Spectrum 21

COMMINUTION
HANDBOOK

Edited by Professor Alban Lynch

COMMINUTION
HANDBOOK
Edited by Professor Alban Lynch

The Australasian Institute of Mining and Metallurgy

Spectrum Series 21

Published by:
THE AUSTRALASIAN INSTITUTE OF MINING AND METALLURGY
Ground Floor, 204 Lygon Street, Carlton Victoria 3053 Australia
 © The Australasian Institute of Mining and Metallurgy 2015

No part of this publication may be reproduced, stored in a retrieval system or

transmitted in any form by any means without permission in writing from the publisher.

The Institute is not responsible as a body for the facts and opinions
advanced in any of its publications.

ISBN 978 1 925100 38 9

Desktop published by Kate Hatch and Claire Stuart for

The Australasian Institute of Mining and Metallurgy

Index prepared by:

Russell Brooks

Front cover images:

Top: courtesy Outotec (see Figure 5.1).
Middle: stock image of an underground grinding mill.
Bottom left and middle: courtesy Magotteaux
(see Figure 7.3 and Figure 9.16).
 Foreword
Alban Lynch is that rarest of combinations: a consummate technologist, a visionary leader
and a clear and persuasive communicator. This book is the latest exemplar of his skills.
Like the best developments, we didn’t know we needed it until we saw it. Now it seems
indispensable.
There is a place for deeply specialist (but rather turgid) texts; for specific-topic
technical papers; and for the ubiquitous Google searches and Wikipedia references.
But where was the handbook for the whole field of comminution – one authoritative
reference source for experts, yet an accessible tool to those new to the field? Who could
convince a list of esteemed but busy experts to volunteer their time; who could link
their individual contributions with insight and eloquence, to make the whole speak
cohesively and more clearly than the sum of its parts? Who else but Alban Lynch?
Just as in his career, traditional boundaries were no obstacle to Alban’s vision for this
book. He assembled experts from all fields of comminution. Most ‘mineral processors’
think of grinding as wet grinding of mineral ores. On the first page Alban gently
reminds us that more than twice this tonnage of cement and coal is subjected to
fine dry grinding. So this text includes contributions on the whole array of crushing,
grinding and classification equipment used to comminute the wide range of materials
demanded by our communities. Of course, the discussion would not be complete
without the explanations of sizing techniques, classification, testing and scale-up
methods, mineral liberation, circuit design, process control, circuit practice and recent
technology developments. By grasping the entire field, we are encouraged to seek new
opportunities by looking outside our personal and professional silos. It should haunt
us that more efficient devices such as high-pressure grinding rolls and tower mills were
common in dry grinding at least 25 years before they were discovered for ore grinding.
This book may help ensure that we don’t let that happen again.
Like Alban, the contributors to this book are dedicated to improving the efficiency of
meeting society’s needs. This is vital work. On the first page we are reminded that the
comminution of ores, cement and coal consumes around three per cent of the world’s
electrical energy, and that society will demand another 25 per cent of these materials by
2040. The energy requirement will increase more than that. Mineral ores are becoming
both lower grade and finer grained, so to produce one tonne of metal we will need to
process more tonnes of rock, and many of those tonnes will require finer grinding. Unless
something changes, the additional 25 per cent product may require closer to double the
energy used in 2010. That is the problem. Chapter 1 concludes by reminding us of the
opportunity: the theoretical energy efficiency of the most commonly used equipment is
still only five per cent.
 This challenge for current and future engineers is clear. Achieving step changes in
comminution efficiency won’t be easy. But it is our job, and society needs us to do it. If we
don’t, who will?
I recommend this text as the starting point for our quest, and Alban’s vision and tireless
work to assemble it as the inspiration.
Joe Pease FAusIMM
 Preface
Comminution is the process in which solid materials are reduced in crushing and grinding
circuits. Comminution machines are the equipment in which the process is carried out.
Billions of tonnes of ore, cement and coal pass through comminution processes annually
and these tonnages are increasing. To meet the rapid growth in demand, significant
advances have occurred in the design and construction of equipment, which have been
mainly due to improvements in energy and materials technologies. Some that have been
adapted to comminution include huge motors and gearless mill drives for the immense
rotating mills, and composite materials for protective linings to minimise breakdowns
due to wear.
In this book the processes and machines involved in modern-day comminution are
discussed, including the ancillary techniques used for the design, control and optimisation
of circuits. Because of advances in comminution technology and the growing population
and demand for materials, the process is now so widespread and intensive that
comminution is one of the technologies on which civilisation depends on a daily basis;
for example, grinding coal for power generation, clinker to make cement for buildings,
and grains to make bread.
Comminution is mainly carried out by mechanical devices, the exception being the first
stage in most mining operations when it is carried out by blasting to extract broken rock
from the rock mass. Blasting is an energy-efficient and cost-effective way of preparing
ores for the crushing and grinding circuits in which the fine particles that are required
for mineral liberation and separation are produced. Blasting is not discussed in this book,
except to mention the link that exists between blasting intensity and the size of the rock
fragments that comprise the feed to semi-autogenous grinding (SAG) mills.
The main processes in comminution are breakage and classification. Breakage is a size
reduction process and it occurs by impact, compression, shear or attrition. Classification
is a size separation process that occurs by screening or by the differential movement of
particles in liquids or gases. Closed circuits, which involve recycling the coarse fraction
from the classifier back to the mill, are commonly used for comminution processes. It is
important to optimise both the sizes of machines in the circuits at the design stage and
the operating conditions of the circuits at the production stage.
There are four themes in discussing comminution in this book:
1. the characteristics of the material to be processed (chapters 2 to 4)
2. the equipment available for processing (chapters 5 to 10)
3. circuit design (chapters 12 and 16)
4. circuit optimisation and control (chapters 11 and 13–15).
 In an effort to cover everything, the topics are discussed concisely and this means
that comments on some important issues have been curtailed or omitted. For further
information, the reader is referred to the symposia on comminution that are organised by
universities and engineering institutions.
The purpose of this book is to present comminution as it is today to those with the
responsibility of improving the technology in the future. The emphasis is on ores, but it is
hoped that there is sufficient discussion of cement and coal to encourage engineers in all
specialities to discuss progress jointly in the future.
There are many people who assisted in the compilation of this book and they are listed
in the following pages. In particular I wish to thank Hakan Dundar, who has contributed
much to several chapters of this book, both technically and mathematically, and has
provided continuing advice and comments. His responses were always fast and relevant.
Thank you for your assistance Hakan. I also wish to thank the AusIMM publishing team,
Stephanie Ashworth, Kristy Burt and Kelly Steele, for their patience, persistence and
dedication.
Professor Alban Lynch HonFAusIMM
 Contributors

Alban Lynch HonFAusIMM

Alban’s career in mineral processing began in 1954 at the Zinc Corporation in Broken Hill.
He then moved to the University of Queensland in 1958, where he worked as a research
engineer for the next 30 years. Amongst his achievements there was his involvement in
the first plant-based research project at Mount Isa Mines. Alban then became Foundation
Director of the Julius Kruttschnitt Mineral Research Centre from 1971 to 1988, where
he specialised in the modelling and control of processes. After that he was Head of the
Department of Mining and Metallurgical Engineering from 1988 to 1993, and then was
a Visiting Professor at universities in Brazil, Mexico, Malaysia and Turkey over a 15-year
period. As well as lecturing, Alban established research programs in these countries,
the most successful of which was in cement clinker grinding at Hacettepe University in
Turkey. This program has grown from a research group of one graduate student in 1999
to a staff of three along with ten graduate students in 2009. The program works closely
with the cement industry.
Alban’s publications include two books on mineral process simulation and control,
The History of Grinding written with Chester Rowland and more recently, the AusIMM-
published History of Flotation. He has also authored over 150 technical papers. Alban has
been the recipient of the AusIMM’s President’s Award and Institute Medal, the AIME’s
Robert H Richards Award and the SME’s Antoine M Gaudin Award. He has also received
the IMPC’s Lifetime Achievement Award and the Order of Australia, and is a member of
the International Mining Technology Hall of Fame (Comminution).

Katie Barns MAusIMM

Katie has worked in the minerals industry since 1994 in a variety of roles, including
flow sheet and plant development, design and commissioning followed by ongoing
optimisation of operations in both Australia and overseas. More recently Katie has
been with Glencore Technology over the past ten years (formerly Xstrata Technology)
as Strategy Manager for the IsaMill Technology business. Here her focus has been on
improving the overall energy efficiency of mineral processing operations in the whole
process chain. Katie has a degree in Chemical Engineering from the University of
Queensland and an MBA from Deakin University.

Dirk Bass
Dirk has been working in the field of instrumentation and control for over 20 years. In
the last 16 years he has specialised in optimising process control systems in a wide range
of industries including food and beverages, manufacturing and mineral processing. Dirk
joined PanAust Ltd in November 2011 as the Senior Advisor – Process Control, where his
current role is to develop the advanced control solutions and process control standards
for the company. He believes that mineral processing plants should be highly automated,
elevating the control room operator’s role from adjusting set points to ensuring process
objectives are achieved.

Duncan Bennett MAusIMM

Duncan commenced his metallurgy career with copper, gold and bismuth at North
Broken Hill Peko’s Warrego operations in the Northern Territory (NT), followed by time
at WMC Resource’s Kambalda nickel concentrator. He then spent four years working
in gold plants in southern Western Australia (WA) with Resolute Ltd. After periods
at Queensland Nickel and managing a private CIP (clean-in-place) carbon-cleaning
business in WA, Duncan moved to Tasmania for eight years where he worked in mill
and general management positions at the Renison Bell and Mount Bischoff tin mines,
including a two-year period at Australian Paper’s Burnie mill. Following a brief foray
back in gold at Hill End in New South Wales (NSW) and Union Reefs in the NT, Duncan
has been with PanAust Ltd since 2010 as Principal Metallurgist overseeing metallurgical
and process development of new and existing copper and gold projects.

Hakan Benzer
Hakan is the Head of Mining Engineering at Hacettepe University, where he also attained
his PhD, and is Director of its International Mining Center. He leads the Hacettepe
Comminution Group and is a member of the Global Comminution Collaborative (GCC),
a research collective addressing the sustainability of comminution in the mining
industry. Hakan has been a visiting academic at JKMRC at the University of Queensland,
and is currently running several projects for the cement and minerals industry. His main
interest is on comminution and classification circuits, where his major research topic is
the modelling of dry grinding and classification circuits.

Alain Broussaud
Alain holds a Master’s degree in Physics and an Engineering degree from the School of
Mines in Nancy, France. He is currently Vice President, Virtual Plant Simulation program
at Metso Minerals and co-founded Metso Cisa (originally Cisa) in 1990. Cisa developed,
marketed and implemented original optimising technology for the mineral processing
industry, including vision systems, acoustic sensors and advanced process control
systems. While being Cisa CEO for 23 years, Alain remained involved in technical matters
and contributed to advanced control projects worldwide. His earlier assignments
include being Plant General Manager at a phosphate company in Senegal, and Deputy
Head of BRGM’s Mineral Technology Department in France, where he created the
UsimPac simulation software package in mineral processing in 1986.

Marcos Bueno GAusIMM

Marcos has eight years’ work and research experience in mining and mineral processing
engineering with in-depth knowledge of the design and optimisation of comminution
circuits and geometallurgy. He has carried out extensive comminution circuit surveys
and pilot plant campaigns in Australia and internationally, and has been involved
in technical studies and process design engineering projects related to most key
mining commodities including gold, copper, iron ore, aluminium, nickel, platinum and
phosphates.

Don Burgess FAusIMM

Don immigrated to Australia and joined Allis Chalmers in 1966, working in the design
office. His interest in crushers and grinding mills was heightened after meeting Fred
Bond on tour in Australia in 1968. Don later became Allis Chalmers’ crushing and
grinding specialist where he has been involved in the design and application of over 300
grinding mills in Australasia, Africa and Europe. In 1994 Don set up his own consulting
business specialising in comminution systems, which involved mill selections, JKSimMet
simulations and design innovations such as fitting discharge grates in 15-foot diameter
rod mills. He has written several technical papers on grinding including one describing an
accurate method of calculating AG/SAG specific energies, an example of which is in this
publication.

Hakan Dundar
Hakan is currently an Assistant Professor in the Department of Mining Engineering
at Hacettepe University in Turkey, where he received his Bachelor, Masters and PhD
degrees in mining engineering. While comminution was the main focus of his Masters
and PhD, he also studied cement grinding (Masters) and high-pressure grinding rolls
in the minerals industry (PhD). Hakan specialises in comminution circuit design and
modelling, and is skilled in the use of modelling and simulation tools to design and
optimise a crushing and grinding circuit, as well as model development. Energy saving
and capacity improvement in a comminution circuit form his main areas of interest.

Udo Enderle
Udo graduated in Mechanical Engineering from the Technical University Munich where
he specialised in design and development. He later joined Netzsch Feinmahltechnik the
same year, and managed projects for paint and ink production plants in Europe, China and
Russia. Udo then spent several years as Head of Mill Design and was later Technical Director
for Netzsch. Since July 2009 he has been Managing Director of Netzsch Feinmahltechnik
where his focus is on technology and development. Udo holds over 100 patents and was
co-inventor of the IsaMill.

Cathy Evans
Cathy is a mineral processing engineer with over 30 years’ experience in the minerals
industry. Since graduating in Mineral Technology from Imperial College, London, she
has worked as a metallurgist in mines in South America and Australia and as an industry-
focused research metallurgist at the Julius Kruttschnitt Mineral Research Centre
(JKMRC) at the University of Queensland. During her career at JKMRC she has applied
her knowledge of processes and ore mineralogy to develop practical methods for
optimising mineral processing operations. Understanding mineral liberation is the key
to optimising comminution and separation processes as an integrated process chain,
and through Cathy’s consulting and research for industry partners while measuring and
modelling mineral liberation, she has developed a broad range of expertise in this area.

Matthew Fitzsimons
Matthew studied a Bachelor of Mechanical Engineering (Aeronautical) at the University of
the Witwatersrand, Johannesburg and graduated in 2001 with four distinctions. He began
his working career at the CSIR in the Defence Aeronautics division where notable
achievements include project managing the upgrade of the High Angle of Attack (HAOA)
capability in the Transonic Wind Tunnel Facility and being nominated for a Technological
Scientific Excellence Award as part of the Continuous Force Development and HAOA
Capability Testing Team. In 2008 he went to Murray and Roberts Steel as a Specialist
Engineer to manage research and development; this encompassed projects involving
product development, capital asset management and production optimisation. Matthew
has been at Multotec as Technical Manager for the past five years.

Bianca Foggiatto MAusIMM

Bianca has eight years of experience in metallurgical and process engineering. She
completed both her Mining Engineering and Masters in Minerals Engineering degrees
at the University of Sao Paulo, where she specialised in the comminution of Brazilian
iron ores. Bianca’s expertise covers management and planning of laboratory test work,
coordination of plant/pilot plant trials and surveys and process design/optimisation.
She has worked at HDA Servicos, a consulting company in Brazil, and with Votorantim’s
technology team where she oversaw all nickel operations and tailings recovery. She
is presently working as a process engineer at Ausenco and completing her PhD at the
University of Queensland’s JKMRC, specialising in comminution and circuit energy
efficiency.

Bodo Furchner
Bodo earned his degree in Process Engineering at the Technical University Munich. After
his studies he worked as a scientist at the Institute for Process Engineering in Munich,
where he undertook his Doctor’s degree. In 1987 he began work for Hosokawa Alpine
in research and development and in 1995 was appointed Manager of the Test Centre
Mechanical Processing. Since 1999, he has been General Manager of the Technical
Division.

Olivier Guyot
Olivier is currently Vice President of Metso Minerals Center for Advanced Technology
(MCAT), and before that was the General Manager of Metso-Cisa. He holds a degree in
Mining Engineering from the Ales School of Mines in France, and possesses 26 years
of experience in creating and delivering innovative technology for the minerals and
metallurgical industry, including modelling, simulation and advanced sensing. Olivier
created OCS© software, a leader in advanced control for mineral processing optimisation,
as well as a series of pioneering vision and audio-based sensors, VisioFroth™ and
VisioRock™.
Cathy Hewett
Cathy is a Materials Consultant Engineer in Perth, Australia and has over 18 years’
experience within the materials, mining and manufacturing spheres. She holds a PhD
in Engineering Science from Monash University, a Bachelor’s degree in Applied Physics
(RMIT University) and a Master of Business Administration (Deakin University). The
common thread in Cathy’s career has been the mitigation of erosive and abrasive wear
using innovative materials solutions alongside engineering design.

Rick Hughes
Rick is Managing Director and Principal Consultant of Microanalysis in Perth, Australia. He
graduated from Robert Gordon University in Aberdeen, Scotland with Honours in Physics,
and has over 27 years’ experience in the industry in all facets of particulate characterisation
and sizing. Rick specialises in forensic particulate science, assisting clients from a broad
spectrum of areas to identify and understand their particulate matter and how this
impacts on their processes.

Cliff King MAusIMM

Cliff is currently the Principal Process Engineer at Preplab Testing Services in Rockhampton,
Queensland. Graduating from the University of Queensland in 1973, he worked in base
metals at Cobar/Broken Hill from 1974 to 1979 before joining the coal industry and
working at Moura, Riverside and Curragh coal mines. Cliff then moved into consulting in
1997, working at Burton, North Goonyella, Moranbah North and Cameby Downs mines.

Deon Kok
Deon obtained his Metallurgical Engineering degree from the North-West University,
South Africa in 1993 and Master’s degree in Business Leadership from the University of
South Africa in 2003. He worked in various industries and capacities (gold, manganese,
steel) before joining Newmont Ahafo Ghana in 2008. He is currently the Process and
Commissioning Manager for the Ahafo mill expansion project.

Greg Lane FAusIMM

Greg has around 30 years’ experience in operations, engineering and design, and study
and project management, with an industry-leading knowledge of concentrator design. In
his current role as Chief Technical Officer of Ausenco, he provides specialist technical and
project development expertise on major projects for clients. The author of more than 40
publications on different aspects of minerals processing and project development, Greg
is a highly sought-after technical expert and world leader in plant design, particularly
comminution and flotation circuits.

Geoffrey Legrand
Geoffrey graduated from the Ecole Nationale Supérieure de Géologie in Nancy, France, and
joined Metso in 2007. He was first involved in several research and development projects
related to the VisioFroth™ technology and then in advanced process control projects
covering grinding, crushing, flotation, thickening and pelletising. Geoffrey has been
involved in projects for companies such as Rio Tinto, Newmont and AngloGold Ashanti in
more than 15 countries. He has extensive expertise in MCAT advanced sensors including
VisioFroth™, VisioRock™ and AudioMill™, and manages the APC engineering team based
in France, which supports Europe, the Middle East and Africa, as well as the hardware
production team providing vision and audio advanced sensors around the world.

Aubrey Mainza
Aubrey has a great deal of experience in the area of comminution and classification. In
addition to a year at Zambia Consolidated Copper Mines, he has spent more than 17 years
working in the comminution research group in the Centre for Minerals Research at the
University of Cape Town. Aubrey has also participated in the design and optimisation of
many mines including all the major platinum and gold mines in Africa and overseas. He is
currently working as an Associate Professor in the Department of Chemical Engineering
at the University of Cape Town and is the Head of Comminution Research and Deputy
Director for the Centre for Minerals Research.

Eddie McLean FAusIMM

Eddie is the Manager of Minerals Consulting at Ausenco, which is based in Brisbane. He is
a graduate in Metallurgy from the University of Queensland and a Fellow of the AusIMM.
At Ausenco since 2002, his roles have included managing regional and national groups of
process engineers and providing technical support to international group offices. Eddie’s
expertise encompasses a range of process design and engineering activities, with much
experience gained in comminution, beneficiation and hydrometallurgical extraction
in the following commodities: precious metals (gold, silver), base metals (copper, lead,
nickel, zinc), mineral sands and industrial minerals.

Gunter Metzner
Gunter started his career with the Measurement and Control Division at Mintek in South
Africa as part of the team that implemented the first successful multivariable controllers
for milling. He joined the De Beers Group in 1994 to build a research group focusing on
automation, monitoring and diagnostics, and was instrumental in developing an expert
control system that successfully increased throughput and stability in diamond recovery.
Gunter has been the Manager of Process Research and Development of DebTech, has
spent some time with the Advanced Systems Group of SGS and is now the Advanced
Process Control Regional Manager for Australasia at Metso’s Minerals Centre for Advanced
Technology (MCAT), where he is involved in technologies for the control and optimisation
of minerals processing plants.

Chris Morley
Chris has 43 years’ experience in design, commissioning, operation and control of
metallurgical and materials-handling plants covering diamonds, coal, gold, silver, iron
ore, vanadium, base metals, platinum and uranium. He has experience in Australia and
internationally, and has established a recognised level of expertise in the engineering of
comminution circuits, especially in the application of high-pressure grinding rolls (HPGR)
systems for several commodities. Working on the Boddington feasibility study, Chris was
involved in the development of a practical high-capacity HPGR-based circuit – one of the
first of its kind – that demonstrated project viability and enabled project execution.

Steve Morrell
Steve is a minerals processing engineer with over 30 years of specialist experience in
comminution, where he has been involved with the design and optimisation of most
major comminution circuits in the world. In 1980 he graduated with a Bachelor's degree
(Honours) in Engineering Science from Imperial College, London, majoring in Metallurgy
and spent the next seven years working on mines throughout Africa. He subsequently
moved to Australia and completed Masters and Doctorate theses in grinding mill
simulation and power draw modelling at JKMRC. Up until 2000, when he left JKMRC to
start his own consultancy (SMCC Pty Ltd), Steve oversaw world-leading research projects
such as the AMIRA P9, High-pressure Grinding Rolls, Fine Grinding and Mine-to-Mill
programs. In 2003 he founded SMC Testing to license the SMC Test® that he developed.

Joe Pease FAusIMM

Joe has worked in the minerals industry since 1982 in a variety of research, operations
management, and technology development roles, including 20 years at Mount Isa copper
and lead-zinc concentrator and smelting operations. He was CEO at Xstrata Technology
for 12 years, and has continued his focus on improving processing efficiency with roles
as CEO of the Cooperative Resource Centre for Optimising Ore Extraction (CRC ORE)
and Board positions with AMIRA, JKMRC and the Ian Wark Institute. Joe is Chairman of
CEEC (Coalition for Eco Efficient Comminution), and is on the Steering Committee of the
Minerals Tertiary Education Council.

Marc Revalor
Marc received his PhD in Engineering Science (design, optimisation and control of rolling
mills) in 2008, and also holds an Engineering degree from the Ecole Centrale in Lyon and
a Master’s degree in Industrial Automation. He has had four years’ experience in process
modelling and control (rolling mills, steel making) at Alcan and ArcelorMittal in France,
before joining Metso Minerals Center for Advanced Technology (MCAT) in January 2009.
Since then, Marc has been involved in advanced process control and modelling projects
worldwide, particularly for concentrators and induration plants. As Technical Expert
for Advanced Process Control and Modelling, he spent a year seconded at Rio Tinto’s
Processing Excellence Centre in Brisbane, and now leads real-time dynamic process
modelling efforts for Metso MCAT.

Etienne Roux
Etienne graduated from the University of Stellenbosch in South Africa, then joined Anglo
American Base Metals in 2003 before going to Newmont Mining Corporation in 2010.
He was responsible for the client-side implementation of the Advanced Process Control
system for the gold mill at Newmont’s Ahafo Project in Ghana, and has been involved
in technical and production roles in commodities such as zinc, copper, lead, titanium
and gold in several countries. Etienne is currently Senior Operations Superintendent at
Horsehead Corporation’s Mooresboro zinc project, where he manages operations of the
Waelz oxide leaching, solvent extraction, bleed treatment and lead recovery plants, as
well as technical support for the zinc electrowinning and casting operations.

John Russell MAusIMM

John attained tertiary qualifications in Mechanical Engineering at the Queensland
Institute of Technology (now QUT) and then joined Mount Isa Mines (MIM) in 1980,
where he spent five years gaining experience in copper and lead smelters and
concentrators, and associated underground mines. Leaving MIM in 1985, he formed
Russell Mineral Equipment (RME), a niche-market engineering enterprise whose core
business is the investigation, design, build and commissioning of specialised mineral
processing equipment for industry. In 2007, John was awarded the Canadian Mineral
Processors Art McPherson Medal for contribution to the advancement of comminution,
and in 2009, an Honorary Doctorate in Engineering from the University of Southern
Queensland. In 2014, he was presented with the Mineral Industry Technique Award by
the AusIMM.

Glenn Schumacher MAusIMM

Glenn is currently Chief Engineer, AGL Energy Group Operations and has responsibility
across its mining and major electricity-generating operations. Prior to joining AGL in late
2013, he was General Manager of NRG Gladstone Operating Services for six years, where
he was responsible for the overall management and leadership of Queensland’s largest
single power station. Glenn has held a number of positions in the electricity generation
industry as well as other areas such as technical services (Tarong Energy Corporation),
production (International Power, Hazelwood Power Station) and maintenance (SILCAR). He
holds a Bachelor of Engineering (Mechanical), Master of Engineering (Power Generation),
Master of Business Administration and Doctor of Engineering (honoris causa).

Spike Taylor
Spike studied a Bachelor of Science (Engineering Extractive Metallurgy) at the University
of the Witwatersrand, Johannesburg and graduated in 1977. He then worked for De Beers
for seven years before moving through the Graduate Metallurgist program and working
in research and development. This was followed by three-and-a-half years at Debswana
Jwaneng commissioning and operating the main treatment plant as an Assistant Plant
Superintendent. Spike has now been at Multotec for over 30 years: the first 18 years in
the sales and marketing of polyurethane and wedge wire screening media and trommel
screens for Multotec Manufacturing, followed by Sales Director for Multotec Rubber from
2002 until 2009 and then Managing Director.

Walter Valery FAusIMM

Walter is Metso Global Senior Vice President of Technology and Innovation. He is a
comminution expert and leader in the industrial implementation of Mine to Mill,
comminution process characterisation, modelling, simulation and optimisation of mineral
processing operations. Walter’s contributions are based on 30 years of fundamental
research, industrial experience and pioneering transformational consulting, and he has
published and presented more than 100 technical papers. He is one of the pioneers in
ore tracking from the blast to the mill and development of geometallurgical systems for
integration and optimisation of mine and processing plants. Walter is currently working
on the development of resource and eco efficient mining processes to extract minerals
more efficiently.

Peter Walker MAusIMM

Peter is a metallurgist with over 30 years’ experience in the design, commissioning
and operation of processing plants and general management of operations in Europe,
Australasia and South America. In recent years he has been responsible for the feasibility
and development of greenfield and brownfield projects in Thailand, Laos and Chile. Peter
has worked for a number of engineering groups, as well major and mid-tier operating
companies with commodities such as lead/zinc, uranium, coal, nickel, copper and
precious metals. He has specific experience in the design and operation of large-scale
comminution circuits involving autogenous and semi-autogenous grinding.

Thomas E Warne
Tom is the President of Schutte-Buffalo Hammermill in Buffalo, New York, who manufacture
over 250 different models of size-reduction equipment. He served as General Manager of
Buffalo Hammermill Corp from 1991 to 2001, and became President of Schutte-Pulverizer
in 2001 where he spearheaded the acquisition of his former employer Buffalo Hammermill
Corp later that year. In 2004, Tom purchased the assets of the combined companies along
with partner James N Guarino. As President and CEO of Schutte-Buffalo Hammermill, he
guided the company to over 300 per cent growth with sales to more than 50 countries
worldwide. In December, 2014 they sold the company, though Tom continues to serve in
the role of Company President.

Mark V Weaver
Mark has been a registered professional engineer since 1989. He earned his Bachelor
of Science in Mechanical Engineering in 1984 and a Master of Engineering from the
University of Alabama, Huntsville, in 1991 while working as Spacelab Mission Lead for the
USMP-series of Spacelab Missions. He joined Polydeck Screen Corporation as Engineering
Manager in 2005, served as Director of Engineering and is now the Director of Research
and Development. During his tenure as Engineering Manager, Mark developed new
designs for trommels now used throughout the world that utilise modular synthetic
screening media.

Jobe Wheeler
Jobe has a Bachelor of Science in Mechanical Engineering from the University of
Buffalo, New York. His industrial experience in manufacturing began at Motorola,
concentrating on equipment maintenance/reliability and process improvements. He
entered the mineral processing industry in 2007 with Derrick Corporation, a leader in
fine-screening technology, with a focus on iron ore processing and comminution circuit
improvements. Since starting at Derrick, Jobe has worked on comminution projects in
the Minnesota Iron Range, Labrador Iron Trough (Canada) and several countries around
the world with a focus on improving classification efficiency and product grade, and
increasing recovery of valuable minerals.

Bob Yench
Bob originally trained as an electrical engineer in Melbourne before moving to Mount
Isa Mines in 1966 where he commenced work as an instrument engineer. He spent the
following 48 years working in instrumentation and process control, primarily in the mineral
processing industry. In 1997 Bob became the inaugural CEO and Managing Director of
MIPAC Engineering (later MIPAC Pty Ltd), a Brisbane-based specialist process control
group. His many roles in the industry over this period have included design, installation
supervision, commissioning and maintenance of process plant control systems. Bob
has had a long-term interest in promoting greater process control knowledge for plant
operators.
 Acknowledgements

STEERING COMMITTEE

Alban Lynch HonFAusIMM, Editor

Diana Drinkwater MAusIMM, Program Director – Accelerated Development
Portfolio, JKTech Pty Ltd, Brisbane, Australia
Peter Tilyard FAusIMM(CP) Metallurgist, Tilyard Metallurgical Services,
Melbourne, Australia

CONTRIBUTORS
Along with the primary contributors listed in the previous pages, the people listed below
assisted with the compilation of this book by providing contributions, comments and
advice. Their involvement in this project is greatly appreciated.
While contributions came from many sources the opinions presented in this book are the
responsibility of the Editor.

Geoffrey Barnett Managing Director, Minco Tech Australia Pty Ltd, Cardiff,
Australia
Miron E Boris Process and General Audits, Thrane Teknikk CJSC, Electrostal
(Moscow), Russia
Rob Coleman MAusIMM, Head Mineral Processing Solutions, Outotec
South-East Asia Pacific, Brisbane, Australia
Eddie De Rivera Managing Director, MIPAC Pty Ltd, Brisbane, Australia
Chris Greet FAusIMM(CP), Manager Minerals Processing Research,
Magotteaux Australia Pty Ltd, Adelaide, Australia
Yaqun He Mineral Processing Manager, China University of Mining and
Technology, Jiangsu, China
Rajiv Kalra MAusIMM, Global General Manager, CITIC Heavy Industries,
Sydney, Australia
Amit Kumar Consultant / Mineral Processing Engineer, Vancouver,
Canada
 Suzanne Lynch-Watson MAusIMM, General Manager, Process and Grinding,
Multotec, Brisbane, Australia
Jeff McKay Manager, Expert Systems Global, Metso Mineral Center for
Advanced Technology (MCAT), South Jordan, Utah, USA
Mark McVey Managing Director, MMD Australia Pty Ltd, Narangba,
Australia
Gavin Pasin Regional Product Manager Asia Pacific – Mill Lining
Solutions, Metso Minerals, Brisbane, Australia
Marc Piccinin Grinding Process Engineer, The Cement Grinding Office,
Verona, Italy
Jerome Portal Export Sales Manager, Fives FCB, Lille, France
Rolf Steinhaus Sampling Specialist, Director – Multotec Process Equipment,
Kempton Park, South Africa
Ron Wiegel Mineral Processing Consultant, Lakeland, Florida, USA
Heather Wilt Director of Marketing and Communications, McLanahan
Corporation, Hollidaysburg, Pennsylvania, USA
Peter Wulff TowerMill Business Development, Eirich Group
Headquarters, Hardheim, Germany
Jawahar M Yardi Cement Engineer, Brisbane, Australia

REVIEWERS
We wish to thank the following reviewers for their helpful feedback.

Chris Bailey MAusIMM, Advisor – Processing and JKSimMet Product

Manager, JKTech, Brisbane, Australia
Ted Bearman Director, Bear Rock Solutions, Perth Australia
Johannes Cilliers Chair of Mineral Processing and Head of Department of
Earth Science and Engineering, Royal School of Mines,
Imperial College, London, UK
Dean David FAusIMM(CP), Technical Director – Process, Amec Foster
Wheeler, Perth, Australia
Bill Johnson FAusIMM(CP), Senior Principal Consulting Engineer,
Mineralurgy and Adjunct Professor, Julius Kruttschnitt
Mineral Research Centre, University of Queensland, Brisbane,
Australia
Emmanuel Manlapig Senior Processing Manager, Julius Kruttschnitt Mineral
Research Centre, University of Queensland, Brisbane,
Australia
Bill McKeague Business Development Manager Asia Pacific, MineSense
Technologies and Owner, Adaptive Solutions, Brisbane,
Australia
Brian McNab MAusIMM(CP), Principal Process Engineer, Amec Foster
Wheeler, Perth Australia
Rob Morrison MAusIMM, Chief Technologist, Julius Kruttschnitt Mineral
Research Centre, University of Queensland, Brisbane,
Australia
Andrew Newell MAusIMM(CP), Executive Consultant, Processing,
RungePincockMinarco, Brisbane, Australia
Sam Palaniandy Senior Research Fellow, Julius Kruttschnitt Mineral Research
Centre, University of Queensland, Brisbane, Australia
David Royston Principal, Royston Process Technology, Brisbane, Australia
John Starkey President and Principal Consulting Engineer, Starkey &
Associates Inc
Walter Valery FAusIMM, Global Senior Vice-President, Metso, Brisbane,
Australia
Elaine Wightman Senior Research Fellow, University of Queensland, Brisbane,
Australia
Bob Yench Business Development Consultant, MIPAC, Brisbane,
Australia
 Sponsors
The AusIMM would like to thank the following sponsors
for their generous support of this voume.

Principal Sponsor

Major Sponsor

General Sponsors
 Constancia copper-molybdenum project, Peru

Ausenco is recognised for

its expertise in all aspects of
comminution.
Ausenco provides process Ausenco’s leading-edge • We utilise other recognised
design, engineering, project comminution experts utilise comminution programmes
construction and asset in-house modelling and and methodologies such as
management services to the simulation programmes to JKSimMet to benchmark and
global resources and energy provide practical solutions for cross-reference performance
industries. comminution circuit design, and outputs.
equipment selection and sizing.
Ausenco’s skilled professionals Their technical efforts are • We take metallurgical test
offer expertise across all phases supported by proven layout and work data, ore resource
of a project lifecycle: test work for engineering design expertise to models and mine schedules
ore characterisation; concept to achieve cost-effective, robust and to design and install circuits
feasibility studies; preliminary to low-risk comminution circuits. that will attain a specified
detailed engineering; plant start- throughput.
up to full commissioning and Ausenco’s assurances to
ramp-up; asset management and the minerals industry for • We commission and provide
optimisation services. comminution circuit design operations support to
and engineering: enable our clients to achieve
The comminution circuit is pivotal optimised performance.
to a project’s success by several • We have developed a
measures: capital cost, operating proprietary, power-based With full-service design and EPC/
cost, energy efficiency, operability, simulation model called EPCM project delivery offices
maintainability and ability to “Ausgrind”, to facilitate circuit that service the minerals and
meet throughput demands in design, equipment sizing and energy industries from major
response to variable ore types selection. cities in Australia, North America,
from the mine. South America and South Africa,
• We are experts in the Ausenco is well positioned to
application of recognised meet the global needs and
ore characterisation challenges of our clients.
methodologies, including
JKDWT, SMC, SPI, Bond,
and Starkey.

www.ausenco.com
Major sponsor profile
JKTech
JKTech Pty Ltd is the technology transfer company for the Sustainable Minerals Institute
(SMI) at The University of Queensland, commercialising research outcomes from the
Centres of the SMI, including the Julius Kruttschnitt Mineral Research Centre.
From JKTech’s formation in 1986, comminution has been the cornerstone of its suite of
product and service offerings, including comminution consulting, JKSimMet simulation
software, breakage characterisation and training. Design and optimisation studies have
been undertaken in hundreds of mineral processing plants around the world, thereby
developing an extensive global database.
JKTech’s consulting services now utilise specialist JKTech software, equipment and
methodologies across comminution, flotation, mineralogy, mining and geometallurgy,
supported by laboratory testing and training courses.
Operating from its Australian head office in Brisbane, JKTech has subsidiary companies
in Chile and South Africa and representatives worldwide.
 Contents

Chapter 1
Comminution – An Overview ....................................................................................... 1

Chapter 2
Mineral Liberation ..........................................................................................................11

Chapter 3
Particle Measurement Techniques............................................................................25

Chapter 4
Ore Comminution Measurement Techniques ......................................................43

Chapter 5
Tumbling Mills..................................................................................................................61

Chapter 6
Compression Machines ................................................................................................79

Chapter 7
High-speed Impact Mills ..............................................................................................99

Chapter 8
Stirred Mills .................................................................................................................... 107

Chapter 9
Mill Liners ........................................................................................................................ 125

Chapter 10
Classifiers ........................................................................................................................ 145

Chapter 11
Comminution Circuits for Ores, Cement and Coal ........................................... 167

Chapter 12
Milling Circuit Calculations....................................................................................... 191

Chapter 13
Modelling Comminution Circuits........................................................................... 215
Chapter 14
Process Control ............................................................................................................. 227

Chapter 15
Case Studies of Control Systems ............................................................................ 245

Chapter 16
Circuit Design ................................................................................................................ 265

Glossary ................................................................................................................................. 303

Index ....................................................................................................................................... 309

 HOME

Chapter 1

COMMINUTION – AN OVERVIEW
Alban Lynch

INTRODUCTION
Comminution is the process of crushing and grinding solid materials into products that
range in size from pebbles to minute particles. It is used in a multitude of industries
worldwide, including high-volume, low-product-cost industries that use ores, coals and
cements, to low-volume, high-product-cost industries that use pigments, confectionery,
cosmetics and pharmaceuticals.
This handbook is concerned with high-volume ores, coals and cements – brittle
materials for which similar machines are used for grinding and sizing classiÀcation.
These materials can be ground dry or as slurries with water. Cement and coal are both
ground dry; however, ores are crushed dry and usually ground wet as this has been
found to be more energy efÀcient. *rinding aids are used in the cement industry to
improve breakage efÀciency by minimising ball coatings. TechniTues such as this,
coupled with the sustainable need to reduce water consumption, has resulted in dry
grinding being reconsidered for ores. The total energy reTuired to grind cement, coal
and ore is high as immense amounts of these materials are used. In 2010, 4 billion tonnes
%t of ore, . %t of cement and  %t of coal were ground to Àne powder in comminution
circuits. These consumed about three per cent of the electrical energy generated
worldwide. Tonnages will continue to rise, and it is expected that by 2040 they will have
increased by 2 per cent above the 2010 value.
6emi-autogenous grinding mills 6$* mills have replaced crushers and primary ball
mills in high-capacity ore plants, and high-pressure grinding rolls +3*5s and vertical
roller mills are used in cement clinker grinding circuits. Comminution machines have
become immense; large 6$* mills are now driven by 2 0: motors and large +3*5s
by 11 0: motor pairs.

CONTEXT
Comminution is usually the most energy-intensive and expensive stage of processing,
but is an indispensable stage to meet product reTuirements. This handbook is designed

Comminution Handbook | Spectrum Series 21 1

 CHAPTER 1 t COMMINUTION – AN OVERVIEW

to aid the selection and operation of the most efÀcient comminution circuit; however,
it assumes that the design engineer is Àrst certain that they are processing the right
material, and to the right size. Before proceeding to comminution design, the engineer
should Àrst check the following to maximise overall efÀciency
• The role of blasting in comminution in this application. Blasting is theoretically
more energy efÀcient at size reduction than crushing or grinding see Table 11.1.
The optimum balance between blasting and downstream comminution will vary
between different ores and different sites, but can have a signiÀcant impact. This
balance should be considered in the early design stages to achieve the most efÀcient
integrated site design. :hile operating sites can subseTuently improve the balance
blasting and comminution often referred to as ¶0ine to 0ill·, and described
in Chapter 11, the beneÀts are limited by the constraints of the already installed
eTuipment for ore handling and comminution. Therefore the beneÀts are likely to be
lower than those from an initial ¶holistic· design.
• The ability to exploit coarse liberation before comminution. :hile most attention
is placed on mineral liberation, site conceptual design should Àrst consider gangue
liberation. If signiÀcant gangue waste can be liberated at a coarse size see for
example )igure 2.4, it is possible to reMect a signiÀcant tonnage before energy-
intensive comminution. The energy for subseTuent mineral liberation can be
signiÀcantly reduced when techniTues such as gravity separation, dense medium
separation, particle sorting or even coarse screening can be used to remove coarse
low-grade waste. These options must be designed in the Áow sheet early since they
dictate the design of ore handling and selection of comminution eTuipment. )or
example, dense medium separation may reTuire a crushing plant and usually cannot
be retroÀt to a large 6$* milling operation.
• :hether ore heterogeneity can be exploited to reMect coarse waste. (merging
developments seek to exploit the natural grade heterogeneity of ore deposits by
reMecting low-grade pockets early in mining operations. The development of grade
sensors for coarse rocks could allow identiÀcation of low-grade intervals of ore
during early ore handling. $lternatively, ¶differential blasting· could be used to blast
low-grade zones coarser and high-grade zones Àner, with coarse low-grade removed
by screening. At the time of writing, these concepts are emerging and undergoing
large-scale site trials, but are yet to be in operation.
The purpose of this handbook is to present a summary of the information available
about the application of comminution to ore, coal and cement, and about the machines
used, which will be useful to engineers who wish to become acTuainted with the
technology.

BACKGROUND
Comminution is an ancient technology dating back to the Stone Age. This can be seen
in 20 000-year-old rock paintings in which pigments were ground Àne to form a durable
surface coating. These could only have been produced by the abrasion and attrition of
coloured minerals such as hematite red, malachite green, limonite yellow and ochres
brown using handstones, activities that had parallels in the ancient methods of grinding
seeds for food preparation. Comminution evolved slowly over several millennia with
emphasis on the more efÀcient use of muscular energy human and animal, the only

2 Comminution Handbook | Spectrum Series 21

 CHAPTER 1 t COMMINUTION – AN OVERVIEW

energy available to break materials. Sources of energy that replaced muscular power –
water, steam, electricity – brought step changes to comminution technology; however,
the demand for energy continued to lead supply. 0achines were invented or modiÀed
to make best use of the limited energy available, and this continued until the latter part
of the 20th century when electricity became plentiful. At this time rapid increases were
occurring in the demand for ores, cements and coals, and ample energy was important in
the rush to build larger machines. A brief chronology of the development of comminution
machines is given in Table 1.1 /ynch and 5owland, 200.

TABLE 1.1
The development of comminution machines.
Era Period Machine
Muscular power Stone Age Mortars and pestles; handstones
2000 Saddlestone mills
Before the Common Era (BCE)
500 Rotary querns for grains
200 Manual edge-roller mills
Common Era (CE)
800 Manual stamp mills for ore
1000 Water-powered rotary querns for grains
Water power
1500 Water-powered stamp mills for ore
1804 Horizontal roller mills (Cornish rolls) for ore
Steam power 1858–1881 Jaw and gyratory crushers for ore
1874–1877 Ball mills; air classifiers for cement
1900 Vertical roller mills for coal
1904 Hydraulic classifiers for ore
1930 Cone crushers for ore
1932 Autogenous mills for ore
Electricity 1948 Hydrocyclones
1960 Semi-autogenous grinding (SAG) mills for ore
1982 High-pressure grinding rolls (HPGR) for cement
1985 High-efficiency air classifiers for cement
1995 High-intensity bead mills for ore

FUNDAMENTALS OF COMMINUTION
Comminution can be deÀned as the reduction of solid materials from a coarser particle
size to a smaller particle size, by crushing, grinding and other processes, and as the
action of reducing a material, especially a mineral ore, to smaller particles or fragments.
In the past it referred to Àner particles rather than coarser pebbles, but it now refers to
both pebbles and particles.
2bMectives of comminution processes for mine waste, ores, cement and coal are to
reduce
• mine waste to a size where it is transportable by conveyor to a deposition area
• coarse rocks to a size at which particles of valuable minerals are liberated and can be
concentrated efÀciently

Comminution Handbook | Spectrum Series 21 3

 CHAPTER 1 t COMMINUTION – AN OVERVIEW

• the pebbles of clinker leaving the cement kiln to a size at which the surface area is
large enough to optimise the reaction of cement with water
• the size of mined coal to particles with surface area that will optimise the burning
rate.
Waste breakage is a special case that is typically carried out in a single stage using a
sizer breaker and is only practised where the waste characteristics and the geography
are suitable. :ith ores and waste, the extraction of rock from deposits by blasting or
excavators is the Àrst stage of breakage. 2re comminution then takes place in a seTuence
of crushing and grinding processes.
Crushing reduces the particle size of run-of-mine 520 ore to such a level that
the grinding mill can produce mineral and gangue as separate or liberated particles.
Crushing occurs by compression of the ore between rigid surfaces, or by impact of rocks
against surfaces. It is dry and is performed in several stages; reduction ratios being small
and ranging from three to six in each stage. The reduction ratio of a crushing stage is
deÀned as the ratio of maximum particle size entering the crusher to the maximum
particle size leaving the crusher. Crushers in use include Maw, gyratory, cone, roll and
impact crushers.
Grinding occurs by compression in roller mills and by abrasion and impact in tumbling
mills when the ore and media are in free motion, such as rods, balls or pebbles. *rinding
is a dry process for cement and coal but usually wet with ores as most concentration
processes are carried out as slurries, although dry grinding does have limited
applications. There is an overlapping range of particle sizes at which it is possible to
choose to crush or grind the ore. At the Àne end of crushing, an eTuivalent reduction
to grinding can be achieved for roughly half the energy and costs reTuired by grinding
)lavel, 1.
Depending upon size and energy considerations, grinding media include tumbling
mills with steel rods, steel balls, or sized ore A* and SA* mills. Stirred or agitated
mills represent the broad category of mills that use a stirrer to provide motion to the
steel, ceramic or rock media. Both vertical and horizontal conÀgurations exist, and since
they can operate with smaller media sizes, they are far more suitable for Àne-grinding
applications than tumbling mills. Stirred mills are thought to be more energy efÀcient
by up to 0 per cent than conventional ball mills Stief, /awruk and :ilson, 1.
They are now widely used for Àne comminution.

Principles
The initiation and propagation of cracks in rock occurs at the atomic level and was
described over 100 years ago Inglis, 11. Similarly, the energy consumed in breakage
occurs when rock fabric is disrupted at the atomic scale. Cracks propagate rapidly
through rock as the stress front moves through the matrix ahead of the crack tip. At the
macro level, rock breakage is achieved by compression, impact and attrition, as has been
discussed.
:hen a particle is broken by compression the products fall into two distinct size ranges
coarse particles, resulting from induced tensile failure, and Ànes from compressive failure
near the points of loading. 0inimising the area of loading can reduce the amount of Ànes
produced. This is often done in compressive crushing machines by using corrugated
crushing surfaces 3artridge, 1. In impact breaking, due to the rapid loading, a

4 Comminution Handbook | Spectrum Series 21

 CHAPTER 1 t COMMINUTION – AN OVERVIEW

particle experiences a higher average stress while undergoing strain than is necessary
to achieve a simple fracture, and tends to break apart rapidly, mainly by tensile failure.
The products are often very similar in size and shape. Abrasion produces particles that
are much smaller than the rocks involved; these rocks are rounded as rougher edges are
smoothed off. /ong-term abrasion can occur when pebbles cannot get out of a SA* mill.
This results in smooth, even polished surfaces.

Theory
Comminution theory focuses on the relationship between energy input and the particle
size produced from a given feed size. The greatest problem is that the machine itself
absorbs most of the energy input to a crushing or grinding machine, and only a small
fraction of the total energy is available for breaking the material. There is a relationship
between energy reTuired for breaking the material and the new surface area produced
in the process; however, this relationship can only be explained if the energy consumed
in creating the new surface can be separately measured. All theories of comminution
assume that the material is brittle, so that no energy is adsorbed in processes such as
elongation or contraction, which is not Ànally utilised in breakage.
n
The general eTuation E = - J.dx/x , in which the energy used is related to change in
particle size, describes the process. Three versions of this eTuation are discussed
1. 9on 5ittinger 1 proposed that
1 - 1
E = k. ` j
x2 x1
ie the energy consumed is proportional to the new surface area produced. E is the
energy input, x1 is the initial particle size, x2 is the Ànal particle size and k is a constant.

2. )riedrich .ick 1 proposed that

x1
E = k.ln ` x j
2
ie the energy consumed is proportional to the reduction achieved in volume of the
particles.

. )red Bond 12 proposed that

1 - 1
E = 2.k. c m
x2 x1
These three eTuations are obtained by substituting 2.0, 1.0 and 1. for n in the general
eTuation. +ukki 11 suggested that the theories could be applied to different parts of
the size reduction curve )igure 1.1.
For practical calculations the criterion of particle size is the size in microns that
0 per cent passes, where
P is the product size
F is the feed size
W is the work input in kilowatt hours per metric tonne
The Bond Ball 0ill :ork Index BB0:i is deÀned as the kilowatt-hours per metric
tonne reTuired to reduce material from theoretically inÀnite feed size to 0 per cent
passing 100 microns. Bond·s eTuation is usually written as

Comminution Handbook | Spectrum Series 21 5

 CHAPTER 1 t COMMINUTION – AN OVERVIEW

ϭϬϲ

ϭϬϱ

ϭϬϰ ZŝƚƚŝŶŐĞƌ͕
ŶĞƌŐLJ/ŶƉƵƚ;ŬtŚͬƚͿ ŶсϮ 'ƌŝŶĚŝŶŐ
ϭϬϯ
ƌƵƐŚŝŶŐ
ϭϬϮ

ϭϬ ŽŶĚ͕ Ŷсϭ͘ϱ

ϭ
<ŝĐŬ͕ Ŷсϭ
ϭϬͲϭ
ϭϬͲϯ ϭϬͲϮ ϭϬͲϭ ϭ ϭϬ ϭϬϮ ϭϬϯ ϭϬϰ ϭϬϱ ϭϬϲ ϭϬϳ
WĂƌƚŝĐůĞ^ŝnjĞ;ŵŝĐƌŽŶͿ

FIG 1.1 – How the theories of Rittinger, Kick and Bond compare with operating practice (image courtesy Hukki, 1961).

10.Wi - 10.Wi
W =c m
P F
where
Wi is the :ork Index

Grindability
Ore grindability refers to the ease with which materials can be comminuted, and data
from grindability tests are used to evaluate crushing and grinding efÀciency. The most
widely used parameter to measure ore grindability is the BB0:i, which is explored
further in Chapter 4. A variation of this reTuires the use of a reference ore of known
grindability (r that is ground for a known time and the power consumption is recorded.
An identical weight of the test ore (t is then ground for a length of time such that the
power consumed is identical with that of the reference ore (r. If t is the ore under test,
then from Bond·s eTuation
10.Wir - 10.Wir 10.Wit - 10.Wit
c m=c m
Pr Fr Pt Ft

The sizing distributions and Wir are known so Wit can be calculated (Berry and Bruce,
1. :hile the Bond test is the best-known grindability test for rod and ball mills, other
grindability tests have been developed that are suitable for the size reduction techniTue
being used. Some of these include
• the Hardgrove test for coal, which uses balls rotating in a groove
• the Drop :eight Test for ores, which uses data from breakage tests carried out with
known impacts
• the SA* 3ower Index (S3I® test, which is based on the Bond test and uses data from
tests carried out in a batch tumbling mill
• a compression grindability test (see Chapter 4.
Testing for grindability Àne feed sizes, typically less than 0. mm, is more difÀcult and
often controversial. 0ore commonly, this work is carried out in a small version of the
Ànal machine, with power being measured directly during the test.

6 Comminution Handbook | Spectrum Series 21

 CHAPTER 1 t COMMINUTION – AN OVERVIEW

BREAKAGE MECHANISMS
Comminution machines can be classiÀed according to the mechanisms working inside
the machine. Particles are broken by different mechanisms of impact, compression,
attrition and abrasion. A schematic view of the mechanisms is shown in Figure 1.2.


 
Impact Abrasion/shear Attrition Compression

FIG 1.2 – Mechanisms by which particles are broken (image courtesy Hakan Dundar).

In most cases more than one breakage mechanism occurs inside the mill, with one
generally being dominant. Impact and compression are more effective for coarse
particles, while attrition and abrasion (or shear are effective methods for Àne particles.
Impact breakage – this refers to the size reduction of a particle achieved by forces acting
on the particle, resulting from the kinetic energy of the grinding media.
Abrasion (or shear) – this occurs when particles rub against each other and the shear
forces create Àner particles. *rinding media can also promote abrasion by rubbing.
Attrition – similar to abrasion, attrition occurs when a smaller particle gets trapped
between larger particles. As the particle edges are chipped and abrasion occurs, the
combination is called attrition.
Compression breakage – breakage by applying compression to a bed of particles is the
most efÀcient method of size reduction in terms of energy utilised. (nergy is transferred
to particles by means of surrounding particles, hence interparticle breakage occurs.
COassiÀcation (si]ing) – in continuous comminution machines smaller particles
are discharged and larger particles are retained for further breakage. This internal
classiÀcation process should not be confused with external classiÀcation in which Àne
particles are separated from coarse particles after they have been discharged from the
mill and the coarse fraction is returned to the mill for further grinding. Other mechanisms
occurring inside mills include disagglomeration of clays and weakly bound particle
aggregates, and delamination of platy materials such as mica and graphite.

CLASSIFICATION MECHANISMS
ClassiÀcation in mineral processing circuits refers to the continuous separation of
particles using techniTues that exploit differences in shape, speciÀc gravity or size. The
mechanisms used for separation are screening and hydraulic and pneumatic separation.

Screening
Screening is the historical method of separating particles on the basis of size. It involves
presenting a stream of particles in either a dry form or as a slurry to a vibrating surface

Comminution Handbook | Spectrum Series 21 7

 CHAPTER 1 t COMMINUTION – AN OVERVIEW

containing apertures through which only the smaller particles can pass. :hen used
in a closed grinding circuit the small particles are the product and the larger particles
are recycled for further breakage. Screening is more efÀcient with slurries than with
dry particles; however, conventional vibrating screens have not been efÀcient when
aperture sizes are less than about 200 microns. High-freTuency vibrating mechanisms
enable screens with aperture sizes as Àne as 40 microns to operate efÀciently in plants.

Hydraulic and pneumatic separation

Smaller or lighter particles fall more slowly through air or water than larger or heavier
particles. Based on this assumption, several machines were devised. The original water
classiÀer was the 5ittinger classiÀer (1 in which slurry Áowed horizontally across
settling boxes of increasing size, with the coarsest particles settling in the Àrst box and
Àner products in later boxes. This was converted into a continuous machine by building a
tank with a sloping base into which feed slurry entered. The slurry overÁowing the tank
contained Àne particles and the coarse particles settled on the base and were dragged
up the slope and discharged. By the late 1940s better pumps were used and centrifugal
force replaced gravity. Slurry was pumped tangentially into the top of a vertical pipe
and two spirals were formed an outer spiral moving down the wall in which the coarser
particles were held, and an inner spiral into which the smaller particles were displaced.
The outer spiral left the cyclone at its base and the inner spiral left through a central pipe
at the top.
In the original air classiÀer for dry separation the feed fell onto a horizontal spinning
disk and was thrown into a thin stream of particles. A rising airstream swept the Àne
particles from the stream out of the unit while the coarse particles fell into the base for
discharge. This principle still applies; however, the separating mechanism is now more
efÀcient. ClassiÀers are discussed further in Chapter 10.

COMMINUTION MACHINES
0any of the machines now used in comminution circuits were invented in the late
19th century because by then water wheels were large enough to provide sufÀcient,
continuous power. Jaw and gyratory crushers, vertical roller mills, ball mills, continuous
classiÀers using both sedimentation and cyclonic separation techniTues were in use by
the early 20th century. Today they are much larger and suit a wide variety of purposes,
including Àne grinding.
This handbook is concerned with ore, cement and coal, all of which have common
processing characteristics
• they are brittle
• they are processed at 4 Bta and are ground to approximately 0 per cent passing
0 microns for further use
• similar eTuipment is used for grinding and classiÀcation whether processes are wet
or dry
• eTuipment size varies from small to immense.
0achines that will be discussed in this handbook are listed in Table 1.2 and include
impact mills, sizers, low-speed and high-speed crushers, tumbling mills, stirred mills,
roller mills, Àne crushers and cyclonic and screen separators.

8 Comminution Handbook | Spectrum Series 21

 CHAPTER 1 t COMMINUTION – AN OVERVIEW

TABLE 1.2
Comminution machines.
Mill type Breakage Reduction F30 to P30 Maximum power (kW) Maximum mill size
Tumbling mills
Rod mill 30 mm – 500 μm
Ball mill – ore
Primary 15 mm – 70 μm 22 000 8.5 m × 13.4 m
Regrind Impact/abrasion 100 μm – 30 μm 740 3.2 m × 4.8 m
Ball mill – cement 50 mm – 70 μm 11 500 5.8 m × 19 m
Ball mill – coal 50 mm – 90 μm 2760 5.5 m × 8.2 m
AG/SAG mills 400 mm – 75 μm 28 000 12.2 m × 8.8 m
Stirred mills
TowerMill – Eirich 3 mm – 20 μm 1120
Vertimill – Metso 6 mm – 20 μm 2300
HIGmill – Outotec 300 μm – 20 μm 5000
Attrition
SMD – Metso 200 μm – 5 μm 1100
IsaMill – Xstrata 500 μm – 10 μm 8000
EiSA mill – Netzsch 250 μm – 10 μm 710
Roller mills
Roller press (HPGR) 80 mm – 40% <1 mm 11 500
Vertical Compression 200 mm – 80% <40 μm 11 000
Horizontal 38 mm – 1 mm 15
High speed crushers
Barmac – Metso 30 mm 600 60 m/s rock speed
MAG’Impact® –
130 mm 400 70 m/s rock speed
Magotteaux
Impact/attrition
Hammer mill 25:1 reduction 1800 75 m/s tip speed
Hammers 52 m/s
Atrita mill – Riley power 75 mm – 35 μm 670
Mill clips 100 m/s
Primary crushers
Gyratory 150–250 mm discharge 1200 5000–9000 mt/h
Compression
Jaw 100–300 mm CSS 400 700–1400 mt/h
Cone crushers
Standard 350–35 mm 1500 2350–2900 mt/h
Shorthead 100–6 mm 1500 1000–1400 mt/h
Compression
Rhodax – Fives FCB 150 mm 900 600 Mt/h
Vibrocone – Sandvik 150 mm 400 400
AG/SAG – autogenous/semi-autogenous grinding; SMD – Stirred Media Detritor; HPGR – high-pressure grinding rolls.

The energy efÀciency of machines varies widely, with compression machines being
the most efÀcient. Approximate values that have been suggested are (Fuerstenau and
Han, 200

Comminution Handbook | Spectrum Series 21 9

 CHAPTER 1 t COMMINUTION – AN OVERVIEW

• autogenous grinding (A*semi-autogenous grinding (SA* mills – three per cent
• primary ball mills – Àve per cent
• high-pressure grinding rolls (HP*5 – 2 per cent
• cone crushers – 0 per cent
• gyratory crushers – 0 per cent.
Production rates may vary widely for the same machine as they depend on the material
processed and the reduction ratio achieved.
Power intensity (k:m can be applied to mills with enclosed breakage zones.
Approximate values are
• Isa0ill – 00
• high-intensity grinding (HI* mill – 200
• Stirred 0edia Detritor (S0D – 10
• hammer mills – 20
• tower mills – 40
• tumbling mills – 21.

BIBLIOGRAPHY
Berry, T F and Bruce, 5 0, 19. A simple method of determining the grindability of ores, Can
Min J, July.
Bond, F C, 192. The third theory of comminution, Trans AIME, 1944.
Flavel, 0 D, 19. Control of crushing circuits will reduce capital and operating costs, Min Mag,
0arch20.
Fuerstenau, 0 C and Han, . 1 (eds, 200. Principles of Mineral Processing,  p (Society for
0ining, 0etallurgy, and (xploration /ittleton.
Hukki, 5 T, 191. Proposal for a Solomonic settlement between the theories of von 5ittinger,
Kick and Bond, Trans AIME (Mining), pp 40–410.
Inglis, C (, 191. Stresses in a plate due to the presence of cracks and sharp comers, Proc Inst Nav
Arch, /9219–20.
Kick, F, 1. 'es Geset] der Proportionalem Ziderstand und 6eine AnZendung (Principle of
Proportional 5esistance and Its Application, (Felix /eipzig, (in *erman.
Lynch, A J and 5owland, C A, 200. The History of Grinding (Society of 0ining, 0etallurgy,
and (xploration /ittleton.
Partridge, A C, 19. Principles of comminution, Mine and Quarry, (JulAug0.
Stief, D (, /awruk, : A and :ilson, / J, 19. Tower mill and its application to Àne grinding,
Min Metall Proc, 4(14–0.
Von Rittinger, P 5, 1. Lehrbuch der Aufbereitungs Kunde ((rnst and Korn Berlin, (in *erman.
Wills, B A and 1apier-0unn, T J, 200. Wills Mineral Processing Technology: An Introduction to the
Practical Aspects of Ore Treatment and Mineral Recovery, seventh edition ((lsevier Amsterdam.

10 Comminution Handbook | Spectrum Series 21

 HOME

Chapter 2

MINERAL LIBERATION
Cathy Evans

INTRODUCTION
The purpose of comminution is to break rock into smaller fragments as part of a process
for which the ultimate aim is to create a product to sell. Considerable amounts of energy
and capital equipment are employed to achieve this in the minerals industry. In some
industrial applications, such as aggregate production, the objective of the comminution
processes is the si]e reduction of rock to a speciÀed particle si]e to create a saleable
product. In mining and mineral processing, the purpose of the overall process is to
produce a saleable mineral concentrate, which, as the name ‘concentrate’ implies, is
higher in grade than the ore being fed into the concentrator.
In mineral processing operations, the purpose of the comminution circuit is to prepare
the ore as a suitable feed for separation processes that upgrade the material by rejecting
the particles that do not contain economic amounts of the target mineral. The implication
is that comminution changes the ore from a population of particles with a relatively
uniform grade (similar to the feed ore), to particles with a range of compositions that
allow them to be separated by composition into high-grade and low-grade streams.
Mineral liberation is the name of this phenomenon, where parent particles with a given
mineral composition break into progeny particles with a range of mineral compositions.
A simple representation of mineral liberation during breakage is shown in Figure 2.1.
As Figure 2.1 demonstrates, a single breakage event can break a particle that consists
of two minerals into progeny particles that vary both in size and mineral composition.
The composition of the progeny can vary from 100 per cent valuable mineral to
100 per cent gangue (that is, waste) mineral. The characteristics of these progeny depend
on factors such as which minerals are present in the ore, their physical properties
and how the minerals are arranged relative to one another; for example, size, shape
and spatial arrangement of the mineral grains, also described as the ore texture and
determined by mineralogical studies.
This chapter describes the aspects of mineral liberation that are of importance in
mineral processing operations.

Comminution Handbook | Spectrum Series 21 11

 CHAPTER 2 t MINERAL LIBERATION

FIG 2.1 – A simple representation of mineral liberation during breakage (image courtesy Cathy Evans).

Liberation terminology
As in any technical Àeld, the world of mineral liberation has a vocabulary that uses
common words in very speciÀc ways. Table 2.1 deÀnes some common terms used in
describing mineral liberation.

TABLE 2.1
Some common terms used in mineral liberation (images courtesy Cathy Evans).
Liberated particles
Contain only one mineral
Grade = 100%
Simplest texture

Composite particles

Contain two or more minerals. If two minerals are present, particle may be
described as ‘locked binary’; if three or minerals are present, may be described
as ‘locked ternary’
0% < grade of mineral of interest <100%
Texture of particle varies depending on relative scale of particle size and original
texture in ore

WHY IS LIBERATION NECESSARY?

To produce a saleable product, most ores have to undergo two types of processes:
1. comminution to physically separate (that is, liberate) the minerals from one another
by breaking the ore particles
2. separation to concentrate the mineral-bearing particles into a saleable product and a
waste stream.

12 Comminution Handbook | Spectrum Series 21

 CHAPTER 2 t MINERAL LIBERATION

Comminution circuits are often the most energy-intensive part of a concentrator. There
is usually a balance that must be struck between grinding the feed Ànely enough to
liberate the minerals and the energy required to achieve this. In all cases, knowing what
liberation characteristics are required is an important part of determining where this
balance lies. Typically, multiple stages of comminution and separation processes are
employed to reject liberated waste minerals from the process as early as possible to
minimise grinding energy requirements. This strategy is discussed in more detail later
in the chapter.
The effectiveness of the separation process depends not just on the mineral composition
distribution of the feed particles generated through breakage and liberation, but also on
the mineral texture of the particles. If a particle is fully liberated, it will be recovered in
separation processes that use surface properties, particle density or chemical leaching
to recover the valuable mineral; however, for some ores – such as low-grade copper
porphyry ores – it is not cost-effective to grind the entire concentrator feed stream to
achieve complete liberation of the target mineral. Therefore, a signiÀcant proportion of
this valuable mineral is present in composite particles. Mineral processing engineers
need to understand how these unliberated particles, which can vary in both composition
(that is, grade of each mineral) and texture, behave in different types of separation
processes.
To illustrate the importance of mineral texture an example is provided in Table 2.2,
where three unliberated particles with quite different textures are shown in cross-
section (as would be measured by an optical microscope or automated mineralogy
system). Although the three particle types contain similar amounts of valuable mineral
(shown as the dark phase), Table 2.2 shows how the recovery of this valuable phase in
a range of mineral separation processes depends on the texture of each particle, not just
its composition. Clearly the design or optimisation of a comminution circuit to generate
a product with an appropriate level of liberation must consider what type of separation
process will be used to separate the ore.
Note that most separation processes do not require the feed stream to consist of
liberated mineral particles. In Áotation, for example, particles with as little as ten per cent
of the mineral on the particle surface can be recovered (Sutherland, 1989). Heap and
column leaching is a special case in which the feed preparation is designed to generate
relatively large particles, with intensive microfractures that allow the leach solution to
reach the surfaces of the target mineral grains. In this special case, the liberation does not

TABLE 2.2
The effect of particle texture on the recovery of unliberated minerals (image courtesy Cathy Evans).
Which unliberated particle texture allows the valuable phase to be recovered?
(All particles are same grade)
Typical particles of
valuable (dark phase)
and gangue (light phase)
shown in cross-section
By flotation DRecovered Not recovered Not recovered
By density separation DRecovered DRecovered DRecovered
By leaching DRecovered Not recovered Not recovered

Comminution Handbook | Spectrum Series 21 13

 CHAPTER 2 t MINERAL LIBERATION

physically divide the grains into separate particles; the microfractures are sufÀcient for
the chosen chemical concentration process to be effective.
:hile mineral liberation is one of the key ore characteristics considered during the Áow
sheet design stage, either explicitly with liberation measurements of the ore or implicitly
in combined grinding/separation tests, it is not often considered in the daily operation
of mineral processing plants. Concentrators generally grind the feed to a selected particle
size distribution because size can be measured easily (including online if required), while
the liberation distribution of the feed cannot. As a result, the importance of measuring
the liberation characteristics of ores in operations tends to be forgotten. Now, with the
increasing availability of liberation distribution data through measurement systems
such as the scanning electron microscope (SEM)-based automated mineralogy systems,
mineral processors are able to conÀrm whether required size and liberation distributions
are being achieved based on quantitative data. Similarly, process optimisation can be
performed in a holistic way, simulating the various parts of the process chain as an
integrated system that is linked by information about the composition of the particles in
the process streams (Wiegel, 1976; Evans et al, 2011, 2013).

HOW DOES LIBERATION HAPPEN?

Liberation may start as soon as an ore is broken into particles, whether by blasting in
the mine or crushing/grinding in the concentrator. The particle size at which liberation
occurs will vary depending on the size of the mineral texture in the ore, and may differ
for different minerals or groups of minerals within the same ore.
As described in Chapter 1, the breakage that takes place in comminution occurs as a
result of fracturing within the rock being processed. These fractures propagate through
the mineral texture of the rock, creating progeny particles and microcracks within these
particles. In some ores the path taken by the fractures is independent of the characteristics
of the rock (for example, mineral type, mineral texture in the rock, degree of alteration),
and this is termed ‘random breakage’. Where random breakage is present, the liberation
of a mineral occurs once the particles are the same size as the grains of that mineral, and
increases as particles become Àner than the original grain size (Wiegel, 2006).
In many ores, the path followed by the fractures during comminution is inÁuenced
by the type of minerals present in the rock and the size, shape and spatial arrangement
of the mineral grains (that is, the ore texture). This behaviour is termed ‘non-random
breakage’. Many of the existing process models of liberation during comminution
assume that breakage is random; however, in practice there is often some degree of non-
random breakage that contributes to liberation. In practice, six types of non-random
fracture processes have been described (King and Schneider, 1998), these being:
1. selective breakage of minerals with different brittleness characteristics
2. preferential breakage where crack branching occurs more frequently in one mineral
3. breakage around grain boundaries
4. liberation of grains by detachment
5. differential breakage where the size breakage function is dependent on the
composition and texture of the parent particle
6. boundary-region fracture where contacts between minerals are highly stressed,
causing preferential breakage in this region.

14 Comminution Handbook | Spectrum Series 21

 CHAPTER 2 t MINERAL LIBERATION

One or more of these non-random fracture processes can affect the liberation
occurring during breakage and therefore the particle composition distribution of the
progeny particles.
Some comminution devices have been observed to promote non-random breakage of
the target mineral in the comminution product with subsequent beneÀts to separation.
At the industrial scale, high-pressure grinding rolls (HPGR) have been shown to
enhance breakage around mineral grain boundaries. As a result, operators have seen
improved gold and copper recovery from leach circuits and Áotation circuits ('unne,
Goulsbra and 'unlop, 1996). The diamond industry has also exploited the grain
boundary breakage of HPGRs to recover large diamonds in ‘a single comminution step,
where all diamonds of all sizes are liberated and preserved’ ('aniel and Morley, 2010).
Researchers have identiÀed other technologies that enhance liberation by promoting
non-random breakage. These include fragmentation by high voltage electrical pulses
(Andres, 1977; Wang, Shi and Manlapig, 2012) and thermal pre-treatment (Veasey and
Fitzgibbon, 1990), both of which enhance liberation by promoting fracturing around
mineral grain boundaries; however, to date neither of these technologies has made the
transition to full-scale industrial adoption.
In designing process Áow sheets, mineral processing engineers commonly seek
to minimise capital and operating costs by rejecting waste minerals as early in the
process route as possible. Process designers exploit the fact that the particle size at
which liberation occurs may differ for different minerals or groups of minerals within
the same ore. This is typically achieved by using a primary grinding stage that grinds
the ore just Àne enough to create liberated gangue particles, and composite particles
that contain enough of the target minerals to be recovered in the primary (or rougher)
separation process.
The concentrate from this Àrst separation stage passes to a further grinding stage
(typically called regrinding); here this smaller tonnage of solids is ground to the Àner
particle size required to liberate the target minerals to the stage where a saleable
concentrate grade can be achieved in the Ànal (or cleaner) separation process. This
‘liberate–separate–liberate–separate’ strategy is shown schematically in the simpliÀed
concentrator Áow sheet (Figure 2.2), where a relatively coarse primary grinding product
is produced, allowing a signiÀcant proportion of the feed mass to be rejected to Ànal
tail in the rougher Áotation section. The rougher Áotation section is operated to obtain
high recovery of the target mineral and the resulting lower grade rougher concentrate
must be ground further to create particles that are of sufÀciently high grade to meet
the Ànal concentrate grade speciÀcations. (Note that the separation processes shown in
the example are 100 per cent efÀcient, a situation that is never true in real operations.)
The beneÀts of applying this strategy can be seen in the example of low-grade porphyry
copper sulÀde ores where, at typical feed grades of less than 1.0 per cent copper, more
than 90 per cent of the feed mass must be rejected to Ànal tail. A further example of this
approach is the platinum industry in South Africa, where several grind-separate stages
may be employed to minimise the overgrinding of gangue minerals and maximise the
recovery of platinum-bearing minerals.
Clearly the strategy shown in Figure 2.2 requires much less energy and is more cost-
effective than the alternative of grinding 100 per cent of the concentrator feed to the size
required to liberate the target minerals to the level where a high-grade Ànal concentrate
can be produced.

Comminution Handbook | Spectrum Series 21 15

 CHAPTER 2 t MINERAL LIBERATION

FIG 2.2 – A concentrator flow sheet showing multiple stages of liberation-separation (image courtesy Cathy Evans).

Another example of how process design exploits liberation at different size scales can
be seen in the Mount Isa lead-zinc ore, which undergoes two distinct stages of liberation,
allowing a preconcentration stage to be deployed before the ore is ground. The texture of
the ore at the scale of centimetres can be seen in Figure 2.3, where the darkest bands are
gangue minerals and the lighter-coloured bands contain the valuable sulÀde minerals
galena and sphalerite.
The process designed to treat this ore crushes the material to the size at which the dark
waste rock is liberated (approximately 13 mm). It then uses density separation to reject
the lower-density liberated waste rock, removing approximately 30–35 per cent of the
feed mass at this early stage in the size reduction process with relatively little loss of
valuable minerals (Munro and Pease, 1988).
The product of this preconcentration stage goes on to a conventional grinding circuit
that grinds the feed to an 80 per cent Àner than 80 —m to liberate the galena, sphalerite
and intergrown waste minerals prior to Áotation separation. The installation of the
preconcentration stage in 1982 increased production capacity of the Mount Isa lead-zinc
concentrator from 2.5 to 4.2 Mt/a (Black and Mutton, 1986), and reduced the energy
intensity (energy consumed per tonne of metal produced) of the operation.

16 Comminution Handbook | Spectrum Series 21

 CHAPTER 2 t MINERAL LIBERATION

FIG 2.3 – An example of ore texture offering mineral liberation opportunities

at coarse particle size (image courtesy Cathy Evans).

MEASURING LIBERATION
A number of techniques are available to measure mineral liberation and particle
composition distributions and these are now brieÁy described.

Bulk laboratory techniques

Where the ore properties are suitable, laboratory separation techniques such as magnetic
separation and density separation can be used together with assays to quantify the
liberation characteristics of an ore. Magnetic separation is used for assessing liberation
in iron ores. These bulk techniques have the advantage of better sample representivity
compared with optical microscopy, since they measure larger numbers of particles in
the bulk samples that are analysed. Additionally, these bulk techniques measure the
particle composition information for the entire particle volume and therefore provide
the data as volume percentage or weight percentage of the mineral of interest.

Optical microscopy
Optical microscopy has been used for many years as a standard method to quantify
mineral proportions and mineral liberation (Jones, 1987). Generally quantitative
measurements are made on polished sections of ore particles. In the early days of
optical microscopy the data was collected by an experienced operator inspecting the
sample visually. The data collected was restricted to assessments of the proportion
of liberated mineral and the proportion locked in binary and ternary particles with
other minerals (Amstutz, 1961). In recent years the development of automated digital
optical microscopes has allowed data collection to be automated. The digital images
collected by these systems can be analysed using image analysis techniques to obtain
more detailed levels of quantitative data than the previous manual point count and
area count techniques could provide.

Comminution Handbook | Spectrum Series 21 17

 CHAPTER 2 t MINERAL LIBERATION

Automated mineralogy systems

Over the past 30 years, automated mineralogy systems based on SEM have been
developed. These systems use automated image acquisition and automated image
analysis.
It is interesting to note that the two dominant systems were both developed in
Australia: the QEM*SEM (now FEI QEMSCAN, short for Quantitative Evaluation
of Minerals by SCANning electron microscopy) developed by CSIRO (Reid and
Zuiderwyk, 1983) and the JKMRC MLA (now FEI MLA, short for Mineral Liberation
Analyser) developed by the Julius Kruttschnitt Mineral Research Centre (JKMRC)
(Gu, 2003). Current systems use specialised software to perform image acquisition
on SEM hardware, followed by image analysis to collect a wide range of detailed
quantitative data for the sample analysed, including mineral liberation data. Some
examples of liberation data generated by these systems and ways in which data are
presented to the user are shown in Figure 2.4. The other useful information generated
by these systems is the colour-coded particle images, which provide a qualitative
visual impression of the data that can assist in interpreting the tabulated data. Some
examples of these particle images are shown back in Table 2.1.

FIG 2.4 – Mineral liberation data examples from an automated mineralogy system (image courtesy Cathy Evans).

One advantage of these SEM-based systems is that mineral identiÀcation is based on

elemental composition data obtained from X-ray spectroscopic analysis of each mineral
grain. These automated systems also have the advantage over manual systems in that
they can operate 24 hours per day, thus analysing many tens of thousands of particles,
improving the statistical representation of the data.
As in optical microscopy, measurements made in automated mineralogy systems
are performed on polished sections of a sample. The use of data collected from two-
dimensional images (that is, areal data) to estimate the characteristics of three-
dimensional objects, such as particles and mineral grains, is subject to stereological
error (also known as stereological bias). In the case of liberation measurements this
error results in an overestimation of the amount of liberated mineral. Mathematical
techniques are available to correct for the stereological error (Barbery, 1991; Lin, Lastra
and Finch, 1999) where required; however, in practice many users have found that the

18 Comminution Handbook | Spectrum Series 21

 CHAPTER 2 t MINERAL LIBERATION

error does not signiÀcantly affect the usefulness of the data. A common approach to
minimise the stereological error is to measure particulate samples in narrowly sized
fractions and then arithmetically combine the results to estimate the overall particle
composition distribution for the unsized sample.

X-ray tomography
The increasing application of X-ray tomography in mineral processing (Lin and Miller,
1996; Evans et al, 2013) permits particle composition and texture to be measured in
three dimensions, generating results that are not affected by stereological bias. These
systems are becoming more common in research and development centres but have
not been installed in industrial sites to date. The mineral discrimination capabilities
of X-ray tomography systems rely on the minerals of interest exhibiting different
X-ray attenuation values. Whether X-ray tomography can quantify liberation for a
given ore depends on the minerals present in the ore. Some mineral pairs require
special measurement techniques to distinguish between them on the basis of X-ray
attenuation alone; an example here is chalcopyrite and magnetite grains, which require
measurements at two energy levels to differentiate between them. Technical advances
over the next few years – including the addition of high-resolution X-ray Áuorescence
analyses – will improve the mineral discrimination and make these systems useful
tools for measuring mineral liberation.

MODELLING MINERAL LIBERATION IN COMMINUTION PROCESSES

The need for a process model of liberation in comminution has been recognised for many
years. While simulation capabilities to predict particle size distributions in comminution
have developed to the point where simulation is commonly used to design and optimise
circuits, liberation models have not developed to the same advanced level.

Heuristic models
In parallel with efforts to develop mathematical models of liberation in comminution,
mineral processing engineers have sought heuristic models (also known as ‘rules of
thumb’) to model liberation in comminution. One simple but powerful heuristic is that
the particle composition distribution in a given size fraction is the same, regardless of the
overall size distribution of the stream from which the sample comes. For example, this
characteristic means that the particle composition distribution in a given size fraction is
the same in both the feed to a comminution device and in its product. This heuristic was
Àrst identiÀed in the literature by Bprubp and Marchand (1984) working with iron ore
but it has also been shown to apply to a wide range of other ore types, including copper
ores and lead-zinc ores. The heuristic also applies irrespective of whether the breakage
device uses impact or compression breakage. The robust nature of the heuristic approach
has seen it used to model liberation in a number of applications, including gold ore and
nickel-copper sulÀde ore (Evans et al, 2011).

Mathematical models of mineral liberation in comminution

While most models of size reduction in milling consider only the breakage of a
homogeneous material, comminution models that include the effect of size reduction
on mineral liberation must consider the breakage of at least two distinct mineral phases,

Comminution Handbook | Spectrum Series 21 19

 CHAPTER 2 t MINERAL LIBERATION

thus making the material heterogeneous. This heterogeneity adds to the complexity of
the modelling problem in two ways:
1. it must now cater for the breakage of two phases, which may have different breakage
and selection functions
2. it must also represent the liberation distribution of particles formed when the
unliberated parent particles are broken into both liberated progeny of each phase
and unliberated progeny with a range of compositions.
Figure 2.1 illustrates the breakage of just three classes of particles in a two-phase ore,
with liberated particles of each phase and one composite class of parent particle containing
both phases. This is a simple case where the challenge for researchers has been to develop
a model that describes the breakage from many composition classes of composite parent
particles into many composition classes of composite progeny particles.
The problem has been described mathematically by Andrews and Mika (1975) and
summarised by King and Schneider (1998) in the diagram shown in Figure 2.5. The
‘attainable region’ in this diagram shows the range of size and composition of progeny
particles it is possible to generate in breaking a parent particle of known mass and
composition denoted by point A. Conversely, the ‘feeder region’ shows the possible
range of size and composition of parent particles that could be the source of a given
progeny particle of known mass and composition denoted by point A.

FIG 2.5 – Diagram showing the relationships possible between the mass and assay
of parent particles and progeny particles (King and Schneider, 1998).

The complexity of modelling breakage and liberation is further increased by the fact
that the breakage of the ore may be random or non-random. In random breakage,
the fracture pattern of a particle is independent of its mineralogical composition and
texture. If an ore does break randomly, useful simpliÀcations can be applied to many
of the published models to obtain a solution. Indeed, many models of liberation in
comminution rely on the assumption of random breakage for their validity. Other
models – such as King and Schneider’s population balance model of liberation in
batch grinding (1998) – do not require any assumptions about random or non-random

20 Comminution Handbook | Spectrum Series 21

 CHAPTER 2 t MINERAL LIBERATION

breakage occurring, but in practice the calculations required to apply the model are
greatly simpliÀed if breakage is assumed to be random.
As mentioned, King and Schneider (1998) summarised six different types of non-
random breakage and of these, preferential breakage is a common phenomenon that
results in the appearance of the preferentially broken phase in the Àner size fractions
after breakage. This can be detected in comminution products by assaying size fractions;
often the sulÀde mineral grade of Àner size fractions is higher than the coarse size
fractions, indicating the presence of non-random breakage. Many (if not most) real ores
and breakage processes exhibit some degree of non-random breakage and the challenge
has been to account for non-random breakage behaviours in modelling liberation.
The mathematical function that describes the transfer of mass from the size and
composition classes of the parent particles to the size and composition classes of
the progeny particles is known as a kernel. In recent times the application of X-ray
tomography in mineral processing has opened up the opportunity to calibrate the
kernel for each ore by measuring the parent particles and their progeny particles in three
dimensions (Bajic, 2014). This, combined with a renewed focus on integrated simulation
capabilities (Evans et al, 2013), may lead to further advances in liberation modelling.

MINERAL LIBERATION IN SEPARATION PROCESSES

As stated previously in this chapter, the overall purpose of the mineral processing
operation is to produce a saleable product. In the case of concentration processes that
physically separate the particles, such as Áotation and density, magnetic and electrostatic
separation, this saleable product is a mineral concentrate for which the composition is
largely one mineral type. The ability to produce a concentrate of saleable grade is closely
linked to the liberation distribution of the feed to the separation process.
A useful technique to assess whether it is possible to create a concentrate of the
required grade from a feed stream with a speciÀc size distribution is to view the
mineralogically limited grade-recovery curve for the stream, as generated by many
automated mineralogy systems. The mineralogically limited grade-recovery data for an
ore represent the theoretical maximum separability of the particles in a perfect separator,
indicating the best possible concentrate grade and recovery that can be achieved for
the physical constraints imposed by the liberation status of the ore at that particle size
distribution. The example curves in Figure 2.6 show that if higher values of cumulative
grade and cumulative recovery are required in the concentrate, the ore must be liberated
further, usually by grinding to a Àner particle size distribution.
Improved liberation allows greater rejection of gangue, which increases concentrate
grade without affecting recovery in this perfect separator, and this shifts the grade-
recovery curve for the more liberated stream to a higher level. Note that in actual
separation processes the grade-recovery values observed are almost always lower than
the mineralogically limited grade-recovery curve since, in practice, separation processes
rarely (possibly never) operate as perfect separators. In addition, complete mineral
liberation is rarely achieved.
The mineral composition of particles can also affect their rate of response in separation
processes. In Áotation separations, the observed Áotation rate of each particle is a function
of the mineral composition of the particle surface, the Áotation rates of the individual
minerals and the particle size (Bartlett and Mular, 1974; Jameson, 2012). In general, as

Comminution Handbook | Spectrum Series 21 21

 CHAPTER 2 t MINERAL LIBERATION

FIG 2.6 – Mineralogically limited grade-recovery curves change

as mineral liberation increases (image courtesy Cathy Evans).

the proportion of the target mineral occupying the particle surface increases, the particle
ÁRWDWLRQ UDWH LQFUHDVHV 7KLV LQIRUPDWLRQ RQ KRZ SDUWLFOH FRPSRVLWLRQ DIIHFWV PLQHUDO
response can provide the basis for separation models used in integrated simulation
FDSDELOLWLHV DOORZLQJ FRPPLQXWLRQ DQG VHSDUDWLRQ FLUFXLWV WR EH VLPXODWHG DV DQ
integrated system (Evans et al

BIBLIOGRAPHY
Amstutz, * & 0LFURVFRS\ DSSOLHG WR PLQHUDOGUHVVLQJQuarterly of the Colorado School of
Mines²
Andres,8/LEHUDWLRQVWXG\RIDSDWLWH²QHSKHOLQHRUHFRPPLQXWHGE\SHQHWUDWLQJHOHFWULFDO
charges, International Journal of Mineral Processing²
Andrews,-5*DQG0LND76&RPPLQXWLRQRIDKHWHURJHQHRXVPDWHULDOGHYHORSPHQWRI
a model for liberation phenomena, in Proceedings XI International Mineral Processing Congress,
&DJOLDULSS²
Bajic, 6  &KDUDFWHULVDWLRQ RI WKH OLEHUDWLRQ NHUQHO 3K' WKHVLV 8QLYHUVLW\ RI 4XHHQVODQG
%ULVEDQH
Barbery, *  0LQHUDO OLEHUDWLRQ PHDVXUHPHQW VLPXODWLRQ DQG SUDFWLFDO XVH LQ PLQHUDO
processing, Les Editions GB4XHEHF
Bartlett,'5DQG0XODU$/'HSHQGHQFHRIÁRWDWLRQUDWHRQSDUWLFOHVL]HDQGIUDFWLRQDO
mineral content, International Journal of Mineral Processing²
Bérubé,0$DQG0DUFKDQG-&(YROXWLRQRIWKHPLQHUDOOLEHUDWLRQFKDUDFWHULVWLFVRIDQ
iron ore undergoing grinding, International Journal of Mineral Processing²
Black,%1DQG0XWWRQ%.7KHGHYHORSPHQWRIWKH+LOWRQPLQH²LQProceedings
13th Congress, The Council of Mining and Metallurgical Institutions6LQJDSRUHSS²
Daniel, 0 - DQG 0RUOH\ &  &DQ GLDPRQGV JR DOO WKH ZD\ ZLWK +3*5" LQ Proceedings
Diamonds – Source to Use 2010 ConferenceSS²7KH6RXWKHUQ$IULFDQ,QVWLWXWHRI0LQLQJ
DQG0HWDOOXUJ\0DUVKDOOWRZQ

22 Comminution Handbook | Spectrum Series 21

 CHAPTER 2 t MINERAL LIBERATION

Dunne, R C, Goulsbra, A and 'unlop, I, 1996. High pressure grinding rolls and the effect on
liberation: comparative test results, in Proceedings Randol Gold Forum 96, 449 p (Randol
International Limited: Golden).
Evans, C L, Andrusiewicz, M A, Wightman, E M, Brennan, M and Manlapig, E V, 2013. Simulating
concentrators from feed to Ànal products using a multi-component methodology, SME Preprint
13–094, SME Annual Meeting, 'enver.
Evans, C L, Wightman, E M, Manlapig, E V and Coulter, B L, 2011. Application of process
mineralogy as a tool in sustainable processing, Minerals Engineering, 24:1242–1248.
Gu, Y, 2003. Automated scanning electron microscope based mineral liberation analysis, Journal of
Minerals and Materials Characterization and Engineering, 2:33–41.
Jameson, G J, 2012. The effect of surface liberation and particle size on Áotation rate constants,
Minerals Engineering, 36–38:132–137.
Jones, M P, 1987. Applied Mineralogy: A Quantitative Approach (Graham and Trotman: London).
King, R P and Schneider, C L, 1998. Mineral liberation and the batch comminution equation,
Minerals Engineering, 11:1143–1160.
Lin, C L and Miller, J ', 1996. Cone beam X-ray microtomography for three-dimensional liberation
analysis in the 21st century, International Journal of Mineral Processing, 47:61–73.
Lin, ', Lastra, R and Finch, J A, 1999. Comparison of stereological correction procedures for
liberation measurements by use of a standard material, Transactions of the Institution of Mining
and Metallurgy, Mineral Processing and Extractive Metallurgy, 108:C127–C137.
Munro, P ' and Pease, J ', 1988. Mount Isa Mines Limited heavy medium preconcentration
plant – six years after commissioning, in Proceedings Third Mill Operators’ Conference, pp 35–42
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Reid, A F and Zuiderwyk, M A, 1983. QEM*SEM: automated image analysis and stereological
applications to mineral processing and ore characterization, Acta Stereologica, 2:205–208.
Sutherland, ' N, 1989. Batch Áotation behaviour of composite particles, Minerals Engineering,
2(3):351–367.
Veasey, T J and Fitzgibbon, K E, 1990. Thermally assisted liberation of minerals – a review,
Minerals Engineering, 3:181–185.
Wang, E, Shi, F and Manlapig, E V, 2012. Mineral liberation by high voltage pulses and
conventional comminution with same speciÀc energy levels, Minerals Engineering, 27:28–36.
Wiegel, R L, 1976. Integrated size reduction-mineral liberation model, Transactions of the Society of
Mining Engineers of AIME, 260:147–152.
Wiegel, R L, 2006. The rationale behind the development of one model describing the size
reduction/liberation of ores, in Advances in Comminution (ed: S K Kawatra), pp 225–242
(Society for Mining Metallurgy, and Exploration: Littleton).

Comminution Handbook | Spectrum Series 21 23

 HOME

Chapter 3

PARTICLE MEASUREMENT
TECHNIQUES
Rick Hughes

INTRODUCTION
Since the inception of modern mineral processing techniques in 1880 (such as closed
grinding circuits and Áotation the methods of characterising particles haYe remained
fairl\ constant ² sieYing Zet assa\ing optical microscop\ and emission spectroscop\
:ith the adYent and increasing use of computer processing in the 180s neZ techniques
(some of Zhich are discussed in this chapter haYe eYolYed from scientiÀc progress
often championed E\ or deYeloped Zithin mineralprocessingrelated Àelds $ll the
techniques mentioned are currentl\ used in comminution to some degree ,n almost
all stages during e[ploration drilling Elasting conYe\ing EeneÀciation and shipping
there is a need to undertaNe measurements of particle si]e composition contamination
or other intrinsic properties in order to quantif\ ore grade or compliance
)igure 1 summarises the Yarious techniques that are used to measure particle si]e as
a function of particle si]e range
3article measurement starts Zith sampling 7his chapter aims to giYe the reader an oYer
YieZ of the importance of accurate suEsampling Zith particular reference to the anal\tical
techniques mentioned )urther information aEout these techniques can Ee oEtained from
*\·s ZorN (18 Zhich underpins all good sampling practice Zhile a detailed account
of sampling methods and particle si]ing techniques can Ee found in $llen (1

SAMPLING
5elatiYe to scientiÀc anal\sis sampling is the art of remoYing a small representatiYe
e[ample (NnoZn as a suEsample from the originating sample material Zith the aim
of minimising an\ ph\sical or chemical differences EetZeen Zhat Zas taNen and Zhat
remains as is statisticall\ MustiÀaEle ,n other Zords the suEsample should Ee e[actl\
the same in as man\ aspects as possiEle to the sample Eut smaller in Yolume and mass

Comminution Handbook | Spectrum Series 21 25

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

FIG 3.1 – Size ranges applicable to measurement techniques

(Hosokawa Alpine Aktiengesellschaft, 2007).

2ften the purpose of suEsampling is to end up Zith a manageaEle fraction that can Ee
practicall\ characterised Zithin the conÀnes of time and cost ,t should Eest represent
the originating material Zithout haYing to anal\se (often conducted destructiYel\ far
greater quantities of it
$n originating sample from site might taNe the form of an entire quarr\ Eench
stocNpile Àlter caNe or conYe\or and range from tonnes to hundreds of thousands of
tonnes $ suEsample of these originating materials should accuratel\ reÁect the original
si]e and composition Zithin a mass in the range of a Nilogram or tZo
7he receiYing laEorator\ conducting the assa\ (for e[ample anal\sing for iron or gold
content Zill haYe to further suEsample the receiYed suEsample to a Yolume and mass
suitaEle for the anal\tical technique perhaps in the range of milligrams to seYeral grams
$ t\pical ;ra\ diffraction (;5' or ;ra\ Áuorescence (;5) anal\sis for e[ample
onl\ uses EetZeen tZo and ÀYe grams to oEtain a Yalid anal\tical result
SuEsampling from Àeld to laEorator\ is therefore critical in oEtaining Yalues that can
Ee interpreted as representing much larger quantities of sample Zhile anal\sing the
least amount possiEle
7here are tZo golden rules to sampling and conÀdence should onl\ Ee placed in
suEsamples that haYe Eeen acquired as folloZs
1 aEstract a suEsample Zhen the sample is in motion (applies equall\ to poZders and
liquids
 sample the entire streamÁoZconYe\orprocess for a short time (or multiple short
times to oEtain the desired suEsample mass
Static suEsampling or stocNpiling Zill ineYitaEl\ lead to Eiases due to segregation
Zhich are harder to oYercome in a static sample compared to a sample in motion

26 Comminution Handbook | Spectrum Series 21

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

Sampling part of a streamÁoZconYe\or for a longer period of time Zill Ee prone to

selectiYel\ aEstracting a Eias (such as Ànes that ma\ haYe migrated due to YiEration
to the Ease of the suEsampling area 7his can lead to a nonrepresentatiYe suEsample
regardless of hoZ long for Zhich the suEsample is acquired
Segregation of particles occurs for a Yariet\ of reasons
x wide particle size distributions alloZing Àner particles to percolate doZnZards under
graYit\ EetZeen larger particles and settle preferentiall\ at the Eottom of the sample
effectiYel\ ¶lifting· larger particles to the top of the sample
x vibration Zhich ma\ segregate Eased on a si]e and densit\ difference
x airÁow and carriaJe velocit\ Zhere Zind Zill carr\ Ànes further from the point of
origin as Zill an\ Yelocit\ associated Zith a moYing sample for e[ample off the end
of a conYe\or
x cKePical cKanJe Zhich ma\ occur on the e[ternal surfaces of a stocNpile due to
oxidation or atmospheric exposure and produce size and compositional changes
compared to the EulN or the reYerse Zhere internal reactions ma\ Ee unaEle to
dissipate their heat as fast as similar reactions on the surface

Representative subsampling
2ne of the most crucial areas of anal\sis is the one that is oYerlooNed the most
suEsampling :hether it is acquiring a large enough suEsample to Ee representatiYe
accuratel\ suEsampling a heterogeneous stocNpile or simpl\ ZithdraZing a solution
from a ÁoZ line incorrect suEsampling Zill Eias an\ anal\sis that folloZs often to an
order of magnitude larger than an\ repeataEilit\ limitations of the anal\sis technique
7here are at least ÀYe t\pes of sampling error
1 settling/segregation causing layering that is unaccounted for in the abstracted
subsample
 an insufÀcient subsample collected to be representatiYe of the Yariations naturally
present (may be acquisition time or mass dependent
 selectiYe Ninetics during subsampling (for example moYing conYeyor sampling and
ÁoZ gradients around Áuid sampling quills/ports
 a changing enYironment (is the external/subsample container the same temperature/
pressure/atmosphere as the material being sampled to preYent physical/chemical
changes postsampling"
 statistical errors intrinsic to the process of subsampling that are unique to the system
being sampled and may not be minimised beloZ a certain Yalue
,t is important to remember that the stocNpile or ÁoZ line that is being monitored may
haYe tonnes or millions of cubic metres of material present Zhich is to be represented
by a subsample of a feZ Nilograms (or tens of Nilograms in the best case or litres
7his Zill be reduced in the analysis process to a feZ tens of grams or millilitres Zhich
may in turn only lead to a total analysed mass in the order of milligrams or microlitres
,t is difÀcult enough to haYe representiYity from beginning to end for one sample
let alone for the hundreds or thousands that may be being processed eYery day in a
typical laboratory
7here are many Yariations on sampling methods $ brief oYerYieZ of the most common
methods is presented in the folloZing sections

Comminution Handbook | Spectrum Series 21 27

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

Scoop sampling
$ Yery basic method of extracting a subsample is Zith a scoop or rounded troZel
(see )igure  7his type of handheld deYice does not typically penetrate all the Zay
through a stocNpile or eYen through a smaller mass such as a bucNet it Zill often only
taNe a surface sample Zhich is unliNely to be representatiYe

FIG 3.2 – Scoop sampler (image courtesy Rick Hughes).

