
an OWC. Previously a scaled impulse turbine model was tested in this test rig [10].
In the present work, a preliminary experiment carried out on monoplane Wells tur-
bine subjected to bidirectional air flow through the piston-cylinder arrangement. The

Fig. 1 Schematic of an OWC

956 K. Kumar et al.

measured performance parameters are rotational speed and the pressure difference
across the turbine. The stroke length and the time period of the oscillating piston
are varied to create different bidirectional airflow profiles similar to different wave
conditions.

2 Working Principle of the Wells Turbine

Oscillating water columns utilize Wells turbine for converting wave energy to elec-
trical energy. Wells turbine is an axial flow bidirectional turbine especially used in
OWC type wave energy converters. The turbine is made up of symmetrical aerofoil
type blades fixed around the hub (Fig. 2). The aerofoil blades are positioned around
the hub with the chord planes perpendicular to the axis of rotation. The turbine
transmits wave energy to the coupled electric generator through rotation.
The Wells turbine working principle is explained in [11]. When the flow is incident
on the turbine blade at an angle of attack (α), a lift force is generated normal to
the direction of relative velocity. Simultaneously a drag force is experienced in the
direction of the relative velocity. Resolving these lift and drag forces (Fig. 2) gives
components of forces in the plane of rotation (tangential force) and normal to the
plane of rotation (axial force), Ftan and Fax , respectively.

Ftan = L sin α − D cos α (1)

Fax = L cos α + D sin α (2)

Whatever may be the direction of incoming flow, the tangential force always acts
in the same direction. Therefore, the turbine rotation always remains in the same
direction regardless of flow direction. This turbine performs optimally up to a value

Rotor
Bi-Directional
airflow L
D
Symmetrical
aerofoil
Outer
casing α
V
(a) (b)

Fig. 2 a CAD model of Wells turbine with shroud, b forces acting on the blade
Experimental Study of an Axial Turbine for Wave Energy Conversion 957

of angle of attack (α), further increase in its value results in flow separation around the
blades. This angle at which flow separation occurs is called as stall angle. Beyond this
angle, the magnitude of lift and drag forces reduces; as a result, efficiency decreases.
One main drawback of Wells turbine is that it is difficult to achieve its operational
speed due to its poor starting characteristics [12].

3 Turbine Design Parameters

The aerodynamic performance of the Wells turbine depends upon various parameters.
These parameters with optimum values are discussed in this section. All the param-
eters are inter-related. Based on the analytical studies, turbine geometric parameters
are given in [11]. All such important parameters are enlisted in Table 1. One of
the most important parameters is solidity, which measures the mutual interaction
between the blades. With the increase in solidity, turbine blades tend to come closer,
thereby hindering flow passage and as a result efficiency of turbine reduces. Previous
experiments show that with an increase in the hub-to-tip ratio beyond certain value
results in loss of efficiency [11]. The aspect ratio and tip clearance play an important
role in turbine performance. The Wells turbine shows high sensitiveness towards tip
clearance.
A Monoplane Wells turbine was modeled and then fabricated for testing in the
laboratory. The 3D CAD model and the actual turbine is shown in Fig. 3. The turbine
is modeled on the recommended values of various design parameters. It is fabricated
by rapid prototyping, and the material used is ABS Plastic (Acrylonitrile Butadiene
Styrene). The turbine geometrical dimensions are listed in Table 2.

Table 1 Recommended
Parameter Analytical Remarks
parameters
description
Solidity (σ ) Nb c/π Rt Efficiency reduces
for σ > 0.5
Aspect ratio (AR) b/c AR ~0.5 is
recommended
Hub-to-tip ratio Rh /Rt Recommended
(h) value h ~0.6
Tip clearance (tc ) % of c ≤2% of c is
recommended
958 K. Kumar et al.

Fig. 3 Actual Wells turbine model

Table 2 Actual turbine parameters

Parameter Description Parameter Description
Aerofoil profile NACA0021 Chord length 36 mm
Number of blades 10 Rotor solidity 0.69
Hub radius 68 mm Aspect ratio 0.8
Hub-to tip ratio 0.69 Tip clearance 2 mm

4 Experimental Setup and Procedure

The experiment is carried out at a laboratory facility in IIT Madras. This experimental
setup consists of a piston-cylinder arrangement which is driven by a motor and
gearbox. Motor rotations are controlled by a variable frequency drive (VFD). The
measured parameters include differential pressure drop across the turbine rotor and
the rotational speed. Table 3 lists the instruments used for measuring the parameters.
The signals from the instrument are recorded by a data acquisition system (DAQ).
The input signals are processed and stored as data in a readable format using LAB
VIEW software.
The VFD runs the motor to generate the required cycle time (CT) of the piston,
which relates to ocean wave time period. The power from the motor shaft is trans-
ferred to the piston rod through worm wheel gearbox (gear ratio 1:40). Ocean wave
height is replicated by piston stroke length (SL). The frequencies corresponding to
different CT of the piston are: 22.5 Hz relates to CT = 4 s and 18.5 Hz to CT =

Table 3 Instrument specifications

Instrument Parameter Range Accuracy Output signal
ABB 266DSH Differential pressure ±10 kPa 0.06% 4–20 mA
Rotational speed Rotational speed 0–5000 rpm ±5% of reading 4–20 mA
frequency converter
(Pepperl + Fuchs)
Experimental Study of an Axial Turbine for Wave Energy Conversion 959

Fig. 4 a CAD model of the turbine assembly, b actual turbine assembly, c the piston-cylinder
system, d complete setup

5 s. The oscillating motion of the piston creates the bidirectional air flow through the
duct. Once the piston starts reciprocating, data is recorded after some time until the
flow is fully developed. The sampling rate of data taken is 100 ms, and it is recorded
for three minutes. The complete experimental setup is shown in Fig. 4.

5 Results and Discussions

All the experiment tests data is recorded by DAQ system, and the comparative results
are plotted as the function of time. Although each data is collected for three minutes
duration, the present result is shown for only one full cycle for the ease of under-
standing. Figure 5 shows the variation of pressure drop and rotational speed of the
960 K. Kumar et al.

2 800
SL=0.4m SL=0.4m
SL=0.5m SL=0.5m
1 600

ΔP (kPa)

N (rpm)
0 400

-1 200

-2 0
0.0 2.5 5.0 0.0 2.5 5.0
T(sec) T (sec)
(a) (b)

Fig. 5 Variation of a pressure drop and b rotational speed for different stroke length at same time
period T = 5 s

turbine for two different stroke lengths (0.4 and 0.5 m) for a fixed time period of 5 s.
In Fig. 5a, the first half cycle shows differential pressure during compression of the
piston and the other half cycle is for suction of the piston. Theoretically, it should be
similar to a sinusoidal curve; however, due to the presence of the supports inside the
duct, the flow faces blockage during the suction stage. As a result, the compression
peak pressure drop is slightly higher than the suction side pressure drop. The turbine
experiences more air flow rate when the stroke length is longer. This is visible in
Fig. 5(b) for the rotational speed. The rotational speed for SL = 0.5 m is much higher
compared to smaller stroke length. For SL = 0.4 m, the turbine rotational speed is
almost zero during the transition from compression to the suction stage. This situ-
ation interrupts the continuous power production from the generator. However, for
SL = 0.5 m, the turbine rotates continuously during both suction and compression
stage of the piston motion.
Figure 6 shows the variation of pressure drop and rotational speed of the turbine
for two different time periods (4 and 5 s) when the stroke length is fixed at 0.4 m.
Higher time period means the piston moves slowly; the rate of air flow is less. So,
the pressure drop is also less when the time period is higher. The same is visible in
Fig. 6a. The variation of rotational speed is shown in Fig. 6b. When the time period is

T=4sec 1200
T = 4sec
2 T=5sec 1000 T = 5sec
N (rpm)

800
ΔP (kPa)

600
0
400
200
-2 0
0.0 2.5 5.0 0.0 2.5 5.0
T (sec) T (sec)
(a) (b)

Fig. 6 Variation of a pressure drop and b rotational speed for different time period at same stroke
length SL = 0.4 m
Experimental Study of an Axial Turbine for Wave Energy Conversion 961

Unidirectional 250 Unidirectional

1.6
Bidirectional Bidirectional
200
0.8

N (rpm)
ΔP (kPa)
150
0.0
100
-0.8 50
0
0.0 2.5 5.0 0.0 2.5 5.0
T (sec) T (sec)
(a) (b)

Fig. 7 Variation of a Pressure drop and b rotational speed for unidirectional and bidirectional
condition at SL = 0.4 m, T = 5 s

less, the piston moves faster, and the transition from compression to suction happens
very fast. So, the turbine speed does not fall to zero and oscillates between 600
and 1000 rpm approximately. However, this rotational speed is less compared to the
design speed of the Wells turbine of similar dimension. This is due to the generator
support, which does not allow uninterrupted airflow to the turbine. For T = 5 s, the
turbine almost stops at the transition from compression to the suction stage.
Figure 7 shows the comparison between unidirectional and bidirectional flow at SL
= 0.4 m and T = 5 s. The piston-cylinder arrangement can be made unidirectional by
opening a valve. In this configuration, the air flows through the turbine only during the
compression stage; whereas, during suction, the air flows through the valve. During
the compression stage, the values for pressure drop and rotational speed are almost
the same for both unidirectional and bidirectional flow. However, for unidirectional
flow, the turbine does not rotate during the suction stage, and the values of pressure
drop and rotational speed remain zero throughout this stage.

6 Conclusion

A preliminary experimental investigation on the performance of monoplane Wells

turbine is carried out for bidirectional air flow, and the results are compared for
two different time periods and stroke lengths of the piston. Due to limited instrument
availability, only the pressure drop and the rotational speed was measured. The turbine
experienced higher pressure drop and higher rotational speed when the stroke length
was high and the time period is low. The generator supports created blockage of
airflow to the turbine, which was the main reason for low rotational speed. Also,
the turbines assembly had a large structural vibration at higher rotational speeds.
So, the results had large deviations from the actual turbine performance. From this
preliminary experience, a modified system is designed, and further tests will be
carried out in future.
962 K. Kumar et al.

Acknowledgements The authors thank the Department of Ocean Engineering, IIT Madras for
providing financial support to carry out the experiment.

References

1. Sannasiraj SA, Sundar V (2016) Assessment of wave energy potential and its harvesting
approach along the Indian coast. Renew Energy 99:398–409
2. Falcao AFDO, Henriques JCC (2016) Oscillating-water-column wave energy converters and
air turbines: a review. Renew Energy 85:1391–1424
3. Raghunathan S, Setoguchi T, Kaneko K (1994) Aerodynamics of monoplane Wells turbine—a
review. Int J Offshore Polar Eng 4(1):68–75
4. Shehata AS, Xiao Q, Saqr KM, Alexander D (2017) Wells turbine for wave energy conversion:
a review. Int J Energy Res 41(1):6–38
5. Gato LMC, Curran R (1996) Performance of the biplane Wells turbine. Trans ASME 118:210–
215
6. Curran R, Gato LMC (1997) The energy conversion performance of several types of Wells
turbine designs. Proc Inst Mech Eng Part A J Power Energy 211(2):133–145
7. Takasaki K, Takao M, Setoguchi T (2014) Effect of blade shape on the performance of Wells
turbine for wave energy conversion. Int J Mech Aerospace, Ind Mechatronics Eng 8(12):2080–
2083
8. Paderi M, Puddu P (2013) Experimental investigation in a Wells turbine under bi-directional
flow. Renew Energy 57:570–576
9. Gato LMC, Webster M (2001) An experimental investigation into the effect of rotor blade
sweep on the performance of the variable-pitch Wells turbine. Proc Inst Mech Eng Part A J
Power Energy, 215(5):611–622
10. George A et al (2018) Experimental analysis of turbine-chamber coupling for wave energy
conversion. Int J Energy Res 42(15):4770–4782
11. Raghunathan S (1995) The Wells air turbine for wave energy conversion. Prog Aerosp Sci
31(4):335–386
12. Inoue M, Kaneko K, Setoguchi T, Raghunathan S (1986) Simulation of starting characteristics
of the Wells turbine. In: 4th fluid Mechanics, plasma dynamics and laser conference, 1986
Investigations into Nonlinear Energy
Sinks for a Stochastic Dynamical
Oscillator

Anuroop Parvathaneni, Dhritimaan Sharma, Dhruv Vashishtha,

Pradeep V. Malaji, and J. Venkatramani

Abstract The paper deals with Nonlinear Energy Sinks (NES), utilizing piezoelec-
tric transduction mechanism, focusing on the degree of effect the auxiliary nonlinear
stiffness has on the performance of the NES and the performance of NES with the
primary system subjected to random excitation. Hence, a parametric sweep of the
auxiliary nonlinear stiffness over a broad range of values has been done and the
variations in primary vibration suppression and voltage generation by the NES have
been observed for its corresponding values. It has been conducted with the NES
attached to a linear primary system and then an essentially nonlinear one. Compari-
son of results and validation of the performance of NES for both the cases have been
performed. Following that, performance of the NES has been investigated when a
linear primary system is subjected to random excitation. Two separate cases have
been utilized to randomize the excitation. Results regarding vibration control and
voltage generated have been derived for both and compared to those obtained for
deterministic excitation. By and large, it is found that NES is successful in protect-
ing a primary system and broadening the operation bandwidth, while satisfyingly
generating voltage, irrespective of the type of excitation on the primary system.

Keywords Nonlinear energy sink (NES) · Energy harvesting · Vibration

suppression

A. Parvathaneni · D. Sharma · D. Vashishtha · J. Venkatramani (B)

Shiv Nadar University, Tehsil Dadri, Greater Noida, U.P., India
e-mail: j.venkatramani@snu.edu.in
P. V. Malaji
BLDEA’s V P Dr. P G Halakatti College of Engineering and Technolog, Vijayapur,
Karnataka 586103, India

© Springer Nature Singapore Pte Ltd. 2021 963

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_84
964 A. Parvathaneni et al.

1 Introduction

Protection of mechanical equipment from excessive vibration is vital as vibrational

damage can severely limit an equipment’s service life. On the other hand, vibration
being such a ubiquitous resource, utilization of vibrational energy for energy har-
vesting can be very useful for small portable equipment or for equipment present in
isolated regions [1, 2]. Conventional energy harvesters have a limitation of harvesting
energy at resonance [3]. Other options consisted of multimodal techniques [3, 4] or a
switch to nonlinear techniques [3, 5]. While tuned Mass Dampers (TMD) and linear
vibration absorbers have been utilized for control of primary structure [6], limitations
such as negligible frequency robustness and poor performance for random vibration
render their utilization an infeasible task [6]. Attachment of auxiliary systems as
energy harvesters for vibration suppression has an advantage of vibration control
and useful energy harvesting [7]. Further investigations into nonlinear alternatives
have led to the development of nonlinear energy sink (NES). While it is successful in
suppressing primary systems’ responses, it can also be coupled with a piezoelectric
vibrational energy harvester (PVEH) [6] to perform simultaneous energy harvesting.
NES has also been found to provide certain unique features such as Strongly Mod-
ulated Response (SMR) [8] and Targeted Energy Transfer (TET) [8]. Such features
allow vibration control in a broadband manner with simultaneous energy harvesting,
making NES a lucrative and more advantageous system when compared to its coun-
terparts. Also, while it has yet not been fully investigated, NES’s ability to increase
operational bandwidth also allows it to effectively suppress primary vibrations, unlike
the linear systems which falter when operated upon by frequency varying excitation
[6].
NES allows passive control of vibration of a primary system and consists of a mass
and viscous damper with essential nonlinearity. Essential nonlinearity refers to an
absence of linear stiffness. The presence of essential nonlinearity or quasi-essentially
nonlinearity [9] is vital as only then can the NES resonate at any frequency and attain
its objective. Quasi-essential nonlinearity refers to a situation in which the linear stiff-
ness of the NES is considerably decreased in comparison to its nonlinear stiffness,
allowing very accurate emulation of essential nonlinearity. It is employed in phys-
ical models as attaining essential nonlinearity is extremely challenging [8]. NES
provides the Targeted Energy Transfer (TET) [8] feature which refers to its ability to
irreversibly transfer energy from the primary system and dissipating it within itself.
However, rather than dissipation, the vibrational energy can be utilized for minor
energy needs. Efforts have been made to convert it into electrical energy and various
transduction models have been introduced to achieve desirable results. By coupling
the NES with a transduction model, the voltage can be generated instead of lim-
iting to only vibration control. Electrostatic, electromagnetic, magnetostrictive and
piezoelectric mechanisms [8] can be utilized with each providing certain advantages
over the other. Most of the study to date has been focused on electromagnetic trans-
ducers [5]. But it has been proposed that piezoelectric vibrational energy harvesters
(PVEH) [5] based on Nonlinear Energy Sinks can also reduce force transmissibility
Investigations into Nonlinear Energy Sinks for a Stochastic … 965

while harvesting energy on a broad frequency scale [9]. It can also be said that it
occupies lesser space while providing better power density [5]. Another area lacking
in investigation has been the performance of an NES for random excitations. Most of
the studies done to date have concentrated on harmonic excitation upon the primary
system which severely limits an NES’s potential to operate satisfyingly on frequency
varying environments while in the meantime, also fails to emulate real-life scenarios
to maximum accuracy. Random vibration has indeed started replacing its harmonic
counterpart when realistic scenarios are concerned and as it is much more capable
of accurately emulating them. Hence, a study is required to understand if an NES
can indeed perform satisfyingly when subjected to random vibrations as only then,
practical application of an NES can be considered on a larger scale.
This work proposes the effect of cubic coefficient of the nonlinear stiffness of the
NES spring on voltage generated by a NES and the structural response of the NES and
the primary spring-mass system. It also investigates NES behaviour for a nonlinear
primary system. The work further investigates into the behaviour of the NES when
the primary system is acted upon by random excitations and analyses the difference
between the voltage generated and structural response of the NES and primary sys-
tem when acted upon by random excitations and when acted upon by deterministic
harmonic excitations. The present study utilizes numerical simulations with the help
of MATLABODE45 function to address these issues. The values of the parameters
used in the study are taken from previous literature [9]. A parametric sweep of the
value of the cubic nonlinear spring stiffness is performed for four different initial
conditions [9]. The interplay and relation between the cubic hardening nonlinear
spring stiffness of the NES and the newly introduced nonlinear spring stiffness of
the primary system are evaluated. The modelling of random vibrations is done by
introducing a deterministic harmonic (sinusoidal) excitation, which depending on
the random variable, was split into two cases to introduce the random vibration,
forcing magnitude or the frequency of forcing. The RMS voltages generated in both
the cases are evaluated and compared with the deterministic harmonic excitation as
well as with the results validated from literature [9]. The time response plots and
time voltage plots hence obtained are presented in this paper.
The paper has been divided into three sections. Section 2 presents the mathematical
model used in this study, followed by results and discussions in Sect. 3. The salient
findings from this study are summarized in Sect. 4.

2 Mathematical Modelling

See Fig. 1.
To study the performance of the piezoelectric vibrational energy harvester (PVEH)
based on the NES, a model is developed consisting of two parts: a primary system
and a PVEH attached to a NES. The governing equations of the linear primary system
with free vibration are given as
966 A. Parvathaneni et al.

y + xp + z y + xp + z

a) b)
ca ca
Piezoelectric k3 Piezoelectric
k3
transducer transducer
k1 R y + xp k1 y + xp
R

cp y cp y
kp kp

Fig. 1 a NES attached to a linear primary system; b NES attached to a nonlinear primary system

   
m p ẍ p + c p ẋ p + k p x p − ca ż + k1 z + k3 z 3 + θ V = 0 (1)

 
m a z̈ + (1 + μ) ca ż + (k1 z + k3 z 3 + θ V ]−μ[c p ẋ p + k p x p ] = 0 (2)

V
+ C S V̇ − θ ż = 0 (3)
R
The governing equations of the nonlinear primary system with free vibration are
given as
   
m p ẍ p + c p ẋ p + k p x 3p − ca ż + k1 z + k3 z 3 + θ V = 0 (4)

 
m a z̈ + (1 + μ) ca ż + (k1 z + k3 z 3 + θ V ]−μ[c p ẋ p + k p x p ] = 0 (5)

V
+ C S V̇ − θ ż = 0 (6)
R
For forced vibration, Eqs. (1) changes to Eq. (7) while Eqs. (2) and (3) are also
utilized
   
m p ẍ p + c p ẋ p + k p x p − ca ż + k1 z + k3 z 3 + θ V = F sin(ωt) (7)

In the governing equations, xa (t) and x p (t) represents the displacement of NES
and the primary mass, respectively. z refers to the relative displacement between the
NES and primary system and is represented as z = xa − x p . m a and m p represent
the mass of the NES and the primary system respectively. ca and c p represent the
damping coefficient of the NES and primary system. k1 and k3 represent the linear
and nonlinear stiffness of the NES spring while k p represents the stiffness of the
primary mass. Mass ratio is denoted by μ = m a /m p . The capacitance and the
Investigations into Nonlinear Energy Sinks for a Stochastic … 967

electromechanical coupling coefficient of the PVEH are represented as C S and θ . R

and V represent the resistive load and voltage.
These above governing equations can also be represented in the state space form,
and numerical simulation is done with the help of MATLAB software using the
ODE45 function to solve the above equations. The initial conditions used are: x p (0)
= X, xa (0) = X, ẋ p (0) = 0, ẋa (0) = 0. X represents the initial displacements given
to the system. Four different values of X are used in the numerical simulation, X =
0.52 mm, X = 1.42 mm, X = 2.65 mm, X = 4.45 mm. Only the first value has been
used when working with the Eqs. (2), (3) and (7) in Sect. 3.3.
Here, the base excitation has been considered to be zero as the base is fixed. The
values of the parameters used in the study are taken from previous literature [9].

3 Results

The focus of this study was to investigate the steady-state performance in addition
to the transient performance of the 2DOF mass-spring system. First, the results of
previous literature were validated [9] which showed that NES can indeed increase
frequency robustness while also decreasing the primary vibrations with the primary
system being subjected to harmonic excitation. It also displayed satisfying results
regarding voltage generation, irrespective of the initial displacement utilized. Hence
the NES was capable of solving the drawbacks present in linear techniques.
Following that, parametric sweep of the auxiliary nonlinear stiffness has been
done for linear and nonlinear primary systems. Logical inferences have been drawn
from the two sets of results and have been compared to understand the difference in
behaviour of NES as per the nature of the primary system. Following that, perfor-
mance of the NES, when the primary system is subjected to random excitation, has
been investigated. Two cases have been used to randomize the excitation. Results for
both voltage generation and primary vibration suppression have been considered. To
understand if NES remains effective for such a case, the results hence obtained have
been compared to the behaviour of the NES when attached to primary system with
deterministic excitation.

3.1 Parametric Sweep of k3 with a Linear Primary System

In this section, Eqs. (1)–(3) have been used. A parametric sweep of k 3 has been
performed as the focus is on understanding if auxiliary nonlinear stiffness plays any
role in voltage generation and primary vibration suppression for a linear primary
system. It can be observed that the value of voltage generated is directly proportional
to the value of displacement of auxiliary mass. Initial displacement (X) or k 3 doesn’t
affect the relation between the two in any way as can be seen in Fig. 2.
968 A. Parvathaneni et al.

Fig. 2 a V versus t plot for X = 0.52 mm; b x versus t plot for X = 0.52 mm; c V versus t plot for
X = 1.42 mm; d x versus t plot for X = 1.42 mm

For X = 0.52 mm, the maximum value of voltage can be obtained for a k 3 value
of 2 × 108 and the maximum displacement of auxiliary mass is also obtained for the
same k 3 value.
For X = 1.42 mm, however, the maximum value of voltage and x a are obtained
at k 3 value of 2 × 107 with a similar observation relation being observed for the two
other initial displacement values 2.65 and 4.45 mm as well. Hence, k 3 doesn’t affect
the interrelation between auxiliary mass displacement and voltage generated in any
way.
For constant k 3 value, greater initial displacements showed higher voltage values
as is observable in Fig. 3. Similar increment in magnitude can be observed for the
other two higher initial displacement values as well.
Also, irrespective of the value of initial displacement, negligible change is
observed in the pattern of voltage generated (with respect to time) when k 3 lies
between 2 × 10 and 2 × 104 . There is a negligible change in magnitude as well.
Variations occur only when k 3 values higher than 2 × 105 are considered. It can be
observed in Fig. 4 where X = 0.52 mm has been considered with similar observations
being made for the other initial displacement values as well.
Investigations into Nonlinear Energy Sinks for a Stochastic … 969

Fig. 3 a V versus t plot for X = 1.42 mm with k 3 = 2 × 106 ; b V versus t plot for X = 0.52 mm
with k 3 = 2 × 106

Fig. 4 a V versus t plot with k 3 = 2 × 107 ; b V versus t plot with k 3 = 2 × 1010 ; c V versus t plot
with k 3 = 2 × 10; d V versus t plot with k 3 = 2 × 104
970 A. Parvathaneni et al.

Fig. 5 a V versus t plot with X = 0.52 mm and k 3 = 2 × 108 ; b V versus t plot with X = 1.42 mm
and k 3 = 2 × 107 ; c V versus t plot with X = 2.65 mm and k 3 = 2 × 106 ; d V versus t plot with X
= 4.45 mm and k 3 = 2 × 106

Noteworthily, no optimum value of nonlinear stiffness could be derived as the

system exhibited maximum voltage for different X values at different k 3 values
(Fig. 5).

3.2 Parametric Sweep of k3 with a Nonlinear Primary System

In this section, Eqs. (4)–(6) have been used. Two cases have been considered with the
value of k 3 being varied while keeping a constant value of k p and vice versa. Its effect
on voltage generation and primary vibration suppression have been investigated.
Here, k p refers to the nonlinear stiffness of the primary system. The primary system
considered is essentially nonlinear in nature as no primary linear stiffness has been
considered.
As can be observed in Fig. 6, for a nonlinear primary system too, primary vibration
has been reduced for both the aforementioned cases. Irrespective of the value of k p
or k 3 considered, with respect to the constant value chosen, primary vibration can be
seen to be dying down.
Investigations into Nonlinear Energy Sinks for a Stochastic … 971

Fig. 6 a x versus t plot with k 3 = 2 × 10 and k p = 4 × 10; b x versus t plot with k 3 = 2 × 10 and
k p = 4 × 1010 ; c x versus t plot with k 3 = 2 × 1010 and k p = 4 × 10; d x versus t plot with k 3 =
2 × 1010 and k p = 4 × 1010

For X = 0.52 mm–

Similar observations have been made for the other three initial displacement values
as well.
Also, a pattern can be observed. Irrespective of k 3 value, the pattern and magnitude
of voltage generated is determined by k p as can be seen in Figs. 7 and 8.
For X = 0.52 mm–
As can be seen, k 3 is irrelevant to the magnitude and pattern of voltage generated.
However, what can be observed throughout is that even for nonlinear primary sys-
tems, voltage is generated by NES and hence, NES can be concluded to be useful
for such systems as well.

3.3 Random Vibrations

The modelling of random vibrations was done by introducing a deterministic har-

monic (sinusoidal) excitation, which depending on the random variable, was split
into two cases to introduce the random vibration, forcing magnitude F or the fre-
quency of forcing ω. The modelling of random vibration was done with the help
972 A. Parvathaneni et al.

Fig. 7 a V versus t plot for k 3 = 2 × 10 for k p = 4 × 10; b V versus t plot for k 3 = 2 × 1010 for
k p = 4 × 10

Fig. 8 a V versus t plot for k 3 = 2 × 10 for k p = 4 × 104 ; b V versus t plot for k 3 = 2 × 10 for
k p = 4 × 1010

of rand() function in MATLAB which generates a random number between 0 and

1. The two cases are (1) Random forcing, where F is the random variable in the
harmonic excitation and F = F mean + σ *random and (2) Random frequency, where
w is the random variable in the harmonic excitation and ω = ωmean + σ *random.
In this formula, σ , represents the noise intensity and random represents the random
number generated by the MATLAB function rand().
Using the governing Eqs. (2), (3) and (7) and values considered from previous
literature [9], one case of initial displacement X = 0.52 mm was considered. Arbitrary
values of F, F mean , ω, ωmean are considered in both the cases and σ is varied from
0.1 to 0.9 in order to study the effect of noise intensity on the voltage generated and
displacements of the NES and primary system. Values of parameters used in case 1
are F mean = 1, ω = 5,σ varying from 0.1 to 0.9 and in case 2 is F = 1, ωmean = 5,
σ varying from 0.1 to 0.9. Using the same F and ω values, the comparison between
Investigations into Nonlinear Energy Sinks for a Stochastic … 973

Fig. 9 Simulation results of voltage generated in case 1 from PVEH with random and deterministic
excitation for F mean = 1, ω = 5 and X = 0.52 mm: a for σ = 0.1; b for σ = 0.5; c for σ = 0.7 and
d for σ = 0.9

the voltage generated by both the deterministic harmonic excitation and the random
excitation in both the cases for different σ values is presented in Figs. 9 and 10. In
case 1, random F, the difference in magnitude between the random and deterministic
voltage generated increases with an increase in σ as shown in Fig. 9. In case 2,
random ω, the voltage generated dies down faster with increase in σ as shown in
Fig. 10. The RMS voltages generated in both the cases are evaluated and compared
with the deterministic harmonic excitation as well as with the results validated from
literature [9]. This is represented in Tables 1 and 2. The RMS voltage generated
increases with increase in σ in case 1 while it decreases with increase in σ in case 2
as shown in Tables 1 and 2.
The structural response of the system is shown in Figs. 11 and 12. It can be
observed in Figs. 11 and 12, that NES displacement is more than the displacement
of the primary mass, and even at higher values of σ the NES protects the primary
system from excessive vibrations in both the random cases as shown in Figs. 11 and
12.
974 A. Parvathaneni et al.

Fig. 10 Simulation results of voltage generated in case 2 from PVEH with random and deterministic
excitation for F mean = 1, ω = 5 and X = 0.52 mm: a for σ = 0.1; b for σ = 0.5; c for σ = 0.7 and
d for σ = 0.9

Table 1 Comparison of
σ value RMS voltage RMS voltage RMS voltage
RMS voltage generated in
(validated) (deterministic) (random)
case 1: random F
σ = 0.1 0.2077 1.0093 1.0414
σ = 0.3 0.2077 1.0093 1.0896
σ = 0.5 0.2077 1.0093 1.1470
σ = 0.7 0.2077 1.0093 1.1945
σ = 0.9 0.2077 1.0093 1.2582

Table 2 Comparison of
σ value RMS voltage RMS voltage RMS voltage
RMS voltage generated in
(validated) (deterministic) (random)
case 2: random ω
σ = 0.1 0.2077 1.0093 0.8883
σ = 0.3 0.2077 1.0093 0.5858
σ = 0.5 0.2077 1.0093 0.4919
σ = 0.7 0.2077 1.0093 0.4221
σ = 0.9 0.2077 1.0093 0.3953
Investigations into Nonlinear Energy Sinks for a Stochastic … 975

Fig. 11 Simulation results of the structural response of the system for case 1 with random and
deterministic excitation for Fmean = 1, ω = 5 and X = 0.52 mm: a for σ = 0.7 and b for σ = 0.9

Fig. 12 Simulation results of the structural response of the system for case 2 with random and
deterministic excitation for F = 1, ωmean = 5 and X = 0.52 mm: a for σ = 0.7 and b for σ = 0.9

4 Conclusion

Based on the NES principle, a piezoelectric harvesting device has been modelled.
Validation of previous literature [9] was done. The transient response and the steady-
state response of the system, regarding energy harvesting and vibration suppression,
were analyzed using numerical simulations. Investigations into effect of auxiliary
nonlinearity when connected to a primary system with deterministic harmonic exci-
tation were made. All in all, the NES successfully suppressed primary vibration and
continued to generate voltage over a wider operational bandwidth, regardless of the
primary system being linear or nonlinear. The degree of effect auxiliary nonlinearity
has on the variation in voltage generation, voltage pattern and its magnitude have
been understood. Random excitation to the system is also modelled and the struc-
tural response as well as the voltage generated have been presented. It was observed
that NES continues to suppress primary vibration just as in deterministic forcing but
976 A. Parvathaneni et al.

sustained voltage generation has been observed, in contrast to that for determinis-
tic model which showed decreasing voltage. Investigations into the effect of noise
intensity (σ ) showed that it is directly proportional to the rms voltage generated for
random F and indirectly proportional to random ω.

