“ANALYSIS & OPTIMIZATION OF MACHINING

PARAMETERS FOR IMPROVEMENT IN SURFACE

FINISH AND MORPHOLOGY FOR TURBINE BLADES
OF MATERIAL TI6AL4V.”

Group Members:

MUHAMMAD ZEESHAN HAMEED (150524)

MUHAMMAD DILAWAR (150551)
AMEER HAMZA JAN KHATTAK (150565)

BE MECHANICAL FALL 15
Project Supervisor
Dr. Hamid Ullah Khan Niazi
Associate Professor

Project Co-Supervisor
Dr. Aamir Raza

DEPARTMENT OF MECHANICAL AND AEROSPACE ENGINEERING

INSTITUTE OF AVIONICS AND AERONAUTICS
AIR UNIVERSITY, ISLAMABAD
 “ANALYSIS & OPTIMIZATION OF MACHINING
PARAMETERS FOR IMPROVEMENT IN SURFACE
FINISH AND MORPHOLOGY FOR TURBINE BLADES
OF MATERIAL TI6AL4V.”

FINAL YEAR PROJECT REPORT

FALL-15

DEPARTMENT OF MECHANICAL AND AEROSPACE ENGINEERING

ii
 ANALYSIS & OPTIMIZATION OF MACHINING
PARAMETERS FOR IMPROVEMENT IN SURFACE
FINISH AND MORPHOLOGY FOR TURBINE BLADES
OF MATERIAL TI6AL4V.

Submitted by:

MUHAMMAD ZEESHAN HAMEED (150524)

MUHAMMAD DILAWAR (150551)
AMEER HAMZA JAN KHATTAK (150565)

Project Supervisor

Assoc. Prof. Dr. Hamid Ullah Khan Niazi

DMAE, IAA

Head of Department

Assoc. Prof. Dr. Jehanzeb Masud

DMAE, IAA

iii
ACKNOWLEDGEMENTS

All regards to Almighty Allah, the most beneficial, the most merciful and creator
of everything what we know and what we don’t.

This project is an outright example of group effort and outstanding backing offered
by the faculty members of Air University, Islamabad and PMO, NESCOM.

The help and technical support given by Associate Professor Dr. Hamid Ullah
Khan Niazi, Deputy DG NESCOM Dr. Aamir Raza and Dr. Muhammad Amin in
the development and implementation of experimental work on “Analysis &
optimization of machining parameters for improvement in surface finish and
morphology for turbine blades of material Ti6Al4V.”

The extraordinary direction and support for the project has been provided by, Dr
Hamid Ullah Khan Niazi, Dr Aamir Raza, Mr. Naeem Jan Khattak and Dr.
Muhammad Amin.

iv
ABSTRACT

Complex contoured parts such as turbine blades, impellers, propellers, helical

rotors, etc. are high functioning components in aerospace, automobiles and other
areas. Modern production facilities continuously try to implement the procedures
that offer a superior product within a short period of time, mainly in the
development of manufacturing procedures for new products. Gas turbines are vital
equipment aviation and power industries. Parts for these marvels of machinery
need to be reliable at extraordinary operating temperatures and pressures.
Conventional metals cannot function in these conditions so new superalloys have
been developed for these applications. These space-age metals such as titanium are
very difficult to work with and require special machining techniques and are
appropriately called difficult to cut metals.

Machining of an aerospace component such as turbine blades has got paramount

importance due to their crucial applications and reliability requirement. For
sensitive and critical applications, it is vital that the mechanical properties of such
materials are analyzed before and after machining. As post machining stresses and
surface work hardening can occur and reduce the life of the component. Surface
roughness is also important in these parts for proper post machining heat
treatments and additional protective coating applications. In light of these
requirements, we are tasked with developing a procedure to optimize the
machining parameters through the design of experiments using “Box-Behnken
design” which will benefit the manufacturing of advance aerospace materials with
reduced surface roughness and machining damage.

Test samples of turbines blades are manufactured using Ti6Al4V by changing the
various machining parameters; i.e. feed rate, axial depth of cut and cutting speed.
This provides us with data that will help to develop a system to machine these
components so that there is minimal surface roughness, highest material removal
rate (MRR), low tolerance, and the best surface quality achievable. This process
will provide an improvement in quality and enhancement in process efficiency and
will also give better tool life, topology and material properties.

v
LIST OF ABBREVIATIONS

Abbreviation Full Form

CNC Computer Numerical Control

DOE Design of Experiment

CAD Computer Aided Design

CMM Coordinate Measuring Machine

SEM Scanning Electron Microscope

CAM Computer Aided Machining

MRR Material Removal Rate

BUE Built up Edge

HCP Hexagonal Closed-Pack

BCC Body Centred Cubic

RSM Response surface method

ISO International standard organization

BBD Box-Behnken design

N Spindle speed

Vc Cutting speed

ae Radial depth of cut

f Feed Rate

fz Feed per tooth

ap Axial depth of cut

h Scallop height / Cusp height

deff Effective diameter

vi
Ra Roughness average

Rt Max-peak-valley height roughness

Rq Mean square average

Ti Titanium

V Vanadium

Al Aluminium

wt Weight

α Alpha

β Beta

vii
LIST OF FIGURES

Number Figure Page#

Figure 1: Fishbone diagram showing effects of machining parameters on

surface roughness [1]. .............................................................................. 2

Figure 2: Turbine blade ......................................................................................... 5

Figure 3: Multi-axis machine axis designation.[3] ................................................ 6

Figure 4: Configuration of CNC machining centre with a rotary axis[2] ............. 7

Figure 5: CC point and CL point identification .................................................... 8

Figure 6: Generation of CNC programs (standard method) [4] ............................ 8

Figure 7: Scallop height of ball end mill .............................................................. 10

Figure 8: Simulation for toolpath verification ..................................................... 10

Figure 9: Cutting path view of Ball end mill[6] ................................................... 12

Figure 10: Thermal conductivity of titanium at various temperature and

pressures [7] ........................................................................................... 15

Figure 11:“Cutting temperatures at the tool-chip interface when end milling

at feed per tooth of 0.15m/tooth”[7]....................................................... 16

Figure 12: Effects of spring back on of the cutting edge region[7]. ................... 17

Figure 13: Flow chart for the methodology of research work ............................. 26

Figure 14: “CoromillⓇ Plura solid carbide ball nose end mill” .......................... 28

Figure 15: The 3N full factorial for 03 factors in BBD........................................ 32

Figure 16: Iso- metric view of cut down model ................................................... 33

Figure 17: Side and top views of model used for experimentation ...................... 34

Figure 18: Operation manager tab on MasterCAM.............................................. 35

Figure 19: Experimental sample machining on CNC machine ............................ 36

Figure 20: Samples after machining ..................................................................... 37

viii
Figure 21: Experimental samples for roughness measurement and SEM
analysis ................................................................................................... 37

Figure 22: Profile of surface texture[18] .............................................................. 41

Figure 23: Graphical representation of Ra[18]..................................................... 41

Figure 24: Graphical representation of Rt & Rz [18] ......................................... 42

Figure 25: Surfcorder test results of roughness average (Ra) .............................. 47

Figure 26: Surface topography of machined samples through SEM .................... 51

Figure 27: Residual plot for linear model............................................................. 54

Figure 28: Residual plot for full quadratic model ................................................ 55

Figure 29: Main effect plots between Ra & input variables ................................. 56

Figure 30: Contour plot for Ra, spindle speed, axial depth & feed rate ............... 57

Figure 31: Surfcorder results for optimized values .............................................. 59

Figure 32: Machined turbine blade according to optimized values ..................... 59

ix
LIST OF TABLES

Number Table Page#

Table 1: Composition of Titanium Alloy Ti6Al4V............................................. 27

Table 2: CoromillⓇ Plura specifications ............................................................... 27

Table 3: Levels for input Parameters/factors....................................................... 31

Table 4:“Box Behnken design” ............................................................................ 31

Table 5: Experimental runs by applying BBD .................................................... 32

Table 6: Comparison between surface finish parameters [17] ............................. 43

Table 7: Ra values for each experimental run ...................................................... 45

Table 8: Analysis of variance (ANOVA) table .................................................... 56

x
TABLE OF CONTENTS

ABSTRACT ......................................................................................................... V

LIST OF ABBREVIATIONS ........................................................................... VI

LIST OF FIGURES........................................................................................ VIII

LIST OF TABLES............................................................................................... X

CHAPTER 1. INTRODUCTION ..................................................................... 1

1.1. Introduction ................................................................................................. 1

1.1.1. Aim ............................................................................................... 3
1.1.2. Objectives ..................................................................................... 3
1.1.3. Steps for achieving our objectives ............................................... 4
1.1.4. Scope, Assumptions and Limitations ........................................... 4
1.2. Brief Description......................................................................................... 5
1.3. CAD Modelling .......................................................................................... 5
1.4. Material Used .............................................................................................. 6
1.5. Machining ................................................................................................... 6
1.5.1. Multi-axis Machines..................................................................... 6
1.6. Toolpath Generation & CNC Programming ............................................... 7
1.6.1. Toolpath Strategies....................................................................... 9
1.6.2. Feasibility Check ........................................................................ 10
1.7. Milling with Ball End Mill ....................................................................... 11
1.8. Analysis and Optimization........................................................................ 12

CHAPTER 2. LITERATURE REVIEW ....................................................... 13

2.1. Difficult-to-Cut Materials ......................................................................... 13

2.2. Titanium .................................................................................................... 13

xi
 2.2.1. Morphology ................................................................................ 14
2.2.2. Tools ........................................................................................... 14
2.2.3. Machining Temperature ............................................................. 14
2.2.4. Mechanical properties ................................................................ 15
2.2.5. Thermal behaviour ..................................................................... 15
2.2.6. Undercut and Springback ........................................................... 16
2.2.7. Distortion.................................................................................... 17
2.2.8. Conclusion.................................................................................. 18
2.3. Optimization Techniques .......................................................................... 18
2.3.1. Design of Experiments ............................................................... 18
2.3.2. Response Surface Method (RSM).............................................. 20
2.4. Finding of Research Area ......................................................................... 23

CHAPTER 3. RESEARCH METHODOLOGY .......................................... 24

3.1. Introduction ............................................................................................... 24

3.2. Research Methodology ............................................................................. 25
3.3. Finding of Input Parameters (Numerical Values) ..................................... 27
3.3.1. 3D- Model of Turbine Blade ...................................................... 27
3.3.2. Material Selection ...................................................................... 27
3.3.3. Cutting tool Selection ................................................................. 27
3.3.4. Spindle Speed (N) ...................................................................... 28
3.3.5. Axial Depth of Cut (ap).............................................................. 28
3.3.6. Feed rate ( f ) .............................................................................. 29
3.3.7. Step over (ae) and Scallop Height (h) ........................................ 29
3.4. Response Surface Method (RSM) ............................................................ 30
3.4.1. Box-Behnken Design (BBD) ..................................................... 30
3.4.2. Application of Box- Behnken Design ........................................ 31