Chute sampling
&hute sample splitters ()igure  come in a range of orientations and sizes from small
handheld units to tabletop deYices 7hey consist of a hopper/feed area Zhere the sample
is placed and alloZed to enter an array of chutes or openings typically there are eight or
more chutes that alternate in directing the falling particulate into receptacles on either
side of the chute (there may be more than tZo receptacles 7he idea is that the chute
splits the sample into tZo representatiYe subsamples that can then be further reduced
by serial chute sampling until the desired subsample mass is obtained hoZeYer bias
can still be introduced into this technique through operators not spreading the incoming
sample oYer all chutes equally or sloZly enough

FIG 3.3 – Chute sampler (image courtesy Rick Hughes).

Spear sampling
$ spear sampler ()igure  is designed to abstract a sample at multiple points along
the spear from a poZder held in a container (for example a bag or bucNet ,t is typically
arranged as a concentric tube Zith openings along its length that may be closed or open
depending on the orientation of the inner tube Some spear samplers use an offset in the
openings so that loZer sampling points open Àrst in a sequence alloZing the sampling
spear to only abstract a subsample from the loZer regions of the material being sampled
7hese types of samplers are Yery common in the grain sampling industry as they enable

28 Comminution Handbook | Spectrum Series 21

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

FIG 3.4 – Concentric tube spear sampler showing holes aligned for sampling (image courtesy Rick Hughes).

the acquiring of representatiYe samples from deep Zithin grain silos or the bacNs of
grain trucNs as Zell as mineral concentrates prior to being dispatched

Coning and quartering

7his is probably one of the oldest subsampling techniques and it has one of the highest
biases of all the techniques coYered in this chapter 7he technique inYolYes placing the
sample in a cone by dropping it from a set height in much the same Zay a stocNpile is
formed $ diYiding plane is brought doZn on the sample pile to diYide it into quarters
7ypically tZo or three of the quarters are subsequently discarded and the process is
repeated Zith the remaining quarters until the desired mass of subsample remains as
demonstrated in )igure 
7he bias in this method is intrinsic to both the size distribution and the shape of the
particles forming the pile and hoZ these factors alloZ larger particles to segregate
from Ànes

A B

C D

FIG 3.5 – Coning and quartering subsampling method (image courtesy Rick Hughes). (A) Cone of powdered
material, (B) flattened cone, (C) quartered cone, (D) select one quarter and repeat the cone and quarter process.

Rotary riffler
7he bias introduced by rotary rifÁing a subsample is minimised by folloZing the
golden rules of sampling a moYing poZder and sampling the Zhole stream for short
periods of time 7his produces the loZest bias of Must about all subsampling techniques
$ Yibrating hopper (top right in )igure  is used to feed the sample at a set ÁoZ rate
into an arrangement of receptacles ² in this case test tubes ² that are rotating at Yariable
speeds 7he sample is moYing continuously doZn the hopper feed and selectiYely falling
into consecutiYe tubes Zith each tube collecting a representatiYe subsample of the original
7he receptacles used in rifÁers range in size from test tubes to gallon drums Zidely
used to prepare charges for test ZorN

Comminution Handbook | Spectrum Series 21 29

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

FIG 3.6 – Rotary micro riffler (image courtesy Rick Hughes).

WET/DRY SIEVING

The sieving technique

SieYing is a size classiÀcation technique that has been used around the Zorld by Yarious
ciYilisations for thousands of years indeed Zet sieYing Zas used by cheese maNers in
the sixth century BCE (Salque et al 01
Calibrated and certiÀed sieYes are readily aYailable betZeen  mm and 1 —m typically
in a ¥ aperture size progression for example  —m  —m 10 —m Extreme patience
needs to be applied to sieYing anything beloZ 0 —m
:et or dry sieYing is a statistical process that is only as Yalid as the time spent ensuring
the process is conducted properly 7he process segregates particles based on their
secondlargest dimension so nonspherical particles may taNe excessiYe agitation to
present an aspect that Zill alloZ the particle to fall through the sieYe opening 7his
often simpliÀed process is easily used (and abused and size cuts may be blurred or
less than optimum through improper use lacN of time spent agitating or neglected and
outofspeciÀcation sieYes 7olerances in the sieYing apertures are fairly generous Zhen
neZ for example a  —m ZoYen Zire sieYe is deemed to be Zithin speciÀcation as long
as the median aperture size falls betZeen 0 and 80 —m and there are not more than
ÀYe per cent of apertures betZeen 1 and 10 —m Electroformed sieYes typically used
in the 1²0 —m sieYe range do not suffer from the same Zide tolerances and may be
purchased Zith Yariously shaped apertures (for example circular or square compared
to the trapezoidal ZoYen mesh sieYes
:oYen nylon mesh sieYes are aYailable betZeen  and  —m and although they are
not supplied as calibrated they are often of a high enough quality that reproducible
results can be obtained using different sections of mesh ,n addition the nylon mesh
does not suffer from the same chemical interactions that may be present Zhen sieYing
acidic or other reactiYe materials

30 Comminution Handbook | Spectrum Series 21

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

%lindinJ Zhere the sieYe entrains trapped particulates in the mesh is the term used to
describe the limited effects of haYing more blocNed openings than there are open ones
3lacing the sieYe in a sonicating bath reYerse bloZing Zith loZpressure compressed
air gentle tapping of the entire sieYe face (not Must the mesh or one side and as a last
resort brushing trapped particulates may return the sieYe to a nonblinded and useful
state Zithout further effort $ny direct application of force (for example the use of
tZeezers or a spatula Zill undoubtedly place the sieYe beyond acceptable tolerances in
an instant *ood care of sieYes is essential if any reproducibility in the operation is to be
maintained 1o pressure or rubbing of any sort should be applied directly to the sieYe
mesh other than a Zater Met or soft paint brush to moYe the particulate around the sieYe

Wet versus dry sieving

7he sharpness of the cut Zith sieYing can be dramatically improYed by conducting the
sieYing Zet either Zith Zater or solYent 7his can remoYe surface charge effects (eYident
Zith some particulates Zhen dry sieYing that cause particles to adhere to the mesh and
not pass through as Zell as assist Zith deagglomeration and disaggregation of coarser
particles into their component particles :et sieYing should be conducted ZhereYer
possible for best comparison and continuity Zith laser diffraction being conducted on
the undersize

Stainless steel versus brass versus nylon mesh sieves

Stainless steel sieYes haYe the adYantage of being slightly more chemically inert than
brass $cidic samples Zill stain and eYentually corrode brass sieYes beyond their
alloZable tolerance in mesh size meaning that they haYe to be disposed of 1ylon mesh
sieYes are the most chemical tolerant but suffer from noncertiÀcation and distort easily

LASER DIFFRACTION PARTICLE SIZE ANALYSIS

Laser diffraction instruments are by far the most common and practical sizing technique
in eYeryday use in both the laboratory and the Àeld 7he technique has been around for
oYer 0 years and came about from the creation of the ¶correlator· a deYice deYeloped
by the 5oyal Signals and 5adar Establishment (5S5E and 0alYern ,nstruments in the
late 10s (3iNe 1

Basic principles
Laser diffraction particle size analysis inYolYes shining a collimated light source
(typically a 0 nanometre (nm laser through a cell carrying the sample (either Zet or
dry and onto a detector that can determine the diffracted angle 0ie (and historically
)raunhofer theory Zas used to interpret the diffraction angle against NnoZn diffraction
angles but complex matrix integrations are noZ completed by an attached computer to
report the resulting size distribution 2ptical parameters (such as refractiYe index and
absorption coefÀcients are required for optimal reÀning of submicron particles
$ll systems on the marNet today comply Zith ,S2 1000 and most offer
1C353art11 (8S )ood and 'rug $dministration 1 compliance for implementing
the setup of standard operating procedures (S23 operating analyses and record
Neeping of data

Comminution Handbook | Spectrum Series 21 31

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

,t should be noted that apart from traceable standards there is often more than one
correct result for particle size distribution analysis )or example agglomerated state
sizing may giYe rise to a distribution that quicNly breaNs doZn as deagglomeration
occurs until equilibrium is reached and either discrete particles or the remaining bonded
particulates cannot be further broNen apart
:here carousels are aYailable for automated analyses it should be noted that S23s
are often not intuitiYe enough to fully interpret results before recording them and
disposing of the analysed subsamples Changing agglomeration/aggregation entrained
air bubbles and the settling of coarse fractions all limit the range and Yariety of samples
that may be analysed in this Zay

Instrument comparison

Malvern Instruments MS2000

7he 0alYern ,nstruments 0astersizer 000 (0S000 measures particles betZeen  mm
and 0 nm and reports size as an equiYalent spherical diameter $ typical plot is shoZn
in )igure 
,n reality all laser diffraction systems suffer from the same basic limitations sample
preparation and sample dispersion 7he +ydro * dispersion system supplied Zith this
instrument has an effectiYe upper capability limit of  mm particles but the stirring
speeds required to representatiYely suspend particles of this size are high and therefore
liNely to induce segregation $ more practical approach is to quantitatiYely sieYe at
sizes betZeen 00 —m and 1 mm and only place the undersize in the laser diffraction
instrument for size analysis
0alYern ,nstruments also manufactures an automated carousel that alloZs for the
unattended analysis of up to  samples

FIG 3.7 – Typical plot from Malvern Instruments MS2000 (image courtesy Rick Hughes).

Coulter LS230
Coulter·s patented polarisation intensity differential scattering (3,'S detector for
nanometre sizing to around 10 nm Zith increased resolution compared to other laser
diffraction instruments is an adYantage in the sizing Àeld if submicron resolution is
important 7rimodal samples at 0 0 and 80 nm can be successfully resolYed using
3,'S technology Zhich complements the loZend sizing capabilities of the laser
diffraction optics

32 Comminution Handbook | Spectrum Series 21

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

Cilas products
Cilas manufactures a range of laser diffraction size analysers as Zell as an optically
based shape analyser 7he top of the line 110 is able to size particles betZeen 0 nm and
 mm and a builtin chargecoupled deYice (CC' camera records images of particles
as they are being sized by the laser optics 7hese images can readily be incorporated into
the size distribution reporting adding Yaluable information on inÁuences of shape and
agglomeration state to the Ànal result
Cilas also produces a dynamic lightscattering instrument called the 1ano 'S Zhich
is a compact instrument capable of measuring size betZeen 0 nm and 10 —m

Fritsch Analysette 22 models

)ritsch manufactures the $nalysette  0icro7ec 3lus modular laser diffraction system
Zhich is capable of sizing betZeen 80 nm and  mm and the $nalysette  1ano7ec
3lus Zhich can size betZeen 10 nm and  mm

Retsch Horiba LA-950/300

5etsch manufactures tZo laser diffraction instruments that are able to measure in
the range of 10 nm to  mm betZeen them 7he units are similar in functionality to
most other laser diffraction systems and may be set up to analyse both Zet and dry
particulates

Micromeritics Saturn Digisizer

7he Saturn 'igisizer is a compact (small bench footprint laser diffraction instrument
that sizes in the range of 0 nm to  mm $n optional automated carousel alloZs for
the unattended analysis of up to 18 samples

SEDIMENTATION (SEDIGRAPH) AND COLUMN SETTLING

Sedimentation refers to the settling action of particles suspended in a Áuid due to some
external force most commonly graYity and is described by StoNes· laZ 3articles of
different sizes (hydrodynamic diameters Zill fall at different rates forming the basis for
the use of techniques such as column settling and instruments such as the SediGraph to
determine effectiYe particle size

Basic principles
7he basic principle of sedimentation studies reYolYes around the fact that Zhen
distributed homogeneously throughout a liquid and made up of the same material
larger diameter particles fall quicNly Zhile smaller particles settle more sloZly So at
any point after the initial homogenising of the sample in liquid (ie shaNing there Zill
be a population distribution of differentsized particles as a result of their different
Yelocities 7his distribution can be calculated using StoNes· laZ Zhich assumes
spherical particles and a laminar ÁoZ of the Áuid around the particles and exactly
describes the relationship betZeen the diameter of the particle and its sedimentation
Yelocity

Comminution Handbook | Spectrum Series 21 33

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

Micromeritics SediGraph
7he SediGraph instrument operates by Àring a collimated beam of soft ;rays through
a Yessel containing suspended particles in a liquid and measuring the transmittance of
the beam at Yarious points in the Yessel oYer time By relating the transmittance of the
beam to the concentration of particles at that point and utilising StoNes· laZ a graph
shoZing the percentage of particles that haYe a particular diameter can be calculated
and plotted1

Settling column
7he settling column experiment uses the same relationship of StoNes· laZ betZeen the
speed of settling and the diameter of the particles in suspension $ large measuring
cylinder is Àlled Zith a homogenised suspension of sample and the settling rate is
recorded as sedimentation occurs $fter  hours the supernatant liquid is siphoned off
and the total Yolume of settled solids and their Zeight is recorded

HYDROCYCLONE METHODS

Hydrocyclones
$ hydrocyclone is a single or arrayed deYice that imparts centrifugal forces on
particulates in suspension forcing larger or denser particles preferentially toZards the
circumference Zhich can then be split out and collected separately from the Ànes 7he
cut point (size is a function of the hydrocyclone setup particle density suspension
media Yiscosity and density and ÁoZ rate
$ hydrocyclone is shoZn in )igure 8
+ydrocylones are limited in their ÁoZ rate if speciÀc cut sizes are required
BanNs of hydrocyclones running in parallel are normally required for highYolume
applications

FIG 3.8 – A typical hydrocyclone configuration (image courtesy Rick Hughes).

1. A far less common instrument is the Quantachrome Microscan, which uses similar principles and virtually the
same construction to the SediGraph to calculate hydrodynamic diameter. At the time of writing, this instrument
was no longer manufactured.

34 Comminution Handbook | Spectrum Series 21

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

Warman Cyclosizer
7he :arman Cyclosizer (noZ manufactured by 0$5C 7echnologies has been around
in Yarious guises since the 10s ,t is a Yery simple arrangement of ÀYe consecutiYe
hydrocyclones that collect the oYersize at the top of the centrifuge (for collection Yia a
YalYe and pass on the undersize directly to the next cyclone 7he typical cut sizes are
predetermined to a large extent by the density of the particles and the ÁoZ Yelocity
as the cyclones· cut points cannot be Yaried indiYidually and are normally betZeen a
top size of 0 —m and an end size of 10 —m 7he underÁoZ can also be collected and is
typically in the order of hundreds of litres for a typical 0 g starting mass
$ charge of around 0 g is the maximum starting mass of the :arman Cyclosizer
Zhich means that Nilogram quantities of sample requiring size fractionation must be
processed in multiple batches

PARTICLE COUNTING
Particle counting differs from other more common sizing techniques in that it is looking
at a number distribution as opposed to a mass/Yolume distribution

Hiac/Royco
7he +iac/5oyco particle counter utilises laser extinction to determine a quantitatiYe
size distribution for loZconcentration suspensions 'epending on the detector the
measurement range falls betZeen 0 —m and 00 —m 7he instrument draZs a knoZn
Yolume of sample through a sampling syringe and through the test cell at a knoZn ÁoZ
rate 7he laser is directed through the sample and the obscuration caused by particles
passing betZeen the laser source and the detector is registered to determine particle
size 7he data is binned into particle count per size range to form a size distribution
Calibration is initially performed using a range of certiÀed reference standards
Yerifying the size signal and the concentration/count Calibration should be reassessed
periodically to take into account any degradation of the laser or detector
5eÁectiYity and transparency of the particles in suspension can affect the measurement
and oYerlapping particles in higherconcentration suspensions may lead to multiple
particles being detected as a single larger particle 'ilution may be necessary for such
highconcentration suspensions 7he morphology of the suspended particulates also
inÁuences the measured particle size as the particles are inÁuenced by the ÁoZ through
the test cell causing them to orient Zith the longest diameter parallel to the direction of
ÁoZ 7he Yiscosity of the suspension medium may also haYe an effect on the ÁoZ rate
and may introduce bubbles to the system Zhich Zould erroneously be measured as
particles 7he limited sample Yolume analysed can cause difÀculties Zith representiYity
so care must be taken Zith sample selection and YeriÀcation of repeatability
+iac 5oyco particle counting can be used to analyse any transparent Áuid including
aqueous samples fuels and oils 7he technique is used for fuel and oil cleanliness
testing by 1$S 18 ($erospace ,ndustries $ssociation 011 and ,S2 01 Zater
quality testing and enYironmental monitoring and many other applications including
troubleshooting Àltration systems

Comminution Handbook | Spectrum Series 21 35

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

Airmet Dusttrak II 8530

7he $irmet 'usttrak ,, 80 aerosol analyser is a Zidely used Àeld instrument that can
measure particle size and mass concentration at the same time 7his compact and sealed
instrument is designed to be left in the Àeld to accumulate data in real time during dust
emission conditions using a laser photometer diode ,nlet Àlters alloZ discrimination
of particulate material for legislated exposure monitoring as P010 (thoracic P04
(respirable P0 and P01

The Coulter Counter

7he Coulter Counter Zas initially deYeloped in the 140s to count blood cells hoZeYer
this nonoptical approach to particle sizing has found use in a much Zider range of
applications including drugs pigments Àllers toners foods abrasiYes explosiYes
clay minerals construction materials coating materials metals and Àlter materials
7he instrument is based on the Coulter principle Zhich is sometimes knoZn as the
electrical sensing zone method 7his uses tZo electrodes separated by an aperture tube
Zith an oriÀce of knoZn diameter $ll the components are immersed in electrolyte
solution and the oriÀce represents a knoZn Yolume of the solution $s particles pass
through this zone they displace a Yolume of solution Zhich creates impedance across
the tZo electrodes that is proportional to the particles· Yolume this signal can be
recorded as a digital pulse 7his data can be easily reprocessed to for example improYe
resolution in a certain section of the resultant size distribution Zithout the need to
reanalyse the sample
7he Coulter Counter·s most signiÀcant adYantage is its ability to directly record the
number population of a sample ,t can measure particle sizes from 04 to 100 —m 7his
range is dictated by the oriÀce size of the aperture tube Zith the ability to measure
particles in the range of tZo to 0 per cent of the diameter of the oriÀce :hile this limits
the dynamic range compared to other techniques it can be addressed by using multiple
tube techniques 7he Coulter Counter is unaffected by colour and refractiYe index and
less prone to issues in sizing arising from shape effects
:hile the most common application of the Coulter Counter still remains in the Àeld
of cell counting the adYantages of the instrument haYe made it a reliable alternatiYe to
particle sizing in a Yariety of Àelds

SURFACE CHARGE TECHNIQUES

Comminution is only as effectiYe as the time that particles remain discrete after grinding
LoZ surface charges on indiYidual particles may facilitate agglomeration occurring as
they collide effectiYely recementing particles together Zhereas high surface charges
preYent agglomeration by alloZing particles to repel each other (much like magnetism
7hese forces called Yan der :aals forces can be relatiYely strong and haYe pronounced
effects on particulate size distributions oYer time
,ncreasing the effectiYe surface charge by the addition of charge modiÀers imparts a
repulsion effect causing particles to repel each other 7hey remain as discrete particles as
they can no longer get close enough to be affected by Yan der :aals forces

36 Comminution Handbook | Spectrum Series 21

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

7he sizing of ground/milled materials/particulates may not be a true reÁection of the

discrete particle size if surface charge inÁuences fail to be fully considered and Zhat is
being sized is agglomerate

Zetasizer
7he 0alYern =etasizer uses surface charge (zeta potential to excite particles in
suspension using an applied electric Àeld 7he resulting motion of particles can be
picked up by a Yariety of sensors

Matec Applied Sciences

0atec $pplied Sciences (0$S deYeloped a noYel technique of determining zeta
potential based on the ultrasonic Yibration intensity of the particles being moYed in
suspension by an applied highfrequency electric Àeld ,t currently markets a range of
three zeta sizing instruments

PHOTOMICROSCOPY-OPTICAL

Basic technique
Limitations Zith the laser diffraction sedimentation and light extinction techniques
such as haYing no information on aspect ratio and relatiYely loZ uppersize limits haYe
seen optical techniques come into their oZn oYer the past decade Computational speed
coupled Zith highspeed highresolution CC' cameras haYe enabled a Zhole neZ area
of rapid analysis preYiously unaYailable
7he technique inYolYes presenting the backlit particles before the camera either as a
falling obMect or on a glass slide and capturing the resulting shadoZ image Controlling
ÁoZ and dispersion are crucial to obtaining meaningful data 'igital images are rapidly
processed to alloZ parameters such as )erret diameter perimeter aspect ratio sphericity
and roundness to all be reported for thousands – if not tens of thousands – of particles
Zithin minutes ,n many instances particle shape analysis by light techniques may
proYide better correlation Zith dry and possibly Zet sieYing and certainly in a highly
reproducible and much faster manner

Manufacturers

Retsch Camsizer
7he 5etsch Camsizer Zill simultaneously measure particle size shape and other
parameters as the dry particulate matter is fed Yia a hopper casting shadoZs onto tZo
highresolution CC' cameras 7he Camsizer has a claimed size range capability of 1 —m
to 0 mm but it is likely that coincidences and representiYity Zould affect the ultimate
resolution at either end of the size scale 7he digital images can be recorded and kept for
later reeYaluation if necessary

Malvern Instruments Morphologi

7he 0alYern ,nstruments 0orphologi technique uses an automated light microscope
to measure particle size shape and a range of other parameters betZeen 0 —m and a
claimed upper limit of seYeral millimetres 7he dispersed sample is placed on a standard

Comminution Handbook | Spectrum Series 21 37

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

microscope slide Zhile the stage is automatically rastered to alloZ Zide area scans of
many particles $n added beneÀt is that transmitted or reÁected light may be used as
Zell as incorporating polarisation for added particle information such as birefringence

Haver and Boecker CPA

7he +aYer and Boecker CP$ system is similar in principle to the 5etsch Camsizer
,t has a claimed size range of betZeen 10 —m and 00 mm although this appears to
not be possible Zithin the setup of a single instrument 7he +aYer and Boecker CP$
system appears to cater for more processtype industries Zith conYeyorstyle sample
presentation units that could be used to sample directly from industrial processes and
proYide realtime particle size data of relatiYely coarse particulate material

Cilas Analysette 28 ImageSizer

7he Cilas $nalysette 8 ,mageSizer is a dynamic image analyser that uses optical methods
to size particles and report shape betZeen 0 —m and 0 mm Cilas also manufactures
an ,mage7ec unit that can optically image particles betZeen 1 —m and  mm ,t claims
that its image sizers Zill size particles in suspension as Zell as emulsions of droplets of
differing refractiYe index to the suspension medium

ELECTRON MICROSCOPY
$s early as the mid180s organisations such as the CommonZealth ScientiÀc and
,ndustrial 5esearch 2rganisation (CS,52 and latterly the -ulius .ruttschnitt 0ineral
5esearch Centre (-.05C at the 8niYersity of 4ueensland had begun to harness the
synergy betZeen automated Zidearea (seYeral square centimetres electron microscopy
imaging and intuitiYe energy dispersiYe spectroscopy (E'S analysis 7he principle
inYolYes potting and polishing or thinsectioning specimens (typically but not limited
to geological samples for analysis and achieYing a Ànal polish that is optically Áat
across particles or regions of interest set in an epoxy resin Particles of carbon/graphite
may be added to the sample particulate mix to aid in particle separation Zithin the
mount 7he resulting particles or regions can be automatically imaged (usually in
backscatter electron mode and softZare can be used to calculate boundaries as Zell as
intraparticle compositional changes 7he same softZare then directs the electron beam
to areas of interest that can be semiquantitatiYely analysed to determine elemental
composition and ratios thus giYing inferred mineralogy E'S systems are often set up
as top of the market options to alloZ for Yery rapid (millions of counts per second
data collection Zhich in turn alloZs for more physical ¶area· to be analysed in a giYen
time Complimentary ;5' (discussed in the folloZing section is alZays recommended
Zhere mineralogy is not fully knoZn Phase liberation efÀciency mineral deportment
and mineral association algorithms can then interpret the particles to determine hoZ
close to the surface (perimeter each phase is for a giYen particle size shape or other
compositional or morphological Yariable
Systems currently on the market include the )E, Company·s 4E0SC$1® (4uantitatiYe
EYaluation of 0inerals by SC$1ning electron microscopy originally deYeloped by
CS,52 Zhich also incorporates elements of -.05C·s 0ineral Liberation $nalysis
(0L$ system 2xford ,nstruments· ,1C$0ineral and 7ESC$1·s ,ntegrated 0ineral
$nalyzer (7,0$ Some such as the )E, and 7ESC$1 systems are supplied on

38 Comminution Handbook | Spectrum Series 21

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

dedicated scanning electron microscopy (SE0 platforms Zhereas the ,1C$0ineral

softZare is capable of acquiring data from any electron microscope platform
$n example data output from an SE0based mineralogical system is shoZn in
)igure  Phases are assigned colours and locations are determined (eg liberation in
relation to the aYailability of surface treatments such as leaching

FIG 3.9 – Wide area mineral phase map (image courtesy Rick Hughes).

X-RAY DIFFRACTION
Crystallographic analysis using ;rays Zas pioneered in the early 100s and Zas soon the
method of choice to rapidly determine mineralogical composition based on crystallinity
7he latest softZare alloZs samples to be rapidly analysed and interpreted against Yast
libraries of knoZn mineral and inorganic/organic phases (see http//ZZZicddcom to
enable fully quantitatiYe analysis for all but the most difÀcult of samples 7he technique
requires that the sample under inYestigation be ground to a Àne poZder as gently as
possible so as not to destroy the crystal structure and effectiYely render the sample
amorphous 7he pressedpoZder sample is then positioned Zithin the instrument and
bombarded Zith monochromated ;rays oYer a range of incident angles $ detector is
then moYed through an arc on the opposite side of the sample through an alignment
knoZn as 'ebyeScherrer geometry to record the ;rays diffracted from the sample
$ny regularity (lattice structure in the crystal/poZder Zithin the grains Zill then alloZ
constructiYe interference to occur and produce a peak Zithin the detector at a speciÀc
takeoff angle 7his angle can be used in Bragg·s laZ to determine the dspacing (or the
separation of the lattice of that particular mineral phase (see )igure 10

FIG 3.10 – Three lattice planes, with d-spacing, showing incoming and diffracted X-rays.
Note the black dots represent atoms (image courtesy Rick Hughes).

Comminution Handbook | Spectrum Series 21 39

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

0ost mineral phases haYe at least three and as many as seYen or more peaks associated
Zith the lattice structure ;5' is a Yery deÀnitiYe method of determining mineral
composition and relatiYe abundance as long as it is crystalline $morphous material
present in the sample Zill only add to the background and cannot be speciated
$ typical ;5' spectrum is plotted in )igure 11 Zhich shoZs the raZ data (black line
and the interpreted mineral composition based on the dspacing peaks (red calcite
blue magnetite and green magnesite associated Zith the unit cell parameters for the
minerals· crystal structure

FIG 3.11 – Typical X-ray diffraction spectrum, interpreted (image courtesy Rick Hughes).

SURFACE AREA ANALYSIS

7he surface area of a poZder Zill determine its rate of reaction as it is a direct measure
of the surface area aYailable to either gases liquids or other solids for interaction
0icropores giYe highly adsorptiYe media (such as actiYated carbon zeolites and other
molecular sieYes their ability to preferentially adsorb or remoYe molecules of a speciÀc
size 7herefore characterisation of surface area and pore structure is a useful tool in
understanding hoZ particulate matter behaYes in industrial processes

The Brunauer–Emmett–Teller theory

Surface area measurements by gas adsorption at cryogenic temperatures are noZ a Yery
routine analysis 0anufacturers such as 0icromeritics 4uantachrome and BELS25P
haYe a range of analysers that proYide fast and accurate surface area pore structure and
chemisorption characteristics
7he Brunauer–Emmett–7eller (BE7 theory Zas deYeloped by three )rench physicists
in 18 ,t describes hoZ the partial pressure of a gas Zould change as a portion of
its Yolume changed to a liquid on the surface of a particle that Zas at or beloZ the

40 Comminution Handbook | Spectrum Series 21

 CHAPTER 3 t PARTICLE MEASUREMENT TECHNIQUES

condensation temperature of that gas 7oday·s instruments typically use nitrogen

Zhich has a footprint of 18 angstroms to adsorb onto the surface of particles and into
Yoids at liquid nitrogen temperatures $ range of Yarying partial pressures is chosen to
obtain a linear portion in the adsorption isotherm Zhich alloZs the calculation from
theory to obtain a singlepoint threepoint or more commonly ÀYepoint BE7 surface
area Yalue $ multipoint isotherm can be used to gain an insight into the pore structure
and characterise any micro (up to 0 nm or meso (0 to 400 nm pores 0acropores
(!400 nm require mercury porosymmetry methods to characterise their distribution
'ifferent gases (such as krypton may be used to improYe the detection limit

The Blaine method

2riginally deYeloped for the cement industry in the early 10s the Blaine method alloZs
the Àneness of poZders to be assessed rapidly Zithout oYerly sophisticated equipment
,t consists of a manometer that applies suction or gas ÁoZ through a packed poZder bed
of knoZn Yolume $ knoZn pressure drop across the packed bed is maintained and the
ÁoZ rate determined 7he method is extremely limited and proYides no characterisation
of the surface area or pore structure in relation to their distribution merely a close
approximation of total surface area

BIBLIOGRAPHY
Aerospace Industries Association, 011 1$S18 – Cleanliness requirements of parts used in
hydraulic systems 'ecember
Allen, 7 1 Particle Size Measurement, Yolumes 1 and  (Chapman and +all London
Gy, P 18 SamplinJ oI Particulate Material, 7Keor\ and Practice (ElseYier $msterdam
Hosokawa Alpine Aktiengesellschaft, 00 +andbooN Powder and Particle ProcessinJ >online@
0 p $Yailable from https//ZZZhosokaZaalpinecom/poZderparticleprocessing/
doZnloadofbrochures/! >$ccessed  0ay 01@
International Organization for Standardization, 1 ,S2 4401 – +ydraulic Áuid poZer –
)luids – 0ethod for coding the leYel of contamination by solid particles 'ecember
International Organization for Standardization, 00 ,S2 1000 – Particle size analysis –
Laser diffraction methods 2ctober
Pike, E 5 1 7he 0alYern correlator case study in deYelopment PK\sics in 7ecKnoloJ\,
10(104–10
Salque, 0 Bogucki P , Pyzel - SobkoZiak7abaka , Grygiel 5 Szmyt 0 and EYershed 5 P
01 Earliest eYidence for cheese making in the sixth millennium BC in northern Europe
>online@ Nature. $Yailable from http//ZZZnaturecom/nature/Mournal/Yaop/ncurrent/
abs/nature118html! >$ccessed 10 -anuary 01@
US Food and Drug Administration, 1 C)5 – Code of )ederal 5egulations 7itle 1 )ood and
'rugs Source  )5 144 0arch

Comminution Handbook | Spectrum Series 21 41

 HOME

Chapter 4

ORE COMMINUTION
MEASUREMENT TECHNIQUES
Alban Lynch, Aubrey Mainza and Steve Morrell

INTRODUCTION
The intrinsic properties of minerals that are important in industrial comminution
processes are grindability, hardness, competence, abrasion resistance and abrasiveness.
Important extrinsic properties are feed rate and feed-sizing distribution. Applying the
general breakage equation, dE = -K.dx/xnWRLQGXVWULDOFRPPLQXWLRQSURFHVVHVLVGLIÀFXOW
because K depends on both the intrinsic and extrinsic properties, and the interactions
may be complicated. The procedure for designing a circuit to comminute a new ore is
GRQHWRGHWHUPLQHWKHLQWULQVLFSURSHUWLHVLQODERUDWRU\WHVWVDQGVHDUFKÀOHVLQRUGHUWR
ÀQGGDWDRQFLUFXLWVWUHDWLQJRUHVZLWKVLPLODUFKDUDFWHULVWLFV(TXLSPHQWVL]HVFDQWKHQ
be selected for the required feed rates and the performance of the circuit can be checked
by simulation. Design companies have databases that are used in the selection process
DQGLWLVXVXDOIRUWKHLUSURFHGXUHVWREHNHSWFRQÀGHQWLDO

BREAKAGE DEFINITIONS
• Grindability LV WKH SURGXFWLRQ RI ÀQHV UHVXOWLQJ IURP WKH FRPPLQXWLRQ RI PDWHULDO
in a standard test. It is important in the selection of rod and ball mills because the
grinding loads in these mills are constant and the product size depends on the feed
rate, feed size and grindability.
• Hardness refers to the resistance of a material to deformation, indentation or
penetration by means such as abrasion, drilling, impact, scratching and/or wear.
Hardness is important in the selection of autogenous grinding and semi-autogenous
grinding (AG/SAG) mills because their performance is dominated by the rate at
which the load breaks down; that is, resistance to deformation.
• Competence is the resistance of the coarser ore sizes to breakage. Ores are usually
UHJDUGHG DV ¶FRPSHWHQW· LI WKH VSHFLÀF HQHUJ\ WKH\ UHTXLUH LQ 6$* PLOOV LV KLJKHU

Comminution Handbook | Spectrum Series 21 43

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

than approximately 8 kWh/short ton (8.8 kWh/t). Low-to-medium competency ores

have historically dominated the mineral processing industry but the trend towards
high mining rates is resulting in ores of higher competence needing to be processed.
• Abrasion resistance is the ability of a rock to withstand frictional forces imposed by
contact with other rocks. Rocks being worn away by contact with other rocks is an
important mechanism in AG/SAG mills, particularly in AG mills where the process
LV UHVSRQVLEOH IRU JHQHUDWLQJ PRVW RI WKH ÀQH HQG RI WKH SURGXFW VL]H GLVWULEXWLRQ
The JKMRC standard abrasion test for ore is discussed later in this chapter as part of
the JK Drop Weight Testing procedure. It gives a measure of how much rock wears
away itself, whereas abrasiveness is a measure of how much a rock can wear away
steel. These two measurements are not interchangeable.
• Abrasiveness refers to wearing down or rubbing away steel by means of friction
with ore. A test for abrasiveness was developed by Allis-Chalmers (Bond, 1963),
which uses a rotating drum into which dry ore samples are placed with an impact
paddle mounted on a centre shaft rotating at a higher speed than the drum. The
drum contains 1.6 kg of -19 +12 mm ore, which is obtained by crushing 5 kg of ore
to -19 mm and screening at 12 mm. The paddle is made from standard alloy steel
hardened to 500 Brinell (where Brinell indicates a measure of metal hardness). The
Abrasion Index is determined by the weight loss of the paddle under standard
operating conditions.

BOND GRINDABILITY TEST

The Bond grindability test is the best known test of its kind and is used worldwide.
It originated in the late 1920s when Fred Chester Bond, a research engineer with Allis
Chalmers, developed a method to predict the energy required in grinding a ton of ore
from a known feed to a known product size, F80 to P80, if the grindability is known.
He collected data from many plants including the ore grindabilities, which were
determined as follows:
• Test equipment was a standard ball mill (305 mm × 305 mm) with smooth liners
rotating at 70 rev/min. The ball charge contained 285 balls with a total mass of
20.1 kg (43 at 36.8 mm, 67 at 29.7 mm, 10 at 25.4 mm, 71 at 19.0 mm, 94 at 15.9 mm).
The mill was in a closed circuit with a sieve.
• The feed (700 ml of -3.36 mm ore) was ground for 100 revolutions and the product
was screened at the required size (eg 74 μm). The oversize was recycled to the mill
and a fresh sample was added to replace the undersize.
• The procedure was repeated with the revolutions being adjusted to achieve
250 per cent recirculation. Six to 12 cycles were typically required before 250 per cent
recirculation occurred in three consecutive cycles.
If Gbp is the ball mill grindability in net g/rev (eg at 74 μm), its value is:
mass of 75 nm undersize in grams
G pb = number of revs of mill
Bond (1953) proposed a third theory of comminution in which:

E = 2.K. (1/ x2 - 1/ x1 )

44 Comminution Handbook | Spectrum Series 21

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

7RDSSO\LWKHGHÀQHG¶:RUN,QGH[·WiIRUDQRUHDVWKHVSHFLÀFHQHUJ\N:KWRQ

UHTXLUHGWRUHGXFHWKHRUHIURPDQLQÀQLWHJUDLQVL]HWR—P)RUDQ\RUHWKH:RUN
Index can be calculated from plant data and determined in a laboratory test:

10 - 10
Plant data W = Wi : c m (1)
P F

where:
W is the energy consumption of the mill in kWh / short ton
Wi is the Work Index
P and F are the 80 per cent passing sizes of the product and feed in μm

44.5
Laboratory data Wib = (2)
0.23
0.82 10 - 10
P1 G bp c m
P F

where:
Wib is the ball mill Work Index in kWh/short ton
Pi is the test sieve aperture size in μm
Gbp is the ball mill grindability in grams/rev
P and F are the 80 per cent passing sizes of the feed and product respectively
Bond’s procedure for mill selection was:
• determine the grindability of the ore in the standard laboratory test
• calculate the laboratory Wi from the grindability and use it as the plant Wi
• knowing Wi and F80, calculate W (kWh/t) for the required value of P80
• calculate the power required for the feed rate and select the mill that delivers the
power required.
The methodology Bond developed for mill selection is still widely used. The
Bond equation was based on data from ball mills used in the 1920s to 1950s that
did not exceed a 4 m diameter. It linked power consumption, mill throughput
and ore grindability, and became widely used for mill selection. Predictions on
the performances of mills that were selected by this procedure were considered
reasonably close to what were observed when variations in the ores being processed
were taken into account.
When 5.5 m diameter ball mills were installed at Bougainville in Papua New Guinea
and Pinto Valley in Arizona in the late 1970s, discrepancies were found to occur.
At Bougainville there were serious disparities between predictions using Bond’s
equation and what were observed, but these were much smaller at Pinto Valley. This led
to a detailed analysis of the performance of large mills (Society for Mining, Metallurgy,
DQG ([SORUDWLRQ  :KLWHQ DQG .DYHWVN\  DQG LW EHFDPH DSSDUHQW WKDW WKH
volume of +5.5 mm particles in the feed to the ball mill (fresh feed and circulating load)
was a crucial factor. This was high for the mills at Bougainville and much lower for mills
DW3LQWR9DOOH\%RWKPLOOVZHUHIHGIURPÀQHFUXVKHUVEXWWKHIHHGUDWHIRUWKH3LQWR
9DOOH\PLOOZDVORZHUDQGWKHEDOOPLOOFLUFXLWJDYHDÀQHUSURGXFW2QFHWKHPP
particles were discharged from the mill they were recycled to the mill feed, increasing
the effect of the coarse feed.

Comminution Handbook | Spectrum Series 21 45

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

When SAG mills and large diameter ball mills were used in series, as is now common,
WKHWUDQVIHUVL]HIURPWKH6$*PLOOWRWKHEDOOPLOOFLUFXLWZDVÀQHUWKDQIURPFUXVKHUWR
ball mill, and the problem presented by 5.5 mm particles was reduced.
The difference between actual and predicted capacities using the Bond equations
highlighted the need to use correction factors in the Bond equation according to
conditions, so design companies developed factors based on their own data; however,
WKH HPHUJHQFH RI YHUWLFDO UROOHU PLOOV DV HQHUJ\HIÀFLHQW FRPPLQXWLRQ PDFKLQHV RYHU
wide ranges of rock/particle sizes may change the concept of how large capacity
comminution circuits are designed.
This discussion has referred to the Bond grindability test and the Bond Work Index for
ball mills. There is a rod mill grindability and a Bond Rod Mill Work Index (BRMWi)
for coarser particles, typically passing 12.7 mm. The test procedure for determining
the rod mill grindability is similar to that for ball mills (Bond, 1961) but the mill
is 0.305 m × 0.610 m (with wave liners of a form described by Bond) and it runs at
40 rev/min. The charge consists of eight rods weighing a total of 33.38 kg. Initially a
 PO VXEVDPSOH RI IHHG LV SUHSDUHG IRU XVH LQ WKH ÀUVW EDWFK JULQG DQG JULQGLQJ
proceeds in cycles to develop a 100 per cent circulating load on the screen representing
the desired product. The procedure for rod mill selection is the same as the procedure
for ball mill selection that has been described. Rod mills were used during the 1950s to
VEXWWKH\KDYHQRZEHHQUHSODFHGE\ÀQHFUXVKHUV7DEOHGLVSOD\VWKHUDQJHVRI
BRMWi values that indicate ore hardness.
TABLE 4.1
Typical ranges of Bond Rod Mill Work Index values (Bailey, 2012).
Property Soft Medium Hard Very hard
Bond Rod Mill Work Index 7–9 9–14 14–20 >20

Comments on the Bond test

• The Bond test is one of the industry standards in characterising the power used in
crushing and grinding using cone crushers, rod and ball mills.
• The power prediction is fairly accurate for devices that generate a product with a
size distribution having a shape similar to the feed. It does not work well for devices
such as the AG/SAG mill and high-pressure grinding rolls where product size
distribution diverges from that of the feed.
• The closing screen for the test should be chosen to generate a test P80 as close
as possible to the P80 required for the circuit being designed. The test P80 will be
DSSUR[LPDWHO\RQH¥VLHYHLQWHUYDOEHORZWKHFORVLQJVFUHHQDSHUWXUH
Many people have applied various factors to the Bond equation in an attempt to make
LW DSSO\ WR GHYLFHV QRW FRQIRUPLQJ WR WKH GHÀQLWLRQ 7KH PRVW FRPPRQO\ XVHG ZDV
GHÀQHGE\5RZODQGDQG.MRV

HARDGROVE GRINDABILITY TEST

The Hardgrove Grindability Index (HGI) test was developed in the 1930s to determine
WKHGHJUHHRIGLIÀFXOW\HQFRXQWHUHGZKHQFRDOVZHUHSXOYHULVHGWRDVL]HUHTXLUHGIRU
HIIHFWLYHFRPEXVWLRQLQDSXOYHULVHGFRDOÀUHGERLOHU5DOSK0+DUGJURYHZRUNHGIRU

46 Comminution Handbook | Spectrum Series 21

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

PDQ\\HDUVRQSXOYHULVHUHIÀFLHQF\DQGGHYLVHGDJULQGDELOLW\WHVWIRUFRDOWKDWEHFDPH
an industry standard. Figure 4.1 reveals a part of the testing apparatus.
The test procedure is to place a 50 g sample of coal in a stationary grinding bowl in
which eight steel balls can run in a circular path. This specimen has been taken from a
larger sample of feed with a maximum size of 4.75 mm and prepared in the size range
1.18 × 0.6 mm. A loaded ring is placed on top of the balls with a gravity load of 29 kg.
After 60 revolutions, the ground sample is sieved to determine the amount of material
passing 74 μm. The HGI is calculated from the equation:

HGI = 13 + 6.93W (3)

where:
W is the weight of the particles smaller than 74 μm in grams (ASTM International,
2012)
The higher the HGI value, the easier it is to pulverise the coal. The HGI value is used as
an indicator for the power consumption by the coal pulveriser. The empirical relationship
0.91
between HGI and the Bond Work Index is Wi = 435/HGI . The usual range of HGI values
is 40 to 70, with easy-to-grind material having the higher values.

FIG 4.1 – Cross-section of Hardgrove Grindability Index test apparatus (AS 1038.20-2002
Figure 1 section AA – reproduced with permission from SAI Global Ltd under Licence 1410-c082).

ZEISEL TEST
In this test, the same apparatus is used as in the Hardgrove test but the grindability
LVH[SUHVVHGLQNLORMRXOHVN-RUNLORJUDPVNJDQGWKHVSHFLÀFVXUIDFHDUHD%ODLQH
number) of the undersized product is measured. The feed is 0.8–1 mm and 30 g of feed
is added to the apparatus at the start of the test. The revolutions are chosen to reduce
the feed to 50 per cent passing 0.125 mm. The undersized is removed and replaced by
new feed. The process is repeated with the number of revolutions selected to produce
50 per cent undersized until a steady state is reached. The Blaine number of the
undersized is then determined. This test is favoured for vertical roller mills because
LW ZRUNV E\ FRPSUHVVLRQ 7KH UHODWLRQVKLS EHWZHHQ VSHFLÀF FRPPLQXWLRQ HQHUJ\

Comminution Handbook | Spectrum Series 21 47

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

DQGVSHFLÀFVXUIDFHLVVKRZQLQ)LJXUH7KH=HLVHOWHVWLVXVHGWRFKDUDFWHULVHWKH
JULQGDELOLW\RIPDWHULDOVWKDWDUHWREHJURXQGLQDYHUWLFDOUROOHUPLOODQH[DPSOHLV
JUDQXODWHGEODVWIXUQDFHVODJ

ZĞƋƵŝƌĞĚƐƉĞĐŝĨŝĐĐŽŵŵŝŶƵƚŝŽŶĞŶĞƌŐLJ ϮϬϬ

ϭϲϬ

ϭϮϬ
;ŬtŚͬƚͿ

ϴϬ

ϰϬ

Ϭ
Ϭ ϱϬϬ ϭϬϬϬ ϭϱϬϬ ϮϬϬϬ ϮϱϬϬ ϯϬϬϬ ϯϱϬϬ ϰϬϬϬ
^ƉĞĐŝĨŝĐƐƵƌĨĂĐĞ;ĐŵϮͬŐͿ

FIG 4.2 – Relationship between specific comminution energy and specific surface (data from Cemtec).

DROP WEIGHT TEST

7KLV WHVW ZDV GHYHORSHG DW WKH -XOLXV .UXWWVFKQLWW 0LQHUDO 5HVHDUFK &HQWUH DQG ZDV
GHVLJQHG WR JHQHUDWH URFN KDUGQHVV GDWD IRU XVH LQ LWV $*6$* PRGHOV 7KH WHVW KDV
EHHQ FRPPHUFLDOLVHG E\ -.7HFK WKH WHFKQRORJ\ WUDQVIHU FRPSDQ\ IRU WKH 6XVWDLQDEOH
0LQHUDOV ,QVWLWXWH 60, DW 7KH 8QLYHUVLW\ RI 4XHHQVODQG DQG LV UHDGLO\ DYDLODEOH
ZRUOGZLGH WKURXJK PRVW PHWDOOXUJLFDO WHVW ODERUDWRULHV $ OLVW RI WKHVH LV DYDLODEOH RQ
-.7HFK·VZHEVLWH1
7KHÀUVWSDUWRIWKHWHVWPHDVXUHVWKHLPSDFWEUHDNDJHDKLJKHQHUJ\WHVW7RFRQGXFW
WKHWHVWNJRIFUXVKHGURFNLQWKHPPVL]HUDQJHRUNJRI34GULOOFRUH
LVUHTXLUHG7KHGURSZHLJKWDSSDUDWXVLVVKRZQLQ)LJXUH,WFRPSULVHVDVWHHOGURS
KHDGZKLFKLVUDLVHGE\DSQHXPDWLFZLQFKWKHQGURSSHGRQWRWKHWDUJHWURFNSDUWLFOH
DQGVXEVHTXHQWO\FUXVKHG7KHWHVWLQYROYHVEUHDNLQJVLQJOHSDUWLFOHVRIYDULRXVVL]HV
ZLWK VSHFLÀF HQHUJLHV LQ WKH UDQJH RI ² N:KW 7KH SDUWLFOH VL]HV XVHG DQG WKH
HQHUJLHVZLWKZKLFKWKH\DUHEURNHQDUHJLYHQLQ7DEOH

TABLE 4.2
Drop Weight Test particle sizes and breakage energies.
Particle size Mean size Specific energies (kWh/t)
(mm) (mm) High Medium Low
53 × 63 57.8 0.4 0.25 0.1
37.5 × 45 41.1 1.0 0.25 0.1
26.5 × 31.5 28.9 2.5 1.0 0.25
19 × 22.4 20.6 2.5 1.0 0.25
13.2 × 16 14.5 2.5 1.0 0.25

1. http://www.jktech.com.au/drop-weight-tester-licensees

48 Comminution Handbook | Spectrum Series 21

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

FIG 4.3 – Drop Weight Tester (image courtesy JKTech Pty Ltd).

7KH UDQJH RI VSHFLÀF HQHUJLHV DUH DFKLHYHG E\ DGMXVWLQJ WKH GURS KHLJKW DQG GURS
head mass.
The product from breaking the rock particles are collected and sized. From the resultant
size distributions a t10SDUDPHWHULVREWDLQHGZKLFKLVGHÀQHGDVWKHSHUFHQWSDVVLQJ
one tenth of the original particle size. Other characteristic tn values can be extracted.
If the tn characteristic values are plotted against t10, a family of curves such as those
shown in Figure 4.4 are obtained.
t10FDQEHFRQVLGHUHGDQLQGH[RIÀQHQHVVDQGLVUHODWHGWRWKHSURGXFWVL]HGLVWULEXWLRQ
that is, if t10 is known, the entire product size distribution can be generated. For a given
rock, t10LVUHODWHGWRWKHVSHFLÀFHQHUJ\DFFRUGLQJWRWKHHTXDWLRQ

ϭϬϬ

ϴϬ

ϲϬ
dϮ
ƚŶ

dϰ
ϰϬ dϭϬ
dϮϱ
ϮϬ dϱϬ

Ϭ
Ϭ ϮϬ ϰϬ ϲϬ ϴϬ
ƚϭϬ;йͿ

FIG 4.4 – t10 versus tn family of curves (image courtesy JKMRC).

Comminution Handbook | Spectrum Series 21 49

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

^-b.Ecsh
t10 = A (1 - e ) (4)
where:
(FV  LVWKHDSSOLHGVSHFLÀFHQHUJ\
t10 is the per cent passing one tenth of the original particle size
A and b are Drop Weight Test parameters that vary according to ore hardness
A detailed description of the Drop Weight Test and the data reduction procedures
that are used can be found in Napier-Munn et al (1996). The parameters A and b have
no physical meaning but it has been found that the product A×b is a useful index of
ore hardness with respect to AG and SAG mills, albeit a qualitative one. Higher values
of A×b indicate softer ore in contrast to most hardness indicators where higher values
indicate harder ores.
To avoid this inverse relationship and to provide more easily understood Drop Weight
Test results, an additional parameter has been included. This parameter, the SAG Circuit
6SHFLÀF(QHUJ\6&6(LVWKHVSHFLÀFHQHUJ\LQN:KWXWLOLVHGE\DVWDQGDUG6$*PLOO
in a closed circuit with a pebble crusher. The standard SAG mill has a 2:1 diameter-
to-length ratio, 15 per cent 125 mm balls, 25 per cent total charge and grate open area
of seven per cent, which is 100 per cent 56 mm pebble ports. The aperture size in the
trommel D50 is 12 mm and the pebble crusher has a closed side setting of 10 mm. The
UHODWLRQVKLSEHWZHHQ6&6(DQG$ðELVVKRZQLQ)LJXUH7\SLFDO6&6(YDOXHVDUH
VKRZQ LQ 7DEOH  7KH 6&6( YDOXHV ZLOO QRW QHFHVVDULO\ PDWFK WKH VSHFLÀF HQHUJ\
required for an existing or planned mill due to the differences in many operating
YDULDEOHV KRZHYHU 6&6( YDOXHV SURYLGH DQ HIIHFWLYH WRRO WR FRPSDUH WKH H[SHFWHG
behaviour of different ores in AG/SAG milling in exactly the same way that the Bond
Ball Mill Work Index is used for ball mill circuits.

FIG 4.5 – The relationship between A×b and specific energy for the standard circuit (image courtesy JKMRC).
The second part of the JK Drop Weight Test is the abrasion-breakage low-energy test,
which uses a tumbling test of a selected single-size fraction.
The standard abrasion test tumbles 3 kg of -55 +38 mm particles for ten minutes at
SHUFHQWFULWLFDOVSHHGLQDPPðPPODERUDWRU\PLOOÀWWHGZLWKðPP
lifter bars. The resulting product is then sized and the t10 value for the product is

50 Comminution Handbook | Spectrum Series 21

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

determined. The geometric mean particle size of the original size fraction -56 +38 mm
is 45.7 mm and the t10 size is 1/10 × 45.7 = 4.57 mm. The abrasion parameter, ta, is then
GHÀQHGDVta = t10/10. The three parameters relevant to AG/SAG milling are A, b and
ta. A and b are used to characterise the impact breakage of the ore and ta is a measure of
WKHUHVLVWDQFHRIWKHRUHWRDEUDVLRQVHH(TXDWLRQ,QERWKFDVHVWKHORZHUWKHYDOXH
the greater the resistance of the ore to that type of breakage. Table 4.3 indicates some
W\SLFDOÀJXUHVIRUWKH-.'URS:HLJKW7HVWSDUDPHWHUVDQGDUHODWLYHPHDVXUHRIZKDW
they mean.

TABLE 4.3
Typical parameters for the JK Drop Weight Test.
Property Very hard Hard Moderate hard Medium Moderate soft Soft Very soft
A×b <30 30–38 38–43 43–56 56–67 67–127 >127
ta <0.24 0.24–0.35 0.35–0.41 0.41–0.54 0.54–0.65 0.65–1.38 >1.38
SCSE >10.7 10.7–9.7 9.7–9.3 9.3–8.4 8.4–7.9 7.9–6.5 <6.5
SCSE – SAG Circuit Specific Energy.

Comments on the JK Drop Weight Test

• JK Drop Weight Tests are good for predicting AG/SAG milling characteristics of
an ore using JKSimMet computer software, provided the ore is brittle and does not
experience much plastic deformation before breaking. Ores that do undergo plastic
deformation, such as those with high clay content, may not be reliably characterised
by a JK Drop Weight Test.
• The JK Drop Weight Test is limited by the size of the particle that can be tested.
Results for larger particles are extrapolated from the data for the size ranges tested.
If an ore is weaker at particle sizes larger than the top size range tested (-63 +53 mm),
the results from a JK Drop Weight Test will be conservative.
• 8VLQJ WKH  HQHUJ\VL]H FRPELQDWLRQ GDWD YDOXHV UHIHU WR 7DEOH  WKH EHVW ÀW
A and b parameters are calculated. They are related to the resistance of the ore to
impact breakage and indicate the range of values that will be encountered when the
mill is operating.
• The main problem with drop-weight results is that they cannot be used directly to
calculate the power requirement for comminution devices. A JKSimMet simulation
must be conducted.
• There may be a small amount of secondary breakage as a result of the weight bouncing
after the initial impact. This is not accounted for in the calculation of energy and the
change it may have on the progeny.

SMC TEST®
The JK Drop Weight Test requires relatively large amounts of sample. Large core is
often not available as it is relatively expensive; smaller diameter core being much more
common. This precludes the use of the drop-weight test for ore hardness variability
testing and orebody hardness mapping. The SMC Test was originally developed so that
estimates of the parameters A and b could be obtained from small drill core with sample
requirements as low as 2–3 kg (Morrell, 2004a). This is achieved by breaking only one

Comminution Handbook | Spectrum Series 21 51

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

particle size using the JK Drop Weight Test, typically one of the three smaller sized
fractions used in the standard drop-weight test. Particles are obtained either by cutting
the core using a diamond saw or by crushing the core. The former is used when drill
core availability is severely limited. The SMC Test was developed to provide estimates
of the parameters A and b used in simulation modelling. To date, over 30 000 tests have
been conducted, making it one of the most popular ore hardness tests in use. A range of
60&7HVWSDUDPHWHUVLVQRZDYDLODEOHZKLFKFDQEHXVHGWRSUHGLFWWKHVSHFLÀFHQHUJ\
demand of conventional crushers, high-pressure grinding rolls (HPGR) and tumbling
mill circuits (Morrell, 2009). This is done using this power-based size-reduction equation
(Morrell, 2004b):

W = 4Mi (x 2 - x 1 )
f (x2) f (x1)
(5)

where:
W LVFLUFXLWVSHFLÀFHQHUJ\N:KW
x2 is 80 per cent passing size for the product (microns)
x1 is 80 per cent passing size for the feed (microns)
Mi is index related to the breakage property of an ore and determined by the SMC
Test (kWh/t)
f(xj) is -(0.295 + xj/1 000 000) where j denotes 1 or 2
)RUXVHLQ(TXDWLRQWKH60&7HVWJHQHUDWHVWKUHH0i parameters used as indexers,
namely:
• Mia – for coarser-particle milling and used for rod mills, AG mills and SAG mills
• Mic – for conventional crushing and used for cone, gyratory and jaw crushers
• Mih – for HPGR size-reduction machines.
(TXDWLRQSUHGLFWVWRWDOVSHFLÀFHQHUJ\RIWKHWXPEOLQJPLOOFLUFXLWIRUH[DPSOHLQD
FLUFXLWFRQWDLQLQJD6$*PLOODQGDEDOOPLOOLWSUHGLFWVWKHFRPELQHGVSHFLÀFHQHUJ\RI
the mills. A further parameter, Mib, is obtained from the raw data generated by a standard
%RQG%DOO0LOO:RUN,QGH[WHVWDQGLVWKHLQGH[IRUÀQHUSDUWLFOHPLOOLQJVXFKDVIRXQG
LQEDOOPLOOV7RREWDLQWKHVSHFLÀFHQHUJ\RIWKH$*6$*PLOODIXUWKHUHTXDWLRQZDV
developed using a large database of operating AG and SAG mills:

S = K.f 80a.DWib.(1+c(1-e-dJ))-1øe.f(Ar).g(x) (6)

where:
6  LVVSHFLÀFHQHUJ\DWWKHSLQLRQ
f80 is 80 per cent passing size of the feed
DWi is drop-weight index (from the SMC Test)
J is volume of balls (per cent)
Ø is mill speed (per cent of critical)
f(Ar  LVDIXQFWLRQRIPLOODVSHFWUDWLRXQVSHFLÀHG
J[  LVDIXQFWLRQRIWURPPHODSHUWXUHXQVSHFLÀHG
a, b, c, d are constants
K is a constant whose value is dependent upon whether a pebble crusher is used
7KHFRQVWDQWVLQ(TXDWLRQSOXVWKHIXQFWLRQVI$r) and g(x) are proprietary to Citic
SMCC Process Technology Pty Ltd.

52 Comminution Handbook | Spectrum Series 21

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

Comments on the SMC Test

• 3DUDPHWHUVIURPWKH60&7HVWDUHJRRGIRUSUHGLFWLQJWKHVSHFLÀFHQHUJ\UHTXLUHPHQW
for a range of comminution devices including conventional crushers, high-pressure
grinding rolls and tumbling mills. The range of devices where the parameters
from this ore characterisation test can be applied is wider than any other known
comminution test.
• The A and B parameters can be used for predicting AG/SAG milling characteristics
of an ore using JKSimMet, provided the ore is brittle and does not experience much
plastic deformation before breaking. Since the test uses the same device as the JK
Drop Weight Test the problems that related to ore characterisation tests using the
JK Drop Weight Test apply to the SMC Test. The methodology for the latter is a
VLPSOLÀHGYHUVLRQRIWKHIRUPHU
• 7KH 6&6( YDOXH GLVFXVVHG LQ WKH 'URS :HLJKW 7HVW VHFWLRQ LV DOVR FDOFXODWHG DQG
reported with SMC Test results.
• The methodology and analysis performed to extract the parameters used in predicting
VSHFLÀFHQHUJ\IRUPRVWRIWKHRUHFKDUDFWHULVDWLRQWHVWVLQFRPPLQXWLRQKDYHEHHQ
published, but how the Mia, Mic, Mih, Mib and DWi are obtained from SMC Test data
is yet to be reported in literature. Functions such as f(Ar) and g(x), as well as the
constants a, b, c, d and e, are proprietary to Citic SMC Process Technology.
• Companies that have used the SMC Test to design comminution circuits have
indicated that the comminution energy requirements estimated from test work
results are reasonably close to that found in the designed operating plant. This is
not surprising, since SMC established a large database of operating comminution
circuits to develop and validate the equation. SMC Test estimates are considered
reliable.

SPI® TEST
SPI stands for SAG Power Index. The SPI test is the commercial name given by the
0LQQRY(;7HFKQRORJLHVFRPSDQ\WRDODERUDWRU\RUHKDUGQHVVFKDUDFWHULVDWLRQWHVWWKH
procedure for which was originally developed by John Starkey.
The procedure involves the use of a 300 mm diameter, 100 mm long laboratory mill
loaded with 25 mm steel balls, as seen in Figure 4.6. Two kilograms of feed material are
crushed to -19 mm (P80 is approximately 12.5 mm) and ground in the mill until a product
P80 of 1.7 mm is obtained. The time in minutes to reach this point is the SPI.
The SPI parameter is used in conjunction with the following equation to predict the
VSHFLÀFHQHUJ\RI6$*PLOOV'REE\%HQQHWWDQG&RVLFN
n.
SAG (kWh/t) = K. (SPI/T 800.5) fsag (7)

where:
T80 is a rock size parameter
fsag is the submodel that incorporates effects such as feed size and pebble crushing
This is an empirical equation that predicts power used by a SAG mill from laboratory
data. SPI is determined experimentally. It is the proprietary nature of all these constants
WKDW PDNH WKH SURFHGXUH GLIÀFXOW WR XQGHUVWDQG KRZHYHU OLNH 0RUUHOO·V SURFHGXUH

Comminution Handbook | Spectrum Series 21 53

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

FIG 4.6 – SPI® test mill (image courtesy SGS).

it must be mentioned because it is widely used. The description here is kept to a

minimum. The submodel fsag and the equation contain 21 constants, enabling the equation
to be applied to AG and SAG mills working under several conditions. These constants
are proprietary to the Swiss company SGS (Société Générale de Surveillance), which
DFTXLUHG0LQQRY(;7HFKQRORJLHVLQ2YHU63,WHVWVKDYHEHHQFRQGXFWHG
making it as popular as the SMC Test.

SAGDESIGN™ TEST
7KH6$*'HVLJQWHVWZDVFUHDWHGVSHFLÀFDOO\IRUWKHVL]LQJRIQHZ6$*DQGEDOOPLOOV
UHIHU WR )LJXUH  7KH WHVW ÀUVW PHDVXUHV WKH 6$* SLQLRQ HQHUJ\ QHHGHG WR JULQG
ore from 80 per cent passing 152 mm to 80 per cent passing 1.7 mm as WSDT in kWh/t.
The second stage of the test measures the Bond Ball Mill Work Index on SAG ground
ore, Sd-BWi in kWh/t. SAGDesign ore feed is prepared from a minimum of 10 kg of
split or whole diamond drill core pieces or run-of-mine ore, by stage-crushing the ore
in a jaw crusher to 80 per cent passing 19 mm. The crushed ore is then ground in a
SAGDesign SAG mill (489 mm inside diameter × 163 mm effective grinding length) that
operates with parameters similar to commercial SAG mills (26 per cent total charge with
11 per cent steel load, 15 per cent ore load and rotation at 76 per cent of critical speed).
The calculated cumulative mill revolutions to 80 per cent passing 1.7 mm is used
WRJHWKHUZLWKWKHPDVVRIWKH6$*PLOOFKDUJHWRFRPSXWHWKH6$*PLOOVSHFLÀFSLQLRQ
energy. The process to determine the revolutions is done in cycles, with the minus
PPPDWHULDOÀQHVUHPRYHGIURPWKHPLOODIWHUHDFKF\FOHXSWRSHUFHQWSDVVLQJ
PP$IWHUWKHÀUVWJULQGLQJWDUJHWLVUHDFKHGWKDWLVSHUFHQWSDVVLQJPP
ERWKWKHÀQHDQGFRDUVHPDWHULDOVDUHUHWXUQHGWRWKHPLOOIRUWKHUHPDLQGHURIWKH6$*
mill-grinding cycles to reach the 80 per cent passing 1.7 mm test end point.
In industrial practice, few if any SAG mills create a product of exactly 80 per cent
passing 1.7 mm. In order to modify this product size to the exact size required for the

54 Comminution Handbook | Spectrum Series 21

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

FIG 4.7 – SAGDesign mill (image courtesy Starkey & Associates Inc).

design, an energy factor is added or subtracted from the SAGDesign test result. This
factor is calculated using the results of the Bond Ball Mill Work Index Test conducted
as part of the SAGDesign test on SAG ground ore and is a function of the size deviation
from 1.7 mm. Factors can also be applied to model alternative mill feed sizes and/or
pebble crushing.
The 1.7 mm closing screen for the SAG test was chosen because it is close to the feed
size F80 used in a Bond Ball Mill Work Index test and matches the coarse transfer size
T80, which is often over 2 mm in large plants. In this way SAG mill energy and ball mill
energy are measured accurately.
The SAGDesign test has some potential advantages over the SPI test in that the feed
size is coarser and arguably better, indicating that the impact of critical size on SAG mill-
VSHFLÀFHQHUJ\LVFRQVLGHUHGZKLOHWKHIHHGPDVVNJDOORZVD%RQG%DOO0LOO:RUN
Index test to be conducted on the SAG test mill product. The SAGDesign test is available
from 11 laboratories around the world. A list is provided on the Starkey & Associates Inc
website.2 SAGDesign technology is open and is patented by Outotec.

SAG variability test

An abbreviated version of the SAGDesign test is the SAG variability test (SVT), intended
for use in geometallurgical throughput study applications. The procedure of the test
LVLGHQWLFDOWRWKHIXOO6$*'HVLJQWHVWH[FHSWWKDWLWLVVWRSSHGDIWHUWKHÀUVWJULQGLQJ
target of 60 per cent passing 1.7 mm is reached. In practice the SVT test is stopped after

2. http://sagdesign.com/home/worldwide-lab-locater

Comminution Handbook | Spectrum Series 21 55

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

three cycles. From this data, the results of a completed SAGDesign test can be predicted.
7KLVUHVXOWLVGHÀQHGDV:SVT and is reported in kWh/t to grind from 80 per cent passing
152 mm to 80 per cent passing 1.7 mm.
This test is intended to provide a cost-effective solution for high volumes of samples
for SAG grinding test work. It also has the advantage that SVT samples can be retrieved
from storage so the remainder of the SAGDesign test procedure can be completed later
on the same sample. This allows process engineers to make an informed decision on
selecting the most appropriate samples for further testing in cases where maximum
accuracy is required on key samples for the project.
$ VXPPDU\ RI WKH YDULDEOH GHÀQLWLRQV UHOHYDQW WR 6$*'HVLJQ DQG 6$* YDULDELOLW\
testing is given below:
F80 80 per cent passing feed size (μm)
T80 80 per cent passing transfer size between the SAG and ball mill (μm)
P80  SHUFHQWSDVVLQJÀQDOSURGXFWVL]H—P
WSDT SAG grindability from F80 = 152.4 mm to T80 = 1700 μm from SAGDesign test
work (kWh/t)
WSVT predicted SAG grindability from F80 = 152.4 mm to T80 = 1700 μm from SAG
variability test work (kWh/t)
Sd-BWi Bond Ball Mill Work Index on SAG ground ore (kWh/t)
(*/  HIIHFWLYHJULQGLQJOHQJWK
WSAG  6$*SLQLRQHQHUJ\IURPVSHFLÀHGF80 to T80 (kWh/t)
WBM  EDOOPLOOSLQLRQHQHUJ\IURPVSHFLÀHGT80 to P80 (kWh/t)
Wtot total pinion energy from F80 to P80 (kWh/t)
6*  RUHVSHFLÀFJUDYLW\

COMPRESSION TEST
7KHPRVWHIÀFLHQWPHWKRGWREUHDNSDUWLFOHVLQWHUPVRIHQHUJ\XWLOLVDWLRQLVWRFRPSUHVV
the particle bed between two plates (Schonert, 1982). Breakage by compression occurs
in vertical roller mills, which have been used since 1900 and in high-pressure grinding
rolls, which were patented in 1982.
In the piston and die test for compression grindability seen in Figure 4.8 (Otte, 1988),
QDUURZVL]HG IUDFWLRQV DUH SUHSDUHG DQG FRPSUHVVHG WR D À[HG UDQJH RI GLIIHUHQW
pressures. Particles in the bed start moving through the voids, then breakage occurs and
this continues until the desired force is reached. The change in the height of the material
bed (displacement) during the test is recorded.
From the force-displacement graph the energy levels corresponding to displacement
are calculated from the following equation:

#
h2
W = f (h) dh (8)
h1

where:
W is work
h is displacement
f is force

56 Comminution Handbook | Spectrum Series 21

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

FIG 4.8 – Piston and die compression tester (image courtesy Hakan Dundar).
Size distributions of the compressed materials are determined by sieve analysis and
the energy–particle size distribution (PSD) relationship can be obtained for the material.
This is used to calculate the PSD for a given energy level.

BOND WORK INDEX VALUES (BOND, 1953)

7KH %RQG %DOO 0LOO :RUN ,QGH[ LV WKH ROGHVW DQG PRVW FRPPRQ IRUP RI GHÀQLQJ WKH
resistance to breakage in terms of energy consumption (kWh/t). An indicative list of
values is given for various materials in Table 4.4. The actual Bond Work Index values for

TABLE 4.4
A selection of Bond Work Index values (Metso, 2012).
Material Value Material Value Material Value
Andesite 18.25 Fluorspar 8.91 Oil shale 15.84
Barite 4.73 Gabbro 18.45 Phosphate rock 9.92
Basalt 17.10 Glass 12.31 Potash ore 8.05
Bauxite 8.78 Gneiss 20.13 Pyrite ore 8.93
Cement clinker 13.45 Gold ore 14.93 Pyrrhotite ore 9.57
Cement raw material 10.51 Granite 15.13 Quartz 13.57
Clay 6.30 Graphite 43.56 Quartzite 9.58
Coal 13.00 Gravel 16.06 Rutile ore 12.68
Coke 15.13 Gypsum rock 6.73 Shale 15.87
Copper ore 12.72 Hematite 12.84 Silica sand 14.10
Diorite 20.90 Lead ore 11.90 Silicon carbide 25.87
Dolomite 11.27 Lead-zinc ore 10.93 Slag 10.24
Emery 56.70 Limestone 12.74 Spodumene ore 10.37
Feldspar 10.80 Magnetite 9.97 Syenite 13.13
Ferro-chrome 7.64 Manganese ore 12.20 Taconite 14.61
Ferro-manganese 8.30 Magnesite 11.13 Tin ore 10.90
Ferro-silicon 10.01 Molybdenum 12.80 Titanium ore 12.33
Flint 26.16 Nickel ore 13.65 Zinc ore 11.56

Comminution Handbook | Spectrum Series 21 57

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

DSDUWLFXODUPDWHULDOFDQEHTXLWHZLGHO\GLVWULEXWHGVRWKHYDOXHIRUDVSHFLÀFPDWHULDO
must be determined by experiment.

SPLIT HOPKINSON PRESSURE BAR TEST

Split Hopkinson pressure bars (SHPB) are used to conduct ballistic impact experiments
and obtain precise measurements of material load response at high strain rates. Bertram
Hopkinson originally put forward the technique in 1914 to measure stress pulses in
EDUV +RSNLQVRQ  :KLOH KH LV FUHGLWHG ZLWK WKH LGHD LW ZDV ODWHU PRGLÀHG E\
Kolsky in 1949 to two bars in series (hence the name Split Hopkinson bar) that could
be used to measure stress and strain. The stress-time responses of ductile materials (for
example, metals) can be used to characterise their strength by conventional engineering
parameters such as yield stress and Young’s modulus.3 The SHPB device can also be
applied to ore-breakage testing by characterising the behaviour via a rock mechanics
approach.
In its simplest from, the SHPB consists of some form of launcher such as a gas gun, an
impactor known as a striker, and two bars in a series usually made out of rigid steel with
a specimen loaded in between them. The bars are mounted in a rigid framework and are
restricted from movement, except in the longitudinal direction where they move freely
DORQJEXVKLQJVRUEHDULQJV)LJXUHVKRZVDSLFWXUHDQGVLPSOLÀHGVFKHPDWLFRIWKH
arrangement as set up at the Blast Impact and Survivability Research Unit (BISRU) at the
University of Cape Town.

FIG 4.9 – Picture and schematic of Hopkinson bar apparatus (image courtesy Aubrey Mainza).

:KHQWKHODXQFKHULVÀUHGWKHVWULNHUVKRRWVRXWRIWKHEDUUHODQGLPSDFWVDJDLQVWWKH
IDFHRIWKHÀUVWEDUFDOOHGWKHLQFLGHQWEDUJHQHUDWLQJDORQJLWXGLQDOVWUHVVZDYHWKDW
propagates along it. When it interacts with the specimen, a portion of this wave follows

3. Young’s modulus is a standard characterisation parameter that describes the elastic properties of a solid
undergoing tension or compression in only one direction.

58 Comminution Handbook | Spectrum Series 21

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

WKURXJKWRWKHVHFRQGURGNQRZQDVWKHWUDQVPLWWHUEDUZKLOHWKHUHPDLQGHULVUHÁHFWHG
back along the incident bar. Both bars are mounted with strain gauges, illustrated in red
in the Figure 4.9 schematic. These output voltage signals are recorded through electronic
DPSOLÀHUVZKLFKDUHGLUHFWO\SURSRUWLRQDOWRWKHLQVWDQWDQHRXVVWUHVVDWWKHLUSRVLWLRQ
These signals can be captured and recorded over the entire loading duration by
computer and used to accurately resolve force to fracture, incident strain energy and
the fraction of this energy absorbed by the specimen during impact. For specimens of
known dimensions, deformation characteristics can be calculated, and through imaging
techniques, crack initiation and propagation can also be studied (Gross and Seelig, 2007).
SHPB is normally used to characterise material behaviour at high strain rates, so while
conventional strength tests normally apply to quasistatic-type loading environments,
this data can apply to dynamic environments. This is useful in the case of ballistic
impacts such as the design of Kevlar jackets for the military, but in comminution it is
not directly used for equipment design. In comminution, the usefulness of the SHPB is
in determining fracture criteria such as probability of fracture, cumulative damage and
strain energy of fracture. The work by Bearman, Briggs and Kojovic (1997) also suggests
that the values of Young’s modulus correlate well with parameters such as fracture
toughness (KIC), which is used in rock mechanics.

BIBLIOGRAPHY
ASTM International, 2012. D409/D409M-12 – Standard test method for the grindability of coal
using the Hardgrove machine method.
Bailey, C W, 2012. JKTech Laboratory Services Bond Rod Mill Index Test.
Bearman, R A, Briggs, C A and Kojovic, T, 1997. The application of rock mechanics parameters to
the prediction of comminution behaviour, Minerals Engineering, 10(3):255–264.
Bond, F C, 1953. Work Indexes tabulated, Mining Engineering, 5:315–316.
Bond, F C, 1961. Crushing and grinding calculations, British Chemical Engineering, 6:378–391.
Bond, F C, 1963. Metal Wear in Crushing and Grinding (Allis Chalmers Publication 07P1701).
Dobby, G, Bennett, C and Cosick, G, 2001. Advances in SAG circuit design and simulation applied
to the mine block model, in Proceedings SAG 2001, pp 221–234, Vancouver.
Engineering Toolbox, 2015. Brinell Hardness Number [online]. Available from: <http://www.
engineeringtoolbox.com/bhn-brinell-hardness-number-d_1365.html> [Accessed: 21 May
2015].
Gross, D and Seelig, T, 2007. Fracture Mechanics, with an Introduction to Micromechanics (Springer-
Verlag: Berlin).
Hopkinson, B, 1914. A method for measuring the pressure produced in the detonation of high
explosives or by the impact of bullets, Philosophical Transactions of the Royal Society, 213:437–456.
Kolsky, H, 1949. An investigation of the mechanical properties of materials at very high rates of
loading, Philosophical Proc Phys Soc London, B62:676.
Levin, J, 1989. The Bond standard grindability test and a proposal for a standard grindability test
IRUÀQHPDWHULDOVJ SAIMM, 89(1):13–21.
McKen, A and Williams, S, 2005. An overview of the small-scale tests available to characterize ore
grindability for design purposes, SGS Minerals Technical Bulletin, 2005–06.
Metso, 2012. Basics in Mineral Processing HandbookHGV3(6DQGJUHQ%%HUJOLQGDQG60RGLJK
pp 3–24.

Comminution Handbook | Spectrum Series 21 59

 CHAPTER 4 t ORE COMMINUTION MEASUREMENT TECHNIQUES

Morrell, S, 2004a. An alternative energy-size relationship to that proposed by Bond for the design
and optimisation of grinding circuits, International Journal of Mineral Processing, 74(1–4):133–
141.
Morrell,6E3UHGLFWLQJWKHVSHFLÀFHQHUJ\RIDXWRJHQRXVDQGVHPLDXWRJHQRXVPLOOVIURP
small diameter drill core samples, Minerals Engineering, 17(3):447–451.
Morrell, 6  3UHGLFWLQJ WKH RYHUDOO VSHFLÀF HQHUJ\ UHTXLUHPHQW RI FUXVKLQJ KLJK SUHVVXUH
grinding roll and tumbling mill circuits, Minerals Engineering, 22(6):544–549.
Napier-Munn, T J, Morrell, S, Morrison, R and Kojovic, T, 1996. Mineral Comminution Circuits –
Their Operation and Optimisation (Julius Kruttschnitt Mineral Research Centre: Brisbane).
Otte, O, 1988. Polycom high pressure grinding principles and industrial application, in Proceedings
Third AusIMM Mill Operators’ Conference, pp 131–136 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Rowland Jr, C A and Kjos, D M, 1980. Rod and ball mills, chapter 12, in Mineral Processing Plant
Design (eds: A L Mular and R B Bhappu), second edition (Society for Mining, Metallurgy, and
([SORUDWLRQ/LWWOHWRQ
Schonert,.0HWKRGRIÀQHDQGYHU\ÀQHFRPPLQXWLRQRIPDWHULDOVKDYLQJEULWWOHEHKDYLRXU
US patent 4357287, 2 November.
Society for Mining, Metallurgy, and Exploration, 1983. Mill design and grinding performance of
ODUJHEDOOPLOOV,DQG,,SUHVHQWHGWR60(²$,0()DOO0HHWLQJ6DOW/DNH&LW\2FWREHU
Whiten, W J and Kavetzky, A, 1984. Studies on scale-up of ball mills, Minerals and Metallurgical
Processing, May, pp 23–28.

60 Comminution Handbook | Spectrum Series 21

 HOME

Chapter 5

TUMBLING MILLS
Alban Lynch

INTRODUCTION
%DOOPLOOVZHUHWKHÀUVWWXPEOLQJPLOOV7KH\ZHUHLQYHQWHGLQWKHVWRJULQGFHPHQW
FOLQNHU EHFDXVH WKH KLJK FOLQNHU SURGXFWLRQ UDWHV RI WKH QHZ URWDU\ NLOQV UHTXLUHG
PDFKLQHV ZLWK PXFK KLJKHU FDSDFLWLHV WKDQ VWDPS PLOOV 6LQFH WKHQ EDOO PLOOV KDYH
GRPLQDWHG ÀQH JULQGLQJ 7XPEOLQJ PLOOV DUH URWDWLQJ F\OLQGHUV WKDW DUH OLQHG RQ WKH
LQVLGH7KHEDUVRQWKHOLQHUVOLIWWKHRUHDQGJULQGLQJPHGLDGXULQJHDFKUHYROXWLRQWR
DKHLJKWDWZKLFKWKHPDVVRIRUHDQGPHGLDGHWDFKHVIDOOLQJLQWRWKHWRHRIWKHPLOODV
VKRZQLQ)LJXUH&ROOLVLRQVRFFXUEHWZHHQWKHRUHJULQGLQJPHGLDDQGOLQHUVDQG
VL]H UHGXFWLRQ RI WKH RUH RFFXUV GXH WR WKH LPSDFW DQG DEUDVLRQ IRUFHV 7KH JULQGLQJ
PHGLDFDQEHVWHHORUDOXPLQDEDOOVWKHRUHLWVHOIRUERWK
&DWDUDFWLQJ DQG FDVFDGLQJ ERWK RFFXU GXULQJ WKH PLOO URWDWLRQ &DWDUDFWLQJ FDXVHV
LPSDFWEUHDNDJHZKLOHFDVFDGLQJFDXVHVDEUDVLRQDQGDWWULWLRQ7KHPLOOORDGWHQGVWR
FDVFDGHDWORZHUVSHHGVDQGFDWDUDFWDWKLJKHUVSHHGV7DEOHVKRZVWKHIDFWRUVWKDW
FDQDIIHFWEDOOPLOOSHUIRUPDQFH

FIG 5.1 – An Outotec single-compartment tumbling ball mill (left; image courtesy Outotec) and
the breakage mechanisms that operate in it (right; image courtesy the Cement Grinding Office).

Comminution Handbook | Spectrum Series 21 61

 CHAPTER 5 t TUMBLING MILLS

TABLE 5.1
The effects of design and operating variables on ball mill performance.
Design/operating variable Effect on ball mill variables
Mill diameter Capacity, feed size, power draw
Mill length Capacity, power draw
Ball load Capacity, power draw
Ball size and distribution Size reduction
Solids per cent Discharge rate, size reduction, power draw

BALL MILLS FOR CEMENT

,Q WKH FHPHQW SURGXFWLRQ SURFHVV WKHUH DUH WZR VWDJHV RI JULQGLQJ UDZ PDWHULDOV
EHIRUHWKHNLOQDQGFOLQNHUDIWHUWKHNLOQ7KHUDZPDWHULDOVDUHJURXQGWRSHUFHQW
SDVVLQJPLFURQVWRRSWLPLVHWKHFKHPLFDOUHDFWLRQVDQGWKHFHPHQWFOLQNHULVJURXQG
WR  SHU FHQW SDVVLQJ  PLFURQV WR SURGXFH FHPHQW &OLQNHU JULQGLQJ FRQVXPHV
D FRQVLGHUDEOH DPRXQW RI HQHUJ\ ² N:KW DQG PLOOV FRQWDLQLQJ RQH RU WZR
FRPSDUWPHQWVDUHWUDGLWLRQDOO\XVHGDOWKRXJKYHUWLFDOUROOHUPLOOVDUHQRZEHLQJXVHG
EHFDXVH RI WKH EHWWHU HQHUJ\ HIÀFLHQF\ WKDW WKH\ RIIHU 6LQJOH FRPSDUWPHQW EDOO PLOOV
DUHXVHGLIWKHFRDUVHFOLQNHUQRGXOHVDUHFUXVKHGÀQHEHIRUHJULQGLQJEXWPLOOVZLWK
WZRFRPSDUWPHQWVVHSDUDWHGE\DVFUHHQDUHXVHGLIWKHFOLQNHULVQRWFUXVKHG,QWZR
FRPSDUWPHQWPLOOVWKHFRDUVHFOLQNHULVEURNHQLQWKHÀUVWFRPSDUWPHQWE\KHDY\LPSDFWV
ZLWKODUJHEDOOVXSWRPP7KHVFUHHQFRQWUROVWKHVL]HRIWKHSDUWLFOHVHQWHULQJWKH
VHFRQG FRPSDUWPHQW ZKHUH VPDOOHU EDOOV XS WR  PP DUH XVHG IRU ÀQH JULQGLQJ
1RUPDOSUDFWLFHLVWRXVHDQDLUVHSDUDWRULQDFORVHGFLUFXLWZLWKEDOOPLOOVWRSURGXFH
ÀQLVKHGFHPHQWDQGFODVVLI\LQJOLQHUVLQWKHVHFRQGFRPSDUWPHQWWRVHJUHJDWHWKHEDOOV
DQGNHHSODUJHUEDOOVQHDUWKHFHQWUDOVFUHHQWRJULQGWKHFRDUVHSDUWLFOHV*ULQGLQJDLGV
DUHXVHGWRLQFUHDVHSURGXFWLRQWKH\ZRUNE\PLQLPLVLQJWKHFRDWLQJRIEDOOVE\ÀQHV
$PRGLÀHGIRUPRIWKHWZRFRPSDUWPHQWPLOOLVWKH7K\VVHQ.UXSS'RXEOH5RWDWRU
LQ ZKLFK WKH QHZ IHHG DQG WKH FODVVLÀHU UHMHFW HQWHU WKH PLOO DW GLIIHUHQW HQGV DQG
WKH SURGXFWV IURP ERWK FRPSDUWPHQWV DUH UHPRYHG IURP WKH FHQWUH RI WKH PLOO IRU
FODVVLÀFDWLRQ7KHVHPLOOVDOVRKDYHHQGGLDSKUDJPVWKURXJKZKLFKFHPHQWLVGLVFKDUJHG
WRWKHVHSDUDWRU,Q3RO\VLXVGRXEOHURWDWRUPLOOVWKHQHZIHHGDQGFRDUVHSURGXFWIURP
WKHVHSDUDWRUHQWHUWKHPLOODWGLIIHUHQWHQGVUHIHUWR)LJXUH
7KH VL]H UHGXFWLRQ WKDW RFFXUV DV FHPHQW SDVVHV WKURXJK D FRQYHQWLRQDO WZR
FRPSDUWPHQW PLOO LV VKRZQ LQ )LJXUH  7KH LQWHUPHGLDWH GLDSKUDJP FRQVLVWV RI
WZRVORWWHGGLVNVDWWDFKHGE\FXUYHGSODWHV)LJXUHZKLFKDFWDVVFRRSVWRPRYH
PDWHULDOWKDWKDVSDVVHGWKURXJKVORWVLQWKHÀUVWGLVFWRWKHVHFRQGFRPSDUWPHQW6ORWV
DUHXVXDOO\DERXWPPZLGH7KHGLDSKUDJPSUHYHQWVEDOOVPRYLQJIURPWKHÀUVWWR
WKHVHFRQGFRPSDUWPHQWDQGUHWDLQVFRDUVHUSDUWLFOHVLQWKHÀUVWFRPSDUWPHQWZKHUH
ODUJHU EDOOV FDQ JULQG WKHP PRUH HIÀFLHQWO\ 7KH  PP DSHUWXUH VFUHHQ RQO\ SHUPLWV
SDUWLFOHVOHVVWKDQPPLQVL]HWRSDVV&RQVHTXHQWO\WKHUHDUHQRFRDUVHSDUWLFOHVLQ
WKHVHFRQGFRPSDUWPHQWDQGVPDOOEDOOVFDQEHXVHGWRLQFUHDVHJULQGLQJUDWHV7\SLFDO
VL]LQJGLVWULEXWLRQVRIEDOOVLQWKHWZRFRPSDUWPHQWVDUHVKRZQLQ7DEOH
7KH ODUJHVW FHPHQW PLOOV QRZ LQ RSHUDWLRQ DUH  P LQ GLDPHWHU DQG  P ORQJ
,QVWDOOHGSRZHUH[FHHGV0:7KHPLOOVDYDLODEOHIURPVHYHUDOFRPSDQLHVDUHJLYHQ
LQ7DEOH

62 Comminution Handbook | Spectrum Series 21

 CHAPTER 5 t TUMBLING MILLS

FIG 5.2 – Examples of two-compartment Polysius mills for grinding cement clinkers –
conventional mill (left) and double rotator mill (right; images courtesy Polysius).

ϭϬϬ
ϵϬ ϭƐƚĐŚĂŵďĞƌ ϮŶĚĐŚĂŵďĞƌ
ϴϬ
ϳϬ
йZĞƚĂŝŶĞĚ

ϲϬ
ϱϬ
ϰϬ
ϯϬ
ϮϬ
ϭϬ
Ϭ
Ϭ ϭ Ϯ ϯ ϰ ϱ ϲ ϳ ϴ ϵ ϭϬ ϭϭ ϭϮ
>ĞŶŐƚŚĂůŽŶŐƚŚĞŵŝůůĂdžŝƐ;ŵͿ
ϰ͘ϳϱŵŵ ϭ͘ϭϴŵŵ Ϭ͘ϰϮϱŵŵ Ϭ͘ϭϬϮŵŵ Ϭ͘ϬϳϮŵŵ Ϭ͘Ϭϯϲŵŵ

FIG 5.3 – Size reduction through a two-compartment cement mill (image courtesy Hakan Dundar).

FIG 5.4 – Centre diaphragm (image courtesy Christian Pfeiffer Beckum).

Comminution Handbook | Spectrum Series 21 63

 CHAPTER 5 t TUMBLING MILLS

TABLE 5.2
Size ranges of grinding balls in two compartments in a cement clinker mill.
Ball size (mm) 17 20 25 30 40 50 60 70 80 90
Compartment 1 (%) 5–15 20–25 20–30 20–25 15–20
Compartment 2 (%) 10–30 20–45 10–25 15–20 5–10 5–10

TABLE 5.3
Sizes of ball mills for grinding cement clinker. Mills may have one or two compartments.
KHD Humboldt FLSmidth Polysius PSP CITIC
Wedag Engineering
Mill diameter (m) 3.0–5.8 3.8–5.8 3.4–5.4 3.4–5.4 2.4–5.8
Mill length (m) 10.0–19.0 13.0–17.0 10.1–17.0 11.0–17.0 9.0–19.0
Installed power (kW) 1100–11 500 2570–9560 1600–7800 1600–7500 900–11 500

%DOOPLOOVKDYHEHHQXVHGIRUPDQ\GHFDGHVLQWKHFHPHQWPDNLQJSURFHVVIRUJULQGLQJ
UDZPDWHULDODQGFHPHQWFOLQNHUEXWWKHUHLVQRZDWUHQGWRUHSODFHWKHPZLWKUROOHU
PLOOVEHFDXVHFRPSUHVVLRQEUHDNDJHXVHVHQHUJ\PRUHHIÀFLHQWO\WKDQLPSDFWEUHDNDJH
2I WKH  PLOOV VROG WR WKH FHPHQW LQGXVWU\ GXULQJ WKH ERRP \HDUV RI ²
SHUFHQWZHUHYHUWLFDOUROOHUPLOOVDQGSHUFHQWZHUHEDOOPLOOV)RUUDZPDWHULDO
JULQGLQJYHUWLFDOPLOOVZHUHLQVWDOOHGLQSHUFHQWRIQHZSODQWVZKLOHEDOOPLOOVZHUH
LQVWDOOHGLQMXVWSHUFHQWPDLQO\IRUSURFHVVLQJDEUDVLYHPDWHULDO:KLOHEDOOPLOOVVWLOO
DFFRXQWIRUSHUFHQWRILQVWDOOHGPLOOVIRUFOLQNHUJULQGLQJWKHLUPDUNHWVKDUHRIQHZ
PLOOVKDVGHFUHDVHGLQUHFHQW\HDUVIURPWRSHUFHQWZKLOHWKHUROOHUPLOOPDUNHW
VKDUHKDVLQFUHDVHGWRSHUFHQW
,Q VRPH SODQWV KLJKSUHVVXUH JULQGLQJ UROOV +3*5 VHH &KDSWHU  DUH LQVWDOOHG
DKHDGRIEDOOPLOOVWRGHFUHDVHHQHUJ\SHUWRQQHE\SURGXFLQJDODUJHDPRXQWRIÀQHV
,QRWKHUSODQWVYHUWLFDOUROOHUPLOOVDUHXVHGWRPDNHÀQLVKHGFHPHQW(QHUJ\XVHLQ
HDFKFLUFXLWLV
• EDOO PLOOV ²  N:KW
• +3*5 SOXV EDOO PLOO ² ² N:KW
• YHUWLFDO UROOHU PLOOV ² ² N:KW
%DOO PLOOV KDYH DGYDQWDJHV IRU JULQGLQJ FHPHQW FOLQNHU 7KH\ JHQHUDWH KHDW WR SDUWLDOO\
GHK\GUDWH J\SVXP WKHLU FDSLWDO FRVW LV ORZHU WKDQ IRU UROOHU PLOOV WKHLU DYDLODELOLW\ LV
KLJK DQG WKH\ DUH HDV\ WR RSHUDWH
&ODVVLÀHUV DUH LPSRUWDQW LQ WKH SHUIRUPDQFH RI FOLQNHU JULQGLQJ FLUFXLWV DQG RQH
PHWKRG RI LPSURYLQJ SURGXFWLYLW\ DW ORZ FRVW LV WR UHSODFH ROGHU FODVVLÀHUV ZLWK PRGHUQ
KLJKHIÀFLHQF\ XQLWV &ODVVLÀHUV DUH GLVFXVVHG LQ &KDSWHU 
:KHQ ODUJHU PLOOV ZHUH EXLOW LQ WKH V WR JULQG PRUH FOLQNHU WKH VL]HV RI WKH PRWRU
DQG SRZHU WUDQVPLVVLRQ V\VWHPV DOVR LQFUHDVHG UHVXOWLQJ LQ PDLQWHQDQFH SUREOHPV
ZLWK WKH SRZHU WUDQVPLVVLRQ V\VWHPV 7KHVH OHG WR WKH GHYHORSPHQW RI JHDUOHVV
PLOO GULYHV LQ ZKLFK WKH PRWRU URWRU HOHPHQWV DUH DWWDFKHG WR WKH PLOO VKHOO DQG DUH
VXUURXQGHG E\ D VWDWLRQDU\ VWDWRU DVVHPEO\ (OHFWURQLFV FRQYHUW WKH LQFRPLQJ FXUUHQW
IURP  +] WR DERXW  +] DQG WKH PLOO VKHOO EHFRPHV WKH URWDWLQJ HOHPHQW RI D ODUJH
ORZVSHHG V\QFKURQRXV PRWRU 7KH ÀUVW JHDUOHVV GULYH ZDV LQVWDOOHG RQ D FHPHQW PLOO

64 Comminution Handbook | Spectrum Series 21

 CHAPTER 5 t TUMBLING MILLS

DW /H +DYUH LQ  DQG PDQ\ ZHUH LQVWDOOHG LQ WKH IROORZLQJ \HDUV *HDUOHVV PLOO
GULYH WHFKQRORJ\ ZDV ZHOO GHYHORSHG E\ WKH WLPH WKDW LW ZDV WUDQVIHUUHG WR ODUJH RUH
PLOOV DURXQG 

BALL/TUBE MILLS FOR COAL

7KHUPDO FRDO LV WKH IXHO XVHG LQ PDQ\ SRZHU VWDWLRQ IXUQDFHV )RUW\RQH SHU FHQW RI WKH
ZRUOG·V WRWDO HOHFWULFLW\ JHQHUDWLRQ LV IURP FRDO ZKLFK PXVW EH JURXQG WR ² SHU FHQW
SDVVLQJ  PLFURQV WR HQVXUH FRPSOHWH FRPEXVWLRQ ZLWKLQ WKH IXUQDFH )DFWRUV FRQVLGHUHG
LQ VHOHFWLQJ D SXOYHULVHU IRU FRDO LQFOXGH
• Grindability ² FRDO ZLWK KLJKHU JULQGDELOLW\ UHTXLUHV KLJKHU JULQGLQJ FDSDFLW\
• Abrasiveness ² D FULWLFDO IDFWRU LQ PLOO ZHDU PLOO DYDLODELOLW\ DQG FRVW RI RSHUDWLRQ
• Mineral matter (ash) amount and nature ² WKLV DIIHFWV WKH DPRXQW RI PDWHULDO WKDW WKH PLOO
PXVW SURFHVV IRU D JLYHQ SURGXFW HQHUJ\ DQG WKH FRPPLQXWLRQ DQG GU\LQJ SURFHVVHV
• Moisture ² DV FRDO PLOOV DUH DOVR GU\LQJ GHYLFHV KLJKPRLVWXUH ORDGV FDQ FDXVH D
UHGXFWLRQ LQ PLOO WKURXJKSXW DQG DQ LQFUHDVH LQ PLOO EORFNDJHV
• Throughput ² WKH PLOO LV D NH\ HOHPHQW LQ WKH IXHO GHOLYHU\ V\VWHP DQG PXVW EH DEOH WR
FRSH ZLWK D ZLGH UDQJH RI WKURXJKSXWV WR PHHW ERLOHU ORDGV
• Mill response ² SRZHU JHQHUDWLRQ LV D G\QDPLF SURFHVV LQ ZKLFK SODQW RXWSXW PRYHV
FRQVWDQWO\ DQG RYHU D ZLGH UDQJH 7KH DPRXQW RI IXHO GHOLYHUHG WR WKH ERLOHU YDULHV
DQG WKH PLOO PXVW UHVSRQG WR FKDQJLQJ ORDGV
• Power requirements ² FRDO PLOOLQJ XVXDOO\ UHSUHVHQWV D VLJQLÀFDQW SRUWLRQ RI WKH HOHFWULF
SRZHU XVHG ZLWKLQ D SRZHU SODQW 0LQLPLVLQJ WKH SRZHU GHPDQG LV D VLJQLÀFDQW
HIÀFLHQF\ PHDVXUH
9HUWLFDO VSLQGOH PLOOV DUH WKH SXOYHULVHUV XVHG LQ PRVW WKHUPDO FRDO SRZHU VWDWLRQV EXW
EDOOWXEH PLOOV DUH SUHIHUUHG IRU YHU\ DEUDVLYH DQG ORZPRLVWXUH FRDOV DQG IRU GLIÀFXOW
PDWHULDOV VXFK DV SHWUROHXP FRNH
$ EDOOWXEH PLOO FLUFXLW LQ D SRZHU VWDWLRQ FRQVLVWV RI WKUHH FRPSRQHQWV D FUXVKHU
GU\HU ZKLFK EUHDNV WKH FRDO IURP WKH PLQH WR OHVV WKDQ  PP DQG GRHV VRPH GU\LQJ
WKH EDOO PLOOGU\HU DQG WKH FODVVLÀHU ZKLFK SURGXFHV ÀQH SDUWLFOHV IRU WKH ERLOHU DQG
UHWXUQV FRDUVH SDUWLFOHV WR WKH PLOO
&RDO EDOOWXEH PLOOV DUH GHVLJQHG VR WKDW IHHG FDQ HQWHU WKH PLOO DW RQH RU ERWK HQGV
+RW DLU ÁRZV LQWR WKH PLOO ZLWK WKH IHHG DQG WUDQVSRUWV WKH GU\ SXOYHULVHG FRDO IURP WKH
FDWDUDFWLQJ ]RQH WR D VHSDUDWRU ,W WKHQ WUDQVSRUWV WKH ÀQH SURGXFW WR WKH ERLOHU 7KH ÁRZ
RI DLU PXVW EH FDUHIXOO\ FRQWUROOHG WR DYRLG VWULSSLQJ WRR PDQ\ FRDUVH SDUWLFOHV IURP WKH
PLOO ,I WKLV RFFXUV WKH FLUFXODWLQJ ORDG ZLOO EH H[FHVVLYH 7DEOH  SUHVHQWV VRPH W\SLFDO
GDWD IRU FRDO EDOOWXEH PLOOV
TABLE 5.4
Sizes and capacities of ball/tube mills for pulverising coal (data from Metso: Babcock Riley).
Operating characteristic Values
Mill diameter (m) 3.8–5.5
Mill length (m) 5.8–8.2
Motor maximum (kW) 820–2760
Throughput (mt/h) 42–141

Comminution Handbook | Spectrum Series 21 65

 CHAPTER 5 t TUMBLING MILLS

7KH ÁRZV RI KRW DLU DQG FRDO WKURXJK WKH PLOO DUH VKRZQ LQ )LJXUH  ,Q &KLQD PRUH
WKDQ  SXOYHULVHUV DUH XVHG LQ SRZHU SODQWV ZLWK  SHU FHQW RI WKHVH YHUWLFDO UROOHU
PLOOV ,Q   EDOOWXEH PLOOV DQG  YHUWLFDO UROOHU PLOOV ZHUH LQVWDOOHG LQ QHZ SRZHU
SODQWV LQ &KLQD

FIG 5.5 – Double-ended ball/tube mill for grinding coal (image courtesy Foster Wheeler Global Power Group).