References

1. Ertuk A, Inman DJ (2011) Broadband piezoelectric power generation on high-energy orbits of

the bistable duffing oscillator with electromechanical coupling. J Sound Vib 330:2339–2353
2. Liuyang X, Tang L, Liu K, R. Mace B., On the use of piezoelectric nonlinear energy sink for
vibration isolation and energy harvesting. In: Conference on smart materials, adaptive structures
and intelligent systems, pp 1–6 (2018)
3. Malaji PV, Ali SF (2018) Analysis and experiment of magneto-mechanically coupled harvesters.
Mech Syst Signal Process 108:304–316
4. Rajarathinam M, Ali S (2018) Energy generation in a hybrid harvester under harmonic excitation.
Energy Convers Manage 155:10–19
5. Kumar KA, Ali SF, Arockiarajan A (2017) Magneto-elastic oscillator: modeling and analysis
with nonlinear magnetic interaction. J Sound Vib 393:265–284
6. Xiong L, Tang L, Liu K, Mace B (2018) Broadband piezoelectric vibration energy harvesting
using a nonlinear energy sink. J Phys D Appl Phys 51:1–2
7. Malaji PV, Rajarathinam M, Jaiswal V, Ali SF, Howard IM (2019) Energy harvesting from
dynamic vibration pendulum absorber. Recent Adv Struct Eng 2:467–478. Springer Singapore
8. Kremer D, Liu K (2014) A nonlinear energy sink with an energy harvester: transient responses.
J Sound Vib 333:4859–4861
9. Zhang Y, Tang L, Liu K (2017) Piezoelectric energy harvesting with a nonlinear energy sink. J
Intell Mater Syst Struct 28(3):307–322
A Novel Banana Leaf Waste-Based
Activated Carbon for Automobile
Emission Control

A. John Presin Kumar, S. Sivakumar, R. Balaji, and Mukesh Nadarajan

Abstract Recycling of plant and vegetable wastes is very much needed for improv-
ing the economy and minimizing wastages of materials. This can be a very beneficial
act toward the progress of society overall. Environmental pollution issues especially
from automobiles are also much decreased by this. Activated carbon is better known
for its characteristics and applications in the areas of gas absorption, water purifica-
tion and as composite reinforcement fillers. In this current research work, an attempt
is made to economically synthesize activated carbon from dried banana leaf waste
taken from a local vegetable market. This is characterized by measuring moisture
content, ash content, bulk density, yield of charcoal, fixed carbon and hardness. Gas
adsorption test is conducted using Atlon gas analyzer for various engine speeds. Its
effect on being porous is analyzed by means of comparing its performance with the
solid ones. Recycled banana leaves are used as the natural resource, and heat activa-
tion is used for converting the so-produced carbon into activated carbon. Potassium
hydroxide is used for synthesizing porous activated carbon. Comparison is made
between solid and porous activated carbons for better adsorbing capabilities for
automobiles. Scanning electron microscopy was done to analyze the morphology of
particles.

Keywords Activated carbon · Potassium hydroxide · Porous structure · Fixed

carbon · Bulk density

A. John Presin Kumar (B) · S. Sivakumar · R. Balaji · M. Nadarajan

Department of Mechanical Engineering, Hindustan Institute of Technology & Science,
Padur, Kelambakkam, Chennai, Tamil Nadu 603103, India
e-mail: johnpk@hindustanuniv.ac.in
S. Sivakumar
e-mail: ssivak@hindustanuniv.ac.in
R. Balaji
e-mail: sridharabalaa@gmail.com
M. Nadarajan
e-mail: revamukesh@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 977

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_85
978 A. John Presin Kumar et al.

1 Introduction

Activated carbon is a type of carbon processed to have minute, low-volume pores.

These pores can increase the surface area of activated carbon particles for better
adsorption capability. Activated carbon particle is shown in Fig. 1. Thus due to its
high degree of microporosity, one gram of activated carbon has a greater surface area.
The amount of activation level may be varied. A desired activation level sufficient
for useful application may be attained from heating.
Many varieties of applications are possible for activated carbons. Such type of
industrial applications involves the use of activated carbon as gas absorbents, in metal
finishing for purification of electroplating solutions and as fillers in composites. Acti-
vated carbon technique is a main purification technique for removing impurities from
nickel plating solutions. Activated carbon can also be effectively used as reinforce-
ment fillers in fabricating various composites such as wood or polymer or metal
composites [1, 2]. Though there are many advantages of activated carbons, they also
have few drawbacks. For instance, the price of activated carbons is considerably
high due to immense energy cost during their synthesis. Therefore, it is always wel-
come to create an economical method of producing activated carbon based on natural
resources.
Every day, an average of over 5000 useful products are developed from plant
wastes in our country. Banana trees always play a big role in our personal life as well
as domestic life. The Indian government in earlier established a target to raise the
forested area to one-third of India’s land area. But increased pressure from humans
stopped this initiative. Examples of these are harvesting wood for fuel and other uses,
and clearing the land for agriculture, and others have prevented the government from
attaining its objective. Wood recycling is the process of converting waste timber into
new and usable products. This is adopted in many countries along the globe. Through
recycling of wood only, we have developed re-useable products, artificial wood, etc.,
which are environment friendly. Banana leaves after their usage in banana mandis
(market) go simply as scrap or wastes as shown in Fig. 1. Dried banana leaves are
a good source of carbonaceous material. Figure 2 shows a useful product such as a
bin, made out of banana leaves (Fig. 3).

Fig. 1 Activated carbon powder

A Novel Banana Leaf Waste-Based Activated Carbon for Automobile … 979

Fig. 2 Dried banana leaf

Fig. 3 A bin made of

banana leaf

The literature survey indicates that activated carbons can be prepared from a
variety of raw materials such as coconut shell, rice husk, walnut shell, banana stem
and animal bones [3–5]. The aim of this paper is to synthesize activated carbons
from dried banana leaves by heat and chemical activation and to investigate their
important characteristics, especially as an efficient gas adsorbing agent. Two forms of
activated carbon were compared for characterization like solid and porous structures.
Porous structure was achieved by using potassium hydroxide KOH during chemical
activation. Figure 4 shows the overall main steps involved in this current work.
980 A. John Presin Kumar et al.

Fig. 4 Steps involved in the methodology

2 Preparation of Activated Carbon

Figure 1 shows the steps involved in the methodology clearly. The dried banana
leaf samples were collected from a local banana and vegetable market in Chrompet,
Chennai, Tamil Nadu, India. These were crushed into small pieces of about 2- to
3-mm diameter for providing large surface area. Then, these particles are sieved to
get rid of dust particles and other impurities. These crushed pieces were put in an
oven, and the temperature was maintained at 105 °C and soaked for 24 h. The dried
samples were taken in airtight plastic bags to avoid re-absorption of moisture from
the atmospheric air. About 103 g of samples was taken in a crucible with tightly
closed tightened lid for ensuring airtight environment. The crucible is placed in the
furnace carefully. The furnace was switched on, heated to a temperature of 460 °C
and held constant for 30 min. The charcoal obtained from pyrolysis process was kept
in an airtight bag in the activation unit, and the furnace was heated to a temperature
of 750 °C. Steam was allowed to flow to it for 30 min by keeping the temperature
constant. Activated carbon thus obtained has very less porosity which is ignorable
and thus considered solid activated carbon particles.
In order to produce porous structured activated carbon [2], chemical activation
was performed using KOH at various concentrations like 0.0, 0.3, 0.5 and 0.7 M
preferably at 0.5 M. All samples were carbonized at a temperature of 600 °C using
a furnace followed by physical activation using CO2 gas at a temperature of 850 °C
which produces porous activated carbon. Thus, both solid and porous activated carbon
particles are obtained. These different forms of activated carbon were compared later
for gas adsorption performances.
A Novel Banana Leaf Waste-Based Activated Carbon for Automobile … 981

3 Characterization of the Activated Carbon

The activated carbon produced was characterized for its moisture content, ash con-
tent, bulk density, yield of charcoal, fixed carbon, hardness and gas adsorption
capability.

3.1 Moisture Content (MC)

Moisture content of activated carbon was calculated by finding the difference in

weight of sample before and after drying. The weight of the empty crucible was
noted down first. Then, total weight of sample and crucible was noted as W s . The
sample was dried in an oven at 105 °C for a time period of 24 h, and then it was
allowed to cool down. The sample along with crucible was weighed again W f , and
the value of ((Ws − Wf ∗ 100))/(Ws − We ) gives the percent of moisture content [1].

3.2 Ash Content (AC)

Dried sample of activated carbon was weighed and recorded as W d . These were
placed in a crucible, and the total weight of crucible plus dry sample was noted and
recorded as W s . Finally, it was placed in a furnace at 500 °C for a time period of 6 h
[1]. It was removed and its content reweighed and recorded W f . These steps were
repeated twice for all the activated carbon samples, and the estimated value of the
ash content was calculated as: ((Ws − Wf ∗ 100))/(Wd ).

3.3 Bulk Density

For calculating this, the activated carbon sample was transferred into a cylinder of
10 ml capacity and the weight of the sample along with cylinder was recorded as
W s [2]. The estimated value of the bulk density was calculated as: weight of sample
(W s )/weight of equal volume of water (W w ).

3.4 Yield of Charcoal (YC)

The yield of charcoal [1] that is available in the activated carbon was estimated as:

WS − AC + MC ∗ 100/WS
982 A. John Presin Kumar et al.

where WS is the original weight of dry sample, AC is the ash content and MC is the
moisture content.

3.5 Fixed Carbon (Fc)

The fixed carbon content in each activated carbon sample was estimated [1] using:

YC − AC − MC ∗ 100/YC

where YC is the yield of charcoal, AC is the ash content and MC is the moisture
content.

3.6 Hardness (H)

This involves the following steps. First, 2 g of activated carbon sample was measured
in a standard sieve of size 40. The sieve was made to shake along with particles for a
time period of 1 h using mechanical shaker at 160 rpm with four steel balls of 6.00-
mm diameter inside the sieve [1]. Finally, the quantity of activated carbon retained
inside the sieve after 1 h was expressed in percentage as hardness.

3.7 Gas Adsorption

To obtain the gas adsorption characteristics of activated carbon, batch adsorption

experiments were conducted [4]. This involves taking into account the effects of the
following parameters like initial concentration, contact time, volume and weight of
adsorbent. The adsorbed amounts of the solution are given by the equation [4] as

Q = (CO − CT) ∗ V /W

where CO is the initial concentration of solution, CT is the final concentration of

solution at time T, V is the volume of solution and W is the weight of adsorbent used.

3.8 Scanning Electron Microscopy

The surface morphology was obtained using scanning electron microscopy (SEM).
A Novel Banana Leaf Waste-Based Activated Carbon for Automobile … 983

3.9 Engine and Gas Analyzer

This current research work was conducted in our institute Hindustan Institute of
Technology and Science, Chennai. Figure 5 shows single cylinder water cooled
naturally aspirated compression ignition engine having maximum horse power (hp)
of 10 hp. The output shaft is connected to an electrical dynamometer which has
maximum capacity of 24 amps. The engine used a mechanical injector, and it is a
constant engine which is designed to run at 1500 rpm [6, 7].
The engine specification is given in Table 1.
Atlon 5 gas analyzer shown in Fig. 6 was used to measure the emission parameter

Fig. 5 Kirloskar DM 10 engine

Table 1 Engine specification

Make and model Kirloskar DM 10
Bore/stroke 102 * 116 mm
Speed 1500
B.H.P 10 hp
Swept volume 661 cc
Compression ratio range 17:5:1
Fuel Diesel
Torque 0.048 KN m
984 A. John Presin Kumar et al.

Fig. 6 Atlon gas analyzer

Table 2 Gas analyzer specification

S. No. Make AVI
1 Model DIGAS 444
2 Display LCD
3 Interface RS 232
4 Power consumption 25 W
5 Operating voltage 230 V 55 Hz

of the exhaust gas. This gas analyzer is capable of measuring 5 parameters: carbon
monoxide (CO), carbon dioxide (CO2 ), hydrocarbon (HC), oxides of nitrogen (NOx )
and oxygen (O2 ) whose specification is given in Table 2.

4 Results and Discussion

The technical analyses of 5 samples are shown in Fig. 3. It should be noted that
all the samples were representative of activated carbon particles produced under
similar conditions [5, 8–11], namely AC1–AC5. These samples were categorized
based on the different combinational values of temperature and soaking duration.
Temperatures were taken like 105, 110, 115, 120 and 125 °C, and soaking times
were taken like 30, 24, 20, 15 and 10 h. The properties were found to be showing
A Novel Banana Leaf Waste-Based Activated Carbon for Automobile … 985

similar values in all the samples indicating isotropic nature [12, 13]. These are shown
in Tables 1 and 2 for comparing the solid and porous samples, respectively.
Table 3 clearly shows various characteristics of activated carbon without porosity
inside them. Being solid particles, these activated carbon particles may have lesser
surface area compared to the porous ones.
From Table 4 values for solid activated carbon, it was clear that the sample AC3
showed promising results indicating the better preparation temperature of 115 °C
and an optimum soaking time of 20 h for all properties other than moisture and ash
removal.
By comparing the two tables, the values of hardness, bulk density and amount
of carbon were slightly better than that of solid activated carbon. Similarly, from
Table 4 values for porous activated carbon, it was clear that the sample AC3 showed
promising results indicating the better preparation temperature of 115 °C and an
optimum soaking time of 20 h for all properties other than moisture and ash removal.
The adsorption of total unburnt gases was first analyzed using the Atlon gas ana-
lyzer, and comparison is made among various performance curves drawn for vary-
ing engine speeds against activated carbon samples prepared at different activation
temperatures. This performance curve is illustrated in Fig. 7. It was found that the
performance was better for full speed of the engine (1500 rpm) and also at a higher
activation temperature of 810 °C. The temperature for 15 different samples chosen
was varying at 30 °C from its predecessor.
The adsorption capacity toward NOx , CO2 and HC was analyzed for various acti-
vated carbon samples. These samples indicate the different temperature and soaking

Table 3 Characteristics of solid activated carbon

Sample No. Moisture Ash content % Bulk density Fixed carbon Hardness
content % g/cm3 %
AC1 11.2 12.2 0.775 88.2 7.34
AC2 11.8 12.6 0.728 88.7 7.11
AC3 10.8 12.1 0.639 89.1 7.56
AC4 10.3 11.9 0.788 88.9 7.23
AC5 11.9 12.3 0.774 89.0 7.44

Table 4 Characteristics of porous activated carbon

Sample No. Moisture Ash content % Bulk density Fixed carbon Hardness
content % g/cm3 %
AC1 10.9 11.3 0.776 89.3 7.35
AC2 10.8 11.6 0.738 89.7 7.31
AC3 10.5 11.1 0.798 89.9 7.66
AC4 10.2 10.9 0.791 88.2 7.25
AC5 10.9 11.4 0.771 89.1 7.64
986 A. John Presin Kumar et al.

Fig. 7 Gas adsorption comparison for different activation temperatures

time durations in oven. The sample showing better gas adsorption indicated clearly
the appropriate temperature and soaking time duration which were optimum for
preparing activated carbon.
The following graph in Figs. 8 and 9 for solid and porous activated carbon, respec-
tively, showed the performance characteristic of activated carbon samples from AC1
to AC5 in observing the harmful gases. For each sample, the average values of 5
trials were plotted. The sample AC3 showed promising characteristic in adsorbing
all three types of gases (Table 5).

Fig. 8 Gas adsorption curves for solid activated carbon

A Novel Banana Leaf Waste-Based Activated Carbon for Automobile … 987

Fig. 9 Gas adsorption curves for porous activated carbon

Table 5 Comparison of gas

Solid activated Porous activated
adsorption %
carbon carbon
% of gas 66 70
adsorption

The effect of higher to lower loading on the gas adsorption performance is shown
as various trials in Fig. 10. Porous activated carbon gave better results in all trials
compared to the solid ones having the values of 68%, 55% and 48%, respectively,
whereas the solid ones were having 65%, 53% and 42%, respectively, for trial 1, trial
2 and trial 3, respectively. The effect of chemical and heat activation on the surface
morphology of activated carbon particles from banana true stems is clearly shown
in Fig. 11. SEM micrographs show that the samples prepared have more pores. The

Fig. 10 Overall gas adsorption performance

988 A. John Presin Kumar et al.

Fig. 11 SEM micrographs of activated carbon particles

sample contains fiber structures such as an agglomeration of fibers. Inclusion of

more pores and agglomeration are important and needed properties for successfully
activated carbon particles.

5 Conclusion

• The effectiveness of dried banana leaves in providing activated carbon was ana-
lyzed in this work by means of comparing the performance among two of its forms
such as solid and porous ones. It was found initially in an overall performance
observation made for different activated carbon samples prepared at different acti-
vation temperatures that the gas adsorption was better for full speed of the engine
(1500 rpm) and also at a higher activation temperature of 810 °C. The temperature
for 15 different samples chosen was varying at 30 °C from its predecessor.
• In the case of solid ones, the sample AC3 had removal efficiency of gases
greater than 66% upon the completion of five consecutive adsorption cycles
corresponding to the oven temperature of 115 °C and 20 h of soaking time.
• The same sample showed promising performance in other properties like hardness,
fixed carbon and bulk density also whereas the moisture removal and ash removal
had been better for the other sample AC4 corresponding to the oven temperature
of 120 °C and 15 h of soaking time duration.
• In the case of porous ones, the sample AC3 had removal efficiency of gases
greater than 70% upon the completion of five consecutive adsorption cycles
corresponding to the oven temperature of 115 °C and 20 h of soaking time.
• The same sample showed promising performance in other properties like hardness,
fixed carbon, bulk density.
• Here also, the moisture removal and ash removal have been better for the other
sample AC4 corresponding to the oven temperature of 120 °C and 15 h of soaking
time duration.
A Novel Banana Leaf Waste-Based Activated Carbon for Automobile … 989

• Dried banana leaves had shown good ability to be a better resource for activated
carbon synthesis. The porosity providing effect of potassium hydroxide solution
was also very clear from the results.

Acknowledgements We would like to express our sincere gratitude and thanks to Dr. A. Velayud-
ham, Scientist ‘G’, Ministry of Defence, CVRDE, Avadi, Chennai, India, and Dr. Ashish Selokar,
Accendere, CL Educate Ltd., New Delhi, India, for their valuable suggestions. We express our
sincere gratitude and thanks for their helps in the whole duration of executing this research work.

References

1. Sunday NJ, Okechukwu NS, Elom N, Anthony O, Michael PT (2018) Quantitative charac-
terization of activated carbon from cow, donkey, chicken and horse bones from Ezzangbo in
Ebonyi State, Nigeria. Amer J Appl Chem Sci Publ Group 6(5):169–174
2. Taer E, Taslim R, Mustika WS, Kurniasih B, Agustino, Afrianda, Apriwandi A (2018) Int J
Electrochem Sci 13:8428–8439
3. Lebedeva MV, Ayupov AB, Yeletsky PM, Parmon VN (2018) Rice husk derived activated
carbon/polyaniline composites as active materials for supercapacitors. Int J Electrochem Sci
13:3674–3690
4. Haji Azaman SA, Afandi A, Hameed BH, Mohd Din AT (2018) J Appl Sci Eng 21(3):317–330
5. Subramani T, Revathi PK (2015) Production of activated carbon from agricultural raw waste.
IOSR JEng (IOSRJEN) 05(05):54–63
6. Udhaya Kumar V, Bharath P, John Presin Kumar A (2017) Experimental Investigation of per-
formance and emission characteristics of ethanol blended diesel with zinc oxide nanoparticles.
Int J Eng Tech 9(3):1524–1531
7. John Presin Kumar A, Bharath V, Balakumar S, Godwin John J (2014) A study on the effect
of the dual fuel mixture of diesel and LPG, CNG gases on diesel engines. Int J Ind Prod Eng
Tech 4(1):11–13
8. Aziz NAA, Mohamed NAA, Mohamad M, Amini MHM, Abdul Aziz MS, Yusoff H, Ismael
Rizman Z (2015) ARPN J. Eng Appl Sci 10(1):376–386
9. Chandramohan D, John Presin Kumar A (2017) Experimental data on the properties of natural
fiber particle reinforced polymer composite material. Elsevier, Data Brief 13:460–468
10. Chandramohan D, John Presin Kumar A (2017) Data-enabled discover. Appl, Cross Mark,
Springer Inter Publ 1–9
11. Tsyntsarski B, Marinov S, Budinova T, FerhatYardim M, Petrov N (2013) Synthesis and
characterization of activated carbon from natural asphaltites. Fuel Processing Tech. Elsevier,
116:346–349
12. Abdul Khalil HPS, Noriman NZ, Ahmad MN, Ratnam MM, NikFuaad NA (2007) J Reinf Plast
Comp 26(3):305–320
13. Flores A, Cagiao ME, Ezquerra TA, BaltaCalleja FJ (2001) J. Appl. Pol. Sci. 79:90–95
14. SubhashreePradhan, Project report, B. Tech-Chem Engineering, Department of Chemical
Engineering, NIT, Rourkela, pp 1–9 (2010)
15. Abubakar M, Alechenu AA, Manase A, Mohammed J (2012) Adv Appl Sci Res 3(5):3089–3096
16. Cobb A, Warms M, Maurer EP, Chies S (2012) IntJ Serv Learn Eng 7(1):93–104
17. Aziz NAA, Mohamed M, Mohamad M, Amini MHM, Aziz MA, Yusoff H, Rizman ZI (2015)
ARPN J Eng Appl Sci 10(1):376–386
Analyzing Different Methods to Increase
the Natural Period of a Compact Wave
Energy Converter

Vishnu Vijayasankar and Abdus Samad

Abstract In this article, we present the numerical analysis performed on a compact

mechanical-based direct drive wave energy converter (WEC), in order to improve its
performance on longer wave periods. For this work, the numerical modeling of the
point absorber is done with the help of an open-source code written in MATLAB,
WEC-Sim and the hydrodynamic parameters of the WEC are determined using the
BEM code AQWA. From the past experience, it is a known fact that the range of
frequencies for which a point absorber can perform satisfactorily is less if they are
small-sized. In this work, we make an effort to increase the operational bandwidth
of a point absorber and to match the natural period of the system with the period of
sea waves, without increasing its size (diameter). During the process, the variation
in different hydrodynamic coefficients and performance parameters, triggered by
the proposed design alteration is studied. In the end, it has been concluded that the
proposed modification was able to improve the operational bandwidth of the device
and the resonant frequency of the system was reduced, as expected.

Keywords Point absorber · Direct drive · Wave energy converter · BEM

1 Introduction

Traditional wave energy converter (WEC) systems like oscillating water columns
(OWC) use a turbine rotating at a high speed to generate electricity [1–3]. However,
since the frequency of oscillations imparted by ocean waves are particularly low, the
complex processes involved in compressing the fluid to attain such a high velocity
proves to be extremely inefficient and less feasible especially if the power requirement
is less or moderate, as in the case of a self-powered navigational or data buoy. A

V. Vijayasankar · A. Samad (B)

Wave Energy and Fluids Engineering Lab, Department of Ocean Engineering, Indian Institute of
Technology Madras, Chennai, India
e-mail: samad@iitm.ac.in
V. Vijayasankar
e-mail: vishnuvijayasankar@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 991

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_86
992 V. Vijayasankar and A. Samad

substitute for these systems would be direct drive systems [4]. The mode of power
conversion in a direct drive system can be mechanical based (using gears) or electrical
based (using linear generators). Even though in the case of a linear generator, there
are only fewer moving parts, the complexity involved in handling a voltage which
varies both in amplitude and frequency makes it difficult to implement.
In the paper, we consider a mechanical-based direct drive wave energy converter
that uses a power take-off (PTO) similar to the design proposed by Ai et al. [5]. The
aforementioned PTO is coupled to a heaving buoy so that together they will perform
like a point absorber WEC. These types of WECs are capable of extracting energy
from a wave of crest length greater than its width [6]. A typical point absorber com-
prises a buoy whose width is less compared to the incident wavelength. In the present
work, the main focus is to improve the performance of such a point absorber WEC
in long period waves without increasing its size or the cost of manufacturing. This
study uses the Boundary Element Method (BEM) code Ansys AQWA to determine
the hydrodynamic coefficient for different test cases which subsequently is fed to an
open-source MATLAB based code that is tailor-made for simulating wave energy
converters (WEC-Sim) [7] to analyze their response in simulated wave conditions.

2 Hydrodynamics of Buoys

In the case of a point absorber, it is convenient to assume the motion of the buoy is
in the heave direction only [8]. Let us consider the buoy makes a displacement z in
the heave direction then, its equation of motion can be written as [9]:

m z̈ = Fe + Fr + Frs + Fdamp (1)

where m is the structural mass of the system, F e is the excitation force imparted to
the buoy by the incoming wave, F r is the wave radiation force, F rs is the hydrostatic
restoring force and F damp is the hydrodynamic damping force.
The excitation force is composed of two different force components, the diffrac-
tion force (F d ) and the Froude–Krylov force (F fk ) [10]. The radiation force consists
of radiation damping and added mass terms.
 
Fr = − −ω2 A(ω) + jωbd (ω) z(ω) (2)

where, ω is the angular frequency, A(ω) is the frequency-dependent added mass and
bd (ω) is the frequency-dependent radiation damping coefficient. The normalized
value of added mass and radiation damping can be obtained as:

A
A (ω) = (3)
ρ
Analyzing Different Methods to Increase the Natural Period … 993

bd
B  (ω) = (4)
ρω

where ρ is the density of the fluid.

2.1 Hydrodynamic Resonance

Resonance occurs when a system is forced to oscillate at a frequency near to its natural
frequency. It is desirable for a wave energy converter to operate near resonance in
order to extract more power [11]. The resonance of a floating buoy occurs for an
angular frequency;

ρgπrb2 + k
ωn = (5)
a+m
2π
Tn = (6)
T
where rb the radius of the buoy, and m is the mass of the buoy.
In many coastal areas like Chennai, the periods of the most percentage of the
waves lies in the range of 4–9 s. So, for any point absorber to work efficiently in those
conditions, their natural period should lie somewhere near the same range. However,
small-sized bodies usually possess a natural period lower than that. From Eqs. (5) and
(6), it is evident that the natural period varies with the structural mass, added mass,
and the buoy radius. Since the added mass varies as the cube of diameter [12], the
natural period can be increased by increasing the diameter of the buoy. However, this
idea contradicts the main objective of this work, i.e. to propose a compact low-cost
design to increase the natural period of the buoy. This leaves us with two options,
either to increase the structural mass of the buoy or to increase the added mass of the
buoy by some other means.

2.2 Theoretical Added Mass

Newman [12] quoted that an approximate value of the added mass formed by a
semi-infinite cylinder with radius r can be found by the formula a = 2.064 ρr 3 .
In 1966, Sarpkaya [13] observed that the added mass formed by a circular plate is
approximately equal to the mass of water in an imaginary sphere enclosing the plate.
Based on these two observations, the following equations were derived to find the
added mass of the system of a floating cylindrical buoy with and without a heave
plate.
994 V. Vijayasankar and A. Samad

Fig. 1 Theoretical added

mass of the system

Added mass

Heave plate

The theoretical added mass of the buoy without heave plate:

 
2 3
A (ω) = ρπ R − r02 Rb (7)
3 b

The theoretical added mass of the buoy without heave plate:

   
8 3 2 3
Ab (ω) = ρ r − 2π R0 r p + π
2
R − r 0 Rb
2
(8)
3 p 3 b

whereRb is the radius of the buoy, r p is the radius of the heave plate and r 0 and R0
are the radii of holes in the buoy and heave plate respectively (Fig. 1).

2.3 Wave Power

Wave power is the power content of the incoming ocean wave per crest lest length
of the wave. Wave power depends on the wave height and wave period and can be
mathematically represented as:

ρg 2 T H 2
Pwave = (9)
32π

where g is the acceleration due to gravity (9.81 m/s2 ), T is the wave period and H is
the wave height (Fig. 2).
Analyzing Different Methods to Increase the Natural Period … 995

Fig. 2 CAD model a Simple buoy b buoy with heave plate

Table 1 Buoy specifications

Buoy specifications Value
Height (cm) 40
Outer diameter (cm) 60
Inner diameter (cm) 15
Material SS

3 Methodology

3.1 Geometry and Meshing

Hydrodynamic software packages such as the ANSYS AQWA provide a modeling

platform sufficient for creating basic shapes and devices. The geometry is discretized
using the panel meshes generated by AQWA. Figure 2a, b shows the geometry of
two different buoy configurations and their corresponding meshes which are used
for the present study.
The buoy dimensions are fixed during the entire study and its details are given in
Table 1.
The mass properties of the geometry like the moment of inertias and the center
of mass can is obtained from SOLIDWORKS. These mass properties have to be fed
to the AQWA solver before starting the analysis.

3.2 Finding Hydrodynamic Coefficients

The surface model generated by the design modeler is analyzed by the AQWA solver
to determine the hydrodynamic coefficients. During the initial phase of analysis, the
mass properties of the geometry, the range of required wave frequencies and wave
directions are to be specified. For the present work, 99 different wave frequencies
with directions ranging from −180° to + 180° were chosen.
996 V. Vijayasankar and A. Samad

3.3 WEC-Sim Code

The response of the buoy for various wave conditions be studied using the WEC-Sim
code. WEC-Sim requires four types of inputs from the user
1. Hydrodynamic data generated by the AQWA.
2. Geometry files of the body.
3. Wave conditions like regular/irregular, wave height, wave period, etc.
4. A Simulink model of the hydrodynamic system.

3.4 Response Amplitude Operator

The Response Amplitude Operator of a body in any direction is the ratio of its motion
amplitude in that direction to the incoming wave amplitude. A floating body can have
RAOs for heave, pitch, and roll motions. In the process of designing a point absorber,
the RAO in heave direction plays a significant role to optimize the power absorption.
RAO for heave motion can be expressed as:

ζa
RAO = (10)
ηa

where, ζa is the heave amplitude of the buoy and ηa is the incoming wave amplitude.

3.5 Test Cases

In the present work, the buoy is subjected to two main test cases; the first one in
which the experiments are done for different masses of the buoy and the second
one in which the experiments are done for the buoy attached with heave plates of
different diameters. Variation in mass refers to increasing or reducing the mass of
buoy without changing its dimension. For testing the effect of mass on the resonance,
three masses (30, 40 and 50 kg) were chosen and for testing the effect of heave plate
diameter, three different diameters (30, 60, and 90 cm) were used. One important
point to be noted is that the AQWA solves the potential flow theory and therefore
does not take into account the drag forces. So, in order to get the exact results, one
has to find the damping coefficients due to drag experimentally and manually include
it in the AQWA code. However, experimentation is out of the scope of the present
work.
Analyzing Different Methods to Increase the Natural Period … 997

4 Results and Discussions

The results of all the test cases developed for the buoy are discussed in this section.
As mentioned in the previous sections, the numerical experiments are first carried
out for different buoy masses without the heave plate and then by connecting heave
plates of different diameters to the buoy after fixing the buoy mass.

4.1 Mesh Dependency

Mesh dependency study is done to find an optimum mesh resolution required to get
an accurate numerical solution at a moderate computational cost. Three different
mesh resolutions; fine, medium, and coarse were considered for this purpose. The
specifications of the different meshes are shown in Table 2.
In the present work, since the added mass plays a vital role in approaching the
conclusions; the values of added mass obtained for different mesh sizes were com-
pared to find an appropriate mesh. As seen in Fig. 3, the added mass values obtained
for the medium quality mesh were the nearest to the theoretically calculated value
of added mass (Sect. 2.3) and it required less computational time compared to the
fine mesh. Hence, the medium-sized mesh was used for all further analyses.

Table 2 Types of meshes

Max. element size (m) No. of elements
used to study mesh
convergence Coarse 0.024 11,103
Medium 0.019 16,911
Fine 0.016 24,615

0.13
Normalized added mass

Fine Medium
Coarse Theorotical
0.12

0.11

0.10
1.6 2.4 3.2 4.0
ω (rad/s)

Fig. 3 Normalized added mass obtained for different mesh resolutions

998 V. Vijayasankar and A. Samad

Table 3 Natural periods and

Buoy mass Nat. period (s) Max. RAO % Diff in
RAOs obtained for different
(kg) natural period
buoy masses
30 1.02 1.64 0
40 1.07 2.45 4.90
50 1.16 2.26 13.72

Fig. 4 RAO for different

30kg 40kg 50kg
buoy masses 3.0

2.3

RAO (m/m)
1.5

0.8

0.0
0.50 1.00 1.50 2.00 2.50
Period (s)

4.2 Numerical Experiments for Different Buoy Masses

The same buoy design is tested for three different masses: 30, 40, and 50 kg. The
value of natural frequencies obtained for different buoy masses is listed in Table 3.
It can be noted that the natural period of the buoy keeps on increasing with an
increase in buoy mass. However, since the dimensions of the buoy are kept constant,
the maximum buoyancy that this particular buoy can provide is also constant. This
limits the maximum mass possible to around 55 kg. Figure 4 shows the variation of
RAO with wave periods for different buoy masses. The peaks are obtained near the
resonant period of corresponding buoy configurations.
As discussed in Sect. 2, added mass is independent on the structural mass of the
buoy. It simply represents the added inertia that develops during motion. Therefore,
added mass remains more or less constant for all the three buoy masses.