CHAPTER 4. EXPERIMENTATION........................................................... 33

4.1. Introduction ............................................................................................... 33

4.2. Selection of Complex Contour Job for Experimentation.......................... 33

xii
4.3. Toolpath Generation & G-code Program.................................................. 34
4.4. CNC Machining ........................................................................................ 35
4.4.1. Rough Machining ....................................................................... 36
4.4.2. Finish Machining ....................................................................... 36
4.5. Summary ................................................................................................... 38

CHAPTER 5. SURFACE QUALITY-A RESPONSE VARIABLE ............ 39

5.1. Introduction ............................................................................................... 39

5.2. Surface Quality ......................................................................................... 39
5.3. Surface Finish & Surface Texture ............................................................ 39
5.4. Surface quality/Surface finish parameters ................................................ 40
5.4.1. Roughness average (Ra) ............................................................. 40
5.4.2. Root-mean-square (rms) roughness (Rq) ................................... 41
5.4.3. Rz ............................................................................................... 41
5.4.4. Max peak-to-valley roughness height ( Rt or Ry or Rmax) ....... 42
5.4.5. Roughness values for various manufacturing processes ............ 42
5.5. Surface Topography .................................................................................. 43
5.6. Morphology of test samples ...................................................................... 44
5.6.1. Roughness Average by Surfcorder............................................. 44
5.6.2. Surface Topography with SEM .................................................. 47
5.7. Summary ................................................................................................... 51

CHAPTER 6. RESULTS AND DISCUSSION.............................................. 52

6.1. Introduction ............................................................................................... 52

6.2. Analysing Response Surface Design ........................................................ 52
6.2.1. Analysis of Variance (ANOVA) ................................................ 55
6.2.2. Contour Plots.............................................................................. 56
6.2.3. Regression model ....................................................................... 57
6.2.4. Analysis of regression model ..................................................... 58
6.2.5. Optimized parameters and verification ...................................... 58
6.2.6. Summary .................................................................................... 60

xiii
CHAPTER 7. CONCLUSIONS AND RECOMMENDATIONS ................ 60

7.1. Conclusions ............................................................................................... 60

7.2. Recommendations ..................................................................................... 61

REFERENCES ................................................................................................... 63

APPENDIX-I ...................................................................................................... 65

APPENDIX II ..................................................................................................... 69

xiv
 CHAPTER 1.
INTRODUCTION
1.1. Introduction

In the modern engineering marvels such as gas turbines advance metal alloys
are being used that face very high temperatures, stresses, corrosive atmosphere
and a potential environment of high vibration. So in these harsh conditions
surface imperfections, machining damages and microcracks can form and
propagate easily and cause catastrophic failure. This can cause damage costing
in millions of dollars, to prevent this from happening regular maintenance is
required. Turbine blades are the most critical parts of a gas turbine. If these
turbine blades are manufactured up to the given standards, then the time
between inspections and maintenance can be reduced drastically.

Previously manufacturing techniques such as die casting, single crystal growth

casting and die forging were used to make these parts. Nowadays multi-axis
machine tools with the application of CAD/CAM software have made possible
to meet the technical developments that these components demand. Through
this project, we have devise the optimum machining parameters values to cater
for roughness and imperfection on the surface and the morphology of the turbine
blades by optimizing the manufacturing parameters, so that surface roughness
can be reduced.

The surface quality of turbine blades plays a vital role in their performance.
Milled surface with excellent surface topography tend to improve fatigue
strength, creep, crack resistance, corrosion resistance etc. Surface roughness can
also effect numerous qualities of these components, such as factors which
causes surface friction, light reflection, heat dissipation capability, lubrication
holding, coatings, fatigue resistance etc. So excellent surface finish is typically
required and suitable procedures are developed to attain the favourable surface
properties.

1
It is a huge challenge to achieve the optimal surface quality of contour shapes
through toolpath generation and CNC machining, which demands to carry
research work in this area. The parameters which affect the surface quality can
be divided into the following categories;

• Machining parameters
• Cutting tool parameters
• Workpiece material properties
• Cutting tool properties

The further detail of the related parameters of the above-mentioned categories

is mentioned in Figure 1.

Figure 1: Fishbone diagram showing effects of machining parameters on surface

roughness [1].

In this project, we have selected titanium alloy Ti6Al4V as the material to

manufacture the turbine compressor blades. Titanium alloys are widely used in
turbine components such as propellers, aerofoils, structures, fuel systems,
hydraulic systems, nacelles and turbine housings. Ti6Al4V alloy accounts for
about half of all alloys used in aerospace applications. Its wide spread usage is
owed to titanium’s high tensile strength to weight ratio, excellent resistance to
corrosion, and good creep strength. Other titanium alloys are used in airplanes,
weaponry, armour plates, marine vessels, spacecraft, rockets and missiles. The
blades are machined on a 4-axis CNC vertical machining centre. The

2
appropriate tools for roughing and finishing are selected. Then using CAM
software MasterCAM toolpaths are generated and simulated. An initial set of
values for the following machining parameters; depth of cut, feed rate, cutting
speed are selected and coolants are also selected based on the tool manufacturers
recommendations. After studying the effects of these parameters on the
experimental samples topology and morphology we have generated optimized
values of the machining parameters and mathematical model through the design
of experiment using Box-Benken design to generate a set of values of the
selected parameters for the final part that can be put into production. The main
objective in the development of these methods is to optimize a procedure within
the existing bounds.

1.1.1. Aim

Complex contour jobs such as turbine blades, impellers, hydraulic system

components, nacelles, propellers and helical rotors require high surface quality
due to their vital performance standards in aerospace, automobile and other such
fields. With separate grinding or polishing operations it is difficult to achieve
high surface finish because of complex profiles and geometrical features of such
parts. Manual grinding operations can harm and deteriorate complex profiles,
geometrical features and tolerances of these parts. Machining of such parts with
high surface quality through milling operations is the main problem because it
is hard to achieve high surface finish through milling operations as compared to
other metal cutting processes. It is desired to achieve optimal surface quality
with milling operations in order to eliminate additional polishing procedures.

1.1.2. Objectives

The core focus of this venture is to accomplish high surface quality through
input machining parameters applied during toolpath generation and CNC
machining. It is required to get surface quality which lies in the
grinding/polishing region through milling operations and also to eliminate the
need for additional grinding/polishing operations. It will enhance the surface
quality and accuracy of complex profiles, tolerances accuracy and geometrical

3
features. On the basis of experimentation and findings, it is required to develop
the mathematical model to apply further for complex contour jobs in ordered to
find the surface quality with same conditions and parameters applied for this
research work. It will be much helpful to manufacture such high functional parts
with an optimal surface finish through this research work.

1.1.3. Steps for achieving our objectives

Following sequential steps have been taken to achieve the scope of this project:

• Study of the mechanical properties of the selected metallic materials like

elastic modulus, fracture toughness, poisons’ ratio, hardness, etc.
• Study of the mechanics of machining process in detail.
• Study of machining parameters like feed rate, cutting depth, cutting
width, cutting speed and their effects on the machining process.
• Study of the cutting tool parameters
• Modelling of a turbine blade using CAD software.
• Preparing Response Surface design using Minitab 16.
• Perform experiments according to the vales generated by the Response
Surface Method.
• CNC toolpath generation using CAM software such as MasterCAM.
• Experimental machining with Ti6Al4V to validate the developed model
using a 4-axis CNC machining centre.
• Measurement of surface roughness measurement using a Surfcorder.
• Inspection of surface topography through a scanning electron micro
scope (SEM).
• Machining of final part for validation of results.

1.1.4. Scope, Assumptions and Limitations

1. Machining parameters that are spindle speed, axial depth of cut and feed rate
were applied as variable parameters while a number of other parameters applied
as constants such as workpiece material, end mill diameter, the cutter material,
and other machining parameters. It is the limitation of this research work and it
can be broad by considering other parameters as variables.

4
2. The CNC machining centre used for this research work has 03-axis
simultaneous movements and 01-axis as positioning axes. The results can be
improved with 05-axis simultaneous machining centre with better positioning
accuracy and spindle run out.

3. The analysis can be performed by applying the functional / application

parameters of such jobs (or on these samples) to find results of the performance
and surface quality of these jobs (or samples).

4. The research results can be improved by applying replications with

experiments and to eliminate the experiments with greater residuals for further
analysis and findings.

1.2. Brief Description

The project is based on two parts. First is the study of the materials and selecting
appropriate machining parameters for initial machining and studying their
effects on the turbine blade part and the second is to analyse our findings and
development of process for manufacturing the turbine blade by optimizing the
machining parameters to get the best surface finish and dimensional accuracies.

1.3. CAD Modelling

The Turbine blade has been modelled in Solid Works 2017. It has a profile
and contour shape of an aerofoil as shown in Figure 2.

Figure 2: Turbine blade

The profile coordinates and dimensions are given in appendix-1

5
1.4. Material Used

Titanium Alloy Ti6Al4V is used to manufacture the turbine blades which was
provided by NESCOM.

1.5. Machining

The parameters which are optimized while machining the part are feed rate,
axial depth of cut and spindle speed. Then their effect on the machining process
are studied.

1.5.1. Multi-axis Machines

Normal CNC machining centres have limitations and can be used for machining
up to 03-axis. For multi-axis machining, CNC machining centres with rotary
axes (A, B & C-axis) are required to get movements more than 3-degrees of
freedom. The addition of one or more rotary axis increase the capabilities of
CNC machining centres to handle the machining of contoured profiles with
better quality and high surface finish. Part holding is also improved as it
eliminated the need for jigs and fixtures. Normally machine builders apply A-
axis to get rotation about X-axis, B-axis to get rotation about Y-axis and C-axis
to get rotation about Z-axis [2]. With a multi-axis machine instead of the typical
XYZ axes, we can have 6 or even more axes drawn in Figure 3 [3].

Figure 3: Multi-axis machine axis designation.[3]

6
A machining centre can be characterised by the location of its revolving axes
as:

• The 2-0 machine that has two rotary axes on the bed shown in Figure 4(a)
• The 1-1 machine has one rotary axis on the bed and one on the tool head
shown in Figure 4(b)
• The 0-2 machine has two rotary axes on the tool head shown in Figure 4(c)

It has been found that with 05-axis simultaneous machining centre (X, Y, Z and
A, B or C), movements in all directions can be achieved and there is no need of
the 6th degree of freedom. Because if 05-degrees of freedom are incorporated
in one system then all movements can be achieved without applying one more
degree of freedom [3].