BALL MILLS FOR ORES

7KH ÀUVW XVH RI EDOO PLOOV WR JULQG RUHV ZDV IRU WKH :LWZDWHUVUDQG JROG RUHV LQ 6RXWK
$IULFD LQ DERXW  )LQHU JULQGLQJ ZDV QHFHVVDU\ EHFDXVH ORVV RI JROG LQ FRPSRVLWH
SDUWLFOHV GXULQJ F\DQLGH OHDFKLQJ ZDV KLJK DQG EHWWHU H[SRVXUH RI WKH JROG WR F\DQLGH
ZDV UHTXLUHG %DOO PLOOV XVHG WR JULQG FHPHQW ZHUH WHVWHG DQG IRXQG WR EH VXFFHVVIXO ZLWK
PLOO GLDPHWHUV RI ² P DQG OHQJWKV RI ² P *ULQGLQJ PHGLD ZHUH ZHDUUHVLVWDQW
SHEEOHV IURP 1RUWK 6HD EHDFKHV EXW WKH\ ZHUH H[SHQVLYH DQG WKH RUH LWVHOI ZDV IRXQG
WR EH VXLWDEOH %DOO PLOOV EHFDPH WKH SUHIHUUHG ÀQHJULQGLQJ GHYLFHV DQG GXULQJ WKH QH[W
 \HDUV PDQXIDFWXUHUV FRQVLVWHQWO\ LQFUHDVHG PLOO VL]HV WR KDQGOH KLJKHU WKURXJKSXWV
DV VKRZQ LQ 7DEOH 
TABLE 5.5
Increase in the size and power of ball mills during the 20th century.
Year 1909 1912 1927 1940 1963 1970 1990 1997 2012
Diameter (m) 1.2 1.9 2.4 3.0 3.9 5.6 6.1 7.3 8.5
Length (m) 2.1 2.3 2.4 2.8 5.5 6.4 9.3 10.5 13.4
kW 11 41 168 447 1491 3169 5593 10 440 22 000

%HIRUH WKH V KLJK SODQW FDSDFLWLHV ZHUH DFKLHYHG E\ RSHUDWLQJ PXOWLSOH EDOO PLOO
FODVVLÀHU FLUFXLWV LQ SDUDOOHO 7\SLFDOO\ WKH IHHG IRU VXFK EDOO PLOOV LV SUHSDUHG E\ ÀQH
FUXVKLQJ RU URG PLOOLQJ $ ÀQH FUXVKHG IHHG LV IHG E\ FRQYH\RU GLUHFWO\ LQWR WKH EDOO PLOO
IHHG ,I WKH EDOO PLOO LV DIWHU D URG PLOO WKH IHHG ZLOO FRQWDLQ D ODUJH DPRXQW RI ÀQHV DQG
ZLOO DOUHDG\ EH LQ D VOXUU\ ,Q WKLV LQVWDQFH WKH IHHG LV ÀUVW VHQW WR D FODVVLÀHU DQG RQO\ WKH
FRDUVH IUDFWLRQ LV VHQW WR WKH EDOO PLOO &LUFXLWV RI WKLV W\SH DUH VWLOO XVHG EXW QRZ XVXDOO\
UHFHLYH IHHG IURP D VHPLDXWRJHQRXV JULQGLQJ 6$* RU DXWRJHQRXV JULQGLQJ $* PLOO

66 Comminution Handbook | Spectrum Series 21

 CHAPTER 5 t TUMBLING MILLS

7DEOH  JLYHV WKH GDWD IURP D FLUFXLW LQ ZKLFK RUH ZLWK D %RQG %DOO 0LOO :RUN ,QGH[
%%0:L RI  N:KW LV JURXQG LQ D  P ð  P EDOO PLOO LQ D FORVHG FLUFXLW ZLWK IRXU
 PP F\FORQHV $ GLDJUDP RI WKH FLUFXLW LV VKRZQ LQ )LJXUH 
7KH HUD RI YHU\ ODUJH EDOO PLOOV VWDUWHG LQ WKH V EHFDXVH QHZ HOHFWULFDO DQG
PHFKDQLFDO HQJLQHHULQJ WHFKQLTXHV PDGH WKHP SRVVLEOH $Q LPSRUWDQW XVH IRU WKHVH
ODUJH PLOOV LV UHJULQGLQJ 6$* PLOO SURGXFW WR HQVXUH DGHTXDWH PLQHUDO OLEHUDWLRQ
7KH ODUJHVW EDOO PLOOV FXUUHQWO\ LQ XVH DUH  P LQ GLDPHWHU E\  P ORQJ DQG UHTXLUH
  0: RI LQVWDOOHG SRZHU UHIHU WR 7DEOH 

TABLE 5.6
Typical stream sizings in a ball mill/cyclone circuit.
Size (mm) Circuit feed Ball mill discharge Cyclone underflow Cyclone overflow
2.36 98.8 98.6 98.5 100
1.18 78.6 96.3 91.5 100
0.60 52.3 90.9 82.2 100
0.30 38.6 76.6 65.3 98.9
0.15 29.9 47.3 35.8 91.3
0.075 23.9 25.4 16.3 72.2
0.038 19.2 16.3 9.9 53.8
Tonnes per hour 110 592 592 110
Solids (%) 95.6 69.4 72.9 34.3

FIG 5.6 – Diagram of the circuit used to gather the data in Table 5.6 (image courtesy Hakan Dundar).

TABLE 5.7
Sizes and motor powers of large ball mills.
Metso Polysius Outotec FLSmidth CITIC
Diameter (m) 7.9 7.3 8.5 8.2 7.9
Length (m) 12.5 12.5 13.4 13.1 13.6
Installed power (MW) 15.0 13.3 22.0 18.6 17.0

Comminution Handbook | Spectrum Series 21 67

 CHAPTER 5 t TUMBLING MILLS

$V EDOO PLOOV EHFDPH ODUJHU WKH VL]H RI WKH GULYH V\VWHPV DOVR LQFUHDVHG DQG WKH
VHOHFWLRQ RI D GULYH V\VWHP EHFDPH D FUXFLDO IDFWRU LQ WKH GHVLJQ RI D FLUFXLW 7KH OLPLW RI
SRZHU WUDQVPLVVLRQ WKURXJK JHDUV LV FXUUHQWO\  N: SHU SLQLRQ DQG   N: IRU D
GXDOSLQLRQ PLOO 7KH DOWHUQDWLYH WR JHDUV LV D ZUDSDURXQG PRWRU RI WKH W\SH GLVFXVVHG
IRU ODUJH FOLQNHU JULQGLQJ PLOOV 7KHUH LV DQ RYHUODS ]RQH RI ² 0: ZKHUH JHDUHG
DQG JHDUOHVV GULYHV FDQ EH XVHG EXW IRU PRUH WKDQ  0: RQO\ WKH JHDUOHVV GULYH KDV
EHHQ XVHG $V HQJLQHHULQJ DQG PDQXIDFWXULQJ LPSURYH JHDUHG GULYHV ZLOO EH DEOH WR
EH XVHG DW KLJKHU SRZHU OHYHOV &XUUHQW HQJLQHHULQJ GHVLJQV DUH ORRNLQJ DW JHDUHG PLOOV
DSSURDFKLQJ  0:

AUTOGENOUS/SEMI-AUTOGENOUS GRINDING MILLS

$* PLOOV DUH WXPEOLQJ PLOOV WKDW SURGXFH JULQGLQJ PHGLD IURP WKHLU RZQ IHHG 7R HQDEOH
WKLV WR RFFXU WKH IHHG LV FRDUVH SHEEOHV DQG URFNV VXFK DV SULPDU\ FUXVKHU GLVFKDUJH
&RPPLQXWLRQ RFFXUV E\ ODUJH URFNV VKDWWHULQJ WKH VPDOOHU SHEEOHV GXULQJ FROOLVLRQV LQ WKH
WRH RI WKH PLOO DQG WKHPVHOYHV EHLQJ DEUDGHG XQWLO WKH\ DUH VPDOO HQRXJK WR EH VKDWWHUHG
$* PLOOV ZHUH XVHG WR JULQG :LWZDWHUVUDQG JROG RUHV LQ 6RXWK $IULFD EHFDXVH WKH RUH
KDG KDUG DQG VRIW FRPSRQHQWV DQG PDGH H[FHOOHQW JULQGLQJ PHGLD EXW RUH YDULDELOLW\
KDG VXFK D ODUJH LQÁXHQFH RQ WKH SURFHVV WKDW LW ZDVQ·W XQWLO  WKDW HQJLQHHUV ZHUH
FRQÀGHQW HQRXJK WR XVH WKH SURFHVV RQ D ODUJH VFDOH ZLWK DQRWKHU RUH 7KLV ZDV LURQ
RUH LQ 4XHEHF DQG WKH SURFHVV ZDV VXFFHVVIXO EHFDXVH WKH RUH ZDV FRQVLVWHQW DQG WKH
JULQGLQJ ORDG UHPDLQHG FRQVWDQW $ SUREOHP LQ PDQ\ ODWHU SODQWV ZDV WKH DFFXPXODWLRQ
RI ² PP SHEEOHV LQ WKH PLOO EHFDXVH WKH\ ZHUH ODUJH HQRXJK WR UHVLVW VKDWWHU ² VR WKH\
RQO\ DEUDGHG VORZO\ ² EXW WRR VPDOO WR SURYLGH XVHIXO EUHDNDJH LPSDFWV 7KH VROXWLRQ WR
EUHDNLQJ WKHVH SUREOHP SHEEOHV ZDV WR DGG ODUJH VWHHO EDOOV WR WKH PLOO ÀYH WR WHQ SHU FHQW
RI PLOO YROXPH WR VKDWWHU WKHP ² DQG WKLV ZDV WKH RULJLQ RI 6$* PLOOV 7KH DGGLWLRQ RI
VWHHO EDOOV EHFDPH DQ LPSRUWDQW SDUW RI 6$* PLOO SUDFWLFH DQG ZLOO EH GLVFXVVHG EULHÁ\

Ball charge and ball size in semi-autogenous grinding mills

Ball charge
6WHHO EDOOV DUH H[WUHPHO\ HIIHFWLYH JULQGLQJ PHGLD IRU EUHDNLQJ KDUG RUHV LQ 6$* PLOOV
EHFDXVH WKH\ FDQ JHQHUDWH YHU\ KLJK NLQHWLF HQHUJLHV IRU LPSDFW EUHDNDJH 7KXV WKH
LQWURGXFWLRQ RI VWHHO EDOOV UHVXOWV LQ HQKDQFHG EUHDNDJH DQG WKH LPSDFW EUHDNDJH WKDW
RFFXUV HQDEOHV D PRUH UDSLG EUHDNGRZQ RI FRDUVHU URFNV WKDQ WKH VORZHU DWWULWLRQ DQG
DEUDVLRQ EUHDNDJH PHFKDQLVP $Q LQFUHDVHG EDOO FKDUJH DOVR UHGXFHV WKH DPRXQW RI
LQWHUPHGLDWHVL]HG URFNV W\SLFDOO\ LQ WKH ² PP UDQJH $V D JHQHUDO UXOH JULQGLQJ RI
ÀQHU SDUWLFOHV LV IDYRXUHG E\ WKH XVH RI VPDOOHU PHGLD KHQFH WKH UHPRYDO RI ² PP
URFN SDUWLFOHV ZLOO UHGXFH WKH DELOLW\ RI D PLOO WR JULQG ÀQHU SDUWLFOHV 7KLV HIIHFW RQ WKH
SURGXFW VL]H GLVWULEXWLRQ FDQ EH UHGXFHG E\ LQFUHDVLQJ WKH EDOO FKDUJH YROXPH WR KLJK
OHYHOV W\SLFDOO\ ! SHU FHQW E\ YROXPH :KHQ WUHDWLQJ VRIW RUHV ZLWK D YHU\ ÀQH IHHG
VL]H GLVWULEXWLRQ 6$* PLOOV RSHUDWH PRUH OLNH D ODUJH EDOO PLOO ZKHUH PRVW RI WKH URFN
SDUWLFOHV DUH RU UDSLGO\ EUHDN LQWR D PXFK VPDOOHU VL]H WKDQ WKH VWHHO EDOOV 7KHUH LV YHU\
OLWWOH URFN WKDW LV KDUG DQG ODUJH HQRXJK WR DFW DV JULQGLQJ PHGLD 7KHUHIRUH KLJK EDOO
FKDUJH YROXPH LV UHTXLUHG WR SURYLGH D VXIÀFLHQW VWHHOJULQGLQJ PHGLD VXUIDFH DUHD DQG
XQXVXDOO\ KLJK OHYHOV DUH XVHG LQ VRPH FLUFXPVWDQFHV  WR  SHU FHQW E\ YROXPH
/LQHU SURÀOHV PXVW EH DGMXVWHG WR EH OHVV DJJUHVVLYH DQG PRUH VLPLODU WR D EDOO PLOO
WR DYRLG OLQHU GDPDJH 7R D ODUJH H[WHQW WKH FKRLFH RI EDOO VL]H ZLOO EH GLFWDWHG E\ WKH

68 Comminution Handbook | Spectrum Series 21

 CHAPTER 5 t TUMBLING MILLS

KDUGQHVV RI WKH RUH DQG LWV IHHG VL]H GLVWULEXWLRQ 6RIWHU DQGRU ÀQHU RUH ZLOO UHTXLUH
VPDOOHU EDOOV ZKLOH KDUGHUFRDUVHU RUH ZLOO UHTXLUH ODUJHU EDOOV

Ball size
/DUJHU GLDPHWHU EDOOV SURGXFH JUHDWHU NLQHWLF HQHUJ\ SHU LPSDFW PHDQLQJ WKDW WKH\
FDQ EUHDN ODUJHU DQG KDUGHU URFNV PRUH HIIHFWLYHO\ KRZHYHU WKHUH DUH IHZHU EDOOV
SHU XQLW YROXPH DQG WKXV OHVV IUHTXHQW EUHDNDJH LPSDFWV WKDQ IRU VPDOOHU GLDPHWHU
EDOOV 3DUWLFXODUO\ IRU KDUG RUHV WKH EUHDNDJH UDWH RI ODUJH SDUWLFOHV LV LQFUHDVHG E\
XVLQJ ODUJHU GLDPHWHU EDOOV GXH WR WKH DGGLWLRQDO NLQHWLF HQHUJ\ $V WKH SDUWLFOH VL]H
GHFUHDVHV WKH HIIHFW RI LQFUHDVHG NLQHWLF HQHUJ\ WKDW ODUJHU EDOOV SURYLGH EHFRPHV
OHVV SURQRXQFHG DQG LV RYHUVKDGRZHG E\ WKH UHGXFWLRQ LQ EDOO QXPEHUV 7KLV LV
SDUWLFXODUO\ QRWLFHDEOH LQ WKH LQWHUPHGLDWH RU FULWLFDO VL]H UDQJH RI ² PP ZKHUH
D ODUJHU QXPEHU RI VPDOOHU EDOOV FDQ UHVXOW LQ KLJKHU EUHDNDJH UDWHV GXH WR WKH KLJKHU
IUHTXHQF\ RI LPSDFWV :KHQ WUHDWLQJ VRIW RUHV ZLWK YHU\ ÀQH IHHG VL]H WKDW LV ZKHQ
RSHUDWLQJ OLNH D ODUJH EDOO PLOO VPDOOHU EDOOV DQG D ODUJH VXUIDFH DUHD DUH PRUH HIIHFWLYH
GXH WR WKH LQFUHDVHG IUHTXHQF\ RI LPSDFWV DQG JUHDWHU VXUIDFH DUHD ZKLFK SURPRWHV
DWWULWLRQ DQG DEUDVLRQ EUHDNDJH PHFKDQLVPV

Characteristics of semi-autogenous grinding mills

$Q H[DPSOH RI D 6$* PLOO ZLWK D JHDUOHVV GULYH DQG WKH VWDQGDUG 6$* PLOOEDOO PLOO
FLUFXLW DUUDQJHPHQW DUH VKRZQ LQ )LJXUH 
7KH DGYDQWDJHV RI $*6$* PLOOV DUH WKDW WKH\
• VLPSOLI\ FLUFXLWV EHFDXVH WKH\ UHSODFH FUXVKLQJ FLUFXLWV DQG URG PLOOV
• FDQ EH VFDOHG XS WR YHU\ ODUJH VL]HV DQG FDSDFLWLHV UHVXOWLQJ LQ ORZHU FDSLWDO DQG
PDLQWHQDQFH FRVWV EXW QRW QHFHVVDULO\ LQ KLJKHU HIÀFLHQFLHV
:KLOH $*6$* PLOOV KDYH GLVDGYDQWDJHV WKH\ DUH FDVHVSHFLÀF DQG WKXV RXWVLGH WKH
VFRSH RI WKLV ERRN
,Q  WKH ODUJHVW 6$* DQG EDOO PLOOV UHTXLUHG  0: DQG  0: PRWRUV UHVSHFWLYHO\
%\  WKLV KDG LQFUHDVHG WR  0: DQG  0: 7KH UDSLG LQFUHDVH LQ RUH SURGXFWLRQ
IURP ORZJUDGH SRUSK\U\ FRSSHU RUHV WKDW RFFXUUHG LQ WKH HDUO\ V UHVXOWHG LQ YHU\
ODUJH $*6$* PLOOV EHLQJ EXLOW ZLWK WKH ODUJHVW 6$* PLOOV QRZ UHTXLULQJ   N:
PRWRUV  P ð  P 6$* PLOOV  P ð  P $* PLOOV /DUJHU PLOOV DUH SODQQHG

FIG 5.7 – A semi-autogenous grinding (SAG) mill (left; image courtesy FLSmidth) and
a typical SAG mill/ball mill circuit (right; image courtesy Hakan Dundar).

Comminution Handbook | Spectrum Series 21 69

 CHAPTER 5 t TUMBLING MILLS

7KH SURGXFW VL]H IURP $* PLOOV WHQGV WR EH ÀQHU WKDQ IURP 6$* PLOOV GXH WR OHVV LPSDFW
DQG PRUH DEUDVLRQ EUHDNDJH EXW ERWK PLOOV DUH XVXDOO\ IROORZHG E\ D EDOO PLOO WR SURGXFH
ÀQH ÁRWDWLRQ IHHG 6$* PLOOV ZLWK EDOOV ÀOOLQJ DV PXFK DV  SHU FHQW RI PLOO YROXPH DUH
WKH SUHIHUUHG PDFKLQHV IRU KLJKFDSDFLW\ FLUFXLWV WKDW FDUU\ RXW ÀQH FUXVKLQJ DQG FRDUVH
JULQGLQJ 6RPH VL]H GLVWULEXWLRQV DURXQG D W\SLFDO FLUFXLW FRQWDLQLQJ D 6$* PLOO EDOO
PLOO DQG K\GURF\FORQHV DUH VKRZQ LQ )LJXUH 

100
90

80

70
60
ƉĂƐƐŝŶŐй

50 Fresh
Feed
40
SAG
30 Product
Trommel
20 U/S
10 Cyclone
O/F
0
0.01 0.1 1 10 100 1000
Ɖ͘ƐŝnjĞ;ŵŵͿ

FIG 5.8 – Typical size distributions in a semi-autogenous grinding (SAG) mill / ball mill circuit. Trommel undersize
(U/S) is the SAG mill product after pebbles are removed. Cyclone O/F – overflow (image courtesy Hakan Dundar).

7KH FKRLFH RI DQ $* RU 6$* PLOO GHSHQGV RQ WKH SURSHUWLHV RI WKH RUH LQ SDUWLFXODU
LWV FRPSHWHQFH ZKLFK LV QRW WKH VDPH DV RUH KDUGQHVV 2UH FRPSHWHQFH LV D PHDVXUH
RI WKH UHVLVWDQFH RI SDUWLFOHV UDQJLQJ EHWZHHQ ² PP WR EUHDNDJH DQG LV UHOHYDQW
WR DXWRJHQRXV JULQGLQJ 2UH KDUGQHVV LV D PHDVXUH RI WKH UHVLVWDQFH RI SDUWLFOHV ÀQHU
WKDQ  PP WR EUHDNDJH DQG LV UHOHYDQW WR EDOO PLOOLQJ $Q $* PLOO UHTXLUHV FRPSHWHQW
RUH WR SURYLGH WKH PHGLD LQ WKH PLOO ,I WKHUH LV LQVXIÀFLHQW FRPSHWHQW RUH LQ WKH PLOO
IHHG RU WKH FRPSHWHQFH LV KLJKO\ YDULDEOH 6$* PLOOLQJ PXVW EH XVHG WR HQVXUH WKDW
VXIÀFLHQW JULQGLQJ PHGLD DUH SUHVHQW 6HOHFWLQJ 6$* PLOOLQJ UHVXOWV LQ H[WUD H[SHQVH IRU
VWHHO EDOOV DQG LQFUHDVHG OLQHU ZHDU KRZHYHU EHFDXVH WKH HQHUJ\ LQWHQVLW\ LV KLJK IRU
6$* PLOOLQJ LW LV XVXDOO\ SRVVLEOH WR UHGXFH FDSLWDO FRVW E\ LQVWDOOLQJ D VPDOOHU PLOO RU
IHZHU PLOOV &ULWLFDO VL]H SDUWLFOHV PD\ EXLOG XS ZLWK HLWKHU $* RU 6$* PLOOLQJ EXW WKH\
FDQ EH PDQDJHG E\ LQVWDOOLQJ SRUWV LQ WKH GLVFKDUJH JUDWH WKURXJK ZKLFK WKH SHEEOHV DUH
UHPRYHG DQG VXEVHTXHQWO\ FUXVKHG
&KRRVLQJ JHDUHG RU JHDUOHVV GULYH V\VWHPV LQ ODUJH $* DQG 6$* PLOOV LV RI SULPH
LPSRUWDQFH 7KH DGYDQWDJHV RI JHDUHG GULYHV DUH ORZHU FDSLWDO FRVW HDVH RI LQVWDOODWLRQ
DQG OHVV PDLQWHQDQFH 7KH SUDFWLFDO OLPLW RI PLOO VL]H IRU D JHDUHG GULYH LV FXUUHQWO\
 P  IHHW $OO PRGHUQ 6$* PLOOV DUH YDULDEOH VSHHG GULYH DQG JHDUHG GULYHV QHHG
WR KDYH WKLV IHDWXUH LQFOXGHG DV D VHSDUDWH LWHP WR WKH PRWRU *HDUOHVV GULYHV KDYH ORZHU
RSHUDWLQJ FRVWV DQG VSHHG FDQ LQKHUHQWO\ YDU\ IURP ]HUR WR  SHU FHQW 9DULDEOH VSHHG
UHGXFHV OLQHU GDPDJH E\ VORZLQJ WKH PLOO ZKHQ RUH FRPSHWHQFH IDOOV DQG DOORZV WKH PLOO
WR EH VSHG XS IRU KLJKFRPSHWHQFH RUHV DQG WR PDLQWDLQ WKURXJKSXW DV OLIWHUV ZHDU
/DUJH 6$* DQG EDOO PLOOV DUH RFFDVLRQDOO\ PRGLÀHG WR KDQGOH VSHFLÀF SUREOHPV
([DPSOHV IURP 2XWRWHF LQFOXGH WKH GHVLJQ DQG PDQXIDFWXUH RI DFLGUHVLVWDQW PLOOV IRU

70 Comminution Handbook | Spectrum Series 21

 CHAPTER 5 t TUMBLING MILLS

XVH ZKHQ JULQGLQJ LV FDUULHG RXW LQ UHWXUQ OLTXRU IURP DQ DFLG OHDFK6;(: VROYHQW
H[WUDFWLRQ DQG HOHFWURZLQQLQJ SURFHVV DQG RI SXOS OLIWHUV WR LPSURYH WKH GLVFKDUJH RI
VOXUU\ IURP KLJKYROXPH 6$* PLOOV 7KH ODUJHVW 6$* PLOOV RIIHUHG E\ WKH PDMRU YHQGRUV
DUH VKRZQ LQ 7DEOH 

TABLE 5.8
Sizes and motor powers of large autogenous and semi-autogenous grinding mills.
Metso Polysius Outotec FLSmidth CITIC
Maximum diameter (m) 12.8 13.4 12.2 12.2 12.2
Maximum length (m) 7.6 6.8 8.8 7.9 11.0
Maximum power (MW) 28.0 25.4 28.0 22.0 28.0

)XOOHU7UD\ORU 6$*$* PLOOV UDQJH LQ VL]H IURP  WR  P  WR  IHHW LQ GLDPHWHU
DQG KDYH SRZHUV DV KLJK DV  0:   KS 7KH\ IHDWXUH K\GURVWDWLF OXEULFDWLRQ
DQG FDQ DFFRPPRGDWH HLWKHU D JHDUHG RU JHDUOHVV GULYH DUUDQJHPHQW DUH W\SLFDOO\ JUDWH
GLVFKDUJH DQG XVH PHWDO OLQHUV

ROD MILLS
5RG PLOOV DUH VLPLODU WR EDOO PLOOV H[FHSW WKDW WKH JULQGLQJ PHGLD DUH IUHHPRYLQJ VWHHO
URGV WKDW H[WHQG WKH OHQJWK RI WKH PLOO &RQVHTXHQWO\ PLOO OHQJWK VKRXOG EH DW OHDVW
 SHU FHQW PRUH WKDQ PLOO GLDPHWHU WR DYRLG GDQJHURXV URG WDQJOLQJ 5RG PLOOV DUH
RSHUDWHG DW D VORZHU VSHHG WKDQ EDOO PLOOV DQG WKH SURGXFW FRQWDLQV IHZHU ÀQHV 7KH IHHG
LV W\SLFDOO\ ² PP DQG UHGXFWLRQ WR ² PP LV FRPPRQ IRU RSHQFLUFXLW PLOOV
5RGV DQG EDOOV LQ SODFH ZHLJK  DQG  NJP UHVSHFWLYHO\ PHDQLQJ WKDW WKHUH
DUH OHVV YRLGV LQ WKH URWDWLQJ PDVV RI URGV WKDQ LQ WKH URWDWLQJ PDVV RI EDOOV 5RG PLOOV
DUH PRUH VHQVLWLYH WKDQ EDOO PLOOV WR H[FHVVLYH KDUG DQG FRDUVH IHHG SDUWLFOHV DQG WKHLU
WHQGHQF\ WR UHMHFW WKHVH IURP WKH URWDWLQJ PDVV OLPLWV WKH IHHG UDWH 3URYLGHG WKHUH LV
QRW H[FHVVLYH FRDUVH PDWHULDO LQ WKH IHHG URG PLOOV FUXVK WKH FRDUVHU VL]HV SUHIHUHQWLDOO\
DQG SURGXFH FRQVLGHUDEO\ OHVV ÀQHV WKDQ EDOO PLOOV IRU WKH VDPH QRPLQDO JULQGLQJ GXW\
7DEOH  VKRZV WKH HOLPLQDWLRQ RI FRDUVHU SDUWLFOHV E\ WKH URG PLOO UHODWLYH WR WKH EDOO
PLOO 5RG PLOOV FDQ EH XVHG ZHW RU GU\ DV WUXQQLRQ GLVFKDUJH RU SHULSKHUDO FHQWUH RU
HQG GLVFKDUJH PLOOV 0LOO VL]HV UDUHO\ H[FHHG  P LQ GLDPHWHU ð  P ORQJ EHFDXVH
 PP GLDPHWHU URGV DUH WKH ODUJHVW VL]H XVXDOO\ DGGHG WR WKH PLOO DQG  P LV WKHLU
PD[LPXP OHQJWK IRU VWUXFWXUDO UHDVRQV 7\SLFDOO\ URGV DUH DYDLODEOH LQ GLDPHWHUV XS WR
 PP EXW VRPH  PP GLDPHWHU URGV DUH DYDLODEOH
5RG PLOOV ZHUH XVHG IRU PDQ\ \HDUV DV WHUWLDU\ FUXVKHUV EXW DUH QRW QRZ XVHG LQ ODUJH
FDSDFLW\ SODQWV WKDW LV PRUH WKDQ  WK 6HFRQGDU\ DQG WHUWLDU\ FUXVKHUV LQ VXFK
SODQWV DUH EHLQJ UHSODFHG E\ 6$* PLOOV WR VLPSOLI\ WKH FLUFXLWV DQG RQO\ VPDOO ÁRZV RI
FRDUVH SDUWLFOHV LQ WKH 6$* PLOO GLVFKDUJH UHTXLUH IXUWKHU FUXVKLQJ KRZHYHU URG PLOOV
DUH VWLOO XVHG LQ VPDOOHU SODQWV XS WR  WK 7KH\ KDYH EHHQ JLYHQ D QHZ OHDVH RI OLIH
E\ WKH XVH RI FRDO DQG SHWUROHXP FRNH VOXUULHV IRU LQWHJUDWHG JDVLÀFDWLRQ FRPELQHG F\FOH
,*&& HOHFWULF SRZHU JHQHUDWLRQ DQG FRJHQHUDWLRQ IDFLOLWLHV %HFDXVH WKH RSHQFLUFXLW
URG PLOO SURGXFW FRQWDLQV IHZHU ÀQHV WKDQ D EDOO PLOO SURGXFW LW PDNHV D VOXUU\ WKDW KDV
D KLJK SHUFHQWDJH RI VROLGV DQG ORZ YLVFRVLW\ ZKLFK FDQ EH SXPSHG GLUHFWO\ LQWR WKH

Comminution Handbook | Spectrum Series 21 71

 CHAPTER 5 t TUMBLING MILLS

TABLE 5.9
Comparison of rod and ball mill feed and product sizings (WI – Work Index).
Rod mill (2.7 m × 4.1 m, 110 t/h WI-14.6) Ball mill (3.2 m × 4.3 m, 155 t/h WI-9.5)
Size (mm) Feed % passing Product % passing Feed % passing Product % passing
19 100 100 100 100
9.5 64 100 86.1 100
4.75 39.2 100 68.5 99.8
2.35 24.1 100 54.2 99.1
1.18 14.4 98.8 40.7 96
0.60 7.7 52.3 27.7 86.3
0.30 3.2 38.6 16.4 68.3
0.15 1.8 29.9 9.3 49.7
0.075 1.3 23.8 5.8 36.5

JDVLÀFDWLRQ UHDFWRU 5RG PLOOV DUH SUHIHUUHG IRU EULWWOH PLQHUDOV VXFK DV VFKHHOLWH DQG
JUDSKLWH WR PLQLPLVH RYHUJULQGLQJ DQG PD[LPLVH ÁRWDWLRQ UHFRYHULHV

ROTARY BREAKERS FOR COAL

5RWDU\ EUHDNHUV DUH XVHG WR UHGXFH WKH VL]H RI UXQRIPLQH FRDO DQG VHSDUDWH WKH FRDO
IURP WKH KDUGHU URFNV 7KH PDFKLQH ZDV GHYHORSHG DQG SDWHQWHG E\ +H]HNLDK %UDGIRUG
RI 5HDGLQJ 3HQQV\OYDQLD DQG WKH ÀUVW WZR %UDGIRUG EUHDNHUV ZHUH EXLOW LQ  7KH
SDWHQWV KDYH ORQJ VLQFH H[SLUHG DQG WKH PDFKLQH LV QRZ EXLOW E\ VHYHUDO PDQXIDFWXUHUV
$ %UDGIRUG URWDU\ EUHDNHU EXLOW E\ 0F/DQDKDQ &RUSRUDWLRQ LV VKRZQ LQ )LJXUH 
%UDGIRUG URWDU\ EUHDNHUV DFKLHYH UHGXFWLRQ E\ UHSHDWHGO\ UDLVLQJ WKH IHHG PDWHULDO DQG
GURSSLQJ LW DJDLQVW VWURQJ SHUIRUDWHG VFUHHQ SODWHV DURXQG WKH LQWHULRU $GMXVWDEOH OLIWHU
VKHOYHV UDLVH WKH IHHG PDWHULDO DQG FRQWURO WKH UDWH RI PDWHULDO PRYHPHQW 7KLV OLIWLQJ DQG
GURSSLQJ DFWLRQ HIIHFWLYHO\ FUXVKHV VRIW WR PHGLXPVWUHQJWK PDWHULDO ZKLFK WKHQ SDVVHV

FIG 5.9 – Rotary breaker 3.6 m diameter × 8.2 m long prior to installation of
dust cover guard (image courtesy McLanahan Corporation).

72 Comminution Handbook | Spectrum Series 21

 CHAPTER 5 t TUMBLING MILLS

WKURXJK WKH VFUHHQ RSHQLQJV WR D FROOHFWLRQ KRSSHU EHORZ +DUG URFN DQG XQFUXVKDEOH
PDWHULDOV DUH GLVFKDUJHG RXW WKH HQG RI WKH F\OLQGHU ZLWK WKH DLG RI D GLVFKDUJH SORZ
7KH EURNHQ FRDO LV W\SLFDOO\ VFUHHQHG DW  PP IROORZLQJ WKH URWDU\ EUHDNHU WKH FRDUVH
FRDO LV SXPSHG WR GHQVH PHGLXP VHSDUDWLRQ F\FORQHV DQG WKH ÀQH FRDO LV FOHDQHG E\
ÁRWDWLRQ 7DEOH  SUHVHQWV GHWDLOV RI WKH FDSDFLW\ RI %UDGIRUG EUHDNHUV

TABLE 5.10
Capacities of Bradford breakers (McLanahan Corporation).
Breaker Aperture size (mm) and throughput (Mt/h)
Dia × L (m) 38 mm 50 mm 63 mm 76 mm 89 mm 102 mm 127 mm 152 mm 203 mm
2.74 × 5.49 303 409 512 529 544 560 622 699 934
3.35 × 6.40 439 593 744 766 788 811 901 1012 1353
3.66 × 8.23 676 913 1145 1180 1214 1249 1388 1560 2082

SCRUBBERS
7ZR URWDU\ VFUXEEHUV DUH VKRZQ LQ )LJXUH  6FUXEEHUV DUH KLJKFDSDFLW\ PDFKLQHV
WKDW DUH SULPDULO\ XVHG WR UHPRYH ZDWHUGLVSHUVLEOH FOD\V GHOHWHULRXV PDWHULDOV DQG
FRDWLQJV IURP D ZLGH YDULHW\ RI KDUG URFN DQG RUH 7KH\ DUH F\OLQGULFDO GUXPV ZLWK
LQWHUQDO OLIWHUV DQG DUH W\SLFDOO\ VXSSRUWHG E\ WUXQQLRQ UROOHUV VKHOO UROOHUV RU WUXQQLRQ
EHDULQJV DW HLWKHU HQG RU E\ D VHULHV RI SQHXPDWLF W\UHV DORQJ WKH VFUXEEHU OHQJWK 7KH
FRQWLQXRXV OLIWLQJ DQG GURSSLQJ DFWLRQ ZLWK ZDWHU KHOSV DEUDGH VFUXE DQG EUHDN GRZQ
GLVSHUVLEOH RUH FRPSRQHQWV 6FUXEEHUV FRQVLVW RI D VROLG VKHOO VHFWLRQ DQG FDQ EH IROORZHG
E\ DQ RSWLRQDO WURPPHO VFUHHQ VHFWLRQ ,Q WKH VROLG VKHOO VHFWLRQ WKH HQHUJ\ LV LPSDUWHG
WR WKH IHHG LQ ZKDW FDQ EH FRQVLGHUHG D ZDVKLQJ DQG DEUDGLQJ DFWLRQ 'LVFKDUJH GDP
ULQJV FDQ EH FKDQJHG WR LQFUHDVH RU GHFUHDVH F\OLQGHU UHWHQWLRQ WLPH DQG WKH DPRXQW RI
HQHUJ\ LPSDUWHG 7KH WURPPHO VFUHHQ LV VXLWHG WR JHQHUDWLQJ D GHZDWHUHG FRDUVH SURGXFW
RU D SDUWLDOO\ GHZDWHUHG ÀQHU SURGXFW ZKLFK FDQ EH WUDQVIHUUHG WR D YLEUDWLQJ VFUHHQ
IRU ÀQDO VL]LQJ 7KH WURPPHO VFUHHQ VHFWLRQ LV DQ LQWHJUDO SDUW RI WKH VFUXEEHU EHFDXVH

FIG 5.10 – Rotary scrubbers showing a scrubbing section and an optional

trommel screen (image courtesy McLanahan Corporation).

Comminution Handbook | Spectrum Series 21 73

 CHAPTER 5 t TUMBLING MILLS

LWV VFUHHQLQJ DQG GHZDWHULQJ FDSDELOLWLHV DUH SURYLGHG WKURXJK LWV FRQWLQXRXV URWDWLRQ
ZLWK WKH VFUXEEHU 7KH WURPPHO VFUHHQ FDQ EH HDVLO\ EDFNZDVKHG DW WKH WRS RI LWV WUDYHO
6FUXEEHUV W\SLFDOO\ RSHUDWH DW ORZ VSHHGV DQG UHO\ RQ WKH RUH FRQWDLQLQJ HQRXJK ODUJH
PHGLD WR SURYLGH WKH QHFHVVDU\ ZDVKLQJ DQG HQHUJ\ LQSXW $ ODFN RI PHGLD FDQ UHVXOW
LQ FOD\ EDOO IRUPDWLRQ ZKLFK FDQ FDXVH SUREOHPV GRZQVWUHDP LQ WKH IRUP RI ORVVHV RU
FRQWDPLQDWLRQ ,I WKHUH LV LQVXIÀFLHQW PHGLD SUHVHQW LQ WKH IHHG D VPDOO TXDQWLW\ RI VWHHO
EDOOV PD\ EH DGGHG SURYLGHG WKDW WKH VFUXEEHU VKHOO LV VXIÀFLHQWO\ VWURQJ $OWHUQDWLYHO\
EDUUHQKDUG ODUJH URFN PD\ EH DGGHG LI DYDLODEOH WR SURYLGH VFUXEELQJ PHGLD 6FUXEEHUV
UDUHO\ KDYH SRZHU LQSXW OHYHOV DERYH DSSUR[LPDWHO\  0: $W WKLV VL]H D VFUXEEHU FDQ
EHJLQ WR ORRN OLNH D EDOO PLOO ZLWK D SLQLRQ GULYH DQG VKHOOPRXQWHG EHDULQJV

GRINDING BALLS
7KH JULQGLQJ EDOOV XVHG LQ YDULRXV W\SHV RI FRPPLQXWLRQ PDFKLQHV KDYH D ZLGH UDQJH
RI VL]HV DQG FRPSRVLWLRQV FDOFXODWHG E\ WKH HTXLSPHQW YHQGRU $FURVV DOO WKH PDFKLQHV
WKH EDOOV· EUHDNDJH PHFKDQLVPV DUH LPSDFW DEUDVLRQ DQG DWWULWLRQ 7KH ZHLJKW RI D EDOO
FKDUJH LV LPSRUWDQW EXW LW LV DOVR QHFHVVDU\ WR KDYH WKH FRUUHFW VL]H GLVWULEXWLRQ RI EDOOV LI
WKH PRVW HIÀFLHQW XVH RI JULQGLQJ HQHUJ\ LV WR EH DFKLHYHG ,Q WKLV FRQWH[W LW LV QRWHG WKDW
 W RI VWHHO EDOOV FRQWDLQV § EDOOV RI  PP GLDPHWHU §  EDOOV RI  PP GLDPHWHU
RU § PLOOLRQ EDOOV RI  PP GLDPHWHU *ULQGLQJ FDXVHV WKH EDOOV WR ZHDU DQG QHZ EDOOV
PXVW EH DGGHG UHJXODUO\ WR PDLQWDLQ WKH FKDUJH OHYHO DQG SRZHU GUDZ LQ D WXPEOLQJ
PLOO 7KH SUHIHUHQFH LV IRU WKH WRS VL]H RI EDOOV LQ D EDOO FKDUJH WR EH ODUJH HQRXJK WR
EUHDN WKH WRS VL]H RI SDUWLFOHV LQ WKH IHHG H[FHSW LQ UHJDUG WR 6$* PLOOLQJ %DOOV LQ WKH
FKDUJH ZLOO DVVXPH D VL]H GLVWULEXWLRQ RYHU D SHULRG RI WLPH DV WKH\ ZHDU DQG UHDFK D
VL]H ZKHUH WKH\ DUH HLWKHU EURNHQ GRZQ TXLFNO\ RU GLVFKDUJHG IURP WKH PLOO DV D ¶VFDW·
:KLOH D FRDUVH IHHG FDQ EH UHGXFHG WR D ÀQH SURGXFW LQ D VLQJOH PLOOFODVVLÀHU FLUFXLW
WKH DOWHUQDWLYH SURFHGXUH RI XVLQJ WZR PLOOFODVVLÀHU FLUFXLWV LQ D VHULHV ZLWK ODUJH EDOOV
DGGHG WR WKH ÀUVW PLOO DQG VPDOO EDOOV WR WKH VHFRQG PLOO LV XVXDOO\ PRUH HIÀFLHQW DQG
SUHIHUDEOH
)UHVK LURQ SDUWLFOHV FRPLQJ IURP WKH DEUDVLRQ RI EDOOV DQG PLOO OLQHUV FDQ DIIHFW
GRZQVWUHDP SURFHVVHV VXFK DV ÁRWDWLRQ E\ SURGXFLQJ D UHGXFLQJ VOXUU\ KRZHYHU
SURJUHVV KDV EHHQ PDGH LQ GHYHORSLQJ PDWHULDOV WR PLQLPLVH ZHDU IRU H[DPSOH
• %DOOV PDGH IRU XVH LQ FHPHQW PLOOV KDYH FKURPLXP FRQWHQWV IURP WHQ WR  SHU FHQW
DQG GLDPHWHUV IURP  WR  PP 7KH ODUJHU EDOOV DUH DGGHG WR WKH ÀUVW FRPSDUWPHQW
• )RU ZHW JULQGLQJ PLOOV WKH UDQJH RI FKURPLXP FRQWHQWV LV WHQ WR  SHU FHQW %DOOV
ZLWK WKH KLJKHU OLPLW DUH XVHG ZKHQ LW LV QHFHVVDU\ WR PLQLPLVH WKH FRQWDPLQDWLRQ RI
ÁRWDWLRQ SXOSV ZLWK LURQ FRPSRXQGV
• $ ORZ FKURPLXP SURGXFW  SHU FHQW &U LV DYDLODEOH LI UHTXLUHG
• .HUDPD[® DUH FHUDPLF EHDGV ² PP LQ GLDPHWHU DQG DUH XVHG LQ VWLUUHG PLOOV WR
SURGXFH YHU\ ÀQH SDUWLFOHV VXFK DV WKRVH UHTXLUHG IRU FDOFLXP FDUERQDWH LQ SDSHU
PDNLQJ DQG IRU WKH SKDUPDFHXWLFDO FRVPHWLF DQG PLQHUDO LQGXVWULHV
• +LJKFDUERQ FKURPLXPDOOR\HG VWHHO URGV DUH DYDLODEOH ZLWK GLDPHWHUV IURP
² PP DQG OHQJWKV IURP ² P

1. The examples given here are products of Magotteaux, a company located in Belgium with global affiliations.

74 Comminution Handbook | Spectrum Series 21

 CHAPTER 5 t TUMBLING MILLS

BUILDING LARGE TUMBLING MILLS

/DUJH WXPEOLQJ PLOOV DUH LPPHQVH PDFKLQHV WKDW UHTXLUH ODUJH PRWRUV DQG WKHUH DUH
IHZ ORFDWLRQV LQ WKH ZRUOG ZKHUH WKH\ FDQ EH EXLOW 7KH FRQVWUXFWLRQ RI ODUJH PLOOV LV DQ
LPSRUWDQW SDUW RI JULQGLQJ WHFKQRORJ\ DQG LV GHVFULEHG EULHÁ\ KHUH WR JLYH DQ LGHD RI
ZKDW LV LQYROYHG 'HVLJQ WHFKQRORJ\ IRU JULQGLQJ PLOOV LV VLPSOH DQG KDV EHHQ DYDLODEOH
IRU PRUH WKDQ D FHQWXU\ ZLWK PDQ\ FRPSDQLHV PDNLQJ PLOOV GXULQJ WKH WK FHQWXU\
,Q WKH HDUO\ V WKHUH ZDV FRQVROLGDWLRQ RI WKH FRPSDQLHV PDNLQJ JULQGLQJ PLOOV DQG
IRU WKH ODVW  \HDUV WKH PDMRU URWDWLQJ HOHPHQW SDUWV IRU ODUJH PLOO VXSSOLHUV LQFOXGLQJ
)/6PLGWK 0HWVR 2XWRWHF DQG 7K\VVHQ.UXSS KDYH EHHQ EXLOW DW &,7,& &HQVD 6SDLQ
&,7,& +HDY\ ,QGXVWULHV &KLQD DQG RWKHU IDFWRULHV &,7,& &HQVD ZDV HVWDEOLVKHG LQ 
LQ QRUWKZHVW 6SDLQ LQ WKH LQGXVWULDO DUHD RI 9LJR DQG LV QRZ SDUW RI &,7,& +HDY\
,QGXVWULHV
$ JULQGLQJ PLOO LV D GUXP FRQWDLQLQJ JULQGLQJ PHGLD WKDW URWDWHV RQ DQWLIULFWLRQ RU
RLO ÀOP EHDULQJV )LQLWH HOHPHQW DQDO\VLV KDV EHHQ GHYHORSHG RYHU  \HDUV DQG WKH
VWUXFWXUDO GHVLJQ FDQ EH FRQÀUPHG XVLQJ WKLV WHFKQLTXH :KLOH IRU PLOO GHVLJQ WKHUH LV QR
OLPLW WR WKH VL]H RI HTXLSPHQW LQ SUDFWLFH WKH OLPLW GHSHQGV RQ WKH DELOLW\ WR PDQXIDFWXUH
WKH YDULRXV FRPSRQHQWV WKDW PDNH XS D PLOO JUHDWHU WKDQ  P LQ GLDPHWHU 7KH FULWLFDO
FRPSRQHQWV DUH WKH PLOO KHDGV DQG WUXQQLRQV PLOO VKHOO ULQJ JHDU DQG SLQLRQ PDLQ
EHDULQJV DQG GULYH PRWRUV (DFK ODUJH JULQGLQJ PLOO LV PDGH WR RUGHU DQG VHFWLRQHG VR
WKDW DOO FRPSRQHQWV FDQ EH WUDQVSRUWHG DQG DVVHPEOHG DW WKH VLWH
7KH IDFLOLWLHV DYDLODEOH DW &,7,& +HDY\ ,QGXVWULHV &KLQD WR EXLOG WKH ODUJH PLOOV LQFOXGH
• FDVWLQJV WKDW HQDEOH SRXUV RI XS WR  W LQ D VLQJOH SLHFH
• DQ   W IRUJLQJ XQLW WKDW LV FDSDEOH RI PDNLQJ  W VLQJOHSLHFH IRUJLQJV
• WZR  P GLDPHWHU FRPSXWHU QXPHULFDO FRQWURO JHDUFXWWLQJ DQG JHDUÀQLVKLQJ
IDFLOLWLHV WKDW DOORZ WKH PDQXIDFWXUH RI JHDUV EHORZ  P LQ GLDPHWHU
• WHFKQLTXHV WR GHVLJQ DQG PDQXIDFWXUH PLOOV XS WR  P ZLWK  0: JHDUOHVV GULYHV
DQG  P ZLWK  ð  0: JHDUHG GULYHV

Components of a mill

Mill heads
0LOO KHDGV DUH PDGH LQ VHJPHQWV HJ  ð ƒ DQG WKHQ EROWHG 7KH PLOO KHDGV WUXQQLRQV
DQG JLUWK JHDUV FDQ EH PDGH IURP HLWKHU VWHHO RU VSKHURLGDO JUDSKLWH LURQ ZKLFK LV D
GXFWLOH LURQ ZLWK LPSDFW DQG IDWLJXH UHVLVWDQFH 0RVW RI WKH ODUJH VWHHO FDVWLQJ VKRSV
DURXQG WKH ZRUOG KDYH FORVHG GRZQ EHFDXVH WKH SULFHV RI VWHHO FDVWLQJV DUH PXFK KLJKHU
WKDQ VSKHURLGDO JUDSKLWH LURQ HYHQ WKRXJK WKH SURSHUWLHV RI VWHHO FDVWLQJV DUH EHWWHU
&,7,& FDVWLQJV IRU 6$*$* PLOOV XS WR  P  IHHW DUH VWHHO ZLWK FDVWLQJ DQG
FRROLQJ LQ VL[ SDUWV )RU  P  IHHW EDOO PLOO KHDGV WKH\ DUH SRXUHG LQ D VLQJOH ƒ
FDVWLQJ DQG WKHQ PDFKLQHG
7KH VKHOO LV D KROORZ F\OLQGHU WKDW QHHGV WR EH PDGH LQ SLHFHV DQG EROWHG DW WKH VLWH 7KH
GLPHQVLRQV OHQJWK ð ZLGWK ð KHLJKW GHWHUPLQH ZKDW FDQ EH WUDQVSRUWHG WR VLWH WR DYRLG
EULGJHV WXQQHOV HOHFWULF RYHUKHDG FDEOHV DQG WKH OLNH $Q\WKLQJ ELJJHU WKDQ  P KDV WR
EH VHFWLRQDOLVHG IRU WUDQVSRUW 7UDQVSRUW UHVWULFWLRQV DQG WKH FDVWLQJ SRXULQJ FDSDFLW\ RI
WKH IRXQGU\ GHFLGH WKH QXPEHU RI VHFWLRQV LQ ZKLFK WKH PLOO KHDGV JHDU DQG VKHOO ZLOO
EH FDVW RU IDEULFDWHG (DFK RI WKH ODUJH JULQGLQJ PLOOV LV PDGH DQG GHVLJQHG HVSHFLDOO\
IRU HDFK ORFDWLRQ

Comminution Handbook | Spectrum Series 21 75

 CHAPTER 5 t TUMBLING MILLS

Shells
7KH VKHOO ÁDQJHV DUH PDGH ZLWK FRQWRXUHG FDVWLQJ DQG WKHQ PDFKLQHG WR WKH FRUUHFW
SURÀOH 0LOO VKHOO SODWHV IRU $* PLOOV DUH UROOHG LQKRXVH DQG WKHQ ZHOGHG LQ DXWRPDWLF
*HUPDQ VXEDUF ULJV RQ IRXU  P GLDPHWHU UROOHUV )RU EDOO PLOO VKHOO SODWHV WKLV LV GRQH
LQ RQH RI WHQ UROOHUV RI  P GLDPHWHU

Trunnions
$V HYHU\ PLOO LV GLIIHUHQW WKH GHVLJQ RI HDFK FRPSRQHQW ZLOO GLIIHU $ VORW RU JURRYH LQ
WKH WUXQQLRQ HQVXUHV WKDW WKH WKUXVW EHDULQJV ÀW QHDWO\ RQWR WKH VLGHV JXDUDQWHHLQJ QR
ODWHUDO PRYHPHQW DV D JLUWK JHDU RU JHDUOHVV GULYH PXVW RSHUDWH RQ D À[HG FHQWUH

Gear and pinion

&DVW VWHHO PDWHULDOV RI FRQVWUXFWLRQ PDGH RI FKURPH QLFNHO DQG PRO\EGHQXP VDWLVI\
WKH VWUHQJWK UHTXLUHPHQWV 2SWLPLVLQJ WKH GHVLJQ RI WKH JLUWK JHDU JHRPHWU\ LQFOXGHV
TXHVWLRQV DERXW PDWHULDO SURSHUWLHV PRGXOH SUHVVXUH DQG KHOL[ DQJOH WRRWK ZLGWK
ÁDQN FRQWDFW SUHVVXUH SLWFK FLUFOH RI SLQLRQ DQG D[LDO IRUFHV )RU PLOO GULYHV D ƒ
SUHVVXUH DQJOH LV SUHIHUUHG DV LW LV DSSUR[LPDWHO\ WHQ SHU FHQW VWURQJHU WKDQ D ƒ SUHVVXUH
DQJOH 7KH ODUJHVW GULYH QRZ RSHUDWLQJ RQ DQ 6$* PLOO LV  ð  0: DQG WKLV ZLOO ULVH WR
 ð  0: ZLWK SUHVHQW GHVLJQV

Gearless drive
$W SUHVHQW WKH JHDUOHVV GULYH LV PRUH FRVWO\ WKDQ WKH FRPSOHWH PHFKDQLFDO SRUWLRQ
PLOO URWDWLQJ HOHPHQW DQG DFFHVVRULHV VXSSOLHG E\ WKH PLOO PDQXIDFWXUHUV *HDUOHVV
GULYH URWDWLQJ HOHPHQWV DUH KHDYLHU WKDQ DQ HTXLYDOHQW JHDU GULYH JULQGLQJ PLOO WKH
FDVW KHDGV DUH KHDYLHU ZLWK D ORQJHU ÁDQJH WR KROG WKH VHJPHQWHG SROHV DQG WKH SROHV
DUH DSSUR[LPDWHO\  SHU FHQW KHDYLHU DQG QHHG WR EH LQGLYLGXDOO\ LQVWDOOHG DIWHU WKH
URWDWLQJ HOHPHQW LV SXW LQ SODFH RQ WKH IRXQGDWLRQ 7KH IRXQGDWLRQ DQG ORDG GDWD RQ
JHDUOHVV GULYHV DUH KLJKHU $LU JDS YDULDWLRQ EHWZHHQ WKH VKHOOURWRU DQG VWDWRU LV D
VRXUFH RI PDJQHWLF SXOO LPEDODQFH DQG WKH H[DFW UXQRXW LV QHHGHG WR UHGXFH WKH ULVN
RI XQEDODQFHG PDJQHWLF SXOOV DURXQG WKH FLUFXPIHUHQFH $IWHU PRXQWLQJ WKH SROHV WKH
WROHUDQFH UHTXLUHG WR PDLQWDLQ FRQFHQWULFLW\ LV OHVV WKDQ  PP 7KH EROWV IRU PRXQWLQJ
WKH SROHV DUH VSHFLDOO\ PDGH VR WKDW WKH SROHV· FRQFHQWULFLW\ UXQRXW FDQ EH DGMXVWHG
ZLWKLQ  PP $ JHDUOHVV GULYH KDV PXOWLSOH SROHV ZKLFK UHTXLUH DFFXUDWH DGMXVWPHQW
RI HDFK SROH WR PDLQWDLQ FRQFHQWULFLW\ ZLWKLQ WKH PRWRU VWDWRU 7KH FRPSOHWH URWDWLQJ
HOHPHQW EHFRPHV WKH URWRU RI WKH ODUJH PRWRU DQG WKH VWDWRU DQG URWRU DUH FDOOHG D
JHDUOHVV GULYH
'XULQJ DFWXDO RSHUDWLRQ WKH DLU JDS DW WKH WRS LQFUHDVHV DQG UHGXFHV DW WKH ERWWRP
WKXV FDXVLQJ DQ XQEDODQFHG PDJQHWLF SXOO )RU PLOOV DERYH  P LQ GLDPHWHU RU RYHU
 0: RI SRZHU WKHUH LV FXUUHQWO\ QR YLDEOH DOWHUQDWLYH WR D JHDUOHVV GULYH )RXU SLQLRQ
GULYHV DUH EHLQJ ORRNHG DW DQG DUH FXUUHQWO\ LQ WKH UHVHDUFK DQG GHYHORSPHQW VWDJH
)RU PLOOV  P RU OHVV LQ GLDPHWHU DQG  0: RU XQGHU LQ SRZHU JHDUGULYHQ V\VWHPV
RIIHU D EHWWHU DOWHUQDWLYH WR JHDUOHVV GULYHV GHOLYHULQJ QHDUO\ HTXLYDOHQW HIÀFLHQF\ PXFK
ORZHU FDSLWDO FRVW ² SHU FHQW FKHDSHU HDVLHU LQVWDOODWLRQ VLPSOHU FRROLQJ VKRUWHU
LQVWDOODWLRQ WLPH DQG D TXLFNHU WZR WR WKUHH PRQWKV RYHUDOO SURMHFW VWDUWXS VFKHGXOH

76 Comminution Handbook | Spectrum Series 21

 CHAPTER 5 t TUMBLING MILLS

BIBLIOGRAPHY
Allen, 7  Particle Size Measurement &KDSPDQ DQG +DOO /RQGRQ
Austin, / * /XFNLH 3 7 DQG .OLPSHO 5 5  The Process Engineering of Size Reduction: Ball
Milling  S 6RFLHW\ IRU 0LQLQJ 0HWDOOXUJ\ DQG ([SORUDWLRQ DQG WKH $PHULFDQ ,QVWLWXWH RI
0LQLQJ 0HWDOOXUJLFDO DQG 3HWUROHXP (QJLQHHUV 1HZ <RUN
Beke, %  The Process of Fine Grinding  S $NDGHPLDL .LDGR %XGDSHVW
Bickle, : + HG  Crushing and Grinding – A Bibliography  S 'HSDUWPHQW IRU 6FLHQWLÀF
DQG ,QGXVWULDO 5HVHDUFK /RQGRQ
Bond, ) &  7KH WKLUG WKHRU\ RI FRPPLQXWLRQ Trans SME/AIME ²
Digre, 0 HG  Proceedings Autogenous Grinding Seminar, YROXPHV  DQG  7URQGKHLP 0D\
 $VVRFLDWLRQ RI 1RUZHJLDQ 0LQHV 7URQGKHLP
Gross, -  &UXVKLQJ DQG JULQGLQJ US Bureau of Mines Bulletin 
Gy, 3 0  Sampling of Particulate Materials: Theory and Practice  S (OVHYLHU $PVWHUGDP
Lowrison, * &  Crushing and Grinding: The Size Reduction of Solid Materials  S %XWWHUZRUWKV
/RQGRQ
Lynch, $ -  Mineral Crushing and Grinding Circuits: Their Simulation, Optimisation, Design and
Control  S (OVHYLHU $PVWHUGDP
Wills, % $ DQG 1DSLHU0XQQ 7 -  Mineral Processing Technology VHYHQWK HGLWLRQ %XWWHUZRUWK
+HLQHPDQQ 2[IRUG

&DWDORJXHV IRU WXPEOLQJ PLOOV DUH DYDLODEOH RQ WKH LQWHUQHW IURP VHYHUDO FRPSDQLHV 7KRVH UHIHUUHG
WR LQ WKLV FKDSWHU LQFOXGH
• )/6PLGWK 'HQPDUN 86$
• 7K\VVHQ.UXSS *HUPDQ\
• 7HQRYD %DWHPDQ 6RXWK $IULFD
• 0HWVR )LQODQG
• &,7,& +HDY\ ,QGXVWULHV &KLQD
• 0F/DQDKDQ &RUSRUDWLRQ 86$

Comminution Handbook | Spectrum Series 21 77

 HOME

Chapter 6

COMPRESSION MACHINES
Alban Lynch

INTRODUCTION
Compression machines are used to break a wide range of different-sized solids, from
run-of-mine ore to  micron particles 7he machines used as the feed becomes Àner
are – in order of particle size – the gyratory crusher, jaw crusher, cone crusher, high-
pressure grinding rolls (HPGRs) and vertical roller mills. The exact sequence of machines
is dependent on the task required, the material types and the necessary throughput.

PRIMARY CRUSHERS

Gyratory crusher
The gyratory crusher was patented in 1881 after years of failure with experimental
machines, and since that time there have been few fundamental changes to the original,
successful design. The crusher consists of a vertical cone suspended at the top and held
by a gyrating eccentric sleeve at the bottom. The eccentric assembly that causes the
gyration gives the crusher an open and closed side setting (OSS and CSS respectively),
and it is this movement that causes rocks to be trapped between the movable surface
(mantle) and the Àxed surface (concave or bowl), resulting in them being nipped and
broken. The mantle then recedes from the concave and the broken particles move down
the chamber. After more crushing cycles, the product is discharged.
A sketch showing how the crusher works is given in Figure 6.1a and an FLSmidth gyratory
crusher is shown in Figure 6.1b. The capacity of the crusher depends on the physical size of
the crusher, the discharge opening (OSS) and the Áowability of the feed. ,n large gyratory
crushers, the crushing chamber is very steep and therefore the Ànes content of the feed
will have a signiÀcant impact on the throughput of a machine. The relationship between
capacity and discharge opening (OSS) is shown in Table 6.1. ,t is important to note that in
Table 6.1, the discharge opening quoted is the OSS. ,n cone crushers, throughput is always
quoted in terms of CSS. Gyratory crushers do not require a feed system as other crushers
do, and trucks can directly discharge their load into the gyratory crusher.

Comminution Handbook | Spectrum Series 21 79

 CHAPTER 6 t COMPRESSION MACHINES

A B

FIG 6.1 – Gyratory crusher: (A) how a gyratory crusher works,

and (B) top-service gyratory crusher (images courtesy FLSmidth).

TABLE 6.1
Capacities (Mt/h) of gyratory crushers with different discharge settings (source: Metso).
Feed opening Discharge opening (mm) kW Crusher weight
(mm) 150 175 200 230 250 (t)
1065 2335 2820 375 119
1370 2885 3145 3485 450 242
1525 4540 5295 5805 600 398
1525 5525 7335 8280 8890 1200 553

Jaw crusher
The jaw crusher was patented in 188 after the Àrst crusher had been built in 18 to
prepare rocks for road building. Essentially, the design has changed little since that time.
A jaw crusher consists of converging movable and Àxed jaws as shown in Figure 6..
The moving jaw is driven through a toggle joint, which exerts immense pressure
on the jaw. Figure 6. shows a double toggle jaw crusher conÀguration, which was
traditionally used for crushing harder feed types as the use of the double toggle
arrangement provided a better crushing action. The need for a double toggle jaw
crusher is indicated by unconÀned compressive strength (8CS) measurements a 8CS
measurement in excess of  0Pa requires a double toggle. Single toggle machines
as shown in Figure 6., were commonly reserved for lighter crushing duties. Over
the last  years, advances in design and fabrication have allowed single toggle jaw
crushers to become the machine of choice, even for harder applications. They offer
higher throughput and reduced mass compared to the double toggle machines. The
capacity of any size of machine depends on its CSS, which can be adjusted. Data
showing the relationship between capacity and CSS for a hard, dry ore are given in
Table 6.. Capacities are dependent on the ore and do not apply to difÀcult ores such
as those containing sticky clays.

80 Comminution Handbook | Spectrum Series 21

 CHAPTER 6 t COMPRESSION MACHINES

FIG 6.2 – Double toggle jaw crusher (image courtesy Pennsylvania Crushing).

FIG 6.3 – Single toggle jaw crusher (image courtesy ThyssenKrupp).

TABLE 6.2
Approximate throughputs of jaw crushers (Mt/h) (data from Terex Corporation).
Feed opening Closed side setting (mm) kW Crusher weight
(mm) 100 150 200 250 300 (t)
1220 × 915 190–210 340–370 485–535 150 37
1220 × 1065 370–410 535–590 660–730 185 45
1500 × 1275 465–515 625–690 775–860 930–1030 250 80

Comparison of jaw and gyratory crushers

Jaw crushers are usually preferred for lower-capacity concentrators and quarries
because they are smaller, less expensive and easier to maintain than gyratory crushers,
which are more expensive and more mechanically complex. Crushing by jaw crusher
is intermittent because of the reciprocating motion of the movable jaw, meaning that
productivity is relatively low compared with a gyratory crusher with the same power
that works continuously. While gyratory crushers have high maintenance costs, their
higher processing capacity means that they are usually selected as the primary crushers
in large-scale concentrators. They also have a relatively high reduction ratio of up to
ten. The costs of jaw crushers and gyratory crushers consist of capital and operating
costs. Generally, the costs of a jaw crusher installation are much lower than that of a

Comminution Handbook | Spectrum Series 21 81

 CHAPTER 6 t COMPRESSION MACHINES

gyratory crusher, but the long-term costs of a jaw crusher tend to be higher. If both jaw
crusher and gyratory crusher can meet the requirements of feed size, range of discharge
openings and processing capacity, it is the initial cost and ongoing operating cost that
are the most important elements in crusher selection. The main advantage of gyratory
crushers is the ability to accept feed directly from trucks; jaw crushers require a feed
system as well as a grizzly to remove oversize rocks.

CONVENTIONAL CONE CRUSHERS

&one crushers work on a similar principle to gyratory crushers but crush much Àner by
using smaller gaps, which requires a different design of the machine. Figure 6.4 shows
how the feed cone crusher Áows over the mantle cone, which rotates eccentrically
inside the concave and is crushed between them.
Cone crushers now in use are commonly derived from the Symons crusher, which was
patented about 1930. This crusher was built as two models: the Standard Symons for
secondary crushing and the Shorthead Symons for tertiary and quarternary crushing.
The reason for the Shorthead Symons was that crushing is more energy efÀcient than
grinding and Àner crushing meant that less energy was required to produce Àne particles
by grinding. With both standard and shorthead crushers available, it became possible to
use secondary and tertiary circuits with an intermediate storage bin, and these circuits
were used for many years. Numerous crusher manufacturers have continued to use this
naming convention, including Metso and FLSmidth. The example shown in Figure 6.5
is the Raptor cone crusher from FLSmidth, which includes both shorthead and standard
units. Table 6.3 summarises the relationship between capacity and CSS for cone crushers.
The Metso Nordberg series of cone crushers show a similar range of capacities for the
same CSS values.
The other main type of cone crusher is generally termed a ‘hydroset’ or bottom-
supported machine and is shown in Figure 6.6.

FIG 6.4 – Sketch of a cone crusher (image courtesy Hakan Dundar).

82 Comminution Handbook | Spectrum Series 21

 CHAPTER 6 t COMPRESSION MACHINES

FIG 6.5 – A7 Raptor XL1300 cone crusher (image courtesy FLSmidth).

TABLE 6.3
Relationship between feed rate and closed side setting for FLSmidth Raptor cone crushers.
Crusher Small Small Large Large
Closed side setting Feed range Closed side setting Feed range
(mm) (Mt/h) (mm) (Mt/h)
Shorthead fine 10 250–350
Shorthead medium 13 325–450 13 770–1200
Shorthead coarse 16 400–550 16 870–1250
Standard fine 25 525–770 25 1200–1900
Standard medium 38 650–1050 38 1700–2100
Standard course 45 780–1250 45 1900–2500

FIG 6.6 – Hydroset-style cone crusher (image courtesy FLSmidth).

Comminution Handbook | Spectrum Series 21 83

 CHAPTER 6 t COMPRESSION MACHINES

SuperÀcially, these crushers have a closer resemblance to a gyratory-style crusher,

but they are designed to fulÀl the same duty as the Symons and its derivatives. ,n the
hydroset-style machines, the discharge setting (CSS) is adjusted via a hydraulic cylinder
under the main shaft. This is different to the Symons-style machines, which are adjusted
via a thread arrangement that moves the concave relative to the crushing head. The
hydroset system provides automatic overload protection to permit the passage of tramp
iron or other uncrushable material. The system then automatically returns the mainshaft
smoothly to its original position (Sandvik).

HIGH CAPACITY CRUSHING CIRCUITS

Different machines are used in a series when a high reduction ratio is required, as shown
in Figure 6., which is the Áow sheet of a  0ta plant crushing a low-grade copper
sulÀde ore at %ougainville Copper Ltd. ,n this example, the machines included two
gyratory crushers, eight secondary cone crushers and 16 tertiary cone crushers. Each
gyratory circuit crushed  th, with the CSS being 1 mm. An average of 66 th
were broken in each secondary crusher operating with a CSS of 6 mm. 9alues for
the tertiary crushers were 41 th operating with a CSS of 6 mm. Typical particle size
distributions (PSDs) for the secondary and tertiary crusher feed and product are shown
in Figure 6.8.
Although %ougainville Copper Ltd only operated between 16 and 18, it was
the highest capacity concentrator of the time and the crushing plant is still the largest
installation of its type. Crushing was in two independent parallel lines of crushers and
screens. The plant operated for 96 per cent of the available time, and the runtime for both
the secondary and the tertiary crushers was 91 per cent, giving an overall utilisation of
8 per cent. The crushing plant achieved  0ta at a P80 of  mm. The %ougainville
crushing plant was an iconic installation. ,ts overall utilisation of 8 per cent, which
was achieved by targeting maximum crushing power in combination with best practice
maintenance, is still considered to be at the top end of the range. ,n most crushing plant
designs, it is now common to target 75 per cent overall utilisation.

FIG 6.7 – Crushing circuit at Bougainville Copper Ltd showing the roles of
compression crushers (image courtesy Bougainville Copper Ltd).

84 Comminution Handbook | Spectrum Series 21

 CHAPTER 6 t COMPRESSION MACHINES

ϭϬϬ
ϵϬ

ĐƵŵƵůĂƚŝǀĞƉĞƌĐĞŶƚƉĂƐƐŝŶŐй
ϴϬ
ϳϬ
ϲϬ
ϱϬ
ƐĞĐ͘ĐƌƵƐŚĞƌĨĞĞĚ
ϰϬ
ƐĞĐ͘ĐƌƵƐŚĞƌƉƌŽĚƵĐƚ
ϯϬ
ƚĞƌ͘ĐƌƵƐŚĞƌĨĞĞĚ
ϮϬ
ƚĞƌ͘ĐƌƵƐŚĞƌƉƌŽĚƵĐƚ
ϭϬ
DĨĞĞĚ
Ϭ
ϭ ϭϬ ϭϬϬ ϭϬϬϬ
ƉĂƌƚŝĐůĞ ƐŝnjĞ;ŵŵͿ

FIG 6.8 – Particle size distributions of flotation feed and secondary and tertiary cone crusher
feeds and products at Bougainville Copper Ltd (image courtesy Bougainville Copper Ltd).

BED BREAKAGE CRUSHERS

%ed breakage compression crushers break rocks by applying pressure to a bed rather
than a layer of particles, and this concept was the basis of a crusher built by engineers at
The ,nstitute of 0ineral Processing (0ekhanobr) in St Petersburg, Russia. Commercial
compression crushers include the Rhodax crusher made by Fives FC% in France and the
9ibrocone crusher made by Sandvik in Sweden.

Rhodax crusher
Fives FC%, which makes the Rhodax crusher, is a 00-year-old French company with
expertise in turnkey cement plants, cement and mineral grinding plants and processing
equipment for the cement and mineral grinding industries. The crusher is an inertial
grinder that works on the pressurised bed fragmentation principle. The fragmentation
force is caused by the rotation of unbalanced masses rather than by an eccentric mass
as is the case in a conventional crusher. This makes it possible to achieve very high
reduction ratios. %y controlling the speed of rotation of the unbalanced masses, the
grinding force can be modiÀed, thus changing the product size distribution.
The components of the Rhodax crusher are
x the bowl subassembly (bowl), which consists of a frame supporting the bowl liner
x the cone subassembly (cone), which is a structure supporting the vertical shaft and
the cone (head) protected by the mantle
x the cone, which is attached to the bowl by rods and ball joints
x the bowl, which is supported by elastic suspensions to minimise the transmission
of vibrations.
The driven part of the Rhodax is the bowl, while the cone subassembly is allowed to
deÁect as a result of the applied grinding force. The mantle (cone liner) is free to rotate
around the centre shaft. The variables that are controlled are the gap opening and the
rotation speed of the unbalanced masses. High-pressure Rhodax crushers have motors
from 0–900 kW and produce 600 th of Àne crushed rock from 150 mm feed. Figure 6.9
presents details of the Rhodax crusher.

Comminution Handbook | Spectrum Series 21 85

 CHAPTER 6 t COMPRESSION MACHINES

FIG 6.9 – Rhodax crusher: (A) components, (B) crushing mechanism (image courtesy Fives FCB).

Vibrocone crusher
The 9ibrocone crusher is similar to the Rhodax crusher and its method of operation is
discussed in 8S patent 795475 (%elotserkovsky, 010). The discharge gap and rotor
speed are controlled by internal sensors to produce the maximum feed rate. The main
shaft is supported by a spherical bearing. An unbalanced weight rotating around the
main shaft causes the mantle to swing without restraint, and the rock is crushed by
multiple compressions between the mantle and the concave (Figure 6.10). The crusher
is choke fed and all particles in the moving bed are compressed, resulting in breakage.
8ltrasonic sensors detect the position and angle of inclination of the mantle and the data
are used to control the crusher by varying the discharge gap and rotational speed of the
shaft. A comparison of the Rhodax and 9icrocone bed breakage compression crushers
is shown in Table 6.4.

TABLE 6.4
Comparison of bed breakage compression crushers.
Sandvik Vibrocone crusher Fives FCB Rhodax crusher
Rated power (kW) 400 Up to 900
Capacity (Mt/h) 200–300 Up to 600
Feed size (max) 150 220
Product size 80% passing 6–8 mm, 50% passing 2–4 mm 0–500 μm to 0–40 mm in closed circuit

86 Comminution Handbook | Spectrum Series 21

 CHAPTER 6 t COMPRESSION MACHINES

A B C
FIG 6.10 – Sketches of the Vibrocone crusher (US patent 7954735). (A) The crusher with its sensors
(R, D1) and controller (D4), (B) an example of the angle and deviation of the mantle moving
without restraint, (C) Vibrocone crusher in operation (images courtesy Fives FCB).

ROLLS CRUSHERS
Rolls crushers represent one of the oldest types of compression machine and consist
of two counter rotating rolls, with the feed material fed into the gap between them.
The rolls surface is mostly smooth or has a superimposed pattern of wear protection.
,n some cases, teeth are afÀ[ed to the surface. $n e[ample of a modern rolls crusher is
shown in Figure 6.11.
Due to the wear displayed when crushing abrasive material, rolls crushers fell out of
favour, with gyratory and cone crushers typically preferred. However, over the last ten
years, rolls crushers have begun to regain some favour due to their ability to deliver very
high throughputs and advances in wear materials and teeth change-out procedures.
The main applications for rolls crushers are usually in oil sands, but a recent application
in iron ore in Brazil has been reported.
$ modern development of the rolls crusher is the H3*R crusher, which is discussed in
more detail in a later section.

FIG 6.11 – Modern rolls crusher (image courtesy ThyssenKrupp).

Comminution Handbook | Spectrum Series 21 87

 CHAPTER 6 t COMPRESSION MACHINES

MINERAL SIZERS
Sizers were developed in the mid-to-late 1970s to break coarse coal efÀciently within
compact dimensions. The basic concept of the sizer is to apply pressure to material and
cause it to fracture along natural planes of weakness. This is done by the use of two
rotors with large teeth on small diameter shafts driven at low speed by a direct high-
torque drive system. This concept is shown in the 00D twin shaft mineral SizerŒ
illustrated in Figure 6.1.
The rotating shafts in the sizer have large teeth attached to them that form scrolls
similar to Archimedean screws. These help to spread material around the machine.
The interlaced toothed rotor design allows undersize to pass through the gaps
generated by the slow-moving shafts to prevent further degradation of already
undersize material. An adjustable breaker bar is positioned centrally in line with and
below the breaker shafts.
The nature of the crushing mechanism for a sizer is shown in Figure 6.1. The material
is initially gripped by the leading faces of opposed rotor teeth. These subject the rock
to multiple point loading, inducing stress into the material to exploit any natural
weaknesses. From initial breakage, the product is further broken in tension by being
subjected to three-point loading, which is applied between the front tooth faces on
one rotor and rear tooth faces on the other. Any lumps of material that still remain

FIG 6.12 – An MMD Sizer (image courtesy MMD).

A B C

FIG 6.13 – Breakage in an MMD Sizer. (A) Stage 1 – initial breakage of large material: teeth closing. (B) Stage 2 – further
breakage of material as the teeth pass each other at the mid-point: teeth passing. (C) Stage 3 – breakage of remaining
oversize from teeth chopping through the breaker bar: teeth chop through breaker bar (images courtesy MMD).

88 Comminution Handbook | Spectrum Series 21

 CHAPTER 6 t COMPRESSION MACHINES

oversize are held by the Àxed teeth of the breaker bar and broken by the chopping
action of the rotors.
The 00D Sizer has a number of sizer, shaft and tooth conÀgurations to enable the
primary, secondary and tertiary reduction of materials. On certain materials, primary
00D sizers can process at rates in excess of 1 000 th. The method of reduction utilising
the weakness of materials in shear and tension offers a controlled three-dimensional
outfeed product with different tooth designs to facilitate different sizes of outfeed in an
efÀcient manner.
A typical reduction ratio in a sizer is 51, and rocks that are 000 mm in size and even
larger can be reduced. Sizers can handle dry hard and wet sticky material simultaneously
through the same machine, even where the moisture content is over 0 per cent. They
are suitable for softer rocks such as bauxite and pisolitic iron ore, but can also process
‘hard rock’ up to 400 0Pa with satisfactory results. For example, metalliferous ores
at CODELCO Andina and El Teniente (approximately 0 0Pa) are currently being
crushed successfully. The rock quality designation (RQD) gives a better indication of
the weaknesses that exist in the rocks to be processed and that can be exploited by the
00D method of reduction.
As Table 6.5 indicates, sizers are smaller than other crushers, with height being a
measure of particular signiÀcance because it is the main parameter deÀning the opening
that must be prepared when a sizer is to be used underground. The 1500 series of sizers,
which is currently the largest series, has a height of 195 mm, but larger capacity sizers
are being developed.
The nominal capacity of each unit is 1000 th of medium-hard limestone. The infeed
size is 750 mm and the product size is 50 mm.

TABLE 6.5
Comparison of primary mineral crushers (MMD).
Equipment Gyratory Jaw Sizer Double roll Impactor
crusher
Dimensions (mm) 11 938 2240 × 1680 1000 series 1800 × 1800 2000 × 2200
Weight (t) 120 170 60 70 85

ROLLER MILLS
%reakage occurs in roller mills by compression, and because this type of breakage is
more energy efÀcient than impact breakage, there is a trend towards using roller mills
for Àne-grinding circuits. The perennial problem with roller mills that limits their use
with hard or abrasive material is wear of the rollers and the grinding surfaces, but this
is now being reduced by the development of new materials for their construction and of
new designs for the wearing surfaces (refer to Figure 6.14).
Roller mills are suitable for grinding friable materials. For example, the Gundlach
machine can grind up to 40 th of 19 mm friable material and yield a 1– mm product.
Sturtevant is another manufacturer of roller mills. Table 6.6 provides some details of
horizontal roller mills.
The largest horizontal roller mills now in operation are those that are used to crush
hot cement clinker, leaving the kiln to the size at which it can be ground in roller mills

Comminution Handbook | Spectrum Series 21 89

 CHAPTER 6 t COMPRESSION MACHINES

A B C

FIG 6.14 – Some crushing roll designs: (A) one inch pin, (B) herringbone roll,
(C) 27 pocket tooth roll (images courtesy Gundlach crushers).

TABLE 6.6
Horizontal roller mills (Sturtevant Inc).
Power (kW) Feed size (mm) Product size Capacity (t/h)
2 19 12.7 mm to 850 micron 0.25 to 2
5 19 12.7 mm to 850 micron 1 to 7
2 × 10 38 19 to 1.6 mm 5 to 44

or ball mills. The hot clinker passes through a cooling zone and is then crushed by
horizontal rolls (refer to Figure 6.15). The rolls have a wide gap to break the coarser
lumps and maintain a high capacity. The kiln and cooling and crushing system operate
at up to 1 500 td. Hot clinker breaks easily, and hammer mills are also used for
breakage.

A B
FIG 6.15 – (A) Cooling system and (B) crushing rolls for hot clinker. Rolls are
up to 4 m long, 8.5 m wide and 0.9 m deep (images courtesy Claudius Peters).

HIGH-PRESSURE GRINDING ROLLS

,nterparticle compression is an efÀcient method of breaking particles that is used in
high-pressure grinding rolls (HPGR). The material is choke fed into the gap between the
spring-loaded, counter-rotating rolls and forms a bed that is compressed. Each particle
in the bed is stressed at every point that it has contact with other particles, meaning that
inter-particle breakage occurs. The compression of the bed is demonstrated in Figure 6.16,

90 Comminution Handbook | Spectrum Series 21

 CHAPTER 6 t COMPRESSION MACHINES

which shows that one roller rotates in Àxed bearings and the other in moveable bearings,
with high pressure applied to the moveable roller via the hydraulic pistons. Table 6.7
presents a comparison of HPGRs in terms of speciÀcations and capacities.
Agglomeration also occurs with compression machines, and a HPGR machine is often
followed by a device to break up the agglomerates. HPGRs are rarely used as stand-
alone machines to produce Ànished cement as HPGR products contain coarse particles
that must be ground further (refer to Figure 6.17). This is easier with compressed rather
than with uncompressed particles because the high pressure forms microcracks that
reduce grindability and minimise the energy required for further grinding. HPGRs
work on similar feed size particles to cone crushers, but the shapes of the product PSDs
differ because of different breakage mechanisms (Figure 6.18).

FIG 6.16 – High-pressure grinding rolls crushing mechanism (image courtesy ThyssenKrupp).

TABLE 6.7
Sizes and capacities of commercial high-pressure grinding rolls.
KHD FLSmidth Koppern Polysius Metso Fives FCB
Diameter (m) 1.0–2.6 0.58–2.7 0.80–2.80 0.95–2.60 0.8–3.0 1.6–4.6
Width (m) 0.5–2.3 0.26–1.85 0.20–1.60 0.65–1.75 0.5–2.0 0.54–1.67
Power (kW) 280–6000 100–5800 150–4000 440–6800 220–11 500 200–3800
Throughput (t/h) 30–4200 100–3000 35–2000 >3000 70–4800 70–1200
The Fives FCB machine refers to the Horomill.

Comminution Handbook | Spectrum Series 21 91

 CHAPTER 6 t COMPRESSION MACHINES

FIG 6.17 – Particle size distribution comparison of high-pressure grinding rolls and cone crushers (image courtesy Weir Minerals).

FIG 6.18 – Feed and product size distributions from high-pressure grinding rolls (image courtesy Hakan Dundar).

As shown in Figure 6.19, HPGRs, ball mills and separators are used in cement clinker
JULQGLQJFLUFXLWVLQYDULRXVFRQÀJXUDWLRQV
6HPLÀQLVKJULQGLQJDQGFORVHGFLUFXLWSUHJULQGLQJDUHWKHPRVWHIÀFLHQWPHWKRGVLQ
WHUPVRIVSHFLÀFHQHUJ\FRQVXPSWLRQ)LQLVKJULQGLQJLVQRWZLGHVSUHDGVLQFHLWJLYHVD
VWHHS36'WKDWUHGXFHVWKHVWUHQJWKRIWKHFHPHQW&ORVHGFLUFXLW+3*5VIROORZHGE\D
EDOOPLOOLVWKHEHVWRSWLRQIRUFHPHQWSURGXFWLRQ
The rolls in HPGRs require periodic reconditioning, and this means extraction and
WUDQVSRUWRIWKHUROOVWKHPVHOYHV$VLQJOHUROODQGDWWDFKHGVKDIWDQGEHDULQJKRXVLQJV
VXEDVVHPEO\ W\SLFDOO\ ZHLJKV RYHU  W $ +3*5 UROOV H[WUDFWRU WUDQVSRUWHU GHYLFH
LV DYDLODEOH ZKLFK XVHV ODVHU PHDVXULQJ WR H[WUDFW WKH UROOV HYHQO\ DQG DFFXUDWHO\ VR
WKDWWKHULVNRIGDPDJLQJWKHSUHFLVLRQPDFKLQHG+3*5IUDPHLVHOLPLQDWHG5XVVHOO
Mineral Equipment).
Briquetting is also carried out in compressed beds. Roller diameters in briquette
SUHVVHV DUH IURP ² P DQG SUHVVXUH LV IURP ² N1 $JJORPHUDWLRQ DQG

92 Comminution Handbook | Spectrum Series 21

 CHAPTER 6 t COMPRESSION MACHINES

A B

C D

FIG 6.19 – Alternative flow sheets in a cement plant using high-pressure grinding rolls, a ball
mill and a separator. (A) Open-circuit pregrinding, (B) hybrid grinding, (C) semi-finish grinding,
(D) closed circuit grinding, (E) finish grinding (images courtesy Hakan Dundar).

breakage occurs in compressed bed machines, and it is usual to follow HPGRs with
a de-agglomerating unit. HPGRs were only used with softer minerals for many years
because hard abrasive ores caused wear on the expensive rollers, but better materials
and new designs for the wearing surfaces have reduced this problem.
A Àner size distribution is produced from the centre of the HPGR than the two edges.
Depending on the product size distribution requirements, edge products are sometimes
recycled to the feed for further crushing.

VERTICAL ROLLER MILLS

Particles are broken in vertical roller mills (9R0s) by spring-loaded rollers compressing
them against a base plate as they Áow across the roller path from the centre to the edge
of the mill. Figure 6.0 shows the comminution zone and separator in a vertical mill.
Particles leaving the grinding zone in a 9R0 fall into a rising air or gas stream that
sweeps them into a separator in the top of the machine. The Àne product leaves the
circuit and the coarse product is returned for further grinding. ClassiÀcation between
the grinding zone and the separator occurs in regions where the particles are being swept
upwards, as well as in the separator. The circulating load is high for example, with coal
it may be 500 per cent at the separator and as high as 000 per cent directly above the
rollers. High speciÀc gravity mineral particles such as silica accumulate in the separator
reject and increase wear on the grinding surfaces. Techniques are being developed to
remove heavy, abrasive materials from the reject before it enters the grinding zone.
,n some machines, described as ballrace mills, balls replace the vertical rollers. 9R0s
have been used to grind coal for more than 100 years and are now used for harder and
more abrasive materials. Table 6.8 gives an indication of the mills now available and
their capacities. The main differences between roller mills are in the comminution zone,
and the various methods used to apply forces to the rollers are shown in Figure 6.1
(%rundiek, 1989).

Comminution Handbook | Spectrum Series 21 93

 CHAPTER 6 t COMPRESSION MACHINES

A B

FIG 6.20 – Vertical roller mill: (A) rollers and separator in a Gebr Pfeiffer mill,
(B) typical particle size distribution of product and reject (images courtesy Gebr Pfeiffer).

TABLE 6.8
Sizes and capacities of vertical roller mills.
Power (kW) Feed rate (t/h)
FLSmidth
Cement 800–11 000 35–685
Slag 900–13 200 25–500
Polysius
Cement raw material 580–4800 90–740
Clinker 3000 Blaine 502–3188 33–209
Granulated Slag 4500 Blaine 700–4450 22–139
Hard coal 50° Hardgrove 30–1250 22–96
Loesche
Coal 400–2400 40–300
Cement 2500–7800 60–340
Ore 7800 >2000
Gebr Pfeiffer
Cement raw material 1600–12 000 250–1400
Granulated slag 2000–6000 Blaine 2500–12 000 70–390
Cement 2000–6000 Blaine 2200–12 000 80–550

With each different method of applying pressure, the rollers track differently.
Uncompacted cement tends to aerate so that a compacting roller leads each grinding
roller and forms a compact mass that can be ground. The PSD of cement is sharper
with VRMs than with ball mills and is in a narrower band. A sharper cut potentially
means less variability in product consistency and more predictable results in product

94 Comminution Handbook | Spectrum Series 21

 CHAPTER 6 t COMPRESSION MACHINES

A B

C D

FIG 6.21 – Methods of applying force to grinding rollers. (A) Raymond ring-roller mill, (B) EVT (CE) roller
mill, (C) Loesche roller mill, (D) MPS roller grinding mill, (E) ring-ball mill (images courtesy Loesche).

performance. %y changing the operating parameters in 9R0s, signiÀcant adjustments

in the PSD, retention time and Àneness of the Ànished cement can be achieved. This can
help with plant operations as production is switched between different cement types.

HOROMILLS
Horomills, seen in Figure 6., were developed by Fives FC% and are a recent addition
to the group of mills that use compression breakage.

FIG 6.22 – The Horomill (image courtesy Fives FCB).

Comminution Handbook | Spectrum Series 21 95

 CHAPTER 6 t COMPRESSION MACHINES

The key components of the Horomill are a shell driven at supercritical speed and a roller
compressing the mineral bed at high pressure by means of hydraulic jacks. Capacities
for different product sizes (Blaine numbers) for cement, slag and raw meal are given in
Table 6.9. The electrical power available with the largest machine is 2940 kW.
7KH +RURPLOO GLVFKDUJH LV HOHYDWHG WR WKH FODVVLÀHU DQG WKH FORVHG FLUFXLW ZLWK WKH
DLU ÁRZV LV VKRZQ LQ )LJXUH  7KH PDWHULDO LV IHG IURP RQH VLGH DQG GLVFKDUJHG
from the other after being ground several times between the shell and the roller as it
passes through the mill. The material centrifuges and scrapers remove the compressed,
comminuted bed from the wall and present it in a loosened form to the roller for further
grinding. The material being ground undergoes a multiple and controlled compression
EHWZHHQWKHUROOHUDQGWKHVKHOO·VJULQGLQJWUDFN$VZLWKRWKHUUROOHUPLOOVWKHVSHFLÀF
HQHUJ\FRQVXPSWLRQLVORZ)LJXUHSUHVHQWVWKH36'IRU+RURPLOOVIRUWKHIHHGDQG
various product streams.

TABLE 6.9
Capacity of Horomills.
Size Indicative capacity t/h
Shell diameter Portland cement Blast furnace slag Raw meal
(mm) 3000 Blaine 4000 Blaine 3500 Blaine 4500 Blaine
2000 20 12 12 8 35–55
2800 50 30 30 20 90–140
3600 95 60 60 40 175–275
3800 110 70 130 90 260
4400 160 100 100 70 255–470

FIG 6.23 – Horomill-TSV® classifier closed circuit (image courtesy Fives FCB).