4.3 Numerical Experiments for Different Heave Plate

Diameters

In the second test case, the buoy mass is conveniently fixed as 30 kg (excluding the
weight of heave plate) and the analysis is done on the buoy fitted with heave plates
of three different diameters, 30, 60, and 90 cm.
It can be seen that the maximum percentage increase in the natural period for the
second test case (89.10%) is much higher than the same in the first test case (13.72%).
Table 4 shows the natural periods for the buoy connected with three different heave
plates. Table 4 also shows an enhancement in the peak power absorbed as the size of
Analyzing Different Methods to Increase the Natural Period … 999

Table 4 Naturalperiods and RAOs obtained for different buoy masses

Plate diameter (cm) Nat. period (s) Max. RAO % Diff in natural period Max. power
30 1.10 2.61 0 19.37
60 1.46 5.26 32.73 36.24
90 2.08 7.50 89.10 79.85

Normalized radiation damping

Without heave plate Heave dia=30 cm Without heave plate Heave Dia=30 cm
0.35 0.020
Heave dia=60 cm Heave dia=90 cm Heave dia=60 cm Heave dia=90 cm
Normalized added mass

0.28 0.016

0.21 0.012

0.14 0.008

0.07 0.004

0 0.000
0 2 4 6 8 0 2 4 6 8
ω (rad/s) ω (rad/s)
(a) (b)

Fig. 5 a Normalized added mass versus angular frequency. b Normalized radiation damping versus
angular frequency

the heave plate increases. Even though this estimate does not include the power lost
due to drag, it shows the dependence of the power absorbed on the diameter of the
heave plate.

Added mass
As shown in Fig. 1, the added mass is heavily dependent on the size of the heave plate
connected at the bottom of the buoy. As the diameter of the heave plate increases,
the added mass increases significantly (See Fig. 5a) since the added mass varies as
the cube of the diameter of the plate.

Radiation damping
Figure 5b indicates that the radiation damping tends to decrease as the heave plate
diameter increases. Since the radiation damping accounts for only 20–30% of the
total damping [14], this doesn’t necessarily mean that the total damping of the system
decreases with an increase in heave plate diameter.

Excitation force
It is evident from Fig. 6a that the excitation force acting on the buoy drops as the size
of the heave plate is increased. This is occurring due to the dependence of excitation
on added mass.
As explained earlier, the heave excitation force is composed of two counteracting
components, the diffraction force (F diff ) and the Froude–Krylov force (F fk ). The
1000 V. Vijayasankar and A. Samad

Normalized excitation force Without heave plate Heave dia=30

0.35 8 Dia = 60 cm Dia = 30 cm Dia = 90 cm
Heave dia=60 Heave dia=90
0.28
6

RAO (m/m)
0.21
4
0.14
2
0.07

0 0
0 2 4 6 8 0.50 1.00 1.50 2.00 2.50
ω (rad/s) Period (s)
(a) (b)

Fig. 6 a Normalized excitation force versus angular frequency. b RAO for different heave plate
diameters

diffraction component increases with an increase in added mass, while at the same
time the more decisive Froude–Krylov component remains more or less constant.
Thus, the net excitation force drops for higher diameters of the heave plate.

Response amplitude operator

The peak value of RAO tends to increase as the heave plate diameter is increased
(See Fig. 6b). This can be explained as a combined effect of two possibilities; (i)
as seen in Eq. (9), the energy content of the wave increases and the wave period
increases and (ii) the resonance occurs at higher wave periods for larger heave plate
diameters.

5 Conclusion

The following conclusions were drawn from the numerical analyses carried out for
the two aforementioned test cases:
1. The resonant period of the buoy increased for both the test cases, but the
percentage of increase is different in the two cases.
2. Increasing the mass of the buoy has little effect on its resonant period when
compared to the addition of a heave plate.
3. In the first test case, an increase in mass has virtually no effect on the added mass
and radiation damping of the buoy.
4. In the second test case, an increase in the heave plate significantly increased
the natural period of the buoy as much as by 89.10%. Also, the excitation force
decreased as the heave plate diameter is increased.
Summing up, the present work was able to successfully fulfil the objective of
improving the performance of the point absorber in longer wavelengths, without
increasing its size.
Analyzing Different Methods to Increase the Natural Period … 1001

References

1. Falcão AFdO (2010) Wave energy utilization: a review of the technologies. Renew Sustain
Energy Rev 14(3):899–918
2. Clément A et al. (2002) Wave energy in Europe: current status and perspectives. Renew Sustain
Energy Rev 6(5):405–431
3. Cruz J (2008) Ocean wave energy current status and future prepectives. Springer, Berlin
4. Baker NJ, Mueller MA (2001) Direct drive wave energy converters. Rev des Energies
Renouvelables 1–7
5. Ai J, Zuo L, Lee H, Liang C (2015) Design, fabrication, simulation and testing of a novel
ocean wave energy converter. In: ASME international manufacturing science and engineering
conference, pp 1–7
6. Budar K, Falnes J (1975) A resonant point absorber of ocean-wave power. Nature
256(5517):478–479
7. WEC-Sim | Open Energy Information. (Online). Available: https://openei.org/wiki/WEC-Sim.
Last accessed 06 Dec 2018
8. Chakrabarti SK (1987) Hydrodynamics of offshore structures. Computational Mechanics
Publications, Southampton
9. Falnes J (2002) Ocean waves and oscillating systems: linear interaction including wave energy
extraction. Cambridge University Press, Cambridge, UK
10. Haslum HA (2000) Simplified methods applied to nonlinear motion of spar platforms.
Nowergian University of Science and Technology, Ph.D
11. Arthur Pecher JPK (ed) (2017) Handbook of ocean wave energy, vol 7. Springer, Aalborg,
Denmark
12. Newman JN (2017) Marine hydrodynamics, 40th Anniversary. The MIT Press, Cambridge
13. Sarpkaya T (2010) Wave forces on offstructure structures, First Edit. Cambridge University
Press, USA
14. Vugts JH (1968) The hydrodynamic coefficients for swaying, heaving and rolling cylinders.
Int Shipbuild Prog 15(167):251–275
Comparison of Hydrogen Yield
from Ball-Milled and Unmilled
Magnesium Hydride in a Batch System
Hydrogen Reactor

J. A. Adeniran, R. S. Fono-Tamo, Esther Titilayo Akinlabi, and T. C. Jen

Abstract Energy plays a crucial role in the economic development of a nation

and its sustenance. In view of the important role of energy to man; energy demand
has been on the increase as the human population increases. Interests in solid-state
hydrogen generation especially from lightweight metals have increased lately due
to high gravimetric and volumetric hydrogen storage/reserve. In this study, effect
of ball milling on hydrogen yield was examined by comparing hydrogen generation
from unmilled MgH2 and ball-milled MgH2 in a hydrolysis reaction carried out in a
batch system hydrogen reactor. Furthermore, the effects of acetic acid concentration
and MgH2 weight on hydrogen generation was investigated. The reaction was carried
out at 30 °C with three substrates weights 0.2 g, 0.4 g and 0.6 g respectively. Ball-
milled MgH2 performed better than unmilled MgH2 by recording higher hydrogen
yield when compared to unmilled MgH2 components with a hydrogen yield of about
0.0194 L relative to 0.0131 L using 0.6 g MgH2 obtained in unmilled MgH2 . The
results from the XRD spectra validates the reduction of crystallite size of ball-milled
MgH2 compared to the unmilled MgH2 . The crystallite size reduces from an average
of 52.4 µm to about 92.25 nm after one-hour ball milling. This development enhances
reaction kinetics by increasing the reaction surface area. Similarly, the fracturing of
the substrate crystals during ball milling increases the nucleation reaction in the
particles thereby increasing the hydrogen release phenomenon.

Keywords Magnesium hydride · Ball milling · Hydrolysis · Hydrogen generation

1 Introduction

Energy remains a pivot for the economic development and sustenance of nations
of the world. With the importance of energy also comes challenges such as high
cost and environmental pollution from some of the conventional energy generation

J. A. Adeniran (B) · R. S. Fono-Tamo · E. T. Akinlabi · T. C. Jen

Department of Mechanical Engineering Science, University of Johannesburg, Auckland Park
Kingsway Campus, Johannesburg 2006, South Africa
e-mail: nikjocrown2000@yahoo.com

© Springer Nature Singapore Pte Ltd. 2021 1003

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_87
1004 J. A. Adeniran et al.

technologies. It is widely believed that energy technologies of the future are those
with sustainability attributes, that is minimal environmental harm, readily accessible
materials, relatively cheap and safe handling.
Hydrogen is one of the veritable renewable energy techniques with increasing
interests from researchers and policymakers in recent years. Among the different
hydrogen generation/storage methods, solid-state hydrogen storage and generation
especially from metal hydrides have attracted renewed interests. The development is
due to the attributes such as high gravimetric and volumetric hydrogen potentials of
most of these categories of materials [1]. Other alluring qualities of metal hydrides
include less harmful environmental reaction by-products, relatively cheaper cost
of experiments, comparative safe handling of the substrates. Magnesium hydride is
among the metal hydrides with increased interests from the experts for hydrogen stor-
age and generation especially for onboard vehicular application. Magnesium hydride
is readily available due to the abundant nature of its main component (magnesium)
regarded as the fourth most abundant element on earth [2]. Similarly, magnesium
hydride has high volumetric hydrogen storage estimated to be about 7.6 wt% [3,
4], the reaction products are considered to be of less harmful effect on the envi-
ronment and the cost of the material is comparatively affordable [5]. However, the
two major challenges facing the use of MgH2 like most metal hydrides are sluggish
reaction kinetics and thermodynamic limitations [6]. Researchers have explored dif-
ferent interventions towards ameliorating the limitations of slow reaction kinetics
and thermodynamics. Among such interventions the applications of catalysts such
noble metals, organic and inorganic acids have been reported to enhance hydrogen
yield [7–10]. Ball milling of metal hydrides for hydrogen generation results in reduc-
tion of crystallite size of the powder, ensures increase in reaction surface area of the
materials [11]. Similarly, the fracturing of the crystals leads to creation of defects
in the crystals thus enhancing nucleation reaction. The increase in nucleation thus
enhances hydrogen generation. In this study, the effect of ball milling on hydro-
gen yield was investigated by comparing hydrogen generation from unmilled and
ball-milled MgH2 via hydrolysis technique.

2 Experimental Design and Method

Figure 1 is a schematic representation of the hydrogen generation experiment. The

major components of the experimental set up include a three-neck round bottom flask
within which the hydrolysis reaction takes place. Other components include temper-
ature regulated water bath (Julabo TW20, Julabo GmbH), the moisture-absorbent
unit, the flowmeter (T1000, Fujikin) and the data logger.
Hydrolysis method was used for hydrogen generation in the study. The substrate
(MgH2 ) was poured inside the round bottom flask, subsequently, the acetic acid-
catalyzed solution in the funnel connected to the reaction vessel was released by
opening the valve on the funnel for the commencement of the hydrolysis reaction.
Other notable components are the thermometer fixed to one of the outer mouths of the
Comparison of Hydrogen Yield from Ball Milled and Un-Milled … 1005

Fig. 1 Schematic representation of hydrogen generation process [12]

three-neck flask while the tube conveying the generated hydrogen from the reaction
vessel is connected to the second outer neck. The line carrying hydrogen from the
reaction vessel was joined to the bubbler for condensation of the gas. The second
opening on the bubbler had another connected to it, through which the condensed
hydrogen gas flows into the flow regulator. It is essential that only hydrogen gas
gets to the flowmeter thus two major components (flow regulator and filter/chemical
separator) were installed on the reactor. The water/moisture in the hydrogen gas is
trapped in the regulator while the gas flows into the filter/chemical separator. The
moisture-absorbent in the design trapped the moisture in hydrogen thus ensuring the
only hydrogen is recorded by the flowmeter. The hydrogen generation was recorded
using the datalogger connected to the flowmeter [13, 14].
The experiment was accelerated with the aid of acetic acid (99.8%, Labchem, SA
(South Africa). The reason acetic acid was chosen as the reaction accelerator is that in
reaction mixture the hydrolysis of the reaction is favour by the large excess of water
employed in the reaction, here water is also serving as the solvent rather than the
acid. This will drive the acetic metal hydrides equilibrium toward the acetic acid, thus
favouring the product formation. The equilibrium will favour the thermodynamically
more stable side and acetic acid is playing the role effectively couple with its pka value
which makes it weaker than other carboxylic acids that could be used for the same
reaction. Furthermore, acetic acid is environmentally friendly as well as the products
of its reaction with MgH2 for hydrogen generation through hydrolysis. Two sets
of substrates were utilized in the study: commercial magnesium hydride powder
99.8% purity (Rockwood Lithium, Germany) without further processing (unmilled)
hereafter referred to as as-received magnesium hydride (AR MgH2 ) and one-hour
ball-milled magnesium hydride (1 h BM MgH2 ). The manufacturer reported average
particle size of 50 µm for the AR MgH2 . Ball milling of the magnesium hydride
1006 J. A. Adeniran et al.

powder was done using Fritsch F6 Planetary mill using tungsten carbide vial and
balls with ball to powder ratio of 30:1. The sample was loaded into the vial inside
glove box operated with argon gas and the sample was ball-milled under argon. The
rotational speed of the ball mill was 300 rpm.
The equation for reaction of magnesium hydride acetic acid-accelerated hydrol-
ysis is represented in Eq. (1):

2MgH2 + 2H2 O + 2CH3 COOH → Mg(OH)2 + Mg(CH3 COO)2

+ 4H2 + Hrxn (1)

The heat of reaction (Hrxn) is approximately −277 kJ/mole.

Three different substrate weights (0.2, 0.4 and 0.6 g) were used in the study.
To ensure uniformity of the substrate’s weight, BM-200 analytical balance (United
Kingdom) with 0.0001 g repeatability was used to measure the substrates. The study
was conducted at 30 °C ambient temperature. Comparison of effect of ball milling and
substrate weight on hydrogen yield was done by conducting the hydrolysis experi-
ment with both AR and one-hour ball-milled MgH2 at the three weight compositions.
Thereafter, the best performing sample was selected for the investigation of catalyst
concentration on hydrogen yield. To ascertain the catalyst concentration effect on
hydrogen generation, the study was carried out using 30 wt%, 40 wt%, 50 wt%,
60 wt% and 70% acetic acid concentrations respectively.

2.1 Material Characterization

The X-ray diffraction (XRD) analysis was accomplished with the aid of the Bruker
AXS D8 Advance instrument (USA). Using Cu-Kα, λ1 = 1.5406 Å, λ2 = 1.5444
Å, 2θ = 10–90o with a step size of 0.035°. Furthermore, α-Al2 O3 was employed as
the standard sample for establishing the impact of the equipment on the peak profile
of parameters. Thereafter, the Rietveld full profile analysis of the data was achieved
using the General Structure Analysis System (GSAS) software [15, 16].

3 Results and Discussion

3.1 Role of Ball Milling/Crystalline Size on Hydrogen

Generation

To ascertain the effect of crystalline size on hydrogen yield, the AR MgH2 and 1 h BM
MgH2 were evaluated for hydrogen generation through three sample weights namely:
0.2 g, 0.4 g and 0.6 g, respectively. Figure 2 shows the result of the comparison
of hydrogen yield from the two materials in three different weight compositions.
Comparison of Hydrogen Yield from Ball Milled and Un-Milled … 1007

Fig. 2 Hydrogen generation from ball-milled and as-received MgH2

Hydrogen generation from the BM samples was higher than AR samples in all the
categories. The development is due to the improvement in the kinetics of the 1 h BM
MgH2 samples due to the increase in surface area relative to the AR MgH2 .

3.2 Sample Weight Effect on Hydrogen Yield

Figure 3 and Table 1 show the combined results of hydrogen yield from the six
substrates used: 0.2 g (AR and 1 h BM MgH2 ), 0.4 g (AR and 1 h BM MgH2 ),
and 0.6 g (AR and 1 h BM MgH2 ), 0.4 g (AR and 1 h BM MgH2 ). From the two
forms of result presentations (Fig. 3 and Table 1), the weight of substrates influenced
hydrogen yield with hydrogen generation proportional to weight of the substrates for
each of the two categories of the six samples (i.e. AR MgH2 and 1 h BM MgH2 ). The
observation agrees with some of the previously reported studies from our research
team [13, 14]. In these sets of experiments, the highest hydrogen yield was 0.0194 L
from 0.6 g 1 h BM MgH2 , however, 0.4 g AR MgH2 produced the lowest hydrogen
yield (0.0096 L) compared to approximately 0.0101 L from 0.2 g AR MgH2 . The
reason for the development is not known now. Similarly, after observing the same
trend of result from the sample when the experiment was repeated and for the fact
that it is the only departure among the six samples, it may be considered an outlier.
1008 J. A. Adeniran et al.

Fig. 3 Optimization of substrate quantity for hydrogen generation

Table 1 Hydrogen
Material Maximum Duration of Ranking of
generation at 30 °C 40 wt%
(Mg H2 ) H2 yield (L) experiment H2 yield
Acetic concentration
(min)
0.6 g 1 h BM 0.0194 60 1
0.6 g AR 0.0131 60 4
0.4 g 1 h BM 0.0176 60 2
0.4 g AR 0.0010 60 6
0.2 g 1 h BM 0.0158 60 3
0.2 g AR 0.0101 60 5

3.3 Optimization of Catalyst Concentration for Hydrogen

Generation

Having observed that 0.6 g 1 h BM MgH2 was the optimum sample for hydrogen
generation, it was selected for optimization of acetic acid concentration on hydrogen
yield. Figure 4 shows the hydrogen yield from the substrate at five different catalyst
concentrations (30, 40, 50, 60 and 70 wt%). The 30 and 40 wt% recorded the best
performance for maximum hydrogen yield at 60 min (approximately 0.0194 L).
However, 30 wt% experiment recorded better hydrogen yield compared to 40 wt%
within 10 min of study suggesting faster kinetics at that point. Moreover, 70 wt%
recorded the fastest kinetics in the first 10 min of experimental observation.
Comparison of Hydrogen Yield from Ball Milled and Un-Milled … 1009

Fig. 4 Hydrogen generation at different catalyst concentrations

3.4 MgH2 Material Characterization

In this section, the results from the XRD analyses of the two samples will be
discussed.
X-ray diffraction (XRD) analyses
XRD characterization of the materials is important to provide useful information on
their crystalline structures and phase information. Figure 5 shows the XRD plots
of AR MgH2 and 1 h BM MgH2 stacked for easy comparison. From the plots, two
major phases were observed namely: Mg and αMgH2 . The αMgH2 is commonly
observed in unmilled MgH2 or samples that have been ball-milled for a short period of
time [17]. Moreover, the crystallite size obtained using GSAS software for Rietveld
analysis reported approximately 52.4 µm for the αMgH2 (MgH2 ) phase, a result
that is close to the 50 reported by the manufacturer. After a 1 h ball milling, the
crystallite size reduces to 92.25 nm. From the structural perspective, the XRD plots
show broadening of peaks in 1 h BM sample and shifting of the peaks to the right
compared to the AR MgH2 an indication of reduction of crystallinity of the materials
[18]. The impact of the crystalline size reduction on hydrogen yield was evident by
the consistent higher hydrogen yield in the 1 h BM samples compared to the AR
samples. The increased hydrogen yield was caused by fracturing of the crystals by
ball milling leading to increased nucleation phenomenon which enhanced reaction
kinetics.
1010 J. A. Adeniran et al.

Fig. 5 X-ray powder diffraction pattern of AR MgH2 and 1 h BM MgH2

4 Conclusion

Ball milling of MgH2 for 1 h reduced the crystallite size of the material 52.4 µm to
92.25 nm and improved hydrogen yield in the study. The result further underscores
the importance of crystalline size reduction to improve the reaction kinetics and
hydrogen generation through hydrolysis. The catalysis of the experiment with acetic
acid being an environmentally friendly acid will also enhance the scale-up potential
of the technology.

Acknowledgements Special thanks to the National Research Foundation (NRF) South Africa for
funding the study of the first author. Authors also thank Dr. M Lototskyy for availing the ball milling
facilities at the South African Institute for Advanced Materials Chemistry, University of the Western
Cape, for this study and for assisting with the XRD and GSAS analyses of the MgH2 powders. The
support of Dr. SN Nyamsi during the ball milling process is also gratefully acknowledged. This
project was funded by the TIA grant of the University of Johannesburg.

References

1. Song MY et al (2014) Hydrogen-storage properties of MgH2 –10Ni–2NaAlH4 –2Ti prepared

by reactive mechanical grinding. J Ind Eng Chem 20(4):1591–1595
2. O’rourke JG, Stevenson DJ (2016) Powering Earth’s dynamo with magnesium precipitation
from the core. Nature 529(7586):387
Comparison of Hydrogen Yield from Ball Milled and Un-Milled … 1011

3. Chao CH, Jen TC (2013) Maximized on-demand hydrogen generator design. Adv Mater Res.
Trans Tech Publications
4. Sakintuna B, Lamrdi-Darkim F, Hirscher M (2007) Metal hydride materials for solid hydrogen
storage: a review. Int J Hydrogen Energy 32(9):1121–1140
5. Huot J, Liang G, Schulz R (2003) Magnesium-based nanocomposites chemical hydrides. J
Alloy Compd 353(1):L12–L15
6. Kojima Y, Kawai Y, Haga T (2006) Magnesium-based nano-composite materials for hydrogen
storage. J Alloy Compd 424(1):294–298
7. Uan JY et al (2009) Evolution of hydrogen from magnesium alloy scraps in citric acid-added
seawater without catalyst. Int J Hydrogen Energy 34(15):6137–6142
8. Akdim O, Demirci UB, Miele P (2009) Acetic acid, a relatively green single-use catalyst for
hydrogen generation from sodium borohydride. Int J Hydrogen Energy 34(17):7231–7238
9. Akdim O, Demirci UB, Miele P (2009) Highly efficient acid-treated cobalt catalyst for hydrogen
generation from NaBH4 hydrolysis. Int J Hydrogen Energy 34(11):4780–4787
10. Gervasio D, Tasic S, Zenhausern F (2005) Room temperature micro-hydrogen-generator. J
Power Sources 149:15–21
11. Webb CJ (2015) A review of catalyst-enhanced magnesium hydride as a hydrogen storage
material. J Phys Chem Solids 84:96–106
12. Adeniran JA et al (2019) Temperature optimized hydrolysis of Acetic acid catalyzed mag-
nesium Hydride for Hydrogen generation in a batch sytem Hydrogen reactor. In: Prabha D
(ed)Transactions on engineering technologies, Springer Nature Singapore Pte: Singapore
13. Jen TC et al (2016) Hydrogen generation from Acetic acid catalyzed magnesium Hydride
using an on-demand Hydrogen reactor. In: International mechanical engineering congress and
exposition. Phoenix, Arizona, USA: American Society of Mechanical Engineers
14. Adeniran JA, et al (2017) Organic Acid-Catalyzed Hydrolysis of Magnesium Hydride for
Generation of Hydrogen in a batch system Hydrogen reactor. In: Proceedings of the world
congress on engineering and computer science
15. Larson AC, Von Dreele RB (1994) Gsas. Report lAUR, pp 86–748
16. Nyamsi SN, Yartysb V, Lototskyya M (2017) Synthesis of Mg 2 FeH 6 assisted by heat treatment
of starting materials. In: 1st Africa energy materials conference. Pretoria: Elsevier
17. Huot J et al (1999) Structural study and hydrogen sorption kinetics of ball-milled magnesium
hydride. J Alloy Compd 293:495–500
18. Grosjean MH et al (2006) Hydrogen production via hydrolysis reaction from ball-milled Mg-
based materials. Int J Hydrogen Energy 31(1):109–119
A Review on the Synthesis of Activated
Carbon from Natural Resources
for Mechanical Applications

A. John Presin Kumar, S. Sivakumar, D. Prasanth, B. Guhanesh,

and A. Ijas Ahamed

Abstract Activated carbon is a material which can be produced from various waste
materials such as plant waste, animal waste and industrial waste resulting in reduc-
tion of material wastage, better utilization of natural materials and environmental
friendliness. Activated carbon plays an important role in many applications in vari-
ous forms like gas absorbent, composite fillers, etc. Activated carbon particles have
many promising applications in various industries like water treatment, dye and sugar
refining, degasification and composite fabrication as fillers. In composites, activated
carbon can play a vital role as effective reinforcement materials for improving prop-
erties. During the last ten years, there had been a good work done in this field with
regard to exploring the possibilities of activated carbon in both production and appli-
cation. This work discusses contribution of several authors in the area of activated
carbon in both synthesis and characterization toward various applications related to
mechanical engineering.

Keywords Activated carbon · Crushing · Cleaning · Drying · Heating · Ball

milling · Sieving · Furnace · Size · Morphology · Nano

A. John Presin Kumar (B) · S. Sivakumar · D. Prasanth · B. Guhanesh · A. Ijas Ahamed

Department of Mechanical Engineering, Hindustan Institute of Technology and Science, Padur,
Kelambakkam, Chennai, Tamil Nadu 603103, India
e-mail: johnpk@hindustanuniv.ac.in
S. Sivakumar
e-mail: ssivak@hindustanuniv.ac.in
D. Prasanth
e-mail: prasanthsekar22@gmail.com
B. Guhanesh
e-mail: bguhanesh@gmail.com
A. Ijas Ahamed
e-mail: ijas9794@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 1013

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_88
1014 A. John Presin Kumar et al.

1 Activated Carbon

In India, it is always appreciable to develop methodologies based on natural resources

for the production of useful materials. Natural resources when explored can provide
raw materials at low cost for various purposes. Activated carbon is a form of carbon
processed in various stages of heating to develop in it small, low-volume pores that
can possibly increase the surface area available for adsorption as shown in Fig. 1.
Due to the very high degree of microporosity that is contained in one gram of acti-
vated carbon, there could be a corresponding higher surface area available for better
adsorption characteristics. This property of porosity provides lots of benefits during
the crucial application areas of activated carbon. There are various levels of activa-
tion which are attainable through methods like chemical activation or heat activation.
An activation level that is just sufficient for a useful application may be obtained by
attaining high surface area as mentioned earlier. According to the requirements for
application, the activation level is chosen and the appropriate method to attain it
is also identified. Further chemical treatment if required often enhances adsorption
properties.
As already mentioned, there are numerous applications for activated carbon. One
such important industrial application involves use of activated carbon particles in
metal finishing for the purpose of purification during electroplating solutions. One
such example is that it is used as the main purification technique for removing organic
impurities from bright nickel or chrome plating solutions. Activated carbon finds an
astonishing usage in medical field also. It is used for treating poisonings and over-
doses in humans and also for animals. Tablets or capsules of activated carbon are

Fig. 1 Activated carbon powder

A Review on the Synthesis of Activated Carbon from Natural … 1015

popularly used in many countries as drug to treat diseases like diarrhea, indiges-
tion and flatulence. The adsorption property of carbon or activated carbon is a much
required property as it makes carbon an important material in the areas where adsorp-
tion is needed. For example, it finds an environment-friendly application of reducing
pollution as mentioned in the following points. Activated carbon has numerous appli-
cations in removing pollutants from air or water streams in the field and in industrial
processes such as: spill water cleanup, groundwater remediation, drinking water fil-
tration, air purification, volatile organic compound capture from painting and other
processes. Activated carbon is a readily allowed substance used by agricultural peo-
ple and organic farmers in both livestock production and wine-making. In livestock
production, it is normally used as a pesticide, animal feed additive, processing aid,
nonagricultural ingredient and disinfectant. Filters along with activated carbon in the
form of coatings are usually used in compressed air and gas purification to remove
oil vapors, odor and other hydrocarbons from the air.

2 Review Summary

There had been a wide variety of synthesis methodologies and characterization

techniques utilized by several authors based on their application scenarios. In all
these works, the commonalities observed were like heat activation, chemical activa-
tion, popularly used chemicals, gas adsorption tests, ash content tests, etc. Synthesis
methodologies can be chosen keeping in mind the budget, the planned application
area of activated carbon, etc. Likewise, activation methods can be planned based on
the properties desired for the activated carbon. As already pointed, more emphasis
may be given for utilizing natural resources like plant materials or animal remaining.
The advantages lying with choosing natural resources are in their plenty of avail-
ability, lower cost compared to artificial ones and most importantly the environment
friendliness. The benefits provided to the society through environmental friendliness
are getting much importance every day as the society is moving toward a much faster
lifestyle through exploitation of fossil fuels.
Activated carbon can be produced from resources having both plant and animal
origins. Initiated carbon can be created from assets having both plant and creature
inceptions. Creature bones, for example, dairy animals, donkey, horse and chicken,
can be utilized to deliver carbon singe for quality evaluation [1]. In this examination,
tests broke down were carbonized in a shut pot at 400 °C and ground into powder
subsequent to cooling. Foreordained amount of every one of the pounded tests was
enacted utilizing hydrochloric corrosive, and another example was initiated by warm
technique. To acquire the ideal weight reduction of the carbons, the carbonization
time frame was shifted for couple of hours. From the outcome, the expansion in
carbonization time caused a calculable increment in the rate weight reduction; this
achieved a most extreme incentive at a specific carbonization time for both carbon
tests at a comparing rate weight reduction of various creature bones separately. The
level of carbon yield of different crude material assets, for example, cow, jackass,
1016 A. John Presin Kumar et al.

Table 1 Variation of carbon

S. No. Raw material % of carbon yield
yield
Acid activation Heat activation
1 Resource 1 ˆ48.9 ˆ37.0
2 Resource 2 ˆ48.6 ˆ36.2
3 Resource 3 ˆ46.3 ˆ34.9
4 Resource 4 ˆ44.8 ˆ30.1

chicken and pony bones upon corrosive and warmth enactment, was in expanding
request as appeared in Table 1. Among the bone examples, chicken bone has the
least fiery remains content for corrosive, heat and non-actuated examples separately.
The distinction in mass densities of corrosive-initiated and warmth-enacted carbons
is minuscule.
The heat activated carbons show higher percentage of hardness than the acid acti-
vated and non-activated carbons. This study indicated that activated carbons from
the animal residues were effective as adsorbents with those obtained from chicken
having better features of adsorbents. Plant matters such as banana stem can also be
a possible resource for producing activated carbon [2]. In this, the author performed
the production of activated carbon electrodes from a banana stem for supercapacitor
cell applications. The increase in pore properties was conducted using a potassium
hydroxide KOH chemical activation agent at low concentrations, whereas physical
activation was performed using CO2 gas at a particular temperature and time period.
The activated carbon electrode was fabricated without the addition of adhesive mate-
rials as can be noticeable in many works [1, 3, 4]. The density, degree of crystallinity,
surface morphology, elemental content and surface area of the electrodes were ana-
lyzed. The study of electrochemical properties focused on the analysis of the specific
capacitance of the supercapacitor cell using cyclic voltammeter. The physical prop-
erties of the activated carbon electrodes were correlated to generate the optimum
conditions for the specific capacitance of the supercapacitor cells.
The ideal explicit capacitance got at a specific grouping of KOH achieved best
an incentive with a particular surface territory. Initiated carbon can be a decent
material for creation of composites for good quality and firmness. Lebedeva et al.
[5] integrated composite materials, carbon/polyaniline, utilizing permeable actuated
carbon delivered from carbonized rice husk [6]. This has a particular surface zone
(Fig. 4). The arrangement of the composite materials was combined by oxidative
aniline polymerization technique within the sight of the actuated carbon in acidic
medium; polyaniline content was changed. Electrochemical properties of the com-
posite materials were researched by cyclic voltammetry, galvano static charge/release
and soundness tests. Results show that properties of the composites are administered
by proportion of the parts: A type of cyclic bends and area of polyaniline crest
esteems emphatically rely upon the polyaniline sum. Gravimetric capacitance devel-
ops with polymer content in the material and achieves greatest incentive, for example,
with particular polyaniline content. In any case, toughness tests demonstrate that the
steadiest is the example with the least polyaniline amount (Fig. 2).
A Review on the Synthesis of Activated Carbon from Natural … 1017

Fig. 2 Rice husk

Coconut shells are often the promising natural material for many applications.
The use of coconut shell (Fig. 5) activated carbon was investigated for a potential
adsorbent for malachite green dye from aqueous solution as done by Haji Azaman
et al. [7]. Abechi et al. studied the effect of various factors, such as initial dye
concentration, contact time, pH and solution temperature [8]. The interaction between
dye molecule and coconut shell activated carbon absorbent was strongly influenced
by the pH of the solution [9]. Maximum adsorption of malachite green dye was
obtained at particular pH, while the point of zero charge was obtained at another value
of pH. The maximum monolayer adsorption capacity was increased with increment
in temperature. The kinetics of adsorption followed a pseudo-second-order kinetic
model. The intra-particle diffusion model was evaluated to determine the mechanism
of the adsorption process (Vijayan et al. [10]). Based on the Boyd plots, the adsorption
of malachite green dye on the coconut shell activated green absorbent was mainly
governed by film diffusion (Fig. 3).
Thermodynamic parameters were determined, and it was found that MG dye
adsorption on coconut shell activated carbon was spontaneous and endothermic in
nature. The conducted reusability test disclosed the decreasing coconut shell activated
carbon performance from higher value of malachite dye removal down to a lower
value of the same after few adsorption/desorption cycles. Activated carbon has wide
variety of applications have done a study in activated carbon [11]. Activated carbon
is used in gas purification, decaffeination, gold purification, metal extraction, water
1018 A. John Presin Kumar et al.