Figure 4: Configuration of CNC machining centre with a rotary axis[2]

1.6. Toolpath Generation & CNC Programming

The CAM software typically produces a set of cutting tip location points ( CL-
point), which are then converted into G & M-codes to represent the machine
motion where the tooltip coincides with the work piece. The CL-point is
positioned on the centreline of the cutter along the surface as shown in Figure
5.

7
 Figure 5: CC point and CL point identification

The generation of CNC-program for multi-axis milling machines generally

follows these two steps; In the first step, the CAM software computes the path
of the milling cutter and a toolpath is generation. Each location of the tool path
is described by its tooltip position in x, y, z coordinates and tool angle in i, j, k
coordinates. Both are converted in a workpiece coordinate system (WCS). In
the next step, the toolpath is generated as a “machine controller independent
cutter location DATA file” (CLDATA), is then changed by an CNC-
postprocessor to the machine specific CNC program i.e. G-code. A vital part of
multi-axis post-processor is the conversion from CLDATA to machine
coordinates [4]. The generation of toolpath and CNC program can be shown
diagrammatically as Figure 6

Figure 6: Generation of CNC programs (standard method) [4]

8
 1.6.1. Toolpath Strategies

Toolpath strategies can be divided into two broad categories such as toolpath
distribution strategies and toolpath calculation strategies.

1.6.1.1 Toolpath Distribution Strategies

The application of these strategies is to machine the entire component. These

distribute the tool path in the dominion of the machined component which is
available with CAD/CAM commercial software. The commonly used toolpath
distribution strategies are;

1. Zigzag or raster curves

2. Contour curves
3. Space-filling curves
4. Sequentially generated curves
1.6.1.2 Toolpath Calculation Strategies

Tool path calculation is the method by which the trajectories of the tool are
constrained to lie on the designed part. Following tool path calculation strategies
are applied for multi-axis machining;

1. Iso-curve approach
2. Iso-planar approach
3. Iso-scallop height approach
In (1) and (2), scallop height varies along the toolpath. Scallop of ball end mill
is displayed in Figure 7. But the maximum value is less than the allowable value
of this factor. So some portions of the part may have a very smooth surface
finish. Iso-scallop height approach generates the toolpath by maintaining the
constant scallop height between each pass on the machined component.
Toolpath generation for complex contour jobs with multi-axis machining is the
main issue for finish machining.

9
 Figure 7: Scallop height of ball end mill

1.6.2. Feasibility Check

Feasibility check is required to find interferences between various bodies

involved in machining. For this purpose, it is required to check any collision,
gouging, and machine stroke-over at cutter contact point (CC-point) for a given
tool orientation. Collision or crash is checked between part surface, tool shank,
tool holder and machine components. The simulation available with CAD/CAM
software is applied to check the collision, gouging and machine stroke over.
Also, the collision avoidance option available with good CAD/CAM software,
is applied to modify the tool location by guarantee the CC-point on the desired
position. The verification of toolpath generated through simulation has shown
in Figure 8.

Figure 8: Simulation for toolpath verification

10
1.7. Milling with Ball End Mill

It is the traditional approach in which ball end mills with either cylindrical or
conical shapes are used. In this method of milling, a curved surface is machined
by the round end of an end mill. The major advantage of ball end milling/point
milling is that almost any type of surface can be point milled. During machining,
the cutting point on the spherical surface changes, which changes the cutting
speed. When the CC-point is at the tip of the tool, the cutting speed nearly equal
to zero which produces a rough surface. So this can be controlled with the
application of tilt angle and yaw angle (also known as lead or lag angle).

Ball nose milling also produces a scalloped surface. The height of the scallop is
directly related to the cutting edge radius and the radial depth of cut or step over.
This problem can be overcome with the application of a constant scallop height
toolpath calculation strategy to optimize the surface quality of complex contour
jobs. Application of perfect control of parameters and scallops has been made
possible to get the optimal surface quality of such complex contour jobs. It is
achievable with the use of a CAM software for toolpath generation and CNC
programming by using various options available within the software [5].

Ball nose end mills are perfect for machining 3-D contoured shapes. These are
usually found in the mold and die making, machining of turbine blades and
general part finishing. To correctly use a ball end mill (with no tilt angle) and
gain the best tool life and part finish. Effective cutting diameter (Deff), shown
in Figure 9, must be used instead of cutter diameter when using a ball nose end
mill that is using axial depth of cut lesser than the full radius of the end mill.
This can calculated using tool diameters and axial depth of cut, also the new
cutting speeds can be calculated[6].

11
 Figure 9: Cutting path view of Ball end mill[6]

1.8. Analysis and Optimization

Once the toolpath generation is complete, the next step is to determine the
significant machining parameters and there working range/levels for which the
optimum value is to be found for the least surface imperfections and maximum
surface finish. Response surface method and Box-Behnken design are used to
study a set of 15 experimental samples with values of the machining parameters
varying from min to max, for each run the surface roughness is measured using
Surfcorder and analysed through Minitab to get the contour plots and the
optimum value for every parameter along with their main effect.

Further, the samples are studied in an SEM to verify the findings and a final test
piece is machined according to the obtained optimum values for the parameters:
feed rate, depth of cut and cutting speed.

12
 CHAPTER 2.
LITERATURE REVIEW
2.1. Difficult-to-Cut Materials

There are many complications which occur during machining of titanium and
other difficult-to-cut metals in industries. The cutting phenomenon while
machining of these alloys is intricate and creates challenges, that as saw-tooth
chips, elevated temperatures, high stresses on the cutting edge, increased tool
wear and undercut parts. These complexities are linked to each other [7].

Numerous properties, such as excellent thermal resistance, high tensile strength

and corrosion resistance, of these metals classify them as superalloys.
Aerospace products are generally manufactured by large material removal from
raw stock. Near net shaping, approaches are hard for most of the these
applications due to intricate contoured profiles, and great tolerance and higher
surface finish desires[7].

2.2. Titanium

Titanium alloys are commonly referred as “difficult-to-machine” materials.

Poor thermal conductivity and the low specific heat of these alloys cause high
cutting temperatures while machining. Also, the tendency of titanium alloys to
retain high tensile strength at elevated temperature and work hardening create
higher machining forces[7]. The Titanium alloy, Ti-6Al-4V also known as
Grade-5 titanium is the most versatile and frequently used metal in aerospace
components.

The low thermal conductivity of Ti-6Al-4V causes high temperatures at the

chip-tool interface. Therefore, enhancements in machining titanium alloy hinge
mostly on controlling the difficulties faced with the natural properties of this
alloy. To increase the machinability of titanium alloys techniques such as
vibration analysis kit, high-pressure flood coolant, cryogenic cooling, thermally
enhanced high conductivity cutting tools and rigid tool holders must be used[7].

13
 2.2.1. Morphology

The ability of titanium alloys to change phase makes the cutting process more
difficult. The hexagonal closed-pack (HCP) α and body-centred cubic (BCC) β
phases are crystal structures of titanium[7].

The Ti alloy, especially Ti-6Al-4V alloy is the most striking and frequently used
metal alloy in the aerospace components [7] and is the material used in this
project. The machining for Ti-6Al-4V is unlike machining of metals. The added
difficulties in the machining of this alloy rise due to its response to the cutting
tool [8].

So, strength at high temperatures, low thermal conductivity, low modulus of

elasticity, strain hardening and high chemical reactivity play important roles in
cutting mechanics of these alloys [8].

2.2.2. Tools

Based on the workpiece hardness, various tools are able to machine Ti-6Al-4V.
Though, a lot of the tougher tools currently available are not suitable for this
alloy owing to chemical reactions which is the cause chemical degradation in
the cutting edge. Also, chips stick easily on the cutter to form a build-up edge
(BUE). Consequently, tool degrades while during machining. Carbide tools are
widely used to machine titanium. Lesser use of ceramic tools is seen in the
industry [7].

2.2.3. Machining Temperature

“It depends on the cutting environments and material properties. The machining
temperature increases with the increase in cutting parameters, such as speed and
feed. On the other hand, the machining temperature increases with the decrease
of the intrinsic properties of workpiece materials, such as the thermal
conductivity, the density and the specific heat of a workpiece. Thus a smaller
value of the product of these intrinsic properties of workpiece materials raises

14
the temperature even at a relatively lower machining speed and smaller
feed”[7].

2.2.4. Mechanical properties

The “thermal conductivity for titanium (18.67 W/m °C at 1000 °C) is the cause
of high temperature while machining. Around 80% of the heat produced is
conducted into the tool due to titanium’s low thermal conductivity. Figure 10
shows that the thermal conductivity of titanium decreases with the increase in
cutting pressure and it has the lowest thermal conductivity around the
temperature of 300°C. This shows that the thermal conductivity of titanium
further declines under cutting conditions where high pressure are involved”[7].

Figure 10: Thermal conductivity of titanium at various temperature and pressures [7]

2.2.5. Thermal behaviour

At”a fairly low cutting speeds (60 m/min), the temperature at the tool chip
interface becomes higher than 1000 K and it can reach as high as 1350 K at 200
m/min. These are around 250 K greater than that of carbon steel”[7].”The heat
produced during machining stays mostly in the workpiece due to low thermal

15
conductivity and high heat generation. While machining titanium only the
cutting edge of the tool takes parts in chip formation and excessive temperature
is noticeable at this point. So, thermal stresses develop when machining
temperature is higher due to poor heat dissipation by chips and the
workpiece”[9].”The dispersal of temperature along the tool-chip interface at
different cutting speeds is given in Figure 11”[7].

Thus various types of cutting tool degradation processes, particularly chemical

reactions, are heightened at elevated temperatures. ”These high cutting
temperatures and chemical reactions reduce tool life, surface quality and cutting
accuracy due to thermal expansion of the tool and workpiece materials, and tool
degradation”[7].

Figure 11:“Cutting temperatures at the tool-chip interface when end milling at feed
per tooth of 0.15m/tooth”[7].

2.2.6. Undercut and Springback

Extremely elastic workpiece material has the ability to deflect and return cutting
tool passes throughout machining. This causes undercut of the parts and causes
dimensional inaccuracies. Figure 12 shows the usual effect of spring back
during machining[7].

16
Low”Young’s modulus, reasonably high hardness and yield strength make
titanium alloy highly elastic which can cause excessive workpiece deflection
and small plastic deformation. Thus, it springs back i.e. there is a bouncing
action as the cutting edge enters the cut during machining processes under
cutting pressure”[7].

Figure 12: Effects of spring back on of the cutting edge region[7].