ENERGY USED IN DIFFERENT CIRCUITS

Before the 1980s, crushing circuits were the predominant mechanism for reducing
PDWHULDOWRDEDOOPLOOIHHGVL]HW\SLFDOO\LQWKHUDQJHRI²PP,QWKHLQWHUYHQLQJ
years, a whole range of alternative circuits have emerged and found varying degrees
RIDFFHSWDQFH7KHPDLQFLUFXLWKDVEHFRPHWKHVHPLDXWRJHQRXVJULQGLQJ6$*EDOO
PLOOFRQÀJXUDWLRQZKHUHFUXVKHUVDUHRQO\XVHGIRUSULPDU\VL]HUHGXFWLRQWRJHQHUDWH

96 Comminution Handbook | Spectrum Series 21

 CHAPTER 6 t COMPRESSION MACHINES

FIG 6.24 – Particle size distribution of streams in a Horomill-TSV® separator circuit

(image courtesy Hakan Dundar, data available from Metso).

a SAG feed with a top size of approximately 250–400 mm rocks and for the crushing of
pebbles generated by the SAG mill. Several circuits are now used to achieve the same
F80 to P80 for an ore, with all requiring different amounts of energy. Many factors are
considered when choosing a circuit, and Table 6.10 gives an indication of the relative
energy consumption of different circuits.
It is interesting to note that the push to reduce energy consumption in milling and
WKH UHODWLYH HQHUJ\ HIÀFLHQF\ RI FUXVKLQJ KDV OHG WR D UHVXUJHQFH LQ QRQ6$*EDVHG
ÁRZVKHHWV5HFHQWH[DPSOHVRIFUXVKHU+3*5EDOOPLOOFLUFXLWVLQFOXGH1HZPRQW
Boddington and Karara Mining in Australia and Cerro Verde in Chile. There is also a
VLJQLÀFDQWLQFUHDVHLQIXOORUSDUWLDOSUH6$*FUXVKLQJRSWLRQV6HYHUDORSHUDWLRQVKDYH
DOUHDG\GHSOR\HGVXFKDQDSSURDFKLQFOXGLQJ,QPHW7URLOXVDQG1HZFUHVW5LGJHZD\
and wider application is being considered to explore options for further energy saving.

TABLE 6.10
Relative energy consumption of different comminution circuits (Marsden, 2011).
Circuit Decrease (%)
Semi-autogenous grinding mills, ball mills Base case
Semi-autogenous grinding mills, pebble crusher, ball mills 6.4
Autogenous grinding mills, pebble crusher, ball mills 22.1
Three-stage crushing, ball mills 25.7
Two-stage crushing, high-pressure grinding rolls, ball mills 34.6
Two-stage crushing, high-pressure grinding rolls, agitator mills ≈41.4
Improved classification ≈46

BIBLIOGRAPHY
Belotserkovsky, K E, 2010. Method for controlling process parameters of a cone crusher, US
3DWHQW6DQGYLN
Brundiek, +  7KH UROOHU JULQGLQJ PLOO ² LWV KLVWRU\ DQG FXUUHQW VLWXDWLRQ Aufbereitungs-
Technik
Marsden, - 2  ,QQRYDWLRQ DQG HQHUJ\ HIÀFLHQF\ LQ FRSSHU H[WUDFWLRQ SDSHU SUHVHQWHG WR
3URFHPLQ&RQIHUHQFH6DQWLDJR'HFHPEHU
Taylor, J C,  5KRGD[LQHUWLDOFRQHJULQGHUJournal of The South African Institute of Mining and
Metallurgy,2FWREHUSS²

Comminution Handbook | Spectrum Series 21 97

 CHAPTER 6 t COMPRESSION MACHINES

Catalogues for crushers referred to in this chapter are available on the internet. The companies are
x Claudius Peters
x FLSmidth
x Gebr Pfeiffer
x Loesche Gmbh
x 0etso Nordberg (Symons crushers)
x 00D Group (Sizers)
x Pennsylvania Crushing
x Russell 0ineral Equipment (R0E)
x Sandvik (9ibrocone crushers)
x Thyssen.rupp.

98 Comminution Handbook | Spectrum Series 21

 HOME

Chapter 7

HIGH-SPEED IMPACT MILLS

Glenn Schumacher, Alban Lynch
and Thomas Warne

INTRODUCTION
In high-speed impact mills, breakage and shattering occurs by fast-moving hammers
impacting on slow-moving rock, or by the collision of fast-moving rocks with other
rocks or a breaker plate. Machines that use both techniques are available, but there has
been a greater application for comminution of moderate-to-low-hardness material of
low abrasion potential. The reason is because wear is a potentially serious problem.
A rule of thumb is that steel hammers are suitable for materials containing no more than
Àve per cent silica doubtful for a silica content of ten per cent to  per cent dangerous
for  per cent to  per cent and prohibitive if it e[ceeds  per cent. 5ock on rock
breakage reduces the problem of wear.
Four common types of high-speed impact mill are reviewed in this chapter.

VERTICAL SHAFT IMPACT CRUSHERS

The %armac rock-on-rock crusher is an e[ample of a vertical shaft impact 96I crusher.
It was devised in  by -im Macdonald, an engineer with the :ellington &ity &ouncil
in New Zealand. The basis for the crusher design was that stones will break if banged
together hard enough, and that steel will be protected from abrasion if covered with a
layer of trapped stones.
The crusher has a rotor that acts as a high-velocity, dry stone pump, hurling a
continuous rock stream into a stone-lined crushing chamber. Broken rock about
² mm in diameter enters the top of the machine from a feeder set and is accelerated
in the rotor to be discharged into the crushing chamber at velocities of up to  ms.
&ollision of high-speed rocks, with rocks falling in a separate stream or with a rock-lined
wall, causes shattering refer to Figure .. The product is typically gravel and sand-
si]ed particles. Barmac crushers are available from  to  k:.

Comminution Handbook | Spectrum Series 21 99

 CHAPTER 7 t HIGH-SPEED IMPACT MILLS

FIG 7.1 – Barmac vertical shaft impact crusher. The dry stone ‘pump’
(rotating drum) is shown in the centre of the machine (image courtesy Metso).

The product size distribution can be controlled by the rotor speed as shown in
Figure ..
In the Magotteau[ MA*·Impact® 96I crusher Figure ., the material to be crushed
falls onto a distributor at the centre of the rotating table. Particles are accelerated by
impellors on the table and driven toward anvils on the peripheral ring where the impact
causes shattering. The ma[imum table diameter is  mm, and four to Àve impellors
and  to  anvils can be used.
Table . gives an indication of the capacities of 96I crushers for different motor sizes.
&apacities are highly dependent on the characteristics of the material being crushed.

ϭϬϬ
ϵϬ &ĞĞĚ
WƌŽĚƵĐƚͲϰϱŵͬƐƚŝƉƐƉĞĞĚ
ϴϬ WƌŽĚƵĐƚͲϲϱŵͬƐƚŝƉƐƉĞĞĚ
ϳϬ
ϲϬ
WĂƐƐŝŶŐ й

ϱϬ
ϰϬ
ϯϬ
ϮϬ
ϭϬ
Ϭ
Ϭ͘ϭ ϭ ϭϬ ϭϬϬ
^ŝnjĞ ;ŵŵͿ

FIG 7.2 – Product size distribution showing the effect of rotor speed
in a Barmac vertical shaft impact crusher (image courtesy Metso).

100 Comminution Handbook | Spectrum Series 21

 CHAPTER 7 t HIGH-SPEED IMPACT MILLS

FIG 7.3 – Magotteaux MAG’Impact® vertical shaft impact crusher (image courtesy Magotteaux).

TABLE 7.1
Examples of verticle shaft impact crusher specifications.
Barmac MAG’Impact®
Motor (kW) 75–600 200–500
Capacity (t/h) 60–477 200–500
Maximum rock velocity (m/s) 85 65

HAMMER MILLS
Hammer mills work on the principle that most materials will crush, shatter or pulverise
upon impact. The hammer mill is the most widely used crusher with thousands
employed worldwide in a large number of industries for primary and secondary
crushing of various material. In the minerals industry, hammer mills are used to crush
materials such as hard coal and lignite, limestone, bau[ite, phosphate rock and other
soft to medium-hard rocks or ores. Mill sizes range from small laboratory units to very
high capacity machines. A large  k: industrial machine is shown in Figure ..

FIG 7.4 – Sandvik 1800 kW hammer mill (image courtesy Sandvik).

Comminution Handbook | Spectrum Series 21 101

 CHAPTER 7 t HIGH-SPEED IMPACT MILLS

A hammer mill consists of a cylindrical chamber containing a horizontal shaft, which

in large industrial mills rotates between  and  revmin. Hammers are attached
to the shaft and have tip speeds up to ² mmin. Material enters the crushing
chamber through a feed chute, usually by gravity, and is shattered by a combination
of repeated hammer impacts, collisions with the walls of the grinding chamber and
particle²particle impacts. Figure . shows how a hammer mill works.
Perforated metal screens or bar grates cover the discharge opening of the mill and retain
coarse material for further grinding while allowing properly sized materials to pass as
Ànished product. The particles that are caught between the hammers and the screen bars
where the gap is small are broken by attrition. The feed to hammer mills is usually in the
range of ² mm and the  per cent passing size of the product is typically about
² mm. The feed rate depends on the hardness of the material and the product size.
In the maMority of hammer mill applications, the key factor determining Ànished
particle size is the screen and because of this, the screen size provides appro[imately
 per cent of the control over the Ànished particle size. The remaining  per cent is
attributed to the force of the impact on the material being processed. Force is determined
by rotor speed and the size and number of hammers.
The distance between the grinding wall and the hammer circle can be controlled in
some hammer mills by a mechanical or a hydraulic mechanism to adjust for wear and
optimise performance. The speed selected for a mill is based on the rotor diameter and
the material being crushed. /arge mills are operated at  to  revmin, smaller
industrial mills at up to  revmin and laboratory mills at  revmin.
:ear is a serious problem with hammer mills when abrasive material is broken. The cost
of wear ² which includes downtime as well as replacement parts ² can be reduced to some
e[tent by using hammers that are reversible, abrasion-resistant wear plates that are easily
replaced, and bar grates or perforated screens that are made from abrasion-resistant steel.
Because of wear, hammer mills are unsuitable for use with hard and abrasive ores.

ƌŽƚŽƌĂŶĚƌŽƚŽƌƐŚĂĨƚ
ĐĂƐŝŶŐ
ŚĂŵŵĞƌ

ŐƌĂƚĞ


FIG 7.5 – Method of operation for a hammer mill (image courtesy Hakan Dundar).

102 Comminution Handbook | Spectrum Series 21

 CHAPTER 7 t HIGH-SPEED IMPACT MILLS

The capacity range of hammer mills designed by two vendors is indicated in

Tables . and ., which refer to their use with coal. These data are indicative only.
Detailed information about the performance of hammer mills can be obtained from the
manufacturers.

TABLE 7.2
Sandvik Mining hammer mills used for coal crushing.
Roll diameter × length (mm) Throughput (t/h)
1000 × 1000 75
1400 × 1800 250
1400 × 2600 400
1600 × 2600 500
1600 × 3400 700
Feed to 250 mm; product size to <1 mm depending on feed material; installed power to 1800 kW.

TABLE 7.3
Schutte Buffalo hammer mills used for coal crushing.
Throughput (t/h) Power (kW) Rev/min (max)
0.25–0.5 7.5 3500
0.5–1.0 15 3500
1.5–5 30 2325
6–15 75 1860
20–50 150 1395
50–100 225 1395
Maximum tip speed is 74 m/s.

ATRITA MILLS (COMBINED HAMMER AND ATTRITION MILLS)

Atrita mills have found applications in power stations. The Duple[ Atrita mill Figure .
is in effect two separate mills mirrored around the centre line of the mill. &oal Áows
from the central feed chute into both sides of the mill, initially passing to the hammer
section. Heavy swing hammers mounted on the shaft fracture the coal through multiple
interactions with both the hammers and striker plates mounted in the casing.
Partially ground particles Áow through a  mm screen to the Àrst effect attrition zone
upstream side of attrition section. Here the centripetal force induced by the spinning
rotor disc moves them outwards and they are further ground by impeller clips attached
to the disc. The coal passes through an annulus between the rotor disc and casing into
the second effect attrition zone where turbulent air currents and the clips continue to
grind the coal. Particles then pass into the e[haust fan, which consists of radial inner and
curved outer fan blades, and are transported to product destination. Further breakage
occurs as particles pass through the fan. Particle size distribution of the products from
each grinding zone, including the fan, are shown in Figure ..
At the *ladstone Power 6tation in 4ueensland, where  mills are operating, the
mill feed is ² mm and the Ànal product is  per cent passing . mm. The mills
at *ladstone have a rotational speed of  revmin. The tip velocity of the crusher

Comminution Handbook | Spectrum Series 21 103

 CHAPTER 7 t HIGH-SPEED IMPACT MILLS

FIG 7.6 – Dual Atrita mill where the coal flows through the hammer and attrition
mills using the same horizontal shaft (image courtesy Glenn Schumacher, AGL).

FIG 7.7 – Dual Atrita mill with product size distribution from each section of the mill (image courtesy Glenn Schumacher, AGL).

hammers is estimated to be 52 m/s and the tip velocity of the grinding clips on the
attrition wheel is calculated at 100 m/s. The coal passes through the hammer mill
VHFWLRQWKHQWKURXJKWKHÀUVWDQGVHFRQGDWWULWLRQVHFWLRQVDQGHPHUJHVDVSXOYHULVHG
fuel from the fan section.

BEATER WHEEL MILLS

Lignite (or brown coal) is widely used in power plants, with an annual worldwide use
of 1000 Mt. Germany is the largest producer and Australia is third with 120 Mt. Beater
wheel mills are used for pulverising lignite and brown coal, which typically contain a

104 Comminution Handbook | Spectrum Series 21

 CHAPTER 7 t HIGH-SPEED IMPACT MILLS

very high moisture content up to  per cent and tend to be Àbrous in nature. High-
speed beater mills fall into two categories:
. a beater wheel mill in which the comminution element consists of a high-speed
armoured beater wheel alone
2. a hammerbeater wheel mill with a primary hammer section in front of the beater
wheel section.
A beater wheel in effect is an armoured impellor that both impacts directly with the raw
coal feed as well as propelling the coal against the armoured mill casing. These actions both
bring about comminution of the feed and also allow the mi[ing of hot gases with the coal
to dry it. At the same time the beater wheel acts as a fan to provide the means of drawing
hot gas from the boiler furnace, which enters the mill with the feed, and propelling the
pulverised mill product pulverised fuel ² PF and associated carrier gas to the site of
combustion in the boiler furnace. Most have a ¶bo[ classiÀer· if any classiÀer is present
at all, which is actually an inertia-type static classiÀer in which separation occurs by a
change in Áow direction. In nearly all major lignite-Àred power plants, the crude lignite is
dried and pulverised in beater wheel mills.
The high moisture load of these coals requires signiÀcant drying to occur in the mill. The
use of hot gases from the boiler furnace results in very large boiler plant and comple[ burner
arrangements to handle the vapour load and circulating gases. To achieve the necessary
drying, mills pulverising lignite brown coal run very hot with gas temperatures around
2²ƒ&. :ith such high moisture loads the mills are required to handle very large gas
volumes as the moisture is converted in water vapour plus the drying boiler furnace
gases and the conditions within the mill are such that corrosion-erosion mechanisms can
occur in these machines. :hilst the coal is generally ¶soft·, the high speeds, high volumes
of material being processed and resulting high velocities when coincidental with some
contaminants in the coal, such as silica, mean that high wear rates can be encountered.

FIG 7.8 – Beater wheel pulveriser at AGL Loy Yang A Power Station,
Victoria, Australia (image courtesy Glenn Schumacher, AGL).

Comminution Handbook | Spectrum Series 21 105

 CHAPTER 7 t HIGH-SPEED IMPACT MILLS

The beater wheel shown in Figure . stands  m high, weighs 2 t, spins at appro[-
imately  revmin within the mill and operates in temperatures around 2²ƒ&.
The pulverised coal is classiÀed and the Àne particles are blown into the boiler to fuel
the furnace Àre. 6ilica in the lignite is abrasive and the beater wheels must be changed
regularly. The largest beater wheel offered by Alstom is  mm high.
The Deutsche Babcock combined hammer and beater wheel mills at the Hazelwood
power station in 9ictoria, Australia each has a  k: motor and  th capacity. There
are  in all ² a signiÀcant amount of comminution devices in one place.

BIBLIOGRAPHY
Worley, -, 6chumacher, * and -uniper, /, 2. (ffects of coal properties upon mill life and
performance, in Proceedings Tenth Australian Coal Science Conference, Brisbane, ² November
Australian Institute of (nergy: Melbourne.

&atalogues for the high-speed impact mills referred to in this chapter are available on the internet.
The companies are:
• Alstom Beater :heel Mills
• Atrita mills
• Magotteau[ MA*·Impact® 9ertical 6haft Impact &rusher
• Metso Barmac 9ertical 6haft Impact &rusher
• 5iley Power
• 6andvik Hammer Mills
• 6chutte Buffalo Hammermill //&.

106 Comminution Handbook | Spectrum Series 21

 HOME

Chapter 8

STIRRED MILLS
Udo Enderle, Katie Barns,
Joe Pease and Bodo Furchner

INTRODUCTION
*ULQGLQJ LV GHÀQHG DV EUHDNLQJ XS D VROLG VWUXFWXUH E\ VXEMHFWLQJ LW WR PHFKDQLFDO
HQHUJ\*ULQGLQJSURFHVVHVFDQEHFODVVLÀHGDVWUXHFRPPLQXWLRQGHDJJORPHUDWLRQRU
GLVDJJUHJDWLRQ6WLUUHGRUDJLWDWRUEHDGPLOOVDUHXVHGIRUJULQGLQJSDUWLFOHVZLWKDIHHG
VL]HVPDOOHUWKDQ²—P7KH\EHORQJWRWKHFODVVRIPDFKLQHVZLWKIUHHO\PRYLQJ
JULQGLQJPHGLDWKHFODVVLFH[DPSOHRIZKLFKLVEDOOPLOOV,QURWDWLQJPLOOVOLNHEDOOPLOOV
HQHUJ\LVLPSDUWHGWRWKHPHGLDE\WKHURWDWLQJVKHOO7KLVOLPLWVWKHHQHUJ\WRWKHVKHOO·V
FULWLFDOVSHHG,QFRQWUDVWLQVWLUUHGPLOOVWKHVKHOOLVVWDWLRQDU\DQGHQHUJ\LVLPSDUWHG
YLDLQWHUQDOURWDWLQJPHFKDQLVPVUHPRYLQJWKHFULWLFDOVSHHGOLPLW
6WLUUHGPLOOVDUHIXUWKHUFDWHJRULVHGLQWRORZVSHHGVWLUUHGPLOOVDQGKLJKVSHHGVWLUUHG
PLOOV/RZVSHHGPLOOVDUHW\SLÀHGE\PLOOVVXFKDV¶WRZHUPLOOV·ZKLFKXVHVWHHOPHGLD
KDYHHQHUJ\LQWHQVLWLHVFORVHWRURWDWLQJPLOOVDQGDUHXVHGIRUFUHDWLQJSURGXFWLQWKH
UDQJHEHWZHHQEDOOPLOOVDQGXOWUDÀQHVWLUUHGPLOOVEURDGO\²—PIRUP+LJK
VSHHG VWLUUHG PLOOV ZHUH GHYHORSHG IRU XOWUDÀQH JULQGLQJ DSSOLFDWLRQV ZLWK SURGXFW
VL]HEHORZPRI—P7KH\KDYHVLJQLÀFDQWO\KLJKHUSRZHULQWHQVLW\PD\EHYHUWLFDO
RU KRUL]RQWDO DQG XVH D YDULHW\ RI VWLUUHU GHVLJQV ZLWK ÀQH FHUDPLF PHGLD ,QLWLDOO\
DSSOLHG WR UHODWLYHO\ VRIW LQGXVWULDO PLQHUDOV WKH\ ZHUH WKHQ GHYHORSHG IRU XOWUDÀQH
JULQGLQJLQWKHPLQLQJLQGXVWU\LQUHVSRQVHWRWKHLQFUHDVLQJO\ÀQHJUDLQHGRUHERGLHV
2QO\VLQFHDERXWKDYHVRPHEHHQDSSOLHGWRFRDUVHUJULQGVDERYHPRI—P
ZKHUHSUHYLRXVO\ORZVSHHGVWLUUHGPLOOVRUURWDWLQJPLOOVZRXOGKDYHEHHQXVHG
6WLUUHG PLOOV FDQ EH IRU HLWKHU ZHW JULQGLQJ ZLWK D VOXUU\ IHHG RU GU\ JULQGLQJ
:HWJULQGLQJLVPRUHHQHUJ\HIÀFLHQWWKDQGU\JULQGLQJ7KLVLVSDUWO\GXHWRWKHHQHUJ\
WKDW WKH FODVVLÀHU LQ WKH GU\ SURFHVVLQJ FRQVXPHV KRZHYHU IRU VRPH DSSOLFDWLRQV ²
IRU H[DPSOH FHPHQW ² D GU\ SURGXFW LV HVVHQWLDO DQG GLFWDWHV D GU\JULQGLQJ SURFHVV
,Q PRVW PLQLQJ DSSOLFDWLRQV WKH XSVWUHDP DQG GRZQVWUHDP SURFHVVLQJ LV ZHW IRU
H[DPSOH EDOOPLOOLQJÁRWDWLRQRUOHDFKLQJVRZHWJULQGLQJLVW\SLFDOO\XVHG,QRWKHU

Comminution Handbook | Spectrum Series 21 107

 CHAPTER 8 t STIRRED MILLS

applications, some ores are dry ground and combusted before leaching (for example,
treating some Nevada gold ores) while grinding coal for combustion. It is possible for a
wet-grinding process to be employed and the product subsequently dewatered and dried.
This will depend on the relative energy and complexity of milling and drying. As the
slurries in the wet mills are highly concentrated, up to 75 per cent solid content, an energy-
efÀcient drying of the slurry is possible. 6een from the aspect of energy consumption, there
may be an advantage for wet grinding including drying compared to dry processing.
2nly about two to  per cent of the energy expended in grinding is efÀciently used
for the generation of new surfaces; most is converted to heat or lost through ineffective
impacts and loads. This emphasises the necessity to improve all grinding practices,
particularly Àne grinding, which is so energy intensive.
Tumbling mills were found to be particularly inefÀcient for Àne grinding because the
slow speed of rotation led to low power intensity and a low rate of ball collisions. In 1928
a stationary grinding chamber was proposed in which spherical grinding media were
moved by a slowly rotating agitator (.wade and 6chwedes 1997). 0ill performance
could be controlled by increasing the agitator speed and using Àner balls.
In 1948 the du Pont Company built a high-speed mill for grinding pigment using
this concept. It was referred to as a ‘sand mill’ because the grinding media originally
consisted of Àne-grained sand fractions. This is why high speed mills are still referred
to as ‘sandmills’ in the 86, although they now use ceramic media si]ed between .1 and
8 mm. The initial stirrer was a vertical cantilevered shaft with solid discs that operated
at speeds of 600 to 800 rev/min (8–11 m/s). The discs were concentric in alignment with
slurry entering from the bottom of the mill and exiting at the top.

COMMINUTION IN STIRRED MILLS

Hans Rumpf divided comminution processes into grinding and fracturing according to
the particle si]e of the feed stocN. 'ifferent machines should be used as the particle si]e
becomes Àner, as shown in )igure 8.1.

FIG 8.1 – Particle size of the feed stock showing the size range where agitator bead mills fit in. The
coarse end of this range is populated by slow-speed stirred mills, the finer end by high-speed stirred mills.
Bead mills are used to disperse both colloidal and coarser particles (image courtesy Hans Rumpf).

108 Comminution Handbook | Spectrum Series 21

 CHAPTER 8 t STIRRED MILLS

,QDJLWDWRUPLOOVWKHEHDGVDUHNHSWLQPRWLRQE\GLVFVRUSHJVDWWDFKHGWRKLJKVSHHG
VWLUUHUVDQGWKHPDLQEUHDNDJHPHFKDQLVPLVWKHDWWULWLRQRISDUWLFOHVE\FROOLVLRQZLWK
HDFKRWKHURUZLWKWKHPHGLD2SHUDWLQJWKHVKDIWDWKLJKHUWLSVSHHGVWUDQVPLWVPRUH
HQHUJ\ IURP PHGLD WR SDUWLFOHV DQG UHVXOWV LQ D ÀQHU JULQG 6RPH DWWDFKPHQWV WR WKH
VKDIWXVHGWRWUDQVPLWHQHUJ\DUHVKRZQLQ)LJXUH)HHGPDWHULDOFDQEHXSWRPP
DOWKRXJKLWLVXVXDOO\PXFKÀQHU3URGXFWVZLOOXVXDOO\EHOHVVWKDQ—PIRUKLJKVSHHG
VWLUUHGPLOOV0LOOVFDQEHIHGE\VOXUU\ZHWJULQGLQJRUGU\VROLGV*ULQGLQJPHGLD
DUHFKRVHQWRVXLWWKHSURFHVVWKH\FDQEHVPDOOVWHHORUFHUDPLFEDOOVRUVL]HIUDFWLRQV
RILQHUWPLQHUDOV²VRPHWLPHVHYHQDIUDFWLRQVFUHHQHGIURPWKHRUH,QHUWPLQHUDOVRU
FHUDPLFEDOOVPD\EHLPSRUWDQWWRGRZQVWUHDPSURFHVVLQJIRUH[DPSOHÁRWDWLRQRU
SURGXFWVLURQFRQWDPLQDWLRQZRXOGDGYHUVHO\DIIHFWWKHFRORXURIFHPHQW

FIG 8.2 – Attachments to a shaft to create agitation (image courtesy Udo Enderle).

Stress energy, stress rate and specific energy

2SWLPXP UHVXOWV DUH REWDLQHG E\ GLIIHUHQW VWUHVVLQJ PHFKDQLVPV GHSHQGLQJ RQ WKH
SURSHUWLHV RI WKH PDWHULDO EHLQJ JURXQG ,I WKH SURFHVV LV WUXH FRPPLQXWLRQ ² WKDW LV
EUHDNDJHLQWRPLQXWHSDUWLFOHV²WKHIHHGSDUWLFOHVPXVWEHVXEMHFWHGWRVXIÀFLHQWHQHUJ\
IURPSUHVVXUHDQGLPSDFWIRUIUDJPHQWDWLRQWRRFFXU,QFRQWUDVWLIWKHREMHFWLYHLVWR
EUHDNWKHFRKHVLRQRIDJJORPHUDWHVWKHQWUDQVIHURIHQHUJ\WKURXJKVKHDULQJIRUFHLV
XVXDOO\VXIÀFLHQW7KHLPSRUWDQWIDFWRUVDUHVWUHVVHQHUJ\VWUHVVUDWHDQGVSHFLÀFHQHUJ\
6WUHVVHQHUJ\LVWKHNLQHWLFHQHUJ\WKDWLVWUDQVIHUUHGWRWKHSDUWLFOHFDXJKWEHWZHHQ
JULQGLQJ EHDGV 6WUHVV UDWH LV WKH DEVROXWH QXPEHU RI LPSDFWV DQG VSHFLÀF HQHUJ\
N:KNJ LV WKH SURGXFW RI VWUHVV HQHUJ\ DQG VWUHVV UDWH WKDW PHDVXUHV WKH HQHUJ\
GLUHFWO\LPSDUWHGWRHDFKXQLWPDVVRIVROLG%\PDWFKLQJWKHJULQGLQJPHGLDVL]HWRWKH

Comminution Handbook | Spectrum Series 21 109

 CHAPTER 8 t STIRRED MILLS

JULQGLQJ DSSOLFDWLRQ WKH VSHFLÀF HQHUJ\ UHTXLUHG FDQ RIWHQ EH UHGXFHG VXEVWDQWLDOO\
'HFUHDVLQJJULQGLQJEHDGVL]HUHGXFHVWKHVWUHVVHQHUJ\RIHDFKLPSDFWEXWLQFUHDVHV
WKHQXPEHURILPSDFWVLQDJLYHQJULQGLQJYROXPHWKDWLVLWLQFUHDVHVWKHVWUHVVUDWH
)RUH[DPSOHWRFUXVKDVROLGEORFNZLWKDKDPPHURQHFRXOGVHOHFWIURPYDULRXVVL]HV
RIKDPPHUDVUHSUHVHQWHGLQ)LJXUH

FIG 8.3 – Crushing a solid block with a hammer (image courtesy Dr Stefan Mende).

$KDPPHUWKDWLVWRRVPDOOZRXOGRQO\WUDQVIHUHQRXJKHQHUJ\WRNQRFNVPDOOFRUQHUV
DQG HGJHV RII WKH EORFN WKHUH ZRXOG QRW EH HQRXJK HQHUJ\ WUDQVIHUUHG WR FDXVH D
IUDFWXUHWKURXJKWKHHQWLUHEORFN7KLVPHDQVWKDWDPLQLPXPVWUHVVHQHUJ\LVUHTXLUHG
WRDGYDQFHWKHFUXVKLQJSURFHVV,IDODUJHVOHGJHKDPPHULVXVHGIRUWKLVH[SHULPHQW
WKHEORFNZRXOGFHUWDLQO\EHFUXVKHG6HYHUDOIUDFWXUHVZRXOGSUREDEO\EHWULJJHUHGLQ
DGGLWLRQWRWKHSULPDU\IUDFWXUHKRZHYHULWLVOLNHO\WKDWWKHWDEOHRQZKLFKWKHEORFN
ZDVVLWWLQJZRXOGDOVREHGDPDJHG7KLVLOOXVWUDWHVWKDWWKHVHOHFWLRQRIJULQGLQJEHDGV
WKDW DUH WRR ODUJH OHDGV WR WUDQVIHUULQJ PRUH VWUHVV HQHUJ\ WKDQ LV UHTXLUHG WR WULJJHU
EUHDNDJH7KHH[FHVVHQHUJ\FDQQRWEHVWRUHGVRLVFRQYHUWHGWRKHDWDQGDVDUXOHWR
LQFUHDVHGZHDURQWKHJULQGLQJPHGLDDQGWKHPDFKLQH
2SWLPXP UHVXOWV FDQ EH H[SHFWHG ZKHQ WKH VWUHVV HQHUJ\ WUDQVIHUUHG URXJKO\
FRUUHVSRQGVWRWKHPLQLPXPDPRXQWRIHQHUJ\UHTXLUHGWRWULJJHUEUHDNDJH6HOHFWLQJ
WKH RSWLPXP JULQGLQJ PHGLD VL]H UHVXOWV LQ D VWUHVV UDWH DW ZKLFK HYHU\ VWUHVV HYHQW
HIIHFWLYHO\ WULJJHUV SDUWLFOH EUHDNDJH 8QGHUVWDQGLQJ VWUHVV HQHUJ\ VWUHVV UDWH DQG
VSHFLÀFHQHUJ\DUHLPSRUWDQWZKHQGHVLJQLQJPLOOVIRUXOWUDÀQHJULQGLQJ

Classifier wheels
,Q GU\ SURFHVVLQJ D FODVVLÀHU VHSDUDWHV FRPPLQXWHG SDUWLFOHV LQWR FRDUVH DQG ÀQH
VWUHDPVDFFRUGLQJWRWKHEDODQFHRIWZRIRUFHVDGUDJIRUFHGXHWRSDUWLFOHPRWLRQLQ
DLUDQGDQLQHUWLDOIRUFHGXHWRWKHURWDWLRQRIDFODVVLÀHUZKHHO7KHVLPSOHVWIRUPRID
FODVVLÀHUZKHHOFRQVLVWVRIWKHKXEWKHYDQHVDQGWKHFRYHUSODWHDVVKRZQLQ)LJXUH
7KHKXEDQGFRYHUSODWHDUHDUUDQJHGLQSDUDOOHOZKLOHDODUJHQXPEHURIYDQHVDUH
PRXQWHGUDGLDOO\RQWKHRXWHUGLDPHWHURIWKHFODVVLÀHUZKHHO7KHVHSDUDWLRQHIÀFLHQF\
RIDFODVVLÀHULPSURYHVZLWKDQLQFUHDVHGQXPEHURIYDQHV7KHVSHHGRIWKHFODVVLÀHU
ZKHHODQGLWVSHULSKHUDOVSHHGGHWHUPLQHWKHPD[LPXPSDUWLFOHVL]HRIWKHÀQHPDWHULDO

Classifier wheel and air jet mill

7KHFRPELQDWLRQRIDG\QDPLFFODVVLÀHUZKHHOZLWKDQDLUMHWPLOOSURGXFHVDIRUPRI
ÀQHJULQGLQJWKDWUHVXOWVLQDQH[DFWPD[LPXPSDUWLFOHVL]H,QFRQWUDVWWRWKHSXUHO\

110 Comminution Handbook | Spectrum Series 21

 CHAPTER 8 t STIRRED MILLS

FIG 8.4 – Classifier wheel showing the hub (bottom disk), the cover
plate (top disk) and the vanes (image courtesy Udo Enderle).

PHFKDQLFDO FODVVLÀHU PLOO WKH FUXVKLQJ HQHUJ\ LQVLGH WKH ÁXLGLVHG EHG MHW PLOO LV
VXSSOLHGE\JDVMHWV,QWKLVXQLWWKHSDUWLFOHVDUHDFFHOHUDWHGLQWKHMHWVWUHDPDQGFROOLGH
ZLWKRWKHUVORZHUSDUWLFOHV7KH\DUHFUXVKHGRQWKHZD\WRWKHFHQWUHDVZHOODVLQWKH
FHQWUHLWVHOI*ULQGLQJLVFRPSOHWHO\DXWRJHQLFVRWKHUHLVQRZHDURQWKHJULQGLQJWRROV
)LJXUHVKRZVWKH+RVRNDZDMHWPLOOZKLFKHPERGLHVDOORIWKHVHIHDWXUHV
:LWKWKHDLUULVLQJLQWKHFHQWUHWKHSDUWLFOHVDUHWUDQVSRUWHGWRWKHFODVVLÀHUZKHHO
ZKLFKGHÁHFWVWKHSDUWLFOHVWKDWDUHWRRFRDUVHDQGRQO\SDVVHVSDUWLFOHVWKDWPHHWWKH
VHWFRQGLWLRQVFODVVLÀHUVSHHGDLUÁRZUDWH7KHLQWHUQDOFODVVLÀFDWLRQUHVXOWVLQDYHU\
VLPSOH RSHUDWLRQ ZKLOH WKH PDFKLQH GHOLYHUV FRQVWDQW SURGXFW VL]H DQG YDULDWLRQV LQ
IHHGPDQLIHVWLQFKDQJLQJWKHWKURXJKSXWUDWH7KDWLVLIIHHGLVFRDUVHURUKDUGHUDQG
WKHFLUFXODWLQJORDGLQWKHPLOOLVLQFUHDVHGWKLVZLOOLPSDFWWKURXJKSXWRIWKHPDFKLQH

FIG 8.5 – Hosokawa Alpine Air jet mill and wheel classifier (image courtesy Hosokawa Alpine).

Comminution Handbook | Spectrum Series 21 111

 CHAPTER 8 t STIRRED MILLS

)RUWKHWUHDWPHQWRIGU\SURGXFWVWKHÁXLGLVHGEHGMHWPLOOFXUUHQWO\KDVWKHPRVWPRGHUQ
WHFKQRORJ\ IRU SURGXFLQJ WKH JUHDWHVW ÀQHQHVV ZLWK GHÀQHG PD[LPXP SDUWLFOH VL]H
OLPLWDWLRQ2SHUDWLQJFRPSOHWHO\ZLWKRXWPHFKDQLFDOJULQGLQJWRROVLWLVDOVRSRVVLEOH
WRSURFHVVHYHQWKHKDUGHVWSURGXFWVXSWR0RKV'HSHQGLQJRQWKHSURGXFWDQG
PDFKLQHVHWWLQJVSURGXFWVL]LQJUDQJLQJIURP—PWR—PFDQEHDFKLHYHG

Superheated steam jet mills

7KH QHHG IRU HYHQ ÀQHU SURGXFWV LQ WKH VXEPLFURQ WR QDQR UDQJH OHG WR D IXUWKHU
WHFKQRORJLFDO DGYDQFH WKDW PDNHV WUXH FRPPLQXWLRQ LQWR WKHVH UDQJHV SRVVLEOH XVLQJ
ÁXLGLVHGEHGMHWPLOOV7KH6LQ6-(7®VWDQGVIRUVWHDP&RQWUDU\WRFRQYHQWLRQDOGU\
JULQGLQJSURFHVVHVZLWKMHWPLOOVVXSHUKHDWHGVWHDPLVXVHGDVWKHJULQGLQJJDVLQWKH
6-(7®V\VWHP)ORZYHORFLWLHVXSWRPVFDQEHDFKLHYHGZLWKVWHDPDQGWKHNLQHWLF
LPSDFW HQHUJ\ RI WKH SURGXFW SDUWLFOHV ZLWKLQ WKH ÁXLGLVHG EHG LQFUHDVHG IRXUIROG
7KHVHKLJKHUÁRZYHORFLWLHVDOORZWKHSURGXFWLRQRIVLJQLÀFDQWO\ÀQHUSDUWLFOHV6LQFH
VWHDPKDVDFRQVLGHUDEO\KLJKHUDFRXVWLFYHORFLW\PVFRPSDUHGWRPVIRU
DLU WKH SRWHQWLDO SHULSKHUDO VSHHG RI WKH ÁRZ ZLWKLQ D FODVVLÀHU ZKHHO LQFUHDVHV DV
ZHOO HQKDQFLQJ WKH DFFHOHUDWLYH IRUFHV LQ SURGXFW FODVVLÀFDWLRQ :LWK WKLV DSSURDFK
WKHGHFLVLYHVWHSKDVEHHQWDNHQWRZDUGVDFWXDOO\FODVVLI\LQJSDUWLFOHVLQWKHVXEPLFURQ
UDQJHDQGWKXVIRUWKHÀUVWWLPHSUHSDULQJWKHPWKURXJKGU\JULQGLQJ

Grinding to nanometre size

7UXH FRPPLQXWLRQ H[WHQGV WR WKH SURGXFWLRQ RI SDUWLFOHV LQ WKH QDQRPHWUH UDQJH
7KHQDQRPHWUHQPLVWKHVPDOOHVWHQJLQHHULQJGLPHQVLRQPP —P—P 
QP'LDPHWHUVRIHYHU\GD\LWHPVLQQDQRPHWUHVDUHVKRZQLQ)LJXUH3DUWLFOHV
LQWKHQDQRPHWUHUDQJHZLOOEHFRPHDSDUWRIHYHU\GD\OLIHLQWKHIXWXUHEHFDXVHWKH\
SURPLVHDGYDQWDJHVLQPDQ\EUDQFKHVRILQGXVWU\IURPFKHPLFDOVDQGSKDUPDFHXWLFDOV
WRPLFURHOHFWURQLFVFHUDPLFVSDLQWVFRDWLQJVDQGFRVPHWLFV7KHHIIHFWWKDWWKH\ZLOO
KDYHRQWKHSURFHVVLQJRIPLQHUDOVFRDORUFHPHQWLVXQNQRZQEXWQRWWREHGLVFRXQWHG
3DUWLFOHVRIQDQRPHWUHVL]HDUHPDGHE\WZRPHWKRGV
 XVLQJWKHFKHPLFDOSURSHUWLHVRIVLQJOHPROHFXOHVFDXVLQJWKHPWRVHOIRUJDQLVHRU
VHOIDVVHPEOHLQWRVRPHXVHIXOFRQIRUPDWLRQWKHERWWRPXSDSSURDFK
 XVLQJ WUDGLWLRQDO PHWKRGV WR EUHDN PDWHULDOV GRZQ LQWR WKH GHVLUHG VL]H WKH WRS
GRZQDSSURDFK

FIG 8.6 – Nanometre scale where 1 mm = 1000 μm, 1 μm = 1000 nm (image courtesy Udo Enderle).

112 Comminution Handbook | Spectrum Series 21

 CHAPTER 8 t STIRRED MILLS

7KH ERWWRPXS DSSURDFK LV PRUH FRPPRQ EXW WKH WRSGRZQ DSSURDFK LV XVHG LI
V\QWKHVLV LV QRW SRVVLEOH RU WRR H[SHQVLYH 7KH +RVRNDZD $OSLQH $+0 PDFKLQH
SURGXFHVSDUWLFOHVGRZQWRQPXVLQJWKLVWRSGRZQDSSURDFK/RQJUHVLGHQFHWLPHV
DQGEHDGVLQWKHWRPPUDQJHDUHXVHGLQDKRUL]RQWDODJLWDWRUPLOODQGVXUIDFH
DFWLYHDJHQWVDUHDGGHGWRPLQLPLVHDJJORPHUDWLRQ

VERTICAL STIRRED MILLS

Tower mills
7KHÀUVWVWLUUHGPLOOVIRUJULQGLQJPLQHUDOVZHUHYHUWLFDOPLOOVZKLFKZHUHGHYHORSHG
LQ -DSDQ DERXW  7KHVH PLOOV ZHUH GHVFULEHG DV WRZHU PLOOV DQG LQ WKH RULJLQDO
WRZHUPLOOWKHPLQHUDOVOXUU\DQGVPDOOEDOOVZHUHVWLUUHGE\DYHUWLFDODJLWDWRUVFUHZ
7KH SDUWLFOHSDUWLFOH DQG SDUWLFOHEDOO FRQWDFWV WKDW RFFXUUHG FDXVHG EUHDNDJH E\
DEUDVLRQDQGDWWULWLRQSURGXFLQJYHU\VPDOOSDUWLFOHV)RU\HDUVWKHPDLQXVHRIWRZHU
PLOOVZDVWRJULQGOLPHVWRQHIRUQHXWUDOLVLQJDFLGLFSXOSVEXWLQ&RPLQFRLQ%ULWLVK
&ROXPELD&DQDGDVWDUWHGXVLQJWKHPWRJULQGÀQHJUDLQHGOHDG]LQFRUHVWRVL]HVWKDW
FRXOGQRWEHDFKLHYHGE\EDOOPLOOVWRLPSURYHPLQHUDOOLEHUDWLRQ7KLVZDVVXFFHVVIXO
DQGVWLUUHGPLOOVQRZGRPLQDWHÀQHJULQGLQJWHFKQRORJ\
7RZHUPLOOVZLWKPPEDOOVZHUHIRXQGWRJULQGPPSDUWLFOHVRIKDUGURFNPP
SDUWLFOHVRIVRIWURFNWROHVVWKDQ—P²ZKLFKZDVÀQHUWKDQSURGXFWVIURPFRQYHQWLRQDO
EDOOPLOOV²DQGHQHUJ\FRQVXPSWLRQZDVUHGXFHGE\²SHUFHQW7KHRULJLQDOPLOOV
DUH QRZ PDUNHWHG ZLWK WKH EUDQG QDPH 7RZHU0LOO E\ WKH (LULFK JURXS LQ *HUPDQ\
0LOO KHLJKWV LQFOXGLQJ WKH VWLUULQJ PHFKDQLVPV DUH  WR  P DOWKRXJK WKH JULQGLQJ
FKDPEHUVDUHDERXWKDOIWKLVKHLJKW0RWRUVUDQJHIURPWRN:6RPHGHWDLOVRI
WKHWRZHUPLOODUHVKRZQLQ)LJXUH

A B C

FIG 8.7 – Eirich TowerMill (image courtesy Eirich). (A) TowerMill, (B) material flow, (C) components.

Vertimill®
9HUWLPLOOV DUH PDGH E\ 0HWVR DQG RSHUDWH RQ WKH VDPH SULQFLSOH DV WKH RULJLQDO WRZHU
PLOO7KHVFUHZLQHDFK9HUWLPLOOVWLUVWKHJULQGLQJEDOOVDQGVOXUU\DWDFRQVWDQWVSHHG
9HUWLPLOOVDUHPDGHLQPDQ\VL]HVZLWKPRWRUVIURPWRN:0LOOKHLJKWVLQFOXGLQJ

Comminution Handbook | Spectrum Series 21 113

 CHAPTER 8 t STIRRED MILLS

WKHGULYHFRPSRQHQWVDUHIURPWRPRIZKLFKDSSUR[LPDWHO\²SHUFHQWLV
WKH JULQGLQJ FKDPEHU GHSHQGLQJ RQ WKH PRGHO 0LOO GLDPHWHUV DUH IURP DERXW  WR
P7KH9HUWLPLOOLVHTXLSSHGZLWK2UHEHGPDJQHWLFOLQHUVZKLFKW\SLFDOO\ODVWIRUWKH
OLIHWLPHRIWKHPLOODVWKHJULQGLQJPHGLDZLOOIRUPDFRQWLQXRXVVHOIUHQHZLQJZHDUOD\HU
SURWHFWLQJWKHOLQHUVDQGWKHPLOOVKHOO)LJXUHVKRZVD9HUWLPLOODQGWKHPL[LQJDXJHU
7KHRQO\UHJXODUZHDULWHPVRQD9HUWLPLOOLQDGGLWLRQWRWKHJULQGLQJPHGLDDUHWKH
OLQHUV ZKLFK PXVW EH UHSODFHG ZKHQ ZHDU LV VHYHUH 7KHVH DUH ODUJH FXUYHG OLQHUV
DQGGHYLFHVKDYHEHHQGHYHORSHGE\5XVVHOO0LQHUDO(TXLSPHQW50(WRPHFKDQLVH
UHOLQLQJZKHUHWKHZHLJKWRIWKHOLQHUKDVFDSDFLWLHVRIDQGNJ
0LOOOLQHUVUHOLQLQJKDQGOHUVDQGUHOLQLQJPDFKLQHVZLOOEHGLVFXVVHGLQ&KDSWHU

FIG 8.8 – Vertimill showing mill component (left) and mixing auger (right) (images courtesy Metso).

HIGmill
7KH+,*PLOOLVSURGXFHGE\2XWRWHF7KHVHPLOOVKDYHEHHQXVHGVXFFHVVIXOO\IRU\HDUV
LQ ÀQHJULQGLQJ DSSOLFDWLRQV IRU OLPHVWRQH DQG LQGXVWULDO PLQHUDOV 7KH\ FRPSULVH
DYHUWLFDOPLOOERG\DVKDIWZLWKJULQGLQJGLVFVVKHOOPRXQWHGFRXQWHUULQJVDQGD
JHDUER[DQGGULYH8SWRSHUFHQWRIWKHJULQGLQJFKDPEHULVÀOOHGZLWKJULQGLQJ
EHDGV5RWDWLQJGLVFVVWLUWKHFKDUJHDQGXSWRGLVFVFDQEHXVHGLQWKHFKDPEHU
GHSHQGLQJRQWKHDSSOLFDWLRQ)HHGVOXUU\LVSXPSHGLQWRWKHPLOODWWKHERWWRPDQG
JULQGLQJRFFXUVE\DWWULWLRQDVLWPRYHVXSWKURXJKFRQVHFXWLYHJULQGLQJVWDJHV7KH
SURGXFW LV GLVFKDUJHG DW WKH WRS RI WKH PDFKLQH ZKHUH LW LV FODVVLÀHG +,*PLOOV DUH
XVHGWRPDNHSURGXFWVRIOHVVWKDQ—P*ULQGLQJEHDGVDUHFHUDPLFRUVWHHODQG
EHDGVL]HLVLQWKHUDQJHRI²PPGHSHQGLQJRQFDQGP&KDPEHUYROXPHVDUH
IURP/WR/ZLWKGULYHVIURPN:WR0:)LJXUHSUHVHQWVGHWDLOV
RI WKH +,*PLOO LQFOXGLQJ W\SLFDO SDUWLFOH VL]H GLVWULEXWLRQV 36'V 1LQH ð  P:
PLOOVDUHFXUUHQWO\RSHUDWLQJLQWKHFDOFLXPFDUERQDWHLQGXVWU\ZKHUHWKHVKDUS36'
WKH\SURGXFHLVQHFHVVDU\7KUHHðP:PLOOVKDYHEHHQVROGWRWKHKDUGURFNPLQLQJ
LQGXVWU\IRUDW\SLFDOFRSSHUUHJULQGSURMHFWLQWKH$PHULFDVZKHUHWKHVKDUS36'LV
DQDGYDQWDJH

114 Comminution Handbook | Spectrum Series 21

 CHAPTER 8 t STIRRED MILLS

A B C

,/'ŵŝůůW^
ϭϬϬ
&ĞĞĚ
ϴϬ WƌŽĚƵĐƚ

WĂƐƐŝŶŐ;йͿ
ϲϬ

ϰϬ

ϮϬ

Ϭ
Ϭ͘ϭ ϭ ϭϬ ϭϬϬ ϭϬϬϬ
WĂƌƚŝĐůĞ^ŝnjĞ;ŵŝĐƌŽŶͿ

FIG 8.9 – Outotec HIGmill (images courtesy Outotec). (A) External view, (B) grinding discs, (C) typical feed and product sizings.

7KH  0: PDFKLQHV DUH  P KLJK DQG  P LQ GLDPHWHU %HFDXVH RI WKH WDOO DQG
QDUURZYHUWLFDOPLOOERG\DUUDQJHPHQWJULQGLQJPHGLDLVHYHQO\GLVWULEXWHGDQGPLQHUDO
SDUWLFOHVUHPDLQLQFRQVWDQWFRQWDFWVLJQLÀFDQWO\LQFUHDVLQJJULQGLQJHIÀFLHQF\7KHLU
PDLQXVHLVIRUJULQGLQJOLPHVWRQHDQGRWKHUVRIWPLQHUDOVFRQWDLQLQJXSWRSHUFHQW
6L2WRDSURGXFWVL]HRI—P7KH\KDYHUHFHQWO\VLQFHEHHQXVHGIRUFRDUVHU
JULQGV DQG IRU UHJULQGLQJ RUH FRQFHQWUDWHV WR LPSURYH OLEHUDWLRQ DQG LQFUHDVH ZDVWH
UHMHFWLRQ 7KH FDVLQJ LV ÁDQJHG YHUWLFDOO\ VR LW FDQ EH VSOLW GRZQ WKH FHQWUH LQWR WZR
KDOYHVWKDWFDQEHPRYHGDSDUWRQDUDLOLQJV\VWHP&KDQJLQJGLVFVDQGOLQHUVHJPHQWV
DVWKH\ZHDUFDQEHGRQHE\WZRVNLOOHGWUDGHSHUVRQQHO7KHW\SLFDOSURGXFWVL]LQJV
EDVHGRQIHHGVL]LQJIRUWKHVHPLOOVLVSUHVHQWHGLQ7DEOH

TABLE 8.1
Feed and product sizings for Outotec’s HIGmill.
Grind classification Feed (F80 μm) Product (P80 μm)
Coarse 100–300 50–100
Fine 50–100 20–60
Ultra-fine <70 <20

Energy-intensive stirred agitator (EiSA) mill

(L6$LVDYHUWLFDOVWLUUHGPLOOPDGHE\WKH1HW]VFK*URXSVHH)LJXUHVRPHWLPHV
VXSSOLHG DV DQ DOWHUQDWLYH WR WKH KRUL]RQWDO ,VD0LOO LQ VRIWHU ORZHQHUJ\ QRQPLQLQJ
DSSOLFDWLRQV VLPLODU WR +,*PLOO DSSOLFDWLRQV ,Q WKHVH DSSOLFDWLRQV WKH\ FDQ SURGXFH
SDUWLFOHVLQWKH²—PUDQJH+HUHDUHVRPHRIWKHIHDWXUHVRIWKH(L6$PLOO
• DQ RSHQ RXWOHW VR QR VHDO V\VWHP LV UHTXLUHG
• WKH LQWHUQDO FODVVLÀHU PHDQV QR H[WHUQDO FODVVLÀFDWLRQ LV QHFHVVDU\
• PLOOLQJ LV XVXDOO\ RSHQ FLUFXLW
• JULQGLQJ PHGLD EHWZHHQ  DQG  PP LV XVHG IRU IHHG EHORZ  —P WR PDNH SURGXFWV
EHWZHHQ  DQG  —P IRU VRIWHU QRQPLQLQJ IHHGV PRUH FRPPRQO\ VPDOOHU PHGLD

Comminution Handbook | Spectrum Series 21 115

 CHAPTER 8 t STIRRED MILLS

FIG 8.10 – Netzsch EiSA mill (images courtesy Netzsch).

ZLWK EHDG VL]H ² PP FDQ EH XVHG IRU YHU\ ÀQH SURGXFW DQG VRIW DSSOLFDWLRQV
VXFK DV OLPHVWRQH ZLWK SURGXFW d   —P
• LQHUW FHUDPLF PHGLD FDQ EH XVHG
7KH YHUWLFDO (L6$ PLOO LV SULFHG EHORZ WKH KRUL]RQWDO ,VD0LOO EHFDXVH LW GRHV QRW
UHTXLUH D VHDOLQJ V\VWHP ,WV GLVDGYDQWDJHV LQ FRPPRQ ZLWK RWKHU YHUWLFDO PLOOV DUH
SRRU DFFHVV IRU PDLQWHQDQFH DQG D KLJKVWDUWLQJ WRUTXH UHTXLUHPHQW /LNH RWKHU KLJK
VSHHG VWLUUHG PLOOV ZLWK WKH UHFHQW H[FHSWLRQ RI WKH ,VD0LOO WKLV OLPLWV LWV SUDFWLFDO
DSSOLFDWLRQ RQ KDUG PLQHUDOV WR XOWUDÀQH JULQGLQJ DSSOLFDWLRQV SURGXFW VL]H EHORZ
 —P DQG VRIWIHHG PLQHUDOV 7DEOH  SUHVHQWV WKH DYDLODEOH VL]H UDQJH DQG FDSDFLWLHV
RI WKH (L6$ PLOO

TABLE 8.2
Netzsch EiSA mill.
Chamber volume (L) Power (kW) Throughput (t/h solids) Throughput (t/h water equivalent)
0–3150 90–1000 3000–35 000 3–35

Stirred media detritor

6RPH ÀQHJUDLQHG VXOÀGH RUHV UHTXLUH ÀQHU JULQGLQJ WKDQ LV DYDLODEOH ZLWK D VFUHZ
DJLWDWRU RI WKH W\SH XVHG LQ WKH 9HUWLPLOO DQG WKH 7RZHU0LOO 7KHUH ZDV D VLPLODU QHHG
IRU WKH ÀQHU JULQGLQJ RI NDROLQ DQG FDOFLXP FDUERQDWH IRU SDSHU FRDWLQJV VR (QJOLVK
&KLQD &OD\ (&& ,QWHUQDWLRQDO GHYHORSHG D KLJKVSHHG DJLWDWRU PLOO XVLQJ VDQG DV WKH
JULQGLQJ PHGLD IRU WKLV SXUSRVH %\  WKLV PLOO ZDV EHLQJ XVHG VXFFHVVIXOO\ ZLWK
FRPSOH[ VXOÀGH RUHV ,W ZDV IRXQG WKDW JULQGLQJ  —P RUH WR  —P UHTXLUHG  WR  PP
VDQG DQG JULQGLQJ LW WR  —P UHTXLUHG  WR  PP VDQG
8WLOLVLQJ HLWKHU VDQG RU FHUDPLF PHGLD 0HWVR·V VWLUUHG PHGLD GHWULWRU 60' LV DQ
H[WHQVLRQ RI WKH (&& PLOO UHIHU WR )LJXUH  7KH 60' LV D ÁXLGLVHG YHUWLFDO VWLUUHG
PLOO WKDW XWLOLVHV WKH URWDWLRQDO HQHUJ\ RI WKH LPSHOORU DUPV WR LPSDUW D KLJKHQHUJ\
PRWLRQ WR WKH PHGLDVOXUU\ PL[WXUH LQVLGH WKH PLOO 7KH VWLUUHU VSHHG LV KLJK HQRXJK WR
GLVWULEXWH WKH PHGLD WKURXJKRXW WKH VOXUU\ UHJDUGOHVV RI PHGLD GHQVLW\ IRUFLQJ SDUWLFOH
DQG PHGLD FRQWDFW 7KH PLOO·V SRZHU LQWHQVLW\ LV RSWLPLVHG WR DFKLHYH HIÀFLHQW JULQGLQJ

116 Comminution Handbook | Spectrum Series 21

 CHAPTER 8 t STIRRED MILLS

FIG 8.11 – Stirred media detritor (images courtesy Metso).

OLPLW ZHDU DQG DOORZ IRU KHDW GLVVLSDWLRQ LQ WKH FDVH RI D KLJKHQHUJ\ JULQG 7KH SRZHU
LQWHQVLW\ N:P LV UHODWLYHO\ KLJK FRPSDUHG WR RWKHU PLOO W\SHV EXW LV UHTXLUHG WR
JHQHUDWH D YRUWH[ RI PHGLD DQG VOXUU\ GXULQJ RSHUDWLRQ DQG EULQJ WKH SDUWLFOHV LQ FRQWDFW
ZLWK RQH DQRWKHU IRU HIÀFLHQW JULQGLQJ KRZHYHU WKH SRZHU LQWHQVLW\ LV QRW VR KLJK WKDW
D FRROLQJ V\VWHP ZRXOG EH UHTXLUHG WR GLVVLSDWH WKH KHDW JHQHUDWHG GXULQJ D KLJK HQHUJ\
JULQG $GGLWLRQDOO\ D ORZHU SRZHU LQWHQVLW\ OLPLWV WKH VKHHU IRUFH RI WKH PHGLDVOXUU\
RQ WKH OLQHUV DQG LPSHOORUV LPSURYLQJ ZHDU OLIH
6WLUUHG PLOOV KDYH EHHQ SURYHQ WR SURYLGH HQHUJ\ VDYLQJV FRPSDUHG WR WUDGLWLRQDO EDOO
PLOOV 7KH ÀQHU WKH SURGXFW UHTXLUHG WKH PRUH HIÀFLHQW VWLUUHG PLOOV DUH WR D EDOO PLOO
7KH DWWULWLRQ JULQGLQJ DFWLRQ YHUWLFDO DUUDQJHPHQW DQG ÀQHU PHGLD VL]H GLVWULEXWLRQ
FRQWULEXWH WR PDNLQJ VWLUUHG PLOOV PRUH HQHUJ\HIÀFLHQW JULQGLQJ PDFKLQHV 7\SLFDOO\
9HUWLPLOO HQHUJ\ VDYLQJV UDQJH IURP  SHU FHQW WR JUHDWHU WKDQ  SHU FHQW FRPSDUHG
ZLWK WUDGLWLRQDO EDOO PLOOV DQG WKH 60' HQHUJ\ VDYLQJV DUH IDU JUHDWHU WKDQ  SHU FHQW
RI EDOO PLOO HQHUJ\ IRU ÀQH JULQGV
7KH WUDQVLWLRQ IURP WRZHU PLOOV WR 60'V W\SLFDOO\ RFFXUV DW SURGXFW VL]HV EHWZHHQ
 DQG  —P 7DEOH  VXPPDULVHV WKH FKDUDFWHULVWLFV RI WKH YDULRXV YHUWLFDO VWLUUHG
PLOOV DYDLODEOH LQ WKH PDUNHWSODFH

TABLE 8.3
Vertical stirred mills.
Manufacturer Motor power (kW) Feed size (mm) Product size (<μm)
TowerMill (Nissan Ehrich) 0.4–1120 10 15
Vertimill (Metso) 11–2237 6 15
Stirred Media Detritor (Metso) 185, 355, 1100 200 5
HIGmill (Outotec) 144–5500 500 20
EiSA mill (Netzsch) 90–710 250 10

Comminution Handbook | Spectrum Series 21 117

 CHAPTER 8 t STIRRED MILLS

HORIZONTAL STIRRED MILLS

7KH PDLQ GLIIHUHQFHV EHWZHHQ YHUWLFDO DQG KRUL]RQWDO VWLUUHG PLOOV DUH WKDW KRUL]RQWDO
PLOOV FDQ RSHUDWH DW KLJKHU PHGLD ÀOOLQJV XS WR  SHU FHQW KLJKHU DJLWDWRU WLS VSHHGV
² PV DQG VPDOOHU PHGLD VL]H  PP WKDQ YHUWLFDO PLOOV 7KH\ DUH HDV\ WR LQVWDOO DQG
KDYH JUHDWHU HQHUJ\ HIÀFLHQF\ WKDQ YHUWLFDO PLOOV 7KH KRUL]RQWDO FRQÀJXUDWLRQ PHDQV
WKDW DOO VWLUUHUV DUH DYDLODEOH WR PRELOLVH PHGLD RQ VWDUWXS 7KLV DOORZV D VLJQLÀFDQW
UHGXFWLRQ LQ WKH VWDUWXS WRUTXH UHTXLUHG FRPSDUHG WR YHUWLFDO PLOOV DQG PHDQV WKDW
KRUL]RQWDO PLOOV FDQ RSHUDWH DW KLJKHU HQHUJ\ GHQVLWLHV LQ WKH PLOOLQJ FKDPEHU DQG
HPSOR\ ODUJHU PRWRUV $V D UHVXOW KRUL]RQWDO VWLUUHG PLOOV FDQ SURGXFH ERWK FRDUVH DQG
ÀQH JULQG VL]HV EDVHG RQ WKH PHGLD VL]H DQG W\SH WKDW LV XVHG
'XH WR WKH KRUL]RQWDO FRQÀJXUDWLRQ WKH PLOOV UHTXLUH D VKDIW VHDO KRZHYHU EHFDXVH RI
WKH VHDOHG GHVLJQ WKH PLOOV FDQ HDVLO\ KDQGOH ÁRZ GHQVLW\ DQG YLVFRVLW\ ÁXFWXDWLRQV RI
WKH PLOO IHHG

IsaMill™
7KH RQO\ KRUL]RQWDO VWLUUHG JULQGLQJ PLOO XVHG H[WHQVLYHO\ LQ WKH PLQHUDOV LQGXVWU\ QRZ
LV WKH ,VD0LOO ,W RULJLQDWHG ZLWK WKH QHHG WR OLEHUDWH WKH VXOÀGHV LQ WKH YHU\ ÀQHJUDLQHG
OHDG]LQF RUHV WKDW 0RXQW ,VD 0LQHV 0,0 ZDV PLQLQJ DW 0RXQW ,VD DQG 0F$UWKXU
5LYHU 1RUWKHUQ 7HUULWRU\ 7XPEOLQJ PLOOV IDLOHG WR JULQG ÀQHO\ HQRXJK DQG \LHOGHG
SRRU ÁRWDWLRQ FRQFHQWUDWHV 0,0 DQG 1HW]VFK &RUSRUDWLRQ DGDSWHG WKH KLJKVSHHG
EDWFK VWLUUHG PLOOV ² XVHG WR JULQG SLJPHQWV ² WR FRQWLQXRXV PLOOV VXLWDEOH IRU KDUG
RUHV DQG QDPHG WKH QHZ PLOO WKH ,VD0LOO $W 0F$UWKXU 5LYHU WKH LQVWDOODWLRQ KDG IRXU
0V  0: HDFK WR JULQG EHORZ  —P (QGHUOH et al  6LQFH LWV GHYHORSPHQW
DW 0RXQW ,VD LQ  LW KDV EHHQ ZLGHO\ DGRSWHG IRU D UDQJH RI PLQHUDO JULQGLQJ
DSSOLFDWLRQV ZRUOGZLGH $IWHU LWV RULJLQDO GHYHORSPHQW IRU XOWUDÀQH JULQGLQJ LW ZDV
IXUWKHU GHYHORSHG LQ WHUPV RI VFDOH DQG WKH DELOLW\ WR SURGXFH FRDUVHU JULQGV 6LQFH WKH
HDUO\ V LW KDV EHHQ DSSOLHG LQ FRQYHQWLRQDO JULQGLQJ DQG UHJULQGLQJ DSSOLFDWLRQV
ZLWK SURGXFW VL]HV DERYH  —P ZKLFK ZHUH QRW SUHYLRXVO\ FRQVLGHUHG WKH GRPDLQ RI
KLJKVSHHG VWLUUHG PLOOV
7KH ,VD0LOO LV D KLJKLQWHQVLW\ VWLUUHG PLOO XVLQJ VPDOO PHGLD VWLUUHG DW JUHDW VSHHG WR
FRPELQH KLJK HQHUJ\ HIÀFLHQF\ ZLWK KLJK EUHDNDJH UDWHV 7KH GLVWLQJXLVKLQJ IHDWXUHV
RI WKH ,VD0LOO DUH WKH KRUL]RQWDO FRQÀJXUDWLRQ WKH KLJKSRZHU LQWHQVLW\ WKH LQWHUQDO
FODVVLÀHU WKDW DFKLHYHV D VKDUS SURGXFW VL]H GLVWULEXWLRQ DQG WKH XVH RI LQHUW PHGLD
$V RI   SHU FHQW RI LQVWDOOHG ,VD0LOOV DUH WDUJHWLQJ SURGXFW JULQG VL]HV RI P
EHWZHHQ  DQG  PLFURQV PRVWO\ RQ KDUG PLQHUDOV VXFK DV VSKDOHULWH S\ULWH DQG
JDOHQD

Operating principles
7KH ,VD0LOO VKRZQ LQ )LJXUH  LV D KRUL]RQWDOO\ VWLUUHG PLOO FRQVLVWLQJ RI D VHULHV
RI VHYHQ WR HLJKW URWDWLQJ JULQGLQJ GLVFV PRXQWHG RQ D FDQWLOHYHUHG VKDIW LW LV GULYHQ
WKURXJK D PRWRU DQG JHDUER[ 7KH GLVFV RSHUDWH DW WLS VSHHGV RI ² PV UHVXOWLQJ LQ
HQHUJ\ LQWHQVLWLHV RI XS WR  N:P
)LJXUH  VKRZV WKH SULQFLSOH RI RSHUDWLRQ 7KH VKHOO LV VWDWLRQDU\ DQG WKH PLOO LV ÀOOHG
ZLWK D VXLWDEOH JULQGLQJ PHGLD 7KH DUHD EHWZHHQ HDFK GLVF LV HVVHQWLDOO\ DQ LQGLYLGXDO
JULQGLQJ FKDPEHU VR DV D UHVXOW WKH PLOO LV HIIHFWLYHO\ HLJKW JULQGLQJ FKDPEHUV LQ VHULHV
7KH PHGLD LV VHW LQ PRWLRQ E\ WKH DFWLRQ RI WKH JULQGLQJ GLVFV ZKLFK UDGLDOO\ DFFHOHUDWH

118 Comminution Handbook | Spectrum Series 21

 CHAPTER 8 t STIRRED MILLS

FIG 8.12 – IsaMill™ layout (image courtesy Glencore).

FIG 8.13 – How the IsaMill operates (image courtesy Glencore).

WKH PHGLD WRZDUGV WKH VKHOO %HWZHHQ WKH GLVFV ZKHUH WKH PHGLD LV QRW VXEMHFW WR WKH
KLJK RXWZDUGV DFFHOHUDWLRQ RI WKH GLVF IDFH WKH PHGLD LV IRUFHG EDFN LQ WRZDUGV WKH
VKDIW FUHDWLQJ D FLUFXODWLRQ RI PHGLD EHWZHHQ HDFK VHW RI GLVFV 7KH JULQGLQJ PHFKDQLVP
ZLWKLQ WKH ,VD0LOO LV RQH RI DWWULWLRQ DQG DEUDVLRQ ZKHUH SDUWLFOH VXUIDFHV DUH FKLSSHG
UHSHDWHGO\ 7KLV EUHDNDJH PRGH SURGXFHV YHU\ ÀQHVL]HG SDUWLFOHV DW D UHODWLYHO\ ORZ
SRZHU FRQVXPSWLRQ GXH WR WKH KLJK SUREDELOLW\ RI PHGLD²SDUWLFOH FROOLVLRQ
7KH HLJKW FKDPEHUV LQ VHULHV HQVXUH WKDW VKRUW FLUFXLWLQJ RI PLOO IHHG WR WKH GLVFKDUJH
LV YLUWXDOO\ LPSRVVLEOH 7KHUH LV D YHU\ KLJK SUREDELOLW\ RI PHGLD²SDUWLFOH FROOLVLRQ DV D
UHVXOW RI WKH KLJKHQHUJ\ LQWHQVLW\ DQG WKH HLJKW FKDPEHUV LQ VHULHV 7KH UHVLGHQFH WLPH
LQ WKH ,VD0LOO LV TXLWH VKRUW FRPSDUHG ZLWK RWKHU JULQGLQJ PLOOV ² W\SLFDOO\ DURXQG  WR
 VHFRQGV 7KLV FDQ KDYH DGYDQWDJHV LQ GRZQVWUHDP SURFHVVLQJ ZKHUH FOHDQ PLQHUDO
VXUIDFHV DUH SURGXFHG DQG TXLFNO\ SDVVHG WR WKH QH[W VWDJH RI SUHSDUDWLRQ

Comminution Handbook | Spectrum Series 21 119

 CHAPTER 8 t STIRRED MILLS

Energy intensity
7KH KLJK WLS VSHHG RI WKH ,VD0LOO UHVXOWV LQ D KLJKHQHUJ\LQWHQVLYH HQYLURQPHQW
(QHUJ\ LQWHQVLW\ RI WKH ,VD0LOO LV VLJQLÀFDQWO\ KLJKHU WKDQ DQ\ RWKHU FRPPHUFLDOO\
DYDLODEOH JULQGLQJ HTXLSPHQW DV VKRZQ LQ 7DEOH  &RPELQLQJ WKH HQHUJ\ LQWHQVLW\
DQG WKH KLJKJULQGLQJ HIÀFLHQF\ OHDGV WR D FRPSDFW PLOO WKDW FDQ ÀW LQWR H[LVWLQJ
SODQWV ZKHUH ÁRRU VSDFH DQG FUDQDJH LV OLPLWHG 0LOOV FDQ EH GLVWULEXWHG WKURXJKRXW
WKH SURFHVVLQJ FLUFXLW DW WKH SRLQWV ZKHUH PLQHUDO OLEHUDWLRQ LV QHHGHG UDWKHU WKDQ
FHQWUDOLVHG LQ WKH JULQGLQJ FLUFXLW

Energy efficiency
7KH HQHUJ\ HIÀFLHQF\ RI WKH ,VD0LOO LV GXH WR LWV DELOLW\ WR XVH VPDOO PHGLD ZKLFK UHVXOWV
IURP VWLUULQJ PHGLD DW D KLJKHU VSHHG 6PDOOHU PHGLD JLYHV KLJKHU JULQGLQJ HIÀFLHQF\
GXH WR WKH KLJKHU VXUIDFH DUHD DYDLODEOH IRU JULQGLQJ 7DEOH  +RZHYHU WKHUH DUH
WKUHH SUDFWLFDO OLPLWV WR XVLQJ VPDOO PHGLD WKH KLJK FRVW WKH DELOLW\ WR EUHDN WKH IHHG DW
DQ DGHTXDWH UDWH DQG WKH DELOLW\ WR UHWDLQ WKH PHGLD ZKLOH GLVFKDUJLQJ WKH SURGXFW
,VD0LOO DGGUHVVHV WKHVH WKUHH FRQVWUDLQWV E\ HPSOR\LQJ FHUDPLF PHGLD KLJKVSHHG
VWLUULQJ WR JHW KLJK EUHDNDJH UDWHV DQG WKH LQWHUQDO FHQWULIXJDO VHSDUDWRU
7KH VHSDUDWRU LV DW WKH GLVFKDUJH HQG RI WKH ,VD0LOO *URXQG VOXUU\ H[LWV WKH FHQWUH RI
WKH VHSDUDWRU DQG PHGLD LV FHQWULIXJHG WR WKH VKHOO DQG SXPSHG EDFN ZLWK VRPH RI WKH
VOXUU\ WR WKH IHHG HQG RI WKH PLOO 7KLV DFWLRQ UHWDLQV PHGLD LQ WKH PLOO ZLWKRXW WKH QHHG
IRU ÀQH VFUHHQV 7DEOH  SUHVHQWV WKH FKDUDFWHULVWLFV RI WKH UDQJH RI ,VD0LOOV

HOSOKAWA ALPINE FINE-GRINDING MILLS

7KH $OSLQH &RPSDQ\ VWDUWHG PDNLQJ ÀQHJULQGLQJ PLOOV LQ  DQG KDV PDLQWDLQHG
D VWURQJ UHVHDUFK DQG GHYHORSPHQW SURJUDP LQ ÀQHJULQGLQJ WHFKQRORJ\ VLQFH WKHQ
7KH FRPSDQ\ SKLORVRSK\ LV WKDW

TABLE 8.4
Comparative energy intensity of grinding technologies.
Mill type Installed power (kW) Mill volume (m³) Power intensity (kW/m³)
Autogenous mill 6400 353 18
Primary ball mill 2600 126 21
Regrind ball mill 740 39 19
Tower mill 1000 12 42
HIGmill 4378 27 160
IsaMill™ 3000 10 300

TABLE 8.5
Comparison of media size for different grinding methods.
Mill type Media size (mm) Number of media/m³ Media surface area (m2/m3)
Ball 20 177 000 222
Tower 12 818 000 370
IsaMill 2 176 500 000 2200

120 Comminution Handbook | Spectrum Series 21

 CHAPTER 8 t STIRRED MILLS

TABLE 8.6
Horizontal stirred mill – IsaMill.
Volume Power Capacity Feed size range Product size Power intensity
(L) (kW) (t/h) (mm) range (μm) (kW/m3)
100 75 8 <1 5–50 750
500 200 25 <1 5–50 400
1000 500 50 <1 5–50 500
5000 1500 150 <1 5–80 300
10 000 3000 300 <1 5–100 300
50 000 8000 800 <1 30–150 160

There is no such thing as a universal mill that optimally meets every requirement in terms
oI Àneness throughput energy eIÀciency Zear contaminationIree grinding and cleaning
etc For this reason $lpine oIIers a Zide spectrum oI diIIerent designs Zhich enaEles an
optimal solution to Ee Iound Ior each proElem speciÀcation +RVRNDZD $OSLQH 
$OSLQH ODWHU +RVRNDZD $OSLQH KDV SURGXFHG WHQV RI WKRXVDQGV RI ÀQHJULQGLQJ PLOOV
DQG URWDU\ JUDQXODWRUV IRU SODVWLF UHF\FOLQJ IRU PRUH WKDQ  \HDUV 'XULQJ WKH ODVW IHZ
GHFDGHV WKH FRPSDQ\ KDV VSHFLDOLVHG LQ ÀQHJULQGLQJ DSSOLFDWLRQV VPDOOHU WKDQ  —P
+RVRNDZD $OSLQH SURGXFHV PLOOV DQG FODVVLÀHUV IRU WKH SURGXFWLRQ RI PLQHUDO ÀOOHUV
:KLOH  \HDUV DJR D ÀOOHU ZLWK D VL]H  SHU FHQW SDVVLQJ  —P ZDV UHJDUGHG DV YHU\
ÀQH WRGD\ D VL]H RI  —P LV D VWDQGDUG UHTXLUHPHQW DQG VRPH SURGXFWV HYHQ KDYH D VL]H
RI  —P $V WKHVH ÀQH ÀOOHUV DUH JHQHUDOO\ XVHG DV GU\ SRZGHUV LW LV HVVHQWLDO WR XVH D GU\
ÀQHJULQGLQJ SURFHVV $ ZLGH YDULHW\ RI GLIIHUHQW DLU FODVVLÀHUV DUH DYDLODEOH WR DFKLHYH
WKHVH ÀQH VL]HV LQ GU\ JULQGLQJ SURFHVVHV
7KH PLOOV DQG FODVVLÀHUV DUH GLVFXVVHG LQ WKH +andEooN oI PoZder and Particle Processing
+RVRNDZD $OSLQH  7KUHH PLOOV WKDW DUH SDUWLFXODUO\ UHOHYDQW WR KLJKFDSDFLW\
PLQHUDO SODQWV ZLOO EH GLVFXVVHG LQ WKLV VHFWLRQ

Horizontal stirred wet AHM mill

7KH +RVRNDZD $OSLQH $+0 PLOO LV DQ DJLWDWHG EDOO PLOO ZLWK DQ HQFORVHG GRXEOHZDOOHG
PLOOLQJ FKDPEHU WKDW LV ERWK KRUL]RQWDO DQG F\OLQGULFDO )LJXUH  ,W LV XVHG IRU ERWK
FORVHGFLUFXLW DQG VLQJOHSDVV ZHW PLOOLQJ RI VXEPLFURQ WR QDQRPHWUH SDUWLFOHV RI KLJK
YDOXH ORZ WRQQDJH DSSOLFDWLRQV 3URGXFW LV IHG WR WKH PLOO LQOHW E\ PHDQV RI D SXPS DQG
H[LWV WKURXJK D VFUHHQ DW WKH PLOO RXWOHW 7KH JULQGLQJ FKDPEHU LV ÀOOHG WR DSSUR[LPDWHO\
 WR  SHU FHQW ZLWK WKH JULQGLQJ PHGLD 7KH VFUHHQ UHWDLQV WKH PHGLD LQ WKH FKDPEHU
EXW DOORZV WKH PLOOHG VOXUU\ WR H[LW 5RWDWLQJ VWLUUHUV SUHYHQW EORFNDJH RI WKH VFUHHQ
0DFKLQH VL]H UDQJHV IURP ODERUDWRU\SLORW SODQW WR ODUJHVFDOH SURGXFWLRQ XQLWV ZLWK
PRWRU VL]HV XS WR  N: 0LOOLQJ EHDGV DUH DYDLODEOH IURP  WR  PP GLDPHWHU LQ
DQ\ PDWHULDO RI FKRLFH
7DEOH  SUHVHQWV VRPH GHWDLOV RI WKH DYDLODEOH PLOO VL]HV

Hosokawa Alpine wet vertical stirred ANR mill

7KH $15 PLOO LV GHVLJQHG IRU XOWUDÀQH JULQGLQJ RI KLJKO\ FRQFHQWUDWHG OLPHVWRQH
VOXUULHV )LQHO\ JURXQG OLPHVWRQH LV XVHG LQ PDQ\ LQGXVWULHV XVXDOO\ DV D ÀOOHU DQG LV
W\SLFDOO\ JURXQG WR D  —P 7KH HFRQRPLF OLPLW RI GU\ JULQGLQJ LV DSSUR[LPDWHO\

Comminution Handbook | Spectrum Series 21 121

 CHAPTER 8 t STIRRED MILLS

FIG 8.14 – Horizontal stirred wet AHM mill and stirrer (images courtesy Hosokawa Alpine).

TABLE 8.7
Horizontal stirred Hosokawa Alpine AHM ball mill.
Chamber volume (L) Power (kW) Throughput (t/h)
5 7.5 0.01–0.05
17 22 0.04–0.2
65 45 0.08–0.4
132 75 0.15–0.8
221 110 0.2–1
516 200 0.4–2
1056 316 0.6–3

 SHU FHQW SDVVLQJ  —P DQG LV FRDUVHU IRU ZHW JULQGLQJ 7KH RSHUDWLQJ PRGH DQG
GHVLJQ RI WKH $15 PLOO SHUPLW WKH SURGXFWLRQ RI KLJKTXDOLW\ OLPHVWRQH ZLWK UHJDUG WR
ÀQHQHVV GHJUHH RI ZKLWHQHVV DQG DEUDVLYHQHVV ,Q JHQHUDO WKH VOXUU\ LV GLUHFWO\ XVHG LQ
SDSHU SURGXFWLRQ
7KH ÁRZ SDWWHUQ WKURXJK WKH YHUWLFDOO\ FRQÀJXUHG $15 LV IURP ERWWRP WR WRS ZKHUHE\
WKH IHHG PDWHULDO LV JURXQG WR WKH UHTXLUHG ÀQH VL]LQJ ZLWK WKH DLG RI FHUDPLF JULQGLQJ
EHDGV ]LUFRQLXP VLOLFDWH RU ]LUFRQLXP R[LGH DV VKRZQ LQ )LJXUH  7KH JULQGLQJ
EHDGV DUH VHSDUDWHG IURP WKH GLVFKDUJLQJ VXVSHQVLRQ DW WKH PLOO KHDG 7KH ORZ
VSHHG DJLWDWRU HQVXUHV DQ H[FHHGLQJO\ ORZHQHUJ\ JULQGLQJ SURFHVV DW ORZ ZHDU UDWHV
DQG KLJK PLOO VHUYLFH OLIH DQG FRUUHVSRQGLQJO\ ORZ WKURXJKSXW UDWH FRPSDUHG ZLWK
PLQLQJ DSSOLFDWLRQV 7KH DJLWDWRU HOHPHQWV DQG WKH JULQGLQJ FKDPEHU ZDOO DUH PDGH
RI ZHDUUHVLVWDQW KDUGHQHG PDWHULDOV (IÀFLHQW FRROLQJ RI WKH PLOO VWDWRU UHPRYHV WKH
KHDW JHQHUDWHG GXULQJ JULQGLQJ 7KH JHRPHWU\ RI WKH PLOO SHUPLWV WURXEOHIUHH VFDOHXS
WR KLJK SURGXFWLRQVFDOH FDSDFLWLHV 7KH SURGXFW OLQH LQFOXGHV VPDOO SLORW PDFKLQHV RI
 N: DQG SURGXFWLRQVFDOH PLOOV XS WR  N: 7KH VSHFLÀF JULQGLQJ HQHUJLHV IRU
JURXQG FDOFLXP FDUERQDWH ÀOOHU ZKHQ WKHUH LV DSSUR[LPDWHO\  SHU FHQW SDVVLQJ  —P
LQ WKH IHHG DUH
•  SHU FHQW SDVVLQJ  —P ² N:KW
•  SHU FHQW SDVVLQJ  —P ² N:KW
•  SHU FHQW SDVVLQJ  —P ² N:KW
$ W\SLFDO PLOO ZLWK D  PP GLDPHWHU DQG  N: GULYH ZLOO SURGXFH ² WK RI
 SHU FHQW SDVVLQJ  —P SURGXFW

122 Comminution Handbook | Spectrum Series 21

 CHAPTER 8 t STIRRED MILLS

FIG 8.15 – Wet agitated ball mill ANR – schematic and production scale mill (images courtesy Hosokawa Alpine).

Hosokawa Alpine dry vertical stirred ATR mill

7KH +RVRNDZD $OSLQH GU\ YHUWLFDO VWLUUHG $75 PLOO SUHVHQWHG LQ )LJXUH  FRPSULVHV
D F\OLQGULFDO FKDPEHU WKDW LV ÀOOHG ZLWK JULQGLQJ EHDGV RI D IHZ PLOOLPHWUHV LQ GLDPHWHU
7KH UDWLR RI JULQGLQJ EHDGV WR SURGXFW LV DGMXVWHG VR WKDW WKH VSDFHV EHWZHHQ WKH JULQGLQJ
EHDGV DUH MXVW ÀOOHG ZLWK SURGXFW 7KH EHG RI PDWHULDO JULQGLQJ EHDGV DQG SURGXFW DUH
WKHQ URWDWHG VORZO\ E\ WKH DUPV DWWDFKHG WR WKH YHUWLFDO VKDIW WKHLU SHULSKHUDO VSHHG
LV DERXW ² PV DQG DW VXFK ORZ VSHHGV WKH HQHUJ\ LQWURGXFHG E\ WKH DJLWDWRU LV
WUDQVPLWWHG PDLQO\ E\ IULFWLRQ LQWR WKH PDWHULDO EHG ,QFRPLQJ PDWHULDO LV ÀQHU WKDQ

FIG 8.16 – Hosokawa Alpine dry vertical stirred ATR mill and the mill mechanism (images courtesy Hosokawa Alpine).

Comminution Handbook | Spectrum Series 21 123

 CHAPTER 8 t STIRRED MILLS

 PLFURQV DQG JULQGLQJ EHDGV DUH LQWURGXFHG IURP DERYH $ GLVFKDUJH VFUHZ DW WKH
ORZHU HQG RI WKH PLOO H[WUDFWV JULQGLQJ EHDGV DQG SURGXFW ZKHUH WKH JULQGLQJ EHDGV DUH
VHSDUDWHG DQG UHWXUQHG WR WKH PLOO
3URGXFW VL]LQJ LV DV ÀQH DV  SHU FHQW SDVVLQJ  —P ZKHQ WKH PLOO LV RSHUDWHG LQ FORVHG
FLUFXLW ZLWK DQ DLU FODVVLÀHU 7KH HQHUJ\ LQWHQVLW\ LV DSSUR[LPDWHO\  N:Pó DQG WKH
FDSDFLW\ LV KLJK LQ UHODWLRQ WR PDFKLQH VL]H 0LOO VL]H UDQJHV IURP D URWRU GLDPHWHU RI
 PP WR  PP WKH PLOO KHLJKW IURP  WR  P DQG WKH RYHUDOO KHLJKW IURP  WR
 P 7KH LQVWDOOHG SRZHUV DUH XS WR  N: 7\SLFDO DSSOLFDWLRQV DUH OLPHVWRQH TXDUW]
DQG RWKHU KDUG PDWHULDOV

BIBLIOGRAPHY
Enderle, 8 :RRGDOO 3 'XII\ 0 DQG -RKQVRQ 1 :  6WLUUHG PLOO WHFKQRORJ\ IRU UHJULQGLQJ
0F$UWKXU 5LYHU DQG 0RXQW ,VD ]LQFOHDG RUHV LQ Proceedings XX International Mineral
Processing Congress $DFKHQ ² 6HSWHPEHU YROXPH  SS  *0'% *HUPDQ\
Hosokawa Alpine,  +andEooN oI PoZder and Particle Processing  S
Hosokawa Alpine,  )LQH LPSDFW PLOOV DQG FODVVLÀHU PLOOV >RQOLQH@ $YDLODEOH IURP KWWSV
ZZZKRVRNDZDDOSLQHFRPSRZGHUSDUWLFOHSURFHVVLQJPDFKLQHVÀQHLPSDFWPLOOVDQG
FODVVLÀHUPLOOV! >$FFHVVHG  -XQH @
Kwade, $ DQG 6FKZHGHV -  :HW FRPPLQXWLRQ LQ VWLUUHG PHGLD PLOOV .21$ PoZder and
Particle Journal ²
Pease, - ' <RXQJ 0 ) &XUU\ ' DQG -RKQVRQ 1 :  ,PSURYLQJ ÀQHV UHFRYHU\ E\ JULQGLQJ
ÀQHU LQ Proceedings MetPlant 2004 Conference SS ² 7KH $XVWUDODVLDQ ,QVWLWXWH RI 0LQLQJ
DQG 0HWDOOXUJ\ 0HOERXUQH
Rule, & 0  6WLUUHG PLOOLQJ ² QHZ FRPPLQXWLRQ WHFKQRORJ\ LQ WKH 3*0 LQGXVWU\ The Journal
of the 6outhern $frican Institute of Mining and Metallurgy ²
Rumpf, +  3UREOHPV RI VFLHQWLÀF GHYHORSPHQW LQ SDUWLFOH WHFKQRORJ\ ORRNHG XSRQ IURP D
SUDFWLFDO SRLQW RI YLHZ PoZder Technology ²

&DWDORJXHV IRU VWLUUHG PLOOV ÀQHJULQGLQJ PLOOV UHIHUUHG WR LQ WKLV FKDSWHU DUH DYDLODEOH RQ WKH
LQWHUQHW 7KH FRPSDQLHV DUH
• (&& VDQG PLOOV (QJOLVK &KLQD &OD\ PLOOV
• +,*PLOO
• +RVRNDZD $OSLQH PLOOV
• ,VD0LOO
• 7RZHU PLOOV YHUWLFDO VWLUUHG PLOOV
• 9HUWLPLOO

124 Comminution Handbook | Spectrum Series 21

 HOME

Chapter 9

MILL LINERS
Spike Taylor, Matthew Fitzsimons,
Cathy Hewett and John Russell

THE ROLE OF LINERS IN TUMBLING MILLS

Particle breakage occurs in a tumbling mill when the mass of solid particles and grinding
balls is detached by gravity from the rising shell of the rotating mill and tumbles onto
the toe of the charge.
Particles are broken down by impact, abrasion and attrition in the charge with the
majority of impact breakage in the toe area. The mill shell and mill ends are abraded if
they are unprotected and in extreme circumstances, this may lead to the destruction of
the mill as shown in Figure 9.1.

FIG 9.1 – Shell cracking that may develop with inadequate lining (image courtesy Metso).

Comminution Handbook | Spectrum Series 21 125

 CHAPTER 9 t MILL LINERS

/LQHUVSURYLGHWKHVDFULÀFLDOOD\HURIZHDUUHVLVWDQWVXUIDFHVLQPLOOV7KH\DUHXVXDOO\
PDGH RI VWHHO UXEEHU RU VWHHOFDSSHG UXEEHU DOVR NQRZQ DV FRPSRVLWH 5XEEHU
SURGXFWVDUHDWWUDFWLYHEHFDXVHRIWKHLUORZZHLJKWEXWWKH\FDQQRWEHXVHGDWWKHKLJK
WHPSHUDWXUHVWKDWRFFXULQFHPHQWPLOOVRULQYHU\ODUJHPLOOVDVWKHIRUFHVDUHWRRJUHDW
IRUWKHUXEEHUWRZLWKVWDQG
/LQHUVLQRUHJULQGLQJPLOOVDUHGHVLJQHGZLWKWKUHHREMHFWLYHV
 SURWHFWWKHVKHOODQGWKHHQGVRIWKHPLOOIURPZHDURUGDPDJH
 WUDQVIHUHQHUJ\HIÀFLHQWO\IURPWKHPRWRULQWRWKHPDVVRIRUHDQGJULQGLQJPHGLDVR
WKDWRUHEUHDNDJHLVRSWLPLVHG
 UHPRYHWKHJURXQGFRQWHQWVHIÀFLHQWO\IURPWKHPLOODWWKHGLVFKDUJHHQG
7KH FRQÁLFWLQJ UHTXLUHPHQWV RI WKH OLQHU WR ERWK SURWHFW WKH PLOO VKHOO IURP LPSDFW
DQG ZHDU DQG WUDQVIHU URWDWLRQDO HQHUJ\ RI WKH PLOO WR WKH PHGLD DQG FKDUJH QHHG
WR EH FDUHIXOO\ EDODQFHG $Q H[DPSOH ZRXOG EH GHVLJQLQJ DQG LQVWDOOLQJ ODUJHU RU
WKLFNHU OLQHUV WR LQFUHDVH WKH OLQHU OLIH ZKLFK ZRXOG UHGXFH WKH PLOO YROXPH DQG
WKURXJKSXW 7KHUH DUH PDQ\ YDULDEOHV WKDW FDQ VLJQLÀFDQWO\ DIIHFW PLOO OLQHU GHVLJQ
DQG SRRU OLQHU GHVLJQ FDQ KDYH D GHWULPHQWDO HIIHFW RQ ERWK PLOOLQJ SHUIRUPDQFH
DQG OLQHU OLIH /LQHU GHVLJQ GUDZV SULPDULO\ RQ SUDFWLFDO H[SHULHQFH IURP PLOOLQJ
RSHUDWLRQV DORQJ ZLWK WKH HYHUGHYHORSLQJ FRPSXWHUEDVHG PRGHOOLQJ RI FKDUJH
PRWLRQ DQG WUDMHFWRULHV DQG ODWHO\ OLQHU ZHDU DQG HQHUJ\ XVDJH 7KH UHODWLYH
RSHUDWLQJ FRVW RI OLQHUV LQ FRPPLQXWLRQ LV JLYHQ LQ )LJXUH  LW YDULHV IURP
²SHUFHQWZKHQQRVWHHOEDOOVDUHXVHGWRVL[SHUFHQWZKHQVWHHOEDOOVDUHXVHG

FIG 9.2 – The operating cost of liners relative to media and energy (image courtesy Metso).

VARIATIONS IN LINERS FOR ORE MILLS

)RUODUJHUPLOOVZLWKZLGHULQOHWWUXQQLRQVDOLQHUKDQGOHUPDFKLQHLVQRZFRPPRQO\
XVHG)RUVPDOOHUPLOOVHVSHFLDOO\EDOOPLOOVWKHOLQHUVPXVWEHLQVWDOOHGE\KDQGDQG
VKRXOG EH VPDOOHU LQ VL]H DQG ZHLJKW 7KH OLQHU KDQGOHUV GLVFXVVHG ODWHU DUH XS WR
WFDSDFLW\PHDQLQJWKHOLQHUVWKHPVHOYHVFDQQRZEHXSWRWSHULQGLYLGXDOOLQHU
7KLVLVQRWUHVWULFWHGWRVKHOOOLQHUVEXWDOVRKHDGOLQHUV,QFUHDVHLQOLQHUVL]HVLJQLÀFDQWO\
UHGXFHVWKHQXPEHUUHTXLUHGWRUHOLQHDPLOODQGWKHDPRXQWRIWLPHUHTXLUHGIRUUHOLQLQJ

126 Comminution Handbook | Spectrum Series 21

 CHAPTER 9 t MILL LINERS

This also allows the opportunity to reduce the number of bolts per liner. These designs
provide great cost beneÀts to the operation as lost operating time reduces production
and decreases revenue.
The four important process objectives in mill liner design are 5oyston, 200:
1. providing the key between the mill charge and the mill shell for charge motion
2. maximising the rate of delivery of thrown grinding media at the toe of the charge to
gain the best milling action
3. avoiding liner and ball damage
4. providing an economic linerwear life.
Shell liner design has been driven in the past by practical operating experiences; for
example, variations of lifter heights, liner spacing and lifter angles all change variables such
as grinding ball trajectories and chargelifting rates, each of which affects mill performance.
Many types of liners are currently available, some of which are now discussed.

MATERIALS OF CONSTRUCTION
The selection of material used for mill liners is dependent on the liner design and
application conditions such as ore type, size of the mill, corrosion environment, size of
balls or ore and mill speed. The most common materials used for mill liner construction
are pearliticchromium molybdenum steel, highchromium white iron, synthetic and
natural rubber and combinations of metallic and rubber materials. General properties
and corresponding applications for each of these are given.
Pearlitic-chromium molybdenum (Cr-Mo) steel has excellent wear characteristics with good
impactabrasion resistance and is generally used for semiautogenous grinding SAG
mill liners. The carbon content is at or slightly greater than the eutectoid composition on
the ironcarbon phase diagram, resulting in a microstructure of Ànegrained pearlite with
varying amounts of hypereutectoid carbides. Pearlitic &rMo steel reTuires controlled
cooling during heat treatment to avoid the formation of hard and brittle martensite or
bainite. Pearlitic &rMo steel work hardens slightly from impact in service, resulting
in a harder surface with a softer and tougher core of the casting. Pearlitic &rMo steel
can have a hardness from 300 to 400 %rinell hardness number %+1 due to variations
of carbon and chromium contents, which can be modiÀed to increase or decrease the
hardness and toughness. A lower carbon content is used when slightly greater impact
resistance is reTuired, such as for discharge grates. A higher carbon content is used for
thinner section liners and applications with less impact that reTuire a higher hardness
for abrasion resistance.
Steel liners have been used in ball mills since they became common in the late 19th
century. They are a medium cost and are the heaviest of all designs so they reTuire a
liner handler. An indication of the market size for these liners is that &,T,&+,& in &hina
has a dedicated factory to manufacture 50 000 t/a of chrome molybdenum steel liners.
High-chromium white iron is considered to have superior abrasion and corrosion
resistance, although it is considered a brittle material. +ighchromium white iron
has a microstructure consisting of iron and alloy carbides in a matrix of tempered
martensite. The hardness, abrasion resistance and fracture toughness depend on the
speciÀc chemical composition and heat treatment used in order to control the carbide
formation as well as the matrix structure and composition. +ighchromium white iron

Comminution Handbook | Spectrum Series 21 127

 CHAPTER 9 t MILL LINERS

generally has a hardness of 00 %+1 or greater resulting in superior abrasion resistance.
The excellent abrasion resistance results in a sacriÀce of toughness, so highchromium
white iron is generally used in autogenous grinding AG mills or ball mills where high
impact resistance is not reTuired.
Fabrication of steel mill liners is a broad subject that will not be covered in this book;
however, a few points concerning production processes will be mentioned because
they are necessary to ensure good Tuality liners. The mould designs are checked for
soundness of liner product using 3D modelling packages such as MAGMASOFT®,
which simulates both the Àlling of the mould plus the solidiÀcation pattern of the
molten metal. 8ltrasonic testing still needs to be used to conÀrm the absence of porosity
inside the liners. Chemical analyses are carried out as well as heat treatment to obtain
the correct microstructure and the best through hardness. Dimensional checks need to
be done to ensure ease of Àtment during the reline.
Rubber (elastomer) mill liners are lighter in weight than metal liners and are a wear
resistant alternative Figure 9.3a. 5ubber for a mill liner should have an elongation of
450 to 00 per cent, a tensile strength of around 1.5–20 MPa and a hardness of between
55 and 0 durometer on the A scale. The properties of rubber allow rubber mill liners to
process hard ore without cracking. The sound dampening effect results in the mill being
less noisy. Installation and the removal of rubber liners is considered easier and more
Áexible due to the weight of the liners. Attachment is done by using aluminium or steel
tracks moulded into the bottom of the lifter bars, which in turn clamp the plates into
place. Forged Tbolts speciÀc to the tracks bolt the rubber lifters to the mill shell. Design
of rubber mill liners can utilise the same methods and programs.
These type of liners were introduced in the late 1930s, and are the lightest of all designs
and have a relatively low to medium cost. They are used in small to mediumsized mills
and can be installed easily. Many improvements in rubber compounds have been made
over the years; nanotechnology research should further improve wear life in the future.
SpeciÀc formulations can be used for different applications, such as a chemicalresistant
rubber for mills that have diesel added to the pulp.
Composites utilising a steel insert achieve a combination of excellent overall toughness
with abrasion resistance in particular sections of the liner Figure 9.3b. These inserts can
be CrMo castings or Tuench and tempered plate, normally 5 mm thick. An increased
wear life for chromium molybdenum steel composite liners can be achieved using high
chromium white iron inserts in lowimpact areas. Metalfaced rubber liners with designs

A B

FIG 9.3 – Mills equipped with (A) rubber and (B) composite liners (image courtesy Metso).