Fig. 3 Coconut shells

purification, medicine, sewage treatment, air filters in gas masks and respirators,
filters in compressed air and many other applications. Due to its high degree of
microporosity, just one gram of activated carbon has a particular surface area, as
determined by gas adsorption. An activation level sufficient for useful application
may be attained solely from high surface area; however, further chemical treatment
often enhances adsorption properties.
One noteworthy mechanical application includes utilization of activated carbon
in the metal finishing field. It is broadly utilized for refinement of electroplating
arrangements. For instance, it is a fundamental sanitization system for expelling nat-
ural contaminations from brilliant nickel plating arrangements. Anticipating on the
accessible crude materials, it is built up that modern interest for actuated carbon can
be met locally. The portrayal of two sorts of manufacturing plant squanders demon-
strated that paper squander was a superior choice over rice husk. The enacted carbon
delivered from pyrolysis of paper factory squander was synthetically actuated with
different initiating specialists like zinc chloride, potassium hydroxide and potassium
chloride. A precise examination of the impact of impregnation proportion, actuation
temperature and enactment time on the properties of the initiated carbon was fin-
ished. The ideal planning conditions together with the affected of these conditions
on the yield and effectiveness of delivered actuated carbon have been examined. The
artificially initiated carbons were described by estimating iodine and yield rate. The
enacted carbon arranged from paper plant squander in this investigation had most
extreme iodine with zinc chloride as the actuating operator.
A composite material is characterized as a blend of at least two materials that
outcomes in preferable properties over the individual segments are utilized alone.
The job of filler material in composites is essential as it advances improvement in
A Review on the Synthesis of Activated Carbon from Natural … 1019

properties. Actuated carbon can be a superior filler material in composites as in the

work. Here, examination was practiced for the effect of started carbon as filler on the
mechanical properties of wood composite. It on a very basic level revolved around
the quality, stress and expulsion of the wood composites. The wood composite has
certain positive conditions over wood, as they are moderate and have the potential
for adaptable plans. The materials have incredible mechanical properties and have a
long organization life. As a result of the significance of mechanical properties, wood
composite materials have been commonly used all through the latest four decades.
Essential and non-helper structured wood composites subject to squeezed wood,
medium thickness fiber board, secured veneer timber, thermoplastic or wood fiber
blends, and are right now used in both inside and outside applications.
With advances in existing development, especially in building field in which the
usage of PC programming, for instance, SolidWorks reenactment is used to test
show cautiously for huge specific information directly off the bat in the arrange-
ment methodology. This sort of programming decreases weight and material cost,
and improves sturdiness and gathering limit of the thing. Reenactment furthermore
enables depiction of the exhibited real structure, and it is directly persistent by tech-
niques for PC. From both proliferation and exploratory system used in this exam-
ination, medium thickness fiber composite models show higher-quality motivating
force than packed wood composite precedents in view of the growing thickness of
the sanctioned carbon filler. For the most part, the rate contrasts among generation
and preliminary technique are lower than 10% which demonstrate that reenactment
is sensible mechanical assembly use to anticipate the nature of wood composites.

3 Inferences

Activated carbon has wide utilization. On the one hand, it tends to be degasified;
then again, it can likewise be filler for composites. Numerous creators have done
work on enacted carbon under various objectives. Exploration and usage of plant
and creature source have been done broadly by numerous creators. Figure 4 portrays
roughly picture scatters of different waste asset usages over the most recent ten years
by different creators in the field of enacted carbon amalgamation. This is just a rough
picture showing a comparative analysis through review of selective works.
Various studies indicated that materials made of activated carbon are performing
well compared to conventional materials. The adsorptive efficiency of each of the
carbons can be enhanced through controlled and well-monitored activation. The study
utilizing animal bones [1] indicated that activated carbons from various animals are
effective as adsorbents with regard to their evaluated properties of moisture content,
ash content, bulk density, fixed carbon, hardness and percentage yield. Activated
carbon gotten from chicken having low ash content is a good indication of a better
adsorbent. Plant materials are also good sources for activated carbon such as banana,
rice and coconut. The preparation method for activated carbon electrodes from banana
1020 A. John Presin Kumar et al.

Fig. 4 Comparative
utilization of various wastes

stems used herein is a very feasible method for the production of supercapacitor cells
as in [2] which gave optimum results.
Figure 5 depicts the observed commonality in the procedural steps of synthesiz-
ing activated carbon from naturally available resources. The steps involved and the
corresponding hardware requirements are also mentioned below. Different authors
on different applications had synthesized activated carbon using activation processes
involving pyrolysis. This is normally grouped as heat activation and chemical activa-
tion method. For most of the applications, activated carbon can be prepared econom-
ically and easily using the following seven-step procedure. The following diagram
depicts this clearly followed by an explanation of what are the exact steps involved
and the corresponding hardware requirements.

Fig. 5 Generalized steps for activated carbon synthesis from natural materials
A Review on the Synthesis of Activated Carbon from Natural … 1021

Steps

1. Collecting naturally available raw material from locality.

2. Crushing into small pieces. (approx. 2–3 mm).
3. Cleaning the pieces through sieving.
4. Drying the pieces by heating in oven mildly. (approx. 100 °C).
5. Soaking it for particular time duration. (approx. 24 h).
6. Initial furnace heating. (approx. 460 °C and held const. for less than an
hour).
7. Final furnace heating. (approx. 750 °C and held const. for less than an
hour).
Hardware requirements:
1. Ball milling machine.
2. Sieving machine.
3. Oven.
4. Furnace.

Rice husk was also a promisingly utilized plant material like banana stem, which
when used for activated carbon has provided many opportunities for making compos-
ite using polymerization [5]. It was shown that the specific capacitance of composite
materials rises with the increase of the polymer amount, and in all cases, its value
exceeds a simple sum of the capacitances of the individual components. Thus, the
combination of the activated carbon material and the conductive polymer provides
the synergistic effect leading to improved capacitive and operational characteris-
tics and allows consideration of such composite materials as promising for the use
as active component of supercapacitor electrodes. Rice husk has been successfully
used by many authors as in [12–14] for polymer bio-composites in addition to other
natural elements like walnut shell and coconut shell, whereas the same authors have
used glass fibers in another work [15] for polymer composites for artificial limp
applications [16, 17].
The coconut shell activated carbon has demonstrated good reusability properties,
with the ability to sustain the removal efficiency of greater values of malachite green
upon the completion of five consecutive adsorption/desorption cycles [7]. In addition
to animal and plant materials, paper wastes are also possible materials for activated
carbons [12] in which the results of the study showed that it is feasible to prepare
activated carbons with relatively high surface areas and pore volumes from paper
waste by direct chemical activation. Activation with zinc chloride produced activated
carbons with better developed porosities than with potassium hydroxide. The iodine
value of the activated carbon product increased with concentration of zinc chloride
solution. As the impregnation time increased, the iodine value rose steeply, reaching a
maximum value. The iodine value also increased with higher activation temperature,
beyond which it gradually decreased, presumably due to excessive carbonization.
1022 A. John Presin Kumar et al.

The porosity of the product was comparable with that of commercially achieved
carbon. The waste-based activated carbon had proper mean pore diameters, pore
volume and micro-details indicating its microporous and mesoporous characters.
Table 2 shows a comparative data of the natural resources used with respect to
different application areas by many authors. Activated carbon not only is capable
of improving strength but also has the ability to relieve stresses [18, 19]; it was
shown that the stress value was increased with the increasing forces applied to the
samples of activated carbon. Other than that, the stress value of plywood without
activated carbon as filler was higher than the one without activated carbon filler. But,
due to the addition of different thicknesses of activated carbon, the stress value of
plywood became lower. This showed activated carbon limiting the energy inside the
plywood sample [20] and thus lowering the stress value of the sample. SolidWorks
simulation determined the sample that has high average displacement and average
stress which indicate good performance of mechanical properties. The addition of
activated carbon as filler increased the average displacement [6, 21, 22] and average
stress which is higher than plywood. Thus, both the methods which were experimental
and simulation showed that the mechanical properties were higher compared to
plywood with the addition of activated filler. It was also proved that SolidWorks
simulation was a suitable tool used to measure strength to get accurate result.
Activated carbon finds better applications as reinforcement fillers for compos-
ites [23–26]. Composites are materials nowadays getting more and more impor-
tance, especially for applications requiring less weight considerations. For instance,
mechanical properties of carbon coir composites depend on many parameters mainly
factors like stress–strain behavior of carbon and matrix phases, the phase volume
fractions, the carbon concentration, the distribution and orientation of the carbon or
reinforcement fillers relative to each other. In many works, it was found that the ten-
sile strength of filler-based composites is higher compared to the other composites
[27–31]. It could be easily predictable that there is always a trend toward carbon
concentration in increasing order. At lower concentration of filler material, the com-
posite specimens exhibit slightly linear behavior before or prior to a sharp failure.

Table 2 Materials used by

Author Natural resource used Application area
authors and their applications
[1, 47] Animal bone Gas adsorption
[1–4] Banana, rice, coconut, Supercapacitor cell
coal
[5, 6] Rice husk and coffee Polymer composites
residue
[7–10] Coconut shell and palm Gas adsorption
shell
[12–14] Paper waste Metal purification
(electroplating)
[15–17] Agricultural raw waste, Wood composite
wood, sawdust
A Review on the Synthesis of Activated Carbon from Natural … 1023

This clearly means that the specimen deformed plastically immediately after elastic
deformation. Normally, the impact performance of the fiber-reinforced composites
depends on many factors including the nature of the individual constituents, interface
between carbon and matrix, the construction and geometry of the composite material
and the surrounding conditions of test. In addition, the impact strength of the com-
posites depends and gets influenced by many factors like toughness properties of the
fillers, the very nature of the interfacial region and frictional component [32–36].
Overall, the mechanical properties of composites filled with activated carbon parti-
cles as reinforcement fillers are influenced by many parameters as discussed. And in
addition to that, the size and morphology of the activated carbon particles are having
very much importance on the resulting mechanical properties of the composites. It is
generally surprising to note that the size of a reinforcement particle is influencing the
overall behavior of the composites [37–41]. Activated carbon particles should gen-
erally be of uniform shape and size as far as concerned in order to achieve betterness
in properties and strength. Surface topography and morphology of activated carbon
particles can be better studied by means of scanning electron microscopy. Recently,
activated carbon particles are getting more importance and influence toward many
environmental applications also like pollution control, agriculture field, etc., through
latest developments in its synthesis procedures and conditions in achieving nano-size
levels [42–46]. Nano-size of the particles can be better promising in achieving even
better surface areas and thereby providing the possibilities of better gas adsorption
and better mechanical properties.

4 Conclusions

• Thus, it was very evident from the work of several authors in the area of activated
carbons that these materials can be better produced through proper selection and
utilization of natural materials for various useful purposes such as natural materials
like coconut shell, rice husk, walnut shell, banana stem and animal bones at an
economical level suitable for our growing country like India.
• The application areas of activated carbon are vast, and especially like gas adsorp-
tion and in composites, the activated carbon plays an enhancing role of improving
the performance characteristics.
• The effective utilization of waste materials which has plant or animal origin creates
lots of benefits not only to the desired application but also to the society as a whole.
• Thus, it promises to be a good future material whose very nature and properties
can be further explored for various better applications and the needs of the society.
• More mechanical properties were discussed with regard to the behavior of
composites made of activated carbon particles in the form of reinforcement fillers.
• The main influencing factors were discussed for composites in relation to the
filled activated carbon fillers.
1024 A. John Presin Kumar et al.

• The relation between mechanical properties of composites made up of activated

carbon particles as fillers with that of size, morphology and topography of particles
was also discussed.
• Recent advancements in the field of activated carbon such as the nano-activated
carbon were discussed with respect to its advantages and applications.
• Activated carbon could thus be a very promising material for many mechanical
applications, especially as reinforcement filler for composites with regard to nano-
size, morphology and recent developments.
• There is a good future scope for activated carbons. More emphasis should be
given while choosing the raw materials based on natural resources. This gives the
benefit of minimizing waste materials which is a very environmental-friendly act.

Acknowledgements We would like to express our sincere gratitude and thanks to Dr. Velayudham,
Scientist ‘G’, Ministry of Defence, CVRDE, Avadi, Chennai, India; Dr. Ashish Selokar, Accendere,
CL Educate Ltd., New Delhi, India, for their valuable suggestions; and R. Balaji, mechanical second-
year student of Hindustan Institute of Technology and Science. We express our sincere gratitude
and thanks for their helps in the whole duration of executing this research work.

References

1. Sunday NJ, Okechukwu NS, Elom N, Anthony O, Michael PT (2018) Quantitative character-
ization of activated Carbon from Cow, Donkey, Chicken and Horse Bones from Ezzangbo in
Ebonyi State, Nigeria. Am J Appl Chem 6(5):169–174
2. Taer E, Taslim R, Mustika WS, Kurniasih B, Agustino A (2018) Afrianda and Apriwandi. Int
J Electrochem Sci 13:8428–8439
3. Renugadevi N, Anisha G, Lalith P (2010) Adv App Sci Res 1(2):102
4. Mane PC, Bhosle AB, Deshmukh PD, Janagam CM (2010) Adv App Sci Res 1(3):212
5. Lebedeva MV, Ayupov AB, Yeletsky PM, Parmon VN (2018) Rice husk derived activated
carbon/polyaniline composites as active materials for supercapacitors. Int J Electrochem Sci
13:3674–3690
6. Aloko DF, Adebayo GA (2007) J Eng Appl Sci 2(2):440–444
7. Haji Azaman SA, Afandi A, Hameed BH, Mohd Din AT (2018) J Appl Sci Eng 21(3):317–330
8. Abechi SE, Gimba CE, Uzairu A, Dallatu YA (2013) Res J Chem Sci 3(7):54–61
9. Cobb A, Warms M, Maurer Chiesa EPS (2012) Int J Ser Lear Eng 7(1):93–104
10. Nsi EW, Akpakpan AE, Ukpong EJ, Akpabio UD (2016) Int J Eng Sci 5(9):38–41
11. Subramani T, Revathi PK (2015) Production of activated Carbon from agricultural raw waste.
IOSR J Eng 5(3):54–63
12. Chandramohan D, Kumar AJP (2017) Experimental data on the properties of natural fiber
particle reinforced polymer composite material. Data in brief, vol 13, pp 460–468
13. Yusuf MI, Ariahu CC, Igbabul BD (2012) Afr J Pure Appl Chem 6(9):123–131
14. Darmawan S, Sofyan K, Pari G, Sugiyant K (2010) J Res 7(2):100–111
15. Chandramohan D, Presin Kumar AJ (2017) Data-enabled Discover. Appl CrossMark. Springer
International Publication, pp 1–9
16. Irle M, Barbu MC (2010) Wood-based panel tech. In: Brunei University Press, London pp 190
17. Lam MK, Zakaria R (2008) (Icenv).Inter Conf Envir pp 1–12
18. Aziz NA, Mazlan M, Marinah M, Mohamad Amini MH, Abdul Aziz MS, Yusoff H, Ismael
Rizman Z (2015) ARPN J Eng Appl Sci 10(1):376–386
A Review on the Synthesis of Activated Carbon from Natural … 1025

19. Kumar A, Gupta A, Sharma KV, Gazali SB 2013 (1988) 8:6231–6241

20. Ambedkar G, Muniyan M (2011) Mater Adv App Sci Res 2(4):283
21. Roh KC, Park JB, Lee CT, Park CW (2008) J Ind Eng Chem 14:247–251
22. Kim T, Jin HJ, Park SS, Kim SJ, Cho SY (2008) J Ind Eng Chem 14(6):714–719
23. Fierro V, Fernandez VT, Celzard A (2007) Microp Mesop Mater 360–363
24. Rodenas MAL, Amoros DC, Solono AL (2003) Carbon 41:267–275
25. Rodenas MAL, Amoros DC, Solono AL (2005) Carbon 43:1758–1767
26. Rodenas MAL, Juan JJ, Amoros DC, Solono AL (2004) Carbon 42:1371–1375
27. Rodenas MAL, Castello DL, Amoros DC, Solono AL (2001) Carbon 39:751–759
28. Castello DL, Rodenas MAL, Amoros DC, Solono AL (2001) Carbon 39:741–749
29. Mastalerz M, Bustin RM (1995) Fuel 74:536–542
30. Otowa T, Nojiyama Y, Miyazaki T (1997) Carbon 35(9):1315–1319
31. Qadeer R, Akhtar S (2005) J Chem 29:95–99
32. Qadeer R, Rehan AH (2002) J Chem 26:357361
33. Sanchez AR, Elguézebal AA, Pliego JA (2005) Mater 85:331–339
34. Solano AL, Gullon IM, Lecea CSM, Talovera BS (2005) Fuel 79:635–643
35. Stavropulos GG, Zabaniotu AA (2005) Mater 82(1–2):79–85
36. Supaluknari S, Larkins FP, Redlich P, Jackson WR (1998) Fuel Proc Technol 19:123–140
37. Teker M, Imamoglu M (1999) J Chem 23:185–191
38. Guzel F, Uzun I (2000) Adsorption of some heavy metal ion from Aqueous solution by activated
Carbon and comparison of percent adsorption result of activated Carbon with those of some
other adsorbent. Turk J Chem 24:291–297
39. Yener J, Aksu Z (1999) Atıksulardaki fenol ve klorofenollerin aktif karbon ve kurutulmuş aktif
çamura adsorpsiyonu. Tr J Eng Environ Sci 23:93–104
40. Zondlo JW, Velez MR (2007) Development of surface area and pore structure for activation of
anthracite coal. Fuel Process Technol 88(4):369–374
41. Prauchner MJ, Reinoso FR (2008) Microp Mesop Mater 109:581–584
42. Najibi H, Chapoy A, Tohidi B (2008) Methane/natural gas storage and delivered capacity for
activated carbons in dry and wet conditions. Fuel 87:7–13
43. Ridha FN, Yunus RM, Rashid M, Ismail AF (2007) Exper Ther Flu Sci 32:14–22
44. Celzard A, Perrin A, Albiniak A, Broniek E, Mareche JF (2007) The effect of wetting on pore
texture and methane storage ability of NaOH activated anthracite. Fuel 86(1–2):287–293
45. Ouederni SBYA (2012) Hydrocarbons gas storage on activated Carbons. Int J Chem Eng Appl
3(3), 81:1777–1803 (2007)
46. Castello DL, Amoros DC, Solano AL, Quinn DF (2008) Carbon 40:989–1002
47. Abubakar M, Alechenu AA, Manase A, Mohammed J (2012) Adv Appl Sci Res 3(5):3089–3096
48. Aznar JS (2011) ARPN J Eng Appl Sci 10(1):376–386
49. Evans MJB, Halliop E, MacDonald JAF (1999) Carbon 37(2):269–274
Combined Casing Groove and Blade Tip
Treatment for Wave Energy Harvesting
Turbine

P. Madhan Kumar , Paresh Halder , and Abdus Samad

Abstract The Wells turbine is a self-rectifying air turbine, used in oscillating water
column (OWC) to harvest wave energy. It produces unidirectional torque as the flow
oscillates inside the OWC chamber. It has inherent disadvantage of narrow operating
range due to stall at high airflow rate. Whereas, a wider operating range is essential
to improve the turbine power output. A casing groove modifies the tip leakage flow
pattern and improves the operating range. In addition, a radiused tip can alter the
tip leakage flow and delay the stall. To enhance the performance further, this paper
investigates the combined effect of tip groove and radiused tip (CG&RT) design
modification. The flow was simulated by solving steady, incompressible Reynolds
averaged Navier–Stokes equations in Ansys CFX 15.0. As expected, the CG&RT
blade enhanced the relative operating range and the turbine power output by 44.4%
and 23.8%, respectively.

Keywords Wells turbine · Casing groove · Radiused blade · Wave energy · OWC

1 Introduction

Wells turbine combined with oscillating water column (OWC) is used to extract power
from ocean wave (Fig. 1). It comprises of symmetrical blades aligned perpendicular
to the flow direction. It suffers narrow operating range due to flow separation at a

P. Madhan Kumar · P. Halder · A. Samad (B)

Wave Energy and Fluids Engineering Laboratory, Department of Ocean Engineering, Indian
Institute of Technology Madras, Chennai 600036, India
e-mail: samad@iitm.ac.in
P. Madhan Kumar
e-mail: madyypts@gmail.com
P. Halder
e-mail: paresh.halder@gmail.com
P. Halder
Quantum Wave Microscopy Unit, Okinawa Institute of Science and Technology Graduate
University, 1919-1 Tancha, Onna-son, Okinawa 904-0495, Japan
© Springer Nature Singapore Pte Ltd. 2021 1027
S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_89
1028 P. Madhan Kumar et al.

Fig. 1 Wells turbine Uni-directional rotation

Hub

Symmetrical blades
low
airf
lic
Cyc

higher flow rate. The airflow inside the OWC is cyclic, and it varies from minimum
velocity to maximum and vice versa. Hence, it is necessary to have a wider operating
range to improve the energy harvest of the turbine.
Many researchers actively pursued different design modifications to enhance the
operating range of Wells turbine. Raghunathan and Tan [1] introduced a biplane Wells
turbine to improve the stall characteristics. The kinetic energy losses at downstream
of Wells turbine can be reduced by implementing guide vanes. The Wells turbine
with guide vanes also showed improved starting characteristics [2]. The blades with
increased thickness improved starting characteristics [3]. Raghunathan and Beattie
[4] used contra-rotating Wells turbine as an alternative to biplane configuration. They
concluded that contra-rotating is superior to biplane turbine regarding operating
range and efficiency. Webster and Gato [5] optimized blade profile to improve the
stall characteristics. Gato and Webster [6] studied Wells turbine with blade sweep
and guide vanes. They reported an enhanced operating range with reduced efficiency.
The blade with variable chord with guide vanes showed better starting character-
istics [7]. The turbine performance can be improved by modifying blade setting angle
[8]. Takao et al. [9] implemented end plates to improve the performance. Mohamed
et al. [10] performed multi-objective optimization to improve efficiency and tangen-
tial force coefficient. The tangential force and efficiency improved by 8.8% and 1%,
respectively. Shaaban and Hafiz [11] studied the Wells turbine with duct geometry.
The optimized duct geometry enhanced the turbine power and efficiency by 14%
and 9%, respectively. The downstream duct suppressed the interaction of tip leakage
flow (TLF) and the suction side (SS). The nature of the interaction between TLF and
primary flow affects the stall characteristic of Wells turbine. Halder et al. [12, 13]
Combined Casing Groove and Blade Tip Treatment for Wave Energy… 1029

Table 1 Turbine geometry

Blade profile NACA0015
[16]
Maximum thickness 15% chord length
Number of blades 8
Chord length (C) 125 mm
Tip radius (Rtip ) 300 mm
Hub radius 200 mm
Solidity 0.64
Tip clearance 1.25 mm (1%C)
Angular velocity 209.44 rad/s

introduced a casing groove treatment to modify the TLF behaviour, thereby delay-
ing stall phenomenon. It improved the turbine power output by 26%. Nazeryan and
Lakzian [14] used a variable thickness blade to weaken the interaction between TLF
and blade SS. The variable thickness blade enhanced torque and operating range.
Kumar et al. [15] used a radiused edge blade tip (REBT) to enhance the stall margin
of Wells turbine. REBT showed weak TLF and primary flow (PF) interaction and
delayed stall.
As explained above, various design modifications are proposed to alter the char-
acteristics of the interaction between TLF and PF. In this study, a combined casing
groove and blade tip (CG&RT) blade are investigated to study their effect on Wells
turbine performance. The performance of the CG&RT blade is computed by solv-
ing steady, and incompressible Reynolds averaged Navier–Stokes (RANS) equations
using commercial CFD solver ANSYS CFX 15.0. The performance parameters of
the CG&RT blade are compared with the reference blade. Moreover, the flowfield
analyses has been explained using the post-processed figures.

2 Reference Geometry

The reference geometry was taken from Torresi et al. [16]. Figure 1 shows the Wells
turbine with eight blades. It has a unidirectional rotation for the cyclic airflow because
of the symmetrical blades. Table 1 provides the details of the reference blade.

3 Casing Groove and Blade Tip Treatment

The reference geometry was CG&RT using CAD software to implement the casing
groove and RETB (Fig. 2). The geometrical parameters for the casing groove were
taken from the study of Halder et al. [12]. The width and depth of the casing groove
were chosen as 40%C and 2%C, respectively. The tip diameter was maintained
1030 P. Madhan Kumar et al.

Shroud 40%C

2%C
1%C

Blade
Fig. 2 Reference (left) and CG&RT (right) geometry

constant for both geometries, and the casing diameter was increased to accommodate
the groove. Additionally, the radius of fillet used in blade is taken as 1%C as suggested
in the work of Kumar et al. [15].

4 Numerical Methodology

The performance parameters of the CG&RT turbine was computed by solving 3D

steady, incompressible RANS equations. A single blade passage was selected as
computational domain (Fig. 3). The periodic boundary conditions are applied to the
lateral faces to account for the full domain. ANSYS ICEM CFD 15.0 was used
to discretize the computational domain with unstructured tetrahedral elements. The
mesh closer to the blade vicinity was provided with higher refinement. In addition,
to resolve the viscous sub-layer region, 20 prism layers with stretching ratio of 1.2
Combined Casing Groove and Blade Tip Treatment for Wave Energy… 1031

Inlet
Shroud

Blade

Hub
Periodic
C
4C interface

Outlet

6C

Fig. 3 Computational domain for flow analyses

were created around the blade. The initial cell height was fixed as 1E-5 m to retain
the y + < 1.
The boundary conditions were imposed in ANSYS CFX-Pre. Table 2 provides
the applied boundary conditions, and the simulations were run in ANSYS CFX 15.0.
High-resolution advection scheme with second-order accuracy is used for spatial
discretization. The k-w SST model was chosen as the turbulence model. It employs
k-w model closer to wall and k-ε model far from the wall [17]. Double precision was
enabled for all simulations to minimize the round-off error.

Table 2 Applied boundary

Working fluid Air
conditions
Nature of flow Incompressible
Reference pressure 1 atm
Inlet Velocity inlet (4–18 m/s)
Outlet Pressure outlet
Blade and hub No-slip wall
Shroud Contra-rotating no-slip wall
Turbulence intensity 5%
Residual criteria 1e–5
Mass imbalance 0.001%
1032 P. Madhan Kumar et al.

Table 3 GCI estimation

Number of elements (106 ) Fine, medium and coarse 6.80, 3.07, 1.39
Average cell size (h) h1 , h2 , h3 0.0052, 0.0068, 0.0089
Grid refinement factor (r) r 21 , r 32 1.3, 1.3
Performance parameter (φ) φ1, φ2, φ3 0.0427, 0.0422, 0.0402
Apparent order p 5.47
Approximate relative error ea21 , ea32 0.0114, 0.0495
Extrapolated values φext
12 , φ 32
ext 0.04294, 0.04295
Extrapolated relative error 21 , e32
eext 0.0035, 0.0152
ext
GCI GCI21
fine , GCI32
medium 0.0043, 0.0192

5 Results and Discussion

5.1 Grid Convergence and Error Estimation

The grid convergence index (GCI) was used to estimate discretization error and
grid convergence. The methodology suggested by Celik et al. [18] is followed to
determine GCI. The GCI can be expressed as

1.25ea21
GCI = p (1)
r21 − 1

where ea21 is the approximate relative error given by φ2φ−φ

1
1
, p is the apparent order of
the mesh employed for the numerical study. In addition, r 21 is the ratio between the
average cell size of the fine and coarse meshes. Table 3 displays the calculation of
GCI.
The GCI decreases with an increase in grid resolution, and it implies that the
discretization error reduces with an increase in grid refinement. The GCI value of
the medium grid is less than 2%. Hence, it can be assumed that the grid convergence
was achieved and the medium grid was retained for all the computations performed
in this study to reduce computational power.

5.2 Numerical Validation

The present numerical results were validation against the experimental [19] and
numerical results [20, 16] available in open literature. The non-dimensional flow
Combined Casing Groove and Blade Tip Treatment for Wave Energy… 1033

Fig. 4 Numerical validation Exp-Curran & Gato (1997)

CFD-Toressi et al. (2004)
1
CFD-Toressi et al. (2008)
Δp*
CFD-Present result
0.75

η, Δp*, T*
η
0.5

0.25 T*

0
0 0.1 0.2 0.3
U*

     
coefficient U ∗ = UUtipA , torque T ∗ = ρωT2 R 5 , pressure  p ∗ = ρω2 Rp 2 and effi-
  tip tip
Tω
ciency η = Q p
were chosen as the parameters to describe Wells turbine perfor-
mance. The notation U A is inlet axial velocity (m/s), U tip is tip velocity (m/s), T is
turbine torque (Nm), Q is the volumetric discharge (m3 /s), and p is static pressure
drop (N/m2 ). Figure 4 shows the validation of present numerical results with the
existing results. It can be seen that the present CFD result matches well with the
existing literature.

5.3 Aerodynamic Performance of CG&RT Blade

Figure 5 displays the performance parameters of the reference and the CG&RT blade.
At low U * , the reference blade shows higher torque compared to the CG&RT blade.
However, the reference blade experiences stall at U * = 0.225. On the contrary, the
CG&RT blade stalls at U * = 0.325 and produces higher torque than reference blade at
U * > 0.225. The reduced pressure drop in the CG&RT blade explains the decreased
torque output. Similarly, the reference blade exhibits higher efficiency compared
to CG&RT blade up to U * = 0.225, and later it reduces due to stall condition.
Furthermore, the CG&RT blade enhanced the relative stall margin and peak torque
by 44.4% and 23.8%, respectively.

5.4 Flowfield Analyses

Figure 6 shows the pressure coefficient (Cp) distribution at blade midspan for dif-
ferent flow coefficients (U * ). The area enclosed by the Cp curve provides the lift
force. At low U * = 0.075, area of the Cp curve is small, which indicates lesser
 1034

0.25
Reference Reference
0.2 CG&RT 0.9 CG&RT

0.15
0.6

T*
Δp*
0.1

0.3
0.05

0 0
0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4
U* U*

0.6 Reference
CG&RT

0.4

η
0.2

0.0
0.0 0.1 0.2 0.3 0.4
U*

Fig. 5 Performance comparison of reference and modified blade

P. Madhan Kumar et al.
Combined Casing Groove and Blade Tip Treatment for Wave Energy… 1035

Fig. 6 Pressure coefficient

distribution at blade midspan
1.0

0.0

Cp
-1.0 Reference
CG&RT
-2.0
0.0 0.2 0.4 0.6 0.8 1.0
X/C
U*=0.07
2.5

-0.5
Cp

-3.5

-6.5 Reference
Reduced blade loading CG&RT

-9.5
0.0 0.2 0.4 0.6 0.8 1.0
X/C
U*=0.22
2.5

-0.5

-3.5
Cp

Stall
-6.5 Reference
CG&RT

-9.5
0.0 0.2 0.4 0.6 0.8 1.0
X/C
U*=0.2

torque. As the U * increases, the area enclosed by Cp curves increases which signi-
fies lift increase. As revealed in Fig. 6, the reference blade shows higher pressure
variation at blade leading edge (LE) compared to the CG&RT blade. The reduced
suction at the CG&RT blade explains the decreased torque shown in Fig. 5. At U *
= 0.225, the suction pressure at LE increases for both blades. However, the CG&RT
blade shows reduced blade loading near the LE and trailing edge (TE). To generate
1036 P. Madhan Kumar et al.

maximum torque, the enclosed Cp curve area should be higher, and the CG&RT
blade shows reduced area and consequently reduced torque compared to reference
blade. At increased U * (0.275), for the reference blade, the suction peak vanishes
and the pressure distribution is almost same along the chord. It implies stall condi-
tion because of the flow separation. In contrast, the CG&RT blade shows increased
suction pressure and higher torque.
Figure 7 displays the Cp contour at blade midspan for different U * . At U * =
0.075, a maximum pressure region is visible near the blade LE for both blades. The
CG&RT blade exhibits reduced pressure at blade LE compared to the reference blade.
It can be noticed that the high pressure region at blade LE shifts towards the TE with
increase in U * . At U * = 0.225, the pressure difference at blade LE is higher for the
reference blade compared to the CG&RT blade, and it corroborates the results shown
in Fig. 6. At U * = 0.275, for reference blade, a low-pressure region engulfs the blade
SS, and it signifies flow separation due to adverse pressure gradient. In contrast, in
CG&RT blade the low-pressure region is confined near the blade LE and produces
higher torque compared to reference blade.
Figure 8 illustrates the axial velocity distribution along the normalized span at
blade LE location. At U * = 0.075, the CG&RT blade shows reduced axial velocity at
hub and tip in contrast to the reference blade. The decreased axial velocity reduces the

Fig. 7 Pressure coefficient Reference CG&RT

contour at blade midspan
Casing
LE SS TE
PS

U*=0.075

Searated flow U*=0.225 Attached flow

U*=0.275

Cp [-]
Combined Casing Groove and Blade Tip Treatment for Wave Energy… 1037

Fig. 8 Normalized span 1

axial velocity distribution Tip
near blade leading edge (LE)

Normalized span
0.75

0.5

0.25
Reference
CG&RT Hub
0
0 5 10 15 20
Axial velocity [m/s]
U*=0.075
1

0.75
Normalized span

0.5

0.25
Reference
CG&RT
0
0 10 20 30 40 50
Axial velocity [m/s]
U*=0.225
1
Normalized span

0.75

0.5

0.25 Reference
CG&RT
0
0 10 20 30 40 50
Axial velocity [m/s]
U*=0.275

relative velocity and angle of attack which reduces the torque generated as shown in
Fig. 5. The casing groove modification is the reason behind the reduced axial velocity
near tip in the CG&RT blade. Moreover, the axial velocity increases with an increase
in U * . At U * = 0.225, the axial velocity increases for both blades and the CG&RT
blade shows reduced axial velocity compared to the reference case. Furthermore, at
1038 P. Madhan Kumar et al.