2.2.7. Distortion

The”distortions in machined parts arise due to the relief of the in-built residual
stress as the material is removed layer by layer throughout the machining
process.”The distortion of parts happens very often during machining of
aerospace parts made of titanium are mostly mass-produced from large stock
material by removing vast sums of material to decrease the number of joins in
the components and make them structurally strong. The foremost reason for
residual stress is the troubles in uniformly cooling large forgings. The external
layers solidify to room temperature first. The outside layers shrink and apply
pressure to the interior layers which are hot and soft. These cause residual
stresses are instabilities in large forgings[7].

17
 2.2.8. Conclusion

Due to high strength, low thermal conductivity, low modulus of elasticity, high
strain hardening and high chemical reactivity[7] the machining of titanium alloy
Ti6Al4V is very difficult. There are difficulties such as spring back, BUE,
distortions, thermal instabilities and chip welding and they need to be catered
during tool path generation and CNC machining.

2.3. Optimization Techniques

2.3.1. Design of Experiments

Design of Experiments is” an organized method to acquire one’s objectives in

a structured manner and can replace haphazardly conducted experimentations.
Fisher’s technique was developed as a statistical approach that allowed
understanding of the effect of one variable on another”[10].”The design and
analysis of experiments circle around the understanding of the effects of
parameter variables on response variable”[11].”In statistical terms, the goal is
to establish a cause-and-effect association between a number of independent
variables and dependent variables called “factors”. These factors are sub-
divided into various values called “levels”. Each level of a factor in a single
experiment is known as a “treatment”. A balanced design occurs when an equal
number of response observations are taken for each treatment of the experiment.
“Replicates” are repetitions for a specific treatment. The number of factor levels
under consideration in an experiment governs its number of treatments.”

Since resources are limited, it's important to gain the most info from every
experiment that is performed. ”Well-designed experiments can yield
significantly more data and mostly require less runs than hit-or-miss or
unplanned experiments. Additionally, a well-designed experiment will
guarantee that you can study the effects that you have recognised as vital”[12].

18
 1.6.1.3 Planning

Cautious planning can aid in avoiding difficulties that can arise during the
execution of the experimental plan. The research necessary in advance of
experimentation is determined by your problem.

• Defining the problem.

• Defining the objective.
• Development of experimental plan that provides significant
information.
• Affirmation that the process and measurements are in control.

1.6.1.4 Screening

In many procedure progress and industrial use, possibly significant varying are
plentiful. Screening decreases the quantity of varying by classifying the
fundamental variables that disturb a products quality. This reduction lets you to
focus on process improvement on the most significant variable, or the
significant factors. The resulting procedures are frequently used for screening:

• Placket-Burman design
• Two-level full factorial design
• Fractional factorial design
• General full factorial designs

1.6.1.5 Optimization

After the identification of the significant factors by screening, optimum

standards for these experimental causes are required. Optimum factor standards
are subject to the process goals. Researchers have adopted many different
techniques and methods to optimize the process such as Engineering Design
Optimization. To attain the best grouping of parameters it is essential to trial a
large number of combinations, which is unfeasible for industries. Optimization
methods are an fascinating answer to reduce the number of combinations of

19
experimental tests. In the last century, many optimization methods have been
established

• Factorial Designs
o Full Factorial Design
o Partial Factorial Design
• Response Surface Designs
o Central Composite Design
o Box- Behnken designs.
• Mixture Designs
• Response Optimization
• Taguchi Designs

1.6.1.6 Design Variables

The variables under consideration are also called design variables. These
include;

• Controlled Variables
• Uncontrolled but observed Variables
• Uncontrolled and unobserved Variables

1.6.1.7 Constraints

Constraints are the bounds given to variables. Constraints help to assure a

parameter or variable stay within the upper and lower limits.

2.3.2. Response Surface Method (RSM)

An important approach in the design of experiments is the response surface

method (RSM). It is initiated that RSM is the best-suited method to apply for
input constraints such as spindle speed (N), axial depth of cut (ap) and feed
rate(f) in order to get response variable in optimize manner/condition. This
method is the grouping of mathematical and statistical techniques used in the
practical study of interactions and optimization, where numerous independent

20
variables influence on dependent variable or response. The connection between
response variables and input variables is determined by approximated
polynomial functions, whilst the model parameters are obtained by a small
number of experimentations consuming the DOE [11].

B. C. Routara et al [1] applied RSM for optimization of surface roughness

constraints such as average roughness, root mean square roughness, skewness,
kurtosis and mean line peak arrangement for materials such as aluminium,
brass, and mild steel. He considered spindle speed, depth of cut and feed rate as
an input variable and applied the full factorial method to find the experimental
runs for these three types of materials separately. He fitted second-order
response surface calculations by consuming Minitab for all three response
variables. He also found that micrographs of surface texture with a scanning
electron microscope (SEM) for optimal conditions of machining had reduced
the surface roughness as compared to un-machined micrographs.

K. Kadirgama et al [13] explained that Box-Behnken design is usually used

when executing non-sequential experimentations and accomplishment the
experiments only once. These designs provide an effective approximation of the
first and second-order constants. Because it has fewer design facts, they are less
affluent to run than central composite designs with the equivalent number of
factors. Box-Behnken design does not have axial points, and it provides surety
that all design points fall within the safe operating zone. He applied this design
to optimize surface roughness by a run of experiments with a different
combination of input variables such as cutting speed, feed rate, axial depth and
radial depth. He found that the optimum value of surface roughness is 0.4261μm
for Al-alloy (AA6061-T6) with an end mill and 90º tool holder. The
corresponding design variables were cutting speed 100 m/min, feed per tooth
0.2 mm/rev, axial depth 0.1mm and radial depth 5.0mm. First order linear
relation was found between surface roughness and input parameters/ factors.

Basim A. Khidhir et al [14] used Box Behnken design of response surface

method to find the sufficient model that related the cutting force and cutting

21
parameters (cutting speed, feed rate, depth of cut and approach angle). He
conducted 27 experiments to turn Nickel based Hastelloy C-276 with ceramic
inserts. Minitab was applied for analysis and regression. He found that linear
with the interaction model between the response variable (cutting force) and
design variables. He concluded that the cutting force is affected knowingly by
the feed rate charted by the depth of cut and then by feed rate with the depth of
cut, and lastly by the cutting speed and method angle. Contour plots were drawn
to find the relationship between input and response variables.

M. Alauddin et al [15] applied central composite design of RSM and then used
regression analysis to find the model of tool life by applying input variables
such as cutting speed, feed rate and axial depth of cut. He conducted 24
experiments after finding the values of input parameters for each run. The
material was used AISI 1020 for vertical milling with cutter Ø 20mm of material
Cobalt-alloyed HSS having 02-flutes. He found that the first order model is only
suitable over a narrow range of the variables. He concluded that the increase in
the speed, feed, and axial depth of cut decrease the tool life and surface
procedure provides a large amount of information with a small amount of
research.

H. Okten et al [16] applied the response surface methodology and genetic

algorithm to optimize the surface roughness. He applied three-level full factorial
design to find the experimental runs by taking input parameters such as feed
rate, cutting speed, axial depth of cut, radial depth of cut and machining
tolerance. Mold material such as aluminium (7075-T) was used for milling
operation with solid carbide Ø10mm end mill PVD ALTiN coated having
04xflutes. The developed response surface model was joined with a established
heritable algorithm. He found that surface roughness of the mold surfaces was
0.412μm before optimization which was condensed to 0.375 μm after
optimization.

22
2.4. Finding of Research Area

From the above literature review, it is found that there is a need to optimize the
surface quality (surface finish & surface texture) for jobs which required multi-
axis machine tool capabilities. Such jobs include turbine blades, propellers,
impellers, helical rotors, etc which are required high performance. Machining
parameters such as cutting speed/ spindle speed, tilt angle, depth of cut, radial
depth of cut, feed rate/feed per tooth, the direction of movement, type of milling
(up milling/down milling) etc. have important effects on surface quality of such
complex contour jobs. In the above literature review, the effect of these
parameters was discussed in face milling or packet/cavity milling as also
discussed in the latest research papers. Moreover, the materials used for
experimentation in these research papers were found aluminium, brass normal
and hardened steels. On the other hand, the complex profile job such as
turbine/compressor blades was manufactured with resources such as Titanium
alloys (aerospace material). It is hereby found that there is required to consider
input parameters such as spindle speed, axial depth of cut and feed rate in
ordered to optimize surface quality (response variable) for such complex profile
jobs. Box Behnken design of RSM should be functional to find the experimental
runs with the application of D-optimality.

In this way it will be a unique research area and not much research work was
carried regarding this area. After this research work, there will be
findings/results related to the most complicated area of manufacturing
technology to get optimize surface finish/surface quality. It is the most dominant
parameter for better performance and functionality of such complex contour
jobs.

23
 CHAPTER 3.
RESEARCH METHODOLOGY
3.1. Introduction

Machining parameters play a vital role in all type of machining procedures such
as milling, turning, grinding, etc. The numerical values of these parameters are
obtained after consulting tables, catalogues and applying certain formulae
keeping in view the workpiece material, tool material and capabilities of
machine tools. Accurate values of these parameters are significant in case of
CNC machining especially when it is required to manufacture of complex
contour parts with multi-axis machining applications. Also, the accurate
selected values of these parameters provide a guaranty to manufacture complex
jobs with the desired quality. The main focus of this project is to study the effect
of machining parameters on surface quality/surface finish. These machining
parameters include mainly spindle rpm, feed rate, feed per tooth, axial depth of
cut, radial depth of cut (step over), and process kinematics.

In the case of multi-axis machining of complex contour jobs, the direction of

feed rate and tool path strategy, type of milling have also influence on surface
quality. From the above-mentioned machining parameters, the most significant
parameters which have a dominant effect on surface quality/surface finish
(minimum surface roughness) and selected for this research work are as under;

1. Spindle speed (N) in rev/min

2. Feed rate (fz) in mm/min
3. Axial depth of cut ( ap ) in (mm)

For surface quality, the response variable selected was Ra (roughness average)
and surface topography was applied for further validation of results/findings as
it is desired to find the effects of input parameters on the surface finish of
complex contour jobs.

24
3.2. Research Methodology

The research methodology is the complete briefing of the procedure through

which the research work will be completed. For this purpose, the 3D model of
the turbine blade has been selected as a complex contour job because of its
varying profiles and complex shape. The material of this blade is selected as
Titanium alloy Ti6Al4V which is used normally in the manufacturing of such
blades.” After finding the numerical values of above-mentioned parameters,
Box-Behnken design (BBD) of response surface method”is applied by using
Minitab 16 software in order to get the values of these parameters for different
experimental runs. Then CAD/CAM software, MasterCAM is applied to
generate toolpath and CNC program for each experimental run by considering
other parameters constant except spindle speed, axial depth of cut and feed rate.
These CNC programs will be executed and samples will be obtained after
machining with CNC machining centre.