128 Comminution Handbook | Spectrum Series 21

 CHAPTER 9 t MILL LINERS

resistant to damage through ball impacts are growing in applications in SAG mills.
They were developed in the 1980s, and are used in medium to large mills. Their cost
lies between rubber and steel linings. The larger composite lifter bars can reTuire lifting
lugs for installation using a mill liner handler.
Fabrication of rubber mill liners is a broad subject that will not be covered in this book.
The basic process is that uncured rubber of the reTuired formulation is put into custom
designed moulds and any metal inserts are prepared by grit blasting followed by the
application of a bonding agent. The rubber is then cured at a high temperature using
high pressure to keep the mould closed. The liners are trimmed to the Ànished size in a
postmoulding stage.
Combination linings incorporating steel lifter bars and rubber shell plates can be used for
weight reduction of the overall lining. A lining can also have steel lifters and rubber grates
to overcome blinding problems associated with steel grates in certain applications. 5ubber
pulp lifters can be installed behind steel grates for weight and cost considerations.

Steel versus rubber

Steel mill liners can withstand the forces associated with the everincreasing dimensions
of the newer generation of mills and can be cast as larger liners commensurate with
the increase in capacity of the mill liner handlers. They can be cast in a variety of
geometries, such as curved grates and pulp lifters, to suit the existing drilling pattern on
a discharge head, while rubber liners can be restricted due to mould constraints. 5egular
inspections by the liner supplier can reveal uneven wear patterns of the liners, which
can be overcome by providing lifter bars of varying heights, with higher sections where
the high wear occurs. Steel mill liners can withstand high temperatures.
5ubber mill liners signiÀcantly reduce the noise levels in a mill environment and can
be manually installed in and removed from the mill. Steel liners have a raised section
around the perimeter of the liner on the bottom surface and around the bolt holes for
ease of fettling the liner after casting, ensuring a good Àt with the mill·s curvature.
This can lead to gold lockup during operation, prompting a high inventory of gold
unable to be recovered until the liner changeout occurs. This raises the security level of
gold theft during mill relines, but is eliminated with rubber liners due to the good Àt
against the shell of the mill. When there are highly acidic conditions in the mill, rubber
liners with stainless steel tracks and attachments need to be used.

POTENTIAL ISSUES WITH LINERS

Peening on liners
A steel surface dimpled with impact craters, often called ¶orange peel·, can sometimes
be seen on steel liners and steel inserts in composite lifter bars. This is detrimental to the
liners· performance and conseTuently their life within the mill. Liners showing this type
of surface are the result of grinding media impacting on steel liners rather than the toe of
the charge. If the mill speed is too fast, the energy imparted to the balls/ore upon release
is too high and the trajectory takes it over the toe, hitting the liners. A slower mill speed
can move the impact point back to the toe. An alternative cause of these impacts is a low
charge volume, which is often associated with the commissioning of a mill. Increasing

Comminution Handbook | Spectrum Series 21 129

 CHAPTER 9 t MILL LINERS

the mill charge volume will increase the toe size thereby moving the toe position, so
the impact position returns to the toe.

Broken/damaged liners
These can result from media impacting directly on steel liners, especially in AG/SAG
mills as the energies of impact are Tuite high because rocks are large. If this still occurs
once the impacts have been resolved, the Tuality of the castings as well as the bolting
and torTuing regimen should be checked. In the case of rubber liners, gouging wear can
be Tuite damaging to their total life. 5ubber liners should therefore be inspected closely
for signs of gouging; for example, from material such as tramp metal.

Excessive liner wear

If excessive liner wear is observed when comparing liner performance between one
installed set and another, there are usually two main causes:
1. properties of the liner material have changed
2. operational parameters within the mill have changed.
Investigation of operational variations and an analysis of the liner should be
undertaken.

Operational changes during liner lifetime

The mill operating conditions are used to initially design the liner shape and proÀle.
Therefore, if operational conditions vary during the life of the liners, it can materially
affect both the liners· performance and the grinding efÀciency within the mill. The design
of the mill liner is as robust as can be; however, large variations in operational conditions
– such as mill speed, input tonnes per hour, or ore size – can have signiÀcant effects on
the life of the liners and the overall functioning of the mill.

Damaged rubber backing

1ormally a 5– mm rubber backing is glued to the mill shell for the liners to be seated
against. This rubber backing can become damaged if the liners move during operation
or if damaged liners allow direct impact by the charge. It needs to be replaced before
new liners are installed during the reline. A polyurethane spray is sometimes applied in
place of the rubber backing for highly acidic milling conditions, as the polyurethane can
be sprayed into the mill bolt holes for protection of these drilled surfaces.

TYPES OF LINERS

Steel mill liners

Steel liners normally have angled ends to prevent the racing wear of the charge·s Àner
section in the gaps at the ends of the liners during the mill·s rotation. This is because steel
liners are designed with small gaps 10 mm between the liners for ease of installation.
These gaps are Àlled with thin rubber extrusions as a method of preventing the gaps
Àlling with ball chips, which would cause difÀculties in removing the worn liners.
The ends against the mill heads are perpendicular to the length of the mill to Àt against

130 Comminution Handbook | Spectrum Series 21

 CHAPTER 9 • MILL LINERS

the corner segment, although this edge can be a saw-toothed shape if this is the profile
of the corner segment.

Solid or integral liners

These have the plates and lifter cast as a single unit. They are therefore easier to install
and will have fewer parts within the mill. This has allowed the development of larger
liners such as the ‘double wide’ and ‘double long’ liners.
The top-hat design (Figure 9.4) is preferred to other integral liners as the symmetrical
cross-section results in uniform solidification of the molten metal and so there is less
possibility of porosity within the liner. The bolt heads are also protected to a higher
degree within the bulk of the lifter.

A B

FIG 9.4 – The (A) top-hat and (B) rail designs are two examples of solid/integral liners (images courtesy Cathy Hewett).

Removable lifters / bolted plate and lifters

In this liner system the lifters are separate from the plates, enabling a change of lifters
if the plate sections show little wear (Figure 9.5). This system can be installed where no
mill liner handler is available as traditional lifting devices can be used for the relines.
A disadvantage is the larger number of liners involved in a reline.

FIG 9.5 – Loose plates with a bolted lifter (image courtesy Cathy Hewett).

Uni-directional and bi-directional liners

A liner may be designed to suit uni-directional or bi-directional milling. For uni-
directional operations the lifter can have different leading and trailing angled faces,
allowing more customisation for actual mill conditions. It may also allow for weight
redistribution on the liners for increased liner life.

Wave liners
Wave liners are designed for ball mills to produce a cascading charge motion (see
Figure 9.6). The safety of the design ensures that there is no overthrow of the charge, so
high-chrome white iron can be used for its hardness and abrasion resistance.

FIG 9.6 – Wave liner (image courtesy Cathy Hewett).

Comminution Handbook | Spectrum Series 21 131

 CHAPTER 9 t MILL LINERS

Rubber mill liners

5ubber mill liners are normally designed to be placed with a slight amount of
compression to prevent gaps and racing wear along a circumferential path. The joints
between the rows of lifter bars and the shell plates can all be in line if the operational
preference is to reline in rings around the mill. The joint between adjacent rows of lifter
bars and the shell plates can also be staggered to prevent racing wear if the practice is to
reline the complete shell lining.

Integral liners – Metso Megaliner

The Megaliner Figure 9. was designed by Metso for large SAG, AG and ball mills that
have large discharge systems, and where modern liner handlers are available. Each liner
covers a large area of up to 4 m2 and contains lifter and shell plate rows. These liners are
lightweight in relation to their size. A liner positioning system is available that provides
for faster liner replacements by giving the operator better visibility of alignment between
liner bolt holes in the Megaliner and the mill shell bolt hole pattern.

FIG 9.7 – Metso Megaliner being placed by a Russell Mineral Equipment machine (image courtesy Metso).

Loose liners
5ubber liners are generally loose sections. The lifter bars are normally unidirectional
but bidirectional lifters can be supplied. Figure 9.8 shows rubber lifter bars and a shell
plate.

FIG 9.8 – Rubber lifter bars and shell plate (image courtesy Multotec).

Lifter bars
5ubber and composite liner bars can be moulded in multiple conÀgurations designed to
suit the application Figure 9.9. 4uench and tempered plate composite liners provide

132 Comminution Handbook | Spectrum Series 21

 CHAPTER 9 t MILL LINERS

FIG 9.9 – Lifter bars (image courtesy Multotec).

good toughness on the lifting surface. Casting insert composites provide better cover for
the top surface of the lifter bars in highenergy applications Figure 9.9.

Plate liners
5ubber plate liners, seen in Figure 9.10, can be moulded to suit any application.
Composite plate liners can be designed to improve wear life costeffectively.

FIG 9.10 – Plate liners (image courtesy Multotec).

Discharge liners
5ubber discharge grates Figure 9.11 can demonstrate reduced ¶pegging· – that is, hole
blockage – during operation as the apertures are more Áexible than steel grates. Composite
material can be added to the grates for additional stiffness and wear resistance if reTuired.

FIG 9.11 – Discharge liners (image courtesy Multotec).

DESIGNING MILL LINERS

Mill liners are designed utilising many tools available to the industry as well as
established liner best practices. It is important to understand mill operating conditions in
order to feed the correct information into the liner design tools. Shell liners are designed
to give optimal grind efÀciency as well as maximum liner life. Liner design variables can
be determined by modelling the trajectory of the charge in the mill using 2D modelling
software such as MillTraj, and 3D modelling such as various discrete element modelling

Comminution Handbook | Spectrum Series 21 133

 CHAPTER 9 t MILL LINERS

(DEM) packages (Powell et al, 2006). Discharge liners are designed to optimise mill
WKURXJKSXWE\FRQWUROOLQJWKHWRWDORSHQDUHDRIWKHPLOODVZHOODVGHÀQLQJDSHUWXUH
size and shape.

2D trajectory or throw prediction

As designs based on experience reached their limit and the computational power of
computers increased, predictions were possible using a 2D slice through the mill to look
at the trajectory or throw of the ore and grinding media within the mill. Figure 9.12
represents a vertical slice through the mill in motion. A major issue in the process design
of shell liners is the ‘throw’; that is the trajectory of balls and ore leaving the liners.
The face angle of the lifter has an important bearing on the point at which the ball or
ore is released and its subsequent trajectory. The main tumbling pattern within an AG
or SAG mill is known as ‘cataracting’. Cataracting is experienced by the portion of the
charge that is raised high enough in the mill to fall back to the toe of the charge leading
to impact, causing fracture of the ore and a coarse but ground product. There are also
areas of attrition and abrasion within the charge, which affect how the mill grinds and
how throughput is achieved as well as how the liners perform. Increasing the face
angle of the lifter is used to change the release point of the balls. When the face angle is
decreased, the space between the liners is reduced. In order to avoid packing (that is,
ÀOOLQJWKHVSDFHEHWZHHQOLQHUVZLWKÀQHRUHDQGWRUHVWRUHWKHOLIWLQJFDSDFLW\RIWKH
lifters, the lifter spacing needs to be increased.
Further DEM has modelled the interactions in three dimensions within the mill and 3D
DEM has become an additional tool in the design of mill liners.

FIG 9.12 – Regions in the mill: cataracting, cascading, attrition

and abrasion, eye, shoulder, toe (image courtesy Cathy Hewett).

134 Comminution Handbook | Spectrum Series 21

 CHAPTER 9 • MILL LINERS

LINERS FOR MILL TYPES

Autogenous grinding mills
AG mills, also referred to as pancake mills due to their length-to-diameter ratio,
always have a grate discharge system to transport the pulp to the outlet trunnion.
Utilising any of the designs discussed in the section ‘Types of Liners’, steel liners may
incorporate a HiLo profile to reduce the downtime to do a reline. This design has
alternating rows of lifter bars as high lifter bars followed by lower lifter bars. When
the high lifter bars are worn the lower lifter bars are replaced with new high lifter bars,
and so on. The impacts of rocks and balls on liners are severe and may cause metal
liners to be distorted. Removing the first row is difficult. This ‘HiLo’ profile also has
the effect of reducing the potential for the large feed-size rocks to pack between the
lifter bars. When considering reline time it is always advantageous to use the smallest
number of the biggest liners possible. Recent studies by the Julius Kruttschnitt Mineral
Research Centre (JKMRC) indicate that it is best to reline more frequently using large
area but thinner, performance-maximised liners. Liner design and fast, reliable safe
relines are important.
AG mills are gaining prominence in diamond mining applications as they are highly
efficient in producing large percentages of fines in a single pass. AG mills in this
application can have rubber liners installed to minimise diamond breakage.

Semi-autogenous grinding mills

As with AG mills, steel liners can be made using any of the designs that have been
discussed, and then incorporating a HiLo profile (Figure 9.13). The discharge system can
incorporate pebble ports in the grate aperture area so that the critical size particles that
may otherwise build up can be discharged. There is an appreciable number of medium-
sized SAG mills using composite liners.

FIG 9.13 – HiLo lifter bar profile (image courtesy Cathy Hewett).

Ball mills
Whilst AG and SAG mills function to break large feed size, the product is relatively
coarse and generally unsuitable for downstream equipment; this is where ball mills
come into the circuits.
If the mill is failing to grind finely enough, or under-grinding the ore, the product is
coarse and may have a low degree of liberation resulting in low recoveries. If the mill is
over-grinding, the efficiency of the mill is not balanced, increasing the power draw and
ultimately the cost. There can also be downstream problems such as sliming of the pulp,
reducing the recoveries.
The ball mill liners affect the way the charge behaves and as such have a significant
impact on the mill and grind efficiencies. Within a ball mill, the main modes of grinding
are attrition and abrasion rather than impact. These comminution modes are a result of

Comminution Handbook | Spectrum Series 21 135

 CHAPTER 9 t MILL LINERS

the cascading tumbling pattern that is most prevalent in ball mills. Cascading describes
the portion of the media that rolls down to the toe of the charge resulting in abrasion
and Àner grinding, but also increased liner wear. To generate this cascading motion,
the mill liners reÁect a more corrugated design with a larger face angle on each lifter,
accentuating the tumbling motion of the charge. 5ubber mill liners are designed to
impact the toe of the charge to provide a normal charge motion.

Rod mills
5od mill liners have a very conservative proÀle to minimise the risk of rod entanglement,
with the lift from the plate section to the top of the thicker wearresistant portion being
as gentle as possible.
The feed head liners are thicker than the discharge head liners as protection against
impact when the rods are introduced into the mill via the discharge trunnion using a
rod charger.
5ubber liners can be used on the shell of a rod mill for noise and weight considerations;
however, the lifter bars must have a rounded crosssection to provide a conservative
charge action. These lifter bars have to be asymmetrical with the track towards the rear
of the lifter bar as protection against the highimpact forces of the rods. The heads still
reTuire steel liners, as rubber liners would undergo too much spearing by the rods.

Scrubbers
Scrubber liners, especially in diamond mining, are rubberlined. The traditional lifter bar
proÀle has been sTuare, even though a cascading action is essential because of the low
rotational speed of scrubbers; however, there has been a move to design faster scrubbers,
so the trajectory needs to be checked in case chamfered lifter bars are reTuired. Scoops
can be supplied to the discharge head for clearing the scrubber of the complete charge
for improved security during relines.

Packing between lifters

Packing affects throw, in effect, by reducing the height of the lifter. As a result, balls and
rocks have a shorter journey on the lifter before they fall off. When packing occurs it is
usually marked at the feed end of the mill and diminishes towards the discharge end.
A form of packing may also occur in all mills due to smaller rocks being picked up by
the lifters as they move through the charge; as a result, a larger rock may sit on smaller
rocks and have less distance to travel.
As the liner wears, packing is gradually reduced and the face angle of the lifter can
be expected to increase concurrently. This combination maintains the throw of ore and
media being directed at the toe of the charge during the life of the liners. It follows
that a strategy of controlled packing can be used to manage throw in the feed end of
the mill.
The main drawback is that packing reduces the Àll level in the mill by one to
three per cent and the lift that can be provided by the mill very signiÀcantly. The main
positive for the limited packing of ore within the mill is it creates a form of rock lining
that can assist in reducing wear on the liner plates. If liner lifters are further spaced out,
the packing can be removed. A high face angle strategy may then be needed to reduce
shell damage from impacts.

136 Comminution Handbook | Spectrum Series 21

 CHAPTER 9 t MILL LINERS

LINERS IN CEMENT MILLS

In twocompartment cement mills, shown in Figure 9.14, the Àrst compartment is for
grinding coarse clinker and larger balls are used see Table 5.2. The second compartment
is for Àne grinding. EfÀciency is increased by the use of small balls and classifying liners.
The liners in the Àrst compartment are usually step or wave type to ensure optimum
lifting of the large balls. The shell lining in the second compartment often has corrugations
for maximum power absorption. Liners are bolted to the shell or held in place by other
means such as wedges. They are built from wearresistant material such as manganese
steel and the typical life is eight to ten years.
Other types of liners are being developed; for example, the Xlift® concept of Magotteaux,
seen in Figure 9.15, which is a Àrstcompartment liner using three interchangeable
plates that each produce a different lifting effect. The three plates enable the lifting
proÀle to be adapted to different operating parameters, thereby optimising grinding
efÀciency. The liners are light plates weighing from 18–24 kg. They are used in mills
from 3.0 to 5.4 m diameter.
Classifying liners shown in Figure 9.1 are used to improve the efÀciency of Àne
grinding in the second compartment. They are designed to move the coarser balls in

FIG 9.14 – Two-compartment FLSmidth cement clinker ball mill showing the centre
diaphragm and liners in the two compartments (image courtesy FLSmidth).

FIG 9.15 – Xlift® liners made by Magotteaux. Cross-section of liners showing the three
types of plates and the liners in place (images courtesy Magotteaux).

Comminution Handbook | Spectrum Series 21 137

 CHAPTER 9 t MILL LINERS

ϲϬ

ϱϬ

Averageballsize(mm)
ϰϬ

ϯϬ

ϮϬ

ϭϬ

Ϭ
Ϭ ϭ Ϯ ϯ ϰ ϱ ϲ ϳ ϴ ϵ
Distancefromintermediategrate(m) /2ndchamber

FIG 9.16 – Magotteaux classifying liners and their effect on ball size distribution (images courtesy Magotteaux).

the second compartment to the feed end where they are most effective in breaking the
larger particles that pass through the diaphragm. This is done by deÁecting balls during
mill rotation towards the feed end. Each classifying liner has a base plate to protect the
mill shell, and a supporting plate to which the deÁector is attached. These liners weigh
25 kg. The standard dimensions of liner plates are 250 mm long and 314 mm around the
mill shell.

INSPECTIONS, WEAR MONITORING AND ONGOING OPTIMISATION

Measurement of the liners within the mill can occur during inspection production
shutdowns or mill relines. This usually involves entering the mill when safe, considering
all lockout and conÀned space procedures. These inspections will look at the state of the
liners, the size of the grate slots, the grinding media if SAG or ball mill, the ore that is
present and the gearing. In addition, measurements of the liners within the mill can be
taken using ultrasonic thickness 8T detection. 5ecently, 3D terrestrial laser scanning
technology has increased in use globally to capture the state of the inside of the mill
with a point cloud. The 3D laser scanning, discussed shortly, has the advantage of
providing information on the entire longitudinal and crosssection proÀles rather than
the laborious multipoint information of UT testing. The UT measurements still remain
useful for an instantaneous analysis of the liners· indicative remaining life.
The liners within the mill should be inspected for any unusual or unexpected wear,
cracks or other damage and a prediction of the remaining life of the liners. The wear
rate of the liners is not linear over their life within the mill and this rate will change
while the liner is in the mill, even with a consistent input/output regime. This is due to
the changing amount of surface exposed to the charge when the liner is new, compared
to the surface exposed to the charge as the proÀle of the liner Áattens with abrasive
wear. The slippage of the charge increases as the liners wear and this accelerates the
wear rate.
These wear measurements, however they are obtained, allow an indicative life in days
and/or tonnes through the mill to be calculated and predicted. This allows production

138 Comminution Handbook | Spectrum Series 21

 CHAPTER 9 t MILL LINERS

shutdowns and mill relines to be planned based on actual rather than historical data.
7KH\DOVRDOORZDQ\GDPDJHGRUFUDFNHGOLQHUVWREHLGHQWLÀHGDQGPRQLWRUHG

CONDITION MONITORING OF LINERS

Mill liners wear by abrasion, which reduces liner thickness gradually until they must
be replaced. This is expensive since the mill must be off-line, and no grinding means
no production from the mine. Hence, the life of liners must be as long as possible,
EXWWKH\PXVWQRWEUHDNRUORVHHIÀFLHQF\LQWUDQVIHUULQJHQHUJ\WRWKHURWDWLQJPDVV
In this context the measurement of liner wear is critical and data from the measuring
device should indicate the remaining life of the liners. The cost, accuracy, reliability and
accessibility of measurement devices are important (Dandotiya et al, 2011).
Four types of device are available:
1. direct measurement device based on a terrestrial 3D laser scanner and data processing
algorithms, creating a 3D thickness map
2. direct measurement line scanner device and data processing algorithms, which are used to
create a 2D ‘slice’ of the mill from which liner condition can be measured
3. mechanical wear reading deviceZKLFKJLYHVWKHSURÀOHDWWKHPHDVXULQJSRLQWXVLQJ
manual inspections
4. acoustic detectors.
6LJQLÀFDQW HFRQRPLF VDYLQJV FDQ EH DFKLHYHG LI WKH GHYLFH LV DEOH WR SHUIRUP
measurements with only minimal stoppage of the mill, and the laser scanner is superior
to the mechanical device in this regard. The 3D laser scanner is also considered to be
the best choice regarding the complete measurement of all the liners’ dimensions.
An ultrasonic device is a promising method for measuring rubber liners with a thickness
of 400 mm; it is preferable to a mechanical device if only the height of the liner is of
interest because it measures quickly.

DIRECT MEASUREMENT OF LINER WEAR

3D laser scanner
This technique is the basis of the Outotec® MillMapper. An output from the MillMapper
that comes from viewing the surface of worn liners is shown in Figure 9.17a. Colour is

A B

FIG 9.17 – (A) MillMapper scan of the feed end liners shown in colour and (B) liners in a tumbling mill (images courtesy Outotec).

Comminution Handbook | Spectrum Series 21 139

 CHAPTER 9 t MILL LINERS

used to identify the degrees of wear, while data for the colour image are also used for
wear analysis. It uses terrestrial laser scanners to measure in situ the condition of liners in
mills, and forecasts reline timing based on wear trend analysis using intelligent software
and 3D laser scan technology. The MillMapper helps minimise the risk of catastrophic
failure by scanning and modelling liner breakage or localised accelerated wear, and it
provides accurate measurement of liner thickness and the weight and condition of the
entire mill surface. Figure 9.17b shows a tumbling mill with liners worn through use.
The software tracks and forecasts wear on the shell liners and on the discharge and feed
end liners. It also provides accurate information on mill charge; and gives measurements
of all wear surfaces to an accuracy of ±3 mm. The software automatically detects high-
wear zones, asymmetric wear patterns, cracked liners, loose plates and broken grates.
With the information it provides, operators can make informed decisions based on
automated, consistent and repeatable measurements followed by rigorous statistical
analysis. $ppro[imately Àve minutes are needed for pure measurement, e[cluding the
time for entering and exiting the mill (Dandotiya et al, 2011).
CrusherMapper is a similar technique developed by Outotec for monitoring the
condition of crusher liners. It measures concave liner thickness, mantle thickness, open
and closed side settings and the thickness of the spider arm to within 3 mm. It is used
for planning liner change and optimising the use of current liners.

2D laser scanner
A relatively new technique being developed by Multotec is the MultoScan system. It uses
a line scanner to map a 2D slice of the mill liner proÀle that is perpendicular to its central
axis at predetermined points along the shell, as well as speciÀc liner measurements on
the heads. Data is compiled to map the proÀle of the mill liners over their operational
life and can be used to predict remaining liner life as well as perform mill trajectories
(see Figure 9.18).

FIG 9.18 – A plot generated using MultoScan software (image courtesy Multotec).

Vertical pins
The device consists of a frame and Àve, nine and 13 or more rods. An example of a Àve-
rod device can be seen in Figure 9.19. These proÀle gauges are used for rubber liners,
with nails being hit into the rubber until they contact the mill shell to provide the base
height for the measurements. It gives a measure of liner proÀle at the measurement
point using manual inspections. Approximately 30 minutes are required for measuring
12 important liners in the mill.

Acoustic monitoring
The noise emissions produced by steel balls when they collide with other balls or mill
liners and lifters are of speciÀc interest from a mill wear perspective because their

140 Comminution Handbook | Spectrum Series 21

 CHAPTER 9 t MILL LINERS

FIG 9.19 – Mechanical wear measuring device (image courtesy Metso).

hardness and speciÀc gravity result in kinetic energy being available to wear the mill
internal components. The impact time and contact area for a steel ball in a typical SAG
mill indicate that the stress of the steel liner is sufÀcient to cause plastic deformation
and breakage of the liner/lifters. Acoustic techniTues can be used to monitor the events
occurring inside a mill. Traditionally, the noise level expressed in decibels has been
used for analysis but an improved system has been developed for analysing noise
emissions using physicsbased models and appropriate signalprocessing techniTues.
The resultant signals include a steelonly signal and signals that classify the intensity of
the events. These have been combined to provide a mill shell steel hits MSS+ signal
with which individual events can be recorded.
Protected microphones are located very close to the mill outer surface so that
background noise is irrelevant. By measuring signals at individual locations around
the mill, the likelihood of accelerated wear and/or cracking of liners and lifters can be
predicted. By integrating the MSS+ signal with time, it is possible to monitor the effects
of operational changes such as mill speed on the wear of the mill liners and lifters.
This acoustic system is being used for process control applications.

RELINING MILLS
Mill relining can only take place when the mill is shut down, and since the mine does not
earn revenue when the grinding circuit does not operate, it is essential that shutdowns
be as brief as possible while achieving the safest possible work practices.
Only basic mechanisation was used during relining in the 1950s when mill diameters
did not exceed 3 m. The work was labourintensive and timeconsuming, and reline
personnel resorted to a variety of methods to assist in removing and installing liners
including cranes, chains and pulleys. By the late 1960s the largest ball mills had more
than tripled in volume, liners were larger and heavier and mining companies were
devising techniTues to improve relining practices. For example, Bougainville Copper
Ltd had 15 mills that were 5.6 m × 6.4 m in the early 1980s and its staff developed
a relining system that used hydraulic tools, hydraulically powered access platforms
and a conventional liner handler. It resulted in faster and less onerous liner changes.
The use of AG and SAG mills was increasing and the heavy liners in them weighing
up to 8 t reinforced the need for better relining techniTues. Many companies now
provide specialist relining eTuipment and services with the objective of safe and
efÀcient relining practice. Mill relining occurs in two phases: worn liner removal and
new liner insertion.

Comminution Handbook | Spectrum Series 21 141

 CHAPTER 9 t MILL LINERS

Removal of worn liners

5emoving worn liners may reTuire:
x Extracting the nuts from the bolts that attach the liners to the mill and driving the
bolts into the mill. The 5ussell Mineral ETuipment 5ME Thunderbolt recoilless
hammer shown in Figure 9.20 is the type of hammer used for this purpose. The
hammer is made in four models that deliver 60 blows per minute at 250–1500 joules
per blow and the model weights are from 20 to 460 kg.
x Collecting and removing the liners that have become detached from the mill.
x Dislodging liners that remain in place due to being peened together. The 5ME
warning that ‘dislodging liners from SAG mills represents the greatest single risk
to mill relining crews and to the planned relining schedule· must be kept in mind
5ussell Mineral ETuipment, 2014b, page 13.

FIG 9.20 – The Russell Mineral Equipment (RME) recoilless hammer is used to
remove and drive bolts into the mill (image courtesy RME).

Insertion of new liners

1ew liners are lifted into position and attached to the mill shell or ends by bolts. Lifting
is carried out by liner handlers for small to medium mills liners weighing 500–1500 kg
or by relining machines for medium to large mills liners weighing 1500–8000 kg, both
seen in Figure 9.21.
The 5ussell 3 mill liner handler reTuires mill entry diameters greater than 800 mm for
units with 1000 kg capacity and greater than 1100 mm for 1500 kg capacity.
Mill liner handlers are made by several companies. The handler made by Maxitool,
which is an engineering company located in north 4ueensland, is shown in Figure 9.22.
Most liner placement mill relining machines in the world 350 units as of 2015 are
made by 5ME.

142 Comminution Handbook | Spectrum Series 21

 CHAPTER 9 t MILL LINERS

A B

FIG 9.21 – Russell Mineral Equipment (RME) (A and B)  single and twin mill liner handlers. Each handler has
three axes of movement. (C) Mill relining machine (images courtesy RME); it has seven or eight axes of movement.

FIG 9.22 – Maxitool liner handler – 1500 kg capacity. It can enter mills through
trunnions greater than 780 mm diameter (image courtesy Maxitool).

BIBLIOGRAPHY
Dandotiya, 5, Lundberg, -, Wijaya, A 5 and Parida, A, 2011. Evaluation of abrasive wear
measurement devices of mill liners, COMADEM, 142:3–1.
Magotteaux, n/d. Xclass® >online@. Available from: http://www.magotteaux.com/products
services/tubemillinternals/classifying/xclass/! >Accessed: August 2014@.
Maxitool Group, n/d. Mill relining [online]. Available from: <http://www.maxitool.com.au/
mining/millrelining/! [Accessed: August 2014].

Comminution Handbook | Spectrum Series 21 143

 CHAPTER 9 t MILL LINERS

Metso Minerals, n/d. 5ubber mill linings. Available from: <http://www.metso.com/services/

milllinersandtrommels/rubbermilllinings/! [Accessed: August 2014].
Outotec, n/d. MillMapper [online]. Available from: <http://www.outotec.com/en/Products
services/ProcesseTuipment/Grindingmills/Linerconditionmonitoring/MillMapper/!
[Accessed: August 2014].
Powell, M S, 2000. MillTraj – liner design software, -.Tech Application Papers, University of
4ueensland.
Powell, M, Smit, I, 5adziszewski, P, Cleary, P, 5attray, B, Eriksson, .G and Schaeffer, L, 2006.
Selection and design of mill liners, in Advances in Comminution ed: S .omar .awatra,
pp 331–36 Society of Mining, Metallurgy, and Exploration: Littleton.
Royston, D, 200. Semiautogenous grinding SAG mill liner design and development, Minerals
and Metallurgical Processing, 243:121–132.
Russell Mineral Equipment, 2014a. Introduction to mill relining brochure.
Russell Mineral Equipment, 2014b. Steps on the mill relining path brochure.
Russell Mineral Equipment, 2014c. 5ME mill relining system product range brochure.

144 Comminution Handbook | Spectrum Series 21

 HOME

Chapter 10

CLASSIFIERS
Alban Lynch, Mark Weaver, Hakan Dundar,
Hakan Benzer and Jobe Wheeler

SIZE SEPARATION IN COMMINUTION PROCESSES

The objective of most comminution processes for ore, cement or coal is to break the material
so that the product is of a suitable size range for downstream processing. The required
VL]HLVXVXDOO\DERXWSHUFHQWSDVVLQJ—PRUÀQHU)RUWKHFRPPLQXWLRQRIRUH
several size-reduction circuits in series may be used depending on the material being
SURFHVVHGDQGWKHUHGXFWLRQLQVL]HWREHDFKLHYHG7KHÀUVWFLUFXLWVLQDFRPPLQXWLRQ
SURFHVVDUHGHVLJQHGWREUHDNWKHFRDUVHUSDUWLFOHVDQGWKHÀQLVKHGSURGXFWLVPDGHLQ
WKHÀQDOFLUFXLW6L]HVHSDUDWLRQRIFRDUVHUSDUWLFOHVLVE\VFUHHQVZKLOHLQWKHÀQDOFLUFXLW
FHQWULIXJDOIRUFHLVXVHGWRVHSDUDWHÀQHSDUWLFOHVLQDSURFHVVNQRZQDVFODVVLÀFDWLRQ
:HWDQGGU\FODVVLÀHUVDVZHOODVVFUHHQVDUHGLVFXVVHGLQWKLVFKDSWHU
)LJXUH  VKRZV D FUXVKLQJ FLUFXLW DQG D VHPLDXWRJHQRXV JULQGLQJ 6$* PLOO 
EDOOPLOOFLUFXLW%UHDNDJHRFFXUVLQWKHFUXVKHUV6$*PLOODQGEDOOPLOO6L]HVHSDUDWLRQ
RFFXUVRQVFUHHQVLQWKHFUXVKLQJSODQWRQWKHVFUHHQWKDWIROORZVWKH6$*PLOORQWKH
WURPPHODWWDFKHGWRWKHEDOOPLOODQGLQWKHF\FORQHWKDWFODVVLÀHVWKHEDOOPLOOGLVFKDUJH
7URPPHOVDUHGLVFXVVHGLQWKHIROORZLQJVHFWLRQIROORZHGE\K\GUDXOLFFODVVLÀHUVDQG
WKHQDLUFODVVLÀHUVZKLFKDUHHPSOR\HGLQGU\JULQGLQJFLUFXLWVVXFKDVFHPHQW+LJK
IUHTXHQF\YLEUDWLQJVFUHHQVQRZRSHUDWLQJDVFODVVLÀHUVDUHGLVFXVVHGLQDODWHUVHFWLRQ
Low-frequency screens are used in crushing and coarse grinding circuits, and will be
discussed in this section.
$QLQGXVWULDOVFUHHQFRQVLVWVRIDPHFKDQLVPWKDWFUHDWHVWKHYLEUDWLRQVVFUHHQPHGLD
on which particles are separated and a rigid frame that holds the mechanism and the
PHGLD 6FUHHQV PD\ EH VLQJOH RU GRXEOH GHFN RFFDVLRQDOO\ WULSOH GHFN DQG FDQ EH
GHVLJQHGWRVXLWWKHWDVNLQFOXGLQJDFXUYHGUDWKHUWKDQDÁDWVXUIDFH7KHVFUHHQVKRZQ
LQ )LJXUHLVPDGHLQVL]HVIURPWR PZLGHDQGWR PORQJ DQGWKH
FDSDFLWLHVUDQJHIURP²WK6FUHHQVRIWKLVW\SHDUHXVHGLQPDQ\FLUFXLWVRQH
XVHEHLQJIRU6$*PLOOGLVFKDUJH

Comminution Handbook | Spectrum Series 21 145

 CHAPTER 10 t CLASSIFIERS

FIG 10.1 – Simplified ore-crushing (top) and comminution circuits (bottom) (images courtesy Hakan Dundar).

FIG 10.2 – A vibrating screen (image courtesy Metso).

146 Comminution Handbook | Spectrum Series 21

 CHAPTER 10 t CLASSIFIERS

Low-frequency screens are produced by many manufacturers and there are small
differences but the common factors are the vibrating mechanism, the screen media and
the rigid frame.

CLASSIFICATION
$FODVVLÀHULVGHVFULEHGE\LWVSDUWLWLRQFXUYHZKLFKVKRZVWKHSHUFHQWDJHRIHDFKVL]H
fraction in the feed that enters the coarse product. This is also known as the Tromp
FXUYHVKRZQLQ)LJXUH,QWKLVFKDSWHUWKHSDUWLWLRQFXUYHZLOOEHXVHGWRGHVFULEH
WKHUHVXOWVRIFODVVLÀFDWLRQ

ϭϬϬ
/ĚĞĂů
ϵϬ
ƐĞƉĂƌĂƚŝŽŶ ZĞĂů
ϴϬ
ƐĞƉĂƌĂƚŝŽŶ
ϳϬ
ƉĂƌƚŝƚŝŽŶй

ϲϬ
ϱϬ
ϰϬ
ϯϬ
ϮϬ
&ŝƐŚ,ŽŽŬ
ϭϬ
LJƉĂƐƐ
Ϭ
G
ϭ ϭϬ ϭϬϬ ϭϬϬϬ
ŵĞĂŶƉĂƌƚŝĐůĞƐŝnjĞ;ŵŝĐƌŽŶͿ

FIG 10.3 – Partition or Tromp curve for a classifier (image courtesy Hakan Dundar).

7KHSDUWLWLRQFXUYHLVGHÀQHGE\WKUHHYDULDEOHVWKHVORSHRUVKDUSQHVVRIWKHOLQHDU
section, the d50 YDOXHWKDWGHÀQHVLWVORFDWLRQRQWKHJUDSKDQGWKHE\SDVVZKLFKLVWKH
IUDFWLRQRISDUWLFOHVWKDWDUHPLVSODFHGLQWRWKHFRDUVHIUDFWLRQ7KHUHDVRQIRUDQLQHIÀFLHQW
closed-grinding circuit, which may be caused by high recirculation, can often be traced to
WKHVKDUSQHVVRUE\SDVVRIWKHSDUWLWLRQFXUYH,QZHWJULQGLQJFLUFXLWVK\GURF\FORQHV
UDNH DQG VSLUDO FODVVLÀHUV DQG ÀQH VFUHHQV DUH FRPPRQO\ XVHG DV FODVVLÀHUV EXW
LQ GU\ FLUFXLWV DLU F\FORQHV DOVR UHIHUUHG WR DV DLU VHSDUDWRUV KDYH OLWWOH FRPSHWLWLRQ
7KHGLVFXVVLRQRIFODVVLÀHUVLQWKLVFKDSWHUZLOOUHIHUWRWKHLUSDUWLWLRQFXUYHV

TROMMELS
7KH ZRUG ¶WURPPHO· LV *HUPDQ IRU ¶GUXP· 7URPPHOV DUH F\OLQGULFDO VWUXFWXUHV ZLWK
VFUHHQLQJ PHGLD IRUPLQJ WKH LQWHULRU VXUIDFH KHUHLQ UHIHUUHG WR DV WKH ¶ZRUNLQJ
VXUIDFH· 7URPPHOV FDQ EH PRXQWHG GLUHFWO\ WR WKH GLVFKDUJH HQG RI D JULQGLQJ PLOO
W\SLFDOIRUFRSSHUJROGDQGRWKHUPLQLQJLQGXVWULHVRUWKH\FDQEHPRXQWHGDQGXVHG
independently – sometimes portably – for many applications such as recycling plants.
Trommels come in a wide variety of sizes, as small as a few centimetres in diameter up
WRDVODUJHDVRUP/HQJWKVRIWURPPHOVDUHW\SLFDOO\EHWZHHQDQGWLPHVWKH
working surface diameter. The structural designs depend greatly on how the trommel
LVPRXQWHG,QPRVWFDVHVWKH\DUHPRXQWHGDWWKHLUIHHGHQGWRWKHGLVFKDUJHRIWKH
PLOO FUHDWLQJ D FDQWLOHYHUHG VWUXFWXUH WKDW PXVW EH VXIÀFLHQWO\ VWURQJ WR VXSSRUW
the trommel’s weight plus the weight of the slurry within it for the design life of the
trommel. Typical fatigue analyses are done for a fatigue life of seven years or more.

Comminution Handbook | Spectrum Series 21 147

 CHAPTER 10 t CLASSIFIERS

The wet surfaces of the trommel structure should be protected from wear by bonded
rubber lining or replaceable protectors.
$ WURPPHO·V ZRUNLQJ VXUIDFH ² WKH DFWXDO VFUHHQLQJ PHGLXP XVHG ² YDULHV JUHDWO\
GHSHQGLQJ RQ WKH QHHGV RI WKH DSSOLFDWLRQ /DUJH 6$* PLOO WURPPHOV PRVW RIWHQ XVH
modular moulded rubber panels. Ball mill and rod mill trommels typically use modular
SRO\XUHWKDQHSDQHOV)LJXUHSUHVHQWVVRPHH[DPSOHVRIWURPPHOV
6PDOO WURPPHOV ZLOO IUHTXHQWO\ XVH ZRYHQ ZLUH VFUHHQ ZHOGHG LQWR SODFH WKRXJK
modular polyurethane is gaining popularity. Modularity allows panels in high wear
]RQHVIHHGHQGWREHFKDQJHGZKLOHSDQHOVLQOHVVHUZHDUDUHDVFDQEHOHIWLQORQJHU
7KHZRYHQZLUHRSWLRQLVW\SLFDOO\RQO\XVHGZKHQWKHWURPPHO·VH[SHFWHGOLIHLVVKRUW
and the entire structure is replaced when worn out.
Ball mill and rod mill trommels are used to separate the slurry from the fragmented
JULQGLQJPHGLDUHPQDQWVEDOOVRUURGVWKDWPDNHWKHLUZD\LQWRWKHWURPPHO7KHVOXUU\
passes through the screening media while the remnants report to the discharge end.
7KLVLVGRQHSULPDULO\WRSURWHFWWKHGRZQVWUHDPSXPSVF\FORQHVIURPGDPDJHFDXVHG
E\WKHVWHHOUHPQDQWV6LQFHWURPPHOVDUHW\SLFDOO\PRXQWHGZLWKWKHLUD[LVKRUL]RQWDO
DGYDQFLQJPHFKDQLVPVFDOOHGVFUROOVDUHVRPHWLPHVDWWDFKHGWRWKHVFUHHQLQJPHGLD
to push the unwanted material from the feed end to the discharge end while the slurry
passes through the slotted openings in the screening panels. The direction of rotation is
something that must be known before designing a ball mill trommel as this determines
what orientation the scrolls must have to properly advance the materials.
6$*WURPPHOVDUHJHQHUDOO\XVHGWRVHSDUDWHODUJHURUHSDUWLFOHVWKDWDUHWRRELJIRU
the downstream ball or rod mills from the smaller ore particles that can be processed
GRZQVWUHDP7KHVPDOOHUSDUWLFOHVRUWKURXJKVJRWKURXJKWKHVFUHHQLQJPHGLDZKLOH
WKHODUJHUSDUWLFOHVRYHUVDUHHMHFWHGRXWRIWKHGLVFKDUJHHQGIUHTXHQWO\JRLQJWRD
FUXVKHU RU UHFLUFXODWHG EDFN LQWR WKH 6$* PLOO 6$* WURPPHOV RIWHQ KDYH UHWDUGLQJ
dams attached to the screening media to slow the material, allowing near-size and
XQGHUVL]HSDUWLFOHVWRGURSWKURXJKWKHVFUHHQLQJPHGLD,QFRPLQJRUHPRYHVWKHODUJHU
SDUWLFOHVWRZDUGDQGRXWRIWKHGLVFKDUJHHQG$VXFFHVVIXOWURPPHOGHVLJQZLOOODVWWKH
desired service life, enable easy installation and removal, allow easy maintenance of the
screening media and be sized to support the application’s production needs.
Trommels are usually washed with high-pressure water sprays. The sprays can be
mounted internally to promote material passing through to the undersize, or mounted
H[WHUQDOO\WRNHHSDSHUWXUHVFOHDQDQGIUHHPDWHULDOWRUHSRUWWRWKHRYHUVL]H

A B C

FIG 10.4 – Trommels. (A) Typical large trommel structure showing rings, structural tubes, diagonal supports and
circumferential modular panel fastening provisions; (B) typical ball mill trommel with advancing scroll panel
layout; (C) typical semi-autogenous trommel structure with flow-retarding dams (images courtesy Polydeck).

148 Comminution Handbook | Spectrum Series 21

 CHAPTER 10 t CLASSIFIERS

7URPPHOVDUHVRPHZKDWXQLTXHDPRQJVWFODVVLÀHUVLQWKDWWKH\FDQEHGHVLJQHGWR
JHQHUDWHPXOWLSOHSURGXFWVLQDVLQJOHSDVV$WURPPHOFDQEHÀWWHGZLWKVPDOOSDQHOV
DWWKHIHHGHQGDQGKDYHSURJUHVVLYHO\ODUJHUSDQHOVPRYLQJWRWKHGLVFKDUJHHQG6XFK
a trommel will usually have one product partition below each screen aperture size with
the smallest particles being collected closest to the feed end. This sort of trommel is
XVHIXOIRUVRUWLQJSHEEOHVLQWRVL]HUDQJHVDIWHU6$*PLOOLQJ

RAKE AND SPIRAL CLASSIFIERS

5DNH DQG VSLUDO FODVVLÀHUV HDFK FRQVLVW RI D UHFWDQJXODU WDQN ZLWK D VORSLQJ EDVH DQG
a reciprocating rake or rotating spiral. This moves sands from the sedimentation zone
up the slope to a discharge launder from which they are returned to the mill for further
JULQGLQJ 7KH ÀQH SDUWLFOHV RYHUÁRZ IURP WKH VHGLPHQWDWLRQ ]RQH DQG PRYH RQWR
further processing.
5DNHFODVVLÀHUVRIWKHW\SHVKRZQLQ)LJXUHZHUHWKHÀUVWFRQWLQXRXVFODVVLÀHUV
ZKHQWKH\ZHUHLQYHQWHGE\-91'RUULQDQGZLGHO\XVHGLQFRPPLQXWLRQFLUFXLWV
XQWLOK\GURF\FORQHVZHUHGHYHORSHG\HDUVODWHU6SLUDOFODVVLÀHUVZHUHDOVRXVHGLQ
pre-hydrocyclone days because they created less turbulence and gave a more precise
VHSDUDWLRQ 'LDPHWHUV RI VSLUDO FODVVLÀHUV XVHG WRGD\ DUH XS WR  PP DQG ODUJH
PDFKLQHVXVHWZRSDUDOOHOVSLUDOVLQRQHWDQN7KHFXWSRLQWVRIUDNHDQGVSLUDOFODVVLÀHUV
DUHFRQWUROOHGE\WKHSHUFHQWVROLGVLQWKHIHHGDQGDUHUDUHO\OHVVWKDQǍP7KH\DUH
XVHGLQSULPDU\RUVHFRQGDU\JULQGLQJLQSODFHRIK\GURF\FORQHVDQGÀQHVFUHHQVEXW
they are easy to operate and will work for long periods with only basic maintenance.
$QH[DPSOHLVWKHUDNHFODVVLÀHUVWKDWZHUHLQVWDOOHGLQOHDG]LQFFLUFXLWVDW%URNHQ+LOO
$XVWUDOLDLQZKLFKRSHUDWHGIRUPRUHWKDQ\HDUV
$OWKRXJK WKH\ SURYLGH EHWWHU FODVVLÀFDWLRQ UDNH FODVVLÀHUV KDYH IDOOHQ RXW RI XVH
DQG KDYH EHHQ ZLGHO\ UHSODFHG E\ K\GURF\FORQHV ZKLFK KDYH D VLJQLÀFDQWO\ VPDOOHU
‘footprint’ and higher utilisation. With power failures, it could take days to unbog a
UDNHFODVVLÀHU2QWKHRWKHUKDQGVSLUDOFODVVLÀHUVÀQGDZLGHDSSOLFDWLRQSDUWLFXODUO\
LQ &KLQHVH SURFHVVLQJ RSHUDWLRQV DQG DUH JHQHUDOO\ OLPLWHG WR VPDOOHU WKURXJKSXW
operations.

A B

FIG 10.5 – Spiral and rake classifiers. (A) Spiral classifier (image courtesy Metso), (B) rake classifiers (image courtesy FLSmidth).

Comminution Handbook | Spectrum Series 21 149

 CHAPTER 10 t CLASSIFIERS

Other classifiers
Many devices have been developed to separate particles that fall at different rates through
ZDWHUEHFDXVHRIGLIIHUHQWVL]HVRUVSHFLÀFJUDYLWLHVKRZHYHURQO\K\GURF\FORQHVUDNH
VSLUDOFODVVLÀHUVDQGÀQHVFUHHQVKDYHEHHQUHJXODUO\XVHGLQFORVHGÀQHJULQGLQJFLUFXLWV
$ UHFHQW GHYLFH LV WKH 5HÁX[TM &ODVVLÀHU ZKLFK VHSDUDWHV VPDOO SDUWLFOHV W\SLFDOO\
PPZLWKRXWWKHXVHRIH[FHVVLYHZDWHU7KLVLQYROYHVDSURFHVVLQDPL[LQJFKDPEHU
LQZKLFKSDUWLFOHVDUHVRUWHGE\WKHIRUFHRIJUDYLW\DQGWKHXSÁRZRIÁXLGLVDWLRQZDWHU
&RDUVHSDUWLFOHVPLJUDWHWRDFHQWUDOXQGHUÁRZYDOYHIURPZKLFKWKH\DUHGLVFKDUJHG
)LQH SDUWLFOHV PLJUDWH WR WKH ODPHOOD VHFWLRQ ZKHUH WKH FKDQQHOV HQKDQFH WKH VHWWOLQJ
rate of misplaced high-density solids that slide down the channels and return to the
PL[LQJ FKDPEHU 7KH RYHUÁRZ FRQWDLQV PRVW RI WKH SURFHVV ZDWHU DQG WKH ÀQH ORZ
density solids. Typical capacity of a large unit for treating iron ore particles in the size
UDQJHWRPPDVZHOODVRWKHULQGXVWULDOPLQHUDOVLVWK

HYDROCYCLONES
%\SXPSVKDGEHFRPHDYDLODEOHWKDWFRXOGPRYHPLQHUDOVOXUULHVDWKLJKYHORFLW\
DQGWKHÁRZVFRXOGEHFRQYHUWHGLQWRURWDU\PRWLRQE\GLUHFWLQJWKHPWKURXJKDWDQJHQWLDO
HQWU\LQWRWKHWRSRIWKHF\OLQGULFDOVHFWLRQRIDF\FORQH7ZRVSLUDOVZHUHIRUPHGDQRXWHU
spiral that moved down the cylinder wall and an inner reverse spiral that moved towards
WKHWRSRIWKHF\OLQGHU+HDY\SDUWLFOHV²E\VL]HRUVSHFLÀFJUDYLW\6*²ZHUHKHOGLQWKH
outer spiral by centrifugal action and discharged through the conical base of the cyclone
DSH[/LJKWHUSDUWLFOHVHQWHUHGWKHLQQHUVSLUDODQGZHUHGLVFKDUJHGWKURXJKDYRUWH[
ÀQGHU,QWKHV.HOO\.UHEVRI(TXLSPHQW(QJLQHHUVGHVLJQHGWKHLQYROXWHHQWU\WR
UHSODFHWKHWDQJHQWLDOHQWU\DQGWKHUHGXFHGWXUEXOHQFHDQGLPSURYHGHIÀFLHQF\PDGH
K\GURF\FORQHVWKHFODVVLÀHURIFKRLFHIRUZHWJULQGLQJFLUFXLWV)LJXUH7KHFXUUHQW
LQYROXWHGHVLJQKDVUHGXFHGWXUEXOHQFHIXUWKHU+\GURF\FORQHVPD\EHVLPSOHLQFRQFHSW
EXWWKHVHGD\VDUHLQHIÀFLHQWLQSUDFWLFHIRUWZRUHDVRQV

A B

FIG 10.6 – The hydrocyclone. (A) Hydrocyclone nomenclature (image courtesy

Hakan Dundar), (B) involute feed entry – 1960 and 2013 (image courtesy FLSmidth).

150 Comminution Handbook | Spectrum Series 21

 CHAPTER 10 t CLASSIFIERS

 WKHDSH[VWUHDPWKDWFDUULHVFRDUVHSDUWLFOHVZLOODOVRFDUU\VRPHÀQHSDUWLFOHVEHFDXVH
these are inevitably present in the water
 WKHYRUWH[ÀQGHUVWUHDPFDUU\LQJWKHÀQHSDUWLFOHVZLOODOVRFDUU\VRPHFRDUVHSDUWLFOHV
WKDWKDYHDORZ6*RUKDYHEHHQPLVSODFHGE\WXUEXOHQFHRUURSLQJDOWKRXJKVFUROOHG
involute designs tend to minimise misreporting of coarse particles.
+\GURF\FORQHV DUH HDV\ WR XVH DQG FDQ KDQGOH WKH ZLGH UDQJH RI ÁRZV UHTXLUHG LQ
FORVHGJULQGLQJFLUFXLWVKHQFHGHVSLWHWKHLULQHIÀFLHQFLHVWKH\ZLOOFRQWLQXHWREHXVHG
DVFODVVLÀHUVIRUPDQ\\HDUV
7KH FDSDFLW\ RI WKH K\GURF\FORQH LV OLPLWHG E\ WKH DELOLW\ RI WKH DSH[ WR GLVFKDUJH
VROLGV,IWKHORDGLVWRRKLJKWKHFRQWHQWRIWKHXQGHUÁRZLQFUHDVHVDQGWKHXPEUHOOD
VKDSHGÁRZEHFRPHVDURSHOLNHVWUHDPRIGHQVHSXOS7KHSHUIRUPDQFHGHWHULRUDWHV
ZLWK FRDUVH SDUWLFOHV EHLQJ GLVSODFHG LQWR WKH ÀQH VSLUDO :KHQ D K\GURF\FORQH LV
RSHUDWLQJSURSHUO\WKHGLVFKDUJHVKRXOGÁDUHRXWLQDQXPEUHOODVKDSH:LWKVLOLFHRXV
RUHVURSLQJRFFXUVDWDERXW²SHUFHQWVROLGV
0DQ\ SDUDPHWHUV LQFOXGLQJ GHVLJQ DQG RSHUDWLQJ KDYH DQ LQÁXHQFH RQ WKH
performance of the hydrocyclones. Those parameters are designed or adjusted with
UHVSHFW WR WDUJHW SURGXFW VL]H LQ JULQGLQJ FLUFXLWV IRU D JLYHQ FDSDFLW\ 7DEOH 
describes the interactions among the cyclone parameters.
Preliminary selection of hydrocyclones is usually done using manufacturers’ charts,
ZKLFKW\SLFDOO\FRQWDLQDUDQJHRIFXWVL]HVDVVRFLDWHGZLWKHDFKK\GURF\FORQH$FKDUW
IURP0LQFR7HFKLVVKRZQLQ)LJXUH7DEOHSURYLGHVVLPLODULQIRUPDWLRQIURP
Metso.
:KHQGHVLJQLQJDK\GURF\FORQHLQVWDOODWLRQWKHDSSUR[LPDWHFXWSRLQWDQGFLUFXODWLQJ
load are estimated by the circuit designer. Manufacturers’ charts are then used to make
a tentative selection of the size and number of the hydrocyclones and simulation can be
FDUULHGRXWWRDFKLHYHÀQDOVHOHFWLRQRIWKHK\GURF\FORQHFOXVWHU
:KHUHRSHUDWLQJÁH[LELOLW\DQGFRQWUROLVQHHGHGWKHQXPEHURIQRUPDOO\RSHUDWLQJ
F\FORQHV VKRXOG EH JUHDWHU WKDQ ÀYH RU VL[ )HZHU F\FORQHV WKDQ ÀYH ZLOO OHDG WR

TABLE 10.1
Interactions between hydrocyclone parameters.
Parameter Influence
Cyclone diameter Cyclone diameter refers to the inside diameter of the cylindrical part at the top. Larger cyclones result
in a low centrifugal force, so a coarser cut size is obtained.
Apex diameter The apex is the most widely replaced part of the cyclone to adjust the cut size. The smaller the apex
diameter, the coarser the cut size. A smaller apex also leads to a low water split to underflow, thus the
density of the underflow is increased.
Vortex finder diameter A smaller vortex finder results in reduced cut size and capacity. The larger the vortex finder, the lower
the water split to underflow.
Inlet pressure Higher inlet pressure leads to higher centrifugal force inside the cylindrical feed chamber, hence
reduced cut size and water split to underflow. The pressure can be increased by switching off one or
more cyclones in the cluster, so a number of the running cyclones determines the pressure.
Feed viscosity Increased viscosity (such as being due to the presence of clay) can dramatically increase the cyclone cut size.
Feed solids specific gravity (SG) High SG solids will cut at a finer size than low SG solids in the same cyclone conditions. In a mixed ore,
quartz and magnetite will have strikingly different partition curves.

Comminution Handbook | Spectrum Series 21 151

 CHAPTER 10 t CLASSIFIERS

FIG 10.7 – The links between hydrocyclone capacity and diameter, operating pressure and d50 (image courtesy Minco Tech).

TABLE 10.2
The links between hydrocyclone diameter, d50, and volumetric capacity (Metso).
Diameter (mm) d50 microns Capacity (US gallons/min) (L/min)
40 5–14 2.5–6 9.5–22.7
100 9–32 20–50 75.7–189.3
250 13–52 50–170 189.3–643.5
350 20–74 100–300 378–1134
420 23–83 200–500 757.1–1893
600 26–95 290–800 1098–3028
900 31–109 700–2100 2650–7949
1200 38–120 1100–3100 4164–11 735

H[FHVVLYHFKDQJHVLQRSHUDWLQJSUHVVXUHDQGÁRZZKHQDVLQJOHF\FORQHLVWXUQHGRQ
or off to enact control.
+\GURF\FORQHVFDQQRWDFKLHYHFXWSRLQWVFRDUVHUWKDQDERXW—PXQOHVVWKHUHDUH
RUHFKDUDFWHULVWLFVVXFKDVORZ6*RUÁDWSDUWLFOHVKDSH
+\GURF\FORQHV LQ FORVHG FLUFXLW FDQQRW EH FRQVLGHUHG ZLWKRXW XQGHUVWDQGLQJ WKH
FRPPLQXWLRQ XQLW DV ZHOO VXFK DV EDOO PLOO SHEEOH PLOO RU URG PLOO )RU H[DPSOH
FKDQJLQJ DSH[ RU YRUWH[ ÀQGHU GLPHQVLRQVKDV OLWWOH HIIHFW RQ WKH XOWLPDWH JULQG VL]H

152 Comminution Handbook | Spectrum Series 21

 CHAPTER 10 t CLASSIFIERS

because the hydrocyclone changes that make separation coarser also make grinding
PRUHHIÀFLHQWDQGYLFHYHUVD7KLVLVZK\K\GURF\FORQHEDOOPLOORSWLPLVDWLRQLVEHVW
conducted by simulation.
7KH FODVVLÀFDWLRQ LV DIIHFWHG E\ SDUWLFOH 6* SDUWLFOH VKDSH DQG VOXUU\ YLVFRVLW\
$SDUWLFOHZLWKDKLJKHU6*QHHGVWREHPXFKÀQHULQRUGHUWRUHSRUWWRWKHK\GURF\FORQH
RYHUÁRZWKDQDSDUWLFOHZLWKDORZHU6*$VDQH[DPSOHIRUDK\GURF\FORQHZLWKDQ
RYHUÁRZFXWVL]HRIP80 —PEDVHGRQWKHJDQJXHWKDWLVW\SLFDOO\VLOLFDEDVHGDQG
KDVDQ6*RIDJDOHQDSDUWLFOHZRXOGQHHGWRKDYHDP80 of 44 μm to report to the
RYHUÁRZZKLOHIRUJROGWKHP80ZRXOGEH—P
&RQVHTXHQWO\PLQHUDOVVXFKDVEDVHPHWDOVXOÀGHVWHQGWREHFRPHRYHUJURXQGDV
WKH\ F\FOH DURXQG WKH PLOOLQJ FLUFXLW XQWLO ÀQH HQRXJK WR UHSRUW WR WKH K\GURF\FORQH
RYHUÁRZ7KLVLVZK\EDVHPHWDOVXOÀGHÁRWDWLRQÀQDOFRQFHQWUDWHVDUHVRPXFKÀQHU
WKDQ WKH SULPDU\ JULQG VL]H DQG IRU H[DPSOH OHDG ÁRWDWLRQ ÀQDO FRQFHQWUDWHV DUH
W\SLFDOO\ OHVV WKDQ  —P 7R WKDW HQG VFUHHQV KDYH EHHQ HPSOR\HG LQ FRPPLQXWLRQ
FLUFXLWV WUHDWLQJ KHDY\ DQG EULWWOH PLQHUDOV VXFK DV JDOHQD LQ RUGHU WR PLQLPLVH
RYHUJULQGLQJZLWKVRPHH[FHOOHQWUHVXOWVVHHWKHVHFWLRQRQKLJKIUHTXHQF\VFUHHQV
)LQDOO\ WKLV SURSHUW\ RI K\GURF\FORQHV WR FRQFHQWUDWH KHDYLHU PLQHUDOV LQ WKH
XQGHUÁRZ LV H[SORLWHG LQ FRPPLQXWLRQ FLUFXLWV WUHDWLQJ SUHFLRXV PHWDOV WKLV LV
SDUWLFXODUO\WKHFDVHZLWKJROGZKHUHDSRUWLRQRIWKHK\GURF\FORQHXQGHUÁRZWHQWR
SHUFHQWLVWUHDWHGLQDJUDYLW\FLUFXLW

FINE SCREENS
)LQHVFUHHQVDUHDOWHUQDWLYHVWRK\GURF\FORQHVIRUVHSDUDWLQJVPDOOSDUWLFOHVRQWKHEDVLV
RIVL]H)RUPDQ\\HDUVWKH\KDYHKDGOLPLWHGXVHLQRUHJULQGLQJFLUFXLWVGXHWRKLJK
VHSDUDWLQJVL]HVDQGORZVFUHHQFDSDFLWLHVKRZHYHUDGYDQFHVLQPDWHULDOWHFKQRORJ\
KDYHDOORZHGWKHPDQXIDFWXUHRIÁDWVFUHHQSDQHOVIURPZLUHYDU\LQJLQVL]HIURP—P
WRPP$Q\VORWVL]HFDQEHVHOHFWHGIRUWKHVFUHHQDQGWKHWROHUDQFHZLOOEH“—P
GHSHQGLQJRQWKHVL]HRIWKHZLUH6FUHHQVDUHW\SLFDOO\XVHGWRPLQLPLVHRYHUJULQGLQJ
DQGGHFUHDVHFLUFXODWLQJORDGVZKHQRUHVFRQWDLQUHODWLYHO\VLJQLÀFDQWOHYHOVRIKLJK6*
DQGEULWWOHPLQHUDOVVXFKDVJDOHQD,QWKLVVHFWLRQVLHYHEHQGVDQGYLEUDWLQJVFUHHQV
including Derrick high frequency screens, will be discussed.

Sieve bends
$VLHYHEHQGFRQVLVWVRIDFRQFDYHFXUYHGVFUHHQPRXQWHGLQDIUDPHZLWKWKHVFUHHQ
RSHQLQJVSHUSHQGLFXODUWRWKHÁRZDVVKRZQLQ)LJXUH+RUL]RQWDOZHGJHZLUHV
form the screen.
$ FXUYHG VFUHHQ KDV D JUHDWHU FDSDFLW\ WKDQ D ÁDW VFUHHQ GXH WR IRUFHV H[HUWHG RQ
PDWHULDOÁRZDJDLQVWWKHFXUYHGVXUIDFH6LHYHEHQGVDUHDOVRNQRZQDV'60VFUHHQV
DIWHUWKH'XWFK6WDWH0LQHVZKRLQWURGXFHGWKHVLHYHEHQGDVDVWDWLFVL]LQJVFUHHQ
IRU ÀQH FRDO VOXUU\ DV LW KDG WHQ WLPHV JUHDWHU FDSDFLW\ WKDQ FRQYHQWLRQDO YLEUDWLQJ
VFUHHQV$VVOXUU\ÁRZVGRZQWKHVFUHHQWKHOHDGLQJHGJHRIHDFKZHGJHZLUHVOLFHV
away a thin layer of water and small-particle solids. Larger particles continue down
WKHVFUHHQVXUIDFHZLWKPXFKRIWKHOLTXLGUHPRYHG6HSDUDWLRQVL]HLVVPDOOHUWKDQ
screen aperture, the size of which will depend on the angle of the sieve bend chosen
for the application.

Comminution Handbook | Spectrum Series 21 153

 CHAPTER 10 t CLASSIFIERS

FIG 10.8 – Sieve bend screening mechanism (image courtesy Hakan Dundar).

The features of sieve bends are:

x small footprint compared to a conventional screen of the same capacity
x modest sizing accuracy – can be used reliably for sizing solid particles down to
200 μm (0.008 inches)
x low maintenance – the wedge wire screen resists pegging and has no moving
parts, minimising maintenance (abrasive solids will reduce screen life and increase
maintenance requirements)
x easy screen replacement – a small screen can be replaced in ten minutes with no tools.
Sieve bends are available with aperture sizes as small as 50 μm, although the minimum
recommended aperture size for static sieve bends is 400 μm. Wear of the wedge wires
UHGXFHV HIÀFLHQF\ EXW WXUQLQJ WKH VLHYH EHQG ƒ ZLOO UHJDLQ WKH RULJLQDO VFUHHQ
performance. This may occur several times before the screen is replaced. Sieve bends are
PRVWHIÀFLHQWDVFODVVLÀHUVZKHQWKHIHHGFRQWDLQVWHQWRSHUFHQWVROLGVE\YROXPH
Variations of the sieve bend are the full circular wedge wire screen often used in South
$IULFDDQGWKHLQFOLQHGÁDWSDQHOZHGJHZLUHVFUHHQVRPHWLPHVXVHGLQWKHPDJQHWLWH
industry.

Vibrating screens

Evolution of screens
6FUHHQLQJ LV WKH SURFHVV RI VHSDUDWLQJ ÀQH SDUWLFOHV IURP FRDUVHU SDUWLFOHV WKURXJK D
screen surface, with the separation size determined by the size of the apertures in the
surface. Screen separation is based primarily on particle size, unlike most other forms
RIFODVVLÀFDWLRQVXFKDVK\GURF\FORQHVZKHUHWKHVHSDUDWLRQLVODUJHO\LQÁXHQFHGE\
particle density. With proper selection of equipment, screens can achieve sharp size
VHSDUDWLRQVZLWKWLJKWWRSVL]HFRQWUROWRWKHXQGHUVL]HDQGPLQLPDOÀQHVE\SDVVWRWKH
RYHUVL]H7KHSUDFWLFHRIVFUHHQLQJLQFRPPLQXWLRQGDWHVEDFNWRWKHGD\VRIPXVFOH
SRZHUHG TXHUQVWRQHV ZKHQ VFUHHQV ZHUH PDGH RI ZRYHQ KD]HO WZLJV IRU FRDUVHU

154 Comminution Handbook | Spectrum Series 21

 CHAPTER 10 t CLASSIFIERS

VHSDUDWLRQVDQGKRUVHKDLUIRUÀQHUVHSDUDWLRQV7KHRYHUVL]HZDVPDQXDOO\UHWXUQHGWR
the querns for further size reduction. Tonnage rates increased when stamp mills began
to be powered by water, wind and then steam, and screen surfaces were constructed
IURP SHUIRUDWHG PHWDO SODWHV IRU GXUDELOLW\ DQG WKHQ IURP ZRYHQ ZLUH PHVK À[HG WR
ZRRGHQIUDPHVIRUÀQHUVHSDUDWLRQV

Wire mesh screens

Wire mesh screen surfaces increased undersize throughput with higher open-area
surfaces. Undersize throughput was also increased by shaking the screen to convey
oversize off the surface and dislodge particles lodged in the surface apertures – a problem
NQRZQ DV EOLQGLQJ $GGLQJ ZDWHU WR FDUU\ WKH ÀQHV WKURXJK WKH VFUHHQ RSHQLQJV
LPSURYHGWKHSURFHVVDQGHOLPLQDWHGWKHUHTXLUHPHQWWRGU\WKHRUHIRUHIIHFWLYHÀQH
screening, but blinding continued to be problematic.
Unique wire mesh weaving methods were developed to reduce blinding, such as
WKH7RQ&DSVFUHHQVXUIDFHVZLWKVORWWHGRSHQLQJVLQWURGXFHGLQE\:67\OHU
(OHFWULF PRWRUV OHG WR WKH LQYHQWLRQ RI YLEUDWLQJ VFUHHQV ZLWK YDULRXV PRWLRQV RQH
RI ZKLFK ZDV WKH 7\OHU +XPPHU VFUHHQ LQWURGXFHG LQ  +XPPHU VFUHHQV ZHUH
used to close ball mill circuits but low capacity and high maintenance continued to be
SUREOHPDWLF%\WKHVWKH'XWFK6WDWH0LQHVZDVXVLQJVLHYHEHQGVIRUGHVOLPLQJ
FRDODQGWKHQWDFRQLWHPLQHVEHJDQH[SHULPHQWLQJZLWKWKHPLQEDOOPLOOFLUFXLWVE\WKH
V'RUU2OLYHUDGGHGUDSSLQJKDPPHUVDQGODWHUSQHXPDWLFYLEUDWRUVWRUHGXFH
EOLQGLQJE\QHDUVL]HSDUWLFOHVEXWFDSDFLWLHVUHPDLQHGORZDQGVHSDUDWLRQHIÀFLHQFLHV
still left much to be desired.

High-frequency screens
7KHVH ZHUH LQWURGXFHG E\ + : 'HUULFN LQ  7KH KLJKIUHTXHQF\ YLEUDWLRQ ZDV
REVHUYHG WR DFFHOHUDWH ÀQH SDUWLFOH VHSDUDWLRQ ZLWK UDSLG SDUWLFOH VWUDWLÀFDWLRQ ZKLOH
maintaining close contact with the screen surface. The development of blinding-resistant
OD\HUHGZLUHPHVK6DQGZLFKVFUHHQVXUIDFHVOHG'HUULFNWRGLVFRYHUZHWVFUHHQLQJFRXOG
be achieved in a relatively short screen length while water passes to undersize, and in
'HUULFNUHOHDVHGWKH0XOWLIHHGVFUHHQLQJPDFKLQHZLWKDÁRZGLVWULEXWRUVSOLWWLQJ
the feed to three short screen decks operating in parallel in a single vibrating frame.
Multifeeds were adopted into grinding circuits, often in conjunction with hydrocyclones
for ores with high-density differences between the valuable and gangue minerals and
HDVLO\ IULDEOH RUHV WKDW DUH GLIÀFXOW WR UHFRYHU DW ÀQHU VL]HV 7KH SUREOHP ZDV KLJK
consumption rates of screen surfaces.
%\ WKH V QHZ PDWHULDOV RI FRQVWUXFWLRQ VXFK DV UXEEHU DQG XUHWKDQH LPSURYHG
VFUHHQ VXUIDFH GXUDELOLW\ KRZHYHU WKHVH PRUH UREXVW VFUHHQ VXUIDFHV RIWHQ VXIIHUHG
IURPEOLQGLQJDQGDORZRSHQDUHD)LQHVL]LQJZLWKXUHWKDQHVFUHHQVXUIDFHVEHFDPH
more practical when Derrick introduced tensioned, blinding-resistant screen surfaces
ZLWKDKLJKRSHQDUHD6LPLODUWR'60VFUHHQVWKHVHVXUIDFHVXWLOLVHGDWDSHUHGUHOLHI
DQJOH WR SUHYHQW EOLQGLQJ EXW WKH DGGHG ÁH[LELOLW\ RI XUHWKDQH DQG D KLJKHU *IRUFH
YLEUDWLQJVFUHHQVXUIDFHGUDVWLFDOO\LPSURYHGWKURXJKSXW%\K\GURF\FORQHVZHUH
EHLQJUHSODFHGE\0XOWLIHHGVZLWKǍPVFUHHQVXUIDFHVRQODUJHUEDOOPLOOFLUFXLWV
,Q'HUULFNUHOHDVHGWKH6WDFN6L]HUVFUHHQ)LJXUHZLWKÀYHVWDFNHGVFUHHQ
GHFNVRSHUDWLQJLQSDUDOOHOZLWKDFRPSDFWIRRWSULQWPRIVFUHHQZLGWKDQGDKLJK

Comminution Handbook | Spectrum Series 21 155

 CHAPTER 10 t CLASSIFIERS

A B
FIG 10.9 – Derrick Stack Sizer screen: (A) operating principle and (B) a Stack Sizer ready for operation (images courtesy Derrick).

frequency low-amplitude linear motion vibration. With durable, non-blinding urethane

Vcreen VurfaceV and improved motor deVignV, Àne Vcreening now requireV minimal
maintenance. 8rethane Vcreen VurfaceV are now available with apertureV down to  Ǎm,
making size separations possible in ranges previously reserved for more traditional
forms of Àne classiÀcation, such as hydrocyclones.

Screen characteristics
Screening equipment should be selected according to the application. Low-frequency
high-amplitude screens are best suited for coarse sizing (>6 mm) applications where
a long stroke is required to dislodge coarse particles from large openings. High-
frequency low-amplitude screens are ideal for Àner separations. 7he low amplitude
allows Àner particles to maintain closer contact with the screen surface for improved
Ànes recovery to undersize. )iner separations also require a thinner bed depth to
reduce Ànes bypass to oversize.
Equipment designs for wet screens are different to dry screens. In wet screening the
Ànes are carried to undersize by water or other Áuid. In general a screen feed density of
20 per cent solids with respect to volume, irrespective of the solids, is recommended.
7his allows for enough water to carry the Ànes to undersize without reducing machine
capacity due to volumetric constraints of the screen drainage rate. In wet screening,
width is more critical than overall screen area. Increased screen width not only reduces
bed depth but for most applications, free water should be drained to undersize within
1 m of screen length. Spray water can be added to improve undersize recovery with
longer screen designs, but without a thin bed, depth Ànes can remain trapped in the
oversize. 7he addition of spray water can also upset the water balance in downstream
operations.
Dry screening requires longer screening as the particles pass to undersize based on the
probability of passing a screen opening. An adequate bed depth should be maintained
to reduce the bouncing of Àne particles. Dry screening requires virtually no moisture

156 Comminution Handbook | Spectrum Series 21

 CHAPTER 10 t CLASSIFIERS

LQ WKH IHHG IRU HIÀFLHQW ÀQHV UHPRYDO (YHQ ORZ OHYHOV RI PRLVWXUH FDQ FDXVH ÀQHV WR
DJJORPHUDWHDQGPLVUHSRUWWRRYHUVL]H6FUHHQLQJEHORZǍPRIWHQUHTXLUHVPRLVWXUH
OHYHOVEHORZSHUFHQW
)DFWRUV WKDW GHWHUPLQH PDFKLQH FDSDFLW\ DUH GU\ VROLGV 6* SDUWLFOH VL]H GLVWULEXWLRQ
SDUWLFOHVKDSHVOXUU\UKHRORJ\YLVFRVLW\SHUFHQWVROLGVIRUZHWSURFHVVHVDQGORRVHEXON
density for dry applications. Manufacturers should be contacted to determine machine
UHTXLUHPHQWV IRU VSHFLÀF DSSOLFDWLRQV 7KLV LV HVSHFLDOO\ WUXH IRU ÀQHU VHSDUDWLRQV RU
PRUHGLIÀFXOWDSSOLFDWLRQVVXFKDVVWLFN\RUÀEURXVPDWHULDO)XOOVFDOHWHVWLQJLVDOZD\V
UHFRPPHQGHGIRUÀQHVFUHHQDSSOLFDWLRQVGXHWRWKHZLGHYDULDELOLW\RIPDFKLQHFDSDFLW\

Comparing high-frequency screens and hydrocyclones in closed

grinding circuits
+LJKIUHTXHQF\VFUHHQVDUHDQDOWHUQDWLYHWRK\GURF\FORQHV,QWKLVVHFWLRQVFUHHQVDQG
K\GURF\FORQHV DUH FRPSDUHG XVLQJ GDWD IURP 2-6& $SDWLW ZKLFK FDUULHV RXW PLQLQJ
DQGSURFHVVLQJRIDSDWLWHQHSKHOLQHRUHVLQWKH.ROD3HQLQVXOD5XVVLD$SDWLWSURGXFHV
XSWRHLJKWPLOOLRQWRQQHV0WSHU\HDURIDSDWLWHFRQFHQWUDWHIURPWZRSODQWV$12)
DQG$12)ZKHUH$12)VWDQGVIRUDSDWLWHQHSKHOLQHRUHGUHVVLQJIDFWRU\7KHUHDUH
WKUHHVL]HVRIEDOOPLOOVLQRSHUDWLRQDQGWKHDYHUDJHIUHVKIHHGUDWHVDUH
 PðPDW²WK$12)
 PðPDW²WK$12)
 PðPDW²WK$12)
&ODVVLÀFDWLRQLQ$12)FLUFXLWVKDVEHHQFDUULHGRXWXVLQJK\GURF\FORQHVDQGVSLUDO
FODVVLÀHUV VKRZQ LQ )LJXUH  EXW WKH KLJKFLUFXODWLQJ ORDG OLPLWHG IHHG UDWHV
7KHSDUWLWLRQFXUYHVIRUWKHVSLUDOFODVVLÀHUDQGWKHSULPDU\K\GURF\FORQHDUHJLYHQLQ
)LJXUH7KHFXUYHIRUWKHK\GURF\FORQHLQGLFDWHVSRRUFODVVLÀFDWLRQDQGVKRZVWKH
UHDVRQIRUWKHKLJKORDG'XULQJ²DWHVWSURJUDPZDVXQGHUWDNHQLQDJULQGLQJ
FLUFXLWLQ$12)WRFRPSDUHK\GURF\FORQHVDQGKLJKIUHTXHQF\VFUHHQV3DUWLWLRQFXUYHV
IRUWKHK\GURF\FORQHVDQGVFUHHQVLQWKHðPPLOOFLUFXLWDUHVKRZQLQ)LJXUH
DQGSDUWLFOHEUHDNDJHUDWHVLQWKHPLOODUHVKRZQLQ)LJXUH7KHSDUWLWLRQFXUYHV
were better with the screens, and breakage rates in the mill were higher because of the
ORZHUFLUFXODWLQJORDG:KHQVFUHHQVUHSODFHGK\GURF\FORQHVLQWKHPðPPLOO
FLUFXLWLQ$12)WKHUHVXOWVZHUHVLPLODU)LJXUHVDQG

FIG 10.10 – Apatit circuit flow sheet in ANOF-2 OJSC Apatit (image courtesy Hakan Dundar and Romke Kuyvenhoven).

Comminution Handbook | Spectrum Series 21 157

 CHAPTER 10 t CLASSIFIERS

ϭϬϬ

ϴϬ

ƉĂƌƚŝƚŝŽŶй
ϲϬ

ϰϬ
ƐƉŝƌĂů
ϮϬ
,ϭϬϬϬ
Ϭ
Ϭ͘Ϭϭ Ϭ͘ϭ ϭ ϭϬ ϭϬϬ
ŵĞĂŶƉ͘ƐŝnjĞ;ŵŵͿ

FIG 10.11 – Partition curves for spiral classifier and the main cyclone
at ANOF-2 (image courtesy Hakan Dundar and Romke Kuyvenhoven).

ϭϬϬ
ƐĐƌĞĞŶͲϮϵϳƚͬŚ&ĞĞĚ
ĐLJĐůŽŶĞͲϰϰϳƚͬŚ&ĞĞĚ
ϴϬ
ƉĂƌƚŝƚŝŽŶй

ϲϬ

ϰϬ

ϮϬ

Ϭ
Ϭ͘Ϭϭ Ϭ͘ϭ ϭ

FIG 10.12 – Partition curves for hydrocyclones and screens in the 3.6 × 4.0 m
mill circuit in ANOF-2 (image courtesy Hakan Dundar and Romke Kuyvenhoven).

ϭϬϬ
DͲƐĐƌĞĞŶͬϮϵϳƚͬŚDĨĞĞĚ
DͲĐLJĐͬ͘ϰϰϳƚͬŚDĨĞĞĚ

ϭϬ
ƌͬĚ

Ϭ͘ϭ
Ϭ͘ϭ ϭ Ɖ͘ƐŝnjĞ;ŵŵͿ ϭϬ ϭϬϬ

FIG 10.13 – Breakage rates in the 3.6 × 4.0 m ball mill in ANOF-2 when cyclones
and screens were used (image courtesy Hakan Dundar and Romke Kuyvenhoven).

7KLVFDVHVWXG\FRQÀUPHGRWKHUVWXGLHVWKDWVKRZHGWKDWEHWWHUSDUWLWLRQFXUYHVGXH
to screens led to smaller circulating loads, higher breakage rates, higher feed rates and
PRUH HIÀFLHQW FLUFXLWV 6FUHHQV GR KDYH GLVDGYDQWDJHV KRZHYHU WKH\ DUH ODUJHU DQG
KDYHDKLJKHUFRVWWKDQF\FORQHVÁRZGLVWULEXWRUVDUHH[SHQVLYHEHFDXVHWKH\UHTXLUH
FDUHIXOGHVLJQDQGFRQVWUXFWLRQDSHUWXUHVFDQEHEORFNHGE\ÀEUHVDQGWKHÁRZWRD
PDFKLQHVKRXOGQRWH[FHHGLWVFDSDFLW\RWKHUZLVHWKHSHUIRUPDQFHGHWHULRUDWHV

158 Comminution Handbook | Spectrum Series 21

 CHAPTER 10 t CLASSIFIERS

ϭϬϬ
ƐĐƌĞĞŶ
ϴϬ ĐLJĐůŽŶĞ

ƉĂƌƚŝƚŝŽŶй
ϲϬ

ϰϬ

ϮϬ

Ϭ
Ϭ͘Ϭϭ Ϭ͘ϭϬ ϭ͘ϬϬ
ŵĞĂŶƉ͘ƐŝnjĞ;ŵŵͿ 
FIG 10.14 – Partition curves for hydrocyclones and screens in the 5.5 × 6.5 m
ball mill circuit in ANOF-3 (image courtesy Hakan Dundar and Romke Kuyvenhoven).
ϭϬϬ
DͲƐĐƌĞĞŶͬϰϬϬƚͬŚDĨĞĞĚ
DͲĐLJĐůŽŶĞͬϯϮϬƚͬŚDĨĞĞĚ

ϭϬ
ƌͬĚ

Ϭ͘ϭ
Ϭ͘ϭ ϭ ϭϬ ϭϬϬ
Ɖ͘ƐŝnjĞ;ŵŵͿ

FIG 10.15 – Breakage rates in the 5.5 × 6.5 m ball mills in ANOF-3 when cyclones
and screens were used (image courtesy Hakan Dundar and Romke Kuyvenhoven).

AIR CLASSIFIERS
7KH SULQFLSOH RI VHSDUDWLQJ ÀQH SDUWLFOHV LQ DLU OLHV LQ WKH XVH RI FHQWULIXJDO IRUFH
JHQHUDWHGE\DYRUWH[ZLWKVWDWLFFODVVLÀHUVRUDURWDWLQJWXUELQHZLWKG\QDPLFFODVVLÀHUV
,QDJULQGLQJFODVVLI\LQJFLUFXLWRQHFDQDVVXPHWKDWIRUSDUWLFOHVL]HVRIOHVVWKDQ—P
DSSUR[LPDWHO\WKHVDPHDPRXQWRIHQHUJ\LVQHFHVVDU\IRUFODVVLÀFDWLRQDVLVUHTXLUHG
for grinding.
,Q WKLV VHFWLRQ WKH W\SHV RI DLU FODVVLÀHUV WKDW DUH LQ XVH DUH VXPPDULVHG EULHÁ\
7KHPRVWSRSXODULVWKHWKLUGJHQHUDWLRQKLJKHIÀFLHQF\FODVVLÀHU$LUFODVVLÀHUVDUHPRVW
ZLGHO\XVHGLQFHPHQWJULQGLQJ,QVWDWLFDLUFODVVLÀHUVWKHUHDUHQRPRYLQJSDUWVDQG
WKHDGMXVWPHQWRIWKHFXWVL]HLVFRPSOHWHO\GHSHQGHQWRQWKHDLUÁRZUDWH,QG\QDPLF
FODVVLÀHUVWKHUHLVDURWDWLQJWDEOHRUDURWRUFDJHGHSHQGLQJRQWKHJHQHUDWLRQHQKDQFLQJ
WKH VHSDUDWLRQ HIÀFLHQF\ RI WKH FODVVLÀHU 7DEOH  RXWOLQHV WKH HIIHFW RI GHVLJQ DQG
RSHUDWLQJSDUDPHWHUVRIWKHDLUFODVVLÀHUVRQWKHLUSHUIRUPDQFH

Static air classifiers

6WDWLFDLUFODVVLÀHUVDUHGHVLJQHGWRVHSDUDWHWKHSDUWLFOHVE\PHDQVRIWKHDLUÁRZRQO\
7KHYHORFLW\DQGGLUHFWLRQRIWKHDLUÁRZGHWHUPLQHVWKHHIÀFLHQF\DQGFXWVL]HRIWKH
VHSDUDWLRQ 6LQFH WKHUH DUH QR PRYLQJ SDUWV LQVLGH WKH VWDWLF DLU FODVVLÀHU WKH\ DUH
UHODWLYHO\LQH[SHQVLYHDQGHDV\WRRSHUDWHEXWSRRULQHIÀFLHQF\

Comminution Handbook | Spectrum Series 21 159

 CHAPTER 10 t CLASSIFIERS

TABLE 10.3
Effect of parameters on air classifier performance.
Parameter Influence
Rotor size Rotor size determines the capacity that the air classifier can handle.
Rotor speed As the rotor spins faster, finer cut size is achieved while circulating load increases.
Airflow rate The airflow rate has a major effect on the bypass to the coarse product. More air reduces the bypass but coarsens the product.
Feed rate Increased feed rate to the separator leads to higher bypass and lower cut size.

Conical air classifiers

7KHVFKHPDWLFYLHZRIWKHFRQLFDOW\SHVWDWLFVHSDUDWRULVJLYHQLQ)LJXUH7KHIHHG
PDWHULDOHQWHUVWKHVHSDUDWRUZLWKWKHDLUÁRZIURPWKHERWWRPDQGLVGLUHFWHGLQVLGHWKH
VHSDUDWLQJFKDPEHUE\WKHVXUURXQGLQJEODGHV$YRUWH[LVJHQHUDWHGZLWKLQWKHDLUÁRZ
and the particles are subjected to centrifugal force. The coarse particles move towards
WKH VLGHV RI WKH VHSDUDWLQJ FKDPEHU DQG OHDYH WKURXJK WKH GLVFKDUJH SLSH 7KH ÀQH
particles that are subjected to a relatively low centrifugal force leave the separator from
the top with the air. The guide blades are adjusted to control the centrifugal force acting
on the particles and cut size.

FIG 10.16 – Schematic views of a conical type air classifier (images courtesy Fives FCB).

Zigzag air classifiers

=LJ]DJ FODVVLÀHUV )LJXUH  FDQ EH HLWKHU LQ WKH VKDSH RI D ]LJ]DJ FURVVVHFWLRQ RU
FRQVLVWRIDVHULHVRILQFOLQHGSODWHVLQDYHUWLFDOD[LVRSSRVLWHHDFKRWKHU7KHPDWHULDOLV
IHGIURPWKHWRSRIWKHDLUFODVVLÀHUDQGWKHDLUÁRZVXSZDUGV7KHÀQHSDUWLFOHVDUHFDUULHG
XSZKLOHWKHFRDUVHSDUWLFOHVPRYHGRZQZKHQJUDYLW\SUHYDLOVRYHUDLUÁRZYHORFLW\

160 Comminution Handbook | Spectrum Series 21

 CHAPTER 10 t CLASSIFIERS

FIG 10.17 – Zigzag classifiers (image courtesy Shapiro and Galperin, 2005).

7KH PRVW UHFHQW ]LJ]DJ FODVVLÀHU LV WKH 9VHSDUDWRU WKDW ZDV GHYHORSHG E\ .+'
+XPEROGW :HGDJ $* IRU FORVHGFLUFXLW DSSOLFDWLRQV RI KLJKSUHVVXUH JULQGLQJ UROOV
+3*5)LJXUH0DWHULDOLVIHGDWWKHWRSRIWKHVHSDUDWRUDQGFDVFDGHVWKURXJK
WKHLQFOLQHGSODWHV7KHDLUÁRZIURPRQHVLGHRIWKHVHSDUDWRULVGLUHFWHGWRWKHFDVFDGLQJ
PDWHULDOE\WKHLQFOLQHGSODWHV7KHDLUZLWKÀQHVOHDYHVWKHVHSDUDWRUIURPWKHWRSZKLOH
WKHFRDUVHPDWHULDOLVGLVFKDUJHGDWWKHERWWRP7KH9VHSDUDWRUZDVGHVLJQHGWRKDQGOH
ODUJHDPRXQWVRIPDWHULDODVZHOODVKLJKPRLVWXUHFRQWHQWV

FIG 10.18 – KHD V-separator (image courtesy KHD Humboldt Wedag Group).

Dynamic air classifiers

7KH IHDWXUHV RI G\QDPLF FODVVLÀHUV DUH WKH URWRU ZKLFK GLVWULEXWHV WKH IHHG PDWHULDO
DQGDQLQWHUQDORUH[WHUQDOIDQSURPRWLQJDLUÁRZLQVLGHWKHVHSDUDWRUWRFDUU\WKHÀQH
SDUWLFOHVRXWRIWKHVHSDUDWLQJFKDPEHU'\QDPLFFODVVLÀHUVDUHJURXSHGDFFRUGLQJWR
WKHLUKLVWRULFDOGHYHORSPHQW7KH0XPIRUG0RRGLHVHSDUDWRUZDVSDWHQWHGLQDQG

Comminution Handbook | Spectrum Series 21 161

 CHAPTER 10 t CLASSIFIERS

PDUNHGWKHVWDUWRIÀUVWJHQHUDWLRQG\QDPLFFODVVLÀHUV7KHVHFRQGDQGWKLUGJHQHUDWLRQV
DSSHDUHG LQ  DQG  UHVSHFWLYHO\ 7KH GHYHORSPHQW RI DLU FODVVLÀHUV KDV EHHQ
concerned with the reduction of d50, the improvement of partition curves and a smaller
E\SDVV7KHLPSURYHPHQWLQSHUIRUPDQFHLVUHFRUGHGLQ7DEOHDQG)LJXUH

TABLE 10.4
Improvement in performance of air classifiers.
Static First generation Second generation Third generation
(1890) (1960) (1980)
d50 (μm) Not applicable 20+ 15–20 10–15
Bypass (%) 60+ 50+ 15–35 5–15
Sharpness 0.65–0.75 0.80–0.85 0.85–0.90 0.95–1.20

FIG 10.19 – Typical partition curves for the three generations of air classifiers (image courtesy Fives FCB).

First generation
,QWKHÀUVWJHQHUDWLRQRIG\QDPLFVHSDUDWRUVWKHPDWHULDOLVIHGRQWRWKHGLVWULEXWLRQ
SODWHWKDWURWDWHVRQWKHVDPHVKDIWZLWKWKHIDQEODGHV)LJXUH
7KH DLUÁRZ LV VXSSOLHG E\ WKH LQWHUQDO IDQ 7KH PDWHULDO LV GLVWULEXWHG LQVLGH WKH
VHSDUDWLQJ FKDPEHU DQG WKH ÀQHV DUH FDUULHG XSZDUGV WKURXJK WKH IDQ EODGHV WR WKH
outer cone. The coarse particles that move towards the walls of the inner cone by means
RIFHQWULIXJDOIRUFHVIDOOWRWKHERWWRP7KHÀUVWJHQHUDWLRQDLUFODVVLÀHUVRSHUDWHZLWK
high bypass values due to the poor ventilation supplied by an internal fan inside the
VHSDUDWLQJFKDPEHUWKHUHIRUHWKH\DUHORZLQHIÀFLHQF\

Second generation
7KHVHFRQGJHQHUDWLRQG\QDPLFDLUFODVVLÀHUV)LJXUHDUHDOVRNQRZQDVF\FORQLF
VHSDUDWRUV8QOLNHWKHÀUVWJHQHUDWLRQWKHVHFRQGJHQHUDWLRQKDVDQH[WHUQDOIDQWKDW
VXSSOLHVDKLJKYROXPHRIDLUWRWKHFODVVLÀHU$VLQWKHÀUVWJHQHUDWLRQWKHPDWHULDOLV

162 Comminution Handbook | Spectrum Series 21

 CHAPTER 10 t CLASSIFIERS

FIG 10.20 – Schematic view of a first-generation dynamic air classifier (image courtesy Fives FCB).

FIG 10.21 – Schematic view of a second-generation dynamic air classifier (image courtesy Fives FCB).

fed to the distributing plate and the coarse particles moving towards the walls of the
LQQHUFRQHDUHFROOHFWHGIURPWKHERWWRP7KHÀQHSDUWLFOHVDUHFDUULHGWRWKHF\FORQHV
ZLWKWKHULVLQJDLUVXSSOLHGE\DQH[WHUQDOIDQ$VLQVWDWLFDLUFODVVLÀHUVDYRUWH[RFFXUV
LQWKHF\FORQHVDQGWKHÀQHSDUWLFOHVDUHVHSDUDWHGIURPWKHDLUDQGGLVFKDUJHGIURPWKH
ERWWRPRIWKHF\FORQHV7KHDLUH[LWLQJIURPWKHWRSRIWKHF\FORQHVLVUHFLUFXODWHG

Third generation
$ WKLUGJHQHUDWLRQ DLU FODVVLÀHU LV SUHVHQWHG LQ )LJXUH  DQG LV NQRZQ DV D KLJK
HIÀFLHQF\VHSDUDWRU
7KHFODVVLÀHUIHHGLVGLVSHUVHGE\DVSLQQLQJGLVWULEXWLRQSODWHDQGIDOOVLQWRWKH]RQH
EHWZHHQVWDWLRQDU\JXLGHYDQHVDQGWKHURWRUFDJH7KHJXLGHYDQHVGLUHFWWKHDLUÁRZ
IURPWKHH[WHUQDOIDQVRWKDWLWWUDQVSRUWVWKHGLVSHUVHGPDWHULDOWRWKHURWRUFDJHZKHUH
FODVVLÀFDWLRQRFFXUV7KHDLUÁRZPD\HQWHUWKHFODVVLÀHUIURPWKHERWWRPRUWKHVLGH
7KH FRDUVHU SDUWLFOHV FROOLGH ZLWK WKH FDJH EODGHV DQG VHWWOH RXW RI WKH VWUHDP 6RPH
coarse material does not enter the cage and the stream leaving at the base of the cone
FRQVLVWVRIWKLVPDWHULDODQGWKHFRDUVHSDUWLFOHVIURPWKHFDJH7KHÀQHSDUWLFOHVOHDYH
from the top or base of the unit according to its design and are collected with cyclones

Comminution Handbook | Spectrum Series 21 163

 CHAPTER 10 t CLASSIFIERS

FIG 10.22 – Schematic views of third-generation dynamic air classifiers (images courtesy The Cement Grinding Office).