U * = 0.275, the reference blade shows increased axial velocity near the tip, and it
indicates stall. Higher axial velocity produces a high angle of attack which signifies
flow separation. In contrast, the CG&RT blade shows reduced axial velocity, and it
is responsible for the delayed stall.
Figure 9 displays the tip leakage vortex at the blade mid-chord section for various
U * . The TLF occurs owing to the pressure variation between the pressure side (PS)
and SS of the tip. Consequently, it rolls into a vortex on the blade SS known as

Fig. 9 Tip leakage vortex at Reference CG&RT

blade mid chord section
Tip

PS SS

Hub
Tip leakage vortex Tip leakage vortex
U*=0.075

Tip leakage vortex Tip leakage vortex

U*=0.225

Tip leakage vortex Tip leakage vortex

breakdown
U*=0.275
Combined Casing Groove and Blade Tip Treatment for Wave Energy… 1039

tip leakage vortex. At U * = 0.075, a clockwise tip leakage vortex is visible on the
SS of blade tip for both blades. As revealed in figure, the strength of tip leakage
vortex increases with U * . The behavior of tip leakage vortex in the CG&RT blade
is altered due to the casing groove and radiused edge tip. At U * = 0.275, the tip
leakage vortex breakdowns in reference blade and the increases the blockage causing
stall. The presence of severe adverse pressure gradient causes the tip leakage vortex
breakdown. In contrast, the tip leakage vortex is stable in the CG&RT blade. It is
confined near the blade tip and offers less blockage to the primary flow. Moreover,
the casing groove increases TLF, and it explains the reduced efficiency (Fig. 5) as
the fluid passing through the tip clearance will not contribute to energy extraction.
Furthermore, coupled with the casing groove and the radiused edge tip can be done
to alleviate the decrement in efficiency.

6 Conclusion

The numerical investigation of Wells turbine with casing groove and radius edge
tip was presented in this study. The problem was solved in ANSYS CFX 15.0,
and the performance parameters obtained were compared with the reference blade.
Moreover, the pressure coefficient, axial velocity distribution and tip leakage vortex
were analysed. The prominent conclusions drawn were listed below.
• The blade with casing groove and radiused edge tip showed a wider operating
range and enhance power output.
• The relative peak torque and stall margin were enhanced by 44.4% and 23.8%,
respectively.
• The CG&RT blade displayed reduced axial velocity at the tip which decreased
the relative velocity and angle of attack, thereby delaying stall.
• Moreover, optimization of the combined casing groove and radiused edge tip
parameters should be performed for further performance enhancement.

References

1. Raghunathan S, Tan CP (1983) Performance of biplane wells turbine. J Energy 7(6):741–742

2. Inoue T, Kaneko M, Setoguchi K (1985) Studies on wells turbine for wave power generator:
part 3; effect of guide vanes. Bul JSME 28(243):1986–1991
3. Raghunathan S, Tan CP (1985) Effect of blade profile on the performance of the Wells self-
rectifying air turbine. Int J Heat Fluid Flow 6(1):17–22
4. Raghunathan S, Beattie WC (1996) Aerodynamic performance of contra-rotating wells turbine
for wave energy conversion. Proc Inst Mech Eng Part A J Power Energy 210(6):431–447
5. Webster M, Gato LMC (2001) The effect of rotor blade shape on the performance of the Wells
turbine. Int J Offshore Polar Eng 11(3):227–230
1040 P. Madhan Kumar et al.

6. Gato LMC, Webster M (2001) An experimental investigation into the effect of rotor blade
sweep on the performance of the variable-pitch Wells turbine. Proc Inst Mech Eng Part A J
Power Energy 215(5):611–622
7. Govardhan M, Chauhan VS (2007) Numerical studies on performance improvement of self-
rectifying air turbine for wave energy conversion. Eng Appl Comput Fluid Mech 1(1):57–70
8. Kaneko K, Setoguchi T, Santhakumar S, Takao M, Kim TH (2003) A modified Wells turbine
for wave energy conversion. Renew Energy 28(1):79–91
9. Takao M, Setoguchi T, Kinoue Y, Kaneko K (2007) Wells turbine with end plates for wave
energy conversion. Ocean Eng 34(11–12):1790–1795
10. Mohamed MH, Janiga G, Pap E, Thevenin D (2011) Multi-objective optimization of the airfoil
shape of Wells turbine used for wave energy conversion. Energy 36(1):438–446
11. Shaaban S, Hafiz AA (2012) Effect of duct geometry on Wells turbine performance. Energy
Convers Manag 61:51–58
12. Halder P, Samad A, Kim JH, Choi YS (2015) High performance ocean energy harvesting
turbine design—a new casing treatment scheme. Energy 86:219–231
13. Halder P, Samad A (2015) Casing treatment of a wave energy extracting turbine. Aquatic
Procedia 4:516–521
14. Nazeryan M, Lakzian E (2018) Detailed entropy generation analysis of a Wells turbine using
the variation of the blade thickness. Energy 143:385–405
15. Kumar PM, Halder P, Samad A (2018) Performance analysis of Wells turbine with radiused
blade tip. In: ASME 2018 37th international conference on ocean, offshore and arctic
engineering. OMAE2018, pp 1–10
16. Torresi M, Camporeale SM, Strippoli PD, Pascazio G (2008) Accurate numerical simulation
of a high solidity Wells turbine. Renew Energy 33(4):735–747
17. Menter FR (1994) Two-equation eddy-viscosity turbulence models for engineering applica-
tions. AIAA J 32(8):1598–1605
18. Celik IB, Ghia U, Roache PJ, Freitas CJ, Coleman H, Raad PE (2008) Procedure for estimation
and reporting of uncertainty due to discretization in CFD applications. J Fluids Eng 130(7)
19. Curran R, Gato LMC (1997) The energy conversion performance of several types of Wells
turbine designs. Proc Inst Mech Eng Part A J Power Energy 211(2):133–145
20. Torresi BFM, Camporeale SM, Pascazio G (2004) Fluid dynamic analysis low solidity Wells
turbine, 59 Congresso ATI, Genvo, Italy, pp 277–288
Design and Development of Wind Tunnel
to Study Smoldering Combustion

S. Sanjana, M. Rakshantha, Yeleti Bunny Venkat, and B. T. Kannan

Abstract Smoldering is flameless combustion that occurs on the surface of the

condensed phase that is very difficult to detect and which causes serious fire risks in
residential and industrial areas by the presence of turbulent winds. Statistics show
that major fire accidents are due to smoldering for example cigarettes. Smoldering
combustion is also a great concern in space facilities like ISS since the zero gravity
environments can increase the rate of smoldering propagation. The main motivation
of this paper is to design and develop the wind tunnel in order to study the effect
of aerodynamic and combustible properties on the smoldering. A wind tunnel is an
effective tool that is used to study the aerodynamic effects of air or smoke when it is
allowed flowing past the solid bodies. The specific measurements that are required
to construct the wind tunnel are mentioned in this paper, as well as the materials
used in various sections of the wind tunnel are discussed. The power source and light
source used for the visualization of the turbulent wind flow past the burning fuel are
discussed. A low-speed tabletop wind tunnel is first designed using CATIA and then
fabricated using necessary tools. The results from these experiments can be used to
know how the combined effect of both laminar and turbulent flow on smoldering
combustion is more harmful than the normal combustion.

Keywords Smoldering · Combustion · Wind tunnel · Design · Flow ·

Aerodynamics

S. Sanjana · M. Rakshantha · Y. B. Venkat · B. T. Kannan (B)

Department of Aerospace Engineering, SRM Institute of Science and Technology, Kattankulathur,
Kancheepuram District 603203, Tamil Nadu, India
e-mail: skyinventorbt@gmail.com
S. Sanjana
e-mail: sanjanasanjay1155@gmail.com
M. Rakshantha
e-mail: rakshanthamohan97@gmail.com
Y. B. Venkat
e-mail: bunnyyeleti@yahoo.com

© Springer Nature Singapore Pte Ltd. 2021 1041

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_90
1042 S. Sanjana et al.

1 Introduction

Kinbara et al. [1] did a study on smoldering combustion downward propagation

through solid materials. The smoldering combustion of sold materials like the paper
was conducted on the surface supported by the air to study the propagation mech-
anism. The sample ahead of the burning front when the smoldering combustion
propagates upward is found to be ascending due to the combustion gases and the
propagation velocity is found to be accelerating and resulting in a wide range of
values. The samples used for studying the downward propagation of smoldering
combustion were a circular rod, incense sticks, and cardboard. They concluded that
the combustion gases found on the sample were the reason for the propagation veloc-
ity increase with respect to the depth of the layer. Ohlemiller et al. [2] worked on
permeable fuel beds to conduct experiments on forward smoldering propagation and
reverse smolder propagation. Wood fiber and polyisocyanurate were used as perme-
able fuels. The sample bed is tangentially insulated in the dispersion section where
the uniform flow of oxygen or nitrogen mixture takes place. This study showed
that the reverse smolder is 10 times faster than thse forward smolder because of
conduction and radiation. And they also concluded that the high mass flow rate of
oxygen increases the consumption of char in both the forward and reverse smolder-
ing. Madara et al. [3] submitted a paper to a journal where the primary focus was on
the design and fabrication process of a small-scale subsonic type wind tunnel which
is an open circuit used to visualize the flow over an object and calculate the lift
and drag coefficient using the experiments. CFD simulations give imperative results
whereas the low circuit subsonic wind tunnel experiments provide a visual under-
standing of the impact of wind on different shapes. The test area was constructed
based on the geometry of the object with a power source and instrumentation for
weight estimation. The profile of the test area includes the cone area and square
test area of different dimensions based on the geometry with the mean velocity of
15 m/s. The straight portion before the test section was made in order to stabilize
the flow before it reaches the testing area. The estimation was made on the wind
tunnel without the aerodynamic object which showed the uniform field which is the
basic requirement for the study of streamlined bodies. The results calculated in a
low-speed subsonic type wind tunnel were nearly accurate to the results calculated
using NASA and MIT wind tunnel.
Mukunda et al. [4] did research on experiments and modeling of smoldering
combustion on incense sticks. The experimental and modeling studies were done on
smoldering as a function of air, velocity, and size. The results from both experimental
and modeling cases were based on the surface smoldering combustion due to heat
transfer of free and forced convection. Blower type wind tunnel was made and the
mixed air of desired oxygen was supplied into the mixing chamber of the wind
tunnel. The sample holder was used to hold the incense sticks and the anemometer
was used to measure the flow inside the wind tunnel. Experimental data on the effects
of smoldering combustion show that the forward smoldering increases with stream
speed and for the reverse smoldering the spread rate are more than the forward
Design and Development of Wind Tunnel to Study Smoldering… 1043

smoldering. The reduction in the smoldering rate in the recirculation zone is due
to the reduction of mass of the oxidizer. Bar-Ilan [5] did an experiment on small
polyurethane foam to study the transition phenomenon of forward smoldering to
flaming. The experiment was carried out on the polyurethane foam in order to find
the velocity flow effect, oxygen concentration and the radiation flux on the transition
at microgravity and low gravity environment. The samples of small size are selected
because the smoldering combustion can self-propagate as the heat is lost to the
environment. The experiments are carried out in a wind tunnel where the fuel is
placed vertically. The lateral sides of the three samples were maintained at some
peak temperature whereas the other three sides were maintained to the upward flow
and radiation flux. From these experiments, it was found that there is a delay in
the transition of smoldering to flaming by decreasing the flow velocity, and also by
increasing the concentration of oxygen and radiation flux. They concluded that the
transition can occur in the char region that was left behind by the smoldering reaction
with the appropriate external conditions.
Buoyancy, Turbulence, and Combustion in Fire Field Modeling were discussed
by Gao et al. [6]. The main purpose of this study was to known the effect of buoyancy,
turbulence and combustion on the fire field model, for example, a test section in the
form of the compartment with a window in the side and door in front. Simulations
were carried out using Computational Fluid Dynamics (CFD). Two cases were stud-
ied under different operating conditions. Case A simulations were carried out with
fire at the center of the room. Case B simulations were carried out at the corner of
the room. Both cases were compared. The flame height is determined by knowing
whether the fuel is burnt or not. The results from case B showed that because of the
closed surface area the flame is height is more thereby increasing the upper ceiling
temperature. In Case A, the flame height is less and doesn’t reach the ceiling as the
combustion propagates uniformly in all directions. They concluded that this is the
reason for the combustion reactions to take place at the ceiling where the heat release
rates are higher.
Schult et al. [7] did a study on porous material to know the propagation and
extinction of smoldering waves caused by the opposed flow that was artificially
generated. The main reason to use porous material is to allow the mixture of gas
(oxygen and nitrogen) to flow through the pores which are present in the porous
material. The cylindrical samples are considered where the gases are allowed to flow
through one end. In the other end, smolder wave is initiated and it flows in the opposite
direction to the flow of the oxidizer. In this paper, they have used graphical methods
to learn about uniform propagation and smolder wave solutions. These results were
compared with the experimental results which were taken by the same author in
his other paper. The analytical results in this paper showed that the propagation
velocity increases with an increase in mass flow rate which perfectly fits with the
experimental results. Stocker et al. [8] did smoldering combustion experiments on
flexible polyurethane foam in microgravity. The experiments were carried out in both
microgravity and normal gravity and the results were compared. Four cases were
conducted by varying the shapes and the environment both for micro and normal
gravity. In this experiment temperature was measured along with the fuel sample at
1044 S. Sanjana et al.

different locations, the video recording was done and post analyses were done on
the char and gas composition. They found that heat loss was less in microgravity
due to no natural convection in microgravity. This resulted in an increase in ignition
temperature in the microgravity. They concluded that the burning of polyurethane
foam gave rise to combustion products like carbon monoxide which is hazardous in
space.
Hadden [9] did a study on the initiation of smoldering combustion and its spread
in peat. The experimental study was done on the vertical samples of peatlands. The
top surface of the sample was ignited by the radiation so that the smoldering com-
bustion propagates downward in the direction opposite to the flow of oxidizer. The
experiments were carried by varying the ignition conditions and the oxygen con-
centrations. The pyrolysis reaction and the oxidation reactions are the parameters
focused in this paper. The char is formed in the pyrolysis oxidation and the char is
converted into ash in the oxidation reactions. It was found that the smoldering front
thickness depends mainly on the low mass flow rate and the mass of the leftover
residues. The wider smoldering front is required to sustain the smoldering combus-
tion. They concluded that the decreased oxygen concentration environment with less
mass flow rate is required to avoid fire accidents. Sato et al. [10] conducted an experi-
ment on cellulosic powder material to study the smoldering spread. The experimental
set up involved a high bulk density cellulosic powder bed over which a stream of air
was supplied in order to study the smolder spread in a horizontal layer. The results
involved the 2D structure of a spread zone, spread rate, cross-sectional views of the
burning region and temperature distributions. From this experiment, it was observed
that the forward smoldering increases the velocity of the air and also increases the
spread rate of the fuel. But above a certain value of the airstream velocity, the spread
rate becomes independent and the smolder zone disappears. From experimental data,
it is found that the balance of heat between the oxidation zone and unburnt part of
the fuel bed is important to the behavior of smoldering spread. They also concluded
that the aerodynamic effects and heat transfer control the behavior of spread rate
over the smolder zone. The design and construction of wind tunnel were made to
understand the combustion better and the behavior of smolder under the influence
of forced flow effect by taking different parameters, i.e., shaped of fuel, different
airspeeds, different sample holders and also the introduction of flow visualization to
study the aerodynamic behavior of the different shapes under smoldering.

2 Wind Tunnel Design

2.1 Geometry and Meshing

The tabletop wind tunnel was initially designed using CATIA V5. The detailed design
was carried out in order to give a clear picture of the fabrication process. The finalized
CAD design of the tabletop wind tunnel is seen in the Figs. 1, 2 and 3.
Design and Development of Wind Tunnel to Study Smoldering… 1045

Fig. 1 Top View

Fig. 2 Side View

Fig. 3 The finalized CAD model of tabletop wind tunnel

The whole length of the wind tunnel is 129 cm and the maximum width is about
29 cm for the inlet to fix the axial fan and 20 cm uniform length for other sections.
Inlet, contraction cone, flow conditioning unit, test section, and outlet are the different
sections of the wind tunnel. Special attention was given on the turbulence intensity,
uniformity, and flow control.
1046 S. Sanjana et al.

Fig. 4 Axial Flow Fan

3 Wind Tunnel Design

3.1 Axial Fan

The axial fan is fixed with the speed regulator of 5 speeds that allow the airflow to
flow in an axial direction. The fan draws the air from the environment and blows
into the wind tunnel. The speed reaches up to 4.7 m/s. The fan rotates to produce
a pressure difference across it as the flow is axial in both entry and exit. The axial
flow fan used in the present work helps in producing the flow inside the wind tunnel
(Fig. 4).
This facilitates in studying the effect of both the laminar and turbulent flow on
the smoldering combustion. The static rise is produced in order to compensate for
the total pressure in the wind tunnel.

3.2 Contraction Cone

The contraction cone is the second part of the tabletop wind tunnel used to study the
smoldering combustion. The contraction cone is used to accelerate the flow and allow
the flow to pass into the test section. This section effectively reduces the velocity
variation and helps in maintaining the flow uniformity across the length of the wind
tunnel. The decrease in the area along the length of the contraction cone helps in
increasing and decreasing velocity and pressure respectively.
Design and Development of Wind Tunnel to Study Smoldering… 1047

Fig. 5 Honeycomb structure

3.3 Flow Conditioning Unit

Flow conditioning unit is used to straighten the air that was drawn from the surround-
ing environment and it is also used to channel the ambient air into the test section.
It comprises of mesh and honeycomb structures. The flow conditioning unit reduces
the turbulence of the air. The main purpose of the mesh is to avoid unnecessary
items to enter the tunnel. There are different types of honeycomb structures used
in the low-speed wind tunnel. For example, honeycomb structures with a circular
cross-section, rectangular cross-section and hexagonal cross-section. The type of
honeycomb structure used in the present work is of the circular cross-section. The
cells are aligned in such a manner that it is in the streamwise direction. The honey-
comb structures reduce the transverse velocity of the flow by reducing the pressure
loss (Figs. 5 and 6).

3.4 Test Section

The test section is the basic and most important component of the wind tunnel. The
incense materials of different shapes are placed inside the test section of the wind
tunnel to study the effect of flow over them. The test section of the rectangular cross-
section is used in this paper. The test section is provided with necessary facilities
1048 S. Sanjana et al.

Fig. 6 Mesh

like glass for visual ability, laser light for flow illumination, sample holder, digital
microscope as a camera, LED light and a black screen to capture clear images (Fig. 7).

Fig. 7 Test section

Design and Development of Wind Tunnel to Study Smoldering… 1049

Fig. 8 Fabricated model

4 Fabrication of the Wind Tunnel

The fabrication of the tabletop wind tunnel was done using wood of thickness
180 mm, nuts, bolts, nails, glass in the test section, mesh and straws for honey-
comb structures. Woodcutter, tape measures, hammer, chisel, nail gun, and clamps
are some of the tools used in the fabrication process. The fan is the basic component
with respect to which the design and fabrication were done. The contraction cone
was the difficult part to fabricate and it was fabricated so that the fan is installed in
such a manner that no vibration takes place. Similarly, the flow conditioning unit and
the outlet were fabricated. Intense care was taken in fabricating the test section of
the tabletop wind tunnel as it was fabricated using glass. The parts were fabricated in
such a way that the parts can be dismantled for easier transportation, to upgrade the
modification of tunnel for various experiments and installed for the experimentation
(Fig. 8).

5 Schematic of Tabletop Wind Tunnel

The schematic diagram (see Fig. 9) gives a clear picture of experimental set up used
to visualize the effect of flow on the smoldering combustion. The laser light of green
color helps in flow visualization inside the wind tunnel test section. The camera used
is a digital microscope that is connected to the computer to record the video.

6 Experimental Setup and Calculations

A sudden impact of the flow on the solid object, when placed in 90°, may not withstand
the flow and can create a disturbance. In order to avoid such circumstances, the ramp
(see Fig. 10) was placed inside the test section of the wind tunnel to develop a smooth
1050 S. Sanjana et al.

Fig. 9 Schematic diagram of tabletop wind tunnel

Fig. 10 Sample holder for boundary layer formation

boundary layer. The different types of sample holders were used for different sample
shapes. Three possible cases were employed inside the test section to carry out the
experimentations. They were along with the flow, transverse to the flow and against
the flow. The angle can be varied by using the string as shown in (see Fig. 13). The
velocity inside the test section was measured using the vent anemometer. The width
of the test section was used, and the Reynolds number had been calculated for the
test section (Figs. 11, 12, 13 and 14) (Table 1).

7 Results and Conclusion

The tabletop wind tunnel was initially designed using CATIA and then fabricated
using different materials and tools. Axial fan, contraction cone, flow conditioning
unit, test section, and outlet are the components of the wind tunnel used. The com-
ponents of the tabletop wind tunnel were fabricated separately and then assembled
to perform experiments. The incense materials with different shapes were used for
the study. Sample holders based on different orientations and shapes were used. The
experimentations were carried under normal gravity conditions (21% oxygen). The
fabricated tabletop wind tunnel was used to study the effect of laminar and turbulent
flow on the smoldering combustion using incense material. Five different fan speeds
were employed over a burning incense material that was placed along with the flow,
transverse to the flow and against the flow. The material is marked at regular intervals
Design and Development of Wind Tunnel to Study Smoldering… 1051

Fig. 11 Sample holders when placed against the flow

of fixed distance each along the whole length. The regression rates were calculated.
The results obtained from the forced flow smoldering combustion experiments were
compared with the results of natural smoldering combustion experiments. The nat-
ural smoldering combustion results were taken for three different orientations. They
were vertical (90°), horizontal (180°), and inverted (270°). In conclusion, the tabletop
wind tunnel was successfully designed and developed to study the effects of forced
flow on the smoldering combustion. The burn rates of the fuel calculated were used
to predict how smoldering combustion is more harmful than normal combustion. In
addition, the tabletop wind tunnel can be used to conduct furthermore experiments
1052 S. Sanjana et al.

Fig. 12 Sample holder when placed along with the flow

Fig. 13 Sample holders

Design and Development of Wind Tunnel to Study Smoldering… 1053

Fig. 14 Sample holders with burning incense materials inside the test section

Table 1 Different speeds

Speed regulator Speed (m/s) Reynolds number inside the
inside the wind tunnel
test section
1 0.8 10.256 × 103
2 1.8 23.077 × 103
3 2.2 28.205 × 103
4 2.3 29.487 × 103
5 2.4 30.769 × 103

to analyze and avoid the causes of residential fires, industrial fires, forest fires, rocket
fires, and building fires.

Acknowledgements The authors acknowledge the usage of Flow Visualization Laboratory present
in the Department of Aerospace Engineering, SRM Institute of Science and Technology (SRMIST)
for the experiments performed towards the present research work.
1054 S. Sanjana et al.

References

1. Kinbara T, Endo H, Sega S (1967) Downward propagation of Smoldering combustion through

solid materials. In: Symposium (International) on Combustion 11(1):525–531
2. Ohlemiller TJ, Lucca DA (1983) An experimental comparison of forward and reverse smolder
propagation in permeable fuel beds. Combust Flame 54:131–147
3. SR Madara, Chithirai Pon Selvan M, JT Varghese (2017) Design and fabrication of low-cost
open circuit subsonic wind tunnel. Int J Innov Technol Res 5(3):6154–6161
4. Mukunda HS, Basani J, Shravan HM, Philip B (2007) Smoldering combustion of incense
sticks—experiments and modeling. J Combust Sci Technol 179(6):1113–1129
5. Bar-Ilan A, Putzeys OM, Rein G, Fernandez-Pello DL (2005) Urban, transition from forward
smoldering to flaming in small polyurethane foam samples. In: Proceedings of the combustion
institute 30:2295–2302
6. Gao Y, Chow WK (2005) Discussion on buoyancy, turbulence, and combustion in fire field
modelling. Int J Architect Sci 6(3):93–105
7. Schult DA, Matkowsky BJ, Volpert VA, Fernandez-Pello AC (1995) Propagation and extinction
of forced opposed flow smolder waves. Combust Flame 101:471–490
8. Stocker DP, Olson SL, Urban DL, Torero JL, Walther DC, Carlos Fernande-Pello A (1996)
Small-scale smoldering combustion experiments in microgravity.In: Twenty-Sixth Symposium
(International) on Combustion/The Combustion Institute, pp 1361–1368
9. Hadden RM, Rein G, Belcher CM (2013) Study of the competing chemical reactions in the
initiation and spread of smoldering combustion in peat. Proc Combust Inst 34:2547–2553
10. Sato K, Sega S Smolder (1989) Spread in a Horizontal Layer of Cellulosic Powder. International
Association for Fire Safety Science. Department of Physics, Faculty of science, Toho University
Spatial Location of Renewable Energy
Plants: How Good Is Good Enough?

Paul A. Adedeji, Stephen Akinlabi, Nkosinathi Madushele,

and Obafemi O. Olatunji

Abstract Land suitability analysis for renewable energy (RE) plants is gradually
becoming multidisciplinary. The integration of Geographical Information System
(GIS) with Multi-Criteria Decision Making (MCDM) tools has proven effective in
such studies. From the trend in the literature, both experts in GIS and non-experts use
these tools. There exist several factors that determine the suitability of a site for an
intended purpose. In this study, factors specific to the reliability and accuracy of the
results of GIS-MCDM approach to siting RE plants are reviewed. The reliability of
the results from this integrated tool is hinged on the accuracy of principal milestones
in the analysis. Four of these were identified in this article: data integrity, sensitivity
analysis, correct use of tools, criteria weight assignment. It was concluded that apart
from the identified dots in achieving high reliability and accuracy of the suitability
analysis, ground verification is highly essential as this helps to verify the reality of
the seemingly virtual analysis.

Keywords GIS · Land suitability analysis · Multi-criteria decision making ·

Renewable energy plant

1 Introduction

Spatial location of renewable energy (RE) plants largely depends on the availability
of sources, which is a geospatial variant and as such vary from one geographical
location to another. Asides climatic factors, RE facility location problem depends on
other criteria, like economic, technological, social, and physiological factors [1–3].

P. A. Adedeji (B) · N. Madushele · O. O. Olatunji

Department of Mechanical Engineering Science, University of Johannesburg, Johannesburg,
South Africa
e-mail: pauladedeji2k5@gmail.com
S. Akinlabi
Department of Mechanical and Industrial Engineering, University of Johannesburg, Johannesburg,
South Africa
Department of Mechanical Engineering, Covenant University Ota, Ota, Nigeria
© Springer Nature Singapore Pte Ltd. 2021 1055
S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_91
1056 P. A. Adedeji et al.

This results in a multi-criteria optimization problem. In recent times, siting RE plants

have been established as a geospatial problem, hence the relevance of Geographical
Information System (GIS) [4–7]. GIS technique fosters the analysis and synthesis
of spatial data. It helps to integrate spatially referenced data for problem-solving
[8]. The technique has found usefulness in several areas like crime area analysis,
facility management and planning, resource monitoring, mineral exploration, water
resource management, energy system facility location and planning, and so on [9–
13]. Rather than the use of analytical models that offer a closed-form of solution,
the GIS technique uses empirical modelling technique with georeferenced data such
that the location component of the system been analyzed is not played down.
The literature is replete with solutions to the “how” and the “what” dimensions
of RE resources, however, the “where” dimension (geospatial component) has lately
attracted attention. GIS is highly effective in locating RE resources with multi-criteria
considerations by means of what is called land suitability analysis. Land suitability
analysis considers several criteria represented as layers in GIS software. These layers
represent constraints to be satisfied, which varies from one energy resource to another.
Presented in Table 1 are some criteria considered in the literature when carrying out
land suitability analysis for specific RE resources.
While some criteria are common to almost all RE farms like distance from residen-
tial areas, the government reserved areas, road networks, some criteria are peculiar
to each RE resource like distance from communication networks peculiar to locating
wind farms [14] and gradient peculiar to solar farms [1, 29, 30], proximity to hot
springs peculiar to geothermal resources [23] and so on. This results from peculiar
interference with and dependence on identified physical features. The results of the
suitability analysis give the best location of RE facilities based on stated criteria.
However, the question of the degree of suitability of the geospatial solution domain
is the focus of this study.
In a view to overcoming climate change and its associated effects on the biosphere,
alternative sources of energy, which are clean and green are explored to replace the
conventional sources. RE resources, which are energy sources that can be replenished
within a short period [30] is fast gaining relevance both at national and global levels
and are deployed as an off-grid, mini-grid, or on-grid system [31]. In any of these,
proximity to energy sources or siting plants in energy hotspot among other factors
is highly essential for the plant’s efficiency, effectiveness, and sustainability. Two
common approaches have been explored in the literature in solving this problem: the
mathematical approach and the GIS-MCDM approach. This study, however, focuses
on the GIS-MCDM approach.
Several techniques have been developed over the years to determine land suitabil-
ity for RE resources. However, the most commonly used technique is the integration
of MCDM techniques with GIS [32]. This technique considers the spatial component
of criteria relative to identified feature raster files often in a projected coordinate sys-
tem. This integrated method has recorded success over time in determining suitable
locations for wind farms [1, 14, 33], solar farms [1, 29, 30, 34], biogas plants [26,
35, 36], geothermal plants [37], hydropower plant [38] and so on. This method is
associated with uncertainties, which can be due to the nature of the data used, human
Spatial Location of Renewable Energy Plants: How Good Is Good … 1057

Table 1 Land suitability criteria for RE resources

Resource Selection criteria References
Wind Energy Wind speed, Elevation, slope, [7, 14–17]
Distance from
1. Energy transmission lines
2. Lakes and rivers
3. Roads
4. Protected Areas
5. Airports and Railway stations
6. Mining sites
7. Fault Lines
8. Radio and TV stations
9. Urban areas
Solar Energy Solar irradiation potential, average temperature, [16, 18]
agrological capacity, land slope, land orientation, plot areas
Distance from:
1. Power lines
2. Substations
3. Main roads
4. Protected areas
Geothermal Energy Heat degree [19–23]
Distance from:
1. Volcano craters
2. Faults
3. Volcanic rocks
4. Hot springs
5. Alteration zones
6. Fumarole
Hydropower Head of hydropower, ground surface slope, conveyance [24]
length-head ratio,
Distance to:
1. Road network
2. Power transmission line
3. Dam reservoir
Biomass/biogas Biomass resource, slope, geomorphological factors, [25–28]
Distance from:
1. Industrial areas
2. Hydrological networks (e.g. rivers, streams, etc.)
3. Flora and fauna
4. Railways, airports, and heliports
5. Protected areas
6. Wetlands
7. Mining areas

judgement, weighting factors used, inherent variability in the model used [39, 40].
However, in reality, criteria for suitability analysis of a RE plant are inexhaustible.
Attempts are often made by modellers to cover critical factors, that largely influence
RE facility location.
1058 P. A. Adedeji et al.

MCDM-GIS approach to land suitability analysis is efficient, however, essential

factors that predicate the reliability of its results most especially, when land suitability
for RE in power generation is considered are discussed briefly. These are fundamental
pitfalls that compromise the accuracy of MCDM-GIS approach to the location of RE
plants. This study aims at carrying out a mini-review on four identified areas to take
into consideration by energy researchers to ensure the reliability of land suitability
analysis in energy research and RE plant location.