Since the purpose of our research work is to optimize the surface quality/surface
finish. Then each sample will be inspected by using Surfcorder to find average
roughness (Ra) and the surface topography will be done with a scanning
electron microscope (SEM) to get micrographs for further analysis. Surface
topography will provide the authenticity of Ra values which will be discussed
next in details. Roughness average is considered as response/output parameter.
The analysis is done using Minitab and the regression model is determined.
After analysing the regression model, the results are found. The research
methodology for this project is described in Figure 13. The complete briefing
of this methodology will be elaborated in the upcoming research work.

25
Figure 13: Flow chart for the methodology of research work

26
3.3. Finding of Input Parameters (Numerical Values)
3.3.1. 3D- Model of Turbine Blade

3D model of the turbine blade was used as a complex contour job for this
research work. It has varying profiles and can be manufactured though multi-
axis machining which is the focus of this research.

3.3.2. Material Selection

Titanium alloys are usually used for manufacturing of turbine blades due to its
heat resistance properties with its low density, high strength and excellent
resistance to corrosion. Its uses are most significant in turbines and airframe
components that are subject to a temperature up to 600 °C. Its high corrosion
resistance is also a valuable property. The titanium alloy (Ti6Al4V annealed
state) was used for this research work which has the chemical composition as
shown in Table 1

Table 1: Composition of Titanium Alloy Ti6Al4V

3.3.3. Cutting tool Selection

The cutter used for this research work was solid carbide CoromillⓇ Plura ball
nose of Ø 6mm (Sandvik Coromant) with PVD AlCrN coating. The end mill is
shown in Figure 14 and the specifications of this cutter are shown in Table 2.

No of flutes Flute length Helical angle Overall length Max rpm

2 10 mm 30 deg 57 mm 80000

Table 2: CoromillⓇ Plura specifications

27
 Figure 14: “CoromillⓇ Plura solid carbide ball nose end mill”

3.3.4. Spindle Speed (N)

By using the catalogue of Sandvik Coromant the cutting speed, Vc in case of

solid carbide ball nose cutter for Titanium alloys machining is 75-150m/min for
finish machining. The cutting speed was selected as m/min for further
calculations of Spindle speed (rev/min).

N = v/∏D

n = vc × 1000 / ∏ × D

Where,

n= Spindle speed

Vc=cutting speed

D=cutting depth

3.3.5. Axial Depth of Cut (ap)

Sandvik Coromant catalogue was used to find the range of axial depth of cut
(also known as axial engagement) for Titanium which varies from 0.05mm to
0.25mm. This can be written as Material Removal Rate (MRR) i.e:

Q = ap × vf / 1000(cm^3 / min)

28
Where,

Q= Material Removal Rate

vf= Feed speed

ap= depth of cut

3.3.6. Feed rate ( f )

Again applied Sandvik Coromant technical data catalogue and found that the
value of feed rate for Titanium material ranging from 200mm to 400mm.

fr = Nntf

Where,

fr= feed rate mm/min

N= Spindle Speed

nt= number of teeth

f= chip load

3.3.7. Step over (ae) and Scallop Height (h)

The step over and scallop height are related parameters. Scallop height was
considered as constant and kept equal to 0.001mm for applying constant scallop
height toolpath (toolpath calculation strategy). It is found that during the
toolpath generation, if the radial depth of cut kept constant with ae = 0.17mm.
Then iso-scallop height (constant scallop height) toolpath will be generated.
This tool path is the best for optimal surface quality. From above findings, the
input parameters are summarized as in Table 3. Step over can be calculated by
the size of the step-over distance and the tool diameter will regulate the scallop
height.

29
Where,

s = Step over value between two cutting passes, in inches

d = Diameter of milling cutter, in inches

3.4. Response Surface Method (RSM)

It is the method of experimental design which is applied to find the association

between different response variables and a set of measureable experimental
variables or features. RSM is essentially a group of mathematical and statistical
method that is valuable for the modelling and investigation of glitches in which
a reply of importance is influenced by numerous variables and the aims are to
optimize the response [11]. It is the most suitable method according to this
research work because it is required to find the optimal surface quality by
considering the effect of three factors as spindle speed, axial depth of cut and
feed rate on surface roughness (response variable). It is also economical to get
results with fewer runs. This method has the ability to further reduce the
experimental runs for optimal results. RSM has further two approaches such as
Box-Behnken design and Central composite design.

3.4.1. Box-Behnken Design (BBD)

This design is applied for performing non-sequential experiments or when we

are planning to perform an experiment once.”It provides an efficient estimation
of the first and second order coefficient. Box Behnken Designs have fewer
design points so that they are less expensive to run than central composite
designs with the same number of factors. These designs can also prove useful if

30
one knows the safe operating zone for the process”[12]. “These designs do not
have axial points, thus one can be sure that all design points fall within a safe
operating zone. These designs also ensure that all factors are never set at their
high levels simultaneously”[1]. From the above discussion, it is found that Box
Behnken design is the most appropriate for the parameters of this project.

3.4.2. Application of Box- Behnken Design

Minitab16 is a commercial statistical analysis software. Here this software was

applied for BBD approach to finding experimental runs of input parameters.
The focus was to find the optimal value of response parameters after performing
experiments according to the runs obtained through this method. The input
parameters were re-arranged for applying Minitab 16 and were recorded as
shown in Table 3

Table 3: Levels for input Parameters/factors

The details of the selection of different available options are shown in Table 4
as under;

Table 4:“Box Behnken design”

31
 Figure 15: The 3N full factorial for 03 factors in BBD

The above Figure 15 shows the data points for the BBD that are used as levels
to determine the experiment run order for the design as shown in Table 5 below
that will be followed and the response variable i.e. average surface roughness is
measured and put into the table and analysed.

“Spindle “Depth
Std Run Pt “Feed Rate “Roughness”
Blocks Speed of Cut
Order Order Type F(mm/min)” Ra(um)
N(rpm)” ap(mm)”
6 1 2 1 8000 300 0.05 1.435

11 2 2 1 6000 200 0.25 0.9944

15 3 0 1 6000 300 0.15 0.6172

1 4 2 1 4000 200 0.15 1.153

14 5 0 1 6000 300 0.15 0.63

3 6 2 1 4000 400 0.15 1.0385

2 7 2 1 8000 200 0.15 1.794

13 8 0 1 6000 300 0.15 0.6531

10 9 2 1 6000 400 0.05 0.6323

7 10 2 1 4000 300 0.25 0.7976

5 11 2 1 4000 300 0.05 0.5644

8 12 2 1 8000 300 0.25 0.3915

12 13 2 1 6000 400 0.25 0.3161

4 14 2 1 8000 400 0.15 0.5163

9 15 2 1 6000 200 0.05 0.7778

Table 5: Experimental runs by applying BBD

32
 CHAPTER 4.
EXPERIMENTATION
4.1. Introduction

The manufacturing of contoured profiles is only possible through multi-axis

CNC machining centres with application of CAD/CAM softwares. For such
parts, CNC program, i.e. G-code is generated after toolpath generation and
simulation. The simulation is very useful for checking collisions either of the
cutter, tool holder, spindle, chuck/clamp or workpiece. It also helps to find if
gouging of the cutter or any other fault occurs during machining. Various
cutting constraints such as radial depth of cut, feed rate, feed per tooth, spindle
speed, tool path strategy, type of machining, etc. are applied during the toolpath
generation.

After verification of toolpath simulation, G-code is then generated by the

appropriate post-processor. This program is then copied to CNC controller for
actual machining.

4.2. Selection of Complex Contour Job for Experimentation

A turbine blade has been chosen as the part to be manufactured for this project.
3D model of turbine blade was cut down for experimentation. This was done
because the material cost is very expensive and the machining process is time
consuming. The cut down model and the model used for experimentation has
revealed in Figure 16 and Figure 17.

Figure 16: Iso- metric view of cut down model

33
 Figure 17: Side and top views of model used for experimentation

This model in Figure 16 was selected to accomplish investigational machining

in order to find out the results for surface roughness corresponding to the
machining parameters for manufacturing the turbine blade. MasterCAM X5 was
used to generate the toolpath for each set of parameters as shown in Table 5.
After verification through tool path simulation, G-code was generated by using
the post-processor compatible with the CNC machining centre (DMG Mori
DMC 1035V Eco line), for each set of experimental run.

4.3. Toolpath Generation & G-code Program

Carbide ball nose end mill of Ø 6mm (kept constant for 15 runs) was selected
for finishing. Values of the parameters applied to generate tool path for 15 runs
are as under;

• Feed rate, f (mm/min) = Variable for each run

• Spindle speed = Variable for each run
• Radial depth of cut, ae = 0.17 mm (Kept constant for 15 runs)
• Scallop height, h = 0.001 mm (Kept constant for 15 runs)
• Tolerance = 0.001 mm (Kept constant for 15 runs)
• Zig zag tool path (Kept constant for 15 runs)
• Down milling (Kept constant for 15 runs)

34
 Figure 18: Operation manager tab on MasterCAM

This post-processor converts toolpaths according to the machine configuration

and specifications and changes them into numerical values of spindle
movements as G-code. All the various constraints such as axial depth of cut,
spindle speed and feed rate, were controlled with G-code on the CNC machine.
The operations management tab of MasterCAM X5 is shown in Figure 18.

Cylindrical blank of Titanium alloy with dimension r =45mm, h=90mm was

used for this experimental machining. 10 experiments per blank were conducted
to save material and reduce machining time

4.4. CNC Machining

After generation of CNC program, the G-code is then transferred to CNC

machine controller via USB memory stick, and copied in the CNC controller.

35
DMG Mori VMC 1035V Eco line, CNC vertical milling mechanism was used
for machining. Detailed specifications of this machining centre is shown in
Appendix II.

4.4.1. Rough Machining

For rough machining, ball nose end mill of Ø12mm, solid carbide & 2x flutes
was recycled and allowance of 1mm was left for finishing pass. Flood coolant
was used for rough machining.

4.4.2. Finish Machining

The parameters were set according to actual experimentation as shown in Table

3. For this purpose, ball nose end mill of Ø 6mm, solid carbide & 2xflutes was
used. The other specification of this cutter has shown in Table 2. Flood coolant
was applied for final machining. The axial depth of cut (ap) was set with CNC
controller of the milling machine as per the value in Table 3 for each
experimental run. For saving time and cost 4-axis milling machine was used. It
means that 3-axis simultaneous movement and 1-axis for positioning were
applied to get samples for each set of parameters. In this way 10 and 6 samples
were obtained after each run. The experimental samples being machined are
shown in Figure 19. The 10 samples after machining are shown in Figure 20.