RUEDJÀOWHUV7KHLPSRUWDQWYDULDEOHVDUHWKHDLUÁRZDQGWKHFDJHVSHHG7KH26HSD®
VHSDUDWRUQRZEXLOWE\)/6PLGWKZDVWKHÀUVWKLJKHIÀFLHQF\VHSDUDWRUDQGZDVEXLOW
LQ-DSDQLQWKHODWHV,WVLPSRUWDQWIHDWXUHVZHUHORZE\SDVVDQGVKDUSVHSDUDWLRQ
'\QDPLF FODVVLÀHUV DUH QRZ PDGH E\ YDULRXV PDQXIDFWXUHUV ZKLOH WKH\ PD\ GLIIHU
VOLJKWO\WKH\DOOLQFRUSRUDWHWKHSULQFLSOHVXVHGLQWKH26HSDXQLW
$FRPSDULVRQZDVFRQGXFWHGRIDWKLUGJHQHUDWLRQFODVVLÀHU3RO\VLXV6HSROZLWKDQ
HDUOLHU JHQHUDWLRQ XQLW E\ VXEVWLWXWLRQ LQ D FRPPLQXWLRQ FLUFXLW ,W VKRZHG D VSHFLÀF
HQHUJ\VDYLQJRIEHWZHHQDQGSHUFHQWDQGDQLQFUHDVHLQRXWSXWEHWZHHQDQG
SHUFHQWLQIDYRXURIWKH6HSROXQLW7KLVXQLWKDGDZKHHOGLDPHWHURIPDQGD
GULYHSRZHURIN:7KHIDQGHOLYHUHGP3KRXURIDLUDQGKDGDN:PRWRU
6FKPLGW)XUFKQHUDQG=DPSLQL

BIBLIOGRAPHY
Allen,7Particle Size Measurement,S&KDSPDQDQG+DOO/RQGRQ
Bradley,'The HydrocycloneS3HUJDPRQ3UHVV2[IRUG
Duhamel,3K&RUGRQQLHU$DQG/HPDLUH'769®WKHKLJKHIÀFLHQF\G\QDPLFFODVVLÀHU
DQGLWVODWHVWGHYHORSPHQWV)&%5HVHDUFK&HQWUH&HPHQWDQG0LQHUDO,QGXVWULHV&5&0

164 Comminution Handbook | Spectrum Series 21

 CHAPTER 10 t CLASSIFIERS

Dündar,+.DOXJLQ$'HOJDGR03DORPLQR$7UNLVWDQO×$$TXLQR%DQG/\QFK$-
6FUHHQV DQG F\FORQHV LQ FORVHG JULQGLQJ FLUFXLWV LQ Proceedings XXVII International Mineral
Processing Congress (IMPC) 2014SDSHU6DQWLDJR*HFDPLQ
Fahlstrom, 3 +  6WXGLHV RI WKH K\GURF\FORQH DV D FODVVLÀHU LQ Proceedings Sixth IMPC,
&DQQHVSS²
Furchner,%DQG=DPSLQL6$LUFODVVLI\LQJLQUllmann’s Encyclopedia of Industrial Chemistry,
VHYHQWKHGLWLRQ:LOH\9&+9HUODJ*PE+ &R:HLQKHLP
Heiskanen,.Particle ClassiÀcationS&KDSPDQDQG+DOO/RQGRQ
Lynch,$-Mineral Crushing and Grinding Circuits, Their Simulation, Optimisation, Design and
ControlS(OVHYLHU$PVWHUGDP
Schmidt, '  +RFKOHLVWXQJVVLFKWHU +LJKHIÀFLHQF\ VHSDUDWRU 6(32/ Zement Kalk Gips,
²
Shapiro,0DQG*DOSHULQ9$LUFODVVLÀFDWLRQRIVROLGSDUWLFOHVDUHYLHZChemical Engineering
and Processing Process IntensiÀcation²
Svarovsky,/HydrocyclonesS+ROW5LQHKDUWDQG:LQVWRQ(DVWERXUQH
Wills, % $ DQG 1DSLHU0XQQ 7 -  Mineral Processing Technology VHYHQWK HGLWLRQ  S
%XWWHUZRUWK+HLQHPDQQ

&DWDORJXHVIRUZHWDQGGU\FODVVLÀHUVDUHDYDLODEOHRQWKHLQWHUQHWIURPVHYHUDOFRPSDQLHVVRPH
RIZKLFKKDYHEHHQUHIHUUHGWRLQWKLVFKDSWHU7KHVHFRPSDQLHVLQFOXGH
x 'HUULFN&RUSRUDWLRQ86$
x )LYHV)&%)UDQFH
x )/6PLGWK'HQPDUN86$
x .+'*HUPDQ\
x 0HWVR)LQODQG
x 0LQFR7HFK$XVWUDOLD
x 7K\VVHQ.UXSS*HUPDQ\
x :HLU0LQHUDOV$XVWUDOLD6FRWODQG

Comminution Handbook | Spectrum Series 21 165

 HOME

Chapter 11

COMMINUTION CIRCUITS FOR

ORES, CEMENT AND COAL
Walter Valery, Hakan Benzer, Cliff King
and Glenn Schumacher

COMMINUTION CIRCUITS FOR ORES

Overview
A mining operation is essentially a series of interconnected processes, with the
performance of each one affecting those downstream. Optimising these processes in
isolation can lead to suboptimal overall performance, so it is important to analyse them
in the context of the whole operation. This is particularly relevant to blasting, crushing
and grinding. lasting is the rst stage of comminution in most mining operations.
t is an energy ef cient and cost effective way of preparing ore for crushing and
grinding in which the ne particles necessary for mineral liberation and separation
are produced.
Improving the overall mine and plant performance can often be achieved by
optimising blast fragmentation, using levels of blasting energy according to the ores
and downstream processes involved. There are many cases where the additional
costs of greater blasting energy and distribution are more than compensated for
by reducing total energy consumption and improving throughput and recovery.
Unfortunately this has resulted in the belief that increasing explosive consumption
necessarily results in optimising the entire value chain (mining and processing).
For true holistic optimisation, blast intensity is not always increased, but is adjusted
to suit the different types of ores as well as the circuit con guration, e uipment, and
installed power downstream. Table 11.1 shows the typical energy consumed and the
cost per tonne of ore breakage processes.

Comminution Handbook | Spectrum Series 21 167

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

TABLE 11.1
Typical energy consumption and cost by unit operation in 2014 (A$).
Specific energy (kWh/t) Cost ($/t)
Drill and blast 0.1–0.25 0.1–0.25
Crushing 0.5–8 0.5–1
Grinding 10–35 2–5

Mine to Mill
The term ‘Mine to Mill’ has been widely used to describe initiatives by mining companies,
explosive suppliers and consultants to optimise mines and plants by changing the blast
intensity. Several ‘Mine to Mill’ initiatives have been implemented over the past 20 years
with varying degrees of success. A range of examples is provided in the AusIMM
Proceedings Mine to Mill Conference by Scott and Morrell (1998).
To be successful, Mine to Mill projects re uire a structured methodology supported by
auditing, surveys, data analysis and training. ene ts include
x maximising throughput (mine and mill)
x better overall process stabilisation and minimum operating cost with minimum
adverse impacts
x developing accurate throughput forecast models to establish optimum blending
strategies
x reducing energy consumption and greenhouse gas emissions.
The methodology developed by Metso Process Technology and Innovation (PTI)
includes
x Scoping – historical information collected on-site to identify problems, bottlenecks
and opportunities.
x Ore characteristics – used to de ne ore domains that have similar fragmentation
properties based on rock structure and strength (see Table 11.2).
x Benchmarking, audits and surveys – audits and surveys are conducted for the key
processes of drill and blast, crushing, grinding, otation leaching and dewatering.
Tracking allows ore source and ore characteristics to be linked to process
performance. The SmartTag™ (Figure 11.1) ore tracking system uses robust passive
radio fre uency transponders ( FI tags) that are detected when they pass antennas
located at critical points in the process ahead of the milling circuit. Each tag has

TABLE 11.2
Ore characterisation measurements.
Rock structure – Rock strength – Floatability –
fragmentation modelling comminution modelling flotation modelling
t
Rock quality designation (RQD) t
Drop weight parameters (A, b, ta) t
Assay measurement
t
Fracture frequency t
SMC Drop Weight Index (DWi) t
Liberation measurement using
t
Joint and plane mapping with Sirovision t
Unconfined compressive strength (UCS) FEI’s Mineral Liberation Analyser or
system t
Point Load Index (PLi) QEMSCAN systems
t
Bond Ball Mill Work Index (BBMWi) t
Assay by size
t
Laboratory flotation tests

168 Comminution Handbook | Spectrum Series 21

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

FIG 11.1 – Ore tracking with SmartTags™ (image courtesy Metso).

a uni ue code that allows it to be tracked from the mine to the primary grinding
mill feed. The data collected are analysed and compared with benchmark data from
mines and processing plants around the world.
x 'eveloping sitespeciÀc process models and simulations – data from benchmarking,
audits, surveys and ore characterisation are used to develop site-speci c models for
key processes, and are integrated to represent the total process.
x Validation and implementation – a detailed plan is developed to implement optimisation
strategies based on mine and plant constraints and cost bene t analysis.
x Sustaining the EeneÀts – incorporating the recommended process changes into
managerial and site operating procedures plus training of operators and engineers
ensures that the bene ts are maintained over the longer term.

Optimising blasting
An important component of the Mine to Mill approach is optimising drill and blast
practices. Better distribution of the energy used during blasting for certain ore types
will often produce a run-of-mine ( OM) si e distribution that has a controlled top
si e and more ne material. Blast intensity is not always increased it is adjusted to
suit different types of ores and provide a consistent and optimised fragmentation. As a
result the primary crusher gap may be reduced since ner OM enables this to be done
without compromising crusher throughput. For semi-autogenous grinding (SAG) mills,
higher throughput may be achieved when the SAG mill feed has as ne a top si e as
possible, the smallest possible amount of 2 to mm intermediate si e material and the
maximum amount of -10 mm nes. This is demonstrated in Figure 11.2.

Comminution Handbook | Spectrum Series 21 169

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

FIG 11.2 – Schematic of ideal semi-autogenous grinding mill feed size distribution (image courtesy Walter Valery).

Alternative comminution technologies – such as high-pressure grinding rolls ( PG )

and stirred mills – and ow sheets are becoming more common due to growing
challenges with regard to supply and cost of energy and water. The holistic Mine to
Mill approach is e ually relevant and important for different technologies and ow
sheet arrangements. There are numerous case studies demonstrating successful
implementation and sustained bene ts from Mine to Mill optimisation for a variety of
operations around the world.

Case study – Compañía Minera Antamina

Antamina is a polymetallic mining complex situated in the central Peruvian Andes
at around 00 m above sea level, producing copper and inc concentrates. It is one
of the main mines in Peru, with a milling capacity that was expanded by 75 per cent
from 2006 to 2012. The advent of harder ores presented a problem, and Antamina
and Metso PTI personnel collaborated to increase mill throughput by integrating and
optimising blast fragmentation, crushing and grinding operations. A comprehensive
review of existing operations at the mine allowed ore sources to be characterised into
ore domains based on rock structure and strength. Blast audits were conducted, and
the blasted material was tracked from the mine through the process using SmartTag™
ore tracking. Two antennas to detect tags were installed one on the primary crusher
product belt and another on the SAG mill feed belt. Process surveys were conducted
while treating the ore from audited blasts, allowing the performance in comminution
circuits to be linked to the ore characteristics and blast fragmentation measured during
the audited blasts.
Metso PTI and Antamina staff also analysed historical data and reviewed the current
operating practices. The data collected were used to calibrate mathematical models of
drilling and blasting, and crushing and grinding operations. The resulting site-speci c
models were linked and simulations were conducted to identify and evaluate optimum
operating strategies for the overall process. The proposed strategies were analysed
in detail with site technical staff and the most cost-effective integrated strategy for
the entire process (mine and plant) was then implemented. This change signi cantly
increased throughput. Following this initial success, further investigation of higher-

170 Comminution Handbook | Spectrum Series 21

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

intensity blasts produced even ner OM fragmentation and subse uently ner feed
to the SAG mill. Additional adjustments were made to the crusher, and SAG and ball
mill circuits to achieve the best overall performance. The additional blasting energy was
more than compensated for by the large energy savings in the crushing and grinding
processes, thus minimising the speci c energy consumption of the entire operation.
At the beginning of the project in 2006, the throughput was 2600 t h with a speci c
energy consumption of 1 k h t. At the end of the project in 2010, the throughput
was 500 t h with a speci c energy of 10.5 k h t. This represents an energy saving of
25 per cent.
More detail on this project can be found in ybinski et al (2011) Samuel, alery and
ybinski (2012) and alery and ybinski (2012).

Comminution circuits for polymetallic sulfide flotation

omminution is applied to polymetallic sul de ores for two reasons
1. to reduce particles to a si e at which they can be processed in otation cells and
particles containing valuable sul de minerals oated
2. to liberate the valuable sul des from waste so that they can be oated in high- uality
concentrates.
SAG and or ball mills are commonly used for the si e reduction of mined ore.
Figure 11. shows a circuit in which an ore containing copper, inc and iron sul des
is oated after grinding to a si e that liberates the copper sul de. This leaves many
composite particles containing inc sul des that re uire further comminution to
improve liberation, which is achieved by oating the copper tailing and grinding
the concentrate to liberate the inc mineral that has reported to it. The high intensity
grinding (HIG) mill (which was described in Chapter 8) is suitable for this purpose.

Cu Flotation
Zn Flotation
SAG Mill
Pyrite Flotation

Ore
Tailing
Thickener
HIG
Mill Py Cleaner
Flotation

Tailings
Py Thickener
Ball Mill Cu Cleaner
Flotation
TC2 Zn Cleaner
Flotation

Py Pressure Filter

Cu Thickener Zn Thickener
Cu Pressure Filter

Py Concentrate

Cu Concentrate Zn Concentrate

FIG 11.3 – Comminution by semi-autogenous grinding, ball and high intensity grinding
mills in a polymetallic sulfide ore flotation circuit (image courtesy Outotec).

Comminution Handbook | Spectrum Series 21 171

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

Fine-grained ores may re uire more than one regrind mill if further liberation is
re uired for particles in several streams. ith complex, ne-grained ores becoming
an important source of metals, and with the development of mills that are able to grind
smaller than 5 μm, this type of circuit and more complex versions of it will become
increasingly important.

Comminution circuits for flotation of low-grade copper sulfide ores

In the circuit shown in Figure 11. , all the ore was ground to the liberation si e of
the copper mineral. If this strategy is applied to a low-grade copper ore, much more
energy would be spent on grinding waste that is liberated early in the process and can
be rejected to tailings, leaving mainly composite particles containing copper minerals.
As shown in Figure 11. , these composite particles can be oated and then reground to
liberate the valuable minerals for further concentration.

Cu Flotation

Ore
HIG Mill Tailings
SAG Mill Thickener

Ball Mill
Cu Cleaner Flotation

Tailings

Concentrate
Thickener

Cu Concentrate
Pressure Filter

FIG 11.4 – Comminution in a typical low-grade copper sulfide flotation circuit (image courtesy Outotec).

COMMINUTION CIRCUITS FOR CEMENT CLINKER

Cement is the major component of the world’s physical infrastructure. Worldwide
production was 3.3 Mt in 2010 and will be 4.2 Mt in 2015. Comminution of the raw
material occurs before the kiln and comminution of the cement clinker after the kiln.
Comminution processes account for more than 60 per cent of the total energy used in the
cement manufacturing process. The nished cement must meet high technical standards
because the structural integrity of the buildings depend on the cement uality. The three-
volume Cement Data Book ( uda, 1985) discusses every aspect of the process in detail.
With cement clinker grinding circuits, the particle si e distribution (PS ) of each
product must be kept at the value that ensures high uality of the nished cement,
which is the critical parameter in cement manufacture. Accurate control of these circuits
is assisted by there being little variation in the clinker leaving the kiln because both the

172 Comminution Handbook | Spectrum Series 21

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

chemical composition of the kiln feed and the kiln operation itself are closely controlled,
reducing disturbances in the mill feed however, variations do occur and an automatic
control system is necessary to recognise and compensate for them.

Single-stage ball milling

Single-stage ball milling is the conventional way of producing nished cement.
A two-chamber ball mill, seen in Figure 11. 5, is used in a closed circuit with a dynamic
air classi er in the rst chamber coarser balls are used for breakage and in the second
chamber ner balls are used. The intermediate grate separating the chambers allows
smaller particles to pass to the second chamber – generally 6–8 mm slot si es are used.
Ground material exits the mill by over ow, although some nes are contained in the
discharge air. The mill lter extracts the nes from the air and this joins the over ow
to feed a dynamic air classi er. Classi er nes is the circuit product, while the coarse
product is recycled to mill for further grinding. In some cases, lter return can enter the
nal product depending on the particle si e.

FIG 11.5 – Single-stage ball mill and separator (image courtesy Hakan Dundar).

Open-circuit pregrinding with high-pressure grinding rolls followed by

single-stage ball milling
Figure 11.6 shows an HPG situated ahead of ball milling to provide pregrinding.
In most of the HPG applications in cement, a portion of the HPG product is recycled
to its feed to increase the nes amount in the HPG , improving the ef ciency of the
compression. In this circuit con guration, the HPG runs in open-circuit mode and
prepares the feed for downstream ball milling.

Hybrid grinding
Hybrid grinding refers to the recycling of a portion of the classi er coarse to the HPG
together with the portion of HPG product as shown in Figure 11.7. As in the open-

Comminution Handbook | Spectrum Series 21 173

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

FIG 11.6 – Open-circuit high-pressure grinding rolls followed

by ball mill separator circuit (image courtesy Hakan Dundar).

FIG 11.7 – High-pressure grinding rolls grinding some reject from the
separator in the ball mill separator circuit (image courtesy Hakan Dundar).

circuit pregrinding application, the HPG prepares the feed for downstream ball
milling. A proportion of classi er coarse product is also used to adjust the nes amount
in the HPG feed for effective grinding.

174 Comminution Handbook | Spectrum Series 21

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

As in the open-circuit pregrinding application, the HPG prepares the feed for
downstream ball milling. A proportion of classi er coarse product is also used to adjust
the nes amount in the HPG feed for effective grinding.

Semi-finish grinding
In semi- nish grinding, the HPG is in closed circuit with the static air classi er, which
is shown as a ‘ -sep’ ( -separator) in Figure 11.8, and the ne product from the -sep
goes to the dynamic air classi er of the ball mill circuit.
The -sep is manufactured by H and is well suited to closed-circuit grinding
applications. Product from the HPG is fed into the top of the separator. While cascading
through the inclined plates it de-agglomerates. The nes are carried with the air ow to
the dynamic air classi er and coarse from the -sep is recycled to the HPG . Particles at
nal product si e in nes from the -sep are separated in the dynamic classi er hence,
some portion of the nal product comes from the HPG circuit as well as from the ball
mill circuit. Since the feed going into the ball mill is signi cantly ner in this case, there
is no need for a two-chamber ball mill.

FIG 11.8 – High-pressure grinding rolls separator circuit with fines going
to the separator in the ball mill circuit (image courtesy Hakan Dundar).

Closed-circuit pregrinding with high-pressure grinding rolls and

single-stage ball milling
In this variation (Figure 11.9), the HPG is used in closed circuit with a static air
classi er as in semi- nish grinding mode. But the static separator nes are directed to
the ball mill instead of the dynamic classi er. The static separator ne product is the
feed to the ball mill. Therefore, as in semi- nish grinding, there is no need for a two-
chamber ball mill.

Comminution Handbook | Spectrum Series 21 175

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

FIG 11.9 – High-pressure grinding rolls separator circuit with fines going
to the ball mill in the ball mill circuit (image courtesy Hakan Dundar).

Finish grinding using high-pressure grinding rolls

It is also possible to produce a nished cement product using only an HPG for
comminution (Figure 11.10). In nish-grinding mode, static and dynamic classi ers
are used in series. The ne product from the static classi er is carried to the dynamic
classi er to produce nished cement. The coarse products from each separator are
combined and returned to the HPG for further grinding.

FIG 11.10 – High-pressure grinding rolls in closed circuit with two separators
in series to produce finished cement (image courtesy Hakan Dundar).

176 Comminution Handbook | Spectrum Series 21

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

Finish grinding using a vertical roller mill

Finished cement product can also be produced in a single unit using a vertical roller
mill (Figure 11.11). The single unit contains both grinding between driven rollers and a
rotating table as well as an air classi er.
Chapter 6 provides a more detailed description of a vertical roller mill.

FIG 11.11 – Finished cement produced by a vertical roller mill–separator in closed circuit.
In the coal industry this mill is known as a vertical spindle mill (image courtesy Hakan Dundar).

COMMINUTION FOR COAL PREPARATION PLANTS

Overview
The extent of comminution in coal circuits depends on the coal type and its use. These
aspects are summarised in Figure 11.12.
Comminution for coal preparation plants is discussed in this section. E uipment for
breaking hard coal is discussed in this chapter, beater wheel mills for breaking lignite
are discussed in Chapter 7.
Coal is broken after mining to facilitate transport, and occasionally for direct use.
Metallurgical coal is prepared for use in ironmaking furnaces by being broken to liberate
the non-coal material and then separating this from the combustible material by gravity
and otation processes.

Coal crushing equipment

Coal is a relatively cheap commodity that is mined in bulk. The mined material sent to
the preparation plant for processing contains both coal and rock from in-seam partings
as well as from the roof and oor of the seam. This material can range in si e from large

Comminution Handbook | Spectrum Series 21 177

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

WŽǁĞƌ
/ŶĚƵƐƚƌŝĂů

Lignite
36%
Lowrankcoal WŽǁĞƌ
47% 64%
Subbitumenous

Anthracite
2% ŽŵĞƐƚŝĐ
Hardcoal /ŶĚƵƐƚƌŝĂů
53% dŚĞƌŵĂů
98% WŽǁĞƌ
Bitumenous /ŶĚƵƐƚƌŝĂů
DĞƚĂůůƵƌŐŝĐĂů
/ƌŽŶΘƐƚĞĞů 
FIG 11.12 – Coal types – percentage of world resources and their uses
(image courtesy Hakan Dundar, drawn from data provided by Professor Alban Lynch).

slabs (up to 1 m) to dust-si ed particles. Prior to treatment it must be crushed, both to

liberate coal contained in composite particles and to reduce the material to a si e that
can be ef ciently treated in the preparation plant. The top si e of the crushed coal will be
determined in the initial bore core test work, which is designed to identify the optimum
product that can be produced from the coal that is, to maximise the saleable product.
Coal crushing machines are shown diagrammatically in Figure 11.13.
The coal crushing system can be as simple as a rotary breaker, either alone or in
combination with a hammer mill or similar primary crusher to reduce the large slabs
to a si e that is suitable for feed to the rotary breaker. These systems work well for coal
containing a high proportion of a bright and brittle component called vitrinite while the
associated rock is tough. When the coal and associated rock are suitable, these systems
are popular because they can crush the coal while rejecting some of the rock that is tough
enough to survive the dropping action of the rotary breaker. A rotary breaker is less
effective for duller, tougher coals and for rocks that break easily.
Other comminution systems use feeder breakers as primary crushers followed by
several slow-speed, high-tor ue si ers (also known as rolls crushers) in series to achieve
the si ing re uired while minimising nes generated. The exit gap on these devices
controls the amount of crushing in the device. The crushing system is designed for
plug ow with no recycle. These systems occasionally suffer blockages from larger
than normal amounts of either very large particles or tough minerals such as calcrete
or siderite entering the system. Trip switches and high-level detectors are relied on to
prevent damage to the e uipment. The disadvantage of the system is that the cause of
the detected blockage must be dealt with manually.
Crushing e uipment employed in coal systems varies depending on the nature of
the coal and the personal preference of the process engineer designing the system.
Examples of crushers used are rotary breakers, hammer mills, si ers, rolls crushers or
feeder breakers either alone or in combination with other devices in the list. Primary

178 Comminution Handbook | Spectrum Series 21

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

A B

C D

E F

FIG 11.13 – Crushers for coarse coal (images courtesy Australian Coal Preparation Society). (A) Rotary breaker,
(B) hammer mill, (C) slow-speed high-torque sizer, (D) single rolls crusher, (E) double rolls crusher, (F) feeder breaker.

crushers tend to be any of the aforementioned devices while secondary units tend to be
rotary breakers or si ers. Additional stages are almost always si ers. Some mines utilise
only primary crushing while others have up to four stages of crushing. The nature of the
coal, the desired nal top si e of the particles and nes minimisation will determine the
e uipment selection.

Bore core testing

The testing of coal bore cores is different from metallic ores due to the need to understand
the physical as well as chemical properties of the coal, enabling different si es to be
bene ciated in different processes. While both testing systems aim to identify the most
economical way of recovering as much of the desirable commodity as practical, the
coal procedure aims to achieve it while controlling crushing to minimise nes while

Comminution Handbook | Spectrum Series 21 179

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

the metallic cores are subject to more extensive crushing and grinding to liberate
the mineral. Initial coal testing identi es the type of properties that determine the
potential uses of the coal and those that indicate the best way to process the coal.
In one core it is possible to identify coal that has coking properties and coal that is
suitable for pulverised coal injection or thermal use. Because of their differences in
price, identifying how much of each can be recovered from the coal, as well as how
best to achieve it, is often critical to the nancial success of a project. etermining the
breakage pattern of the coal is critical to identifying potential products.
Because coal is relatively low priced, it must be processed as cheaply as practical.
The processing cost is cheaper for coarser coal and for simple processes. Therefore
crushing should be minimised to what is re uired to achieve the top si e that ensures
recovering the optimum product. A further complication in determining the optimum
top si e for some coals is that further crushing will often liberate particles with better
ualities, but the increased nes in the product may result in higher processing costs,
higher product moisture (that is, lower product values) and handling dif culties.
The nes may re uire special handling techni ues for storage and transport while the
high moisture level may result in market rejection of the product coal. Higher cost of
treatment and handling needs to be balanced by a higher coal selling price.

Testing of core breakage

Breakage of cores establishes the natural breakage of the coal and the effect of additional
breakage and attrition on plant feed si ing (Swanson, Fletcher and Partridge, 1993).
Core is re uired with as large a diameter as can be justi ed economically, at least as
large as the proposed top si e of the coal. Since this will not be known initially, the rst
core may be tested to estimate the diameter re uired from the physical properties of
the coal. Core diameters can vary from 50 mm to 200 mm. The test work is designed to
simulate the natural breakage of the coal, which will occur in mining and handling as
it is transported and stockpiled. The procedure can vary depending on the Hardgrove
Grindability Index (HGI) (AS 1038.20-2002) of the coal but is designed to determine the
‘natural si ing’ it will reach after normal handling (mining, hauling and stockpiling).
The test uses a series of drops from 2 m onto a steel plate to shatter the coal until it
reaches its ‘natural si ing’, de ned as the point at which continued dropping results in
only minor breakage compared to the breakage occurring during the rst drops.
Change in the breakage pattern is usually very marked and easily identi ed by
graphing the cumulative percentage passing a suitable screen si e versus number of
drops. The plot shows an initial rapid rise to a plateau, beyond which there is very
little increase. Based on the top si e re uired for the separation process, any particles
above the nominated top si e are crushed and recombined with the drop-shattered ner
material. The si ing of this material is then determined and used as the raw coal si ing.
For soft coals it is possible that the handling of the core will have broken the material to
its nal si ing and the drop-shatter procedure described will not be re uired. Physical
and chemical testing of the particles is then undertaken. The testing procedure utilised
will depend on the coal’s rank, its HGI and the proposed use of the coal however, the
breakage portion of the coal testing is crucial to its evaluation.

180 Comminution Handbook | Spectrum Series 21

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

Coal grinding circuits in power plants

Modern coal- red central generating plants (power stations) use a pulverised fuel (PF)
coal- ring system that is operated on demand. Coal pulverising circuits are also used in
other applications such as cement kilns and for direct injection of coal into blast furnaces
in the steel industry. This is called pulverised coal injection (PCI).
In coal- red power stations the grinding circuits have the dual role of preparing the
fuel and delivering it to the furnace, and the feed rate is set by the electricity network
of which the power station is a part. Conse uently, process control falls into two areas
1. control of the milling system as a key component of the process undertaken by the
whole power-generation unit
2. control within the milling system to ensure safe and ef cient operation of the milling
plant.
The pulveriser grinds incoming coal from a top si e of between 50 and 75 mm to
70 per cent passing 75 μm. In power stations the product of the milling circuit is PF,
which is blown by hot primary air via a set of delivery pipes into the boiler furnace
where combustion takes place.
The general arrangement of a coal milling system in a power plant is illustrated in the
circuits shown in Figure 11.14. Coal milling is a dry process and the transport medium
is hot, dry (primary) air this delivers a constant ne, dry product to the boiler furnace.
As the PF rate in direct ring applications is critical to the energy release in the boiler
furnace and the generating capacity of the power plant, the mill effectively controls the
rate of energy conversion from the chemical energy in the fuel to the thermal energy in
the generated steam.
The mill system has four main functions
1. grind the coal
2. dry the coal
3. separate the PF
4. transport the PF.

Power plant milling systems

These systems generally fall into the two categories shown in Figure 11.14
1. pressure systems in which the mill is pressurised by forcing the primary air into the
mill
2. suction systems in which the mill is under suction from an extraction-type fan after
the mill.
ownstream problems such as poor combustion, slagging fouling of the boiler
furnace and boiler gas path erosion start with power plant ‘milling’ processes (drying,
comminution, classi cation and transport). A better understanding of the milling
system and its interaction with the processes would assist in improving overall plant
performance.
Figure 11.15 illustrates typical ows to and from a power plant coal mill, in this case a
vertical spindle mill ( SM) type. ertical spindle mills include vertical roller mill ( M)
designs and ring–ball designs.

Comminution Handbook | Spectrum Series 21 181

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

FIG 11.14 – Mill system arrangements (images courtesy Glenn Schumacher, AGL).
(A) Vertical spindle mill type (suction example), (B) ball/tube mill type (pressure example).

182 Comminution Handbook | Spectrum Series 21

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

FIG 11.15 – Power plant coal mill inputs and outputs – vertical
spindle mill type (image courtesy Glenn Schumacher, AGL).

The ‘just-in-time’ nature of power station coal milling re uires a milling system that
incorporates classi ers to control the PF si ing. With some mill types (for example,
SMs) the classi er is enclosed within the mill body, but other mill types (such as ball
tube mills) use external classi ers. In addition to the classi cation that takes place within
the classi er itself, other classi cation processes occur within the mills. There can be
a number of recirculation loops within a mill or milling system, and the situation can
be further complicated by the existence of mill reject systems that reject undesirable
material (tramp metal and non-coal bearing rock) from the mill.
The recirculation circuits that exist within the mill are illustrated in Figure 11.16.
With vertical spindle mills the coal from the feeder enters the mill through a chute
passing down through the classi er and is delivered directly to the grinding one. ry,
hot air (primary air – PA) enters the mill through a plenum under the grinding table
and passes up through a port ring on the outer periphery of the grinding table. The PA
creates a uidising one that collects ne material from the grinding one and transports
the ground material upwards inside the mill body. Heavy non-coal material passes into
the mill rejects discharge system and is rejected from the mill. The ground coal collected
by the uidising PA travels upwards inside the mill however, the velocity decreases
above the direct uidising one and larger particles fall back to the grinding one.
Smaller particles are entrained in the PA ow and are classi ed. Particles in the coarse
stream are returned to the mill for regrinding those in the ne stream exit the mill and
ow through the system, which distributes the PF to the burners.

Comminution Handbook | Spectrum Series 21 183

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

FIG 11.16 – Power plant coal mill internal circuits for coal and air (coal and
air enter through the same chute) (image courtesy Glenn Schumacher, AGL).

CONTROL OF COAL MILLS IN POWER STATIONS

The milling system is the fuel-handling system within the power station. The real-
time demand for fuel to feed the combustion process is dictated by the demand for
electricity from the network to which the power station is connected. The generation of
electricity by individual power stations is controlled on an automatic basis by demand
within the electricity network. Fluctuations in demand cascade back along the energy
transformation process as set out in Figure 11.17.
A power station that burns black coal and produces 660 MW from a single unit
(a ‘unit’ being a single boiler, turbines and generator, and necessary plant such
as mills and fans) will use about 300 t of coal per hour at full load. epending on
coal properties and mill types, these plants are tted with several mills per unit for
example, the 660 MW unit may be tted with seven ball tube type mills (typically
six mills will be in service to meet maximum demand). Each mill will pulverise about
50 t h of coal reducing the si e from 50–75 mm to 75 μm. SMs (see Figure 11.16) are
also used and are usually preferred because compression is more energy-ef cient than
impact and attrition.
Two components are important in the control of comminution circuits in power
plants
1. grinding the PF re uired by the network (this affects the entire power station)
2. ensuring safe and ef cient operation of the milling plant.

184 Comminution Handbook | Spectrum Series 21

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

FIG 11.17 – The role of the milling system in power station operation (image courtesy Glenn Schumacher).

Control of the milling system variables

It is essential that the milling system is operated safely since it affects the entire power
station. This system is effectively the main fuel delivery mechanism to the combustion
process, and the safe management of the combustion process depends upon the ability
to control the delivery of the fuel.
The burner management system (BMS) is key to ensuring a safe, orderly operating
seTuence in the startup and shutdown of fuelÀring eTuipment, of which the milling
system is central. The BMS also protects the plant against malfunction of fuelÀring
equipment and associated systems during operation. The safety features of the BMS are
designed to provide protection in most common emergency situations.
The relationship between control elements in the plant and milling system is shown
in Figure 11.18.
Control of power station electrical output can be traced back to the introduction of
fuel to the combustion process in the boiler furnace. This delivery is a function of the
milling system. Demand on the output of the station is varied on an automatic basis by
the automatic governor control. The ‘pull’ demand on energy initiated by the electricity
network ultimately is responded to by increasing or decreasing the delivery of fuel
(coal) to the furnace of the boiler, where the energy within the coal is transferred to the
water in the boiler generating steam. Obviously the management of the combustion
process through the boiler burner system (by the BMS) is critical. The ‘pull’ demand
on the milling system is addressed by either increasing or decreasing both coal feed to
the mill as well as primary air (the transport medium) Áow to the mill. The increasing
decreasing demand on the milling system results in the change of fuel delivery to the
mill. Details of the control elements of the milling system in a coalÀred power station
are illustrated in Figure 11.19.

Comminution Handbook | Spectrum Series 21 185

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

FIG 11.18 – Relationship between control elements and the milling system (image courtesy Glenn Schumacher).

Control within the milling system

Pulverised feed coal ring differs from other coal combustion technologies primarily
because of the much smaller particle si e used and the resulting high combustion rates.
The combustion rate of coal as a solid fuel is, to a large extent, controlled by the total
particle surface area. By pulverising coal to a nominal 75 μm the coal can be completely
burned within one to two seconds. This approaches the rate for oil and gas.
In contrast, other ring technologies use crushed coal (lumps) of various si es and
provide substantially longer combustion one residence times (up to 60 seconds or
longer). As SM types constitute the largest group of mills now found in power
stations, this discussion of control within the milling system will focus on this mill
type.
The movement of solids and air through a typical SM is shown in Figure 11.16.
Control parameters for the milling system are the rate at which coal is fed to the mill
and the variables of the classi cation process. These variables affect the uality of the
PF because this depends on the PS as well as the throughput rate of the mill.
In coal mills, the transport medium through the mill is air that is heated to aid drying
of the coal, one exception being the use of hot boiler furnace gases in brown coal mills.
The importance of the ow rate of PA to the mill is also critical as it affects the through-
put of the mill and the classi cation mechanisms within the mill and mill classi er.

186 Comminution Handbook | Spectrum Series 21

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

FIG 11.19 – Control elements in the milling system (image courtesy Glenn Schumacher).

With power station units being built to operate for long, continuous periods (typically
two to four years before maintenance shutdowns), the plants are often designed with
an additional mill to ensure full output. This redundancy is to allow out-of-service
mill maintenance to be possible without interrupting the unit output however, mill
performance over the operating campaign between overhauls naturally deteriorates.
Whilst some online control mechanisms, such as classi er adjustments on SMs,
allow adaptation to varying mill performance, the fact remains that issues such as mill
wear create some control challenges. The introduction of additional parameters such as
variable speed mill drives may offer more adaptation capability however, such changes
are not common as they are seen as adding complexity to power station mill systems.

CONTROL OF COMMINUTION CIRCUITS IN COAL PREPARATION PLANTS

Separation of coal and rock in preparation plants is carried out as coarse as possible to
minimise treatment costs. Crushing is carried out to achieve the top si e re uired while
minimising the generation of ner particles and this involves crushing in a controlled
manner using relatively slower processes compared to those employed in metallic ores,
which re uire a ner nal si ing. It is not unusual for the crushing of raw coal to consist
of three or four stages. Because coal and the associated sedimentary rocks are relatively
soft or brittle, the crushing is easy enough. Conse uently, the control system re uired in
coal preparation plants is very simple.

Comminution Handbook | Spectrum Series 21 187

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

Proportional-integral-derivative (PI ) control is widely accepted in industry. PI

controllers are successful over a wide range of operating conditions and they are
straightforward. The PI algorithm consists of three basic coef cients – proportional,
integral and derivative – which are varied to achieve optimal response. Closed loop
systems, the theory of classical PI and the effects of tuning a closed loop control system
are discussed in the ational Instruments (2011) article ‘PI theory explained’.
Sensors, such as belt weighers and bin level detectors, are used to sense feed rates and
provide input to PI controllers to regulate the output of variable speed feeders such as
apron, belt or plate types. Particle top si e is determined by an exit gap on the machine
or a screen aperture that is part of the crushing device. It is usually a plug ow system
with little or no recycle. The exception to this is the occasional use of an impact crusher
and screen on bypass coal. In this system the crushed coal is screened and oversi e
recycled to the crusher feed. These systems are usually simple enough, with the driver
of the loader feeding the system visually controlling it.

CONTROL OF COMMINUTION CIRCUITS IN CEMENT PLANTS

The production of high- uality cement starts with the preparation of a mixture of
materials which, when roasted at 1400°C in a kiln, forms clinker that can be transformed
into high- uality cement by ne grinding. The mixture must have the correct chemical
composition while the ground clinker must have the correct PS . The raw materials will
vary in composition – a potential source of problems in the kiln product – but the rapid
analytical techni ues now available enable variations to be detected and compensated
for during the kiln operation. Conse uently, the clinker that is fed to the grinding circuit
is reasonably stable.
Modern high-capacity cement plants operate at up to 12 000 t d, with the grinding
mills in common use being two compartment ball mills and high-pressure systems
HPG , vertical roller mills ( M) and Horomills®. Ball mills have diameters up to
5.7 m and Ms have tables up to 7.5 m in diameter. Several additives are available
for use in the grinding circuit to ensure that the neness necessary for a speci c cement
can be achieved.
Control of production re uires control of product uality and the measure of uality
is the strength of the nal cement. This involves control of the PS of the separator nes
although high cement uality does not always mean high concrete uality as water,
aggregate, ller and chemicals are other major variables in concrete uality. In the
cement industry, there is a tendency to move the understanding of uality from cement
to concrete because tests now emphasise concrete strength.
The objectives of the automatic control of clinker grinding circuits are to
x Obtain maximum throughput for the energy used (about 70 per cent of the total
energy used in the cement production plant is for grinding raw material and cement
clinker). In a ball mill circuit this re uires the mill lling to be maintained at the
optimum level by control of the feed rate.
x Maintain the PS of the nes, which is the nished cement at a constant value by
control of the separator speed and the nes’ PS .
On an hour-to-hour basis, grinding circuit operation of cement production is controlled
by cement uality, and the particle si e or Blaine surface area of the product is the

188 Comminution Handbook | Spectrum Series 21

 CHAPTER 11 t COMMINUTION CIRCUITS FOR ORES, CEMENT AND COAL

parameter used for control.1 In a typical plant the operators follow the results coming
from the laboratory on an hourly basis to control the production rate. The disadvantage
of this type of procedure is that the response time is too long and in most cases the
operators have to periodically adjust the production to compensate for the uctuations
in the cement uality.
Mill sound level is a measure of mill lling and can be monitored by microphones.
ecreasing sound amplitude and increasing power indicates that the mill is lling
conversely, increasing sound and decreasing power indicates that the level is falling.
This can be built into a control system if necessary.
In addition to feed rate control, the ratio of additives going in to the cement control is
critical for uality assurance and an automatic loop is used to keep the ratio constant as
advised by the laboratory manager. The PS of the nished cement can be measured on
a sample off-line using a laser si er or, more fre uently, by the amount retained on 32 or
45 microns using a rapid sieving techni ue such as an Alpine air jet sieve. The PS can also
be measured online using the Hosokawa Alpine Optisi er and it can be controlled by the
rotor speed. A PI controller is suitable for this loop. umerous control techni ues have
been developed to predict the PS of the cement and to use this as a basis for controlling
the mill and the separator processes simultaneously, but they are not widely adopted.

BIBLIOGRAPHY
Duda, W H, 1985. Cement Data Book, three volumes, 444 p (Bauverlag GMBH Wiesbaden and Berlin).
National Instruments, 2011. PI theory explained online . Available from http www.
ni.com white-paper 3782 en Accessed 29 March 2011 .
Outotec Solutions, 2012. Outotec Mineral Processing Solutions online . Available from http
www.outotec.com .
Rybinski, E, Ghersi, , avila, F, inares, , alery, W, ankovic, A, alle, and ikmen, S,
2011. Optimisation and continuous improvement of the Antamina comminution circuit,
in Proceedings Fifth International Conference on Autogenous and Semi-Autogenous Grinding
Technology, SAG2011, paper #130.
Samuel, M, alery, W and ybinski, E, 2012. Mine-to-mill mastery Antamina boosts throughput
for hard ores with support from Metso PTI, CIM Magazine, 7(7) 44–46.
Scott, A and Morrell, S, 1998. Proceedings Mine to Mill Conference, 248 p (The Australasian Institute
of Mining and Metallurgy Melbourne).
Standards Australia, 2002. AS 1038.20-2002 Coal and coke – Analysis and testing – Higher rank
coal – Hardgrove grindability index, Part 20.
Swanson, A , Fletcher, I S and Partridge, A C, 1993. Improved prediction of si e distributions
and their effects in material handling and coal preparation, ACA P project 1290, uly.
Valery, W and ybinski, E, 2012. Optimisation process at Antamina boosts production and energy
ef ciency helping a large copper inc mine meet the economic challenge of processing harder
ore types, Engineering and Mining Journal (E&MJ), September, pp 116–120.

1. The Blaine test gives an indication of the surface area of cement particles, which is important because cement
works by adhering to sand and gravel. Grinding cement finer gives a higher surface area and produces a stronger
cement, but requires more energy.

Comminution Handbook | Spectrum Series 21 189

 HOME

Chapter 12

MILLING CIRCUIT
CALCULATIONS
Hakan Dundar, Don Burgess, Bianca Foggiatto,
Marcos Bueno, Eddie McLean and Greg Lane

INTRODUCTION
&RPPLQXWLRQ FLUFXLWV FRQWDLQ JULQGLQJ PLOOV DQG FODVVLÀHUV 7KH\ DUH OLQNHG E\ WKH
FLUFXODWLQJ ORDG ZKLFK LV WKH VWUHDP FRQWDLQLQJ WKH FRDUVHUVL]HG SDUWLFOHV IURP WKH
FODVVLÀHUV W\SLFDOO\ WKH XQGHUÁRZ WKDW LV UHWXUQHG WR WKH PLOOV IRU IXUWKHU JULQGLQJ
7KH FLUFXODWLQJ ORDG LV RIWHQ PXFK ODUJHU LQ WRQQHV SHU KRXU WKDQ WKH IUHVK IHHG VR
FDQ KDYH D VHULRXV HIIHFW RQ WKH HIÀFLHQF\ RI WKH PLOO DQG WKH FDSDFLW\ RI WKH FLUFXLW
'HWHUPLQLQJWKHFLUFXODWLQJORDGDQGWKHHQHUJ\UHTXLUHGSHUWRQQHWRJULQGPDWHULDO
IURPF80 to P80 DUH LPSoUtDQt FDOFXODtLoQV IoU HYHU\ PLOOLQJ FLUFXLt 7KH ÀUVt VHFtLoQ LQ tKLV
FKDStHU LV FoQFHUQHG ZLtK tKHVH FDOFXODtLoQV

CLASSIFIER CALCULATIONS – MASS BALANCING AND CIRCULATING LOAD

7KH NH\ to XQGHUVtDQGLQJ HYDOXDtLQJ DQG PoGHOOLQJ tKH SHUIoUPDQFH oI FOoVHG
JULQGLQJ FLUFXLtV LV DFFXUDtH PDVV EDODQFLQJ 7KHUH DUH DOZD\V HUUoUV LQ VDPSOLQJ
DQG PDVV EDODQFLQJ JLYHV tKH PoVt OLNHO\ HVtLPDtHV oI PDVV ÁoZ UDtHV oI GLIIHUHQt VL]H
SDUtLFOHV LQ DOO tKH FLUFXLt VtUHDPV 0DVV EDODQFLQJ FDQ EH SHUIoUPHG DUoXQG DQ\ GHYLFH
tKDt VHSDUDtHV tKH IHHG LQto GLIIHUHQt SUoGXFtV VXFK DV D SUoSHUO\ oSHUDtLQJ FODVVLÀHU
oU DUoXQG DQ\ QoGH Dt ZKLFK tZo GLIIHUHQt VtUHDPV DUH FoPELQHG VXFK DV D SXPS VXPS
LQ ZKLFK FLUFXLt IHHG DQG EDOO PLOO GLVFKDUJH DUH FoPELQHG to IoUP K\GUoF\FOoQH IHHG
,t LV tKH oQO\ UHOLDEOH PHtKoG IoU HYDOXDtLQJ PHDVXUHG FLUFXLt ÁoZ UDtHV DQG FODVVLÀHU
SHUIoUPDQFH $t VtHDG\ VtDtH tKH DPoXQt oI PDtHULDO HQtHULQJ tKH FODVVLÀHU LQ HDFK VL]H
HTXDOV tKH DPoXQt OHDYLQJ Lt LQ ÀQH DQG FoDUVH SUoGXFtV 7KHUHIoUH tKH PDVV EDODQFH FDQ
EH SHUIoUPHG oQO\ IoU VtHDG\VtDtH FoQGLtLoQV 7KH IoOOoZLQJ HTXDtLoQV GHÀQH DOO ÁoZ
UDtHV DUoXQG tKH K\GUoF\FOoQH EDVHG oQ )LJXUH 
F=U+O 

Comminution Handbook | Spectrum Series 21 191

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

K͕Ž 
ŝ

&͕Ĩ 
ŝ

h͕Ƶ 
ŝ


FIG 12.1 – Hydrocyclone streams (image courtesy Hakan Dundar).

F˜f i = U˜ui + O˜oi 

ZKHUH
F LV tKH FODVVLÀHU IHHG UDtH
U LV tKH FODVVLÀHU FoDUVH SUoGXFt ÁoZ UDtH
O LV tKH FODVVLÀHU ÀQH SUoGXFt ÁoZ UDtH
fi, ui, oi DUH tKH DVVD\ oU PDVV IUDFtLoQ LQ LtK VL]H UDQJH oI IHHG, XQGHUÁoZ DQG oYHUÁoZ
UHVSHFtLYHO\
&oPELQLQJ (TXDtLoQV  DQG  JLYHV tKH VLPSOLÀHG IoUP oI tKH oYHUÁoZIHHG
2) UDtLo, ZKLFK FDQ EH GHVLJQDtHG DV ǃ (TXDtLoQ  7KHQ, SUoYLGHG tKH VL]H
GLVtULEXtLoQV oU DVVD\V DUoXQG D FODVVLÀHU DQG tKH ÁoZ UDtH oI Dt OHDVt oQH VtUHDP DUH
NQoZQ, tKH PLVVLQJ ÁoZ UDtHV DUoXQG tKH FODVVLÀHU FDQ EH FDOFXODtHG 6oPH PLVVLQJ VL]H
GLVtULEXtLoQV DUoXQG tKH FLUFXLt FDQ DOVo EH FDOFXODtHG SUoYLGHG tKDt VXIÀFLHQt PHDVXUHG
YDOXHV DUH DYDLODEOH ([DPSOHV oI tKH PDVV EDODQFH HTXDtLoQV DUH DV IoOOoZV

)ƛIi )2ƛXi2ƛRi

)ƛIi )ƛXi²2ƛXi2ƛRi

)ƛIi²)ƛXi 2ƛRi²2ƛXi

)ƛIi – ui 2ƛRi – ui)

f-u
b= O = i i 
F oi - ui
8VLQJ PHDVXUHG GDtD, (TXDtLoQ  ZLOO JLYH D UDQJH oI YDOXHV IoU ǃ IoU HDFK VL]H
FODVV KoZHYHU, E\ GHÀQLtLoQ ǃ LV tKH IUDFtLoQ oI totDO IHHG UHSoUtLQJ to tKH oYHUÁoZ
VtUHDP DQG FDQ oQO\ KDYH oQH tUXH YDOXH 7KHUHIoUH, DV D VtDUtLQJ SoLQt LQ tKH PDVV
EDODQFLQJ SUoFHVV, (TXDtLoQ  LV GHULYHG to FDOFXODtH tKH EHVt HVtLPDtHG YDOXH oI ǃ E\
FoQVLGHULQJ DOO VL]H FODVVHV toJHtKHU

192 Comminution Handbook | Spectrum Series 21

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

7KH GLIIHUHQFH EHtZHHQ tKH IHHG DQG tKH VXP oI tKH SUoGXFtV IoU DQ\ VL]H FODVV LV tKH
PDVV ÁoZ HUUoU IoU tKDt VL]H FODVV, DV VKoZQ LQ (TXDtLoQ 

Ti = fi ]1 - b g ui - boi 

7o ÀQG tKH EHVt YDOXH oI ǃ tKDt PLQLPLVHV tKH totDO PDVV ÁoZ HUUoU, tKH GHULYDtLYH oI
tKH VTXDUHG HUUoU VXP LV VHt to ]HUo

d / T2i = 2 / _ f - ]1 - b gu - bo i_ - o + u i
i i i i i
db

b=
/ _oi - uii_ fi - uii 
/ _oi - uii2
$ItHU tKH FDOFXODtLoQ oI ǃ, tKH VL]H GLVtULEXtLoQV FDQ EH DGMXVtHG E\ HOLPLQDtLQJ tKH
HUUoUV tKDt FDQ FoPH IUoP VDPSOLQJ DQG VL]H DQDO\VLV (TXDtLoQ  FDOFXODtHV tKH HUUoU
ZKLOH tKH IoOOoZLQJ HTXDtLoQV DUH XVHG to GLVtULEXtH tKH HUUoU to tKH ZKoOH VL]LQJ GDtD

fti = fi - Ti 

O
ut i = ui + b 1 - l Ti 
F
O
oti = oi + b l Ti 8
F
ZKHUH
'i GHQotHV tKH HUUoU
tf LV tKH FoUUHFtHG SHU FHQt UHtDLQHG YDOXH IoU IHHG
i
ut i LV tKH FoUUHFtHG SHU FHQt UHtDLQHG YDOXH IoU XQGHUÁoZ
oti
LV tKH FoUUHFtHG SHU FHQt UHtDLQHG YDOXH IoU oYHUÁoZ
$ VLPSOH H[DPSOH LV SUoYLGHG LQ 7DEOH , ZKLFK JLYHV tKH PHDVXUHG SHU FHQt UHtDLQHG
YDOXHV DUoXQG D FODVVLÀHU ZLtK 0 tK IHHG UDtH

TABLE 12.1
Measured size distributions around the classifier (per cent retained).
Size class (mm) Classifier feed (f) Classifier underflow (u) Classifier overflow (o)
+0.177 0.00 0.00 0.00
-0.177/+0.125 5.68 7.78 0.04
-0.125/+0.088 12.19 16.52 0.47
-0.088/+0.063 19.90 26.15 2.98
-0.063/+0.044 19.47 29.10 9.64
-0.044/+0.031 12.38 11.18 15.64
-0.031/+0.022 8.30 4.68 18.11
-0.022/+0.016 4.48 1.78 11.98
-0.016 17.60 2.81 41.14
Total 100.00 100.00 100.00

Comminution Handbook | Spectrum Series 21 193

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

7DEOH  SUoYLGHV tKH PDVV EDODQFH FDOFXODtLoQV IoU HDFK VL]H FODVV 7KH VL]LQJV LQ HDFK
VtUHDP DUH FoUUHFtHG Vo tKDt FDOFXODtLoQV IoU HYHU\ VL]H FODVV \LHOGV tKH VDPH FLUFXODtLQJ
OoDG

TABLE 12.2
Mass balance calculations using Equations 12.4 to 12.8.
Size class oi-ui fi-ui (oi-ui) × (oi-ui)2 Di tf ut i ot i
i
(mm) ( fi-ui)
+0.177 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
-0.177/+0.125 -7.74 -2.10 16.25 59.91 0.65 5.03 8.20 0.27
[0.04–7.78] [5.68–7.78] [-7.74 × -2.10] [(-7.74)2] [5.68-(1–0.355) [5.68–0.65] [7.78 + [0.04 + 0.355
× 7.78- (1–0.355) × 0.65]
(0.355 × 0.04)] × 0.65]
-0.125/+0.088 -16.05 -4.33 69.50 257.60 1.36 10.83 17.40 0.95
-0.088/+0.063 -23.17 -6.25 144.81 536.85 1.97 17.93 27.42 3.68
-0.063/+0.044 -19.46 -9.63 187.40 378.69 -2.73 22.20 27.34 8.67
-0.044/+0.031 4.46 1.20 5.35 19.89 -0.38 12.76 10.93 15.51
-0.031/+0.022 13.43 3.62 48.62 180.36 -1.14 9.44 3.94 17.71
-0.022/+0.016 10.20 2.70 27.54 104.04 -0.92 5.40 1.19 11.65
-0.016 38.33 14.79 566.90 1469.19 1.19 16.41 3.58 41.56
∑1066.37 ∑3006.53 => β = O/F = 1066.37/3006.53 = 0.355

&ODVVLÀHU oYHUÁoZ ÀQH SUoGXFt ÁoZ UDtH

2 = ) ð ǃ = 0 tK ð 0 = 88 tK
&ODVVLÀHU XQGHUÁoZ FoDUVH SUoGXFt ÁoZ UDtH
8 = ) ² 2 = 0 tK88 tK = 8 tK
7KH FoPSDULVoQ oI PHDVXUHG DQG PDVV EDODQFHG FoUUHFtHG VL]H GLVtULEXtLoQV DUH
JLYHQ LQ )LJXUH 

ϭϬϬ

ϴϬ
ůĂƐƐŝĨŝĞƌ
&ĞĞĚ
ƉĂƐƐŝŶŐй

ϲϬ
ůĂƐƐŝĨŝĞƌ
ŽĂƌƐĞ
ϰϬ
ůĂƐƐŝĨŝĞƌ
&ŝŶĞ
ϮϬ
ĐŽƌƌĞĐƚĞĚ

Ϭ
Ϭ͘Ϭϭ Ϭ͘ϭ ϭ
Ɖ͘ƐŝnjĞ;ŵŝĐƌŽŶͿ

FIG 12.2 – Measured and corrected size distributions around the classifier (image courtesy Hakan Dundar).

194 Comminution Handbook | Spectrum Series 21

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

Calculation of actual and corrected partition coefficients

7KH HIÀFLHQF\ oU SDUtLtLoQ FXUYH UHSUHVHQtV tKH VHSDUDtLoQ HIÀFLHQF\ oI D FODVVLÀHU, DQG
LV FDOFXODtHG XVLQJ (TXDtLoQV  DQG 0

U $ ui
Pui = # 100 
F $ fi

O $ oi
Poi = # 100 0
F $ fi

ZKHUH
Pui LV tKH DFtXDO SHUFHQtDJH oI VL]H IUDFtLoQ i LQ tKH IHHG tKDt UHSoUtV to FoDUVH SUoGXFt
Poi LV tKH DFtXDO SHUFHQtDJH to tKH ÀQH SUoGXFt
7KHUHIoUH
Pui = 00 ² Poi 

&DOFXODtLoQ VtHSV IoU DQ DFtXDO SDUtLtLoQ FXUYH IoU D PDVV EDODQFHG GDtD VHt DUH JLYHQ LQ
7DEOH  ,Q tKLV H[DPSOH tKH IHHG to tKH FODVVLÀHU LV  tK, FoDUVH SUoGXFt LV 08 tK
DQG ÀQH SUoGXFt LV 8 tK 7KH VL]H GLVtULEXtLoQV DUoXQG tKH FLUFXLt DUH DOVo JLYHQ
$FtXDO SDUtLtLoQ FXUYHV to FoDUVH DQG ÀQH SUoGXFt DUH JLYHQ LQ )LJXUH , ZKLFK
VKoZV tKDt tKH FXUYHV LQtHUVHFt Dt 0 SHU FHQt

TABLE 12.3
Partition curve calculations.
Size range Feed Underflow Overflow Actual Pui Actual Poi
(μm) fi (%) F × fi ui (%) U × ui oi (%) O × oi (U × ui/F × fi) × 100 (O × oi/F × fi) × 100
+425 0.00 - 0.00 - 0.00 - - -
-425/+300 2.39 3.66 3.48 3.65 0.02 0.01 99.74 0.26
-300/+212 2.15 3.29 3.13 3.28 0.02 0.01 99.72 0.29
-212/+150 4.32 6.61 6.25 6.55 0.12 0.06 99.10 0.88
-150/+106 12.55 19.20 17.96 18.82 0.79 0.38 98.02 1.98
-106/+74 15.39 23.55 21.33 22.35 2.47 1.19 94.93 5.06
-74/+53 17.04 26.07 21.80 22.85 6.69 3.22 87.63 12.37
-53/+38 12.31 18.83 12.80 13.41 11.24 5.42 71.22 28.76
-38/+25 9.19 14.06 6.19 6.49 15.71 7.57 46.14 53.85
-25/+18 5.70 8.72 2.06 2.16 13.61 6.56 24.75 75.22
-18/+12 5.50 8.42 1.32 1.38 14.59 7.03 16.44 83.57
-12/+8.6 3.38 5.17 0.72 0.75 9.16 4.42 14.59 85.38
-8.6/+6.2 2.23 3.41 0.52 0.54 5.95 2.87 15.97 84.06
-6.2/+3 3.13 4.79 0.93 0.97 7.91 3.81 20.35 79.61
-3/+1.8 1.80 2.75 0.61 0.64 4.39 2.12 23.21 76.83
-1.8 2.92 4.47 0.90 0.94 7.33 3.53 - -
Total 100 153.00 100 104.80 100.00 48.20 - -

Comminution Handbook | Spectrum Series 21 195

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

100

80
WĂƌƚŝƚŝŽŶй

60
ActualPui

40 ActualPoi

20

0
1 10 100 1000
DĞĂŶƉĂƌƚŝĐůĞ ƐŝnjĞ;ђŵͿ

FIG 12.3 – Actual partition curves for the coarse and fine products (image courtesy Hakan Dundar).

)oU tKH XQGHUÁoZ oU FoDUVH SUoGXFt, tKH SDUtLtLoQ FoHIÀFLHQt GHFUHDVHV ZLtK GHFUHDVLQJ
SDUtLFOH VL]H to D PLQLPXP DQG tKHQ PD\ VtDUt LQFUHDVLQJ 7KLV SKHQoPHQoQ LV tHUPHG
¶E\SDVV· 6oPH SoUtLoQ oI tKH IHHG PDtHULDO E\SDVVHV to tKH FoDUVH SUoGXFt UDtKHU tKDQ
SDVVLQJ tKUoXJK tKH VHSDUDtLoQ ]oQH IoU H[DPSOH, E\ EHLQJ HQtUDLQHG EHtZHHQ ODUJHU
SDUtLFOHV DQG Qot KDYLQJ tKH oSSoUtXQLt\ to EH FODVVLÀHG 7KH E\SDVV FoQtDLQV VoPH
ÀQHU SDUtLFOHV tKDt VKoXOG UHSoUt to tKH ÀQH SUoGXFt VtUHDP EXt DFtXDOO\ UHSoUt to tKH
FoDUVH SUoGXFt VtUHDP DQG, DV D UHVXOt, tKH SDUtLtLoQ FoHIÀFLHQt LQFUHDVHV Dt tKHVH VL]HV
,Q tKH QXPHULFDO H[DPSOH tKH E\SDVV LV  SHU FHQt, ZKLFK LV tKH PLQLPXP YDOXH
oI tKH SDUtLtLoQ FoHIÀFLHQt 7KH DVVXPStLoQ LV tKDt LI tKH E\SDVV LV  SHU FHQt, tKHQ
 SHU FHQt oI HYHU\ VL]H IUDFtLoQ LQ tKH IHHG JoHV to tKH FoDUVH SUoGXFt GXH to tKH
E\SDVV PHFKDQLVP ,t LV DOVo QotHG tKDt ¶E\SDVV· LV VoPHtLPHV FDOOHG ¶ZDtHU VSOLt· VLQFH
tKH IUDFtLoQ oI ZDtHU tKDt UHSoUtV GLUHFtO\ to tKH FoDUVH SUoGXFt ZLtKoXt FODVVLÀFDtLoQ LV
DSSUo[LPDtHO\ SUoSoUtLoQDO to tKH IUDFtLoQ oI ÀQHV tKDt Go tKH VDPH
7KH QH[t VtHS LQ tKH SDUtLtLoQ FXUYH FDOFXODtLoQ LV to tDNH tKH E\SDVV oXt oI tKH DFtXDO
SDUtLtLoQ FXUYH 7KH QHZ FXUYH LV FDOOHG tKH ¶FoUUHFtHG SDUtLtLoQ FXUYH· DQG Lt toXFKHV
tKH [D[LV Dt LtV PLQLPXP YDOXH 7KH FoUUHFtHG SDUtLtLoQ FXUYH LV FDOFXODtHG XVLQJ
(TXDtLoQ  )LJXUH  JLYHV tKH FoUUHFtHG SDUtLtLoQ FXUYH FDOFXODtLoQ
7KH SDUtLtLoQ YDOXHV PXVt EH SOottHG DJDLQVt tKH PHDQ VL]H oI tKH SDUtLFOHV LQ HDFK VL]H
LQtHUYDO $V tKH VLHYHV XVHG DUH LQ D JHoPHtULF SUoJUHVVLoQ, tKH JHoPHtULF PHDQ VL]H LV
XVHG

U $ ui - _F $ fi ) bypassi P - bypass
Puci = # 100 or Puci = ui # 100 
F $ f - _ F $ fi ) bypassi
i
100 - bypass

Comparison of the actual and ideal efficiency curves

(IÀFLHQF\ FXUYHV DUH t\SLFDOO\ FKDUDFtHULVHG E\ tKH E\SDVV, DV ZHOO DV tKH FXt VL]H DQG tKH
VOoSH oI tKH HIÀFLHQF\ FXUYH
7KH FXt VL]H oU VHSDUDtLoQ VL]H oI D FODVVLÀHU oItHQ UHIHUUHG to DV G0 LV GHÀQHG DV
tKH SDUtLFOH VL]H Dt ZKLFK 0 SHU FHQt oI tKDt VL]H UHSoUtV to HLtKHU FoDUVH SUoGXFt oU

196 Comminution Handbook | Spectrum Series 21

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

100

ActualPui
80
CorrectedPuci

WĂƌƚŝƚŝŽŶй
60

40

20

0
1 10 100 1000
DĞĂŶƉĂƌƚŝĐůĞ ƐŝnjĞ;ђŵͿ

FIG 12.4 – Actual and corrected partition curves for the coarse product (image courtesy Hakan Dundar).
ÀQH SUoGXFt tKH VKDUSQHVV oI tKH VHSDUDtLoQ LV UHIHUUHG to DV tKH VOoSH oI tKH OLQHDU
SDUt oI tKH SDUtLtLoQ FXUYH $V tKH SDUtLtLoQ FXUYH VtHHSHQV, tKH FODVVLÀHU PoYHV toZDUG
LGHDO VHSDUDtLoQ LQ tHUPV oI HIÀFLHQF\ ,Q LGHDO VHSDUDtLoQ, ZKLFK PHDQV 00 SHU FHQt
HIÀFLHQF\, DOO SDUtLFOHV DEoYH tKH FXt VL]H DUH FODVVLÀHG DV FoDUVH ZKLOH DOO SDUtLFOHV EHOoZ
DUH FODVVHG DV ÀQH )LJXUH  JLYHV tKH DFtXDO DQG FoUUHFtHG SDUtLtLoQ FXUYHV DQG LGHDO
VHSDUDtLoQ IoU tKH FXt VL]H oI DUoXQG 0 —P, ZKLFK ZDV GHtHUPLQHG IoU tKH QXPHULFDO
H[DPSOH 7KH ÀJXUH DOVo VKoZV tKH FoUUHFtHG SDUtLtLoQ FXUYH DV GHVFULEHG 7KH FoUUHFtHG
d0 UHIHUUHd to DV d0F LV DOZD\V FoDUVHU tKDQ tKH DFtXDO d0


 Ideal
separation


Partition %

Corrected

curve

 Actual
separation

)LVK

+RRN

%\SDVV

  G GF  
Mean particle size (micron)

FIG 12.5 – Actual and ideal efficiency curves (image courtesy Hakan Dundar).

POWER CALCULATIONS USING BOND GRINDABILITIES

0DQ\ PHtKodV KDYH EHHQ dHYHOoSHd IoU dHVLJQLQJ FUXVKLQJ DQd JULQdLQJ FLUFXLtV
%oQd·V PHtKod LV ZLdHO\ XVHd IoU SUHdLFtLQJ tKH HQHUJ\ UHTXLUHPHQt oI FUXVKHUV DQd
Uod DQd EDOO PLOOV, ZKLOH D QXPEHU oI PHtKodoOoJLHV KDYH EHHQ SUoSoVHd IoU DXtoJHQoXV

Comminution Handbook | Spectrum Series 21 197

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

JULQdLQJ $* DQd VHPLDXtoJHQoXV JULQdLQJ 6$* PLOOV (DFK PHtKod LQYoOYHV XVLQJ
ODEoUDtoU\ tHVtV to dHtHUPLQH tKH PDtHULDO JULQdDELOLt\ dHVFULEHd LQ &KDStHU  %DVHd
oQ tKH PDtHULDO JULQdDELOLtLHV DQd HPSLULFDO IoUPXODH, tKH VSHFLÀF HQHUJ\ oI PLOOLQJ
FLUFXLtV LV FDOFXODtHd
%oQd·V LQYHVtLJDtLoQV tDUJHtHd XQdHUVtDQdLQJ tKH UHODtLoQVKLS EHtZHHQ tKH
FoPPLQXtLoQ HQHUJ\ DQd tKH SUodXFt VL]H oEtDLQHd IoU D JLYHQ IHHd VL]H +H dHYHOoSHd
DQ HTXDtLoQ %oQd,  tKDt FDQ EH DSSOLHd to FUXVKHUV DQd Uod DQd EDOO PLOOV ,Q ,
%oQd SXEOLVKHd HTXDtLoQV IoU FDOFXODtLQJ tKH :oUN ,QdH[ :L IUoP ODEoUDtoU\ tHVt
UHVXOtV 7KH &UXVKLQJ :oUN ,QdH[ LV oEtDLQHd IUoP OoZHQHUJ\ LPSDFt tHVtV, ZKLOH Uod
DQd EDOO PLOO ZoUN LQdLFHV DUH oEtDLQHd ZKHQ HTXLOLEULXP LV UHDFKHd LQ D OoFNHdF\FOH
tHVt, DV dHVFULEHd LQ &KDStHU  7\SLFDO YDOXHV oI tKHVH ZoUN LQdLFHV ZHUH SXEOLVKHd
E\ %oQd  DQd DUH SUHVHQtHd LQ &KDStHU  7KH DSSUoDFK FoQVLVtV oI tKH IoOOoZLQJ
VtHSV
 dHtHUPLQH tKH oUH FKDUDFtHULVtLFV :L YLD D VtDQdDUdLVHd ODEoUDtoU\ tHVt
 SUHdLFt tKH VSHFLÀF HQHUJ\ DQd tKH SoZHU dUDZ oI tKH LQdXVtULDO PDFKLQH
 oEtDLQ tKH SoZHU dUDZ IoU tKH UHTXLUHd tKUoXJKSXt IUoP tKH UHODtLoQVKLS
P=TðW
ZKHUH
P LV tKH SoZHU dUDZ N:
W LV tKH VSHFLÀF HQHUJ\ N:Kt
T LV tKH VoOLdV tKUoXJKSXt tK
$FFoUdLQJ to %oQd, tKH VSHFLÀF HQHUJ\ W LV SUoSoUtLoQDO to tKH QHZ FUDFN tLS OHQJtK
JHQHUDtHd LQ SDUtLFOH EUHDNDJH DQd LV HTXDO to tKH ZoUN UHSUHVHQtHd E\ tKH SUodXFt
PLQXV tKDt UHSUHVHQtHd E\ tKH IHHd 7KLV UHODtLoQVKLS LV H[SUHVVHd DV IoOOoZV
10 - 10
W = Wi e P80 F80 o # 1.1 (Wi = kWh/st) 

7KH &UXVKLQJ :oUN ,QdH[ FDQ DOVo EH XVHd to dHtHUPLQH tKH UHTXLUHd oSHQVLdH
VHttLQJV 266 IoU MDZ FUXVKHUV DQd J\UDtoU\ FUXVKHUV, oU FOoVHdVLdH VHttLQJV &66 IoU
FoQH FUXVKHUV IoU D JLYHQ SUodXFt VL]H
7KH VSHFLÀF HQHUJLHV FDOFXODtHd IoU Uod DQd EDOO PLOOV UHTXLUH FoUUHFtLoQV to DYoLd
LQFoUUHFtO\ HVtLPDtLQJ tKH ZoUN LQSXt SHU toQQH oI PDtHULDO ,QLtLDOO\, %oQd SUoSoVHd
FoUUHFtLoQV IoU FDVHV ZKHUH tKH IHHd ZDV VFDOSHd SULoU to JULQdLQJ /DtHU, LQ ,
5oZODQd DQd .MoV UHYLHZHd %oQd·V FoUUHFtLoQV DQd UHFoPPHQdHd tKH XVH oI HLJKt
HIÀFLHQF\ IDFtoUV IoU oSHUDtLoQV tKDt dLIIHU IUoP tKH VSHFLÀHd FoQdLtLoQV DQd oStLPXP
IHHd VL]HV oI D Uod DQd EDOO PLOOLQJ FLUFXLt 5oZODQd DQd .MoV, 80 7KHVH LQFOXdH
x () ² dU\ JULQdLQJ
x () ² oSHQ FLUFXLt
x () ² PLOO dLDPHtHU
x () ² FoDUVH IHHd
x () ² ÀQH IHHd
x () ² UHdXFtLoQ UDtLo IoU Uod PLOOV
x () ² UHdXFtLoQ UDtLo IoU EDOO PLOOV
x ()8 ² Uod PLOO LQHIÀFLHQF\

198 Comminution Handbook | Spectrum Series 21

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

7KH UHTXLUHd SoZHU dUDZ LV FoUUHFtHd E\ PXOtLSO\LQJ tKH FDOFXODtHd SoZHU dUDZ E\
tKH HIÀFLHQF\ IDFtoUV

Preqd = P # EFi 

Efficiency factors
7KH QDtXUH DQd YDOXH oI HDFK IDFtoU LV DV IoOOoZV 5oZODQd DQd .MoV, 80
x () LV XVHd ZKHQ dU\ JULQdLQJ LV FoQVLdHUHd 3oZHU PXVt EH LQFUHDVHd EHFDXVH dU\
JULQdLQJ FoQVXPHV 0 SHU FHQt PoUH SoZHU tKDQ ZHt JULQdLQJ () = 
x () LV DSSOLHd ZKHQ D EDOO PLOO LV oSHUDtHd LQ DQ oSHQ FLUFXLt 9DOXHV IoU () DUH
OLVtHd LQ 7DEOH 
x () LV tKH PLOO dLDPHtHU HIÀFLHQF\ IDFtoU DQd Lt XVHV D EDVH PLOO LQtHUQDO dLDPHtHU oI
8 It  P DV UHIHUHQFH () YDOXHV DUH FDOFXODtHd IUoP tKH IoOOoZLQJ HTXDtLoQV
8 0.2
EF3 = b IoU ' LQ IHHt 8 = 8 It LQVLdH OLQHUV
Dl

2.44 0.2
EF3 = b IoU ' LQ PHtUHV
D l
,I ' • 8 P  It, tKHQ () = 0 PD[LPXP
x () LV XtLOLVHd ZKHQ H[tUD SoZHU LV UHTXLUHd IoU oYHUVL]Hd IHHdV 7KH UHIHUHQFH IHHd
VL]H LV JLYHQ E\ tZo HTXDtLoQV

Rod mills " F0 = 16000 # 13 nm

RMWi

Ball mills " F0 = 4000 # 13 nm

RMWi
ZKHUH
Wi LV LQ N:KVKoUt toQ Wi FDQ EH %0:L oU 50:L
,I F < F0, () LV Qot DSSOLFDEOH
F - F0
Rr + ]BMWi - 7g # f p
F0
EF4 =
Rr
ZKHUH
Rr LV tKH UHdXFtLoQ UDtLo F/P
F and P aUH tKH IHHd VL]H F80 and SUodXFt VL]H P80 Ln —P

TABLE 12.4
EF2 values.
Product size control reference EF2 Product size control reference EF2
% passing % passing
50 1.035 90 1.40
60 1.05 92 1.46
70 1.10 95 1.57
80 1.20 98 1.70

Comminution Handbook | Spectrum Series 21 199

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

x () LV tKH ÀnHnHVV oI JULnd IaFtoU ,t LV aSSOLHd ZKHn P80 LV ÀnHU tKan  —P

EF5 = P80 + 10.3

1.145 # P80

x () LV XVHd ZKHn tKH Uod PLOO UHdXFtLon UatLo 5U LV not oStLPXP tKat LV, tKH
UHdXFtLon UatLo LV OoZHU tKan oStLPXP 7KH oStLPXP UHdXFtLon UatLo LV FaOFXOatHd
XVLnJ tKH IoOOoZLnJ HTXatLon
L
Rr0 = 8 + b 5 #
Dl

ZKHUH
L LV tKH Uod OHnJtK
D LV tKH PLOO LnVLdH OLnHU dLaPHtHU Ln It
,I Rr0 ²  < Rr < Rr0 + , Lt PHanV tKat () LV not aSSOLHd
2tKHUZLVH
2
_ Rr - Rr0i
EF6 = 1 +
150

x () LV XVHd ZKHn tKH EaOO PLOO UHdXFtLon UatLo LV OHVV tKan VL[ 7KLV oFFXUV SaUtLFXOaUO\
Ln UHJULndLnJ
EF7 = Rr - 1.22
Rr - 1.35

x ()8 FonVLdHUV tKH Uod PLOO LnHIÀFLHnF\ ZKHn Ln dLIIHUHnt FLUFXLt FonÀJXUatLonV
x ()8 =  LI a VLnJOH Uod PLOO FLUFXLt IoOOoZV an oSHn FUXVKLnJ FLUFXLt
x ()8 =  LI a VLnJOH Uod PLOO FLUFXLt IoOOoZV a FOoVHd FUXVKLnJ FLUFXLt
x ()8 =  LI a UodEaOO PLOO FLUFXLt IoOOoZV an oSHn FUXVKLnJ FLUFXLt
x ()8 = 0 LI a UodEaOO PLOO FLUFXLt IoOOoZV a FOoVHd FUXVKLnJ FLUFXLt

Mill power calculations using the Bond method and efficiency factors
3oZHU FaOFXOatLonV dLVFXVVHd Ln tKLV VHFtLon aUH EaVHd on %ond·V PHtKod, ZKLFK LV
dHVFULEHd Ln &KaStHU  and UHIHUUHd to Ln &KaStHUV  and  7KH IoOOoZLnJ FLUFXLtV aUH
dLVFXVVHd
x oSHnFLUFXLt Uod PLOO &aVH $
x Uod PLOO  EaOO PLOO FLUFXLt &aVH %
x FOoVHdFLUFXLt VLnJOHVtaJH EaOO PLOO &aVH &
x oSHnFLUFXLt VLnJOHVtaJH EaOO PLOO &aVH '
x IHHd FontaLnLnJ H[FHVVLYH ÀnHV &aVH (
x 6$* PLOO  EaOO PLOO FLUFXLt at tKH &adLa OoZJUadH FonFHntUatoU &aVH )
:oUN ,ndH[ LV dHÀnHd aV tKH energy required in kWh/t to UHGXFHDSDUWLFOHRILQÀQLWHVL]H
to 380 of 100 μm ,I tKH NnoZn :oUN ,ndH[ LV 0 N:Kt, tKH tKHoUHtLFaO VSHFLÀF HnHUJ\ LV
aOVo 0 N:Kt %Xt aOO PatHULaOV KaYH an HnHUJ\ YaOXH dHSHndLnJ on SaUtLFOH VL]H %ond,
 Vo tKHUH LV aOZa\V HnHUJ\ FUHdLtHd to F80 tKat ZLOO UHdXFH tKH ÀnaO VSHFLÀF HnHUJ\
FaOFXOatLon

200 Comminution Handbook | Spectrum Series 21

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

The importance of efficiency factors

%ond·V ZoUN ZaV FaUULHd oXt at a tLPH ZKHn tKH OaUJHVt EaOO PLOO ZaV  P dLaPHtHU
and 00 N: ZaV a OaUJH PotoU 'ata aYaLOaEOH IUoP tKHVH PLOOV LndLFatHV tKat tKHUH
ZaV a OLnHaU UHOatLonVKLS EHtZHHn IHHd UatH and SoZHU, SUoYLdHd tKat tKH ¶KaUdnHVV·, P80
and UHdXFtLon UatLo ZHUH FonVtant ,n tKH PLd0V OaUJHU EaOO PLOOV 8 P dLaPHtHU
ZHUH EXLOt E\ $OOLV &KaOPHUV IoU XVH at %oXJaLnYLOOH ,VOand, 3aSXa 1HZ *XLnHa and
&\SUXV 3LPa, $UL]ona 7KH &\SUXV PLOOV tUHatHd FonYHntLonaO oUHV EXt tKH %oXJaLnYLOOH
PLOOV tUHatHd KaUdHU, FoaUVHU oUHV 7KH t\SLFaO %oXJaLnYLOOH IHHd VL]H Kad F80 oI 0 ǍP
and tKH PLOO JUoXnd to P80 oI 80 ǍP 7KH VSHFLÀF HnHUJ\ ZaV KLJK and Ln an HIIoUt to
oYHUFoPH tKLV LnHIÀFLHnF\, tKH IHHd VL]H ZaV UHdXFHd oYHU tLPH to F80 oI  ǍP 7KLV
VtUatHJ\ KHOSHd to UHdXFH tKH SUoEOHP EXt tKH tKUoXJKSXt ZaV OHVV tKan UHTXLUHd
7KH &\SUXV UHVXOtV ÀttHd tKH NnoZn SattHUn oI IHHd UatHSoZHU UHOatLonVKLSV EXt
tKH %oXJaLnYLOOH UHVXOtV VKoZHd tKat a VHULoXV dHFUHaVH Ln IHHd UatH, aEoXt 0 SHU FHnt,
ZaV nHFHVVaU\ to aFKLHYH tKH UHTXLUHd P80 7KH UHVXOt ZaV tKat FoUUHFtLon IaFtoUV EaVHd
on NnoZn SOant UHVXOtV ZHUH LntUodXFHd to HnVXUH tKat nHZ FLUFXLtV XVLnJ OaUJH PLOOV
oU JULndLnJ KaUd oU FoaUVH oUH PHt SUodXFtLon UHTXLUHPHntV 7KHVH IaFtoUV aUH noZ
FoPPonO\ XVHd and FoPSanLHV KaYH FoPSLOHd tKHLU oZn VHtV oI IaFtoUV, ZKLFK aUH
SUoSULHtaU\ 7KHUH ZHUH aOVo UHTXHVtV IUoP ZLtKLn $OOLV &KaOPHUV IoU %ond to NHHS KLV
ZoUN FonÀdHntLaO EXt KH LnVLVtHd on SXEOLFatLon and KLV SaSHUV ZHUH a OaUJH LnÁXHnFH
on $OOLV &KaOPHUV· VXFFHVV ZLtK PLOO VaOHV

The use of efficiency factors

%ond and FoOOHaJXHV XVHd HIÀFLHnF\ IaFtoUV to PodLI\ tKHLU HPSLULFaO HTXatLonV and tKH
FaVH VtXdLHV VKoZ tKH SUoFHdXUH

Case A – Rod mill in open circuit

5od PLOO IHHd LV tKH XndHUVL]H IUoP tKH VFUHHn Ln tKH FUXVKLnJ SOant (IÀFLHnF\ IaFtoUV
aUH UHTXLUHd to HnVXUH tKat tKH 80 SHU FHnt SaVVLnJ SUodXFt VL]H LV aFKLHYHd
,nSXt data
x FaSaFLt\ ² 0 dtK
x FUXVKHU :L ² 8 N:Kt, Uod PLOO :L ²  N:Kt, EaOO PLOO :L ²  N:Kt
x F80 ²  000 ǍP, P80 ² 00 ǍP
x () = 1$
x () = dHtHUPLnHd aItHU SUHOLPLnaU\ VHOHFtLon oI PLOO
x () = 0
x () = 1$ VLnFH P80 LV OaUJHU tKan  —P
x () = dHtHUPLnHd aItHU PLOO VHOHFtLon
x () = 1$
x ()8 =  LI Uod PLOO IHHd LV VFUHHn XndHUVL]H LI not aSSO\ a  IaFtoU
x : =  ð _10 1100 - 10 16 000 i  ð 0 ð  = 0 N:t LnFOXdLnJ () IaFtoUV
x SLnLon SoZHU = 0 ð 0 = 0 N: = 88 KS
x ,' = LntHUnaO dLaPHtHU and / = OHnJtK

Comminution Handbook | Spectrum Series 21 201

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

Alternative 1
7KH SUHOLPLnaU\ PLOO VHOHFtLon PadH IUoP a taEOH oI PLOO VL]HV 5oZOand and .MoV,
80 LV
x ,' = 0 It  P, / = 00 It  P
x () = 0 FaOFXOatHd aFFoUdLnJ to tKH SUHOLPLnaU\ VHOHFtLon oI PLOO dLaPHtHU
x nHZ SLnLon SoZHU = 0 ð 0 = 88 KS
x PLOO VHOHFtLon IUoP taEOH LV  It  P ,', / = 00 It HIIHFtLYH JULndLnJ OHnJtK (*/
at 0 SHU FHnt Uod FKaUJH ZLtK 80 N: PotoU
x () =  FaOFXOatHd aFFoUdLnJ to tKH VHOHFtHd PLOO dLPHnVLonV
x VHOHFtHd PLOO  It ,', 0 It (*/ oSHUatLnJ at 0 SHU FHnt Uod FKaUJH FonVXPLnJ
0 N: at tKH SLnLon

Alternative 2
7KH SUHOLPLnaU\ PLOO VHOHFtLon PadH IUoP tKH taEOH LV
x ,' = 0 It  P, / = 0 It (*/ 8 P
x tZo PLOOV ZLOO EH UHTXLUHd, aUoXnd 000 N: HaFK
x () = 0 FaOFXOatHd aFFoUdLnJ to tKH SUHOLPLnaU\ VHOHFtLon oI PLOO dLaPHtHU
x nHZ SLnLon SoZHU = 0 ð 0 =  KS oU  KS HaFK
x PLOOV VHOHFtHd IUoP taEOH aUH tZo at 0 It ,', / = 0 It 8 P (*/ at 0 SHU FHnt
Uod FKaUJH and 8 N: PotoUV tKH totaO SoZHU aYaLOaEOH LV LnVXIÀFLHnt
x to aFKLHYH tKH UHTXLUHd SoZHU dUaZ tKH PLOO OHnJtK LV LnFUHaVHd aV IoOOoZV OHnJtK to
 It 8 P and dLaPHtHU IUoP  It to 0 It  P
x () =  FaOFXOatHd aFFoUdLnJ to tKH VHOHFtHd PLOO dLPHnVLonV
x VHOHFtHd PLOOV tZo at  It ,' E\  It (*/ at 0 SHU FHnt Uod FKaUJH and 0 N: PotoUV
totaO SoZHU LV 00 N:, FoPSaUHd to onH PLOO at 80 N:
x VHOHFtHd PLOO  It ,',  It (*/ at  SHU FHnt Uod FKaUJH FonVXPLnJ 8 N: at tKH
SLnLon

Case B – Rod mill / ball mill circuit

%aOO PLOOV oSHUatLnJ ZLtK Uod PLOOV aUH HIÀFLHnt 7KH VtandaUd %ond PHtKod oI SoZHU
FaOFXOatLon LV noUPaOO\ XVHd 5od PLOOV oSHUatH LnHIÀFLHntO\ EXt tKH Uod PLOO SUodXFt
FontaLnV H[FHVVLYH ÀnHV tKat FaXVH tKH EaOO PLOO to UXn YHU\ HIÀFLHntO\ :KHn tKH Uod PLOO
LnHIÀFLHnF\ LV FoPELnHd ZLtK tKH EaOO PLOO HIÀFLHnF\, Lt UHVXOtV Ln a FLUFXLt HIÀFLHnF\ FOoVH
to 00 SHU FHnt 1oUPaOO\ tKH totaO SoZHU LV VSOLt  SHU FHnt to Uod PLOO,  SHU FHnt to
EaOO PLOO
,nSXt data
x FaSaFLt\ ² 00 dtK
x FUXVKHU :L ² 8 N:Kt, Uod PLOO :L ²  N:Kt, EaOO PLOO :L ²  N:Kt
x rod mill F80 ²  000 Ǎm, P80 ² 00 Ǎm Eall mill F80 ² 00 Ǎm, P80 ²  Ǎm

Rod mill calculation

x : = :i ð _10 1100 - 10 16 000 i
x : =  N:Kt inFlXdinJ () IaFtorV

202 Comminution Handbook | Spectrum Series 21

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

x () = 1$, () = 0, () = 0, () = 1$, () = 0, () = 1$, ()8 = 0 rod
mill IHHd iV VFrHHn XndHrVi]H
x Sinion SoZHr = 0 N: at  SHr FHnt rod FKarJH ZitK a dHnVit\ oI  tm 0 mm
diamHtHr rodV at 0 SHr FHnt FritiFal VSHHd ZitK nHZ linHrV
x mill Vi]H inVtall a  It  m ,' ð 0 It  m (*/ oYHrÁoZ rod mill ZitK a
00 N: motor

Ball mill calculation

$ ZHt JrindinJ Eall mill in FloVHd FirFXit ZitK F\FlonHV rHFHiYHV F\FlonH XndHrÁoZ
7KH rod mill SrodXFt iV IHd to tKH Eall mill diVFKarJH VXmS
x : = :i ð _ 10 75 - 10 1100 i
x : =  N:Kt inFlXdinJ () IaFtorV
x () = 1$, () = 0, () = 0, () = 1$, () = 0, () = 1$, ()8 = 1$
x Sinion SoZHr =  N: at  SHr FHnt Eall FKarJH, 0 mm to  mm EallV at  SHr FHnt
FritiFal VSHHd ZitK nHZ linHrV
x mill Vi]H inVtall an 8 It 8 m ,' ð 0 It  m (*/ oYHrÁoZ Eall mill ZitK a
0 N: motor

Case C – Closed-circuit single-stage ball mill

6inJlHVtaJH Eall millV arH inHIÀFiHnt ZKHn FomSarHd ZitK %ond FalFXlationV iI tKH mill
IHHd matHrial iV larJH and Kard andor dHYoid oI ÀnHV
,nSXt data
x FaSaFit\ ² 00 dtK
x FrXVKHr :i ² 8 N:Kt, rod mill :i ²  N:Kt, Eall mill :i ²  N:Kt
x Eall mill IHHd F80 ² 8000 Ǎm, P80 ²  Ǎm
x : = :i ð _ 10 75 - 10 8000 i
x : =  N:Kt nHt () = 0
x () = 1$, () = 1$, () = 0, () = 0, () = 0, () = 1$, () = 0,
()8 = 1$
x Sinion SoZHr = 0 N: in tKiV H[amSlH, EHFaXVH tKH rod mill :i iV JrHatHr tKan tKH
Eall mill :i and F80 iV larJH at 8000 Ǎm, inFrHaVH tKH Sinion SoZHr E\ tHn SHr FHnt to
80 N:
x mill oSHratinJ FonditionV ²  SHr FHnt Eall FKarJH,  mm to 80 mm EallV at  SHr FHnt
FritiFal VSHHd ZitK nHZ linHrV
x mill Vi]H inVtall a 0 It ,' ð  It  m (*/ oYHrÁoZ Eall mill ZitK a 00 N:
motor

Case D – Open-circuit single-stage ball mill

%all millV oSHratinJ in oSHn FirFXit arH inHIÀFiHnt 7o oEtain an 80 SHr FHnt SaVVinJ
SrodXFt Vi]H a 0 SHr FHnt inHIÀFiHnF\ IaFtor mXVt EH addHd to tKH SroFHVV SoZHr
,I a 0 SHr FHnt SaVVinJ SrodXFt Vi]H iV rHTXirHd, a 0 SHr FHnt IaFtor mXVt EH addHd
7KH HIÀFiHnF\ IaFtorV FoYHrinJ 0 SHr FHnt to 8 SHr FHnt SaVVinJ rHTXirHmHntV ZHrH
JiYHn SrHYioXVl\ and arH rHSHatHd KHrH aV 7aElH  Ior FomSlHtHnHVV
() alVo aSSliHV aV dHVFriEHd in ¶VinJlHVtaJH Eall mill in FloVHd FirFXit·

Comminution Handbook | Spectrum Series 21 203

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

TABLE 12.5
Efficiency factors.
Product size % passing Efficiency multiplier Product size % passing Efficiency multiplier
50 1.035 90 1.40
60 1.05 92 1.46
70 1.10 95 1.57
80 1.20 98 1.70

Case E – Feed gradation containing excessive fines

,I tKH IrHVK IHHd to tKH mill FontainV H[FHVViYH ÀnHV aEoYH a normal Jradation ² aV
oFFXrV ZitK a rod mill diVFKarJH ² tKiV Zill EHnHÀt tKH Eall mill VtaJH EHFaXVH tKH Eall
mill HnHrJ\ inSXt Zill EH lHVV to aFKiHYH tKH rHTXirHd mill SrodXFt P80 ,n SraFtiFH, F80 is
rHdXFHd ZKHn tKH IHHd Jradation is SlottHd on loJloJ SaSHr E\ ForrHFtinJ tKH sloSH dXH
to tKH H[FHssiYH ÀnHs

Case F – SAG mill / ball mill circuits

$*6$* Eall mills arH inHIÀFiHnt ZKHn FomSarHd to rodEall mill FirFXits and an
inHIÀFiHnF\ IaFtor mXst EH addHd dHSHndinJ on tKH 6$* mill SrodXFt si]H $s tKH $*
6$* mill SrodXFt T80 transIHr si]H 80 SHr FHnt SassinJ EHFomHs smallHr, tKH rHTXirHd
HIÀFiHnF\ IaFtor rHdXFHs 7Kis is disFXssHd ZitK rHIHrHnFH to tKH 6$* mill  Eall mill
FirFXit at &adia nHar 2ranJH, 1HZ 6oXtK :alHs 7KH FonFHntrator Zas FommissionHd in
8 SroFHssinJ a loZJradH FoSSHrJold orH &irFXit sXrYH\s rHSortHd a 6$* sSHFiÀF
HnHrJ\ oI 8 N:Kt and Eall mill sSHFiÀF HnHrJ\ oI 80 N:Kt 'XnnH et al, , 00
+art et al, 00 7KH T80 Zas not JiYHn so tKis Zas EaFNFalFXlatHd to EH 00 μm EasHd
on tKH Eall mill sSHFiÀF HnHrJ\ oI 80 N:Kt and Eall mill P80 oI 0 ȝm 7KH orH sSHFiÀF
JraYit\ 6* Zas 

Ball mill specific energy

7Zo  It ,' ð  It (*/ oYHrÁoZ Eall mills arH XsHd to Jrind tKH 6$* mill SrodXFt at
&adia ²  SHr FHnt FritiFal sSHHd &6, 0 SHr FHnt Eall FKarJH (aFK mill draZs a total
SoZHr oI 80 N: at tKH Sinions, rHTXirinJ a 0 000 N: dXal Sinion driYH
,nSXt data
x FaSaFit\ ² 0 dtK HaFK mill, $ðE =  Kard orH
x FrXsKHr :i ²  N:Kt, rod mill :i ²  N:Kt, Eall mill :i ²  N:Kt
x Eall mill F80 ² 00 Ǎm FalF, Eall mill P80 ² 0 Ǎm
x () ² 1$, () ² 1$, () ² 0, () ² 0, () ² 0, () ² 1$, () ² 0,
()8 ² 1$
x : = :i ð _10 170 - 10 1500 i
x : = 80 N:Kt or 80 N: SHr mill
x notH tKH 5od 0ill :orN ,ndH[ is XsHd doZn to a P80 = 000 Ǎm

204 Comminution Handbook | Spectrum Series 21

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

SAG mill specific energy

7KH 6$* PLOO VL]H LV  IW ,' ð  IW (*/ ÀWWHG ZLWK D   N: ULQJ PRWRU ,W ZDV
UHSRUWHGWKDWWKHPLOOZDVGUDZLQJN:DWWKHVKHOOJLYLQJDVSHFLÀFHQHUJ\RI
N:KW
,QSXWGDWD
x FDSDFLW\²GWK6*²$ðE²':LN:KP
x FUXVKHU:L²N:KWURGPLOO:L²N:KWEDOOPLOO:L²N:KW
x F²ǍPT²ǍPIURPEDOOPLOOVSHFLÀFHQHUJ\ N:KW

Calculation of mill power

7KH SURFHGXUH LV WR FDOFXODWH %RQG VSHFLÀF HQHUJ\ LQ WKUHH SKDVHV GHÀQHG E\ WKH
SHUFHQWSDVVLQJVL]HVǍPWRǍPǍPWRǍPDQGǍP
WRǍP
x ǍPWRǍP²XVH%RQGHTXDWLRQDQGFUXVKHU:L²N:KW
x : N:KW
x ǍPWRǍP²'%&50:LHTXLYDOHQWIURP)LJXUHDRI%XUJHVV
LVN:KW
x : N:KW
x ǍPWRǍP²XVH%RQGHTXDWLRQDQGURGPLOO:L²N:KW
x : N:KW
x 7RWDO%RQGVSHFLÀFHQHUJ\ N:KW
)RUPRUHLQIRUPDWLRQRQHIÀFLHQF\IDFWRUVVHH%XUJHVV
7KH'%&IDFWRUIRU6$*SURGXFWRIǍPLVUHIHUWR)LJXUHE
x '%&6$*PLOOVSHFLÀFHQHUJ\ ð N:KW
x FRUUHFWLRQIRU6* N:KW
x 6$*VSHFLÀFHQHUJ\  N:KW

ϯϬϬϬ
ϮϳϱϬ
ϮϱϬϬ
ϮϮϱϬ
ϮϬϬϬ
ϭϳϱϬ
T80μm

ϭϱϬϬ
ϭϮϱϬ
ϭϬϬϬ
ϳϱϬ
ϱϬϬ
ϮϱϬ
Ϭ
ϭ ϭ͘ϱ Ϯ Ϯ͘ϱ

A B DBCFACTOR

FIG 12.6 – (A) DBC graph to convert Ab values to equivalent Rod Mill Work Indices (RMWi) for the 25 000 μm to
3000 μm SG calculation; (B) DBC factors for given transfer size T80 (μm) (images courtesy Don Burgess Consulting).