2 Reliability Factors for Geospatial Approach to RE Plant

Location

2.1 Data Integrity

Data integrity emphasizes data accuracy and consistency in quality and recency.
The recency of the geospatial data either raster or vector plays a significant role in
the reliability of the suitable sites obtained for RE plant. The advent of GIS as a
potent tool for spatial analysis has led to the proliferation of geographic databases
globally. Individual data creator also makes efforts in populating these databases
with digital maps of physical features and attributes; an example is raster files for
road networks, places, locations, etc. The accuracy of these data, the expertise and
integrity of the data creator is highly significant to the reliability of the study which
uses the data. Government organizations also populate databases for public use. They
often update these data with recent developments about the geospatial location of
features. The analysis which makes use of data from trusted database (e.g. national
database) however, without recency may likely falsify the results of land suitability
analysis of RE plants. For example, protected areas in the database of a year may
not be completely accurate for a study in the next year. Certain areas may have been
considered as non-protected areas due to political or anthropogenic activities. The
recency of the data reveals the present status of surfaces or features.
Data quality, on the other hand, depends on several factors among which is data
projection. Incorrect projection of the data presents the wrong spatial location of the
site. Geospatial data used for spatial analysis are obtained from metadata, which is
either in the form of raster or vector [41]. These metadata contain information that
describes the data like the projection (either spherical or projected coordinate system).
The quality of the input data into a model determines the quality and reliability of
its output [39].
Spatial Location of Renewable Energy Plants: How Good Is Good … 1059

2.2 Sensitivity Analysis

Aside tractability and feasibility, a good model is expected to be robust, which sensi-
tivity analysis constitutes one of the tools in achieving this. As part of good modelling
practice, it is expected that a modeller gives a level of confidence to which the model
efficiently represents the system [39]. Sensitivity analysis helps to account for inac-
curacy, ambiguity, and imprecision of GIS-multi-criteria evaluation process [42] and
also enhances understanding of land suitability results for informed decision making
[43]. It is a related practice to uncertainty analysis, and it is aimed at identifying
principal inputs significantly responsible for output variations [44]. Many empirical
facility location models follow this trend. However, few GIS integrated models, espe-
cially in RE plant consider this. Sensitivity analysis in spatial models differs from
that of analytical and other empirical models due to the nature of the data. There
are several methods, which have been developed over the years for performing sen-
sitivity analysis on spatial models. Some of these are extensively discussed in [39,
45, 46]. However, in RE farm site suitability problems, a pixel by pixel validation
to ensure satisfactory results is sometimes used. Here, a cadastral plot is selected at
random and exported as a single layer, where a similar process of adding factors to its
attribute table is performed [18]. The result is compared with the main analysis for
consistency. The simple weighted model aggregate scoring method of essential fac-
tors of influence is also commonly used in RE plant suitability analysis. This method
is efficient in RE plant land suitability analysis because sometimes, the suitability of
a site for RE farm may be due to an aggregate of trivial factors rather than crucial
and technical factors [47]. Sensitivity analysis in RE farm location also enables the
modeller to identify principal factors, which aggregates to the ranking score. This
analysis is highly essential before actual ground verification.

2.3 Choice of Appropriate Tool

GIS software like the ArcMap is multi-functional and so contains many tools, which
can be harnessed for geospatially related problems in many fields. The choice of
an appropriate tool for each analysis determines the validity and reliability of the
results. For example, in determining land suitability for a RE plant using GIS-MCDM
approach, several criteria are considered to ensure that the proposed site for the plant
does not interfere with physical or environmental features. To ensure this, exclusion
rules are developed with specific distances from physical features of significance
to the study [7, 14]. Exclusion criteria use inequalities, which require it been care-
fully specified using appropriate tools. A commonly used distance function is the
Euclidean distance. This calculates a straight-line distance from one point to another
as a crow would fly. It has found usefulness in the RE land suitability analysis in
the specification of distance from criteria like waterways, roads, seismic, volcanic
1060 P. A. Adedeji et al.

faults [2, 18, 48], etc. The Euclidean distance tool contains several fields like max-
imum distance. The maximum distance displays raster files with subclasses to the
maximum distance specified. Further exclusion criteria with distances greater than
the already specified distance do become irresponsive to such criteria. Incorrect use
of tools as such may lead to misleading results and wrong location for RE plant.

2.4 Subjectivity in Assigning Criteria Weights

Determining suitable land areas for RE plants requires a combination of critical,

subjective, and objective inputs/criteria [49]. Criteria weight assignment contributes
immensely to the controversy and uncertainty associated with MCDM-GIS approach
to land suitability analysis in RE farm [46]. The impact of this is significant when
several decision-makers are involved. Rather than one set of weights, a range of
weights is derived. A slight perturbation in the weights can significantly affect the
ranking of possible alternatives, which in some cases causes inaccuracy in the results
and undesirable consequences [50].
While objective criteria are easily represented in quantitative forms, subjective
criteria come in qualitative forms. This is often observed when maps are to be clas-
sified into fuzzy classes [23, 51]. On the other hand, critical inputs/criteria emanate
from the successful integration of objective and subjective inputs/criteria. They form
principal criteria in the model for further analysis. In spatial RE farm location prob-
lem, subjectivity in decision making is demonstrated when site choices are to be
made based on criteria of influence [52]. Commonly used MCDM techniques for
this include, analytic hierarchy process (AHP), analytic network process (ANP),
technique for order processing by similarity to ideal solution (TOPSIS), elimination
and choice expressing reality (ELECTRE), multi-criteria optimization and compro-
mise solution (VIKOR) and their integrated fuzzy hybrids, as decision support sys-
tem. These methods incorporate weight assignment techniques like rating, ranking,
trade-off, pairwise comparison [23]. In RE studies, rating, ranking, and pairwise
comparison techniques are commonly used for weight assignment and selection of
alternatives. This process is subjective and may likely result in a disparity in the
model result. Experts’ knowledge is leveraged, as there is no technical documenta-
tion for this. These methods are chosen due to their simplicity in computation, though
computation becomes complex for large criteria size.
One of the problems associated with subjectivity transformation is inconsistency
in the formulation of the criteria matrix on the part of the modeller. To avoid this
pitfall, Eigen method for criteria matrix is recommended [49], such that a square
matrix of criteria has an aggregate of its eigenvalues equal to its trace. The trace is
the sum of the diagonal elements of the criteria matrix.
Spatial Location of Renewable Energy Plants: How Good Is Good … 1061

3 Conclusions

It is observed in the recent decade that studies on suitability analysis for RE plants
using GIS-MCDM as a tool is gradually becoming multidisciplinary. This study,
therefore, focused on specific milestones in the analysis, which contributes to the
reliability and credibility of results obtained. The study discussed the relevance of
data integrity with a focus on obtaining recent data from credible sources. We also
established the importance of sensitivity analysis and the use of appropriate tools
in ensuring the reliability of the result. Effort is being made to reduce subjectivity
in transforming qualitative variables to quantitative equivalence based on experts’
knowledge. The study identified Eigen matrix method with singularity of the trace
elements as a method of ensuring consistency in subjective to the objective transfor-
mation process. Asides these identified dots in achieving high reliability and accuracy
of the results, ground verification is highly essential as this helps to verify the reality
of the seemingly virtual analysis.

References

1. Janke JR (2010) Multicriteria GIS modeling of wind and solar farms in Colorado. Renew
Energy 35:2228–2234. https://doi.org/10.1016/j.renene.2010.03.014
2. Villacreses G, Gaona G, Martínez-Gómez J, Jijón DJ (2017) Wind farms suitability loca-
tion using geographical information system (GIS), based on multi-criteria decision making
(MCDM) methods: the case of continental Ecuador. Renew Energy 109:275–286. https://doi.
org/10.1016/j.renene.2017.03.041
3. Palmer D, Gottschalg R, Betts T (2018) The future scope of large-scale solar in the UK :
site suitability and target analysis. Renew Energy 2018:1–11. https://doi.org/10.1016/j.renene.
2018.08.109
4. Bonham-Carter GF (1994) Geographical information systems for geoscientists: modeling with
GIS. Comput methods Geosci vol 13, New York: Pergamon
5. Noorollahi Y, Itoi R (2007) Geothermal power plant site selection with environmental
consideration in Namafjall area, North Iceland. Trans Geotherm Resour Counc 31:193–198
6. Moghaddam MK, Meentemeyer RK, Samadzadegan F, Sharifi MA, Itoi R (2013) Spatial data
analysis for exploration of regional scale geothermal resources. J Volcanol Geotherm Res 266
7. Noorollahi Y, Yousefi H, Mohammadi M (2016) Multi-criteria decision support system for
wind farm site selection using GIS. Sustain Energy Technol Assess 13:38–50. https://doi.org/
10.1016/j.seta.2015.11.007
8. Iyengar JV (1998) Application of geographical information systems. J Int Inf Manag 7
9. Case C, Hawthorne TL (2013) Served or unserved? A site suitability analysis of social services
in Atlanta, Georgia using geographic information systems. Appl Geogr 38:96–106. https://doi.
org/10.1016/j.apgeog.2012.11.012
10. Calvert K, Mabee W (2015) More solar farms or more bioenergy crops? Mapping and assessing
potential land-use conflicts among renewable energy technologies in eastern Ontario, Canada.
Appl Geogr 56:209–221. https://doi.org/10.1016/j.apgeog.2014.11.028
11. Castro-santos L, Prado G, Sim T (2019) Planning of the installation of offshore renewable
energies : a GIS approach of the Portuguese roadmap. 132:1251–1262. https://doi.org/10.1016/
j.renene.2018.09.031
1062 P. A. Adedeji et al.

12. Tien Bui D, Pradhan B, Lofman O, Revhaug I, Dick OB (2012) Landslide susceptibility map-
ping at Hoa Binh province (Vietnam) using an adaptive neuro-fuzzy inference system and GIS.
Comput Geosci 45:199–211. https://doi.org/10.1016/j.cageo.2011.10.031
13. Boulos MNK (2004) Towards evidence-based, GIS-driven national spatial health Kingdom vol
50, pp 1–50
14. Atici KB, Simsek AB, Ulucan A, Tosun MU (2015) A GIS-based multiple criteria decision
analysis approach for wind power plant site selection. Util Policy 37:86–96. https://doi.org/10.
1016/j.jup.2015.06.001
15. Al-Yahyai S, Charabi Y, Gastli A, Al-Badi A (2012) Wind farm land suitability indexing using
multi-criteria analysis. Renew Energy 44:80–87. https://doi.org/10.1016/j.renene.2012.01.004
16. Ali S, Taweekun J, Techato K, Waewsak J, Gyawali S (2019) GIS based site suitability assess-
ment for wind and solar farms in Songkhla, Thailand. Renew Energy 132:1360–1372. https://
doi.org/10.1016/j.renene.2018.09.035
17. Bili A, Vagiona DG (2018) Use of multicriteria analysis and GIS for selecting sites or onshore
wind farms: the case of Andros Island (Greece). Eur J Environ Sci 8:5–13
18. Sánchez-Lozano JM, Teruel-Solano J, Soto-Elvira PL, Socorro García-Cascales M (2013)
Geographical information systems (GIS) and multi-criteria decision making (MCDM) methods
for the evaluation of solar farms locations: case study in south–eastern Spain. Renew Sustain
Energy Rev 24:544–556. https://doi.org/10.1016/j.rser.2013.03.019
19. Kiavarz M, Jelokhani-Niaraki M (2017) Geothermal prospectivity mapping using GIS-based
ordered weighted averaging approach: a case study in Japan’s Akita and Iwate provinces.
Geothermics 70:295–304. https://doi.org/10.1016/j.geothermics.2017.06.015
20. Kiavarz M, Samadzadegan F, Noorollahi Y (2014) Geothermics Spatial analysis and multi-
criteria decision making for regional-scale geothermal favorability map. Geothermics 50:189–
201. https://doi.org/10.1016/j.geothermics.2013.09.004
21. Alsuhaibani Z, Hepbasli A (2013) Environmental effects future aspects of geothermal energy
in Saudi Arabia. Energy Sources Part A Recover Util Environ Eff 7036. https://doi.org/10.
1080/15567036.2010.497805
22. Watts MD (1987) Geothermal exploration of Roccamonfina volcano, Italy. Geothermics
16:517–528
23. Yalcin M, Kilic GF (2017) A GIS-based multi criteria decision analysis approach for exploring
geothermal resources: Akarcay basin (Afyonkarahisar). Geothermics 67:18–28. https://doi.org/
10.1016/j.geothermics.2017.01.002
24. Ahmadi H, Shamsai A, Storage P, Selection S (2009) Preliminary site selection of pumped
storage hydropower plants—a GIS-based approach. AUT J Model Simul 2009:25–32
25. Perpiña C, Martínez-Llario JC, Pérez-Navarro Á (2013) Multicriteria assessment in GIS envi-
ronments for siting biomass plants. Land Use Policy 31:326–335. https://doi.org/10.1016/j.
landusepol.2012.07.014
26. Sahoo K, Mani S, Das L, Bettinger P (2018) GIS-based assessment of sustainable crop residues
for optimal siting of biogas plants. Biomass Bioenerg 110:63–74. https://doi.org/10.1016/j.
biombioe.2018.01.006
27. Thompson E, Wang Q, Li M (2013) Anaerobic digester systems (ADS) for multiple dairy
farms: a GIS analysis for optimal site selection. Energy Policy 61:114–124. https://doi.org/10.
1016/j.enpol.2013.06.035
28. Brahma A, Saikia K, Hiloidhari M, Baruah DC (2016) GIS based planning of a biomethanation
power plant in Assam, India. Renew Sustain Energy Rev 62:596–608. https://doi.org/10.1016/
j.rser.2016.05.009
29. Uyan M (2013) GIS-based solar farms site selection using analytic hierarchy process (AHP)
in Karapinar region Konya/Turkey. Renew Sustain Energy Rev 28:11–17. https://doi.org/10.
1016/j.rser.2013.07.042
30. Asakereh A, Soleymani M, Sheikhdavoodi MJ (2017) A GIS-based Fuzzy-AHP method for
the evaluation of solar farms locations: case study in Khuzestan province, Iran. Sol Energy
155:342–353. https://doi.org/10.1016/j.solener.2017.05.075
Spatial Location of Renewable Energy Plants: How Good Is Good … 1063

31. Bhandari B, Lee KT, Lee CS, Song CK, Maskey RK, Ahn SH (2014) A novel off-grid hybrid
power system comprised of solar photovoltaic, wind, and hydro energy sources. Appl Energy
133:236–242. https://doi.org/10.1016/j.apenergy.2014.07.033
32. Cebi S, Ilbahar E, Atasoy A (2016) A fuzzy information axiom based method to determine the
optimal location for a biomass power plant: a case study in Aegean region of Turkey. Energy
116:894–907. https://doi.org/10.1016/j.energy.2016.10.024
33. Kabu IK (2016) Wind farm site selection assessment in the greater Accra, Volta and Eastern
Regions of Ghana : a GIS spatial multi-criteria assessment 2016
34. Chen CR, Huang CC, Tsuei HJ (2014) A hybrid MCDM model for improving GIS-based solar
farms site selection. Int J Photoenergy 2014. https://doi.org/10.1155/2014/925370
35. Franco C, Bojesen M, Hougaard JL, Nielsen K (2015) A fuzzy approach to a multiple criteria
and geographical information system for decision support on suitable locations for biogas
plants. Appl Energy 140:304–315. https://doi.org/10.1016/j.apenergy.2014.11.060
36. Höhn J, Lehtonen E, Rasi S, Rintala J (2014) A geographical information system (GIS) based
methodology for determination of potential biomasses and sites for biogas plants in southern
Finland. Appl Energy 113:1–10. https://doi.org/10.1016/j.apenergy.2013.07.005
37. Noorollahi Y, Itoi R, Fujii H, Tanaka T (2008) GIS integration model for geothermal exploration
and well siting. Geothermics 37:107–131. https://doi.org/10.1016/j.geothermics.2007.12.001
38. Setiawan D (2015) Potential sites screening for mini hydro power plant development in Kapuas
Hulu, West Kalimantan: a GIS approach. Energy Procedia 65:76–82. https://doi.org/10.1016/
j.egypro.2015.01.034
39. Crosetto M, Tarantola S, Saltelli A (2000) Sensitivity and uncertainty analysis in spatial
modelling based on GIS. Agric Ecosyst Environ 81:71–79. https://doi.org/10.1016/S0167-
8809(00)00169-9
40. Crosetto M, Tarantola S (2001) Uncertainty and sensitivity analysis: tools for GIS-based model
implementation. Int J Geogr Inf Sci 15:415–437. https://doi.org/10.1080/13658810110053125
41. Gigović L, Pamučar D, Božanić D, Ljubojević S (2017) Application of the GIS-DANP-MABAC
multi-criteria model for selecting the location of wind farms: a case study of Vojvodina. Serbia
Renew Energy 103:501–521. https://doi.org/10.1016/j.renene.2016.11.057
42. Malczewski J (2004) GIS-based land-use suitability analysis: a critical overview. Prog Plann
62:3–65. https://doi.org/10.1016/j.progress.2003.09.002
43. Xu E, Zhang H (2013) Spatially-explicit sensitivity analysis for land suitability evaluation.
Appl Geogr 45:1–9. https://doi.org/10.1016/j.apgeog.2013.08.005
44. Mavromatidis G, Orehounig K, Carmeliet J (2018) A review of uncertainty characterisation
approaches for the optimal design of distributed energy systems. Renew Sustain Energy Rev
88:258–277. https://doi.org/10.1016/j.rser.2018.02.021
45. Bonin O (2006) Sensitivity analysis and uncertainty analysis for vector geographical applica-
tions. 7th Int Symp Spat Accuracy Assess Nat Resour Environ Sci 319–328
46. Feizizadeh B, Blaschke T (2014) An uncertainty and sensitivity analysis approach for GIS-
based multicriteria landslide susceptibility mapping. Int J Geogr Inf Sci 28:610–638. https://
doi.org/10.1080/13658816.2013.869821
47. Cevallos-Sierra J, Ramos-Martin J (2018) Spatial assessment of the potential of renewable
energy: the case of ecuador. Renew Sustain Energy Rev 81:1154–1165. https://doi.org/10.
1016/j.rser.2017.08.015
48. Al Garni HZ, Awasthi A (2017) Solar PV power plant site selection using a GIS-AHP based
approach with application in Saudi Arabia. Appl Energy 206:1225–1240. https://doi.org/10.
1016/j.apenergy.2017.10.024
49. Sule D (2001) Logistics of facility location and allocation. vol 20010505. https://doi.org/10.
1201/9780203910405
50. Feizizadeh B, Blaschke T (2013) GIS-multicriteria decision analysis for landslide susceptibility
mapping: comparing three methods for the Urmia lake basin. Iran Nat Hazards 65:2105–2128
1064 P. A. Adedeji et al.

51. Miller A, Li R (2014) A geospatial approach for prioritizing wind farm development in North-
east Nebraska, USA. ISPRS Int J Geo-Inf 3:968–979. https://doi.org/10.3390/ijgi3030968
52. Kazak J, van Hoof J, Szewranski S (2017) Challenges in the wind turbines location process
in Central Europe—the use of spatial decision support systems. Renew Sustain Energy Rev
76:425–433. https://doi.org/10.1016/j.rser.2017.03.039
Analysis of Pendulum-Based Nonlinear
Energy Sink for Energy Harvesting

Pradeep V. Malaji

Abstract Passive method is one of the best-suited methods for vibration reduction
of the primary structure. Energy harvesting converts vibration energy into useful
electrical energy which can be used to power up sensors used for monitoring. Non-
linear energy sinks (NES) are analyzed to passively reduce vibration of the structure
as well as energy harvesting simultaneously in this article. Current work considers
the applicability of common pendulum as the NES for mitigation linear primary
structure and electromagnetic energy conversion. It is observed that the pendulum
NES can overcome one of the main limitations of conventional tuned mass damper
(TMD) designs, as it has the capability to mitigate primary structure excitation and
harvest electrical energy over a wide range. This is because the pendulum has the
capability to operate both in linear and nonlinear zones that can be utilized into a
resonance of the primary oscillator. For small input excitation, the pendulum acts as
a conventional tuned mass damper. However, for larger energy input, the pendulum
acts in the nonlinear zone as NES to reduce primary mass vibration. Thus, a pen-
dulum can dissipate energies in a relatively broad spectrum. This article presents a
numerical analysis of pendulum NES for control and energy harvesting.

Keywords Nonlinear energy sink · Pendulum · Energy harvesting · Passive control

1 Introduction

Protection of structures, machines and vehicles from excessive vibration is vital as

damage due to vibration can severely affect equipment’s service life. With emerge
of new sensing and control technologies can provide information about the health of
the structure. These tools can help to detect the initial damage in the system [1].

P. V. Malaji (B)
BLDEA’s V P Dr. P G Halakatti College of Engineering and Technology, Vijayapur, Karnataka
586103, India
e-mail: pradeepmalaji@gmail.com

© Springer Nature Singapore Pte Ltd. 2021 1065

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_92
1066 P. V. Malaji

Vibration isolation by passive methods has been widely studied by many

researchers. The main limitation of conventional linear isolators is that they are effec-
tive only if the natural frequencies of these isolators are nearer or below the excitation
frequency [2, 3]. This requires a low spring stiffness which makes the primary struc-
ture to undergo large static deflection. To address this issue, researchers analyzed
anti-resonant vibration which leads to the generation of anti-resonance frequencies
beyond the natural frequency [4]. This might have an advantage over conventional
isolator but still has a limitation on the isolation bandwidth with a single narrow
anti-resonance frequency.
To overcome the above limitations, nonlinearity has been used in designs in order
to improve the bandwidth of the vibration isolation system. Various nonlinear isola-
tors have been reviewed by Ibrahim [5] which indicates active research in this area.
Nonlinear damper and nonlinear stiffness are the two kinds of nonlinear elements
that have been studied [6, 7].
The nonlinear energy sink (NES) refers to an essentially nonlinear system con-
sisting of a small mass attached to a primary mass to attenuate the vibration under
various excitations [8]. NES provides the targeted energy transfer (TET) feature
which refers to its ability to irreversibly transfer energy from the primary system and
dissipating it within itself [9]. However, rather than dissipation, the vibration energy
can be converted into useful electric energy for minor energy needs. The process of
converting vibration energy into electric energy is called as energy harvesting.
Utilization of vibration energy for energy harvesting can be very useful for small
portable equipment or for equipment present in isolated regions [10, 11]. Conven-
tional energy harvesters have a limitation of harvesting energy at resonance [13].
Other options consisted of multimodal techniques [13, 15] or a switch to nonlin-
ear techniques [13, 14, 16]. While tuned mass dampers (TMD) and linear vibration
absorbers have been utilized for control of primary structure [17], limitations such as
negligible frequency robustness and poor performance for random vibration render
their utilization an infeasible task [17]. Attachment of auxiliary systems as energy
harvesters for vibration suppression has an advantage of vibration control and use-
ful energy harvesting [12]. Further investigations into nonlinear alternatives have
led to the development of nonlinear energy sink (NES). While it is successful in
suppressing primary systems’ responses, it can also perform simultaneous energy
harvesting. Such features allow vibration control in a broadband manner with simul-
taneous energy harvesting, making NES a lucrative and more advantageous system
when compared to its counterparts. Also, while it has yet not been fully investigated,
NES’s ability to increase operational bandwidth also allows it to effectively suppress
primary vibrations unlike the linear systems which falter when operated upon by
frequency varying excitation [17].
This paper analyzes a two-degree vibration isolation system, consisting of pri-
mary mass coupled with a pendulum NES with electromagnetic conversion system.
Amplitude of primary structure and voltage output from the NES are presented when
the primary system is under harmonic base excitation. Parametric study is conducted
to explore the effect of these on NES performance. Comparison of NES with TMD
is also presented to enumerate the advantages of NES.
Analysis of Pendulum-Based Nonlinear Energy Sink for Energy … 1067

This paper has been divided into four sections. Section 2 describes the NES
harvester model with mathematical equations used for the simulations. Section 3
deals with the results and discussion followed by summarizing and providing the
relevant and observed conclusions in Sect. 4.

2 System Model

The system model is shown in Fig. 1. The primary system has mass M, damping c1
and linear stiffness k. The pendulum is mounted on the primary mass with provision
for electromagnetic conversion due to pendulum oscillation when primary system is
subjected to support excitation x g . It has mass m, mechanical and electrical damping
c2m and c2e, respectively, and length l.
Lagrangian of the system in terms of potential and kinetic energy is written as,

  1
L = (M + m) ẋ g2 + ẋ 2 + M ẋ g ẋ − kx 2 − mgl(1 − cos θ )
2
1 2 2 
+ m l θ̇ + 2 ẋ g ẋ + 2 ẋl θ̇ cos θ + 2 ẋ g l θ̇ cos θ (1)
2
Damping as through Rayleigh’s dissipation function is taken as,

1 
D= C1 x 2 + (C2m + C2e )θ 2 (2)
2
From the Lagrangian equation of motion with the voltage generated is written as:
where B is magnetic flux density, L is coil inductance and d is a gap between coil
and magnet.

(M + m)ẍ + ml cos θ θ̈ − ml 2 sin θ θ̇ 2 + c1 ẋ + K x = −(M + m)ẍ g

Fig. 1 System model with

pendulum NES harvester
1068 P. V. Malaji

ml 2 θ̈ + ml ẍ cos θ + (c2m + c2e )l 2 θ̇ + mgl sin θ = ml ẍ g cos θ (3)

v = B Ld θ̇

An equation of motion in non-dimensional form is written as,

ü + ε cos θ θ̈ − ε sin θ θ̇ 2 + ζ1 u̇ + u = f Ω 2 cos(Ωτ )

θ̈ + cos θ ü + (ζ2m + ζ2e )θ̇ + r 2 sin θ = f Ω 2 cos(Ωτ ) cos θ (4)

v = θ̇

The non-dimensional parameters are,

 
k g
τ= ω12 t, x= ul, ω1 = , ω2 =
M +m l
m ω
xg = X g cos(ωt), ε = ,Ω =
(M + m) ω1
c1 c2m c2e
ς1 = , ς2m = , ς2e =
mω1 mω1 mω1
Xg ω2
f = ,r = (5)
l ω1

Equation 4 will be solved numerically using ODE45 in MATLAB to obtain results

which will be presented in the next section.

3 Results and Discussion

In order to study the effect of the pendulum, NES on the primary system and voltage
output from NES parametric study on the system are considered. It is to be noted
that the parameters of primary systems are not changed. Therefore, the influence of
frequency ratio r, electrical damping ξ2e , mass ratio ε and excitation level f is con-
sidered. Following parameters shown in Table 1 are used unless otherwise changes
are mentioned.
Figure 2 shows the response of the primary structure without any absorber. The
structure exhibits peak amplitude at resonance as expected. To safeguard the structure
from failure at resonance, it is necessary to minimize the response of the primary
structure at resonance. This can be done by adding secondary structure as shown in
Fig. 1.
The response of the primary structure and voltage generated from NES is shown
in Fig. 3. With the introduction of NES, the amplitude of the primary system at
resonance has been reduced drastically. Ratio of primary and pendulum structure
Analysis of Pendulum-Based Nonlinear Energy Sink for Energy … 1069

Table 1 Parameters used for

Parameter Value
simulation
Mass ratio ε 0.2
Ratio of frequencies of secondary and primary system 0.9
r
Damping factor of primary system ξ1 0.025
Damping factor of secondary system ξ2m 0.01
Electrical damping factor of secondary system ξ2e 0.05
Ratio of excitation amplitude to length of pendulum f 0.03

1.5

1
u

0.5

0
0.5 1 1.5
Ω

Fig. 2 Primary system response without absorber

1 2
r=0.9 F R
r=1 F R
r=1.1 F R

1
v

0.5
u

r=0.9 F R
r=1 F R
r=1.1 F R
0 0
0.5 1 1.5 0.5 1 1.5
Ω Ω
(a) (b)

Fig. 3 Effect of frequency ratio of primary system and pendulum NES a on primary system,
b voltage generated from NES. (Continuous line-F-forward sweep, dotted line-R-backward sweep)
1070 P. V. Malaji

frequency r is varied to study its effect. With an increase in r, the response curve
of primary and NES structure shifts toward right. For r < 1, the amplitude u of the
primary system is almost zero at resonance when r > 1 the amplitude of primary
structure increases as shown in Fig. 3a. Multi-frequency voltage can be generated
from NES indicating broadband harvesting. Spring softening effect can be observed
at the second frequency which further increases frequency band as in Fig. 3b.
Effect of electrical damping on system performance is shown in Fig. 4. The elec-
trical damping is varied between 0.005 and 0.02. With the increase in electrical damp-
ing, the amplitude of primary structure away from resonance decreases indicating
vibration reduction. This also will affect the bandwidth of the voltage generated by
NES. Hence, an optimal value of electrical damping has to be selected to optimize
the amplitude of the primary structure and voltage generated.

1 1.5
2e =0.005 F R
2e =0.01 F R
2e =0.02 F R 1
v
u

0.5
2e =0.005 F R
0.5 2e =0.01 F R
2e =0.02 F R

0 0
0.5 1 1.5 0.5 1 1.5
Ω Ω
(a) (b)
Fig. 4 Effect of electrical damping ξ2e a on primary system, b voltage generated from NES

4
2 f=0.01 F R
f=0.01 F R
f=0.03 F R
f=0.03 F R
f=0.05 F R
f=0.05 F R

2
u

0 0
0.5 1 1.5 0.5 1 1.5
Ω Ω
(a) (b)
Fig. 5 Effect of excitation amplitude a on primary system, b voltage generated from NES, ξ2e =
0.01
Analysis of Pendulum-Based Nonlinear Energy Sink for Energy … 1071

Figure 5 shows the effect of excitation amplitude f on the system which is varied
from 0.01 to 0.05. With the increase in excitation level, amplitudes of primary struc-
ture away from resonance increase as expected whereas the amplitude at resonance
remains the same as shown in Fig. 5a. Voltage amplitude and bandwidth from NES
increase with the increase in excitation level as shown in Fig. 5b. Spring softening
effect at the first frequency can also be observed at higher excitation.
Mass of the pendulum m has to be much smaller than the mass of primary system
M. The effect of mass ratio on the system performance is shown in Fig. 6. Increases
in mass ratio shift the response curve toward left. For the low value of mass ratio ε =
0.1, the amplitude of primary structure at resonance will be more compared to when
ε = 0.2 as shown in Fig. 6a. Voltage generated by NES at a low mass ratio will have
higher amplitude at both frequencies and peaks are nearer as shown in Fig. 6b.
To highlight the benefit of the NES with compared to the classic TMD, even at r =
1, a comparison is shown in Fig. 7. For the sake of comparison, all other parameters
are kept the same for both NES and TMD. The amplitudes with both NES and TMD

1 3
ε =0.1 F R ε=0.1 F R
ε =0.2 F R ε=0.2 F R

2
0.5
v
u

0 0
0.5 1 1.5 0.5 1 1.5
Ω Ω
(a) (b)

Fig. 6 Effect of mass ratio a on primary system, b voltage generated from NES. ξ2e = 0.01

1 NES Forward 2
NES Backward
T MD
v

0.5 1
u

NES Forward
NES Backward
T MD
0 0
0.5 1 1.5 0.5 1 1.5
Ω Ω
Fig. 7 Comparison of TMD and NES at r = 1 a primary system, b voltage generated from NES.
ξ2e = 0.01, r = 1
1072 P. V. Malaji

at resonance remain the same for the primary structure, whereas TMD produces
comparatively low amplitudes away from the resonance. The voltage generated by
NES has higher amplitude and better bandwidth compared to TMD.

4 Conclusion

From the analysis carried out in the manuscript, it can be mainly concluded that
regular damped pendulum with electromagnetic conversion provision can be used
as a nonlinear energy sink to mitigate vibration of primary structure and harvest
useful energy from NES. Also, NES exhibits better performance at a broader range.
Parametric study is carried out to understand the effect of parameters on system
performance. Further, a theoretical and experimental study with an optimization
study will be considered in the future.