Figure 19: Experimental sample machining on CNC machine

36
 Figure 20: Samples after machining

The samples were parted off into sets of 2 for surface roughness measurement
and SEM analysis and are shown in Figure 21.

Figure 21: Experimental samples for roughness measurement and SEM analysis

37
4.5. Summary

For tool path generation, scallop height was reserved continuous at 0.001mm
and then the radial depth of cut (ae) was set at 0.017 after calculations. The
tolerance was taken to be 0.001mm during tool-path generation. The toolpaths
were generated for 15 runs separately for rough and finish machining. G-code
was generated first for rough machining and then finish machining was
performed with separate G-code program for each run. The samples were
obtained after performing CNC machining.

38
 CHAPTER 5.
SURFACE QUALITY-A RESPONSE VARIABLE
5.1. Introduction

In the case of turbine blades, high surface quality assures the outstanding
performance and efficient applications of these components. Morphology and
surface quality include surface texture, surface finish, topography, tolerance of
surface, surface roughness etc. have been described in this chapter to gain a
basic understanding and to differentiate among these parameters.” At the end,
the average surface roughness (Ra) of the samples is measured and then surface
topography of each sample was done with a scanning electron microscope.”

5.2. Surface Quality

The term surface quality is used in broad sense. It includes surface texture,
surface finish, topography, tolerances of surface, surface roughness and other
such related terms.”It has been taken as the response variable in this project.
Usually surface finish and surface quality are used as synonyms of each other
in manufacturing.”

5.3. Surface Finish & Surface Texture

Surface finish is the permissible deviance from a perfectly flat surface that is
made using a manufacturing process [17].”When any method is used to
manufacture a part, there will be some roughness on the surface. This roughness
may be produced by a cutting tool making small grooves at the surface or by
the grains of the grinding wheel. Each cutting pass produces its own groove on
the surface. It depends on the choice of cutting tool, speed of the tool,
environmental conditions, and definitely by which material that is used.”

So it's required to make a compromise among function and cost of

manufacturing. Surface finish generally refers to the overall description of the

39
surface including the texture, the flaws of the materials, and any coatings
applied. However, the term does not include errors of form [17].

“Surface texture is a pattern of the surface which deviates from a normal

surface. The deviations can be repetitive or random and can be the end result
from roughness, waviness, and flaws”[16]. It has two simple uses in
engineering, i.e to control the manufacturing process and its effects on the
performance of work piece. It is normally divided into three components such
as roughness, waviness, and form. From this discussion it is clear that surface
texture is a broader and more accurate term as compared to surface finish.

5.4. Surface quality/Surface finish parameters

Surface finish can be defined by various parameters.”Owing to the requirement

for various parameters in a wide variety of machining operations, a huge amount
of new parameters have been developed. The most commonly used parameters
of surface finish are roughness average (Ra), root mean-square-average (Rq),
max-peak-valley height roughness (Rz) and maximum height of the profile (Rt
or Ry or Rmax).”

5.4.1. Roughness average (Ra)

It is also known as the arithmetic mean roughness value, AA (arithmetic

average) applied in the USA or CLA (center line average) applied in the UK.
Ra is the generally recognized as the mostly used parameter internationally.”It
is the arithmetic mean of the deviation of the roughness profile from the mean
line[18]. Ra is preferably used in order to evaluate gradual surface changes. Yet,
different profile shapes cannot be measured using the Ra value. The measured
value for Ra is always lesser than that of the Rz value measured on the same
roughness profile due to the arithmetic formula shown below”[18].

40
Where Ra= the arithmetic average deviation from the mean line
L= the sampling length
Y= the ordinate of the profile

It is the arithmetic mean of the departure of the roughness profile from the mean
line. “An example of surface profile is shown in Figure 22 and the graphical
representation of Ra is shown in Figure 23.”

Figure 22: Profile of surface texture[18]

Figure 23: Graphical representation of Ra[18]

5.4.2. Root-mean-square (rms) roughness (Rq)

Root-mean-square roughness corresponding to Ra is,

Ra and Rq are the usually the surface finish parameters used in industries.

5.4.3. Rz

“Rz is the average difference between the five highest peaks and five deepest
valleys, within one sampling length. It is frequently applied in Europe”[18].

41
 Figure 24: Graphical representation of Rt & Rz [18]

5.4.4. Max peak-to-valley roughness height ( Rt or Ry or Rmax)

“It is the distance between two lines parallel to the mean line that contacts the
extreme upper and lower points on the profile within the roughness sampling
length”[18].The graphical representation of Rt and Rz has shown in Figure 24

5.4.5. Roughness values for various manufacturing processes

Table 6 provides equivalent values for roughness average Ra,”the approximate

Rt values, the N numbers, the obsolete triangular representations, and some
manufacturing finishes. There is no mathematical relationship between Ra and
Rt .” The Rt values are approximate and should only be used for guidance (Ref:
ISO 1302:1992 which uses N numbers & BS EN ISO 1302:2002 where N
numbers have not included).

42
 Table 6: Comparison between surface finish parameters [17]

5.5. Surface Topography

The term surface topography itself represents all of the spatial structure of peaks
and valleys that exist on a surface. It includes the terms as roughness, waviness,
error of form and flaws [19]. Once again the roughness consists of the closely
spaced irregularities and those can be due to cutting tool marks or can be
produced by way of the grit of the grinding wheel. Roughness, consequently,
has unique structure which can be related to those causes. The waviness includes
more broadly spaced irregularities, which is probably produced by vibration or
chatter in the machine. Errors of the form encompass long-duration or noncyclic
deviations in the surface profile. These can be produced by errors within the
machine bed or spindles, or uneven mounting of the part in the machine. Finally,
flaws are discrete and infrequent irregularities which included cracks, pits, and
scratches. “It is important to realize that the surface topography may affect the
finicality of the component. Therefore, the relationship between the machining
process used and the resulting topography produced need to be determined”
[19].

3D surface topography analysis is carried out usually by two methods i.e. “the
scanning electron microscopy (SEM) method and the use of a 3D surface

43
mapping. The main advantage of using SEM method is the real life surface
features can be seen more clearly. However, it is difficult to establish true
visualization of the surface features because the SEM view is always viewed in
a 2D projection. This problem is solved by using a 3D surface mapping”[19].

5.6. Morphology of test samples

The morphology of the samples was studied after CNC machining to the values
of average surface roughness and the surface topography features.

5.6.1. Roughness Average by Surfcorder

The values of response variable, i.e. roughness average (Ra) was measured
using Surfcorder SE 1700α. The test samples of each run were inspected using
this instrument and readings of Ra was taken and recorded as in Table 7.

Depth
Spindle Feed Rate
Std Run Pt Block of Cut Roughnes
Speed F(mm/min
Order Order Type s ap(mm s Ra(um)
N(rpm) )
)
6 1 2 1 8000 300 0.05 1.435

11 2 2 1 6000 200 0.25 0.9944

15 3 0 1 6000 300 0.15 0.6172

1 4 2 1 4000 200 0.15 1.153

14 5 0 1 6000 300 0.15 0.63

3 6 2 1 4000 400 0.15 1.0385

2 7 2 1 8000 200 0.15 1.794

13 8 0 1 6000 300 0.15 0.6531

10 9 2 1 6000 400 0.05 0.6323

7 10 2 1 4000 300 0.25 0.7976

5 11 2 1 4000 300 0.05 0.5644

8 12 2 1 8000 300 0.25 0.3915

12 13 2 1 6000 400 0.25 0.3161

44
 4 14 2 1 8000 400 0.15 0.5163

9 15 2 1 6000 200 0.05 0.7778

Table 7: Ra values for each experimental run

These Ra values are measured perpendicular to the cutter movement direction.

Ra values in this direction are more definite for surface quality. The results of
the samples obtained by Surfcorder are shown in Figure 25. The detailed
specifications of Surfcorder has shown in Appendix II.

Sample# 01 Sample# 02 Sample# 03

Sample# 04 Sample# 05 Sample# 06

45
Sample# 07 Sample# 08 Sample# 09

Sample# 10 Sample # 11 Sample #12

Sample#13 Sample #14 Sample #15

46
 Figure 25: Surfcorder test results of roughness average (Ra)

5.6.2. Surface Topography with SEM

The machined samples are further inspected through SEM (scanning electron
microscope) model, JEOL JSM-6490 for surface topography. The detailed
specifications of this SEM has shown in Appendix II. Magnifications at 30x was used
to study the macro features of surface roughness. The SEM pictures of surface

topography of machined samples are shown in

Sample# 1 Sample# 2

Sample# 3 Sample# 4

47
Sample# 5 Sample# 6

Sample# 7 Sample# 8

Sample# 9 Sample# 10

Sample# 11 Sample# 12

48
 Sample# 13 Sample# 14

Sample# 15

Figure 26.

Sample# 1 Sample# 2

Sample# 3 Sample# 4

49
Sample# 5 Sample# 6

Sample# 7 Sample# 8

Sample# 9 Sample# 10

Sample# 11 Sample# 12

50
 Sample# 13 Sample# 14

Sample# 15

Figure 26: Surface topography of machined samples through SEM

5.7. Summary

Surface quality is used as the response variable in this project. Other terms
related to surface quality as surface finish, surface texture, and surface
topography has been explained above. The roughness average (Ra) is one of the
key parameter in case of surface quality which was measured for each sample.
Surfcorder was used to measure the values of Ra. It was found that Ra values
change significantly by varying the machining parameters. Trends can be seen
as the surface quality improves and further statistical analysis can give us the
optimum values of these parameters. Also surface topography analysis was
carried out with a scanning electron microscope to further validate the findings
and results.

51
 CHAPTER 6.
RESULTS AND DISCUSSION
6.1. Introduction

This section explains the statistical investigation by using Minitab 16. For this

section the evaluation option available with RSM is conceded out to finalize the

proper model with the current values of machining and response parameters.

Graphical plots such as main effect plot and contour plots are drawn. Analysis

of variance (ANOVA) is held out to find important variables among input

variables for this research work. The fitted values of average roughness were

compared with the results obtained by Surfcorder and SEM. At the end of the

chapter the regression model is presented and analysed.

6.2. Analysing Response Surface Design

Minitab is used as the statistical tool to analyse the design. Response surface

design was applied to discover the fitted model. For this purpose, Table 7 was

applied and roughness average was taken as the response parameter. The

analysis was done using un-coded units. It is said that the model will be best

fitted if it contains following characteristics;

• Lack- of-fit will be larger as compared to other options.

• Error “S “will be min as compared to other options.

A huge value of R-sq does not essentially indicate that the regression model is
a worthy one. Some regression model builders prefer R-sq (adj) value. When R-
sq and R-sq (adj) vary dramatically, there is a greater chance that in important
terms are present in the model [11]. The important values of models obtained
with Minitab 16 have been shown next.