Comminution Handbook | Spectrum Series 21 205

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

x ForrHFtion Ior $ðE =  is 1$ as P80 !0 Ǎm

x Ànal '%& 6$* mill SoZHr is 80 N:Kt
7KH sSHFiÀF HnHrJiHs oI tKH 6$* and Eall mills Fan EH FomSarHd to similar YalXHs
rHSortHd E\ 'XnnH et al , 00 and +art et al 00, all as dHsFriEHd in %XrJHss 0

Additional methods for predicting power in circuits

%arratt and $llan 8 SroSosHd tKH Xtilisation oI %ond FomminXtion ZorN indiFHs
and an HIÀFiHnF\ IaFtor I6$* Ior dHtHrmininJ tKH sSHFiÀF HnHrJ\ oI $* or 6$* mill
EasHd FirFXits 8sinJ a Silot Slant and oSHratinJ Slant dataEasH, %arratt 8 dHÀnHd
an HmSiriFal IormXla to dHtHrminH tKH sSHFiÀF SoZHr rHTXirHmHnts Ior FirFXits inFlXdinJ
6$* and Eall mills 7KH mHtKod rHTXirHd a SroSosHd F80 IHHd si]H to tKH 6$* mill, tKH
Ànal Jrind si]H and an HstimatH oI tKH transIHr si]H EHtZHHn tKH tZo mills )or a tZo
staJH $*6$* and Eall mill FirFXit
1 - 1 1 - 1 1 - 1
WSAG = 1.25 =10WiC e PC FC o + 10WiR e PR FR o kr + 10WiB e PB FB oG
1 1
- 10WiB e 110 - PSAG o kb
1 1
Wball = 10WIB e P - PSAG o kb
B

7KH sXEsFriSts 6$*, &, 5 and % rHIHr to 6$* mills, FrXsKHrs, rod mills and Eall mills,
rHsSHFtiYHl\
6inFH tKHn, otKHr mHtKodoloJiHs KaYH EHHn dHYHloSHd Ior FalFXlatinJ sSHFiÀF HnHrJ\
Ior $* and 6$* mills 6omH oI tKHsH mHtKodoloJiHs arH dHsFriEHd, ZitK mHtKods XsHd
E\ 60&& 0orrHll, 00 and $XsHnFo /anH, )oJJiatto and %XHno, 0 dHsFriEHd in
dHtail 7KH sHlHFtHd mHtKodoloJiHs arH XsHd in ZorNHd H[amSlHs SrHsHntHd in &KaStHr 

MacPherson
0aF3KHrson 8 FrHatHd a JrindinJ tHst tKat Zas FondXFtHd in a FontinXoXs laEorator\
mill in FlosHd FirFXit ZitK a FlassiÀHr Xntil stHad\statH Zas aFKiHYHd $t tHst FomSlHtion,
si]H anal\sHs ZHrH FarriHd oXt and tKH sSHFiÀF HnHrJ\ FalFXlatHd XsinJ SoZHr draZ and
tKroXJKSXt inIormation 7KH tHst also SroYidHd insiJKt into KoZ tKH KardHr FomSonHnts
EXild XS in tKH mill FKarJH (mSiriFal ForrHFtions ZHrH aSSliHd to tKH laEorator\ mill
:orN ,ndH[ to dHÀnH tKH :orN ,ndH[ Ior Slant dHsiJn

GrindPower
7KH mHtKodoloJ\ SroSosHd E\ %arratt and $llan 8 Zas inForSoratHd into a
FomSXtHrEasHd SroJram Ior tKH si]inJ and sHlHFtion oI JrindinJ FirFXits FallHd
*rind3oZHr 0attKHZs and %arratt, 

Julius Kruttschnitt Mineral Research Centre (JKMRC)

7KH -.05& mHtKodoloJ\ is EasHd on tKH XsH oI modHls aYailaElH in tKH -.6im0Ht
SroJram 0orrison and 5iFKardson, 00 1aSiHr0Xnn et al, 00 7KH FomminXtion
modHls, ZKiFK arH EasHd on -.05&·s rHsHarFK oXtFomHs, arH FaliEratHd ZitK HTXiSmHnt
and orHsSHFiÀF SaramHtHrs oEtainHd Irom sXrYH\ data and orHsSHFiÀF SaramHtHrs
Irom tHsts ,nitiall\, tKHsH ZHrH tKH SHndXlXm tHst, 'roS :HiJKt 7Hst -.':7, 6$*

206 Comminution Handbook | Spectrum Series 21

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

mill FomminXtion tHst 60& morH rHFHntl\, Irom tKH rotar\ ErHaNaJH tHst -.5%7 7KH
6$* mill and Eall mill SoZHr draZs arH FalFXlatHd indHSHndHntl\ XsinJ sHmiHmSiriFal
HTXations as a IXnFtion oI mill JHomHtr\, sSHHd and load

SAG Power Index

7KH 6tarNH\ 6$* 3oZHr ,ndH[ 63,® tHst 6tarNH\ and 'oEE\,  Zas dHYHloSHd as
tKH $*6$* HTXiYalHnt oI tKH :orN ,ndH[ tHst $ 6$* mill is oSHratHd in FlosHd FirFXit
and tKH timH rHTXirHd to Jrind tKH matHrial to a P80 oI  mm is XsHd to SrHdiFt $*6$*
sSHFiÀF HnHrJ\ rHTXirHmHnts $dMXstmHnt IaFtors arH XsHd iI IHHd si]H dHYiatHs Irom
tKH ranJH HstaElisKHd or iI tKH HIIHFt oI a SHEElH FrXsKHr is FonsidHrHd 7KH rHlationsKiS
EHtZHHn sSHFiÀF HnHrJ\ and tKH timH rHTXirHd to Jrind tKH matHrial is dHÀnHd E\ tKH
IolloZinJ HTXation
0.5 n
SAG kWh t = K _SPI T80 i fSAG
ZKHrH
fSAG is tKH sXEmodHl tKat inForSoratHs HIIHFts sXFK as IHHd si]H and SHEElH FrXsKinJ
7KH sXEmodHl fSAG and tKH sSHFiÀF HnHrJ\ HTXation arH aYailaElH ZitKin tKH &((7
FomminXtion HFonomiF HYalXation tool FomminXtion simXlator .osiFN, 'oEE\ and
%HnnHtt, 00

SAGDesign™
7KH 6$*'HsiJn tHst Zas dHYHloSHd E\ 6tarNH\ and sXSSortHd E\ 'aZson /aE and
2XtoNXmSX noZ NnoZn as 2XtotHF to oYHrFomH tHFKniFal limitations oI tKH 63, tHst
6tarNH\, +indstrom and 1adasd\, 00 $FFordinJ to tKH aXtKors, its main limitation
Zas tKat tKH 63, tHst rHTXirHd onl\ a  NJ tHst, so man\ samSlHs ZHrH nHFHssar\ to
aFKiHYH dHsirHd aFFXraF\ lHYHls 7KH 6$*'HsiJn tHst SroFHdXrH XsHs a larJHr laEorator\
6$* mill tKan tKH 63, tHst and rHTXirHs aSSro[imatHl\ 0 NJ oI samSlH ZitK FoarsHr toS
si]H 7KH JroXnd matHrial JHnHratHd E\ tKH 6$* mill is XsHd in a sXEsHTXHnt Eall mill
%ondst\lH :orN ,ndH[ tHst %\ FondXFtinJ tKHsH tZo tHsts, tKH total JrindinJ HnHrJ\ Fan
EH FalFXlatHd

OMC’s method
6idall and 3Xtland 00 SXElisKHd tKH 20& 2rZa\ 0inHral &onsXltants mHtKod to
FalFXlatH JrindinJ SoZHr rHTXirHmHnt EasHd on laEorator\ tHsts tKat dHÀnH tKH FoarsH
and ÀnH KiJK and loZHnHrJ\ ErHaNaJH FKaraFtHristiFs +iJKHnHrJ\ ErHaNaJH tHsts
sXFK as tKH aXtoJHnoXs mHdia FomSHtHnF\ tHst $0&7, -.':7, 63, or 6$*'HsiJn
tHst arH tZinnHd ZitK a loZHnHrJ\ ErHaNaJH tHst sXFK as %ond %all 0ill :orN ,ndH[
%0:i 6tandardisHd IHHd and SrodXFt si]Hs arH XsHd to FalFXlatH tKH oYHrall HnHrJ\
rHTXirHmHnts XsinJ %ond·s IormXla and an I6$* HIÀFiHnF\ IaFtor 7KH 20& dataEasH is
tKHn XsHd to sHlHFt YiaElH oStions oI FirFXit FonÀJXration Ior dHtailHd anal\sis

DBC’s method
%XrJHss 0 dHYHloSHd '%&·s 'on %XrJHss &onsXltinJ mHtKodoloJ\ Ior FalFXlatinJ
JrindinJ HnHrJ\ XsinJ %ond·s FomminXtion tKHor\ 7Kis is tKHn XsHd to FalFXlatH tKH
sSHFiÀF HnHrJ\ oI indiYidXal FomminXtion ranJHs XsinJ ErHaNaJH FKaraFtHrisation
indiFHs oEtainHd Irom %ond·s FrXsKHr tHsts, rod and Eall mill tHsts and tKH -.':7

Comminution Handbook | Spectrum Series 21 207

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

7KH SaramHtHrs oEtainHd Irom tKH -.':7 arH FonYHrtHd to JrindaEilit\ :orN ,ndiFHs
and $* or 6$* mill sSHFiÀF HnHrJ\ is FalFXlatHd 7Kis is ForrHFtHd ZKHn Yariations
oFFXr to standard SaramHtHrs sXFK as 6*, $ðE YalXHs and mill asSHFt ratio

Morrell’s method
$FFordinJ to 0orrHll 00, tKH sloSH oI tKH Slot Irom an 60& 7Hst Ior EroNHn SrodXFt
SHr FHnt oI XndHrsi]H aJainst tKH inSXt HnHrJ\ is rHlatHd to tKH strHnJtK oI tKH roFN 7Kis
is tKHn XsHd to JHnHratH a strHnJtK indH[ ':i, ZKiFK is XsHd to HstimatH tKH sSHFiÀF
HnHrJ\ oI a ZidH ranJH oI FomminXtion FirFXits
7KH sSHFiÀF HnHrJ\ oI tKH $* or 6$* mill Fan EH dHtHrminHd XsinJ tKH IolloZinJ
HTXation
a
S = KF80 DWi b ^1 + c ]1 - e-dJ gh-1s e f ^ Ar h

ZKHrH
S is tKH sSHFiÀF HnHrJ\ at tKH Sinion
F80 is 80 SHr FHnt SassinJ si]H oI tKH IHHd
DWi is tKH strHnJtK indH[
J is tKH YolXmH oI Ealls SHr FHnt
s is tKH mill sSHHd SHr FHnt oI FritiFal sSHHd
fAr is a IXnFtion oI mill asSHFt ratio
a, b, c, d and e arH Fonstants
K is a IXnFtion ZKosH YalXH dHSHnds on ZKHtKHr a SHEElH FrXsKHr is in tKH
circuit
7KH sSHciÀc HnHrJ\ oI tKH $* or 6$* mill can EH dHtHrminHd E\ comSarinJ tKH
oSHratinJ conditions Ior tKH circuit undHr considHration ZitK a larJH dataEasH oI
oSHratinJ $* and 6$* mills 7KH si[ constants and tKH Iunction fAr, ZKicK arH usHd to
coYHr diIIHrHnt oSHratinJ conditions in tKH HTuation, arH SroSriHtar\
7KH 60& SoZHrEasHd aSSroacK usHs :orN ,ndicHs rHlatHd to conYHntional crusKHrs,
KiJKSrHssurH JrindinJ rolls +3*5s and tumElinJ mill SroSHrtiHs 0orrHll, 00,
00 7Kis is donH usinJ tKH SoZHrEasHd si]H rHduction HTuation

- x f (x1)k
x i
W = 4 # Mi # a x 2f_ 2
1

ZKHrH
W is tKH circuit sSHciÀc HnHrJ\ N:Kt
Mi is an indH[ rHlatHd to tKH ErHaNaJH SroSHrt\ oI an orH and is dHtHrminHd
E\ tKH 60& 7Hst N:Kt
x and x arH tKH 80 SHr cHnt SassinJ si]Hs in μm Ior tKH circuit Sroduct and IHHd
7KH Iunction is dHSHndHnt on tKH P80 and can EH calculatHd usinJ tKH IolloZinJ HTuation
0orrHll, 00
f _ x j i = - _0.295 + x j 1000000i

Mi SaramHtHrs arH JHnHratHd E\ tKH 60& 7Hst Mic is tKH indH[ Ior conYHntional
crusKinJ and Mih is tKH indH[ Ior +3*5 si]H rHduction Mia is tKH :orN ,ndH[ rHlatHd
to coarsH JrindinJ, sucK as Iound in rod, $* and 6$* mills 7KH %0:i tHst raZ data

208 Comminution Handbook | Spectrum Series 21

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

is usHd to H[traSolatH tKH :orN ,ndH[ rHlatHd to ÀnH JrindinJ, Mib 7KH coarsH ranJH
coYHrs tKH staJHs Srior to Eall millinJ and ÀnH ranJH rHlatHs to conYHntional Eall millinJ
7KH si]H sHlHctHd as tKH diYision EHtZHHn coarsH and ÀnH JrindinJ Zas 0 mm
Mi SaramHtHrs arH usHd to SrHdict total sSHciÀc HnHrJ\ oI tKH comminution circuit
7KH sum oI Wc, Wh, Wa and Wb rHsults in tKH total sSHciÀc HnHrJ\ to oEtain a Ànal Sroduct
)or H[amSlH, iI tKH circuit contains an $* or 6$* mill and a Eall mill, tKH mHtKodoloJ\
SrHdicts tKH comEinHd sSHciÀc HnHrJ\ oI tKH mills 7KH total sSHciÀc HnHrJ\ WT to
rHducH Srimar\ crusKHr Sroduct to Ànal Sroduct in si]H is JiYHn E\

WT = Wa + Wb + Wc + Wh + Ws
ZKHrH
Wa is tKH sSHciÀc HnHrJ\ to Jrind coarsHr SarticlHs in tumElinJ mills
Wb is tKH sSHciÀc HnHrJ\ to Jrind ÀnHr SarticlHs in tumElinJ mills
Wc is tKH sSHciÀc HnHrJ\ Ior conYHntional crusKinJ
Wh is tKH sSHciÀc HnHrJ\ Ior +3*5s
WV is tKH sSHciÀc HnHrJ\ corrHction Ior si]H distriEution
0orrHll 0 SuElisKHd YaluHs attriEutHd to tKH constants as IolloZs
x K = 0 Ior all circuits tKat do not contain a rHc\clH SHEElH crusKHr and 0 ZKHrH
circuits KaYH a SHEElH crusKHr 7Kis SaramHtHr is onl\ usHd Ior coarsH SarticlH JrindinJ
in tumElinJ mills
x K = 0 Ior all crusKHrs oSHratinJ in closHd circuit ZitK a classiI\inJ scrHHn ,I tKH
crusKHr is in oSHn circuit Ior H[amSlH, +3*5s and SHEElH crusKHrs in an $*6$*
circuit, K taNHs tKH YaluH oI 
x Sc is a coarsH orH KardnHss SaramHtHr tKat is usHd in Srimar\ and sHcondar\ crusKinJ
situations 7Kis SaramHtHr Zas includHd to alloZ Ior tKH dHcrHasH in orH KardnHss
tKat EHcomHs siJniÀcant in rHlatiYHl\ coarsH crusKinJ aSSlications sucK as Srimar\
and sHcondar\ conHJ\rator\ circuits ,n tHrtiar\ and SHEElHcrusKinJ circuits, it is
normall\ unnHcHssar\ and taNHs tKH YaluH oI unit\ ,n IullscalH +3*5 circuits ZKHrH
IHHd si]Hs tHnd to EH coarsHr tKan usHd in laEorator\ and SilotscalH macKinHs, tKH
SaramHtHr Kas also EHHn Iound to imSroYH SrHdictiYH accurac\ 7KH SaramHtHr is
dHÀnHd E\ tKH IolloZinJ HTuation
0.2
Sc = Ks # ^ k1 # k2h-
ZKHrH
Ks is a macKinHsSHciÀc constant tKat taNHs tKH YaluH oI  Ior conYHntional crusKHrs
and  in tKH casH oI +3*5s

Ausgrind
/anH, )oJJiatto and %uHno 0 dHscriEHd a comminution mHtKodoloJ\ usinJ $usJrind
$usHnco·s inKousH mHtKod to dHsiJn comminution circuits 7KH SroJram t\Sicall\
usHs %ond :orN ,ndH[Hs, 60& 7Hsts andor -K':7s as inSut SaramHtHrs ,I ErHaNaJH
cKaractHrisation tHst ZorN Zas conductHd usinJ otKHr mHtKodoloJiHs sucK as 0ac3KHrson
or 63,, HmSirical IormulaH arH usHd to corrHlatH ZitK $usJrind modHl inSut SaramHtHrs
7KH JrindinJ circuit sSHciÀc HnHrJ\ is calculatHd usinJ tKH %ond Iormula and
an HIÀciHnc\ Iactor, I6$*, ZKicK is sSHciÀc to tKH sHlHctHd circuit conÀJuration and

Comminution Handbook | Spectrum Series 21 209

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

orHdHSHndHnt 7KH HmSirical HTuations Ior dHtHrmininJ tKH HIÀciHnc\ Iactor I6$* in
$usJrind calculations ZHrH dHriYHd Irom $usHnco·s dataEasH %uHno and /anH, 0
7KH I6$* rHlationsKiSs ma\ also EH adMustHd ZKHn sHcondar\ crusKinJ or intHnsiYH
ElastinJ SracticHs arH in SlacH tKat is, Ior ÀnHr circuit IHHd si]Hs, F80 )iJurH  sKoZs
tKH $usJrind I6$* rHlationsKiSs Ior diIIHrHnt circuit conÀJurations and IHHd si]Hs
7KH %ond sSHciÀc HnHrJ\ is calculatHd usinJ tKH &:i, &rusKinJ :orN ,ndH[ 5:i 5od
0ill :orN ,ndH[, and %:i to dHtHrminH tKH sSHciÀc HnHrJ\ rHTuirHmHnts oI a crusKHr
² rod mill ² Eall mill circuit Ior a Sroduct si]H P80 oI 0 μm as SHr tKH %ond IormulaH
ZitKout an\ corrHction Iactors listHd E\ %ond, 5oZland and otKHrs ,I tKH Sroduct si]H P80
is diIIHrHnt Irom 0 μm, tKH HnHrJ\ diIIHrHncH is calculatHd as SHr tKH %ond rHlationsKiS
7KH $usJrind total sSHciÀc HnHrJ\ calculation is dHscriEHd in tKH IolloZinJ HTuation
7otal (cs = >%ond (cs to 0 —m fSAG ² F80-HIIHct@ “ >%ond (cs to Ànal P80@

ZKHrH
7otal (cs is tKH circuit sSHciÀc HnHrJ\
%ond (cs is tKH sSHciÀc HnHrJ\ calculatHd as SHr tKH %ond rHlationsKiS
7KH EasH casH 6$* mill sSHciÀc HnHrJ\ is calculatHd as a Iunction oI orH comSHtHnc\ in
tKis casH mHasurHd E\ ':i &KanJH in ':i ErinJs a SroSortionatH cKanJH in 6$* (cs
$ sHriHs oI Iactors arH tKHn aSSliHd to adMust tKH EasH casH 6$* sSHciÀc HnHrJ\ Ior orH
JHotHcKnical SroSHrtiHs, tKH oSHratinJ conditions, mill dimHnsions Ior H[amSlH, asSHct

FIG 12.7 – Ausgrind fSAG and efficiency improvement due to size reduction (image courtesy Ausenco).

210 Comminution Handbook | Spectrum Series 21

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

ratio and otKHrs 6omH oI tKHsH adMustinJ Iactors arH sKoZn in )iJurH 8 7KH 6$*
sSHciÀc HnHrJ\ calculation is dHscriEHd in tKH IolloZinJ HTuation
6$* (cs = >%asH &asH 6$* (cs@ >adMustinJ Iactors@

FIG 12.8 – Adjusting factors to SAG Ecs (images courtesy Ausenco).

Comminution Handbook | Spectrum Series 21 211

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

ZKHrH
6$* (cs is tKH 6$* mill sSHciÀc HnHrJ\
7KH SHEElH crusKHr sSHciÀc HnHrJ\ is calculatHd sHSaratHl\ and YHndor data is usHd Ior
HTuiSmHnt sHlHction 7KH Eall mill sSHciÀc HnHrJ\ is calculatHd as tKH rHsidual usinJ tKH
IolloZinJ HTuation
%all mill (cs = 7otal (cs ² 6$* (cs

)or circuits ZitKout a 6$* mill, tKH Eall mill sSHciÀc HnHrJ\ is HTual to tKH total
JrindinJ circuit sSHciÀc HnHrJ\ 6imilarl\, Ior circuits ZitKout Eall mills, tKH 6$* mill
sSHciÀc HnHrJ\ HTuals tKH total JrindinJ circuit sSHciÀc HnHrJ\ 0ill sKHll and motor si]H
sHlHction considHrs tKH sSHciÀc HnHrJ\ rHTuirHmHnts and tKH sSHciÀHd Slant tKrouJKSut,
toJHtKHr ZitK tKH H[SHctHd oSHratinJ HnYHloSH oI HacK mill

BIBLIOGRAPHY
Barratt, ' -, 8 $n uSdatH on tHstinJ, scalHuS and si]inJ HTuiSmHnt Ior autoJHnous and
sHmiautoJHnous JrindinJ circuits, in 3URFHHGLQJV$GYDQFHVLQ$XWRJHQRXVDQG6HPLDXWRJHQRXV
Grinding Technology (SAG 1989) Hds $ / 0ular and * ( $Jar, SS ² 8niYHrsit\ oI %ritisK
&olumEia 9ancouYHr
Barratt, ' - and $llan, 0 -, 8 7HstinJ Ior autoJHnous and sHmiautoJHnous JrindinJ a
dHsiJnHr·s Soint oI YiHZ, 0LQHUDOVDQG0HWDOOXUJLFDO3URFHVVLQJ, 80²
Bond, ) &,  7KH tKird tKHor\ oI comminution, 7UDQVDFWLRQRIWKH$,0(, 8²
Bond, ) &,  :orN indH[Hs taEulatHd, 0LQLQJ(QJLQHHULQJ, ²
Bond, ) &,  &rusKinJ and JrindinJ calculations, Sart ,, %ULWLVK&KHPLFDO(QJLQHHULQJ, 8²8
Bueno, 0 and /anH, *, 0 $ rHYiHZ oI 0 \Hars oI $*6$* Silot trials, in 3URFHHGLQJV
,QWHUQDWLRQDO $XWRJHQRXV *ULQGLQJ 6HPLDXWRJHQRXV *ULQGLQJ DQG +LJK 3UHVVXUH *ULQGLQJ 5ROO
Technology (SAG 2011) Hds K 0aMor, % & )lintoII, % KlHin and K 0c/Hod, 9ancouYHr
Burgess, ', 0 $ mHtKod oI calculatinJ autoJHnoussHmiautoJHnous JrindinJ mill sSHciÀc
HnHrJiHs usinJ a comEination oI %ond ZorN indicHs and -ulius KruttscKnitt SaramHtHrs, tKHn
aSSl\inJ HIÀciHnc\ Iactors, in 3URFHHGLQJV WK $XV,00 0LOO 2SHUDWRUV· &RQIHUHQFH, SS ²
7KH $ustralasian ,nstitutH oI 0ininJ and 0HtallurJ\ 0HlEournH
Dunne, 5, &KittHndHn, 5, /anH, * and 0orrHll, 6,  7KH &adia Jold coSSHr SroMHct H[Sloration
to startuS, SaSHr SrHsHntHd at 60( $nnual &onIHrHncH, 'HnYHr, 0arcK
Dunne, 5, 0orrHll, 6, /anH, *, 9alHr\, : and +art, 6, 00 'HsiJn oI tKH 0Ioot 6$* mill
installHd at tKH &adia Jold coSSHr minH, in 3URFHHGLQJV,QWHUQDWLRQDO&RQIHUHQFHRQ$XWRJHQRXV
DQG 6HPLDXWRJHQRXV *ULQGLQJ 7HFKQRORJ\ 6$*  Hds ' - %arratt, 0 - $llan and
$ / 0ular, SS ² 8niYHrsit\ oI %ritisK &olumEia 9ancouYHr
Hart, 6, 9alHr\, :, &lHmHnts, %, 5HHd, 0, 6onJ, 0 and 'unnH, 5, 00 2Stimi]ation oI tKH
&adia +ill 6$* mill circuit, in 3URFHHGLQJV ,QWHUQDWLRQDO &RQIHUHQFH RQ $XWRJHQRXV DQG 6HPL
DXWRJHQRXV *ULQGLQJ 7HFKQRORJ\ 6$*  Hds ' - %arratt, 0 - $llan and $ / 0ular,
SS ²0 8niYHrsit\ oI %ritisK &olumEia 9ancouYHr
Kosick, *, 'oEE\, * and %HnnHtt, &, 00 &((7 comminution Hconomic HYaluation tool Ior
comminution circuit dHsiJn and Sroduction SlanninJ, in 60($QQXDO0HHWLQJ, SrHSrint 0²
6ociHt\ Ior 0ininJ, 0HtallurJ\, and ([Sloration /ittlHton
Lane, *, )oJJiatto, % and %uHno, 0 3, 0 3oZHrEasHd comminution calculations usinJ
$usJrind, in 3URFHHGLQJV 7HQWK ,QWHUQDWLRQDO 0LQHUDO 3URFHVVLQJ &RQIHUHQFH (Procemin 2013)
Hds 0 $lYarH], $ 'oll, : KracKt and 5 Ku\YHnKoYHn, SS 8² *Hcamin

212 Comminution Handbook | Spectrum Series 21

 CHAPTER 12 t MILLING CIRCUIT CALCULATIONS

MacPherson, $ 5, 8 $ simSlH mHtKod to SrHdict tKH autoJHnous JrindinJ mill rHTuirHmHnts
Ior SrocHssinJ orH Irom a nHZ dHSosit, 7UDQVDFWLRQVRIWKH$,0(, ²0
Matthews, % ' and %arratt, ' -,  *rindSoZHr comSutHr EasHd SroJram Ior tKH si]inJ and
sHlHction oI JrindinJ circuits, in 3URFHHGLQJV6HFRQG&DQDGLDQ&RQIHUHQFHRQ&RPSXWHU$SSOLFDWLRQV
LQWKH0LQHUDO,QGXVWU\ Hds 5 3oulin, 5 & 7 3aNalnis and $ / 0ular, SS ² &anadian
,nstitutH oI 0ininJ, 0HtallurJ\ and 3HtrolHum 0ontrHal
Morrell, 6, 00 3rHdictinJ tKH sSHciÀc HnHrJ\ oI autoJHnous and sHmiautoJHnous mills Irom
small diamHtHr drill corH samSlHs, 0LQHUDOV(QJLQHHULQJ, ²
Morrell, 6, 00 5ocN cKaractHrisation Ior KiJK SrHssurH JrindinJ rolls circuit dHsiJn, in 3URFHHGLQJV
,QWHUQDWLRQDO$XWRJHQRXVDQG6HPLDXWRJHQRXV*ULQGLQJ7HFKQRORJ\, 9ancouYHr, SS ²8
Morrell, 6, 00 3rHdictinJ tKH oYHrall sSHciÀc HnHrJ\ rHTuirHmHnt oI crusKinJ, KiJK SrHssurH
JrindinJ roll and tumElinJ mill circuits,0LQHUDOV(QJLQHHULQJ, ²
Morrell, 6, 00 3rHdictinJ tKH sSHciÀc HnHrJ\ rHTuirHd Ior si]H rHduction oI rHlatiYHl\ coarsH IHHds
in conYHntional crusKHrs and KiJK SrHssurH JrindinJ rolls, 0LQHUDOV(QJLQHHULQJ, ²
Morrell, 6, 0 8sinJ tKH 60& 7Hst® to SrHdict comminution circuit SHrIormancH >onlinH@
$YailaElH Irom <KttSZZZsmctHstinJcomdocumHnts8sinJBtKHB60&B7HstSdI!
>$ccHssHd $uJust 0@
Morrison, 5 ' and 5icKardson, - 0, 00 -K6im0Ht a simulator Ior anal\sis, oStimisation
and dHsiJn oI comminution circuits, in 0LQHUDO 3URFHVVLQJ 3ODQW 'HVLJQ 3UDFWLFH DQG &RQWURO
3URFHHGLQJV Hds $ / 0ular, ' 1 +alEH and ' - %arratt, SS ²0 6ociHt\ Ior 0ininJ,
0HtallurJ\, and ([Sloration /ittlHton
Napier-Munn, 7 -, 0orrHll, 6, 0orrison, 5 ' and KoMoYic, 7, 00 0LQHUDO&RPPLQXWLRQ&LUFXLWV
7KHLU2SHUDWLRQDQG2SWLPLVDWLRQ -ulius KruttscKnitt 0inHral 5HsHarcK &HntrH %risEanH
Rowland Jr, & $ and KMos, ' 0, 80 5od and Eall mills, in 0LQHUDO 3URFHVVLQJ 3ODQW 'HVLJQ
Hds $ / 0ular and 5 % %KaSSu, sHcond Hdition, cKaStHr  6ociHt\ Ior 0ininJ, 0HtallurJ\,
and ([Sloration /ittlHton
Siddall, % and 3utland, %, 00 3rocHss dHsiJn and imSlHmHntation tHcKniTuHs Ior sHcondar\
crusKinJ to incrHasH millinJ caSacit\, in 60( $QQXDO 0HHWLQJ, SrHSrint 0²0 6ociHt\ Ior
0ininJ, 0HtallurJ\, and ([Sloration /ittlHton
Starkey, - and 'oEE\, *,  $SSlication oI tKH 0innoYH[ 6$* 3oZHr ,ndH[ at ÀYH &anadian
6$* Slants, in 3URFHHGLQJV,QWHUQDWLRQDO&RQIHUHQFHRQ$XWRJHQRXVDQG6HPLDXWRJHQRXV*ULQGLQJ
Technology (SAG 1996) Hds $ / 0ular, ' - %arratt and ' $ KniJKt, SS ²0 8niYHrsit\ oI
%ritisK &olumEia 9ancouYHr
Starkey, -, +indstrom, 6 and 1adasd\, *, 00 6$*'HsiJn tHstinJ ² ZKat it is and ZK\ it ZorNs,
in 3URFHHGLQJV,QWHUQDWLRQDO&RQIHUHQFHRQ$XWRJHQRXVDQG6HPLDXWRJHQRXV*ULQGLQJ7HFKQRORJ\
(SAG 2006) Hds 0 - $llan, K 0aMor, % & )lintoII, % KlHin and $ / 0ular, SS 0²
8niYHrsit\ oI %ritisK &olumEia 9ancouYHr

Comminution Handbook | Spectrum Series 21 213

 HOME

Chapter 13

MODELLING
COMMINUTION CIRCUITS
Hakan Dundar

INTRODUCTION
To carry out a simulation of a comminution circuit, the usual practice is to link the models
of each process unit together using a process model of each individual stream. A typical
simulation model of a stream contains, as an e[ample, the ÁoZ rates of each component
of interest. )or instance, typical components might Ee the ÁoZ rates of the Zater and the
mass for each si]e fraction of ore. )or convenience, the user interface might shoZ the total
mass ÁoZ of the dry solids, per cent solids and the si]e distriEution of the ore in per cent
retained. +ence to simulate the ÁoZ sheet, start Ey converting the feed stream into ÁoZ
rates Ey si]e and Zater ÁoZ. As ÁoZ rates can simply Ee added at a node, the Eall mill
product from the last iteration can Ee added to the feed stream Zith additional Zater if
reTuired and the total ÁoZ passed to the hydrocyclone model as its feed stream. ,n this
manner, a complete circuit simulation can be carried out and iterated to a numerical steady
state. This simulation strategy is called ‘sequential modular’. Another effective strategy is
to split the recycle streams into an input and an output version and iterate the circuit
until the tZo streams match. This strategy is called ‘tearing’. 0ular and 5ichardson 
provide a good description of various strategies for simulation. The rest of this chapter
describes some typical models and hoZ their parameters can be ‘Àtted’ or derived from
measured data.

PERFECT MIXING BALL MILL MODEL

A si]e reduction process in a comminution machine has been deÀned by many
researchers in different Zays using mathematical equations. 6everal modelling
approaches to predict the product size distribution of a comminution machine have
been developed so far. The most common and Zellaccepted approach for the mills
Zas proposed by :hiten ,  as the perfect mi[ing mill model. This approach

Comminution Handbook | Spectrum Series 21 215

 CHAPTER 13 t MODELLING COMMINUTION CIRCUITS

considers the mill as a perfectly mi[ed volume Zhere breakage occurs and the product
discharged from the mill is a function of the particle size. As shoZn in )igure .,
continuous breakage of the particles in a mill can be deÀned as
x breakage of the coarser particles into a particular size fraction
x further breakage of the particles in the particular size fraction
x particles at a particular size comes into the system Zith the neZ feed
x discharge of the particular size from the mill.
For a particular size fraction of i, transport into a breakage zone, breakage and transport
out for continuous milling is represented in Figure ..
A simple mass balance equation can be Zritten for a particular size fraction of i as
folloZs
i
pi = fi + / aijrjs j - risi ;i $ j .
j=1

Zhere
pi is the amount of i size product from the mill
fi is the amount of i size material in neZ feed
aij is the breakage function (generation of i size material after the breakage of coarser
j size material)
rj is the breakage rate of j size
sj is the amount of j size material Zithin the mill
ri is the breakage rate of i size
si is the amount of i size material Zithin the mill
The product from the mill is deÀned as a function of size. ,f the mill content is knoZn,
the product can be calculated using the folloZing equation (di is the discharge rate for
the i size material)
pi = di $ si (.)

FIG 13.1 – Product generation in a population balance model (image courtesy Tim Napier-Munn).

216 Comminution Handbook | Spectrum Series 21

 CHAPTER 13 t MODELLING COMMINUTION CIRCUITS

Then the mass balance equation can be Zritten as folloZs

i r r
pi = fi + / aij dj p j - di pi ;i $ j (.)
j=1 j i

,n most of the cases Zhere the mill content cannot be measured, the ratio rd can be
calculated for the given feed and product size distributions, and the breakage function.
The rd is regarded as tonnes per hour broken per tonne in the mill. The equations
beloZ shoZ the mass balances around some size intervals
r1 r
For i  p1 = f1 + >a1,1 p1H - 1 p1
d1 d1

r1 r r
For i  p2 = f2 + >a2, 1 p1 + a2, 2 2 p2H - 2 p2
d1 d2 d2

$ $
$ $
$ $
r1 r2 r r
For i = n p n = fn + >a n, 1 p +a p + g + a n, n n p nH - n p n (.)
d1 1 n, 2 d2 2 dn dn

2nly the rd parameters need to be knoZn to provide a complete Zay of calculating
the product sizes from a feed sizing and a suitable appearance function. This Zill be
demonstrated in the example given in the section titled ‘Fitting example for ball mill –
hydrocyclone circuit’.

EFFICIENCY CURVE MODEL FOR CLASSIFIERS

A size separation process can be described by its partition curve that shoZs the per cent
of each size fraction in the feed reporting in the coarse product. The variables that deÀne
it are the slope or sharpness of the linear section of the curve, the d value (Zhich
deÀnes its location on the graph) and the bypass, Zhich is the fraction of particles that
are misplaced into the coarse fraction. Figure . shoZs a typical partition curve for
classiÀers.

ϭϬϬ
/ĚĞĂů
ϵϬ
ƐĞƉĂƌĂƚŝŽŶ ZĞĂů
ϴϬ
ƐĞƉĂƌĂƚŝŽŶ
ϳϬ
ƉĂƌƚŝƚŝŽŶй

ϲϬ
ϱϬ
ϰϬ
ϯϬ
ϮϬ
&ŝƐŚ,ŽŽŬ
ϭϬ
LJƉĂƐƐ
Ϭ
G
ϭ ϭϬ ϭϬϬ ϭϬϬϬ
ŵĞĂŶƉĂƌƚŝĐůĞƐŝnjĞ;PŵͿ

FIG 13.2 – A typical partition curve for classifiers showing the parameters (image courtesy Hakan Dundar).

Comminution Handbook | Spectrum Series 21 217

 CHAPTER 13 t MODELLING COMMINUTION CIRCUITS

Whiten (1966) modelled the partition curve with the following equation:
R V
S 1 + b $ b) $ di $ a W
S f
d50c ]e - 1g W
p
Pi = 100 - R S W
S )
a$b $ i
d
W
S e d +e -2
50c
a
W
T X
where:
R indicates the bypass of the separator
a is the sharpness of the separation
ǃ LVWKHÀVKKRRNEHKDYLRXURIWKHFODVVLÀFDWLRQ
di is the mean particle size
d50c is the corrected cut size
ǃ* is the parameter that protects the d50c; that is, for Pi = 0.5(100-R) then di = d50c. Bypass
PHDQVWKHSURSRUWLRQRIWKHIHHGWKDWLVQRWFODVVLÀHGDQGLVVLPSO\WUDQVIHUUHGWRWKH
FRDUVHXQGHUÁRZSURGXFW$W\SLFDOK\GURF\FORQHE\SDVVLVSHUFHQWWRSHUFHQW
The sharpness of the separation ǂLVUHSUHVHQWHGE\WKHOLQHDUVHFWLRQLQ)LJXUH
,WLVUHDGIURPWKHHIÀFLHQF\FXUYHDQGLVW\SLFDOO\WZRWR$QLGHDOVHSDUDWLRQZRXOG
have an ǂYDOXHRIWHQRUJUHDWHUKRZHYHUZKHQWKHÀVKKRRNHIIHFWLVSUHVHQWWKHǃ
value interacts with the ǂ value.
7KHÀVKKRRNREVHUYHGLQWKHÀQHHQGRIWKHSDUWLWLRQFXUYHLVWKHSRUWLRQRIWKHYHU\
ÀQHSDUWLFOHVLQIHHGUHSRUWLQJWKHFRDUVHSURGXFW7KHÀVKKRRNLVUHSUHVHQWHGE\WKH
parameter ‘ǃ·LQWKH:KLWHQHIÀFLHQF\FXUYHPRGHODQGLWLVEDFNFDOFXODWHGE\GDWDÀW
8QOLNHWKHE\SDVVWKHǃ does not indicate a real value that is shown on the curve; it only
FRUUHVSRQGVWRWKDWYDOXH,IWKHUHLVQRÀVKKRRNǃ is zero. In some cases it can go up
to three or four.
7KLVPRGHOFDQEHH[WHQGHGWRK\GURF\FORQHFODVVLÀHUVXVLQJWKHDSSURDFKRI/\QFK
and Rao (1975).

FITTING EXAMPLE FOR BALL MILL – HYDROCYCLONE CIRCUIT

,QWKLVVHFWLRQDEDOOPLOO²K\GURF\FORQHFLUFXLWLVFRQVLGHUHGIRUPRGHOÀWWLQJ)LJXUH
JLYHV WKH ÁRZ VKHHW RI WKH FLUFXLW DQG PDVV EDODQFHG WRQQDJHV WKDW ZLOO EH UHTXLUHG
GXULQJPRGHOÀWWLQJ7KHGDWDFRPHIURPDVDPSOLQJVXUYH\SHUIRUPHGLQDJROGSODQW
Table 13.1 provides the mass balanced size distributions around the circuit.

FIG 13.3 – Ball mill – hydrocyclone flow sheet and mass balanced tonnages.
Water is added at the sump (image courtesy Hakan Dundar).

218 Comminution Handbook | Spectrum Series 21

 CHAPTER 13 t MODELLING COMMINUTION CIRCUITS

TABLE 13.1
Mass-balanced size distributions around the circuit.
Cumulative passing %
Size range class Hydrocyclone Hydrocyclone Hydrocyclone Ball mill
(μm) feed overflow underflow discharge
-3350/+2360 100.00 100.00 100.00 100.00
-2360/+1700 98.95 99.99 98.64 99.61
-1700/+1180 97.28 99.95 96.51 99.30
-1180/+850 94.46 99.89 92.90 98.80
-850/+600 91.37 99.82 88.95 98.06
-600/+425 88.21 99.77 84.91 96.80
-425/+300 84.52 99.71 80.19 94.38
-300/+212 79.11 99.55 73.29 89.43
-212/+150 73.08 99.38 65.59 83.29
-150/+106 63.27 99.22 53.04 71.99
-106/+74 51.10 97.49 37.89 57.45
-74/+53 38.35 89.91 23.68 42.06
-53/+38 29.29 78.48 15.00 31.32
-38/+25.5 23.31 66.76 10.05 24.41
-25.5/+18 19.45 56.97 7.40 20.04
-18/+12.5 16.56 48.32 5.75 16.94
-12.5/+9 14.33 41.90 4.55 14.56
-9 12.46 36.67 3.40 12.45

Ball mill fitting

According to the ÁoZ sheet, the hydrocyclone underÁoZ is the ball mill feed, Zhich is f
in the mass balance equation. The ball mill product is represented by p.
Table . gives the calculation of the f and p values for each size class. They are
both calculated by multiplying the mill throughput Zith retained per cents of feed and
product streams.
A sizeindependent breakage or appearance function aij Zas used in this example and
is shoZn in Table .. 6ize independent breakage means that each size class gives the
same progeny size distribution after breakage Zhen scaled to itself. +ence, the expected
progeny fraction at a particular ratio to the parent particle is the same; that is, a, = a,
or a, = a,, and so on.
After calculating the f and p values in Table . and using the breakage function in
Table ., rd values are calculated for each size as shoZn in Table .. &alculated
rd values Zere plotted against size in Figure .. The rd increases Zith particle size
and, depending on the ball size distribution, starts decreasing after a certain size, Zhich
means that the impact energies provided by the ball distribution in the mill are not
sufÀcient to break coarser particles as effectively as Àner ones.

Comminution Handbook | Spectrum Series 21 219

 CHAPTER 13 t MODELLING COMMINUTION CIRCUITS

TABLE 13.2
Calculation of the f and p values for each size class.
Size range class (μm) Size interval, i fi (t/h) pi (t/h)
-3350/+2360 1 2.44 = (100 – 98.64)/100 × 179.3 0.70 = (100 – 99.61)/100 × 179.3
-2360/+1700 2 3.82 = (98.64 – 96.51)/100 × 179.3 0.56 = (99.61 – 99.30)/100 × 179.3
-1700/+1180 3 6.47 0.90
-1180/+850 4 7.08 1.33
-850/+600 5 7.24 2.26
-600/+425 6 8.46 4.34
-425/+300 7 12.37 8.88
-300/+212 8 13.81 11.01
-212/+150 9 22.50 20.26
-150/+106 10 27.16 26.07
-106/+74 11 25.48 27.59
-74/+53 12 15.56 19.26
-53/+38 13 8.88 12.39
-38/+25.5 14 4.75 7.84
-25.5/+18 15 2.96 5.56
-18/+12.5 16 2.15 4.27
-12.5/+9 17 2.06 3.78
-9 Subsieve 6.10 22.32
Total 179.3 179.3

TABLE 13.3
Size-independent breakage function.
i j 1 2 . . 14 15 16 17
1 0
2 0.193 0
3 0.157 0.193 .
4 0.126 0.157 .
5 0.101 0.126 .
6 0.082 0.101 .
7 0.066 0.082 .
8 0.053 0.066 .
9 0.043 0.053 .
10 0.035 0.043 .
11 0.028 0.035 .
12 0.022 0.028 .
13 0.018 0.022 . .
14 0.015 0.018 . . 0

220 Comminution Handbook | Spectrum Series 21

 CHAPTER 13 t MODELLING COMMINUTION CIRCUITS

TABLE 13.3 CONT …

j 1 2 . . 14 15 16 17
i
15 0.012 0.015 . . 0.193 0
16 0.010 0.012 . . 0.157 0.193 0
17 0.008 0.010 . . 0.126 0.157 0.193 0
Subsieve 0.031 0.039 0.524 0.65 0.807 1
Total 1 1 1 1 1 1 1 1

TABLE 13.4
Calculation of r/d values for each size class using mass balance equation.
Size interval, i ri/di Mass balance equation for the given feed and product
particle size distribution and breakage function
r r
1 2.49 0.70 = 2.44 + <0 $ 1
d
$ 0.70F - d
1
$ 0.70
1 1

r r
2 6.42 0.56 = 3.82 + <0.193 $ 2.49 $ 0.70 + 0 $ d
2
$ 0.56F - 2
d
$ 0.56
2 2

3 7.26
4 5.86
5 3.60
6 1.90
7 0.85
8 0.66
9 0.35
Similar calculations to sizing interval -1700/+9 μm
10 0.24
11 0.11
12 0.05
13 0.04
14 0.02
15 0.011
16 0.008
r r
17 0.006 3.78 = 2.06 + <0.008 $ 2.49 $ 0.70 + 0.010 $ 6.42 $ 0.56 + g + 0 $ d
17
$ 3.78F - 17
d
$ 3.78
17 17

The operating variables of the mill have an impact on the breakage rate. Depending on
the operating variables the rd curve shifts upZards, doZnZards, leftZards, rightZards
and so on. Therefore a smooth rd curve is required for simulation purposes. 6ince it
is not possible to deÀne the rd curve Zith one unique equation, cubic spline function
is used to obtain a smooth curve. 6imply, three or four data points on the curve are
determined and used as an input to the cubic spline function. The data points are varied
in order to achieve the best Àt. Figure . gives the smoothed (Àtted) rd curve after
cubic spline. An example of calculating a cubic spline function is provided in 1apier
0unnHWDO (). Table . gives the calculated and Àtted rd values.

Comminution Handbook | Spectrum Series 21 221

 CHAPTER 13 t MODELLING COMMINUTION CIRCUITS

ϭϬ

ϭ
ƌͬĚ

Ϭ͘ϭ

Ϭ͘Ϭϭ
DƌͬĚ

Ϭ͘ϬϬϭ
ϭϬ ϭϬϬ ϭϬϬϬ ϭϬϬϬϬ
ŵĞĂŶƉĂƌƚŝĐůĞƐŝnjĞ;PŵͿ

FIG 13.4 – Calculated r/d (image courtesy Hakan Dundar).

ϭϬ

ϭ
ƌͬĚ

Ϭ͘ϭ

Ϭ͘Ϭϭ ĐĂůĐƵůĂƚĞĚƌͬĚ

ĨŝƚƚĞĚƌͬĚ
Ϭ͘ϬϬϭ
ϭϬ ϭϬϬ ϭϬϬϬ ϭϬϬϬϬ
ŵĞĂŶƉĂƌƚŝĐůĞƐŝnjĞ;PŵͿ

FIG 13.5 – Smoothed r/d curve (image courtesy Hakan Dundar).

TABLE 13.5
Calculated and fitted r/d values.
Size interval, i Calculated ri/di Fitted ri/di
1 2.49 2.80
2 6.42 6.30
3 7.26 8.00
4 5.86 6.44
5 3.60 3.64
6 1.90 2.07
7 0.85 1.17
8 0.66 0.66
9 0.35 0.37
10 0.24 0.21
11 0.11 0.12
12 0.05 0.07

222 Comminution Handbook | Spectrum Series 21

 CHAPTER 13 t MODELLING COMMINUTION CIRCUITS

TABLE 13.5 CONT …

Size interval, i Calculated ri/di Fitted ri/di
13 0.04 0.04
14 0.02 0.02
15 0.011 0.012
16 0.008 0.006
17 0.006 0.004

According to the Àtted rd values the product from the ball mill is backcalculated to
check the Àt quality. Table . gives the calculation steps of the product and the results
are plotted in Figure ..
As can be seen from Figure . the predicted product size distribution is a bit coarser
than the experimental one at Àne sizes. This results from the poor Àtting of rd at Àne
sizes in Figure ..

Hydrocyclone fitting
The actual and corrected partition curves for the hydrocyclone are calculated as
discussed in previous sections (Table .). 3artition curves Zere plotted starting from
TABLE 13.6
Prediction of the product using fitted r/d.
Size interval, i Calculated pi Mass balance equation for the given feed particle
(t/h) size distribution, breakage function and fitted r/d
1 0.64 p1 = 2.44 + [0 $ 2.80 $ p1] - 2.80 $ p1

2 0.57 p 2 = 3.82 + [0.193 $ 2.80 $ 0.64 + 0 $ 6.30 $ p2] - 6.30 $ p2

3 0.83
4 1.23
5 2.25
6 4.12
7 8.02
8 11.84
9 20.90
Similar calculations to sizing interval -1700/+9 μm
10 27.81
11 28.47
12 20.04
13 13.33
14 8.68
15 6.24
16 4.84
17 4.24 p17 = 2.06 + [0.008 $ 2.80 $ 0.64 + 0.010 $ 6.30 $ 0.57 + g + 0 $ 0.004 $ p17] - 0.004 $ p17
17
Subsieve 15.27 = 179.3 - / pi
i=1

Comminution Handbook | Spectrum Series 21 223

 CHAPTER 13 t MODELLING COMMINUTION CIRCUITS

ϭϬϬ
ϵϬ
ϴϬ
ϳϬ
ƉĂƐƐŝŶŐй ϲϬ
ϱϬ
ϰϬ
ϯϬ
ϮϬ DĚŝƐĐŚ͘
ϭϬ ĐĂůĐ͘DĚŝƐĐŚ͘
Ϭ
ϭ ϭϬ ϭϬϬ ϭϬϬϬ ϭϬϬϬϬ

FIG 13.6 – Experimental and calculated (predicted) product size distributions (image courtesy Hakan Dundar).

TABLE 13.7
Actual and corrected partitions.
Size class (μm) Mean size Partition (Uu/Ff) Corrected partition
-3350/+2360 2811.76 = 3350 # 2360 100.00 100.00
-2360/+1700 2003.00 99.30 98.80
-1700/+1180 1416.33 99.67 99.43
-1180/+850 1001.50 99.52 99.17
-850/+600 714.14 99.54 99.21
-600/+425 504.98 99.59 99.29
-425/+300 357.07 99.30 98.80
-300/+212 252.19 99.42 99.00
-212/+150 178.33 99.60 99.31
-150/+106 126.10 96.92 94.70
-106/+75 89.16 86.77 77.23
-75/+53 63.05 74.59 56.27
-53/+38 44.88 64.45 38.81
-38/+25.5 31.13 53.45 19.88
-25.5/+18 21.42 44.45 4.39
-18/+12.5 15.00 41.90 0.00
-12.5/+9 10.61 47.88 10.29

  microns to eliminate the scatter at coarser sizes (Figure .). As can be seen
from Figure . the bypass of the hydrocyclone exceeds  per cent and dc is less
than  microns.
The actual partition curve is then Àtted to an efÀciency curve model to backcalculate
the model parameters such as ǂ, ǃ and ǃ by iterative solution. Table . gives the
calculation of the Àtted partition curve using model parameters. Figure . gives the
actual and Àtted partition curves for the hydrocyclone.

224 Comminution Handbook | Spectrum Series 21

 CHAPTER 13 t MODELLING COMMINUTION CIRCUITS

ϭϬϬ
ϵϬ
ϴϬ
ϳϬ 3350 ) 2360

ƉĂƌƚŝƚŝŽŶ й
ϲϬ
ϱϬ
ϰϬ
ϯϬ
ϮϬ ĂĐƚƵĂů
ϭϬ ĐŽƌƌĞĐƚĞĚ
Ϭ
ϭ ϭϬ ϭϬϬ ϭϬϬϬ
ŵĞĂŶƉĂƌƚŝĐůĞƐŝnjĞ;PŵͿ

FIG 13.7 – Actual and corrected partition curves for the hydrocyclone (image courtesy Hakan Dundar).

TABLE 13.8
Fitted partition curve and model parameters.
Mean size Fitted partition Fitted parameters
R V R = 41.90
S b1 + 1.65 $ 3.40 $ 2811.76 l $ ^e 0.64 - 1h W
100.00 = 100 - 41.90 $ S
56.52 W α = 0.64
2811.76 S 2811.76 W
S e0.64 $ 3.40 $ 56.52 + e0.64 - 2 W β = 1.65
T X β* = 3.40
2003.00 100.00 d50c = 56.52
1416.33 100.00
1001.50 100.00
714.14 100.00
504.98 100.00
357.07 100.00
252.19 99.94
178.33 99.27
R V
S b1 + 1.65 $ 3.40 $ 126.10 l $ ^e 0.64 - 1h W
S 56.52 W
126.10 96.04 = 100 - 41.90 $
S 0.64 $ 3.40 $ 126.10 W
S e 0.64
56.52 + e - 2 W
T X
89.16 88.01
63.05 75.71
44.88 62.91
31.13 52.16
21.42 46.08
15.00 44.29
10.61 44.96

The sharpness of the efÀciency curve is Zidely used to assess the performance of the
separation; hoZever, the correlation betZeen the operating variables and the model
parameters should be assessed in order to use it for simulation purposes.

Comminution Handbook | Spectrum Series 21 225

 CHAPTER 13 t MODELLING COMMINUTION CIRCUITS

ϭϬϬ
ϵϬ
ϴϬ
ƉĂƌƚŝƚŝŽŶ й ϳϬ
ϲϬ
ϱϬ
ϰϬ
ϯϬ
ϮϬ ĂĐƚƵĂů
ϭϬ ĨŝƚƚĞĚ
Ϭ
ϭ ϭϬ ϭϬϬ ϭϬϬϬ
ŵĞĂŶƉĂƌƚŝĐůĞƐŝnjĞ;PŵͿ

FIG 13.8 – Actual and fitted partition curves (image courtesy Hakan Dundar).

These models are straightforZard to implement in modern spreadsheets. They can

also be linked into circuits Zithin a spreadsheet. For closed circuit simulations, run the
spreadsheet in singlestep mode and Zatch for convergence.
As a note of caution, the sizes used in the example breakage functions are in a square
root of tZo decreasing series. +ence, it is essential to use a similar decreasing series for
the size fractions in the simulation.

CONCLUSIONS
The techniques outlined in this chapter provide poZerful tools for modelling and
optimising simple existing circuits based on Zellmeasured data. The use of such models
for design requires a substantial collection of measured data (or orespeciÀc pilot plant
data) and extensive experience in operations as Zell as modelling of comminution
processes.
For complex circuits, ÁoZ sheet based systems Zith integrated data analysis and
modelling capabilities are available commercially. Training in hoZ to use these systems
is also essential.

BIBLIOGRAPHY
Lynch, A - and 5ao, T &, . 0odelling and scale up of hydrocyclone classiÀers, in PURcHHdiQJV
;,,QWHUQDWiRQDO0iQHUDOPURcHVViQJ&RQJUHVV, &aglari, paper .
Mular, A / and 5ichardson, - 0, . 0etallurgical balances, in 'HViJQ DQd ,QVWDOODWiRQ RI
&RQcHQWUDWiRQDQd'HZDWHUiQJ&iUcXiWV, chapter , pp – (6ociety for 0ining, 0etallurgy,
and (xploration /ittleton).
Napier-Munn, T -, 0orrell, 6, 0orrison, 5 D and .ojovic, T, . 0iQHUDO&RPPiQXWiRQ&iUcXiWV
7KHiU2SHUDWiRQDQd2SWiPiVDWiRQ,  p (-ulius .ruttschnitt 0ineral 5esearch &entre Brisbane).
Whiten, : -, . /ecture notes for :inter 6chool on 0ineral 3rocessing, Department of 0ining
and 0etallurgical (ngineering, University of 4ueensland (unpublished).
Whiten, : -, . A matrix theory of comminution machines, &KHP(QJ6ci, –.
Whiten, : -, . Ball mill simulation using small calculators, 7KH$XV,00PURcHHdiQJV, –.

226 Comminution Handbook | Spectrum Series 21

 HOME

Chapter 14

PROCESS CONTROL
Bob Yench, Gunter Metzner and Alban Lynch

INTRODUCTION
Every process plant involves process control. Process control can be manual, automatic
or, as is usually the case, a mixture of both. The process engineer designs the process
to meet the production requirements of the plant. It is then arranged into a coherent
SODQW IRUPDW DV GHSLFWHG E\ WKH SURFHVV ÁRZ GLDJUDP 3)' 7KH SULPDU\ DLP RI
SURFHVVFRQWUROLVWRPD[LPLVHHIÀFLHQF\LQWKHSURFHVVZKLFKFDQEHGHÀQHGDVVDIHO\
DFKLHYLQJPD[LPXPSURGXFWLRQDWPLQLPXPFRVWZLWKPLQLPXP¶RIIVSHF·SURGXFW
,QGHFLGLQJZKLFKRSHUDWLRQVZLWKLQWKHSODQWZLOOEHPDQXDOO\RSHUDWHGDQGZKLFK
ZLOO EH XQGHU DXWRPDWLF FRQWURO WKH SURFHVV HQJLQHHU ² SHUKDSV LQ FRQVXOWDWLRQ ZLWK
SODQWRSHUDWRUVDQGLQVWUXPHQWHQJLQHHUV²ZLOOZHLJKXSPDQ\IDFWRUVLQFOXGLQJWKH
FULWLFDOLW\RIWKDWSRUWLRQRIWKHSURFHVVDKLJKFULWLFDOLW\SRLQWVWRDXWRPDWLFFRQWURO
DQGWKHQHHGIRUWKHRSHUDWRUWRPDLQWDLQDKDQGVRQIHHOIRUWKHSURFHVVZKHUHDKLJK
QHHGSRLQWVWRPDQXDOFRQWURO,GHDOO\RSHUDWLQJPHWDOOXUJLVWVVKRXOGIROORZWKHSODQW
FRQWUROV\VWHPGHYHORSPHQWIURPFRQFHSWWRFRPPLVVLRQLQJDQGEH\RQG7KH\QHHG
WRKDYHHQRXJKNQRZOHGJHWRDVNSHUWLQHQWTXHVWLRQVWKURXJKRXWWKLVSURFHVVEHFDXVH
the plant control system is the single most important tool available in getting the best
RXWRIWKHSODQW8QIRUWXQDWHO\IHZPHWDOOXUJLVWVDUHJLYHQWKHRSSRUWXQLW\WRWDNHD
FRQWURO V\VWHP IURP FRQFHSW WR RSHUDWLRQ ZLWK GHVLJQ DQG LPSOHPHQWDWLRQ QRUPDOO\
handled by a separate project team and then given to operations.
)LJXUH  VKRZV WKH OD\HUV LQ FRQWURO IXQFWLRQV 6RPH RI WKH IXQFWLRQV LQ WKH
OD\HUVDUHSHUIRUPHGDXWRPDWLFDOO\ZKLOHRWKHUVDUHSHUIRUPHGPDQXDOO\E\SHRSOH
At the top of the hierarchy of controls for a grinding circuit is, for example, senior
PDQDJHPHQWDQGDWWKHERWWRPLVÀHOGLQVWUXPHQWDWLRQ7KHEDVHRIWKHKLHUDUFK\LV
the grinding circuit.
Plant operators and metallurgists are often required to operate over several layers
RI WKH KLHUDUFK\ RI FRQWUROV SDUWLFXODUO\ LQ D SODQW ZLWK PLQLPDO SURFHVV FRQWURO DQG
DXWRPDWLRQ6LQFHWKHFRQWUROV\VWHPLVWKHSULPDU\WRROIRUPDQDJLQJWKHRSHUDWLRQRI

Comminution Handbook | Spectrum Series 21 227

 CHAPTER 14 t PROCESS CONTROL

FIG 14.1 – Layering of process control (image courtesy Philip Thwaites).

WKHSODQWPHWDOOXUJLVWVPXVWKDYHNQRZOHGJHDQGSUDFWLFDOVNLOOLQDQXPEHURIDUHDV
LQFOXGLQJ
x the process that is being controlled
x ÀHOGLQVWUXPHQWDWLRQDQGDFWXDWRUV
x relevant process control techniques and skills; for example, loop tuning
x capability of the plant control systems; for example, distributed control system
'&6
0DQDJLQJ RSHUDWLQJ DQG PDLQWDLQLQJ D SODQW SURFHVV FRQWURO V\VWHP LV D WHDP
EDVHG XQGHUWDNLQJ UHTXLULQJ NQRZOHGJH LQ SURFHVV FRQWURO WHFKQLTXHV DQG ÀHOG
LQVWUXPHQWDWLRQDVZHOODVVNLOOVLQSURJUDPPLQJGDWDDQDO\VLVDQGRSHUDWRUWUDLQLQJ
,QDGGLWLRQWKHSODQWPHWDOOXUJLVWPXVWKDYHWKHSURFHVVFRQWURONQRZOHGJHWRHQVXUH
process performance is optimised.

PROCESS MEASUREMENT
To control a process it must be possible to measure relevant parameters of the process.
7KHVH FRPPRQO\ LQFOXGH ÁRZ WHPSHUDWXUH SUHVVXUH DQG GHQVLW\ IRU YDULRXV SDUWV
RIWKHSURFHVV)RUPDQ\RIWKHVHSDUDPHWHUVWKHUHZLOOEHDFKRLFHRIPHDVXUHPHQW
techniques, so it pays to take care in selecting the best device for the purpose. It is
critical that issues such as sizing, materials, environmental protection, installation
requirements and access for maintenance are fully considered for each installation.
2QO\ZKHQDOORIWKHVHPDWWHUVDUHFRQVLGHUHGZLOOWKHUHEHFRQÀGHQFHLQORQJWHUP
reliable and repeatable measurements.
Process measurement techniques do not focus on ultimate accuracy. Good accuracy
is important but repeatability is essential. Repeatable measurements provide the
FRQWUROV\VWHPZLWKLQIRUPDWLRQIRUDVWDEOHRSHUDWLRQ,WLVDFRPPRQPLVFRQFHSWLRQ
DPRQJSODQWRSHUDWRUVWKDWIRUH[DPSOHLI\RXDUHZHLJKLQJPLOOIHHGRQDFRQYH\RU
ZHLJKWRPHWHUWKHWRWDOWRQQHVFDQEHXVHGIRUPHWDOEDODQFHDFFRXQWLQJRUFXVWRG\
transfer purposes. In general, a different class of instrumentation is required for these
purposes.

228 Comminution Handbook | Spectrum Series 21

 CHAPTER 14 t PROCESS CONTROL

Flow measurement
6LQFH ÁRZ LV RQH RI WKH PRVW FRPPRQ PHDVXUHPHQWV UHTXLUHG LQ PLQHUDO SURFHVVLQJ
DQGWKHUHLVDZLGHYDULDWLRQLQSULQFLSOHVHPSOR\HGZKDWIROORZVLVDVKRUWGLVFXVVLRQ
RQVRPHFRPPRQÁRZPHDVXUHPHQWSULQFLSOHV
7KHIROORZLQJ¶FRQWLQXRXV·ÁRZPHDVXUHPHQWGHYLFHVDQGWHFKQLTXHVDUHFRPPRQO\
XVHGIRUSURFHVVPHDVXUHPHQWDQGFRQWURO$OOFRQWLQXRXVÁRZPHDVXUHPHQWWHFKQLTXHV
DUH DSSOLFDWLRQVSHFLÀF DQG UHTXLUH FDUHIXO VHOHFWLRQ VL]LQJ DQG SRVLWLRQLQJ LQ WKH
SURFHVVWRDFKLHYHWKHLUGHVLJQVSHFLÀFDWLRQV7KHUHDUHDOVRSRVLWLYHGLVSODFHPHQWÁRZ
PHDVXUHPHQWGHYLFHVVXFKDVGRVLQJSXPSVRUSDGGOHW\SHÁRZPHWHUVZKLFKDUHPRUH
commonly used in chemical dosing or batch systems.
0DJnHtLFÁoZPHtHU²FRPPRQO\XVHGIRUFRQGXFWLYHOLTXLGVVXFKDVZDWHURUPLQHUDO
VOXUU\ 7KH FRQGXFWLYH OLTXLG ÁRZLQJ WKURXJK DQ HOHFWURPDJQHWLF ÀHOG FUHDWHV DQ
HOHFWULFDORXWSXWWKDWLVSURSRUWLRQDOWRWKHDYHUDJHÁRZUDWH,WPXVWEHLQVWDOOHGLQD
ULVLQJOLQHWRHQVXUHDIXOOSLSHDQGKDVDQRUPDOÁRZUDQJHWXUQGRZQRI
2ULÀFHSODtH²XVHGRQJDVHVDQGQRQDEUDVLYHOLTXLGV7KHSODWHREVWUXFWVWKHÁRZDQG
FUHDWHVDSUHVVXUHGURSWKDWLVSURSRUWLRQDOWRWKHÁRZUDWH,WPXVWEHRQDULVLQJOLQHIRU
OLTXLGDSSOLFDWLRQVDQGKDVDQRUPDOÁRZUDQJHWXUQGRZQRI
$YHUDJLnJSLtot²XVHGRQJDVHVDQGFOHDQOLTXLGV7KLVFRQVLVWVRIDKROORZEDULQVWDOOHG
across a diameter of the process pipe. The bar creates a pressure drop proportional to
WKHÁRZUDWH6XSSOLHUVFODLPWXUQGRZQFDSDELOLW\EXWLQSUDFWLFHWKLVLVGLIÀFXOWWR
DFKLHYHEHFDXVHRIWKHORZGLIIHUHQWLDOSUHVVXUHJHQHUDWHGE\WKHVHGHYLFHV
9oUtH[ÁoZPHtHU²WKHPRVWFRPPRQW\SHRIYRUWH[PHWHULQWKHPLQHUDOSURFHVVLQJ
LQGXVWU\LVWKHYRUWH[VKHGGLQJGHYLFHZKLFKXWLOLVHVDVSHFLDOO\VKDSHGEDUDFURVVWKH
GLDPHWHURIWKHSLSHWR¶VKHG·YRUWLFHVLQWKHÁRZLQJOLTXLG7KHVHYRUWLFHVDUHGHWHFWHG
E\VHQVRUVRQWKHEDFNRIWKHVKHGGHUEDU7KH\DUHUHODWLYHO\FKHDSDQGUREXVWZLWK
D JRRG WXUQGRZQ EXW DUH QRW VXLWDEOH IRU YLVFRXV RU GLUW\ ÁXLGV 7KH\ DUH JHQHUDOO\
XQVXLWDEOHIRUOLQHVODUJHUWKDQPPGLDPHWHU
&oULoOLV PHtHU ² D UHODWLYHO\ UHFHQW GHYLFH WKDW LV FDSDEOH RI PHDVXULQJ PDVV ÁRZ RI
OLTXLGV VOXUULHV DQG KLJKSUHVVXUH JDVHV ZLWK D ODUJH GHJUHH RI DFFXUDF\ DQG D WXUQ
GRZQRI7KHÁXLGSDVVHVWKURXJKDWXEHWKDWLVVXEMHFWHGWRDYLEUDWLQJIRUFHDQG
WKHUHVLVWDQFHWRWKHYLEUDWLQJIRUFHLVSURSRUWLRQDOWRWKHPDVVÁRZRIÁXLGWKURXJKWKH
WXEHV,WLVOLPLWHGWRDPD[LPXPRIDURXQGPPGLDPHWHUOLQHDQGLVEHVWVXLWHG
to clean liquids.

CONTROL ELEMENTS
7KHFRQWUROV\VWHP¶SURFHVVHV·WKHLQSXWVLJQDOVIURPWKHPHDVXUHPHQWGHYLFHVDSSOLHV
the logic that is programmed into its processors and then sends signals out to the plant
¶FRQWUROHOHPHQWV·WRFRQWLQXHRUPRGLI\WKHSURFHVVRSHUDWLRQ0RVWFRPPRQO\WKHVH
DFWXDWRUVIDOOLQWRRQHRIWZRFDWHJRULHVYDULDEOHVSHHGGULYHVRUDFWXDWRUVIRUYDOYHVRU
dampers. Variable speed drives are used to control conveyors, mills, pumps and fans
ZKLOHDFWXDWRUVFRQWUROYDOYHDQGGDPSHUSRVLWLRQVWRYDU\WKHÁRZRIÁXLGVDQGJDVHV
$VKRUWGLVFXVVLRQRQWKHUROHRIDFWXDWRUVLQDSURFHVVSODQWIROORZV

Comminution Handbook | Spectrum Series 21 229

 CHAPTER 14 t PROCESS CONTROL

Actuators
Processes are regulated by applying a control signal to an actuator in the process.
7KHÁRZRIOLTXLGVRUJDVHVLQWKHSURFHVVSLSHZRUNPD\EHUHJXODWHGE\DFRQWURO
valve that is automatically positioned by the control system. In every case a device
DFWXDWRULVUHTXLUHGWRDPSOLI\WKHORZOHYHOFRQWUROVLJQDOWRDOHYHOWKDWFDQFRQWURO
WKHSURFHVV7KHSRZHUVRXUFHIRUDFWXDWRUVFDQEHSQHXPDWLFFRPSUHVVHGDLULVPRVW
FRPPRQK\GUDXOLFRUHOHFWULF
0RVWYDOYHVZLWKDFWXDWRUVDUHDOVRÀWWHGZLWKDFXUUHQWWRSQHXPDWLFYDOYHSRVLWLRQHU
This device regulates the supply of air to the actuator in response to an electrical current
signal from the control system. It also incorporates mechanical feedback to ensure that
WKHDFWXDWRULVSRVLWLRQHGLQSURSRUWLRQWRWKHFXUUHQWVLJQDO7ZRW\SHVRIDLUDFWXDWRU
are available.
&\OLnGHU DFtXDtoUV ² WKLV DFWXDWRU LV QRZ WKH PRVW FRPPRQO\ XVHG DLU DFWXDWRU LQ
WKH SURFHVV LQGXVWULHV 7KH SRVLWLRQLQJ IRUFH LV FUHDWHG E\ D SLVWRQ RSHUDWLQJ ZLWKLQ
D F\OLQGHU 7KH\ DUH JHQHUDOO\ GRXEOH DFWLQJ ZKHUH VXSSO\ SUHVVXUH DLU LV DSSOLHG WR
HDFKHQGRIWKHF\OLQGHU7KH\UHVSRQGTXLFNO\DUHKLJKO\SRZHUIXODQGDUHVLPSOHWR
FRQVWUXFWDVZHOODVEHLQJHDV\WRPDLQWDLQDQGFDOLEUDWH
'LDSKUDJPDFtXDtoUV²WKHVHZHUHSRSXODULQWKHHDUOLHUGD\VRISQHXPDWLFWUDQVPLVVLRQ
ZKHUHWKHVLJQDOZDVSUHVHQWHGGLUHFWO\WRWKHGLDSKUDJPZLWKRXWDQ\FRQYHUVLRQRU
DPSOLÀFDWLRQ7KHIRUFHFUHDWHGE\WKHGLDSKUDJPLVRSSRVHGE\DVXEVWDQWLDOVSULQJE\
ZKLFKPHDQVWKHVKDIWWDNHVXSDSRVLWLRQLQSURSRUWLRQWRWKHVLJQDOSUHVVXUH$VHSDUDWH
,3FXUUHQWWRSQHXPDWLFFRQYHUWHUPXVWEHXVHGLIDIHHGEDFNSRVLWLRQHULVQRWÀWWHG
(OHFtULFDFtXDtoUV²WKHVHDUHFRPPRQLQWKHSRZHULQGXVWU\$QHOHFWULFPRWRUSRZHUHG
by mains voltage is used as the driving component. This motor is reversible and is
controlled by a converter/positioner using mechanical feedback. Electric actuators are
uncommon in the process industries due to their high initial cost.
9DOYHV²WKHVHGHYLFHVLQVRPHIRUPDUHWKHSULPDU\SURFHVVFRQWUROGHYLFHLQWKHSURFHVV
industries. There are many types of valves available, and industry itself has an even
JUHDWHUQXPEHURIGLIIHUHQWDSSOLFDWLRQVIRUWKHVHGHYLFHV7KH\IDOOLQWRWZRFDWHJRULHV
on/off valves or regulatory valves. The interests of the measurement and control industry
DUHODUJHO\FRQÀQHGWRWKRVHRQRIIRUUHJXODWRU\YDOYHVWKDWDUHRSHUDWHGE\DUHPRWHO\
JHQHUDWHGVLJQDODQGQRWKDQGRSHUDWHG7KHPDMRUW\SHVRIYDOYHVXVHGLQWKHSURFHVV
LQGXVWULHVLQFOXGHEDOOEXWWHUÁ\JOREHGLDSKUDJPSLQFKSOXJDQGJDWHYDOYHV(DFKLV
suited to a particular range of applications according to the process conditions. A poor
choice in valve type, sizing, construction materials or positioning in the process for
DQ\ JLYHQ DSSOLFDWLRQ FDQ UHVXOW LQ DQ\WKLQJ IURP LQHIÀFLHQW SURFHVV FRQWURO DQG KLJK
maintenance costs through to catastrophic failure in critical applications.

TRANSMISSION
:KLOHWKHUHDUHPDQ\¶ORFDO·PHDVXUHPHQWGHYLFHVIRUH[DPSOHSUHVVXUHJDXJHVDQG
¶ORFDO· FRQWURO HOHPHQWV IRU H[DPSOH KDQGRSHUDWHG YDOYHV LQ HYHU\ SURFHVV SODQW
there are also many measurement and control devices that are required to either send or
UHFHLYHVLJQDOVWRIURPDFHQWUDOLVHGFRQWUROV\VWHP6XFKFRPPXQLFDWLRQVDUHDFKLHYHG
by producing a robust signal at the sending end and conveying it over one of a number
of transmission mediums to the receiving device.

230 Comminution Handbook | Spectrum Series 21

 CHAPTER 14 t PROCESS CONTROL

Many modes of transmission have been developed and employed over the years.
$SRSXODUPHGLXPXSWRWKHPLGVZDVSQHXPDWLFXVLQJFRPSUHVVHGDLULQDUDQJH
RI ² SVL SRXQGV SHU VTXDUH LQFK 9DULRXV HOHFWULFDO FXUUHQW DQG YROWDJH VLJQDOV
ZHUH VRPHWLPHV XVHG 7KHVH GD\V PRVW VLJQDOV DUH WUDQVPLWWHG HOHFWULFDOO\ XVLQJ WKH
XQLYHUVDOO\DFFHSWHG²P$WUDQVPLVVLRQVLJQDO$SURFHVVSUHVVXUHLQWKHUDQJHRI
VD\²N3DZRXOGEHFRQYHUWHGLQWRDQHTXLYDOHQW²P$3QHXPDWLFVLJQDOVDUH
VWLOOXVHGLQPDQ\LQVWDQFHVWRDFWXDWHUHJXODWLQJGHYLFHVVXFKDVGLDSKUDJPDFWXDWHG
FRQWUROYDOYHV7KHFRPPDQGVLJQDOLVFRQYH\HGHOHFWULFDOO\WRWKHÀHOGXVLQJ²P$
WUDQVPLVVLRQ$WWKHGLDSKUDJPDFWXDWRUWKHVLJQDOZLOOEHFRQYHUWHGE\DQ,3FXUUHQW
WRSQHXPDWLFFRQYHUWHUWRDSQHXPDWLFVLJQDOLQWKHUDQJHRI²N3D
As has been noted, compressed air is used in many instances to provide the actuating
SRZHUIRUUHJXODWLQJGHYLFHVVXFKDVDXWRPDWLFFRQWUROYDOYHVDQGSRVLWLRQLQJDFWXDWRUV
7KHVH GHYLFHV UHTXLUH GU\ FOHDQ DLU WR ZRUN HIÀFLHQWO\ &RPSUHVVHG DLU SURGXFHG
IRUJHQHUDOSODQWXVHLVQRWVXLWDEOHIRUXVHZLWKLQVWUXPHQWDWLRQHTXLSPHQW*HQHUDO
SODQWFRPSUHVVHGDLUPXVWEHIXUWKHUÀOWHUHGDQGGULHGEHIRUHXVHIRULQVWUXPHQWDWLRQ
SXUSRVHVDQGWKHQRUPDOSUDFWLFHLVIRUDVHSDUDWHÀOWHUGU\HUXQLWWREHLQVWDOOHGDQG
air to be reticulated via a dedicated instrument air ring main. Instrument air systems
QHHGWREHFDUHIXOO\GHVLJQHGIRUH[DPSOHKDYLQJWKHDSSURSULDWHÀOWHUVDQGGU\LQJ
HOHPHQWVIRUWKHDPELHQWFOLPDWHLQVWDOOHGVXFKDVOLQHVEHLQJVORSHGWRZDWHUWUDSV
DQGFRPPLVVLRQHGWKRURXJKFOHDQLQJEHIRUHFRQQHFWLRQWRLQVWUXPHQWDWLRQWRSURYLGH
ORQJWHUPUHOLDEOHVHUYLFH
7KH GHYHORSPHQW RI GLJLWDO QHWZRUNV LV VHHLQJ UDSLG FKDQJHV LQ WKH PRGHV RI
GDWD WUDQVPLVVLRQ 0DQ\ FKDQQHOV RI LQIRUPDWLRQ FDQ QRZ EH TXLFNO\ DQG FKHDSO\
WUDQVPLWWHG DFURVV D VLQJOH FRQGXFWRU WKXV VLJQLÀFDQWO\ UHGXFLQJ WKH EXON DQG FRVW
RI ÀHOG ZLULQJ 7KH DSSDUHQW EHQHÀWV RI VXFK V\VWHPV ZLOO QRW EH UHDOLVHG LQ DOO
applications and the need for a higher level of trade skills to install, maintain and
IDXOWÀQGPHDQVWKDWWKH\VKRXOGEHFRQVLGHUHGRQDFDVHE\FDVHEDVLV7KHSURFHVV
DQGLQVWUXPHQWDWLRQGLDJUDP3 ,'XVHVGLIIHUHQWOLQHVW\OHVWRGLIIHUHQWLDWHVLJQDO
WUDQVPLVVLRQ W\SHV 7KHVH VW\OHV KDYH LQWHUQDWLRQDO VWDQGDUGV IRU DXWRPDWLRQ ,6$
applied to them.

PROCESS CONTROL SYSTEMS AND DEVICES

Many process measurements are made and much data is collected and stored in the
average process plant. In order to provide useful information to operating personnel
and automatically control these variables, a means of conveying this information to
WKH RSHUDWRU DQG DXWRPDWLFDOO\ PDQLSXODWLQJ WKH ÀHOG UHJXODWLQJ GHYLFHV LV UHTXLUHG
7KHV\VWHPVXVHGWRFDUU\RXWWKHVHIXQFWLRQVFDQEHGLYLGHGLQWRWZREURDGJURXSV
 3DnHOPoXntHG LnGLFDtLnJ UHFoUGLnJ DnG FontUoOOLnJ GHYLFHV. The application of this
JURXS LV QRZ OLPLWHG WR VPDOO SURFHVV SODQWV UHTXLULQJ OLPLWHG PHDVXUHPHQW DQG
FRQWURODQGODUJHUSODQWVUHTXLULQJORFDODX[LOLDU\ÀHOGFRQWUROSDQHOV
2. 6FUHHnEDVHGPLFUoSUoFHVVoUEDVHGFontUoODnGGDtDFoOOHFtLnJV\VtHPV. The introduction
DQG UHÀQHPHQW RI WKH PLFURSURFHVVRU LQ WKH HDUO\ V VDZ UDSLG FKDQJHV RFFXU
LQ SURFHVV FRQWURO HTXLSPHQW DYDLODEOH WR LQGXVWU\ 7ZR SURFHVV FRQWURO FRQFHSWV
GHYHORSHGVHSDUDWHO\ FDPHIURPGLIIHUHQW HQJLQHHULQJ GLVFLSOLQHV 7KH SRZHU DQG
YHUVDWLOLW\RIWKHPLFURSURFHVVRUZDVUHFRJQLVHGE\ERWKJURXSV

Comminution Handbook | Spectrum Series 21 231

 CHAPTER 14 t PROCESS CONTROL

One group represented the process industries that relied heavily on regulatory
RU DQDORJXH FRQWURO 6RPH RI WKHVH LQGXVWULHV LQFOXGH WKH SHWURFKHPLFDO JHQHUDO
FKHPLFDOSXOSDQGSDSHUSRZHUDQGVWHDPUDLVLQJLQGXVWULHV$OORIWKHVHUHTXLUHG
WKH PRGXODWLQJ RI YDULDEOHV WR ÀQLWH YDOXHV DQG FRQWLQXRXV RSHUDWLRQ $QDORJXH
IXQFWLRQV SUHYLRXVO\ FDUULHG RXW LQ SDQHOPRXQWHG HTXLSPHQW ZHUH UHÁHFWHG LQ WKH
ORJLFRIWKHQHZPLFURSURFHVVRUGHYLFHVDQGWKHGHYHORSPHQWZDVLQÁXHQFHGJUHDWO\
E\ WKH LQSXW IURP SURFHVV FRQWURO SHRSOH 7KXV WKH '&6 HYROYHG %HLQJ EDVLFDOO\
D GLJLWDO GHYLFH LW ZDV HDV\ WR LQFRUSRUDWH VHTXHQWLDO ORJLF DQG WHFKQLTXHV VXFK DV
PRWRU FRQWURO KRZHYHU LWV SULQFLSDO SXUSRVH ZDV SURFHVVLQJ DQDORJXH VLJQDOV DQG
providing modulating control.
The second group evolved from the electrical engineering task of automated plant
FRQWURO ZKHUH WKH UHTXLUHPHQW ZDV SULPDULO\ WR VZLWFK LQGXVWULDO FRPSRQHQWV RQ RU
RII DQG LQWHUORFN WKH FRQWURO DFFRUGLQJO\ 7KH SURJUDPPDEOH ORJLF FRQWUROOHU 3/&
ZDVGHYHORSHGWRXVHWKHVSHHGDQGUHOLDELOLW\RIWKHPLFURSURFHVVRUWRUHSODFHUHOD\V
WLPHUVDQGLQWHUFRQQHFWLQJZLULQJ7KHVHTXHQWLDOFRQWUROIHDWXUHVZHUHGHYHORSHGDQG
UHÀQHGWRDKLJKGHJUHHDQGWKH3/&EHFDPHDSRZHUIXODQGUHOLDEOHUHSODFHPHQWIRU
HOHFWULFDOKDUGZDUH6RPHDQDORJXHSURFHVVLQJDELOLW\ZDVGHVLJQHGLQWRWKHHDUOLHU3/&
GHYLFHVEXWWKHUDQJHRIIXQFWLRQVZDVOLPLWHGDQGUHJXODWRU\WHFKQLTXHVZHUHGLIÀFXOW
to implement.
2YHUUHFHQW\HDUVWKH3/&KDVGHYHORSHGUDSLGO\LQSRZHUYHUVDWLOLW\DQGUHOLDELOLW\
7KH FRVW SHU XQLW RI DQDORJXH LQSXWRXWSXW ,2 KDV UHGXFHG VLJQLÀFDQWO\ DQG WKH
UHJXODWRU\ FRQWURO FDSDELOLW\ DQG HDVH RI FRQÀJXUDWLRQ KDV LPSURYHG GUDVWLFDOO\
The operator interface devices these days are mostly standard personal computers
ZLWKDSSURSULDWHVXSHUYLVRU\FRQWURODQGGDWDDFTXLVLWLRQ6&$'$VRIWZDUHLQVWDOOHG
7KHDELOLW\WRLQWHJUDWHWKHVHTXHQWLDODQGUHJXODWRU\FRQWUROORJLFKDVVHHQWKH3/&JDLQ
VLJQLÀFDQWDFFHSWDQFHIRUPDQ\VPDOOWRPHGLXPVL]HGSODQWVQHHGLQJERWKVHTXHQWLDO
and regulatory control.
)RUVPDOOHUDSSOLFDWLRQVVRIWZDUHLVDYDLODEOHWKDWDOORZVFRQWURORIDVLPLODUFDSDELOLW\
WR WKH 3/& WR EH LQFRUSRUDWHG LQWR WKH VWDQGDUG SHUVRQDO FRPSXWHU ZLWK D QHWZRUN
connection to a related I/O device.

Comparison of distributed control system and programmable

logic controller

Distributed control system

7KHIHDWXUHVRID'&6DUHDVIROORZV
x aimed at process plants requiring predominantly modulating control
x lesser capability in high speed sequential logic and equipment control
x KLJKHUFRVWFRPSRQHQWVOHDGWRDSSDUHQWKLJKHUV\VWHPFRVWV
x KLJKHUOHYHOSURJUDPPLQJFRQÀJXUDWLRQODQJXDJH
x superior reporting, trending and historical storage facilities
x integrated control/operator interface functions
x VXSHULRUOHYHOVRIVHFXULW\DQGEXLOWLQUHGXQGDQF\

232 Comminution Handbook | Spectrum Series 21

 CHAPTER 14 t PROCESS CONTROL

Programmable logic controller

7KHIHDWXUHVRID3/&DUHDVIROORZV
x DLPHGDWSURGXFWLRQOLQHSODQWVUHTXLULQJSUHGRPLQDQWO\HTXLSPHQWZLWKVWRSVWDUW
control
x lesser modulating control capability
x ORZHUFRVWFRPSRQHQWU\OHDGVWRORZHUEDVLFV\VWHPFRVWV
x ORZHUOHYHO SURJUDPPLQJ ODQJXDJHV KHQFH SURJUDPPLQJ RI FRPSOH[ UHJXODWRU\
FRQWUROHYHQZKHQSRVVLEOHPD\EHGLIÀFXOWDQGRUWLPHFRQVXPLQJ
x RSHUDWRULQWHUIDFHPRVWO\SURYLGHGE\VHSDUDWHVRIWZDUH6&$'$SDFNDJHV
x YHU\KLJKVSHHGORJLFFDSDELOLW\

MEASUREMENT AND CONTROL COSTS – ‘RULE OF THUMB’

The appropriate extent and sophistication level of a process control system for a process
SODQWLVGHWHUPLQHGDIWHUFRQVLGHULQJDQXPEHURIIDFWRUV7KHVHLQFOXGH
x VRSKLVWLFDWLRQ DQG FRPSOH[LW\ RI WKH SURFHVV IRU LQVWDQFH FRQFUHWH SODQW YHUVXV
QXFOHDUSRZHUSODQW
x SURFHVVKD]DUGVDQGFRQVHTXHQFHVDVDERYH
x SODQWORFDWLRQSUR[LPLW\RIWHFKQRORJ\VXSSRUW
x RSHUDWRUVNLOOOHYHOVDELOLW\WRRSHUDWHDQGPDLQWDLQWHFKQRORJ\
7KH WRWDO FRVW RI WKH LQVWUXPHQWDWLRQ ² FRQWURO DQG HOHFWULFDO GHVLJQ VXSSO\
LQVWDOODWLRQ DQG FRPPLVVLRQLQJ ² ZLOO EH  WR  SHUFHQW RI WRWDO SODQW GHVLJQ DQG
construction costs.

Things to remember
x /RRNDWWKHSURFHVVDVDZKROHZKHQDVVHVVLQJWKHFRQWUROUHTXLUHPHQWV&RQVLGHU
WKDWWKHSODQWPXVWRXWSXWDSURGXFWGHULYHGIURPUDZPDWHULDOSUHVHQWHGDWWKHLQSXW
and recognise the control as an overall system rather than a collection of individual
SURSRUWLRQDOLQWHJUDOGHULYDWLYH3,'ORRSV
x )RFXVRQLGHQWLI\LQJWKHFULWLFDOSURFHVVSDUDPHWHUVWKDWUHTXLUHPHDVXUHPHQWDQG
control.
x 2SHUDWH \RXU SODQW ZLWK WKH EDVLF FRQWUROV LQ SODFH IRU D SHULRG RI WLPH DIWHU
FRPPLVVLRQLQJ,URQRXWEDVLFRSHUDWLQJLVVXHVDQGWKHQZKHQWKHSODQWKDVVHWWOHG
GRZQUHYLVLWWKHFRQWUROV\VWHPFRQÀJXUDWLRQDQGWXQLQJDQGFRQVLGHUWKHQHHGIRU
KLJKHUOHYHOFRQWURO
x The advanced regulatory control or advanced control systems may be included in
WKHXSIURQWGHVLJQ7KLVHQVXUHVWKDWWKHUHJXODWRU\FRQWUROOD\HUZLOOLQWHUIDFHZHOO
ZLWKWKHDGYDQFHGSURFHVVFRQWURO$3&OD\HU<RXUSULPDU\SODQWFRQWUROV\VWHP
'&6 RU 3/& ZLOO FRPH ZLWK WKH IDFLOLWLHV DQG FDSDELOLWLHV WR IXOÀO YLUWXDOO\ DOO RI
\RXU FRQWURO UHTXLUHPHQWV 'R QRW FRPPLW WR DGGLQJ ¶VPDUW· RU ¶H[SHUW· FRQWURO
packages to your plant unless you have exhausted the capability of your installed
FRQWUROV\VWHP<RXKDYHSDLGIRUWKHLQVWDOOHGV\VWHPZLWKZKLFK\RXURSHUDWRUV
DQG PDLQWHQDQFH SHRSOH DUH IDPLOLDU DYRLG LQWURGXFLQJ DQRWKHU V\VWHP ZLWKRXW
exhausting the possibilities of the incumbent system.

Comminution Handbook | Spectrum Series 21 233

 CHAPTER 14 t PROCESS CONTROL

x <RXUSODQWFRQWUROV\VWHPLVWKHVLQJOHPRVWLPSRUWDQWWRRO\RXKDYHWRJHWWKHEHVW
out of your plant. Take the time and trouble to understand it.
x Process measurement techniques do not focus on ultimate accuracy. Good accuracy
is important but repeatability is essential. Repeatable measurements provide the
FRQWUROV\VWHPZLWKWKHLQIRUPDWLRQIRUDVWDEOHRSHUDWLRQ6WDELOLW\LVNH\WR\RXU
operation.
x 'UDZ XSRQ WKH H[SHULHQFH RI WKH RSHUDWRUV 7KH\ DUH RIWHQ D JRRG VRXUFH RI
NQRZOHGJH RI WKH SODQW·V LQWULFDFLHV DQG WKHLU PDQXDO RSHUDWLQJ WHFKQLTXHV FDQ
inform selecting an appropriate control strategy.

GRINDING CIRCUIT CONTROL

7KH ÀUVW VWHS LQ VHWWLQJ XS D FRQWURO V\VWHP IRU DQ\ JULQGLQJ FLUFXLW LV WR GHÀQH LWV
REMHFWLYH )RU RUHV FHPHQWV DQG FRDOV WKH SURGXFW VL]HV DUH W\SLFDOO\ LQ WKH UDQJH
²SHUFHQWSDVVLQJ—PEXWFLUFXLWREMHFWLYHVDUH
x oUHV²WROLEHUDWHWKHYDOXDEOHPLQHUDOVIURPZDVWHVRWKDWWKH\FDQEHFRQFHQWUDWHG
LQWRKLJKTXDOLW\SURGXFWVWKDWDUHVXLWDEOHIRUVPHOWLQJRUOHDFKLQJSURFHVVHV
x FHPHntV²WRJULQGFHPHQWFOLQNHUWRJHWKHUZLWKWKHDGGLWLYHVWRSDUWLFOHVWKDWDUHÀQH
enough to form cements that meet statutory requirements
x tKHUPDOFoDOV²WRLQFUHDVHWKHVXUIDFHDUHDDQGUHDFWLYLW\RIWKHFRDOVWRDOORZIRUPRUH
HIÀFLHQWDQGHIIHFWLYHFRPEXVWLRQ
)LJXUHVKRZVDQLQVWUXPHQWHGVHPLDXWRJHQRXVJULQGLQJ6$*PLOO²EDOOPLOO
circuit.
&KDQJHVLQPDWHULDOVHJKDUGQHVVRUIHHGVL]HRULQRSHUDWLQJFRQGLWLRQVHJIHHGUDWH
FDXVHYDULDWLRQVLQSURGXFWVL]HVWKDWOHDGWRLQHIÀFLHQFLHVLQGRZQVWUHDPSURFHVVHVDQG
corrective actions are required to cancel these disturbances. The task of the metallurgist
LVWRGHÀQHWKHUHJLRQLQZKLFKWKHSODQWVKRXOGRSHUDWHDQGWKHWDVNRIWKHRSHUDWRULV
WRPDLQWDLQWKHSURFHVVZLWKLQWKLVUHJLRQ
6RPH FLUFXLWV DUH FRQWUROOHG E\ RSHUDWRUV ZKR FKDQJH YDULDEOHV PDQXDOO\ EXW LQ
WKH ODUJH DQG FRPSOH[ FLUFXLWV DXWRPDWLF FRQWURO LV D QHFHVVLW\ IRU VDIH DQG HIÀFLHQW
RSHUDWLRQ$XWRPDWLFFRQWUROKDVWZRLPSRUWDQWIXQFWLRQVLQWKDWLW
 DOORZVWKHRSHUDWRUWR¶GLDOXS·WKHUHJLRQZKHUHWKHSURFHVVVKRXOGRSHUDWH
2. NHHSVWKHSURFHVVZLWKLQWKLVRSHUDWLQJUHJLRQGHVSLWHH[WHUQDOGLVWXUEDQFHVWKDWFDQ
VKLIWWKHSURFHVVDZD\IURPLW
7KHVHWDVNVVKRXOGEHDFKLHYHGZLWKWKHKLJKHVWSULRULW\RQKXPDQVDIHW\IROORZHG
by equipment protection. Automatic control systems should therefore be considered
DVWRROVWRDVVLVW²UDWKHUWKDQUHSODFH²WKHRSHUDWRU7KHQH[WVHFWLRQSURYLGHVVRPH
examples of common automatic control loops used in grinding and the disturbances
that they can counteract.