References

1. Ali SF, Ramaswamy A (2009) Optimal dynamic inversion based semi-active control of
benchmark bridge using MR dampers. Struct control Health Monitor 16:564–585
2. DeBra DB (1992) Vibration isolation of precision machine tools and instruments. CIRP Ann
Manuf Technol 41:711–718
3. Rivin EI (2003) Dynamic properties of vibration isolation systems. In: Passive vibration
isolation, Chapter 1, ASME press, New York, USA
4. Yilmaz C, Kikuchi N (2006) Analysis and design of passive band-stop filter-type vibration
isolators for low frequency applications. J Sound Vibration 291:1004–1028
5. Ibrahim RA (2008) Recent advances in nonlinear passive vibration isolators. J Sound Vib
314:371–452
6. Peng Z, Lang Z, Zhao L, Billings SA, Tomlinson GR, Guo P (2011) The force transmissibility of
MDOF structures with a non-linear viscous damping device. Int J Non-Linear Mech 46:1305–
1314
7. Ahn HJ (2008) Performance limit of a passive vertical isolator using a negative stiffness
mechanism. J Mech Sci Technol 22:2357–2365
8. Vakakis AF (2001) Inducing passive nonlinear energy sinks in vibrating systems. J Vib Acous
123:324–332
9. Kopidakis G, Aubry S, Tsironis GP (2001) Targeted energy transfer through discrete breathers
in nonlinear systems. Phys Rev Lett 87:165501
10. Ertuk A, Inman DJ (2011) Broadband piezoelectric power generation on high-energy orbits of
the bistable duffing oscillator with electromechanical coupling. J Sound Vib 330:2339–2353
11. Liuyang X, Tang L, Liu K, Mace BR (2018) On the use of piezoelectric nonlinear energy
sink for vibration isolation and energy harvesting. In: Conference on smart materials, adaptive
structures and intelligent systems, pp 1–6
12. Malaji PV, Rajarathinam M, Jaiswal V, Ali SF, Howard IM (2019) Energy harvesting from
dynamic vibration pendulum absorber. Rec Adv Struct Eng 2:467–478 (Springer Singapore)
13. Malaji PV, Ali SF (2018) Analysis and experiment of magneto-mechanically coupled
harvesters. Mech Syst Signal Process 108:304–316
14. Malaji PV, Ali SF (2017) Magneto-mechanically coupled multiple energy harvesters. In:
1st International Conference on Power Electronics, Intelligent Control and Energy Systems
(ICPEICES 2016), IEEE, pp 1–5
Analysis of Pendulum-Based Nonlinear Energy Sink for Energy … 1073

15. Rajarathinam M, Ali S (2018) Energy generation in a hybrid harvester under harmonic
excitation. Energy Convers Manage 155:10–19
16. Kumar KA, Ali SF, Arockiarajan A (2017) Magneto-elastic oscillator: modeling and analysis
with nonlinear magnetic interaction. J Sound Vib 393:265–284
17. Xiong L, Tang L, Liu K, Mace BR (2018) Broadband piezoelectric vibration energy harvesting
using a nonlinear energy sink. J Phys D Appl Phys 51:1–2
Effect of Input Velocity on the Output
of Vertical Axis Wind Turbine (VAWT)

Sankgeeth Vennila Sigamani, Shami Jose Jose Robin,

Arun Prakash Chandran, and B. Anand Ronald

Abstract Vertical axis wind turbines (VAWT) are finding applications in the
advanced scientific fields and seem to be a promising and rapidly growing inno-
vation. A rarely researched field regarding VAWT is on analyzing the effects of
various input velocities for bidirectional wind flow on the vertical axis wind turbines
due to the complex and unsteady wind flow characteristics. In the current study, three-
dimensional simulations are carried out, and the static and dynamic results are then
reported for different values of rotational speed. The outcomes like output velocities
and pressure are analyzed for a range of input velocities in a Savonius wind turbine.
Different input velocities in the range of 10 m/s, 15 m/s, 20 m/s and 25 m/s were
studied. The model for the rotor was designed on Solidworks 2018, and numerical
analysis was carried out for a bidirectional wind input on commercial CFD software,
ANSYS Fluent v16.0. From the analysis, the output velocity and pressure graphs
were plotted with respect to the input air velocity and compared.

Keywords Renewable energy · VAWT · Velocity · Wind turbine

1 Introduction

Over the past few years, development and modernization in the form of natural
alterations and over-exploitation of available resources have led to global concern
over pollution, global warming and possible depletion of non-renewable sources of
energy, such as oil. This has led us to the exploration of renewable and sustainable
energy resources. Conventionally, wind energy is harnessed on a large scale using
horizontal axis wind turbines (HAWTs) spread over a large area. But, HAWT has
several disadvantages like the requirement of large wind speeds, large area, high ini-
tial cost and more maintenance. Vertical axis wind turbine (VAWT), unlike a HAWT,
is one whose axis of rotation is vertical with respect to ground. This arrangement

S. Vennila Sigamani · S. J. Jose Robin · A. P. Chandran (B) · B. Anand Ronald

Department of Mechanical Engineering, Sri Sivasubramaniya Nadar College of Engineering,
Kalavakkam, Chennai, Tamilnadu 603110, India
e-mail: arunprakashc@ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 1075

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_93
1076 S. Vennila Sigamani et al.

allows the generator and gearbox to be located close to the ground facilitating easy
maintenance and repair. They have gained more interest for their excellent adapt-
ability to the urban environment in terms of visual impact and noise pollution. Being
axis symmetric, they are omnidirectional turbines, which respond well to changes
in wind direction. A typical wind turbine, which utilizes natural wind power, faces
problems regarding lack or excess of wind speed. But in this case, a considerable
amount of wind energy is produced due to the pressure difference created by the
moving vehicles on the highways. We need a system to recycle otherwise wasted
energy of highway traffic. Moreover in remote locations extending power supply
would itself be a tiring task which can be solved by having this type of turbine which
could generate power enough to illuminate the road cautions or indicators needed
for the road users. By setting VAWTs in highways, artificial wind can also be used
to produce electricity. The other major advantage of this system is that it is depen-
dent only on the number of vehicles on the road and not on other resources such as
sunshine or wind, which are seasonal or location specific. Out of the two types of
VAWT, called Darrieus VAWT and Savonius VAWT, the latter is known as quite a
wind power source because of its lowest tip speed ratio and compact size. Savonius
rotor was firstly developed and introduced by Finnish inventor S.J. Savonius. Though
the Savonius rotor has low aerodynamic efficiency, it is preferably used because it
runs at wind speeds as low as 2 m/s.

2 Literature Review

Although the concept of vertical axis wind turbines (VAWTs) was proposed by
Darrieus as early as 1931, the research and development in this area are still of interest
and in progress nowadays. In order to determine the performance of the VAWT,
analytical and numerical approaches, as well as experimental tests, are generally used
among others [1]. The Sutherland report on the stream tube model and vortex models
allows one to analyze the aerodynamic response of VAWTs [2]. The first simplest
approach developed by Templin is the single stream tube numerical model. Templin
calculates the performance for a single blade whose chord equals the sum of the
chords of the actual rotor’s blade through the application of the blade element theory
[3]. Strickland extends Templin’s approach into a multiple single stream tube model
by considering a number of adjacent smaller steam tubes, applying the conservation
of momentum and blade element theory to each steam tube. Further improvement
in the multiple steam tube model has been reported by Paraschivoiu, namely the
double-multiple steamtube model (DSMT) [4]. A second class of models is those
based on vortex representation of the blades and their wakes. Strickland et al.’s
model is based on the replacement of the rotor blade by a bound vortex filament,
called a lifting line, whose strength is a function of the azimuthal position [5]. Wang
presents a potential flow 2D vortex panel model. Experimental tests can be performed
by means of wind tunnel or open-field measurement in order to evaluate the real
performance of wind turbines [6]. Bedon describes the results of an experimental
Effect of Input Velocity on the Output of Vertical Axis Wind … 1077

campaign of measurements conducted on a VAWT with a helical blade having a radius

of 1 m with a NACA-0018 airfoil section, tested in an open field. With the advent of
computers, many authors have investigated the feasibility and accuracy of different
computational fluid dynamics (CFD) approaches [7]. Wolfe uses CFD to investigate
the two-dimensional aerodynamic performance of typical wind turbine airfoils with
the standard turbulence model. Carrigan et al., fixing the tip speed ratio (TSR) of
the wind turbine, developed an iterative design system to maximize the torque for
different airfoil cross sections and solidities in two-dimensional CFD simulations. A
CFD performance model based on blade element momentum theory was presented
by Raciti Castelli et al. for evaluating the aerodynamic forces of straight blade VAWT
[8]. Most of the authors have the adopted Reynolds averaged Navier–Stokes equations
(RANS) approach with different turbulence models. However, Iida et al. used a model
with large-eddy simulations (LES), proving that the as-obtained results are in good
agreement at high TSR with that of momentum theory. Only recently have some
researchers carried out three-dimensional CFD simulations. Howell et al. provided
a three-dimensional CFD model based on the Re-Normalization Group (RNG) κ-
turbulence model, and the effects of turbine solidity, roughness and tip vortices were
considered.

3 Experimental Setup

Savonius VAWT is a drag-type device, consisting of two scoops. Looking down on

the rotor from above, a two-scoop machine would look like an “S” shape in cross
section. Because of the curvature, the scoops experience less drag when moving
against the wind than when moving with the wind. The differential drag causes the
Savonius turbine to spin. Effect of various parameters on the performance of Savonius
rotor was studied in the previous literature such as aspect ratio, use of end plates, use
of deflecting plates, overlap ratio and number of blades. The current experimental
model consists of a mild steel rod of length 700 mm and a diameter of 20 mm
supporting the blades. The blades are made of aluminum with an “S-”shaped profile
and thickness of 2 mm. Each blade is curved approximately to a diameter of 260 mm
and a length of 335 mm. The entire setup is supported on a mild steel base. A model
was made in Solidworks 2018 using the appropriate dimensions. The Navier stokes
equation was used as the governing equation and a 6 Degree of Freedom (6DOF)
solver was used. Subsequently, the analysis was carried out in ANSYS Fluent v16.0.

4 3D Modeling

Using Solidworks 2018, the 3D part is generated in the fabricated dimensions by

comprehensively considering structural features, processing requirements and solu-
tion calculation amount. For the purpose of reducing the number of elements and
1078 S. Vennila Sigamani et al.

Fig. 1 Shaft rod profile

workload, when the finite model is established various reasonable simplifications

were carried on, various petty features like fillet and chamfering were given up. An
optimum aspect ratio (Ç 0.5) and power coefficient [2, 5] were considered.

4.1 Create the Shaft Rod Profile

The shaft rod has been modeled using Solidworks 2018 software and is shown in
Fig. 1.

4.2 Generate the Blade Design (Fig. 2)

The CAD model of the blade design is shown in Fig. 2.

5 CFD Model

The sequence of operations involved in the modeling and analysis is shown in the
flowchart (Fig. 3 and 4).
The profile of the rotor is symmetric and is extruded off the 2D geometry. The
analysis was done in ANSYS Fluent v16.0. The finite element method (FEM) is used
to perform the analysis. The finite element method is a powerful tool to obtain the
numerical solution of a wide range of engineering problem. The method is general
Effect of Input Velocity on the Output of Vertical Axis Wind … 1079

Fig. 2 Design of the blade

Fig. 3 Steps involved in analysis of various blade velocities for VAWT

enough to handle any complex shape or geometry, for any material under differ-
ent boundary and loading conditions. The generality of the finite element method
fits the analysis requirement in the current application. It is a method of piecewise
approximation in which the structure or body is divided into small elements of finite
dimensions called finite elements, and then, the original body or the structure is
considered as an assemblage of these elements connected at the finite number of
1080 S. Vennila Sigamani et al.

Fig. 4 Importing the file geometry to the design module

joints called nodal points or nodes. The constraints of the body are given through a C
program. The enclosure has inlets and outlets diagonally on opposite sides of a box
enclosure. The adjacent inlet and outlet are separated by a small region of nil effect
considering practical conditions of bidirectional fluid flow. Re-meshing (as shown in
Fig. 5) is carried out for each time step and smoothing is defined. The time steps are
taken as 0.0025 s considering the accuracy and calculation time. A dynamic mesh is
defined for the fluid domain. The procedures are repeated for each input velocities
(Fig. 6 and 7).

Fig. 5 Meshed view

Effect of Input Velocity on the Output of Vertical Axis Wind … 1081

Fig. 6 Setting up of boundary conditions

Fig. 7 Importing in CFD Fluent

6 Results

The results are calculated and plotted (Fig. 8).

1082 S. Vennila Sigamani et al.

Fig. 8 Input velocity versus output velocity and pressure

For the Savonius wind turbine with a blade thickness of about 2 mm and by fixing
the temperature, the input velocity is changed. The corresponding output velocity and
pressure are noted. As seen from the graph, we increased the input velocity, which in
turn reduced the pressure (on the blades) drastically and increased the output velocity
linearly. We found that, at 15 m/s input velocity, the pressure and the output velocity
are near to each other which make the turbine mechanically efficient. For a particular
input velocity, both low and high pressure makes the blade unstable. And thus at
15 m/s, the turbine is efficient.

7 Conclusion

In this project, we investigated different velocities which exert on the vertical axis
wind turbine. Here, the analysis is taken for the different velocity values such as
10, 15, 20, 25 m/s. The Savonius wind turbine is designed in Solidworks and which
is imported in ANSYS Fluent launcher for the flow analysis, and the variations in
the pressure and velocity values are obtained for each condition. The wind speed is
taken into account, and it is given as angular velocity to the fan. An analysis was
performed on ANSYS Fluent simulating the process, and it was concluded that, in
case of complex bidirectional wind flow, a Savonius rotor with a curved blade of
Effect of Input Velocity on the Output of Vertical Axis Wind … 1083

2 mm thickness is efficient at 15 m/s of input velocity with maximum mechanical

efficiency and minimum pressure difference, thus attaining stable speed faster and
has a higher rate of acceleration.

References

1. Darrieus G (1931) Turbine having its rotating shaft transverse to the flow of the current. U.S.
Patent1, 835, 018, 8 Dec 1931
2. Sutherland HJ, Berg DE, Ashwill TD (2012) A retrospective of VAWT technology; Sandia
Report, SAND 2012-0304; Sandia National Laboratories: Albuquerque, NM, USA, Jan 2012
3. Islam M, Ting DSK, Fartaj A (2008) Aerodynamic models for Darrieus-type straight-bladed
vertical axis wind turbines. Renew Sustain Energy Rev 12:1087–1109
4. Templin RJ (1974) Aerodynamic performance theory for the NRC Vertical-Axis Wind Turbine;
Technical Report N-76-16618; LTR-LA-160; National Aeronautical Establishment: Ottawa,
ON, Canada, 1 June 1974
5. Paraschivoiu I (1988) Double-multiple stream tube model for studying vertical-axis wind
turbines. J Propuls Power 4:370–377
6. Brusca S, Lanzafame R, Messina M (2014) Design of a vertical-axis wind turbine: how the
aspect ratio affects the turbine’s performance. Int J Energy Environ Eng 5:333–340
7. Strickland J, Webster B, Nguyen T (1979) A vortex model of the Darrieus turbine: an analytical
and experimental study. J Fluids Eng 101:500–505
8. Wang L, Zhang L, Zeng N (2007) A potential flow 2-D vortex panel model: applications to
vertical axis straight blade tidal turbine. Energy Convers Manag 48:454–461
Radiative Heat Transfer of Magnetic
Nanofluid Flow Past a Porous Inclined
Plate: A Mathematical Model

M. Shanmugapriya

Abstract This paper numerically investigates radiative magnetohydrodynamic

boundary layer flow of nanofluids over a porous inclined plate in the presence of
viscous dissipation. The governing partial differential equations are converted into
nonlinear differential equations by using a suitable similarity transformation, which
are solved numerically using Runge–Kutta–Fehlberg fourth-fifth-order method along
with shooting technique. Two types of nanoparticles are discussed, namely copper
(Cu) and alumina (Al2 O3 ) in the base fluid of water. Maxwell Garnetts and Brinkman
models are used for the effective thermal conductivity and dynamic viscosity of the
nanofluids, respectively. The velocity and temperature distributions are obtained for
various governing physical parameters and are presented graphically. In addition,
the evolution of skin friction coefficient and Nusselt number values with selected
parameters is presented. Comparison is made with previously published work and
found to be in good agreement.

Keywords Nanofluid · Thermal radiation · Magnetohydrodynamic flow · Heat

transfer · Inclined plate

1 Introduction

A nanofluid are fluids that contain small volumetric quantities of nanometer-

sized (diameter of about 1–100 nm) particles called nanoparticles. These flu-
ids are engineered colloidal suspensions of nanoparticles in a base fluid like oil,
water and ethylene glycol. Nanoparticles are made from various materials, such as
Cu, Ag, Au, Fe, H g, T i metals and non-metallic, Al2 O3 , Cu O, T i O2 , Si O2 , etc.
The convectional heat transfer base fluids have poor heat transfer abilities due to
their low thermal conductivity. To upgrade the thermal conductivity of these fluids,
a lot of theoretical and experiential analysis has been made. Choi [1] first used the

M. Shanmugapriya (B)
Department of Mathematics, Sri Sivasubramaniya Nadar College of Engineering, Chennai
603110, India
e-mail: shanmugapriyam@ssn.edu.in

© Springer Nature Singapore Pte Ltd. 2021 1085

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_94
1086 M. Shanmugapriya

word nanoparticle to improve the thermal conductivity of base fluids. Eastman et al.
[2] have studied that the thermal conductivity of the base fluid (water) was increased
up to 60% when CuO nanoparticles of volume fraction 5% were added to the water.
Furthermore, Eastman et al. [3] have also noticed that the effective thermal conduc-
tivity has shown to be increased 40% when copper nanoparticles of volume fraction
less than 1% are added to the ethylene glycol or oil. Xuan et al. [4] reported that with
a small volumetric fraction of nanoparticles less than 5% the thermal conductivity of
the base fluid is enhanced by 10–20%. Xie et al. [5] have studied that Al2 O3 -ethylene
glycol-based nanofluid thermal conductivity is increased in the range 25–30% when
alumina nanoparticles are added. Therefore, these nanofluids are capable of promot-
ing energy efficiency, improving thermal conductivity and thermal properties of heat
transport of base fluid. Nanofluids have therefore found significant applications in
major process industries, including material and chemicals, food and drink, oil and
gas, paper and printing and textiles.
The word magnetohydrodynamics has several industrial applications such as
MHD generators, pumps, petroleum technologies and liquid metal cooling blan-
ket for fusion reactor. The MHD flow in electrically conducting fluid can control
the rate of heat transfer at the surface, and hence, desired cooling effect can be
achieved. Magneto-nanofluid has specific applications in biomedicine, optical mod-
ulators, magnetic cell separation, magneto-optical wavelength filters, silk flood sepa-
ration, nonlinear optical materials, drug delivery, etc. A magnetic nanofluid has both
liquid and magnetic properties. Hamad et al. [6] introduced a one-parameter group
to represent similarity reduction for the problem of MHD-free convection flow of a
nanofluid past a vertical semi-infinite flat plate. Sudarsana Reddy et al. [7, 8] have
discussed the effect of chemical reaction on MHD natural convection boundary layer
flow, heat and mass transfer characteristics of nanofluids over a rotating disk through
a porous medium and vertical cone, respectively.
The radiative effects have important applications in physics and engineering. The
radiative heat transfer effects on different flows are very important in space tech-
nology and high-temperature processes, and very little is known about the effects of
radiation on the boundary layer of a radiative MHD fluid past a body. Shanmugapriya
[9] has studied the effect of thermal radiation Cu-water nanofluid past a moving
wedge using Gyarmati’s variational principle. Shanmugapriya et al. [10] investigated
the effect of entropy generation and thermal radiation of Cu-water nanofluid over a
moving wedge using the shooting technique together with Runge–Kutta fourth-order
integration scheme. Thiripathi Thumma et al. [11] have presented heat source/sink
effects in addition to viscous dissipation for two different magnetic nanofluid flows
from a nonlinear inclined stretching/shrinking sheet. Sudarsana Reddy et al. [12]
have discussed the influence of thermal radiation and chemical reaction on MHD
boundary layer flow over a nanofluid saturated inclined porous plate under boundary
conditions. Sudarsana Reddy et al. [13] have reported MHD boundary layer flow, heat
and mass transfer analysis of nanofluid over an inclined vertical plate saturated by
the porous medium with thermal radiation and heat generation/absorption by using
versatile, extensively validated, variational finite element method.
Radiative Heat Transfer of Magnetic Nanofluid Flow Past … 1087

The main objective of this work is to address the impact of thermal radiation and
viscous dissipation on MHD boundary layer flow and heat transfer over a porous
inclined plate saturated by Cu-water and Al2 O3 -water-based nanofluid. The effects
of the magnetic parameter, mixed convection parameter, radiation parameter, porous
medium parameter, Eckert number, suction parameter, Prandtl number and angle of
inclination on velocity and temperature fields are presented via figures and tables.

2 Mathematical Formulation of the Model

Consider the steady two-dimensional laminar MHD flow and heat transfer of an
incompressible nanofluid past a porous inclined plate with a linear stretching veloc-
ity u w (x) = ax where a(> 0) being the stretching constant. The stream function ψ
can be defined as u = ∂ψ
∂y
and v = − ∂ψ∂x
, and similarly, the temperature at the surface
and free stream are denoted as Tw and T∞ , respectively. The viscous dissipation and
radiation effects are considered in the energy equation. The Rosseland approxima-
tion is used to describe the radiative heat flux in the energy equation. The uniform
magnetic field B0 is applied to the flow direction. The basic governing equations for
such type of flow can be written as (Tiwari and Das [14]).

∂ 2ψ ∂ 2ψ
− =0 (1)
∂ y∂ x ∂ x∂ y
 
∂ψ ∂ 2 ψ ∂ψ ∂ 2 ψ ∂ 3ψ g(ρβ)n f (T − T∞ ) cos α
− = γ n f +
∂ y ∂ x∂ y ∂ x ∂ y2 ∂ y3 ρn f
2   
σ B0 ∂ψ γn f ∂ψ
− − (2)
ρn f ∂ y K ∂y
   2 2
∂ψ ∂ T ∂ψ ∂ T ∂2T μn f ∂ ψ
− = αn f 2 +  
∂y ∂x ∂x ∂y ∂y ρc p n f ∂ y 2
1 ∂qrad
−  (3)
ρc p n f ∂ y

subject to the boundary conditions

∂ψ ∂ψ
= Uw (x) = ax, = Vw , T = Tw at y = 0 (4)
∂y ∂x
∂ψ
→ 0, T → T∞ as y → ∞ (5)
∂y

Here, u, v stand for components of velocity in x, y directions, respectively, T is

the temperature of the nanofluid in the boundary layers, μn f is the viscosity of the
1088 M. Shanmugapriya

nanofluid, ρn f  is the
 density of the nanofluid, αn f is the thermal diffusivity of the
μ
nanofluid and ρc p n f is the heat capacitance of the nanofluid, γn f = ρnnff symbolizes
the kinematic diffusivity of the nanofluid, βn f is the thermal expansion coefficient of
the nanofluid, α is the angle of inclination, B0 is the strength of the magnetic field and
K is the constant permeability of the porous medium. By the use of the Rosseland
estimation for radiation in Eq. (3), we get the following form of the equation.
   2
∂ψ ∂ T ∂ψ ∂ T ∂2T μn f ∂ 2ψ
− = αn f 2 +  
∂y ∂x ∂x ∂y ∂y ρc p n f ∂ y2
1 16σ ∗ T∞ ∂ T
3 2
+  (6)
ρc p n f 3k ∗ ∂ y 2

The relation between the physical quantities of nanofluid, water, copper and
aluminum oxide is written as follows:
The thermophysical characteristics of the nanofluid with nanoparticle concentra-
tion ϕ are defined as (Oztop et al. [15])
μf
μn f = , ρn f = (1 − ϕ)ρ f + ϕρs ,
(1 − ϕ)2.5
κn f      
αn f =  , ρc p n f = (1 − ϕ) ρc p f + ϕ ρc p s ,
ρc p n f

(ρβ)n f = (1 − ϕ)(ρβ) f + ϕ(ρβ)s ,

   
κn f ks + 2k f − 2ϕ k f − ks
=    (7)
κf ks + 2k f + 2ϕ k f − ks

The following similarity transformations are invoked:


√ a T − T∞
Ψ = x aγ f f (η), η = y , (η) = (8)
γf Tw − T∞

Using these variables in Eqs. (2) and (6), we get the following nonlinear ordinary
differential equations:
   

ρs ρβs
f  + (1 − ϕ)2.5 1−ϕ+ϕ f f  − f 2 + λ 1 − ϕ + ϕ θ cos α − M f 
ρf ρβ f
− K1 = 0 (9)

  
4 kf ρβs
1 + R θ  + Pr 1−ϕ+ϕ
3 kn f ρβ f
Radiative Heat Transfer of Magnetic Nanofluid Flow Past … 1089
⎡ ⎤
Ec  
⎣ f θ  − f θ +  f  2 ⎦=0 (10)
ρc
(1 − ϕ)2.5 1 − ϕ + ϕ ρc ppsf

subject to the boundary conditions

f (η) = S, f  (η) = 1, θ (η) = 1 at η = 0 (11)


f = 0, θ (η) = 0 at η → ∞. (12)

σ B02
where the prime denotes differentiation with respect to η,M = aρ f
is the magnetic
gβ f (Tw− T∞ )x 3 Uw (x)x
parameter, (Gr x ) f = ϑ 2f
is the local Grashof number, (Rex ) f = ϑf
(Gr x ) f gβ(Tw −T∞ )
is the local Reynolds number, λ = (Rex )2f
= a2
is the mixed convection
4σ ∗ T∞ 3 ϑf
parameter, R = k ∗ kn f
is the radiation parameter, K 1 = aK
is the porous medium
Uw (x)2
  −1
parameter,Ec = c p f (Tw −T∞ )
is the Eckert number, S = −vw (x) aϑ f 2 is the suction
γf
parameter,Pr = αf
is the Prandtl number and α is the angle of inclination.

3 Skin Friction Coefficient and Nusselt Number

The important physical quantities in this study are skin friction coefficient C f and
the Nusselt number N u x , which are defined by the following equations:
 
μn f ∂u
Cf = ,
ρ f Uw (x)2 ∂ y y=0
 ∗
 4
−kn f ∂∂Ty − 4σ
3k ∗
∂T
∂y
y=0 y=0
N ux = (13)
k f (Tw− T∞ )/x

Using Eqs. (8) and (13), the skin friction coefficient and the Nusselt number can
be expressed as


1 1 
−1 −kn f 4R 
(Rex ) f C f =
2
f (0), (Rex ) f N u x =
2
1+ θ (0). (14)
(1 − ϕ)2.5 kf 3
1090 M. Shanmugapriya

Table 1 Thermal properties

Properties Base fluid (water) Cu Al2 O3
cp (J/kg K) 4179 385 765
ρ (kg/m3 ) 997.1 8933 3970
κ (W/m K) 0.613 400 40
β × 105 (k−1 ) 21 1.67 0.85


Table 2 Comparison of − f  (0) and −θ (0) for various values of M
M Ishak et al. [16] Anjali Devi et al. [17] Present result
  
− f (0) −θ  (0) − f (0) −θ  (0) − f (0) −θ  (0)
0.0 0.5607 1.0873 0.5608 1.0873 0.5606 1.0873
0.1 0.5658 1.0863 0.5659 1.0863 0.5659 1.0863
0.2 0.5810 1.0833 0.5810 1.0833 0.5810 1.0833
0.5 0.6830 1.0630 0.6830 1.0630 0.6830 1.0630
1.0 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000
2.0 1.8968 0.8311 1.8968 0.8311 1.8968 0.8311
5.0 4.9155 0.4702 4.9155 0.4703 4.9155 0.4703

4 Numerical Method

Equations (9)–(12) are solved numerically using Runge–Kutta–Fehlberg fourth-fifth-

order method along with shooting technique. A comparison of the obtained numerical
results with previously published data is performed. The values of skin friction coef-
ficient f  (0) for various values of M are compared with those obtained by Ishak
et al. [16] and Anjali Devi et al. [17], which is shown in Table 2 and observed that
both the results are in good agreement.

5 Results and Discussion

Graphical distributions and table representation of the influence of various physical

parameters on velocity, temperature profiles, local skin friction and local Nusselt
number are presented in Figs. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, and 12 and Table 3. In
this paper, we computed the above results by fixing the value Pr = 6.2, K 1 = 1 and
S = 0.5. Figures 1 and 2 represent the dimensionless velocity and temperature for
different values of the magnetic parameter for both types of Cu-water and Al2 O3 -
water nanofluids. It is clearly shown in Fig. 1, the velocity of the nanofluid decreases
with increasing strength of the magnetic field for both the types of nanofluid. The
 
Table 3 Values of − f  (0) and −θ (0) for various values of M, ϕ, α, λ, R, Ec, S and Pr .

M ϕ α λ R Ec S Pr − f  (0) −θ (0)
Cu Al2 O3 Cu Al2 O3
◦
0.0 0.01 45 1.0 1.0 0.1 0.5 6.2 1.1782 1.1922 2.4666 2.4579
0.5 1.2884 1.3017 2.4326 2.4241
1.0 1.5716 1.5834 2.3452 2.3371
◦
0.5 0.02 45 1.0 1.0 0.1 0.5 6.2 1.3230 1.3461 2.3638 2.3479
0.03 1.3549 1.3872 2.2977 2.2748
0.04 1.3843 1.4251 2.2339 2.2044
◦
0.5 0.01 0 1.0 1.0 0.1 0.5 6.2 1.1668 1.1770 2.4591 2.4514
◦
30 1.3901 1.4012 2.4091 2.4011
◦
60 1.6986 1.7111 2.3300 2.3212
◦
0.5 0.01 45 1.5 1.0 0.1 0.5 6.2 1.2220 1.2324 2.4472 2.4395
2.0 1.1537 1.1638 2.4618 2.4542
2.5 1.2185 1.0960 2.4680 2.4682
◦
0.5 0.01 45 1.0 2.0 0.1 0.5 6.2 1.2481 1.2619 1.7621 1.7554
Radiative Heat Transfer of Magnetic Nanofluid Flow Past …

3.0 1.2156 1.2298 1.4167 1.4108

4.0 1.1883 1.2029 1.2018 1.1964
◦
0.5 0.01 45 1.0 1.0 0.2 0.5 6.2 1.2829 1.2963 2.3031 2.2930
0.3 1.2774 1.2909 2.1752 2.1636
0.4 1.2720 1.2855 2.0490 2.0357
◦
0.5 0.01 45 1.0 1.0 0.1 1.0 6.2 1.6612 1.6799 3.3482 3.3365
3.0 3.4628 3.5093 7.7218 7.695
5.0 5.4588 5.5355 12.4381 12.3964
◦
0.5 0.01 45 1.0 1.0 0.1 0.5 1 1.0854 1.1011 0.6889 0.6852
10 1.3224 1.3352 3.4448 3.4337
100 1.5891 1.6008 23.4968 23.4393
1091
1092 M. Shanmugapriya

Fig. 1 Velocity profiles for

different values of M

Fig. 2 Temperature profiles

for different values of M

Fig. 3 Velocity profiles for

different values of ϕ
Radiative Heat Transfer of Magnetic Nanofluid Flow Past … 1093

Fig. 4 Temperature profiles

for different values of ϕ

Fig. 5 Velocity profiles for

different values of λ

Fig. 6 Temperature profiles

for different values of λ
1094 M. Shanmugapriya

Fig. 7 Velocity profiles for

different values of R

Fig. 8 Temperature profiles

for different values of R

Fig. 9 Velocity profiles for

different values of Ec
Radiative Heat Transfer of Magnetic Nanofluid Flow Past … 1095

Fig. 10 Temperature
profiles for different values
of Ec

Fig. 11 Velocity profiles for

different values of α

Fig. 12 Temperature
profiles for different values
of α
1096 M. Shanmugapriya

reductions in the velocity are accompanied by simultaneous reductions in the hydro-

dynamic boundary layer thickness for both the copper–water and alumina–water
nanofluids. Increasing the values of the magnetic parameter (M) is to increase the
temperature distribution. Therefore, the thermal boundary layer thickness increases
with the increasing values of M. The effect of the volume fraction parameter (ϕ) on
velocity, temperature profiles is illustrated in Figs. 3 and 4. It is shown in Fig. 3, for
the increased range of volume fraction (ϕ) is to suppress the velocity distribution for
both the types of Cu-water and Al2 O3 -water nanofluids and the opposite phenomena
occur in the temperature profile.
The effect of the mixed convection parameter (λ) for both the types of Cu-water
and Al 2 O3 -water nanofluids is plotted in Figs. 5 and 6. It is observed that the increase
in the mixed convection parameter enhances the velocity but diminishes the tem-
perature for both the nanofluids. Figures 7 and 8 show the deviation of velocity and
temperature fields for different values of the radiation parameter (R). It is noticed
that increasing the radiation parameter improves the hydrodynamical and thermal
boundary layer. The impact of the viscous dissipation parameter on the velocity and
the temperature is depicted in Figs. 9 and 10. It is found that the hydrodynami-
cal boundary layer and the thermal boundary layer fall for the increasing values of
the Eckert number (Ec). Figures 11 and 12 are plotted to investigate the effect of
angle of inclination α for both Cu-water and Al2 O3 -water nanofluids. The velocity
of the nanofluid suppresses with increasing the values of α. Physically increase in
the inclination angle decreases the buoyancy forces, so velocity profile decreases.
We concluded from Fig. 11 that the nanofluid velocity is maximum at the vertical
position of the plate, and the opposite phenomena occur in the nanofluid temperature
for both copper–water and alumina–water.

6 Conclusion

The boundary layer flow and heat transfer of a nanofluid over a porous inclined
plate with the effect of viscous dissipation, radiation and magnetic field have been
studied numerically. The nonlinear ordinary differential equations with given bound-
ary conditions were solved using Runge–Kutta–Fehlberg fourth-fifth-order method
along with shooting technique. This study investigates the effect of various physical
parameters on the fluid flow and heat transfer characteristics. The following results
can be drawn from the present study.
• When the magnetic parameter (M) raises, it reduces the velocity, skin friction
coefficient and enhances the temperature, Nusselt number due to the effect of
Lorentz force.
• The hydrodynamical and thermal boundary layer was increased by increasing
radiation parameter (R), and this trend is reversed in the viscous dissipation
parameter (Ec).
Radiative Heat Transfer of Magnetic Nanofluid Flow Past … 1097

• Skin friction coefficient decreases for increasing the values of the suction
parameter (S) and Prandtl number (Pr ).
• Suction parameter (S) and Prandtl number (Pr ) increases the thermal boundary
layer thickness.
• It is found that, in general, the non-dimensionless heat transfer rate for the copper–
water nanofluid is more compared with the alumina–water nanofluid.