52
• Linear”Model:

S = 0.368508 PRESS = 3.09380

R-Sq = 33.72% R-Sq(pred) = 0.00% R-Sq(adj) = 15.64%

Source DF Seq SS Adj SS Adj MS F P

Lack-of-Fit 9 1.49312 1.49312 0.165902 501.15 0.002

• Full Quadratic Model:

S = 0.154352 PRESS = 1.89685

R-Sq = 94.71% R-Sq(pred) = 15.83% R-Sq(adj) = 85.20%

Source DF Seq SS Adj SS Adj MS F P

Lack-of-Fit 3 0.11846 0.11846 0.039487 119.28 0.008

The detailed output by using Minitab 16 software with RSM has shown in
Appendix I . The fitted model is decided on the basis of lack of fit values for F
& P, error(S), R-sq and R-sq (adj) as applied below.

1) From above it is obvious that P-values of Lack-of-Fit for full quadratic are
small as compared to linear models. Also there is dramatically difference
between R-sq and R-sq (adj) values for linear models as compared to the full
quadratic model. Error “S” is higher in linear model as compared full quadratic
models. linear model is not not adequate

2) Comparison of linear & Quadratic model;

a) P- value of Lack-of-Fit for linear with square model is slightly greater than
linear model but F- value of Lack-of-fit is less for linear as compared to full
quadratic model. i.e full quadratic model is more adequate.

b) The difference between R-sq and R-sq (adj) is greater for linear model as
compared to full quadratic model. i.e full quadratic is more adequate.

c) S-value (standard deviation value in Minitab) for linear model is greater for
linear model as compared to full quadratic model. i.e full quadratic is more
adequate.

53
d) Residual plots in both of the models are almost same as shown in

Figure 27 and Figure 28. But the linear model has more deviation of residual
from the mean.”

Figure 27: Residual plot for linear model

54
 Figure 28: Residual plot for full quadratic model

From above discussion it is clear that full quadratic model is the most adequate
model for experimental factors as spindle speed, axial depth of cut, feed rate
and response variable as surface roughness.

6.2.1. Analysis of Variance (ANOVA)

Analysis of variance was supported out to discover the significance of input

parameters on roughness average. Simple ANOVA was performed by
considering individual factors and then sum up their results as shown in Table
8. It is evident that axial depth of cut (ap) is significant for roughness average
(Ra) values as its P-value for F-statistic is 0.010 which is fewer than 0.05. Since
α= 0.05 is the significance value for 95% confidence level. P-value for spindle
speed and feed rate are 0.497 and 0.575 correspondingly, which is not close to 0.05
and may not be considered as significant.

It means that with small variation in depth of cut (ap) there is large change of
roughness average (Ra). The above mentioned results have also supported by
main effect plot as shown in Figure 29.

55
 Source DF SS MS F P

Spindle speed 1 0.04253 0.01279 0.54 0.497 Marginally significant

(RPM)
Feed rate(mm) 1 0.61383 0.00858 0.36 0.575 Marginally significant

Axial depth( mm) 1 0.10349 0.39290 16.49 0.010 Considered as

significant

Table 8: Analysis of variance (ANOVA) table

Main Effects Plot for Ra(um)

Data Means

N(rpm) F(mm/min)
1.20

1.05

0.90

0.75

0.60
Mean

4000 6000 8000 200 300 400

ap(mm)
1.20

1.05

0.90

0.75

0.60
0.05 0.15 0.25

Figure 29: Main effect plots between Ra & input variables

6.2.2. Contour Plots

From the Figure 30 we can see how the surface roughness improves with
changing each parameter.

56
 Contour Plots of Ra(um)
F(mm/min)*N(rpm) ap(mm)*N(rpm) Ra(um)
400 0.25
< 0.0
0.0 – 0.3
350 0.20
0.3 – 0.6
0.6 – 0.9
300 0.15 0.9 – 1.2
1.2 – 1.5
250 0.10 1.5 – 1.8
> 1.8
200 0.05
Hold Values
4000 5000 6000 7000 8000 4000 5000 6000 7000 8000
N(rpm) 8000
ap(mm)*F(mm/min) F(mm/min) 400
0.25
ap(mm) 0.25

0.20

0.15

0.10

0.05
200 250 300 350 400

Figure 30: Contour plot for Ra, spindle speed, axial depth & feed rate

The dark blue region is showing surface roughness less than 0.3 µm that can be
accomplished by setting the constraints according to the shown values.

6.2.3. Regression model

Minitab16 was again applied to find the regression model which is based on
experimental results. Naturally the Ra value is not zero if all the four input
parameters are taken zero. So the option of Y-intercept was applied. The output
obtained in the form of regression equation/ model is as under

Estimated Regression Coefficients for Ra(um)

Term Coef SE Coef T P

Constant 0.8748 1.35265 0.647 0.546
N(rpm) -0.0002 0.00027 -0.733 0.497
F(mm/min) -0.0033 0.00550 -0.600 0.575
ap(mm) 16.6580 4.10195 4.061 0.010
N(rpm)*N(rpm) 0.0000 0.00000 3.791 0.013
F(mm/min)*F(mm/min) 0.0000 0.00001 2.334 0.067
ap(mm)*ap(mm) -14.0804 8.03272 -1.753 0.140
N(rpm)*F(mm/min) -0.0000 0.00000 -3.768 0.013
N(rpm)*ap(mm) -0.0016 0.00039 -4.136 0.009
F(mm/min)*ap(mm) -0.0133 0.00772 -1.726 0.145

S = 0.154352 PRESS = 1.89685

R-Sq = 94.71% R-Sq(pred) = 15.83% R-Sq(adj) = 85.20%

57
Analysis of Variance for Ra(um)

Source DF Seq SS Adj SS Adj MS F P

Regression 9 2.13451 2.13451 0.237167 9.95 0.010

Linear 3 0.75985 0.46211 0.154036 6.47 0.036
N(rpm) 1 0.04253 0.01279 0.012794 0.54 0.497
F(mm/min) 1 0.61383 0.00858 0.008579 0.36 0.575
ap(mm) 1 0.10349 0.39290 0.392905 16.49 0.010

Square 3 0.55794 0.55794 0.185979 7.81 0.025

N(rpm)*N(rpm) 1 0.33860 0.34234 0.342342 14.37 0.013
F*F(mm/min) 1 0.14613 0.12984 0.129837 5.45 0.067
ap(mm)*ap(mm) 1 0.07320 0.07320 0.073203 3.07 0.140

Interaction 3 0.81672 0.81672 0.272239 11.43 0.011

N(rpm)*F(mm/min) 1 0.33826 0.33826 0.338259 14.20 0.013
N(rpm)*ap(mm) 1 0.40749 0.40749 0.407491 17.10 0.009
F(mm/min)*ap(mm) 1 0.07097 0.07097 0.070969 2.98 0.145

Residual Error 5 0.11912 0.11912 0.023824

Lack-of-Fit 3 0.11846 0.11846 0.039487 119.28 0.008
Pure Error 2 0.00066 0.00066 0.000331
Total 14 2.25363

6.2.4. Analysis of regression model

High T-value and low P-value shows that there is strong relationship between
Ra & ap. With a good model (satisfying all the statistical assumptions), Ra can
be predicted by a combination of input parameters as spindle speed, axial depth
of cut and feed rate.

6.2.5. Optimized parameters and verification

Optimized values for the machining parameters are given by Minitab 16 as:

N(rpm) = 6523.24
F(mm/min) = 400
ap(mm) = 0.25

and the predicted response is:

Ra(um) = 0.1937

For verification of this optimized value, another experimentation was performed

separately and then roughness average (Ra) was obtained with Surfcorder
shown in Figure 31. The values of roughness average is determined to be

58
1.924µm and 0.1687µm for two reading on a complete turbine blade shown in
Figure 32 .

Figure 31: Surfcorder results for optimized values

The slight error may be due to the fact that the values of Ra values are in microns
and due to small variation in environment or other parameters. The slight
difference between Ra values in both of the cases is also due to the information
that the regression model expresses the investigational and statistical results in
mathematical form and there is some error present in the model.

Figure 32: Machined turbine blade according to optimized values

59
 6.2.6. Summary

Minitab16 was applied to analyse the model. It was found that linear model is
the most adequate for these input and response parameters. Different plots such
as main effect plot, interaction plots and contour plots were drawn and found
out the effects of different combination of these parameters upon response
parameter. The discussions had been made at various stages to find the
inferences where required. The regression model then established with Minitab
16 to get the relation between response variable (roughness average) and input
parameters as, spindle speed, axial depth of cut and feed per tooth. The optimum
solution of the parameters was found and verified.

CHAPTER 7.
CONCLUSIONS AND RECOMMENDATIONS
7.1. Conclusions

The values of roughness average (Ra) of 15x machining samples obtained

through Surfcorder vary from 1.8~ 0.3µm. It is found that values of Ra change
by varying the selected machining parameters and their combination. It shows
that with constant scallop height tool path generation and CNC machining, it is
possible to get excellent surface quality/ surface finish.

From contour plots in Figure 30, it is concluded that the roughness average Ra
can be obtained even equal to less than 0.3 μm at spindle speed near (N) 6000
rpm with keeping highest values of both depth of cut (ap) 0.25 mm and feed rate
(F) 400 mm/min.

The optimum results of surface quality (min value of Ra 0.1687 μm) have been
achieved with machining sample where spindle speed (N) 6523 rpm, depth Of
Cut (ap) 0.25 mm & Feed rate (F) 400 mm/min shown in Table 7. During the
tool path generation & CNC machining constant scallop height (h= 0.001mm
kept constant for 15 runs) with tolerance of 0.001 mm (Kept constant for 15
runs) and Zigzag tool path strategy was applied. The surface topography has

60
shown that the micro-graphic structure of sample of Sample No. 13 is the best
as compared to micro-graphic of other samples. It shows that surface
topography has further support and validates the Ra values obtained by using
Surfcorder.

The regression model was developed and analysed. It is concluded on the base
of results that this model is accurate and robust and can be applied for complex
contour jobs with same condition and parameters which are considered in this
research work. Optimum values of the machining parameters were calculated
and verified to give the best possible surface finish.

7.2. Recommendations

It is here by recommended to apply this method to achieve surface quality of

complex contour jobs comparable to the grinding operations. For best surface
quality (min value Ra = 0.1687 μm), the recommended parameters are spindle
speed(N) 6523 rpm, depth of cut (ap) 0.25 mm & feed rate (F) 400 mm/min
(Table 7). Constant scallop height (h= 0.001mm) strategy should be applied
during the tool path generation and CNC machining. It is good approach for
such high performance jobs as turbine blades, impellers, propellers, helical
rotors, etc. because separate grinding operation may harm the profile accuracy,
geometrical features and tolerances. So it is here by recommended that such
high functional jobs should be machined with multi-axis capabilities set up to
produce qualified jobs with grinding less surface quality and separate grinding
operations can be eliminated as done in this research work.