Acknowledgements The author wishes to thank the Management of Sri Sivasubramaniya Nadar
College of Engineering for providing the necessary facility to carry out the present work.

References

1. Choi SUS (1995) Enhancing thermal conductivity of fluids with nanoparticles, developments
and applications of non-Newtonian flows. In: Siginer DA, Wang HP (eds) FED-vol 231/MD,
vol 66, pp 99–105. The American Society of Mechanical Engineers
2. Eastman JA, Choi SUS, Li S, Thompson LJ, Lee S (1997) Enhanced thermal conductivity
through the development of nanofluid. In: Komarneni S, Parker JC, Wollenberger HJ (eds)
Nanophase and Nanocomposite Materials II. Pittsburg, PA, MRS, pp 3–11
3. Eastman JA, Choi SUS, Li S, Yu W, Thompson LJ (2001) Anomalously increased effective
thermal conductivities of ethylene glycol-based nano-fluids containing copper nano-particles.
Appl Phys Lett 78:718–720
4. Xuan Y, Lin Q (2003) Investigation on convective heat transfer and flow features of nanofluids.
J Heat Trans 125:151–155
5. Xie H, Wang J, Xi T, Liu Y, Ai F, Wu Q (2002) Thermal conductivity enhancement of
suspensions containing nanosized alumina particles. J Appl Phys 91:4568–4572
6. Hamad MA, Pop I, Ismail AI (2011) Magnetic field effects on free convection flow of a nanofluid
past a vertical semi-infinite flat plate. Nonlinear Anal Real World Appl 12:1338–1346
7. Sudarsana Reddy P, Sreedevi P, Chamkha AJ (2017) MHD boundary layer flow, heat and
mass transfer analysis over a rotating disk through porous medium saturated by Cu–water and
Ag–water nanofluid with chemical reaction. Powder Technol 307:46–55
8. Sudarsana Reddy P, Suryanarayana Rao KV (2015) MHD natural convection heat and mass
transfer of Al2 O3 —water and Ag–water nanofluids over a vertical cone with chemical reaction.
Proc Eng 127:476–484
9. Shanmugapriya M (2018) Analysis of heat transfer of Cu-water nanofluid flow past a moving
wedge. J Inf Math Sci 10(1&2):287–296
10. Shanmugapriya M, Sangeetha P (2018) Entropy generation analysis of copper–water nanofluid
flow over a moving wedge. Desalination Water Treatment 121:14–21
11. Thumma T, Anwar Beg O, Kadir A (2017) Numerical study of heat source/sink effects on
dissipative magnetic nanofluid flow from a non-linear inclined stretching/shrinking sheet. J
Mol Liquids 232:159–173
12. Sudarsana Reddy P, Sreedevi P, Chamkha AJ (2016) Heat and mass transfer flow of a nanofluid
over an inclined plate under enhanced boundary conditions with magnetic field and thermal
radiation. Heat Transfer-Asian Res 46(7):815–839
13. Sudarsana Reddy P, Chamkha AJ, Al-Mudhaf A (2017) MHD heat and mass transfer flow of a
nanofluid over an inclined vertical porous plate with radiation and heat generation/absorption.
Adv Powder Technol 28(3):1008–1017
14. Tiwari RK, Das MK (2007) Heat transfer augmentation in a two-sided lid-driven differentially
heated square cavity utilizing nanofluids. Int J Heat Mass Trans 50(9–10):2002–2018
1098 M. Shanmugapriya

15. Oztop H, Abu-Nada E (2008) Numerical study of natural convection in partially heated
rectangular enclosures filled with nanofluids. Int J Heat Fluid Flow 29(5):1326–1336
16. Ishak A, Nazar R, Pop I (2008) Hydromagnetic flow and heat transfer adjacent to stretching
vertical sheet. Heat Mass Trans 44:921–927
17. Anjali Devi SP, Suriyakumar P (2017) Effect of magnetic field on Blasius and Sakiadis flow
of nanofluids past an inclined plate. J Taibah Univ Sci 11(6):1275–1288
A GA-ANFIS Model for the Prediction
of Biomass Elemental Properties

Obafemi O. Olatunji, Stephen Akinlabi, Nkosinathi Madushele,

and Paul A. Adedeji

Abstract The elemental composition of biomass is a significant property, which

determines the energy content of biomass feedstock. This article develops a pre-
diction model based on a hybrid adaptive neuro-fuzzy inference system (ANFIS)
optimized with genetic algorithm (GA). The model inputs were the proximate con-
stituents of biomass which are ash, fixed carbon, and volatile matter. These were used
to predict the hydrogen (H), oxygen (O) and carbon (C) content of biomass fuels.
The proposed algorithm was evaluated based on some known performance metrics.
The root mean squared error (RMSE), mean absolute deviation (MAD), mean abso-
lute percentage error (MAPE), coefficient of correlation (CC), mean absolute error
(MAE) are 3.673, 2.4609, 5.1757, 0.9464, 0.309 at computation time (CT) of 33.65
secs for carbon (C); 0.6293, 0.4168, 8.3011, 0.75581, 0.0716 at CT of 40.21 secs
for hydrogen (H); 4.4538, 3.1042, 13.3983,0.9167, 0.9899 at CT of 33.57 secs for
oxygen (O), respectively. Regression analysis was also carried out to determine the
level of dependence among the correlated variables. The model performance shows
that GA-ANFIS can be applied in the computation of the elemental composition of
biomass for strategic decision-making.

Keywords GA-ANFIS · Biomass feedstock · Efficient utilization · Elemental

composition

1 Introduction

As of 2017 the United Nation, UN projected population growth from 7.6 billion
to 9.8 billion by 2050 [1]. The clean and sustainable energy is the viable solution

O. O. Olatunji (B) · N. Madushele · P. A. Adedeji

Department of Mechanical Engineering Science, University of Johannesburg, Johannesburg,
South Africa
e-mail: tunjifemi@gmail.com
S. Akinlabi
Department of Mechanical and Industrial Engineering, University of Johannesburg, Johannesburg,
South Africa

© Springer Nature Singapore Pte Ltd. 2021 1099

S. Vijayan et al. (eds.), Trends in Manufacturing and Engineering Management,
Lecture Notes in Mechanical Engineering,
https://doi.org/10.1007/978-981-15-4745-4_95
1100 O. O. Olatunji et al.

to achieve safer and energy-efficient world since the rising population is leading to
increasing concentration of CO2, which was reported at 1.4% as at 2017 as the highest
ever recorded. Considering the yardsticks such as availability, sustainability, energy
efficiency, applicability, environmental impact, and flexibility, biomass is promising
option to address over-dependence on fossil fuel and provide attractive feedstocks for
energy generation [2]. Biomass elemental composition is an indispensable parameter
required for the design and operation of biomass-to-fuel conversion technologies [3].
The carbon, hydrogen, oxygen content, and other constituents in negligible quantity
have an impact on the gross heating value, the emission, and corrosion of conver-
sion equipment. The knowledge of elemental composition helps in: calculating the
material balance in bioprocess, development of biomass handling equipment such
as boilers, gasifiers, combustors, estimating the volume and constituents of combus-
tion gases, and fixing the amount of oxidant required for biomass gasification [4].
However, experimental determination of elemental composition using the essential
apparatus is becoming very expensive and time-consuming compared to proximate
analysis which can be conducted much easily. Therefore, the rapid and economical
approach would be to develop a model which can be used to estimate the elemental
constituent from the proximate value since it has been established that there is a rela-
tionship between these two experimental methods. Against all these odds, artificial
intelligence, AI, has emerged with a potential to assist in fulfilling the objectives of
biomass-to-energy exploration which centres around the estimation of the proper-
ties of biomass, such as the elemental composition and enthalpy of combustion [3,
5–10]. AI has the capability for modelling of complex systems with high degree of
accuracy. AI can be grouped into artificial neural network (ANN), adaptive neuro-
fuzzy inference systems (ANFIS), evolutionary algorithms (EA), and support vector
machines (SVMs) [11–16]. Several optimization methods have been proposed by
researchers to improve the accuracy of ANFIS [17–19], one of these methods is the
genetic algorithm, GA. GA is an optimization technique which can be applied to
solve problems built on natural selection process among the population such that a
population of candidate solutions is advanced towards improved solutions [20–22].
It was developed in the 1970s by Holland with the basic idea of abstract modelling of
natural evolution [23]. The GA has been extensively used in different fields [24–26]
and most closely in the prediction of the heating value of biomass [13], but there
is no any know GA application for the prediction of the elemental composition of
biomass which covers different sources. There are quite a few published models
which predict the elemental composition of solid biomass based on proximate anal-
ysis [3, 5, 8–10, 27–30]. Nhuchhen [30] developed a correlation which estimated the
carbon, oxygen, and hydrogen constituent of untreated and torrefied biomass on the
basis of proximate analysis. Similarly, Mahmudul et al. [9] developed an empirical
correlation towards the prediction of the elemental composition of torrefied woody
biomass. Also, the correlation developed by Vakkilainen [29] is only limited to black
liquor. Parikh et al. [10] and Shen et al. [5] considered a data set for proximate values
which are volatile matter, fixed carbon, ash for an equivalent elemental composition
of carbon, hydrogen, and oxygen. The range of the data considered is as shown in
Table 1. It can be observed that the present study has a wider datapoint coverage
A GA-ANFIS Model for the Prediction … 1101

Table 1 Datapoint coverage for the proximate and ultimate properties of biomass
Reference Constituents Datapoint range %
[10] C 36.2–53.1
H 4.4–8.3
O 31.4–49.5
VM 57.2–90.6
FC 4.7–38.4
ASH 0.12–77.7
[5] C 36.2–53.1
H 4.7–6.6
O 31.4–48.0
VM 57.2–90.6
FC 9.2–32.8
ASH 0.1–24.6
Present study C 22.35–92.86
H 0.52–11.42
O 2.32–52.06
VM 5.04–89.60
FC 7.40–94.73
ASH 0.10–56.10

compared to previous studies in the same domain. The performance of these corre-
lations was evaluated as shown in Table 2. The contribution of ash content in the
prediction of elemental composition was considered by Shen [5] which made the
correlation to be more robust with better prediction than Parikh et al. [10].
All the mentioned correlations are linear and have not satisfactorily predicted
the properties of the biomass since the elemental constituent of biomass and other
dependent variables do not constantly follow a linear pattern, particularly when tested
with various samples. Ghugare et al. [3] has proposed a hybrid genetic programming
algorithm, support vector regression (SVR), and artificial intelligence in order to

Table 2 Correlations for the estimation of elemental composition

Reference Equation* R2 RMSE MAPE
[10] C = 0455VM + 0.637FC 0.810 6.763 4.990
H = 0.062VM + 0.052FC 0.500 1.082 19.734
O = 0.476VM + 0.304FC 0.890 1.530 26.374
[5] C = 0.46VM − 0.095ASH + 0.635FC 0.745 7.025 5.149
H = 0.06VM + 0.01ASH + 0.059FC 0.353 1.169 13.11
O = 0.469VM − 0.023ASH + 0.34FC 0.828 6.839 28.304
1102 O. O. Olatunji et al.

predict the elemental composition biomass. The result of their investigation showed
that the model developed was more accurate than the existing linear and polynomial
models.
Since there is no known GA-ANFIS model which has been used to predict the
elemental composition of solid biomass with such a large data coverage, this article
proposed a prediction model based on a hybrid adaptive neuro-fuzzy inference system
(ANFIS) optimized with GA. The model inputs were the proximate properties of
biomass which are ash content (ASH), fixed carbon (FC), and volatile matter (VM).
These were used to predict the hydrogen (H), oxygen (O), and carbon (C) content of
biomass fuels.

2 Materials and Methods

2.1 Collection of Data

In predicting the elemental constituents of biomass-based on the proximate analysis

results, the component of proximate which are VM, FC, and ash in weight percent-
ages (dry basis) were used as input variables to estimate the C, H, and O as the
output variables. The ranges of the proximate value are shown as the present study in
Table 1. A cumulative of 830 datapoints were extracted from the database compiled
by Ghugare et al. [3]. After randomization, the data were divided into training and
testing at ratio 7:3, respectively, for the development of the model.

2.2 Theory of ANFIS

ANFIS combines the dual benefits of fuzzy logic and the neural network in a sin-
gle framework, which enables it to solve complex problems in a way that optimal
parameters of membership function are obtained from input to output mapping [15,
31]. It is a variant of the Takagi–Sugeno fuzzy inference system [32, 33]. It consists
of input–output parameters and fuzzy rules of Takagi–Sugeno class [15]. It is a class
of adaptive multilayer feed-forward network which assumes a framework of two
inputs H1 , H2 and output G such that the entries are evaluated by linguistic variable
(A1 , B1 ). The first-order Takagi–Sugeno fuzzy model can be expressed as:
Principle 1: If H1 is A1 AND H2 is B1 then G 1 = b1 J1 + c1 J2 + z 1 .
Principle 2: If H1 is A2 AND H2 is B2 then G 2 = b2 J1 + c2 J2 + z 2
where b1, c1 , z1 , b2, c2 , z2 are output function parameters. In this article, the two rules
above were adapted with input Ji (i = 1 . . . 3), output G i (i = 1 . . . 3) and cluster
C j ( j = 1 . . . 10) with the same weights using a fuzzy c-means clustering method.
Typically, the ANFIS network has five layers [16] in the following sequence:
fuzzification layer, multiplication layer, rationalization layer, defuzzification layer,
A GA-ANFIS Model for the Prediction … 1103

and summation layer. It should be noted that the first and fourth layer include the
adaptive parameters capable of modification and update based on a learning tech-
nique used. The estimated output ultimately depends on the five layers in the ANFIS
after the completion of the training phase. From previous studies, the model per-
forms optimally when the fuzzy rules based on the Gaussian membership function
are applied. This study also applied the Gaussian membership function on the input
variables. The readers can refer to [15, 16, 22, 25, 32, 34, 35] for further informa-
tion on ANFIS models. In this study, elemental constituents comprising of three
inputs which are VM, FC, and ash were mapped to the output which is the principal
elemental constituents (C, H, O) of biomass.

2.3 Model Optimization Based on Genetic Algorithm (GA)

GA is an evolutionary heuristic optimization algorithm which is based on population

[20, 31, 36–38]. The process which leads to the generation of new solution in GA
has a semblance of the natural selection process of living organisms. Usually, the
recurrent application of three evolutionist operators occurs in three phases which
are: selection, crossover, and mutation which all assist in the attainment of new
solutions. Mutation can be defined as a mechanism through which a population is
diversified in order to avoid stagnation [39, 40]. In GA, a population of candidate
solutions is advanced to improve the solutions. Each potential solution has a set
of characteristics termed genotype, which can be mutated. The objective function
governs the selection of individuals for a new generation. During the selection process
of the elites for the next generation, a fitness value is obtained for each member of the
population. This value determines which member proceeds to the next generation.
Consequentially, the optimized generation of the potential solutions is used in the
following iteration [41, 42]. Apart from population density and the optimal number of
iterations, the selection, crossover, and mutation techniques must be considered in GA
decision-making process. Various selection methods may adopt different probability
of assignment in order to ensure population diversity and obtain an optimal solution.
The two most popular selection criteria are: tournament and roulette wheel. This study
adopts the roulette wheel for the selection criteria during parameter optimization in
the hybrid model. Crossover method should be selected carefully such that a simple
method which may lead to unusable chromosome for the complex optimization
problem is avoided. The concept of mutation has been comprehensively discussed
by Gen et al. [39] Holland [23], Pal et al. [40]. The selection of mutation technique
and probability will ensure that the diversity of the population is maintained.
The antecedent and consequence of a typical ANFIS model are optimized by the
training function. This optimization is performed on the adaptive layers which are the
first and the fourth layers [36]. In this investigation, the GA algorithm was written
in MATLAB (R2015a) and used to train the ANFIS model. An iterative training
process aimed at achieving optimal parameters for the antecedent was performed.
1104 O. O. Olatunji et al.

The Gaussian membership function (Eq. 1) was used due to its less disruptive effect
and robustness when applied to the system [43, 44].

∅(z) = e− 2 ( )
1 z−μ 2
σ (1)

where ∅ = Gaussian function, σ and μ = the membership function parameters.

2.4 Implementation of ANFIS-GA Model

Shown in Fig. 1 is the flow diagram of GA-ANFIS model developed for the prediction
of C, H, O components of biomass-based on VM, FC, and ash contents as the input

Initialize GA

Generate

Initial population

Evaluate individual
fitness

Rank individual fitness

Is any stopping criteria Yes Use global optimal value

met? for ANFIS parameters

No
Generate new population
Selection
Crossover ANFIS-GA
Mutation End

Fig. 1 Flow diagram for ANFIS-GA implementation

A GA-ANFIS Model for the Prediction … 1105

Table 3 Learning and

User-defined parameters Value
network optimization
parameters Number of rules 20
FIS type Takagi–Sugeno
Initial population size 20
Population type Double vector
Selection function Roulette wheel
Mutation rate 0.15
Crossover function Heuristic
Mutation function Gaussian
Crossover percentage 0.4
Number of iterations 600
Stopping criteria Minimum training error

variables. Genetic operators are applied to facilitate breeding in GA. Using GA, the
ANFIS model was trained until the optimal solution was obtained.
The initial conditions used for the computation is presented in Table 3. The con-
vergence values, which minimize the training error objective function, were made a
stopping criterion. The computation was done on a desktop computer workstation
with configuration 64 bits, 32 GB RAM Intel (R) Core (TM) i7 5960X. In general
terms, the principle guiding the GA-ANFIS is as highlighted in the steps below:
Step 1: Generate the first-string population
Step 2: Assign and determine the fitness of each member of the population.
Step 3: Choose a membership pair from the population string for breeding
Step 4: Rank the fitness of the members of the population
Step 5: Allocate the pair into passing population membership
Step 6: Repeat step 5, until temporary population membership is filled up
Step 7: Replace the existing population with members of passing population and
some members of existing population to complete the membership
Step 8: Use the optimal population value for the ANFIS parameters then end
GA-ANFIS
Step 9: If the termination criteria are not met, return to step 2.

2.5 Model Performance Evaluation

Statistical estimation of errors was carried out to evaluate the performance of GA-
ANFIS model which was designed in this study as follows: computation time (CT)
was also estimated as a measure of the overall prediction time as to evaluate the
time cost of the forecasting and to further underline the economic significance of
this model.
1106 O. O. Olatunji et al.

Correlation Coefficient (CC):

n   2
yk − yk
CC = 1 − i=1
n (2)
i=1 (yk − yk )2

Root Mean square Error (RMSE):


N   2
k=1 yk − yk
RMSE = (3)
N

Mean Absolute Deviation (MAD):

1
N
MAD = |yk − ȳk | (4)
N k=1

Mean Absolute Percentage Error (MAPE)

N  
1  yk − ȳk 
MAPE = × 100 (5)
N k=1  yk 

3 Results

3.1 Model Validation

To satisfactorily predict the elemental component of biomass, the model was trained,
and the prediction capacity of the models was evaluated at the testing phase. Each
elemental composition was considered independently during the training and test-
ing phase. Each predicted elemental constituents and corresponding experimental
values were plotted against the data index. Figures 2, 3 and 4 were plotted for car-
bon, hydrogen, and oxygen, respectively, at both testing and training phase to show
the prediction progression from training to testing. The results show a satisfactory
concurrence between the actual and predicted elemental constituents with a slight
variation at training and testing phases. The concurrency between the actual and
predicted carbon and oxygen revealed a closer pattern compared to hydrogen. In
case of hydrogen, the lower accuracy could be as a result of experimental errors and
narrower data range when compared to other experimental constituents [3].
A GA-ANFIS Model for the Prediction … 1107

Fig. 2 a Testing phase for carbon b Training phase for carbon

3.2 Model Evaluation

For each model, the error was statistically calculated for the predicted and actual
elemental constituents at training and testing stages as presented in Table 4. The
RMSE, MAD, and STD are consistently lower for hydrogen at the training and
testing phase when compared to carbon and oxygen. However, despite the lowest
RMSE, MAD, and STD reported for hydrogen, the magnitude of MAPE and CC
1108 O. O. Olatunji et al.

Fig. 3 a Testing phase for hydrogen b Training phase for hydrogen

shows that the model does not predict as much data correctly as the models for
carbon, though it does better than the oxygen. The computation time of the carbon
and oxygen were approximately the same, but significantly different from hydrogen.
This further validates the initial observation that the hydrogen model is not as accurate
and economical as the others. The MAPE revealed that around 80% of the datapoint
were correctly predicted for hydrogen. Of all the three constituents, carbon was the
most accurately predicted with 95% of the data set followed by hydrogen (>80%)
and oxygen (>70%). This is at variance with various models (GP-MLP, GP-SVR,
A GA-ANFIS Model for the Prediction … 1109

Fig. 4 a Training phase for Oxygen b Testing phase for Oxygen

MLP-SVR) developed by Ghugare [3] which proposed that the MAPE of carbon and
oxygen constituents is generally and constantly better than hydrogen for all testing
cases. The CT results revealed that the model can be tested in less than 1 min, though
closely related to the computing power of the computer used. The result is important
in the conversion of biomass to energy since the properties of biomass feedstock
determine the energy quality.
The comparative analysis of the GA-ANFIS model with the models developed
by other authors is indicated in Table 5. The RMSE of hydrogen is significantly
1110 O. O. Olatunji et al.

Table 4 Predicted Elemental components (C, H, O)

Evaluation parameters Hydrogen Carbon Oxygen
Training Testing Training Testing Training Testing
RMSE 0.696 0.6293 3.673 3.3923 4.2235 4.4538
MAD 0.4309 0.4168 2.4609 2.383 2.85 3.1042
MAPE 9.4125 8.3011 5.1757 4.9733 11.5911 13.3983
Error STD 0.6966 0.6265 3.6725 3.385 4.1843 4.3512
CT (s) 40.205 33.65 33.57
CC 0.7861 0.7559 0.9159 0.9464 0.9024 0.9167

Table 5 Performance of GA-ANFIS compared to other models

Models RMSE MAD MAPE CC
Hydrogen
This study 0.629 0.417 8.301 0.756
Parikh et al. [10] 1.082 – 19.734 0.500
Shen et al. [5] 1.169 – 22.276 0.353
Carbon
This study 3.392 2.383 4.973 0.946
Parikh et al. [10] 6.763 – 4.990 0.810
Shen et al. [5] 7.025 – 5.149 0.745
Oxygen
This study 4.454 3.104 13.398 0.917
Parikh et al. [10] 6.301 – 26.374 0.890
Shen et al. [5] 6.839 – 28.304 0.829

lesser than that of the other two models, while the CC is also significantly higher
[5, 10]. Also, the MAPE calculated based on Shen et al. [5] and Parikh et al. [10]
are pointed of a lower order of magnitude with about half of the value estimated in
this study. For carbon, CC is lower and RMSE is also significantly lower by more
than 50% for the model developed in this study. However, the MAPE was close
as against a significant disparity in case of hydrogen constituent prediction model.
Also, for oxygen model developed in this study, RMSE value shows a significant
order of difference and the CC, though not as much as RMSE. But MAPE showed
that the other linear correlation could only accurately predict around 50% of the
data. This further underlines the significance of the model developed in this study.
The overall observation revealed that the GA-ANFIS model developed in this study
provides more satisfactory performance for all the elemental components which were
predicted in compared to the existing linear models.
A GA-ANFIS Model for the Prediction … 1111

Fig. 5 Regression for C

constituent

3.3 Post-regression Analysis

The regression analysis was based on the test results for all the predicted elements.
The objective is to determine the model of greatest fitness based on minimum error
criterion and the data dispersion along the line of the best fit (Figs. 5, 6 and 7). It can
be inferred from the plots that carbon in Fig. 5 shows the lowest dispersion compared
to other constituents with regression equation given as: F(C)predicted = 0.83F(C)exp
+ 8.2 in %weight.

4 Conclusion

The elemental constituents of biomass were predicted based on the proximate val-
ues using ANFIS model which was optimized by GA. The accuracy of the model
developed in the study was more satisfactory compared to the other linear models
since it resolved nonlinear dependencies inherent in biomass properties. Regression
analysis was also carried out to determine the level of dependence among the corre-
lated variables and the data scattering along the line of best fit. The report shows that
GA-ANFIS can be used in the computation of the elemental composition of biomass-
based on proximate properties. This result is significant for strategic decision-making
regarding biomass-to-energy conversion operation.
1112 O. O. Olatunji et al.

Fig. 6 Regression for H

constituent

Fig. 7 Regression plot for

oxygen constituent

References

1. WorldNuclearAssociation (2018) World energy needs and nuclear power. http://www.

world-nuclear.org/information-library/current-and-future-generation/world-energy-needs-
and-nuclear-power.aspx. Accessed on 8 Dec 2018
A GA-ANFIS Model for the Prediction … 1113

2. Callejón-Ferre A, Velázquez-Martí B, López-Martínez J, Manzano-Agugliaro F (2011) Green-

house crop residues: energy potential and models for the prediction of their higher heating value.
Renew Sustain Energy Rev 15(2), 948–955
3. Ghugare SB, Tiwary S, Tambe SS (2017) Computational intelligence based models for predic-
tion of elemental composition of solid biomass fuels from proximate analysis. Int J Syst Assur
Eng Manage 8(4):2083–2096
4. Tumuluru JS, Wright CT, Boardman RD, Yancey NA, Sokhansanj S (2011) A review on
biomass classification and composition, co-firing issues and pretreatment methods. In: 2011
Louisville, Kentucky, American Society of Agricultural and Biological Engineers, p 1, 7–10
Aug 2011
5. Shen J, Zhu S, Liu X, Zhang H, Tan J (2010) The prediction of elemental composition of
biomass based on proximate analysis. Energy Convers Manage 51(5):983–987
6. Uzun H, Yildiz Z, Goldfarb JL, and S. Ceylan, “Improved prediction of higher heating value
of biomass using an artificial neural network model based on proximate analysis. Bioresour
Technol vol 234, pp 122–130, Jun 2017
7. ObafemiO, Stephen A, Ajayi O, Mashinini P, Nkosinathi M (2018) Experimental investigation
of thermal properties of Lignocellulosic biomass: a review. IOP Conference Series: Materials
Science and Engineering, vol 413, no. 1, p 012054
8. Posom J, Sirisomboon P (2017) Evaluation of lower heating value and elemental composition
of bamboo using near infrared spectroscopy. Energy 121, 147–158
9. Hasan M, Haseli Y, Karadogan E (2018) Correlations to predict elemental compositions and
heating value of torrefied biomass. Energies 11(9):2443
10. Parikh J, Channiwala SA, Ghosal GK (2007) A correlation for calculating elemental
composition from proximate analysis of biomass materials. Fuel 86(12–13):1710–1719
11. Vikhar PA (2016) Evolutionary algorithms: a critical review and its future prospects. In: 2016
International conference on Global Trends in Signal Processing, Information Computing and
Communication (ICGTSPICC), IEEE, pp 261–265
12. Gong S, Sasanipour J, Shayesteh MR, Eslami M, Baghban A (2017) Radial basis function
artificial neural network model to estimate higher heating value of solid wastes. Energy Sources
Part A Recovery Utilization Environ Effects 39(16):1778–1784
13. Baghban A, Ebadi T (2019) GA-ANFIS modeling of higher heating value of wastes: application
to fuel upgrading. Energy Sourc Part A Recovery Utilization Environ Effects 41(1):7–13
14. Baghban A, Kashiwao T, Bahadori M, Ahmad Z, Bahadori A (2016) Estimation of natural gases
water content using adaptive neuro-fuzzy inference system. Petroleum Sci Technol 34(10):891–
897
15. Jang JS (1993) ANFIS: adaptive-network-based fuzzy inference system. IEEE Trans Syst Man
Cybernetics 23(3):665–685
16. Akkaya E (2016) ANFIS based prediction model for biomass heating value using proximate
analysis components. Fuel 180:687–693
17. Bäck T, Fogel DB, Michalewicz Z (2018) Evolutionary computation 1: basic algorithms and
operators. CRC press
18. Wang GG, Deb S, Gao XZ, Coelho LDS (2016) A new metaheuristic optimisation algorithm
motivated by elephant herding behaviour. Int J Bio-Inspired Comput 8(6):394–409
19. Kachitvichyanukul V (2012) Comparison of three evolutionary algorithms: GA, PSO, and DE.
Indus Eng Manage Syst 11(3):215–223
20. Malhotra R, Singh N, Singh Y (2011) Genetic algorithms: concepts, design for optimization
of process controllers. Comput Inf Sci 4(2):39
21. Sarkheyli A, Zain AM, Sharif S (2015) Robust optimization of ANFIS based on a new modified
GA. Neurocomputing 166:357–366
22. Karaboga D, Kaya E (2018) Adaptive network based fuzzy inference system (ANFIS) training
approaches: a comprehensive survey. Artificial Intelligence Rev 1–31
23. Holland JH (1992) Adaptation in natural and artificial systems: an introductory analysis with
applications to biology, control, and Artificial Intelligence. MIT press
1114 O. O. Olatunji et al.

24. Rouhibakhsh K, Darvish H, Sabzgholami H, Goodarzi MS (2018) Application of ANFIS-

GA as a novel and accurate tool for estimation of interfacial tension of carbon dioxide and
hydrocarbon. Petrol Sci Technol 36(15):1143–1149
25. Esfandyari M, Esfandyari M, Jafari D (2018) Prediction of thiophene removal from diesel
using [BMIM][AlCl4] in EDS Process: GA-ANFIS and PSO-ANFIS modeling. Petroleum Sci
Technol 1–7
26. Armaghani DJ, Mohamad ET, Hajihassani M, Yagiz S, Motaghedi H (2016) Application of
several non-linear prediction tools for estimating uniaxial compressive strength of granitic
rocks and comparison of their performances. Eng Comput 32(2):189–206
27. Fried A, Padouvas E, Rotter H, Varmuza K (2005) Prediction of heating values of biomass fuel
from elemental composition. Analytica Chimica Acta 544(1–2):191–198
28. Yi L, Feng J, Qin YH, Li WY (2017) Prediction of elemental composition of coal using
proximate analysis. Fuel 193:315–321
29. Vakkilainen E (2000) Estimation of elemental composition from proximate analysis of black
liquor. Paperi ja puu 82(7):450–455
30. Nhuchhen DR (2016) Prediction of carbon, hydrogen, and oxygen compositions of raw and
torrefied biomass using proximate analysis. Fuel 180:348–356
31. Yaghoobi A, Bakhshi-Jooybari M, Gorji A, Baseri H (2016) Application of adaptive neuro fuzzy
inference system and genetic algorithm for pressure path optimization in sheet hydroforming
process. Int J Adv Manuf Technol 86(9–12):2667–2677
32. Takagi T, Sugeno M (1985) Fuzzy identification of systems and its applications to modeling
and control. IEEE Trans Syst Man Cybern 1:116–132
33. Sugeno M, Kang G (1988) Structure identification of fuzzy model. Fuzzy Sets Syst 28(1):15–33
34. Rezakazemi M, Dashti A, Asghari M, Shirazian S (2017) H2-selective mixed matrix membranes
modeling using ANFIS, PSO-ANFIS, GA-ANFIS. Int J Hydrogen Energy 42(22):15211–
15225
35. Suparta W, Alhasa KM (2016) Modeling of tropospheric delays using ANFIS. Springer
36. Haznedar B, Kalinli A (2016) Training ANFIS using genetic algorithm for dynamic systems
identification. Int J Intell Syst Appl Eng 4(1):44–47
37. Qdais HA, Hani KB, Shatnawi N (2010) Modeling and optimization of biogas production from
a waste digester using artificial neural network and genetic algorithm. Res Conserv Recycl
54(6):359–363
38. Jacob S, Banerjee R (2016) Modeling and optimization of anaerobic codigestion of potato
waste and aquatic weed by response surface methodology and artificial neural network coupled
genetic algorithm. Bioresour Technol 214:386–95, Aug 2016
39. Gen M, Cheng R, Lin L (2008) Network models and optimization: multiobjective genetic
algorithm approach. Springer Science & Business Media
40. Pal SK, Wang PP (2017) Genetic algorithms for pattern recognition. CRC press
41. Mayilvaganan MK, Naidu K (2011) Comparison of membership functions in adaptive-network-
based fuzzy inference system (ANFIS) for the prediction of groundwater level of a watershed.
J Comput Appl Res Dev 1(1):35–42
42. Yagiz S, Ghasemi E, Adoko AC (2018) Prediction of rock brittleness using genetic algorithm
and particle swarm optimization techniques. Geotech Geol Eng 36(6):3767–3777
43. Antonio LM, Coello CAC (2018) Coevolutionary multiobjective evolutionary algorithms:
survey of the state-of-the-art. IEEE Trans Evol Comput 22(6):851–865
44. Corne D, Lones MA (2018) Evolutionary algorithms. In: Handbook of Heuristics, pp 1–22