The regression model is recommended to apply with varying values of input

parameters in case of complex contour jobs with factors considered constant
during the experimentation. So we can achieve optimal surface quality with this
regression model without going tedious, expensive and time consuming
experimentation. After calculating minimum Ra value (maximum surface
finish) with this model, the same parameters can be applied for actual machining
of complex contour jobs through multi-axis machining applications.

61
62
REFERENCES

[1] B. C. Routara, A. Bandyopadhyay, and P. Sahoo.,2009, “Roughness

modeling and optimization in CNC end milling using response surface
method: Effect of workpiece material variation,” International Journal
of Advanced Manufacturing Technology.

[2] M. Lendel.,2007, Master Cam X2 4 &5 axis mill training tutorials. In

house solutions Inc, 2007.

[3] Mastercam.,2018, “Mastercam Multiaxis Toutorial,” June. 2018.

[4] B. Lauwers, P. Dejonghe, and J. P. Kruth.,2003, “Optimal and collision

free tool posture in five-axis machining through the tight integration of
tool path generation and machine simulation,” CAD Computer Aided
Design.

[5] M. Lendel.,2005, Master cam X training tutorials-Mill applications. In

house solutions Inc, 2005.

[6] H. Solutions, “Ball Nose Milling Strategy.” [Online]. Available:

http://www.helicaltool.com/secure/Content/Documents/Tech_BallNose
MillingStrategy.pdf.

[7] A. Pramanik and G. Littlefair.,2015, “Machining of titanium alloy (Ti-

6Al-4V)-theory to application,” Machining Science and Technology, 19,
1, pp. 1–49.

[8] D. R. Chichili, K. T. Ramesh, and K. J. Hemkeer.,1998, “The high-strain-

rate response of alpha-titanium: Experiments, deformation mechanisms
and modelling,” Acta Materialia, 46, pp. 1025–1043.

[9] A. Pramanik.,2014, “Problems and solutions in machining of titanium

alloys,” International Journal of Advanced Manufacturing Technology,
70, 5–8, pp. 919–928.

63
[10] R. Fisher.,1935, The Design of Experiments, 9th ed. Macmillan, 1935.

[11] D. C. Montgomery, Design and Analysis of Experiments, 5th ed. John

Wiley and Sons, INC. New York.

[12] Design of Experiments. Minitab Inc.

[13] K. A. A.-E.-H. K.Kadirgama, M.M.Noora, N.M.Zuki.N.M, M.M.

Rahmana, M.R.M. Rejaba, R. Dauda.,2008, “Optimization of Surface
Roughness in End Milling on Mould Aluminium Alloys (AA6061-T6)
Using Response Surface Method and Radian Basis Function Network,”
Jordan Journal of Mechanical and Industrial Engineering.

[14] Basim A. Khidhir and Bashir Mohamed.,2009, “Selecting of Cutting

Parameters from Prediction Model of Cutting Force for Turning Nickel
Based Hastelloy C-276 Using Response Surface Methodology,”
European Journal of Scientific Research.

[15] M. Alauddin, M. A. El Baradie, and M. S. J. Hashmi.,1997, “Prediction

of tool life in end milling by response surface methodology,” Journal of
Materials Processing Technology.

[16] H. Öktem, T. Erzurumlu, and H. Kurtaran.,2005, “Application of

response surface methodology in the optimization of cutting conditions
for surface roughness,” Journal of Materials Processing Technology.

[17] Y. S. Lee and T. C. Chang.,1995, “Application of computational

geometry in optimizing 2.5D and 3D NC surface machining,” Computers
in Industry.

[18] M. S. Lou, J. C. Chen, and C. M. Li.,1998, “Surface Roughness

Prediction Technique For CNC End-Milling,” J Ind Technol, 15.

[19] C. K. Toh.,2004, “Surface topography analysis in high speed finish

milling inclined hardened steel,” Precision Engineering.

64
APPENDIX-I
The analysis was done using uncoded units.

Estimated Regression Coefficients for Ra(um)

Term Coef SE Coef T P

Constant 0.8748 1.35265 0.647 0.546
N(rpm) -0.0002 0.00027 -0.733 0.497
F(mm/min) -0.0033 0.00550 -0.600 0.575
ap(mm) 16.6580 4.10195 4.061 0.010
N(rpm)*N(rpm) 0.0000 0.00000 3.791 0.013
F(mm/min)*F(mm/min) 0.0000 0.00001 2.334 0.067
ap(mm)*ap(mm) -14.0804 8.03272 -1.753 0.140
N(rpm)*F(mm/min) -0.0000 0.00000 -3.768 0.013
N(rpm)*ap(mm) -0.0016 0.00039 -4.136 0.009
F(mm/min)*ap(mm) -0.0133 0.00772 -1.726 0.145

S = 0.154352 PRESS = 1.89685

R-Sq = 94.71% R-Sq(pred) = 15.83% R-Sq(adj) = 85.20%

Analysis of Variance for Ra(um)

Source DF Seq SS Adj SS Adj MS F P

Regression 9 2.13451 2.13451 0.237167 9.95 0.010
Linear 3 0.75985 0.46211 0.154036 6.47 0.036
N(rpm) 1 0.04253 0.01279 0.012794 0.54 0.497
F(mm/min) 1 0.61383 0.00858 0.008579 0.36 0.575
ap(mm) 1 0.10349 0.39290 0.392905 16.49 0.010
Square 3 0.55794 0.55794 0.185979 7.81 0.025
N(rpm)*N(rpm) 1 0.33860 0.34234 0.342342 14.37 0.013
F(mm/min)*F(mm/min) 1 0.14613 0.12984 0.129837 5.45 0.067
ap(mm)*ap(mm) 1 0.07320 0.07320 0.073203 3.07 0.140
Interaction 3 0.81672 0.81672 0.272239 11.43 0.011
N(rpm)*F(mm/min) 1 0.33826 0.33826 0.338259 14.20 0.013
N(rpm)*ap(mm) 1 0.40749 0.40749 0.407491 17.10 0.009
F(mm/min)*ap(mm) 1 0.07097 0.07097 0.070969 2.98 0.145
Residual Error 5 0.11912 0.11912 0.023824
Lack-of-Fit 3 0.11846 0.11846 0.039487 119.28 0.008
Pure Error 2 0.00066 0.00066 0.000331
Total 14 2.25363

Obs StdOrder Ra(um) Fit SE Fit Residual St Resid

1 1 1.153 1.039 0.134 0.114 1.48
2 2 1.794 1.766 0.134 0.028 0.36
3 3 1.039 1.066 0.134 -0.028 -0.36
4 4 0.516 0.631 0.134 -0.114 -1.48
5 5 0.564 0.519 0.134 0.046 0.59
6 6 1.435 1.303 0.134 0.132 1.71
7 7 0.798 0.930 0.134 -0.132 -1.71
8 8 0.392 0.437 0.134 -0.046 -0.59
9 9 0.778 0.938 0.134 -0.160 -2.07
10 10 0.632 0.650 0.134 -0.018 -0.23
11 11 0.994 0.977 0.134 0.018 0.23
12 12 0.316 0.156 0.134 0.160 2.07
13 13 0.653 0.633 0.089 0.020 0.16
14 14 0.630 0.633 0.089 -0.003 -0.03
15 15 0.617 0.633 0.089 -0.016 -0.13

R denotes an observation with a large standardized residual.

65
Predicted Response for New Design Points Using Model for Ra(um)

Point Fit SE Fit 95% CI 95% PI

1 1.03874 0.133673 ( 0.69512, 1.38235) ( 0.51386, 1.56362)
2 1.76616 0.133673 ( 1.42255, 2.10978) ( 1.24128, 2.29104)
3 1.06634 0.133673 ( 0.72272, 1.40995) ( 0.54146, 1.59122)
4 0.63056 0.133673 ( 0.28695, 0.97418) ( 0.10568, 1.15544)
5 0.51877 0.133673 ( 0.17516, 0.86239) (-0.00611, 1.04366)
6 1.30295 0.133673 ( 0.95933, 1.64657) ( 0.77807, 1.82783)
7 0.92965 0.133673 ( 0.58603, 1.27327) ( 0.40477, 1.45453)
8 0.43712 0.133673 ( 0.09351, 0.78074) (-0.08776, 0.96201)
9 0.93769 0.133673 ( 0.59407, 1.28130) ( 0.41281, 1.46257)
10 0.65009 0.133673 ( 0.30647, 0.99370) ( 0.12521, 1.17497)
11 0.97661 0.133673 ( 0.63300, 1.32023) ( 0.45173, 1.50149)
12 0.15621 0.133673 (-0.18740, 0.49983) (-0.36867, 0.68109)
13 0.63343 0.089115 ( 0.40436, 0.86251) ( 0.17528, 1.09159)
14 0.63343 0.089115 ( 0.40436, 0.86251) ( 0.17528, 1.09159)
15 0.63343 0.089115 ( 0.40436, 0.86251) ( 0.17528, 1.09159)

66
67
68
APPENDIX II
Equipment details used for research work
1- CNC machining centre, DMG Mori, DMC 1035v Eco line

Technical Data

Travel X 1035 mm

Travel Y 560 mm

Travel Z 510 mm

A-axis 360 deg

Spindle Speed 10000 rpm

Spindle Taper SK 40

CNC Controller Sinumerik 810 D

69
2- SURF CORDER SE 1700 Α

70
3- SCANNING ELECTRON MICROSCOPE JSM-6490LV

The JSM-6490LV “is a high-performance, scanning electron microscope with

a high resolution of 3.0nm. The low vacuum mode (which can be accessed by
the click of a mouse), allows for observation of specimens which cannot be
viewed at high vacuum due to excessive water content or due to a non-
conductive surface. Its asynchronous five-axis mechanically eucentric stage
with compeucentric rotation and tilt can accommodate a specimen of up to 8-
inches in diameter. Standard automated features include Auto Focus/Auto
Stigmator, Auto Gun (saturation, bias and alignment), and Automatic Contrast
and Brightness.” Other features include:

• Fast, unattended data acquisition

• Streamlined design
• Compact footprint
• Customized toolbars for repetitive functions
• Enhanced SE imaging
• Super conical lens
• LV secondary electron detector (optional)

The JSM-6490LV is used in varied applications with several options that

increase its versatility.

71
4- “CoroMillⓇ Plura solid carbide ball nose end mill for Profiling”

72