Regulatory control loops – ore feed rate

A common example of a control loop used to regulate the feed rate of dry ore to a mill
LQ WKLV FDVH D 6$* PLOO FDQ EH VHHQ LQ )LJXUH  7KH RUH LV GHOLYHUHG WR WKH IHHG
FKXWHRIWKHPLOOYLDDÀ[HGVSHHGFRQYH\RU7KHRUHLVZHLJKHGRQWKHFRQYH\RUXVLQJD
EHOWZHLJKHUW\SLFDOO\DZHLJKWRPHWHUZKLFKSURGXFHVDWRQQHVSHUKRXUWKVLJQDO
7KLVVLJQDOEHFRPHVWKHSURFHVVYDULDEOHRIDFRQWUROOHUZKHUHLWLVFRPSDUHGWRWKHVHW

234 Comminution Handbook | Spectrum Series 21

 CHAPTER 14 t PROCESS CONTROL

ToFlotation
FromMine

Cyclones
Loader (onlyoneshown)
(Scats) PT
Coarse Loader
Ore Stockpile (COS)
(COS) *

FC SAGMill DT
COSFeed (variablespeed)
FT Trommel MF
Hopper Feed
ApronFeeder Water Screens JT
(variableSpeed) FT
ControlValve

FeedConveyor
WT2 BallMill
FeedChute

JT
WC
WT1
LT LC *
AT
*
/ŶƐƚƌƵŵĞŶƚƐĂŶĚŽŶƚƌŽůůĞƌƐ *
PT Ͳ Cyclonefeedpressure(kPa)
DT Ͳ Cyclonefeeddensity(%solids) FC CycloneFeed
FT Ͳ Volumetricflow(m3/h) CycloneFeed Pump(variableSpeed)
MF Ͳ Solidsflowcalculation(t/h) Hopper FT Hopper
LT Ͳ Hopperlevel(%) Water
LC Ͳ Hopperlevelcontroller ControlValve
FC Ͳ Waterflowcontroller
JT Ͳ SAG/BallMillpower(kW)
WT1 Ͳ SAGload/weight(tonnes)
AT Ͳ SAGnoiselevel(dB)
WT2 Ͳ SAGfeedrate(t/h)
WC Ͳ SAGfeedratecontroller
* Ͳ Operatorsetpoint

FIG 14.2 – Instruments and controllers installed on a circuit containing a semi-autogenous grinding (SAG) mill and a ball mill.

point that is the desired feed rate. If the feed rate is above the set point the feed rate
FRQWUROOHUUHGXFHVWKHVSHHGRIWKHDSURQIHHGHUDQGLIWKHIHHGUDWHLVEHORZWKHVHW
point the apron feeder speed is increased. The signal from the controller sent to the
DSURQ IHHG UDWH FRQWUROOHU LV FDOOHG WKH PDQLSXODWHG YDULDEOH ,QIRUPDWLRQ IURP ÀHOG
LQVWUXPHQWV  DQG WR DFWXDWRUV FDQ EH WUDQVPLWWHG E\ ² P$ VLJQDOV RU GLJLWDOO\ YLD
VSHFLDOLVHGGLJLWDOQHWZRUNV
7KLV FRQWURO ORRS DOVR FRXQWHUDFWV GLVWXUEDQFHV WR WKH PLOO IHHG UDWH ,I WKHUH ZHUH
QRGLVWXUEDQFHVWKHFRQWUROOHUZRXOGVLPSO\DGMXVWWKHDSURQIHHGHUVSHHGXQWLOWKH
IHHG UDWH ZDV HTXDO WR WKH VHW SRLQW DQG QR IXUWKHU DGMXVWPHQWV ZRXOG EH UHTXLUHG
KRZHYHUWKHUHDUHGLVWXUEDQFHVWKDWDIIHFWWKHFRQWURORIHYHQWKLVUHODWLYHO\VLPSOH
SURFHVV7KH\DUH
x VKRUWWHUP ² VHFonGV to PLnXtHV ² ZKLFK LQFOXGH YDULDWLRQV LQ EXON GHQVLW\ DQG
particle size
x longer term ²KoXUVtoZHHNV²ZKLFKLQFOXGHZHDUDQGRUHW\SHFKDQJHV
The feed rate controller is required to counteract all these disturbances and keep the
feed rate as close as possible to the set point.

Regulatory control loops – water feed rate

)LJXUHDOVRVKRZVDQRWKHUFRQWUROORRSXVHGLQJULQGLQJFLUFXLWV7KLVORRSDGMXVWV
WKH ÁRZ UDWH RI ZDWHU WR WKH 6$* PLOO IHHG FKXWH 7KH ZDWHU ÁRZ UDWH FRQWUROOHU LV

Comminution Handbook | Spectrum Series 21 235

 CHAPTER 14 t PROCESS CONTROL

DQRWKHUH[DPSOHRIDEDVLFUHJXODWRU\FRQWUROOHU,IWKHZDWHUÁRZUDWHLVDERYHWKHVHW
SRLQWWKHFRQWUROOHUFORVHVWKHFRQWUROYDOYHDQGLILWLVEHORZWKHVHWSRLQWWKHFRQWUROOHU
opens the control valve.
:KLOHWKHZDWHUÁRZUDWHFRQWUROORRSLVVLPSOHLWLVVWLOOVXEMHFWWRGLVWXUEDQFHVWKH
PRVWFRPPRQRIZKLFKLVFDXVHGE\FKDQJLQJOLQHSUHVVXUHGXHWRRWKHUSODQWVHFWLRQV
UHTXLULQJPRUHRUOHVVZDWHU7KHRUHIHHGUDWHDQGZDWHUÁRZUDWHFRQWUROORRSVPXVW
EH OLQNHG WRJHWKHU WR HQVXUH WKDW WKH GHQVLW\ SHU FHQW VROLGV E\ ZHLJKW HQWHULQJ WKH
6$*PLOOUHPDLQVFRQVWDQW6LQFHGHQVLW\RIWKHPLOOIHHGFDQQRWEHGLUHFWO\PHDVXUHGLW
PXVWEHLQIHUUHGIURPH[LVWLQJPHDVXUHPHQWV8VLQJWKLVDSSURDFKWKHDPRXQWRIZDWHU
UHTXLUHGWRDFKLHYHDWDUJHWIHHGGHQVLW\FDQEHFDOFXODWHGXVLQJWKHIROORZLQJHTXDWLRQV

Wd = Wt²)m 

'f = Wd:d)Z 

,IWKHUHTXLUHGIHHGGHQVLW\WDUJHWLV'tWKHDPRXQWRIIHHGZDWHU)Z to achieve this

WDUJHWFDQEHFDOFXODWHGE\VXEVWLWXWLQJ'tLQWR'fLQ(TXDWLRQDQGVROYLQJIRU)Z

)Z :d't²:d 

ZKHUH
)Z LVWKHUHTXLUHGIHHGZDWHUÁRZUDWHPKWRDFKLHYHDWDUJHWGHQVLW\ 't
Wd LVWKHIHHGUDWHRIGU\VROLGVJRLQJWRWKH6$*PLOOWK
Wt LVWKHWRWDOIHHGUDWHRIVROLGVSOXVZDWHUPHDVXUHGE\WKHZHLJKWRPHWHUWK
)m LVWKHDVVXPHGIHHGPRLVWXUHSHUFHQWZDWHUE\ZHLJKW
' f LVWKHFDOFXODWHGIHHGGHQVLW\SHUFHQWVROLGVE\ZHLJKW
't LVWKHWDUJHWRUUHTXLUHGIHHGGHQVLW\SHUFHQWVROLGVE\ZHLJKW
7R LPSOHPHQW (TXDWLRQ  LQ D FRQWURO V\VWHP D FDOFXODWLRQ EORFN LV XVHG 7KLV
EORFNXVHVWKHPHDVXUHG6$*IHHGUDWHWKHDVVXPHGIHHGPRLVWXUHDQGWKHWDUJHWIHHG
GHQVLW\WRFDOFXODWHWKHUHTXLUHGIHHGZDWHUÁRZUDWH1RUPDOO\FDOFXODWLRQEORFNVDUH
SURJUDPPHGXVLQJDVLPSOLÀHGFRPSXWHUODQJXDJHZLWKD%$6,&OLNHV\QWD[7KHOLQHV
of computer code in these calculation blocks are often limited in number and complexity,
ZKLFK LV GRQH GHOLEHUDWHO\ WR DYRLG RYHUORDGLQJ WKH V\VWHP &DOFXODWLRQ EORFNV DUH
LQWHQGHGIRUVPDOOVLPSOHFDOFXODWLRQVVXFKDV(TXDWLRQ
)LJXUH  DOVR VKRZV WKH FRQWURO VWUDWHJ\ IRU WKH 6$* PLOO WKDW FRPSULVHV WKH
IROORZLQJFRQWUROORRSVDQGFDOFXODWLRQV
x 6$*PLOOIHHGUDWHFRQWUROORRS
x 6$*PLOOIHHGZDWHUFRQWUROORRS
x IHHGZDWHUFDOFXODWLRQEORFNZKLFKOLQNVWKHPHDVXUHGIHHGUDWHWRWKHIHHGZDWHU
ÁRZUDWHVHWSRLQW
This simple example indicates the type of regulatory control used in comminution
FLUFXLWV ,Q D W\SLFDO ZHW JULQGLQJ SURFHVV WKHUH DUH PDQ\ VXFK FRQWURO VWUDWHJLHV
SHUIRUPLQJERWKVLPSOHDQGPRUHFRPSOH[FRQWUROIXQFWLRQVZKLFKFDQEHFRQWUROORRSV
and calculations.

236 Comminution Handbook | Spectrum Series 21

 CHAPTER 14 t PROCESS CONTROL

Logic-based control
$QRWKHU LPSRUWDQW FRQWURO V\VWHP IXQFWLRQ LV ORJLFEDVHG FRQWURO VXFK DV VHTXHQFHV
DQG LQWHUORFNV ZKLFK LV XVHG WR VWRSVWDUW SODQWV IRU H[DPSOH D JULQGLQJ FLUFXLW
VWRSVWDUWLQGLYLGXDOXQLWVRIHTXLSPHQWDSXPSRUSURWHFWHTXLSPHQWIURPGDPDJH
VWRSDSXPSLIKRW/RJLFFRQWUROLVLPSOHPHQWHGLQWKHSODQWFRQWUROV\VWHPWRJHWKHU
ZLWK²EXWDWDORZHUOHYHOWKDQ²FRQWUROORRSVDQGFDOFXODWLRQV

RESPONSE OF THE SEMI-AUTOGENOUS GRINDING MILL

TO CHANGE IN FEED RATE
7KH UHDFWLRQ RI 6$* PLOO ORDG DQG SRZHU WR LQFUHDVLQJ IHHG UDWH LV DQ H[DPSOH RI D
complex response to change in a controllable variable.
$VD6$*PLOOIHHGUDWHLVLQFUHDVHGERWKWKHORDGDQGSRZHULQFUHDVHKRZHYHURQFH
an optimum feed rate has been achieved, continued increase in the feed rate leads to
LQFUHDVLQJORDGEXWGHFUHDVLQJSRZHUEHFDXVHWKHFHQWUHRIJUDYLW\RIWKHFKDUJHPRYHV
WRZDUGVWKHFHQWUHRIWKHPLOO)LJXUH,IWKHIHHGUDWHFRQWLQXHVWRLQFUHDVHZKLOH
SRZHULVORZRUIDOOLQJDFODVVLFDOPLOORYHUORDGWKH6$*PLOOFDQORDGWRWKHSRLQW
ZKHUHWKHIHHGLVIRUFHGRXWRIWKHPLOODQGWKHSURFHVVKDVWREHVWRSSHG7KHUHODWLRQVKLS
EHWZHHQ IHHG UDWH DQG SRZHU ZDV WKH EDVLV RI WKH :LOOLDPVRQ FRQWUROOHU GHYHORSHG
LQ 6RXWK $IULFD LQ  WR FRQWURO WKH IHHG UDWH WR DXWRJHQRXV JULQGLQJ $* PLOOV
,WZRUNHGLQIRXUPLQXWHF\FOHVDVIROORZV
 IRUWZRPLQXWHVRUHLVIHGWRWKHPLOO
2. PLOOSRZHULVUHDGDQGFRPSDUHGZLWKWKHSRZHUDWWKHVDPHSRLQWLQWKHSUHYLRXV
cycle
 LISRZHUKDVGHFUHDVHGWKHPLOOLVRSHUDWLQJDWH[FHVVLYHORDGDQGIHHGLVWXUQHGRII
 LISRZHUKDVLQFUHDVHGRUUHPDLQHGVWDWLRQDU\WKHPLOOKDVVSDUHFDSDFLW\DQGIHHG
is left on.
&XUUHQWEHVWRSHUDWLQJSUDFWLFHLQYROYHVWKHPLOOIHHGUDWHEHLQJJUDGXDOO\UHGXFHGDV
WKHSRZHUSHDNLVDSSURDFKHGDQGWKHQVWDELOLVHV7KHDGGLWLRQDOFRPSOH[LWLHVRI6$*
control require the control logic to consider many measurements in making a decision
LQYROYLQJVHYHUDOPDQLSXODWHGYDULDEOHVLQFOXGLQJ
x load or bearing pressure
x PLOOSRZHUGUDZ

FIG 14.3 – The relationships between feed rate, power and load in a semi-autogenous
grinding mill (image courtesy Jeff McKay; adapted from Williamson, 1960).

Comminution Handbook | Spectrum Series 21 237

 CHAPTER 14 t PROCESS CONTROL

x mill motor amps

x mill motor temperature
x recycle tonnage
x VFUHHQORDGLQJRUDPSHUDJH
x mill sound
x GRZQVWUHDP OLPLWDWLRQ VXFK DV EDOO PLOO FLUFXLW FRQVWUDLQWV SURGXFW SDUWLFOH VL]H
ÁRWDWLRQFLUFXLWFRQVWUDLQWVDQGVRRQ
%DVHGRQFRPELQDWLRQVRIWKHVHYDULDEOHVWKHORJLFWKHQPDNHVGHFLVLRQVWRFRQWUROWKH
VHWSRLQWVRI
x 6$*IHHGUDWH
x 6$*PLOOVSHHG
x 6$*IHHGGHQVLW\RUZDWHUDGGLWLRQ
x 6$*IHHGHUUDWLRVRUZKLFKIHHGHUVLQFUHDVHRUGHFUHDVHXQGHUWKHFXUUHQWFRQGLWLRQV
6$*UHJXODWRU\FRQWUROWKHQ
x takes the mill feed rate set point and controls feeder speeds and conveyor speed to
bring the actual feed rate indication to the desired set point
x takes the desired mill speed set point and adjusts the output to the mill motor to
control the mill speed rev/min at the desired set point
x DGMXVWVWKHZDWHUÁRZUDWHVHWSRLQWWRSURYLGHWKHSURSHUUDWLRRIZDWHUWRIHHGWR
DFKLHYHWKHGHVLUHG6$*PLOOGHQVLW\LWDOVRFRQWUROVWKHRXWSXWWRWKHZDWHUYDOYHWR
SURYLGHWKHZDWHUÁRZUDWHGHVLUHGE\WKHGHQVLW\FRQWUROOHU
7KHUHODWLRQVKLSEHWZHHQIHHGUDWHDQGPLOOORDGDQGSRZHUPRYHVDFFRUGLQJWRRUH
hardness and size of the rock feeding the mill, so an effective advanced controller must
automatically adapt to the changing conditions.

CONTROL OF A CRUSHER
The feed rate to the primary grinding circuit is the critical variable in a concentrator
since it determines the economics of the entire mineral processing enterprise.
3ULPDU\ JULQGLQJ LV SUHFHGHG E\ FUXVKLQJ VR FUXVKHU FRQWURO LV UHIHUUHG WR EULHÁ\
7KHUHJXODWRU\FRQWURORIDFUXVKHULQYROYHV
x maintaining feed rate at the desired set point
x starting and stopping the crusher
x protecting the crusher from metal or other damage
x choke feeding the crusher.
$FRQHFUXVKHUPD\EHOLPLWHGE\WKHYROXPHRIIHHGFRPLQJWRLWERZOOHYHOWKH
KDUGQHVVRIWKHRUHSRZHUGUDZRUWKHSK\VLFDOOLPLWDWLRQVRIWKHFUXVKHUK\GUDXOLF
SUHVVXUH RU PRWRU DPSV &UXVKHU ORJLF PXVW WKHQ GHWHUPLQH ZKHQ WR LQFUHDVH RU
GHFUHDVHIHHGUDWHEDVHGRQ
x ERZOOHYHO
x SRZHUGUDZ
x K\GUDXOLFSUHVVXUHLIWKHFUXVKHULVDK\GURVHWFUXVKHUDQGWLJKWHQRUORRVHQWKHJDS
RIWKHFUXVKHUGHSHQGLQJXSRQWKHGHVLUHGSURGXFWVL]HDQGKRZKDUGWKHFUXVKHU
LVZRUNLQJ

238 Comminution Handbook | Spectrum Series 21

 CHAPTER 14 t PROCESS CONTROL

$GMXVWPHQWRIWKHJDSVHWWLQJZRXOGWKHUHIRUHLQYROYHORRNLQJDWFUXVKHUDPSVDQG
the rock size distribution generated by the crusher.
7KXVWKHUHLVDJDLQDQRQOLQHDUFRQWUROZLWKPXOWLSOHLQSXWVDQGPXOWLSOHRXWSXWV

ROLE OF THE METALLURGIST

Defining process control objectives

$QLPSRUWDQWWDVNRIWKHSURFHVVFRQWUROPHWDOOXUJLVWLVWRGHÀQHWKHREMHFWLYHVRIWKH
FRQWUROVWUDWHJ\-XVWOLNHDQRSHUDWRUEDVHGFRQWURODXWRPDWLFFRQWUROVWUDWHJLHVPXVW
EHGHVLJQHGWRIROORZDFOHDUVHWRIREMHFWLYHV%HIRUHWKHGHVLJQDQGLPSOHPHQWDWLRQ
of any process control strategy these objectives must be developed. The development
RIFRQWUROREMHFWLYHVLVSRVVLEO\WKHPRVWLPSRUWDQWVWHSWRZDUGVWKHGHYHORSPHQWRID
VXFFHVVIXOFRQWUROVWUDWHJ\&RQVLGHUWKHVLPSOHH[DPSOHRIFRQWUROOLQJWKHOHYHOVLQD
KRSSHURUVXPSWKDWIHHGVDQHVWRIK\GURF\FORQHVLQFORVHGFLUFXLWZLWKDEDOOPLOO
x 7KH ÀUVW VWHS LV WR GHÀQH WKH SURFHVV YDULDEOHV DQG PDQLSXODWHG YDULDEOHV
The manipulated variables are adjusted by the control system to achieve some
REMHFWLYH ZLWK WKH SURFHVV YDULDEOHV )RU WKH KRSSHUOHYHO FRQWURO VWUDWHJ\ WKHUH LV
RQHPDQLSXODWHGYDULDEOHSXPSVSHHGRQHSURFHVVYDULDEOHKRSSHUOHYHODQGWKH
UHTXLUHGREMHFWLYHWRPDLQWDLQWKHKRSSHUOHYHODWDVHWSRLQWXVXDOO\SHUFHQW
x ,QLWLDOWKLQNLQJVXJJHVWVWKDWWKLVREMHFWLYHFDQEHDFKLHYHGE\DFRQYHQWLRQDO3,'
FRQWUROOHU WXQHG WR PDLQWDLQ WKH KRSSHU OHYHO FORVH WR WKH VHW SRLQW  SHU FHQW
:KLOHWKLVDSSURDFKWRWXQLQJZLOOUHVXOWLQWKHOHYHOUHWXUQLQJTXLFNO\WRVHWSRLQW
IROORZLQJ D GLVWXUEDQFH LW ZLOO UHVXOW LQ ODUJH FKDQJHV WR SXPS VSHHG ZKLFK ZLOO
FDXVHODUJHFKDQJHVLQF\FORQHIHHGSUHVVXUHDQGF\FORQHFXWSRLQW²QRWDJRRGUHVXOW
for a grinding circuit that is required to produce a stable product size distribution.
x The actual control objective is completely different. The hopper should be considered
DVDVXUJHYHVVHOORFDWHGEHWZHHQWZRSURFHVVLQJXQLWVEDOOPLOODQGWKHF\FORQHV
,WVIXQFWLRQLVWRDEVRUEÁRZUDWHÁXFWXDWLRQVIURPWKHEDOOPLOODQGQRWSDVVWKHP
RQ WR WKH F\FORQHV 7KHUHIRUH WKH REMHFWLYH RI WKH KRSSHUOHYHO FRQWUROOHU LV QRW WR
KROGWKHOHYHOWRVHWSRLQWEXWWRFKDQJHWKHSXPSVSHHGDVOLWWOHDVSRVVLEOHZKLOH
PDLQWDLQLQJWKHKRSSHUOHYHOEHWZHHQVDIHDQGDFFHSWDEOHOLPLWV7KLVW\SHRIOHYHO
control is commonly termed surge tank level control or averaging level control.
7XQLQJDQGVHWXSRI3,'FRQWUROOHUVWRDFKLHYHWKLVW\SHRIUHVSRQVHLVGHVFULEHG
LQPDQ\UHIHUHQFHVVXFKDVWKH0HWVR([SHUWXQHZHEVLWHKWWSZZZH[SHUWXQH
FRPWXWRUKWPO
x 7KHRSHUDWLQJREMHFWLYHVIRUWKHSXPSFDQQRZEHUHYLVHGDORQJWKHIROORZLQJOLQHV
x PDLQWDLQWKHKRSSHUOHYHOSURFHVVYDULDEOHEHWZHHQSHUFHQWDQGSHUFHQW
E\DGMXVWLQJWKHSXPSVSHHGPDQLSXODWHGYDULDEOH
x WXQH WKHOHYHOFRQWUROOHUWRPLQLPLVHSUHVVXUHÁXFWXDWLRQVLQF\FORQHSUHVVXUH
ZKLOHNHHSLQJWKHKRSSHUOHYHOEHWZHHQSHUFHQWDQGSHUFHQW
,W FDQQRW EH HPSKDVLVHG HQRXJK WKDW JRRG SURFHVV FRQWURO ZLOO RQO\ UHVXOW IURP
ZHOOGHÀQHGFRQWUROREMHFWLYHVDQGUHOLDEOHLQVWUXPHQWPHDVXUHPHQWV

Comminution Handbook | Spectrum Series 21 239

 CHAPTER 14 t PROCESS CONTROL

Practical skills required

To implement and maintain successful control strategies in a large modern comminution
FLUFXLWUHTXLUHVNQRZOHGJHRIWKHFLUFXLWDQGWKHVNLOOVHWWRXVHWKLVNQRZOHGJHWKDWRQO\
a team can provide. The metallurgist is part of this team but cannot be expected to
possess all the skills required. Often these are not formal teams but individuals brought
together to implement a particular process control project. This is a small company
DSSURDFK ZKHUH WKHUH LV RIWHQ D VKRUWDJH RI VNLOOHG SHRSOH 6RPH PHPEHUV RI WKHVH
ephemeral teams are contractors or consultants engaged from outside the company.
/DUJHFRUSRUDWLRQVRIWHQKDYHWKHEHQHÀWRISHUPDQHQWWHDPVRUHYHQZKROHSURFHVV
FRQWUROGHSDUWPHQWVWKLVSUDFWLFHLVFRPPRQLQWKHFKHPLFDORLODQGSRZHULQGXVWULHV
but less so in mining.
6RPHRIWKHDUHDVUHTXLULQJVXFKVNLOOVLQFOXGH
x 3UoFHVVNnoZOHGJH²LW·VLPSRVVLEOHWRFRQWURODSURFHVVZLWKRXWLWEHLQJXQGHUVWRRG
7KH UROH RI WKH SURFHVV FRQWURO PHWDOOXUJLVW LV WR SURYLGH WKLV SURFHVV NQRZOHGJH
It may be necessary to have a process specialist on the team, particularly for large
projects or if complex processes are involved.
x &ontUoO oEMHFtLYHV ² IRU WKLV WDVN WKH SURFHVV VSHFLDOLVW ZLOO XVH LQIRUPDWLRQ IURP
equipment suppliers, original design information, data and information from
subsequent process upgrades, production requirements, plant maintenance
specialists and, of course, process operators. This list provides an indication of the
UDQJHRILQIRUPDWLRQUHTXLUHGWRGHYHORSDZRUWKZKLOHVHWRIFRQWUROREMHFWLYHV
x )LHOG LnVtUXPHntDtLon DnG DFtXDtoUV ² GLVFXVVHG LQ WKH VHFWLRQ RQ FRQWURO HOHPHQWV
reference to industrial instruments and actuators can be found in many volumes;
IRUH[DPSOH'XQQ$QDOOWRRFRPPRQPLVWDNHLVWRDWWHPSWWRFLUFXPYHQW
a malfunctioning instrument or actuator by retuning the control loop rather than
UHSDLULQJ RU UHSODFLQJ WKH YDOYH DVVHPEO\ 7KLV DSSURDFK LV XQVXFFHVVIXO IRU WZR
reasons
 $ZDWHUÁRZFRQWUROOHUZLWKDIDXOW\YDOYHFDQQRWEHVXFFHVVIXOO\WXQHG
2. ,IWKHIDXOW\YDOYHLVÀ[HGWKHDGKoFWXQLQJFRQVWDQWVZLOOEHXQOLNHO\WRSURYLGH
RSWLPDO FRQWURO 7KH SUDFWLFH RI PRGLI\LQJ FRQWURO VWUDWHJLHV RU UHWXQLQJ
FRQWUROOHUVLQVRPHZD\WR¶JHWDURXQG·SRRULQVWUXPHQWDWLRQVKRXOGEHDYRLGHG
7KH UROH RI WKH PHWDOOXUJLVW LQ WKLV DUHD RI HYHULQFUHDVLQJ FRPSOH[LW\ LV QRW WR EH
DQ H[SHUW EXW WR EH DZDUH RI WKH PHDVXUHPHQW SULQFLSOHV DQG WHFKQLTXHV DQG WKHLU
limitations.

PROCESS CONTROL TECHNIQUES AND SKILLS

There is a range of standard and advanced process control techniques that are required
to implement and maintain control strategies in comminution circuits. The process
FRQWURONQRZOHGJHDUHDVLQFOXGH
x /ooS tXnLnJ ² WKH DELOLW\ WR WXQH D UDQJH RI FRQWURO ORRSV LV D IRXQGDWLRQ VNLOO
&RPPLQXWLRQFLUFXLWVOLNHDOOSURFHVVHVKDYHVRPHXQLTXHFRQWUROORRSVWKDWPXVWEH
WXQHGDSSURSULDWHO\$FRPPRQPLVWDNHLVWRLQFRUUHFWO\WXQHKRSSHUOHYHOFRQWUROOHUV
0F0LOODQOLVWVFRPPRQFRQWUROORRSVDQGKRZWKHVHVKRXOGEHWXQHG

240 Comminution Handbook | Spectrum Series 21

 CHAPTER 14 t PROCESS CONTROL

x 5HJXODtoU\FontUoO²DVRXQGNQRZOHGJHRIERWKEDVLFDQGDGYDQFHGUHJXODWRU\FRQWURO
WHFKQLTXHVLVUHTXLUHGWRGHVLJQDQGLPSOHPHQWFRQWUROVWUDWHJLHV6SLW]HUOLVWV
regulatory control techniques and explains them in more detail.
x &DSDELOLt\ oI tKH SODnt FontUoO V\VtHP 3&6 ² WKH 3&6 LV WKH WRRO XVHG WR LPSOHPHQW
control strategies. These systems provide more than just regulatory control; they also
SURYLGHDJUDSKLFDOLQWHUIDFHIRUWKHRSHUDWRUDQGWKHDELOLW\WRLPSOHPHQWORJLFEDVHG
FRQWURO([SHULHQFHGSURFHVVFRQWUROPHWDOOXUJLVWVPXVWKDYHDWKRURXJKNQRZOHGJH
RIWKHFDSDELOLW\RIWKHLU3&6SDUWLFXODUO\WKHUHJXODWRU\FRQWUROFDSDELOLW\

ADVANCED PROCESS CONTROL SYSTEMS

$3& V\VWHPV LQ RQH IRUP RU DQRWKHU DUH SUHVHQW LQ YLUWXDOO\ DOO PLQHUDO SURFHVVLQJ
SODQWVRIVLJQLÀFDQWVL]H3URFHVVFRQWUROLQPLQHUDOSURFHVVLQJFDQEHEURNHQGRZQLQWR
WZRJHQHUDOFDWHJRULHVUHJXODWRU\RUVWDELOLVLQJFRQWURO and supervisory or advanced
control VHHWKHFRQWUROKLHUDUFK\)LJXUH
The purpose of regulatory or stabilising control is to hold a process at given set points
VXFKDVIHHGUDWHWKWRWKHSODQWÁRWDWLRQFHOOOHYHOVDQGDLURUZDWHUÁRZUDWHV7KLV
in itself is a challenging endeavour.
6XSHUYLVRU\ RU DGYDQFHG SURFHVV FRQWURO LV JHQHUDOO\ FRQVWUXFWHG RQ WRS RI WKH
regulatory control so that it feeds set points or outputs to the plant regulatory control
V\VWHP8QOLNHUHJXODWRU\FRQWUROZKRVHMRELWLVWRPDLQWDLQWKHSURFHVVDWDJLYHQVHW
point and stabilise the process, the objective of advanced process control is to optimise
and push the process to squeeze out the most value. This is done by creating logic that
IRFXVHVRQ
x increased production
x improved and controlled product grade
x reduced consumables such as reagents, energy or grinding media.
Advanced process control systems generally include some degree of expert control
V\VWHPVUXOHEDVHGFRQWUROHVWDEOLVKHGRQGHWDLOHGNQRZOHGJHDERXWWKHEHKDYLRXURI
WKHV\VWHPE\DSURFHVVH[SHUWEXWPD\DOVRHPSOR\DGDSWLYHFRQWUROEDVHGRQSURFHVV
PRGHOV SK\VLFDO RU QHXUDO QHWZRUN WKDW HVWLPDWH XQPHDVXUHG RU XQUHOLDEOH SURFHVV
SDUDPHWHUVDQGRSWLPLVHUVZKLFKHVWLPDWHRSWLPDOSURFHVVVHWSRLQWV
:KLOH UHJXODWRU\FRQWUROJHQHUDOO\GHDOVZLWK VLPSOHUDQGPRUHSUHGLFWDEOH FRQWURO
UHDFWLRQV DGYDQFHG SURFHVV FRQWURO LV W\SLFDOO\ FRPSOH[ GHDOLQJ ZLWK PXOWLSOH
observations or inputs and multiple reactions or outputs.

Rule-based expert systems (fuzzy and crisp)

,QPLQHUDOVSURFHVVLQJWKHPRVWZLGHO\XWLOLVHGDGYDQFHGFRQWUROVDUHH[SHUWV\VWHPV
([SHUWV\VWHPVDUHGHVLJQHGWRHPXODWHZKDWDQH[SHUWRSHUDWRUZRXOGGR7KHSULPDU\
PRWLYDWLRQ IRU LPSOHPHQWLQJ H[SHUW FRQWURO LV WR LQFUHDVH SODQW SURÀWDELOLW\ WKURXJK

1. The most widely used expert systems in mineral processing include Gensym’s G2© software, KnowledgeScape’s
KSX© product and Metso’s OCS© and OCS-4D© software tools. All of these have been utilised for many years,
decades even, and have solid track records at plants around the world. Most DCS vendors also have some APC
tools embedded in or offered with their control systems, but these may be limited in scope or accessibility to
process personnel.

Comminution Handbook | Spectrum Series 21 241

 CHAPTER 14 t PROCESS CONTROL

LQFUHDVHG SURGXFWLRQ RI VDOHDEOH SURGXFW ZLWK UHGXFHG FRQVXPSWLRQ RI SODQW
FRQVXPDEOHVVXFKDVUHDJHQWVDQGEDOOFKDUJHDQGRULQFUHDVHGSURGXFWTXDOLW\
([SHUWV\VWHPVDUHUXOHEDVHGDSSOLFDWLRQVHPSOR\LQJ¶LI«WKHQ«·ORJLFWRSHUIRUP
anything from simple to complex actions. While virtually all expert systems utilise an
¶LI«WKHQ«·VWUXFWXUHWKHH[DFWV\QWD[IRUFRQVWUXFWLQJFRQWUROORJLFGHSHQGVXSRQ
WKHVSHFLÀFWRROHPSOR\HG([SHUWV\VWHPVW\SLFDOO\IXQFWLRQRQDPDFURVHFRQGWLPH
VFDOHPRUHWKDQRQHVHFRQGVFDQWLPHVDQGDUHWKXVEHVWVXLWHGIRUDFWLRQVW\SLFDOO\
WDNHQE\RSHUDWRUVUDWKHUWKDQUHSODFLQJ3,'RUUHJXODWRU\FRQWUROORRSV%HFDXVHRIWKH
UHODWLYHO\VLPSOHSURJUDPPLQJODQJXDJHLI«WKHQ«H[SHUWV\VWHPVFDQEHTXLFNO\
learnt and utilised by process engineers and process control personnel.
([SHUW V\VWHPV SURYLGH D WRRO WKDW QRW RQO\ DLGV LQ SURJUDPPLQJ FRPSOH[ PXOWL
LQSXWPXOWLRXWSXWVWUDWHJLHVEXWDOVRSURPRWHVWKHORQJHYLW\DQGVXUYLYDELOLW\RIWKH
system after its development. This is due to plant personnel being able to understand
DQG PRGLI\ WKH H[LVWLQJ ORJLF ZLWK PLQLPDO WUDLQLQJ $QRWKHU VWURQJ DGYDQWDJH RI
expert systems is their inherent ability to explain their actions. As they use a humanistic
DSSURDFKWRGHFLVLRQPDNLQJLILVWUXHWKHQGRWKHIROORZLQJWKHH[HFXWLRQRIWKHLU
ORJLFVKRZVH[DFWO\KRZWKHDSSOLFDWLRQUHDFKHGDGHFLVLRQ6RPHH[SHUWV\VWHPVHYHQ
VDYHWKLVORJLFVRWKDWWKHXVHUFDQJREDFNLQWLPHDQGVHHMXVWZKDWKDSSHQHGDQGZK\
at any given moment in the recent past. This feature is particularly useful as it is rare for
VRPHRQHWREHPRQLWRULQJWKHFRQWUROORJLFZKHQLQWHUHVWLQJHYHQWVRFFXU

Logic types
7KHW\SHRIORJLFXWLOLVHGE\PRVWH[SHUWV\VWHPVPD\EHFDWHJRULVHGDV¶FULVS·RU¶IX]]\·
ORJLF 7KH PRVW ZLGHO\ XVHG H[SHUW V\VWHPV HPSOR\ D FRPELQDWLRQ RI ERWK FULVS DQG
fuzzy logic.
&ULVSORJLFDVWKHQDPHLPSOLHVLVORJLFZKHUHWKHSUHPLVHLISDUWRIDUXOHLVHLWKHU
 SHU FHQW WUXH RU  SHU FHQW IDOVH  RU  7KLV UHVXOWV LQ GLVWLQFW EODFN RU ZKLWH
DFWLRQV$QH[DPSOHRIFULVSORJLFLV
,I6$*SRZHU!6$*SRZHUKLJKOLPLWDQG6$*SRZHUUDWHRIFKDQJH! 
7KHQ6$*IHHGUDWHVHWSRLQW63FKDQJH WK
,Q IX]]\ ORJLF WKH SUHPLVH RI D UXOH FDQ KDYH D GHJUHH RI WUXWK EHWZHHQ ]HUR DQG
SHUFHQWDQGWKLVGHJUHHRIWUXWKIXOQHVVLVXVHGWRZHLJKWWKHFRQFOXVLRQFRPSDUHG
WR FRQFOXVLRQV IURP RWKHU UXOHV )X]]\ ORJLF DOORZV WKH SURJUDPPHU WR H[SUHVV WKH
control strategy in easily understood and agreed upon terminology and therefore is
PXFKHDVLHUWRXQGHUVWDQGZKHQUHDGLQJWKHFRQWUROORJLF)RUH[DPSOHDIX]]\UXOH
PLJKWVWDWHWKHIROORZLQJ
,I6$*SRZHULVKLJKDQG6$*SRZHUUDWHRIFKDQJHLVQRWGHFUHDVLQJ
7KHQWKHIHHGUDWH63FKDQJHLVDODUJHGHFUHDVH
7KH GLIIHUHQFH EHWZHHQ WKH WZR UXOHV LV WKDW WKH FKDQJH DSSOLHG LQ WKH VHFRQG UXOH
PD\YDU\IURPDVPDOODPRXQWWRDODUJHDPRXQWGHSHQGLQJRQWKHGHJUHHWRZKLFKWKH
SUHPLVHLVWUXH6$*SRZHULVKLJKDQGQRWGHFUHDVLQJ,QWKHÀUVWUXOHDWKIHHG
UDWHFXWZLOORFFXUQRPDWWHULIWKHSRZHUDQGLWVUDWHRIFKDQJHLVEDUHO\RYHUWKHOLPLWV
KLJKOLPLWDQGUDWHRIFKDQJHRI]HURRUIDULQH[FHVVRIWKHOLPLWV

242 Comminution Handbook | Spectrum Series 21

 CHAPTER 14 t PROCESS CONTROL

7KHÁRZRIORJLFLQDQH[SHUWV\VWHPLVW\SLFDOO\DUHVXOWRIHLWKHUFRQFOXVLRQVIURP
RQHUXOHEHLQJXVHGLQWKHSUHPLVHRIDQRWKHUUXOHRULQJUDSKLFDOFRQQHFWLRQVEHWZHHQ
blocks of logic.

Model-based control
Process models can be used as soft sensors to provide the necessary estimation for a
FRQWUROOHGYDULDEOHWKDWLVQRWDFWXDOO\PHDVXUHGLQWKHÀHOGVHHVHFWLRQRQ¶6RIWVHQVRU
LQ2&6·LQ&KDSWHU0RGHOVFDQDOVREHLQYROYHGLQSURFHVVFRQWUROLQDPXFKVLPSOHU
RUPRUHFRPSOH[ZD\
On the simpler side, regulatory control and advanced regulatory control make use
RIPRGHOVIRUDQXPEHURIWHFKQLTXHV,QIHHGIRUZDUGFRQWUROIRULQVWDQFHDQLQYHUVH
DQGRIWHQVLPSOLÀHGPRGHORIWKHSURFHVVFDQEHXVHGWRFDOFXODWHVRPHSDUWRIWKH
FRQWURODFWLRQOHDYLQJWKHWULPPLQJWRWKHIHHGEDFNFRQWUROOHU'HDGWLPHFRPSHQVDWLRQ
DOVRPDNHVXVHRIDVLPSOLÀHGSURFHVVPRGHOWRSUHGLFWDFRQWUROOHGYDULDEOHDQGKHOS
VWDELOLVH FORVHGORRS V\VWHPV ZLWK SXUH GHOD\V EHWZHHQ WKH DFWLRQ DQG WKH PHDVXUH
RILWVHIIHFWV7KLVLVIRUH[DPSOHWKHFDVHRIPDWHULDOÁRZFRQWURORQEHOWFRQYH\RUV
ZKHUHWKHUHLVDSXUHWUDQVSRUWGHOD\EHWZHHQWKHDFWXDWRUVIHHGHUVDQGWKHVHQVRU
EHOWVFDOH'HDGWLPHFRPSHQVDWLRQLVDSDUWLFXODUFDVHRIWKHPRUHJHQHUDOLQWHUQDO
PRGHOFRQWUROWHFKQLTXHIRUZKLFKWKHFRQWUROOHUHPEHGVDPRGHORIWKHV\VWHP
,QDGYDQFHGSURFHVVFRQWUROWKHSUHGLFWLYHFDSDELOLW\RIPRGHOVLVPDLQO\XVHGLQWZR
FRQWUROWHFKQLTXHV
 PRGHOEDVHGRSWLPLVLQJFRQWURO
2. model predictive control.
7\SLFDO PRGHOEDVHG RSWLPLVLQJ FRQWURO DSSOLFDWLRQV ZLOO PDNH XVH RI D PRGHO DQG
RSWLPLVDWLRQDOJRULWKPWRÀQGWKHEHVWVHWRIPDQLSXODWHGYDULDEOHYDOXHVWRPD[LPLVH
DQHFRQRPLFIXQFWLRQXQGHUDVHWRIFRQVWUDLQWV)RUD6$*PLOOWKLVZRXOGWUDQVODWHLQWR
ÀQGLQJ WKH EHVW IHHG UDWH IHHG GHQVLW\ DQG PLOO VSHHG WR PD[LPLVH WKURXJKSXW ZKLOH
NHHSLQJSURGXFWVL]HSHEEOHVSURGXFWLRQSRZHUGUDZDQGPLOOORDGLQUHDVRQDEOHUDQJHV
Model predictive controllers have found good success in chemical and petrochemical
industries, but their application in minerals processing has been less pervasive.

BIBLIOGRAPHY
Dunn,:& )XnGDPHntDOV oI ,nGXVtULDO ,nVtUXPHntDtLon DnG 3UoFHVV &ontUoOS0F*UDZ
+LOO%RRN&RPSDQ\1HZ<RUN
McMillan, * .  7XnLnJ DnG &ontUoO /ooS 3HUIoUPDnFH $ 3UDFtLtLonHU·V *XLGH, third edition,
$SSHQGL[.S,QWHUQDWLRQDO6RFLHW\RI$XWRPDWLRQ1RUWK&DUROLQD
Spitzer,':$GYDnFHG 5HJXODtoU\ &ontUoO $SSOLFDtLonV DnG 7HFKnLTXHVS0RPHQWXP
3UHVV1HZ<RUN
Williamson,-(7KHDXWRPDWLFFRQWURORIJULQGLQJPHGLDLQSHEEOHPLOOV-oXUnDO oI tKH 6oXtK
$IULFDn ,nVtLtXtH oI 0LnLnJ DnG 0HtDOOXUJ\²

Comminution Handbook | Spectrum Series 21 243

 HOME

Chapter 15

CASE STUDIES OF
CONTROL SYSTEMS
Dirk Baas, Duncan Bennett and Peter Walker1
Alain Broussaud, Geoffrey Legrand, Deon Kok,
Etienne Roux, Olivier Guyot and Marc Revalor2

INTRODUCTION
(YHU\ FRPPLQXWLRQ FLUFXLW FDQ EHQHÀW IURP HIIHFWLYH SURFHVV FRQWURO 7KH FRQWURO
V\VWHPVDW3KX.KDPDQG$KDIRGLVFXVVHGLQWKLVFKDSWHUDUHFRQFHUQHGZLWKORZ
JUDGHJROGRUHDQGORZJUDGHFRSSHURUHFLUFXLWVZKLFKPXVWKDYHKLJKWKURXJKSXWV
WR JHQHUDWH VXIÀFLHQW UHYHQXH LI PLQLQJ RSHUDWLRQV DUH WR EH YLDEOH 7KH FRQWURO RI
FRPPLQXWLRQFLUFXLWVIRUFRDOLQSRZHUVWDWLRQVZKLFKPXVWEHJURXQGÀQHHQRXJKWR
EXUQUDSLGO\LQWKHIXUQDFHDQGFHPHQWFOLQNHUZKLFKPXVWEHJURXQGÀQHHQRXJK
WRPDNHFRQFUHWHWKDWPHHWVVWDWXWRU\VWUHQJWKUHTXLUHPHQWVDUHGLVFXVVHGEULHÁ\LQ
&KDSWHU
&RPPLQXWLRQFLUFXLWVIRUORZJUDGHFRSSHUDQGJROGRUHVXVXDOO\FRQWDLQERWKVHPL
DXWRJHQRXVJULQGLQJ6$*PLOOVDQGEDOOPLOOVDQGRIWHQSURFHVVRUHVIURPGLIIHUHQW
GHSRVLWV 7KH FRPPLQXWLRQ UDWHV RI SDUWLFOHV LQ 6$* PLOOV GHSHQG RQ WKH RUH LQ WKH
PLOODQGFKDQJHLQRUHW\SHLQHYLWDEO\FKDQJHVWKHKDUGQHVVDQGIHHGVL]HRIWKHRUH

1. Phu Kham case study: this case study was originally published as: Baas, D, Bennett, D and Walker, P, 2014.
Developing process control standards for optimal plant performance at PanAust Limited, in Proceedings 12th
AusIMM Mill Operators’ Conference, pp 325–334 (The Australasian Institute of Mining and Metallurgy: Melbourne).
2. Ahafo case study: this case study is a modified version of a paper originally published as: Broussaud, A, Legrand, G,
Kok, D, Roux, E, Guyot, O and Revalor, M, 2011. Integrated advanced grinding control system at Newmont Ahafo,
in Proceedings SAG Conference 2011. It is reproduced here with the kind permission of the Canadian Institute of
Mining, Metallurgy and Petroleum.

Comminution Handbook | Spectrum Series 21 245

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

7KH WKHPH RI WKH $KDIR FDVH VWXG\ LV PLQLPLVLQJ WKH HIIHFW RI YDULDWLRQV LQ WKH
SURSHUWLHV RIRUHW\SHVZKLOHWKHFDVHVWXG\DW3KX.KDPOLQNVFRPPLQXWLRQFLUFXLW
FRQWUROZLWKWKHFRQWURORIWKHHQWLUHPLQLQJDQGSURFHVVLQJRSHUDWLRQ
3URFHVVFRQWUROLVDUDSLGO\PRYLQJWHFKQRORJ\EHFDXVHQHZPDWKHPDWLFDOWHFKQLTXHV
DUH EHLQJ GHYHORSHG IRU XVH LQ FRQWURO V\VWHPV QHZ VHQVRUV DUH EHFRPLQJ DYDLODEOH
DQG WKH FDSDFLW\ RI FRPSXWHUV LV LQFUHDVLQJ &RQWURO WHFKQLTXHV EDVHG RQ VHQVRUV
DQG PDWKHPDWLFDO WHFKQLTXHV WKDW KDYH EHHQ DYDLODEOH IRU \HDUV DUH UHIHUUHG WR DV
FRQYHQWLRQDOFRQWUROZKLOHWKRVHWKDWKDYHEHHQGHYHORSHGUHFHQWO\DUHUHIHUUHGWRDV
DGYDQFHGFRQWURO7KHDGYDQFHGFRQWUROV\VWHPDW$KDIRLVGLVFXVVHGDVDQH[DPSOHRI
WKHW\SHRIFRQWUROV\VWHPDSSOLHGWRODUJHVFDOHPLQHUDOSURFHVVLQJRSHUDWLRQV

GRINDING CIRCUIT CONTROL AT THE PHU KHAM MINE IN LAOS

Introduction
3DQ$XVW/LPLWHGLVDFRSSHUDQGJROGSURGXFHUZLWKRSHUDWLRQVLQ/DRVLQ6RXWK(DVW
$VLDDQGSURMHFWVLQ&KLOHDQG3DSXD1HZ*XLQHD7KHSURGXFLQJDVVHWVLQ/DRVDUH
WKH 3KX .KDP FRSSHU²JROG RSHUDWLRQ DQG WKH %DQ +RXD\[DL JROG²VLOYHU RSHUDWLRQ
7KH ORFDWLRQV RI 3KX %LD 0LQLQJ·V 3KX .KDP FRSSHU²JROG DQG %DQ +RXD\[DL JROG²
VLOYHU RSHUDWLRQV DUH VKRZQ LQ )LJXUH  7KH 3KX .KDP RSHUDWLRQ FRQVLVWV RI DQ

FIG 15.1 – Location of PanAust’s Phu Kham copper–gold and

Ban Houayxai gold–silver operations in Laos (image courtesy PanAust Ltd).

246 Comminution Handbook | Spectrum Series 21

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

RSHQSLWPLQHDQGDQ0WDFRQFHQWUDWRU7KHRSHUDWLRQXWLOLVHVFRQYHQWLRQDOVKRYHO
PLQLQJDQGWUXFNKDXODJHWRWKHSURFHVVLQJIDFLOLWLHV
7KH 3KX .KDP FRSSHU²JROG GHSRVLW LV D FRPSOH[ PLQHUDORJLFDO V\VWHP ZLWK
PLQHUDOLVDWLRQSUHVHQWLQLURQULFKVNDUQVVLOLFDULFKVWRFNZRUNDQGDOWHUHGGLVVHPLQDWHG
VW\OHV 7KH GHSRVLW KDV XQGHUJRQH QXPHURXV IDXOWLQJ IROGLQJ DQG DOWHUDWLRQ HYHQWV
FUHDWLQJ VLJQLÀFDQW YDULDELOLW\ LQ PLQHUDORJ\ DQG KDUGQHVV ZLWKLQ RUH EORFNV
7KH JHRORJ\ DQG PLQHUDORJ\ LQFOXGLQJ JDQJXH PLQHUDOV DUH GHVFULEHG LQ PRUH
GHWDLOE\%HQQHWW&UQNRYLFDQG:DONHU
7KH 3KX .KDP FLUFXLW FRQVLVWV RI D SULPDU\ J\UDWRU\ FUXVKHU  ð  P
 ð  LQFKHV IHG GLUHFWO\ E\ ' KDXO WUXFNV &UXVKHU GLVFKDUJH UHSRUWV WR D
FUXVKHGRUHVWRFNSLOHZLWKDQDSSUR[LPDWHOLYHFDSDFLW\RI0WYLDDQRYHUODQG
FRQYH\RU5HFODLPHGRUHIURPWKHVWRFNSLOHIHHGVWKH0:PGLDPHWHUðP
HIIHFWLYH JULQGLQJ OHQJWK (*/  IW Ǘ ð  IW (*/ 6$* PLOO RSHUDWLQJ LQ FORVHG
FLUFXLWZLWKSHEEOHVUHF\FOHGEDFNWRWKH6$*IHHG1RSHEEOHFUXVKLQJLVLQVWDOOHG
'RZQVWUHDPRIWKH6$*PLOOLVWKHEDOOPLOO²F\FORQHFLUFXLWFRQVLVWLQJRIWZR0:
 P GLDPHWHU ð  P (*/  IW Ǘ ð  IW (*/ EDOO PLOOV RSHUDWLQJ LQ FORVHG
FLUFXLWZLWKWZRFOXVWHUVRIPPGLDPHWHUF\FORQHV
&\FORQHRYHUÁRZVUHSRUWWRWKHÁRWDWLRQFLUFXLWZKLFKFRQVLVWVRIDEDQNRIWHQP
WDQN FHOO URXJKHUV DQG D EDQN RI ÀYH  Pó 5RXJKHU WDLOV DUH PL[HG ZLWK WKH FOHDQHU
VWDJH WDLOLQJV DQG IRUP WKH ÀQDO FRPELQHG WDLOLQJV VWUHDP ZKLFK ÁRZV E\ JUDYLW\ LQWR
D VXEDTXHRXV WDLOLQJV VWRUDJH IDFLOLW\ 5RXJKHU FRQFHQWUDWH LV FODVVLÀHG LQ D FOXVWHU RI
 PP UHJULQG F\FORQHV ZLWK RYHUÁRZ UHSRUWLQJ WR D -DPHVRQ &HOO XVHG LQ D FOHDQHU
IHHG VFDOSLQJ GXW\ 8QGHUÁRZ LV UHJURXQG LQ WZR RSHQFLUFXLW 0 ,VD0LOOŒ UHJULQG
PLOOV DQG UHMRLQV WKH RYHUÁRZ VWUHDP &RQFHQWUDWH IURP WKH -DPHVRQ &HOO UHSRUWV WR WKH
ÀQDO FRQFHQWUDWH WKLFNHQHU -DPHVRQ &HOO WDLOLQJV IHHG D FRQYHQWLRQDO WKUHHVWDJH FOHDQLQJ
FLUFXLW ZLWK WKH ÀUVW VWDJH LQ RSHQ FLUFXLW &LUFXLW GHWDLOV DQG SURFHVV GHYHORSPHQW
LQIRUPDWLRQ DUH SURYLGHG E\ %HQQHWW &UQNRYLF DQG :DONHU 
7KH 3KX .KDP FLUFXLW LV VKRZQ VFKHPDWLFDOO\ LQ )LJXUH 
3DQ$XVW KDV UHFRJQLVHG WKH EHQHÀWV RI XVLQJ SURFHVV FRQWURO DFURVV DOO IDFHWV RI LWV
RSHUDWLRQV LQ RUGHU WR PD[LPLVH FDSLWDO HIÀFLHQF\ DQG DVVHW XWLOLVDWLRQ 3URFHVV FRQWURO
VWUDWHJLHV FRQWURO PHWKRG GHYHORSPHQWV DQG VWDQGDUGV IRU SURGXFWLRQ FRQWURO DQG
UHSRUWLQJ DUH DSSOLHG WR H[LVWLQJ VLWHV DQG QHZ SURMHFWV

Process control strategy and approach

7KH SURFHVV FRQWURO VWUDWHJ\ LV WR PD[LPLVH UHWXUQ IURP WKH FRPSDQ\·V DVVHWV E\
DXWRPDWLRQ DQG FRQWURO RI DOO VWHSV LQ WKH YDOXH FKDLQ IURP WKH PLQH WR WKH ÀQDO SURGXFW
VDOH 3URFHVV FRQWURO VWUDWHJ\ GHYHORSPHQW IROORZV D WRSGRZQ DSSURDFK ZLWK SURFHVV
GULYHUV DQG NH\ LQSXWV DQG RXWSXWV GHWHUPLQHG IRU HDFK VWHS RI WKH SURGXFWLRQ FKDLQ
.H\ SHUIRUPDQFH PHDVXUHV DUH DOLJQHG DW HDFK VWHS EDVHG RQ ÀQDO SURGXFW TXDQWLW\ DQG
TXDOLW\ WR EH GHOLYHUHG WR WKH FXVWRPHU
3URFHVV FRQWURO ZDV ÀUVW LGHQWLÀHG DV DQ RSHUDWLRQV LPSURYHPHQW RSSRUWXQLW\ LQ HDUO\
 DQG DQ LQLWLDO DXGLW DQG UHYLHZ RI SODQW SURFHVV FRQWURO ZDV XQGHUWDNHQ LQ WKH
3KX .KDP FRQFHQWUDWRU 7KH DXGLW LGHQWLÀHG WKDW PRVW FRQWURO ORRSV LQVWDOOHG GXULQJ WKH
RULJLQDO SODQW GLVWULEXWHG FRQWURO V\VWHP '&6 SURJUDPPLQJ ZHUH EHLQJ UXQ PDQXDOO\
ZLWK RQO\ D IHZ RI WKH ORRSV WKDW ZHUH UXQQLQJ DXWRPDWLFDOO\ SURYLGLQJ WKH LQWHQGHG
SURFHVV FRQWURO SHUIRUPDQFH 7KH PDLQ UHDVRQ IRU ORRSV EHLQJ LQ PDQXDO ZDV WKH SRRU

Comminution Handbook | Spectrum Series 21 247

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

FIG 15.2 – Phu Kham simplified process flow diagram (image courtesy PanAust Ltd).

SHUIRUPDQFH RI LQVWDOOHG LQVWUXPHQWDWLRQ DQG FRQWURO HOHPHQWV SDUWLFXODUO\ ÁRZPHWHUV

DQG FRQWURO YDOYHV $ ERWWRPXS DSSURDFK WR SODQW SURFHVV FRQWURO LPSURYHPHQW ZDV
DGRSWHG EHFDXVH ZLWKRXW DFFXUDWH DQG UHOLDEOH PHDVXUHPHQW GHYLFHV LQ SODFH SURFHVV
FRQWURO LV QRW SRVVLEOH WR LPSOHPHQW 7KLV LV GHPRQVWUDWHG LQ )LJXUH  ZKHUH WKH ERWWRP
RI WKH S\UDPLG UHSUHVHQWV ÀHOG LQVWUXPHQWDWLRQ WKDW PHDVXUH SURFHVV SDUDPHWHUV VXFK DV
ÁRZ OHYHO DQG GHQVLW\ DQG FRQWURO HOHPHQWV VXFK DV YDOYHV DQG YDULDEOH VSHHG GULYHV XVHG
WR PDQLSXODWH SURFHVV YDULDEOHV ,I WKH EDVH HOHPHQWV KDYH VKRUWFRPLQJV LQ SHUIRUPDQFH
DQG UHOLDELOLW\ WKLV ZLOO ÁRZ XS WKH KLHUDUFK\ SUHYHQWLQJ KLJKHU OHYHOV IURP IXQFWLRQLQJ
DV LQWHQGHG 7KH VWDELOLVLQJ DQG UHJXODWRU\ FRQWURO FRYHUV VLQJOH ORRS FRQWUROOHUV VXFK
DV OHYHO RU ÁRZ ORRSV ,Q VRPH FDVHV WKHVH ORRSV DUH HQKDQFHG E\ XVLQJ WHFKQLTXHV VXFK
DV IHHGIRUZDUG DQG GHDG WLPH FRPSHQVDWLRQ $GYDQFHG FRQWURO FRPELQHV D QXPEHU
RI VLQJOHORRS FRQWUROOHUV LQWR D FRQWURO PRGHO ZKLFK KDQGOHV D FRPSOH[ RSHUDWLQJ XQLW
VXFK DV D 6$* PLOO RU ÁRWDWLRQ FLUFXLW 7KH DGYDQFHG FRQWURO OHYHO QRUPDOO\ PDNHV XVH RI
PRGHOEDVHG PRGHOSUHGLFWLYH UXOHEDVHG RU IX]]\ ORJLF FRQWURO VRPHWLPHV XVLQJ WKH
PHDVXUHPHQW RI SUR[\ YDULDEOHV WKDW DUH FRUUHODWHG WR NH\ YDULDEOHV +ROLVWLF FRQWURO LV
DQ H[WHQVLRQ RI DGYDQFHG FRQWURO LW FRPELQHV D QXPEHU RI XQLWV WRJHWKHU VXFK DV ÁHHW
PDQDJHPHQW DQG FUXVKLQJ RU JULQGLQJ DQG ÁRWDWLRQ DQG GHWHUPLQHV XQLW RSHUDWLQJ
LQSXWV DQG RXWSXWV UHTXLUHG WR DFKLHYH ÀQDO SURGXFW TXDQWLW\ DQG TXDOLW\ WDUJHWV
7KH ERWWRPXS LPSURYHPHQW DSSURDFK ZDV DSSOLHG LQ HDFK DUHD RI WKH SURFHVVLQJ SODQW
VWDUWLQJ ZLWK WKH FUXVKHU 7KH UDWLRQDOH ZDV WKDW E\ UHGXFLQJ XQLW RXWSXW YDULDELOLW\ IURP
WKH EHJLQQLQJ RI WKH SURFHVV SHUIRUPDQFH RI WKH GRZQVWUHDP XQLWV ZLOO EH LPSURYHG
,Q VRPH DUHDV WKH VLQJOH ORRS FRQWUROOHUV DQG UHJXODWRU\ FRQWURO ORRSV SURYLGH DGHTXDWH
SURFHVV FRQWURO ZKLOH LQ RWKHUV DGYDQFHG RU KROLVWLF FRQWURO LV UHTXLUHG

248 Comminution Handbook | Spectrum Series 21

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

FIG 15.3 – Hierarchy of control (image courtesy PanAust Ltd).

3ULRU WR  WKH FRQWURO URRP RSHUDWRU ZRXOG EH JLYHQ D VHW RI UHJXODWRU\ ORRS VHW
SRLQWV RU UDQJHV DV D JXLGH WR DFKLHYH WKH SODQW VKLIW REMHFWLYHV RU WDUJHWV 7KH RSHUDWRU
ZRXOG FRQWLQXDOO\ UHHYDOXDWH WKH VHW SRLQWV WR HQVXUH WKH SODQW SHUIRUPDQFH ZDV ZLWKLQ
WKH UHTXLUHG UDQJH DV ZHOO DV PRQLWRU WKH SHUIRUPDQFH RI WKH UHJXODWRU\ ORRSV ,Q HIIHFW
WKH RSHUDWRU ZDV DQ LQWHJUDO SDUW RI WKH SURFHVV FRQWURO V\VWHP ZKLFK OHDGV WR FRQVLGHUDEOH
FRQWURO YDULDELOLW\ ERWK ZLWKLQ DQG DFURVV VKLIWV DV RSHUDWRUV DUH XQDEOH WR PRQLWRU WKH
SURFHVV FRQWLQXDOO\ DQG RSHUDWRU UHVSRQVH WR D JLYHQ VLWXDWLRQ FDQ EH LQFRQVLVWHQW
%\ OLIWLQJ WKH OHYHO RI DXWRPDWHG FRQWURO WKURXJK UHOLDEOH GDWD DFTXLVLWLRQ DQG FRQWURO
RI WKH SURFHVV WKH RSHUDWRU LV UHPRYHG IURP UHJXODWRU\ FRQWURO GHFLVLRQV DQG LV RQO\
UHTXLUHG WR PRQLWRU WKH SHUIRUPDQFH RI WKH DGYDQFHG FRQWUROOHUV (OHYDWLQJ WKH UROH RI
WKH RSHUDWRU UHTXLUHV WKH SURFHVVLQJ SODQW WR EH HIIHFWLYHO\ DQG UHOLDEO\ DXWRPDWHG ZLWK
WKH LQWHUIDFH GHVLJQHG WR LQFUHDVH WKH RSHUDWRU·V VLWXDWLRQDO DZDUHQHVV RI WKH SODQW ZKLOH
QRW EHLQJ ÁRRGHG ZLWK H[FHVVLYH LQIRUPDWLRQ RU ZDUQLQJV
6WDUWLQJ DQ RSHUDWLRQ ZLWK WKH EHVW FRQWURO PHWKRGRORJLHV GHYHORSV D FXOWXUH WKDW
DFFHSWV DXWRPDWLRQ DQG FRQWURO ,QWURGXFLQJ DGYDQFHG FRQWURO PHWKRGV GXULQJ DQ
RSHUDWLRQ·V OLIH F\FOH UHTXLUHV ZLQGRZV RI RSSRUWXQLW\ IRU LPSOHPHQWDWLRQ DQG PXFK
FKDQJH PDQDJHPHQW LQSXW DQG RSHUDWRU WUDLQLQJ WR DYRLG SXVKEDFN LQ DGRSWLQJ WKH QHZ
PHWKRGV $OO RI WKHVH VWHSV LQFUHDVH WKH WLPH WR DFKLHYH FRQWURO EHQHÀWV DQG ZLWKRXW
VLJQLÀFDQW UHVRXUFHV DSSOLHG WR EHGGLQJLQ DGYDQFHG FRQWURO WKH RSHUDWLRQ ZLOO TXLFNO\
UHWXUQ WR LWV IRUPHU VWDWH
7KH ÀUVW SKDVH RI WKH ERWWRPXS SURFHVV ZDV WR HQDEOH VLPSOH VLQJOHORRS FRQWUROOHUV
WR RSHUDWH DV SHU GHVLJQ 7KH SUHPLVH ZDV WKDW UHJXODWRU\ DQG VWDELOLVLQJ DGYDQFHG
FRQWURO FDQ RQO\ EH LPSOHPHQWHG LI WKH LQVWUXPHQWDWLRQ DQG EDVLF FRQWURO LV IXQFWLRQLQJ
HQKDQFHPHQW ZLWK VWDQGDUG WHFKQLTXHV VXFK DV IHHG IRUZDUG UDWLR FRQWURO DQG GHDG
WLPH FRPSHQVDWLRQ EHFDPH SRVVLEOH
7KH VHFRQG SKDVH ZDV WR DSSO\ PHWKRGV RI VWDELOLVLQJ DGYDQFHG FRQWURO WR LPSURYH
WKH SHUIRUPDQFH RI WKH SODQW·V NH\ DUHDV VXFK DV FUXVKLQJ FRQYH\LQJ JULQGLQJ ÁRWDWLRQ

Comminution Handbook | Spectrum Series 21 249

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

DQG WKLFNHQLQJ ,PSURYHPHQWV LQ DUFKLWHFWXUH DQG GDWD VWRUDJH V\VWHPV ZHUH DOVR
LPSOHPHQWHG DOORZLQJ GHWDLOHG SURFHVV DQDO\VLV WR EH XQGHUWDNHQ RYHU ORQJ WLPH
SHULRGV $SSO\LQJ DGYDQFHG FRQWURO DQG GDWD DQDO\VLV KDV UHGXFHG YDULDELOLW\ LQ SURFHVV
RXWSXWV DQG GLVWXUEDQFH WR LQSXWV WR GRZQVWUHDP XQLWV RI WKH SURFHVV :KHUH FRPPRQ
SURFHVVHV H[LVW EHWZHHQ RSHUDWLQJ VLWHV VXFK DV 6$* PLOOLQJ DGYDQFHG FRQWURO VWUDWHJLHV
KDYH EHHQ XWLOLVHG UHJDUGOHVV RI GLIIHUHQFHV LQ KDUGZDUH DQG VRIWZDUH SODWIRUPV
7KH VXFFHVV RI DSSO\LQJ LPSURYHG DXWRPDWLRQ KDV SURPSWHG 3DQ$XVW WR GHYHORS
SURFHVV FRQWURO VWUDWHJLHV DQG VWDQGDUGV WKDW ZLOO EH XVHG IRU DOO SURMHFWV 7KH VWUDWHJLHV
DUH LQWHQGHG WR EH DSSOLHG DFURVV DOO SURGXFWLRQ DQG SURGXFW ORJLVWLFV DFWLYLWLHV WR
LQWHJUDWH DQG FRQWURO HDFK OLQN LQ WKH EXVLQHVV FKDLQ

Operations management strategy

7KH REMHFWLYHV RI WKH 3DQ$XVW FRQWURO VWUDWHJ\ DUH
x PD[LPLVLQJ DVVHW XWLOLVDWLRQ DQG SURGXFWLYLW\
x LPSURYLQJ UHOLDELOLW\ RI DVVHWV DQG SODQV
x SURGXFLQJ D WDUJHWHG TXDOLW\ SURGXFW WKURXJK WKH PLQLQJ FKDLQ IURP PLQLQJ WKURXJK
WR WKH ÀQDO FRQFHQWUDWH
x PLQLPLVLQJ LQYHQWRULHV DQG DVVRFLDWHG FRVW
x DQ HPSKDVLV RQ GDWD FROOHFWLRQ DQDO\VLV DQG SURFHVV FRQWURO WR HOLPLQDWH GHIHFWV DQG
DFKLHYH ORZFRVW DQG FRQWLQXRXV GHERWWOHQHFNLQJ
7KH RSHUDWLRQ RI PLQLQJ DQG SURFHVVLQJ SURGXFWLRQ DVVHWV UHTXLUHV FRQWURO ÁH[LELOLW\
WR EDODQFH XQLW LQSXWV DQG RXWSXWV IRU H[DPSOH XVLQJ PLOOLQJ DQG FUXVKHU PHWULFV WR
HQVXUH EOHQGLQJ DQG FDSDFLW\ WDUJHWV DUH DFKLHYHG 7KH ÁH[LELOLW\ PXVW DOVR DOORZ IRU
VLWHVSHFLÀF RSHUDWLQJ VWUDWHJLHV IRU H[DPSOH FKDQJLQJ WKH PLQLQJ SODQ RYHU VKRUW WLPH
SHULRGV WR LPSURYH PDWHULDOV KDQGOLQJ RU PLQHUDORJLFDO LVVXHV DIIHFWLQJ GRZQVWUHDP
SURFHVVLQJ SHUIRUPDQFH DQG PHWDO RXWSXW
7KH REMHFWLYH RI WKH VWUDWHJ\ LV WR FRQWLQXRXVO\ RSWLPLVH DQG UHGXFH FRQVWUDLQWV
H[SHULHQFHG E\ DQ\ RSHUDWLQJ XQLWV LPSDFWLQJ RQ SURGXFWLYLW\ ,I DQ RSHUDWLQJ XQLW
EHFRPHV D ERWWOHQHFN DQG IDLOV WR DFKLHYH LWV RSHUDWLQJ REMHFWLYHV WKH FRQWURO VWUDWHJ\ RI
D SURFHVV PD\ QHHG WR FKDQJH (DFK RSHUDWLQJ XQLW PXVW KDYH D ÁH[LEOH VHW RI RSHUDWLQJ
VWUDWHJLHV WKDW FDQ HDVLO\ EH LPSOHPHQWHG )LJXUH  VKRZV KRZ LQIRUPDWLRQ PXVW ÁRZ
EHWZHHQ XQLWV WR HQDEOH DOO XQLWV RI WKH RSHUDWLRQ WR DOLJQ ZLWK RSHUDWLRQDO REMHFWLYHV
7KH RSHUDWLQJ VWUDWHJLHV HQFRPSDVV QRW RQO\ DXWRPDWLRQ DQG SURFHVV FRQWURO IXQFWLRQV
EXW DOO DVSHFWV RI SURGXFWLRQ PDQDJHPHQW VXFK DV GULOO DQG EODVW GHVLJQV VWRFNSLOH DQG
UHKDQGOH PLQLPLVDWLRQ ZHW RUH KDQGOLQJ DQG IHHG EOHQGLQJ

System design
(DFK XQLW RI WKH SURFHVV KDV LQSXWV DQG RXWSXWV GHÀQHG LQ PHDVXUDEOH WHUPV VXFK DV
WRQQHV WKURXJKSXW SDUWLFOH VL]H DQG JUDGH ZKLFK GHWHUPLQH WKH OHYHO RI LQVWUXPHQWDWLRQ
UHTXLUHG 7KH XQLW·V VWUDWHJLHV VKRXOG DOVR EH ÁH[LEOH VR DV WR PHHW WKH ÀQDO SURGXFWLRQ
TXDQWLW\ DQG TXDOLW\ WDUJHWV 'XULQJ SURFHVV FRQWURO GHVLJQ WKH SURFHVV REMHFWLYHV DQG
VWUDWHJLHV IRU FRQWURO PXVW EH GHÀQHG 7KH FRQWURO VWUDWHJ\ WKHQ GHWHUPLQHV WKH SK\VLFDO
PHDVXUHPHQWV DQG ORFDWLRQ RI LQVWUXPHQWDWLRQ RQOLQH VDPSOLQJ DVVD\V GDWD FROOHFWLRQ
DQG UHWHQWLRQ UHSRUWLQJ DQG WKH FRQWURO V\VWHP UHTXLUHG 7KH VDPH SURFHVV DSSOLHV WR
DOO IXQFWLRQDO XQLWV RI WKH RSHUDWLRQ 8VLQJ VWDQGDUGV GXULQJ GHVLJQ HQVXUHV WKDW SURYHQ
HTXLSPHQW LV VHOHFWHG DQG LQVWDOOHG FRUUHFWO\ RU WKDW DOORZDQFH LV PDGH IRU HTXLSPHQW

250 Comminution Handbook | Spectrum Series 21

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

FIG 15.4 – Flexible operating strategies and objectives (image courtesy PanAust Ltd).

WR EH UHWURÀWWHG LI MXVWLÀHG E\ VLWH FRQGLWLRQV 7KH GHWDLOHG UHTXLUHPHQWV HQVXUH WKDW WKH
SURFHVV FRQWURO ZLOO EH DXWRPDWHG DFFRUGLQJ WR WKH GHVLJQ DQG IXQFWLRQLQJ DV VRRQ DV
FRPPLVVLRQLQJ KDV EHHQ FRPSOHWHG

Operational control design

7KH FRQWURO GHVLJQ FRPPHQFHV DIWHU GHYHORSLQJ WKH SURFHVV GHVLJQ FULWHULD DQG SURFHVV
ÁRZ GLDJUDPV IRU HDFK SURGXFWLRQ DQG ORJLVWLFV XQLW 'HWDLOHG FRQWURO GHVLJQV DUH
WKHQ GHYHORSHG IRU PHHWLQJ SURFHVV GHVLJQ REMHFWLYHV 6WDQGDUGV IRU FRQWURO V\VWHP
DUFKLWHFWXUH VRIWZDUH KDUGZDUH DQG UHSRUWLQJ DUH XVHG WR VSHFLI\ WKH HTXLSPHQW
UHTXLUHPHQWV FRQWURO V\VWHP IXQFWLRQDO VSHFLÀFDWLRQV DQG ÀQDO GHVLJQ 7KH VDPH
VWDQGDUGV DSSO\ IRU HDFK SURGXFWLRQ XQLW ZKLFK IDFLOLWDWHV LQWHUXQLW FRPPXQLFDWLRQV
DQG SUHYHQWV VRIWZDUH KDUGZDUH RU QHWZRUN LQFRPSDWLELOLW\ ,QWHJUDWLQJ DOO XQLWV
WRJHWKHU GXULQJ WKH GHVLJQ SKDVH RI D SURMHFW DQG XVLQJ SURYHQ FRQWURO PHWKRGV WR UHGXFH
YDULDELOLW\ KHOSV WR GHFUHDVH UHTXLUHG LQWHUXQLW VXUJH FDSDFLW\ ZKLFK FDQ VLJQLÀFDQWO\
UHGXFH FDSLWDO DQG RSHUDWLQJ FRVWV

Equipment
7KH GHVLJQ VWDQGDUGV DUH DSSOLHG WR WKUHH EURDG JURXSV RI LQVWUXPHQWDWLRQ DQG
HTXLSPHQW
*URXS $ 0DQGDWRU\ PLQLPXP HTXLSPHQW IRU DOO SURMHFWV UHTXLUHG WR PHHW FRQWURO
VWDQGDUGV
*URXS % 3URFHVVVSHFLÀF HTXLSPHQW WKDW LV PDQGDWRU\ IRU VRPH SURFHVVHV EXW QRW
RWKHUV
*URXS & 1RQPDQGDWRU\ HTXLSPHQW WKDW PD\ EH HFRQRPLFDOO\ MXVWLÀDEOH EDVHG RQ
VLWHVSHFLÀF UHTXLUHPHQWV RU QRW UHTXLUHG GXULQJ HDUO\ \HDUV DQG LQVWDOOHG DW
D ODWHU GDWH

Comminution Handbook | Spectrum Series 21 251

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

*URXS $ DQG % PDQGDWRU\ HTXLSPHQW LV LQFOXGHG LQ GHVLJQ IURP SUHOLPLQDU\ SURMHFW

VWXG\ VWDJHV *URXS & HTXLSPHQW PD\ QRW EH GHÀQHG RU LQFOXGHG LQ GHVLJQ XQOHVV LW
LV MXVWLÀHG E\ VLWHVSHFLÀF UHTXLUHPHQWV (DFK SURFHVV FRQWURO VWUDWHJ\ VSHFLÀHV WKH
LQVWUXPHQWDWLRQ UHTXLUHG WR PHHW WKH SURFHVV FRQWURO REMHFWLYHV
([DPSOHV RI *URXS $ PDQGDWRU\ HTXLSPHQW LQFOXGH ÁHHW PDQDJHPHQW V\VWHPV DQG
'&6 RU VXSHUYLVRU\ FRQWURO DQG GDWD DFTXLVLWLRQ 6&$'$ V\VWHPV ZHLJKWRPHWHUV
RQ FUXVKHU GLVFKDUJH DQG PLOO IHHG FRQYH\LQJ YDULDEOH VSHHG GULYHV RQ SXPSLQJ
HTXLSPHQW OHYHO WUDQVPLWWHUV RQ WDQNV DQG KRSSHUV ÁRZ DQG GHQVLW\ WUDQVPLWWHUV RQ
NH\ VWUHDPV DQG PHWDOOXUJLFDO VDPSOHUV *URXS % H[DPSOHV LQFOXGH SURFHVVVSHFLÀF
PDQGDWRU\ HTXLSPHQW VXFK DV RQVWUHDP DQDO\VLV IRU ÁRWDWLRQ RU RQVWUHDP F\DQLGH
FRQFHQWUDWLRQ DQDO\VHUV IRU JROG SURFHVVLQJ *URXS & QRQPDQGDWRU\ H[DPSOHV LQFOXGH
SDUWLFOH VL]H DQDO\VHUV RU VOXUU\ SDUWLWLRQLQJ XQLWV WKDW PD\ EH MXVWLÀHG IRU SULPDU\ RU
UHJULQG SURFHVVHV

Reliability
$ FRPPRQ IDLOLQJ UHFRJQLVHG GXULQJ FRQWURO VWDQGDUGV GHYHORSPHQW DQG UHFRUGHG
GXULQJ WKH  3KX .KDP DXGLW LV WKDW ÀHOG LQVWUXPHQWDWLRQ IRU SURFHVV FRQWURO LV
RIWHQ QRW FRQVLGHUHG FULWLFDO UHFHLYLQJ D ORZ PDLQWHQDQFH SULRULW\ EHFDXVH LW GRHV QRW
DFWXDOO\ VWRS WKH SURFHVV 7KH UHVXOW RI GHJUDGLQJ ÀHOG LQVWUXPHQWDWLRQ LV LQFUHDVLQJ
RSHUDWRU LQSXW DQG D PRYHPHQW WRZDUG PDQXDO FRQWURO LQFUHDVLQJ SURFHVV YDULDELOLW\
7KH DFFXUDF\ DYDLODELOLW\ DQG FRQVHTXHQFH RI IDLOXUH IRU HDFK PHDVXUHPHQW DQG
LQVWUXPHQW QHFHVVDU\ IRU FRQWURO PXVW EH VSHFLÀHG GXULQJ GHVLJQ DFFRUGLQJ WR WKH
LPSDFW RQ FRQWUROOHG RXWSXWV RI UDWH DQG TXDOLW\ 7KLV ZLOO GHWHUPLQH WKH W\SHV RI
LQVWUXPHQWV LQVWDOOHG ZKHWKHU EXLOWLQ UHGXQGDQF\ LV UHTXLUHG DQG D PDLQWHQDQFH
FULWLFDOLW\ UDWLQJ
5HOLDELOLW\ DQG SHUIRUPDQFH LV DFKLHYHG E\ VHOHFWLQJ ÀWIRUSXUSRVH DQG KLJKTXDOLW\
HTXLSPHQW WKDW LV FRUUHFWO\ LQVWDOOHG FDOLEUDWHG DQG PDLQWDLQHG &RQWURO PRGHOV
DUH GHVLJQHG ZLWK LQVWUXPHQW IDLOXUH FRQVLGHUDWLRQV LQ SODFH ,I DQ LQVWUXPHQW IDLOV
WKH FRQWURO PRGHO VKRXOG DGDSW WR WKH ORVV RI D PHDVXUHPHQW DQG QRW UHVSRQG ZLWK
LQDSSURSULDWH YLJRXU 'LIÀFXOWWRPDLQWDLQ HTXLSPHQW VXFK DV 6$* PLOO ORDG FHOOV FDQ
EH EDFNHG XS ZLWK EHDULQJ RLO SUHVVXUH WUDQVPLWWHUV WR HQVXUH FRQWURO LV PDLQWDLQHG
RWKHUZLVH UHGXQGDQW HTXLSPHQW PD\ EH MXVWLÀHG WR NHHS FULWLFDO FRQWURO IXQFWLRQV
RSHUDWLQJ

Control rooms
7KH 3KX .KDP SURFHVVLQJ RSHUDWLRQV ZHUH LQLWLDOO\ FRQWUROOHG E\ SODQW FRQWURO URRPV
ORFDWHG LQ WKH SODQW DUHDV 7KH PDLQ FRQWURO URRP ZDV ORFDWHG DW D FHQWUDO SRVLWLRQ QHDU
WKH JULQGLQJ FLUFXLW DQG RSHUDWRUV ZHUH DQ LQWHJUDO SDUW RI WKH SURFHVV FRQWURO V\VWHP
E\ ZDONLQJ WKH SODQW $W 3KX .KDP WKH FUXVKLQJ DQG ÁRWDWLRQ FLUFXLWV ZHUH FRQWUROOHG
IURP VPDOO FRQWURO KXWV ZLWKLQ WKH DUHD ZLWK RSHUDWRUV VSHQGLQJ WLPH RXW LQ WKH ÀHOG
DQG FRPLQJ EDFN WR PDNH FRQWURO DGMXVWPHQWV 7KH GLVDGYDQWDJHV RI ORFDO FRQWURO DUH
WKDW HDFK SURGXFWLRQ XQLW LV UXQ LQGHSHQGHQWO\ ZLWK PLQLPDO LQWHUDFWLRQ ZLWK RWKHU
SURGXFWLRQ XQLWV HYHQ ZLWKLQ WKH SURFHVV SODQW $Q\ FKDQJHV WR RUH EOHQGV ZHUH XVXDOO\
D VXUSULVH DQG FRQWURO ZDV UHDFWLYH %HQFKPDUNLQJ RI RWKHU RSHUDWLRQV VKRZHG WKDW
VLJQLÀFDQW DGYDQWDJHV ZHUH DFKLHYDEOH E\ KDYLQJ D FRPPRQ FHQWUDO FRQWURO URRP %RWK
3KX .KDP DQG %DQ +RXD\[DL QRZ KDYH D FHQWUDO FRQWURO URRP ZKHUH WKH PLQLQJ ÁHHW
PDQDJHPHQW V\VWHP DQG SURFHVVLQJ SODQW FRQWURO V\VWHPV DUH ORFDWHG

252 Comminution Handbook | Spectrum Series 21

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

7KH FHQWUDOLVHG FRQWURO IDFLOLWLHV DUH ORFDWHG LQ QHXWUDO JURXQG ZLWKLQ WKH DGPLQ
LVWUDWLRQ EXLOGLQJ DQG DUH GHVLJQHG WR HQKDQFH FRPPXQLFDWLRQ EHWZHHQ EXVLQHVV XQLWV
DQG HQDEOH KROLVWLF RSHUDWLRQV FRQWURO DQG UHSRUWLQJ )LJXUH  VKRZV WKH 3KX .KDP
FHQWUDOLVHG FRQWURO URRP WKDW DOORZV GLUHFW LQWHUDFWLRQ EHWZHHQ PLQLQJ DQG SURFHVVLQJ
DQG DUHD VXSHUYLVRUV $W IXWXUH RSHUDWLRQV WKH XVH RI UHPRWH FRQWURO URRPV PD\ EH
FRQVLGHUHG WR DYRLG WKH FRVWV RI WUDQVSRUWLQJ DQG DFFRPPRGDWLQJ RSHUDWRUV DW VLWH

FIG 15.5 – Phu Kham central control room (image courtesy PanAust Ltd).

Operational systems overview

'DWD FROOHFWLRQ VWRUDJH DQG UHSRUWLQJ ZHUH LGHQWLÀHG DV DQ RSHUDWLRQV LPSURYHPHQW
RSSRUWXQLW\ LQ  GXULQJ WKH 3KX .KDP  0WD XSJUDGH SURMHFW ZKHQ LW ZDV IRXQG
WKDW DFFHVVLQJ KLVWRULFDO RSHUDWLQJ GDWD IURP H[LVWLQJ V\VWHPV WR GHYHORS WKH SURMHFW
GHVLJQ ZDV FXPEHUVRPH 7R GULYH FRQWLQXRXV LPSURYHPHQW NH\ RSHUDWLQJ LQIRUPDWLRQ
PXVW EH UHDGLO\ DYDLODEOH WR DOO OHYHOV RI SHUVRQQHO ZLWKLQ WKH RSHUDWLRQ HQDEOLQJ
SHUIRUPDQFH DQDO\VLV DQG HIIHFWLYH SODQQLQJ DQG PDQDJHPHQW FRQWURO 'DWD KLVWRU\ DQG
DQDO\VLV DQG UHSRUWLQJ V\VWHPV KDYH EHHQ LPSOHPHQWHG DV D FRPSRQHQW RI WKH VWDQGDUG
SURGXFWLRQ PDQDJHPHQW V\VWHPV
6RPH RI WKH LPSOHPHQWHG FRQWURO PRGHOV DW 3KX .KDP DUH GHVFULEHG EULHÁ\ LQ WKH
IROORZLQJ VHFWLRQV

Fleet control and reporting

$ ÁHHW PDQDJHPHQW V\VWHP DW WKH PLQH PRQLWRUV LQGLYLGXDO SLHFHV RI HTXLSPHQW DQG
RSWLPLVHV ÁHHW XWLOLVDWLRQ $OO PDMRU SLHFHV RI PLQLQJ HTXLSPHQW DUH PRQLWRUHG DQG
FRQWUROOHG E\ WKH ÁHHW PDQDJHPHQW FRQWURO V\VWHP LQ WKH FHQWUDO FRQWURO URRP 7KH ÁHHW
PDQDJHPHQW V\VWHP LV LQWHJUDWHG ZLWK WKH SURFHVVLQJ SODQW WR HQVXUH WKH FUXVKLQJ SODQW
LV QRW ZDLWLQJ IRU RUH DQG RUH FKDUDFWHULVWLFV RXWSXWV RI WKH JUDGH FRQWURO DQG JHRORJ\
PRGHOV DUH IHG IRUZDUG WR WKH FRQFHQWUDWRU LQ UHDO WLPH

Crushing
7KH SULPDU\ REMHFWLYH RI WKH FUXVKHU FRQWURO VWUDWHJ\ LV WR PRYH WKH RUH WKURXJK WKH
FUXVKHU WR WKH FRDUVH RUH VWRFNSLOH ZLWK PLQLPDO UDWH YDULDWLRQ 7KH FUXVKHU FDQ EH VHW

Comminution Handbook | Spectrum Series 21 253

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

XS WR DFKLHYH WKH UHTXLUHG SDUWLFOH VL]H GLVWULEXWLRQ DQG IRUVDNH WKURXJKSXW E\ VKXWWLQJ
WKH FORVHG VLGH VHWWLQJ RU PD[LPLVH WKURXJKSXW E\ RSHQLQJ WKH FORVHG VLGH VHWWLQJ
7KH FUXVKHG RUH YDXOW XQGHU WKH J\UDWRU\ FUXVKHU LV ÀWWHG ZLWK D UHOLDEOH OHYHO VHQVRU
DQG WKH FUXVKHG RUH FRQYH\RU LV ÀWWHG ZLWK D ZHLJKWRPHWHU ZKLFK LV ORFDWHG DV FORVH
DV PHFKDQLFDOO\ SRVVLEOH WR WKH FUXVKHG RUH YDXOW GLVFKDUJH DSURQ IHHGHU 7KH FUXVKHG
RUH ZHLJKW FRQWUROOHU PDNHV XVH RI GHDG WLPH FRPSHQVDWLRQ WR FRQWURO WKH ORDGLQJ UDWH
RQWR WKH RYHUODQG FRQYH\RU HQVXULQJ WKH PD[LPXP UDWH LV PDLQWDLQHG ZLWKRXW WKH EHOW
EHLQJ RYHUORDGHG WKLV LPSURYHV FRQYH\RU DYDLODELOLW\ WKURXJK UHGXFHG VSLOODJH DQG EHOW
GDPDJH 7KH PLQLQJ ÁHHW PDQDJHPHQW V\VWHP UHFHLYHV LQIRUPDWLRQ IURP WKH FUXVKHU
LQFOXGLQJ WKH FXUUHQW FUXVKHU WKURXJKSXW VHW SRLQW WR SURYLGH LQSXW LQWR WUXFN QXPEHUV
IRU WKH FUXVKHU WR PHHW SURGXFWLRQ WDUJHWV

Grinding control
7KH  0: 3KX .KDP 6$* PLOO LV XVHG LQ D VWDQGDUG 6$* DQG EDOO PLOO FRQÀJXUDWLRQ
DV GHVFULEHG E\ &UQNRYLF et al  ZLWK WKH 6$* PLOO SHEEOHV UHFLUFXODWLQJ WR 6$*
PLOO IHHG ,QLWLDOO\ WKH RQO\ DXWRPDWLF ORRSV LQVWDOOHG ZHUH 6$* PLOO WRQQHV IHHG UDWH DQG
ZDWHU DGGLWLRQ 7KH RSHUDWRU ZRXOG FRQWLQXDOO\ DGMXVW WKH IHHG UDWH VHW SRLQW PLOO VSHHG
DQG 6$* PLOO GHQVLW\ WR FRQWURO WKH 6$* PLOO ZHLJKW 7KH IHHG UDWH FRQWUROOHU GLG QRW
SURYLGH D VWDEOH RXWSXW DQG ZRXOG RVFLOODWH LI WKH RSHUDWRU FKDQJHG WKH VHW SRLQW RU WKH
VSHHG UDWLR EHWZHHQ WKH WZR IHHGHUV ORFDWHG XQGHU WKH FRDUVH RUH VWRFNSLOH
7KH GHYHORSHG JULQGLQJ FLUFXLW FRQWURO VWUDWHJ\ LV WR PD[LPLVH WKURXJKSXW ZKLOH
NHHSLQJ WKH SURGXFW JULQG VL]H DV ÀQH DV SRVVLEOH 6$* DQG EDOO PLOO 6$% FLUFXLWV DUH
XVHG DW ERWK WKH 3KX .KDP DQG %DQ +RXD\[DL RSHUDWLRQV ZLWK WKH FRQWURO VWUDWHJLHV
LGHQWLFDO DW ERWK SODQWV GHVSLWH GLIIHUHQW KDUGZDUH DQG VRIWZDUH SODWIRUPV 7KH 6$*
PLOO FRQWURO PRGHO ZLOO YDU\ WKH 6$* PLOO IHHG UDWH DQG VSHHG WR PDLQWDLQ WKH UHTXLUHG
PLOO ZHLJKW VHW SRLQW 7KH RSWLPDO PLOO ZHLJKW LV VHW DFFRUGLQJ WR PLOO YROXPH ZLWK
FRPSHQVDWLRQ PDGH IRU OLQHU ZHDU
$ FULWLFDO FRPSRQHQW WR DFKLHYLQJ JRRG 6$* PLOO FRQWURO LV WR KDYH VWDEOH DQG IDVW
DFWLQJ FRQWURO RI WKH IHHG UDWH 'XULQJ LQLWLDO PRGHO GHYHORSPHQW WKH ZHLJKWRPHWHU
RXWSXW ZDV IRXQG WR KDYH D VHFRQG ODJ ÀOWHU 7KLV ZDV UHPRYHG WKH IHHGHU UDWLR
FRQWURO ORJLF FKDQJHG DQG GHDGWLPH FRPSHQVDWRUV LQFOXGHG UHVXOWLQJ LQ D IDVWDFWLQJ
VWDEOH IHHG UDWH FRQWUROOHU
7KH G\QDPLFV RI WKH 3KX .KDP 6$* PLOO ZHUH WKHQ LGHQWLÀHG E\ XVLQJ WKH PHWKRGV
GHÀQHG E\ =KX  ZKHUH WKH EHKDYLRXU RI WKH SURFHVV LV XVHG WR LGHQWLI\ G\QDPLF
UHVSRQVH UDWKHU WKDQ GHULYDWLRQ IURP ÀUVW SULQFLSOHV 7KH DQDO\VLV XVHG KLVWRULFDO
GDWD IURP DOO VHW SRLQW FKDQJHV WKH RSHUDWRUV KDG PDGH RYHU D VL[PRQWK SHULRG
7KH EHKDYLRXU RI WKH PLOO LQ UHVSRQVH WR FKDQJHV LQ PLOO VSHHG IHHG UDWH DQG PLOO GHQVLW\
YDULDEOHV XQGHU GLIIHUHQW FRQGLWLRQV ZDV DQDO\VHG WR GHYHORS WKH FRQWURO PRGHO IRU WKH
6$* PLOO 7KH FRQWURO PRGHO ZDV LPSOHPHQWHG LQ -DQXDU\  DQG RQO\ UHTXLUHG PLQRU
WXQLQJ FKDQJHV EHIRUH JDLQLQJ RSHUDWRU FRQÀGHQFH 7KH XWLOLVDWLRQ RI WKH PRGHO ZDV
WUDFNHG E\ WKH '&6 DQG UHFRUGHG LQ WKH KLVWRU\ WR HQVXUH WKH UHDVRQV IRU WXUQLQJ RII WKH
FRQWURO PRGHO FRXOG EH UHYLHZHG ZLWK WKH RSHUDWRU UHVSRQVLEOH
$Q DQDO\VLV RI WKH SHUIRUPDQFH RI WKH 3KX .KDP 6$* PLOO FRQWURO ZDV SHUIRUPHG E\
FRPSDULQJ WZR KRXU SHULRGV EHIRUH 1RYHPEHU  DQG DIWHU )HEUXDU\  WKH
FRQWURO PRGHO ZDV LPSOHPHQWHG ZLWK GDWD EHLQJ ORJJHG HYHU\ ÀYH VHFRQGV 'XH WR
FKDQJHV LQ RUH FKDUDFWHULVWLFV RYHU VKRUW SHULRGV DW 3KX .KDP WKH PHDVXUH RI FRQWURO

254 Comminution Handbook | Spectrum Series 21

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

SHUIRUPDQFH XVHG ZDV FRHIÀFLHQW RI YDULDQFH ZKLFK LV WKH VWDQGDUG GHYLDWLRQ YDOXH DV
D SHUFHQWDJH RI WKH PHDQ YDOXH
3ULRU WR LPSOHPHQWDWLRQ RI WKH FRQWURO PRGHO WKH 6$* PLOO ZHLJKW WHQGHG WR GULIW
DFURVV D EURDG UDQJH DV VKRZQ LQ WKH IUHTXHQF\ KLVWRJUDP LQ )LJXUH  7KH DYHUDJH
PLOO ZHLJKW FKDQJHG IURP  W EHIRUH LPSOHPHQWDWLRQ WR  W DIWHU LPSOHPHQWDWLRQ
ZKLOH WKH VWDQGDUG GHYLDWLRQ IURP WKH DYHUDJH ZHLJKW UHGXFHG IURP  W  SHU FHQW
FRHIÀFLHQW RI YDULDQFH WR  W  SHU FHQW FRHIÀFLHQW RI YDULDQFH LQGLFDWLQJ D VLJQLÀFDQW
LPSURYHPHQW LQ FRQWURO
7KH 6$* PLOO SRZHU VKRZQ LQ WKH )LJXUH  IUHTXHQF\ KLVWRJUDP GHPRQVWUDWHV
WKH LPSURYHG SRZHU GUDZ FRQWURO ZLWK UHGXFHG GHYLDWLRQ IURP WKH PHDQ 3ULRU WR
LPSOHPHQWDWLRQ RI WKH FRQWURO PRGHO WKH PHDQ 6$* SRZHU GUDZ ZDV  N: ZLWK
D VWDQGDUG GHYLDWLRQ RI  N:  SHU FHQW FRHIÀFLHQW RI YDULDQFH $IWHUZDUGV 6$*
PLOO PHDQ SRZHU GUDZ ZDV   N: ZLWK D VWDQGDUG GHYLDWLRQ RI  N:  SHU FHQW
FRHIÀFLHQW RI YDULDQFH 7KH UHGXFHG VWDQGDUG GHYLDWLRQ ZDV DJDLQ DQ LQGLFDWLRQ RI
LPSURYHG 6$* PLOO FRQWURO
7KH 6$* PLOO FRQWURO V\VWHP WKURXJKSXW SHUIRUPDQFH UHYLHZ SUHVHQWHG LQ )LJXUH 
LQGLFDWHV WKDW SULRU WR LPSOHPHQWDWLRQ RI WKH PRGHOEDVHG FRQWURO WKH WKURXJKSXW
DYHUDJH ZDV  WK ZLWK D VWDQGDUG GHYLDWLRQ RI  WK  SHU FHQW FRHIÀFLHQW RI
YDULDQFH )ROORZLQJ FRQWURO PRGHO LPSOHPHQWDWLRQ WKH IHHG UDWH LQFUHDVHG WR  WK
ZLWK D VWDQGDUG GHYLDWLRQ RI  WK  SHU FHQW FRHIÀFLHQW RI YDULDQFH

6000
5000
Frequency

4000
3000
Before
2000
After
1000
0
300
315
330
345
360
375
390
405
420
435
450
465
480

SAGMillWeight(tonnes)
FIG 15.6 – Semi-autogenous grinding (SAG) mill weight histogram (image courtesy PanAust Ltd).

5000

4000
Frequency

3000

2000 Before
1000 After
0
6000
6800
7600
8400
9200
10000
10800
11600
12400
13200
14000
14800

SAGMillPower(kW)

FIG 15.7 – Semi-autogenous grinding (SAG) mill power histogram (image courtesy PanAust Ltd).

Comminution Handbook | Spectrum Series 21 255

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

2000

1500

Frequency
1000
Before
500 After

0 1100
1175
1250
1325
1400
1475
1550
1625
1700
1775
1850
1925
2000
2075
2150
SAGMillFeedRate(t/h)

FIG 15.8 – Semi-autogenous grinding (SAG) mill feed rate histogram (image courtesy PanAust Ltd).

$V LV ZHOO NQRZQ PD[LPLVLQJ PLOO SRZHU PHDQV LQFUHDVHG WKURXJKSXW ZKLFK ORZHUV
RSHUDWLQJ FRVWV DQG LPSURYHV FRQFHQWUDWH SURGXFWLRQ WKDW LV UHYHQXH DOO RWKHU WKLQJV
EHLQJ HTXDO 7KH GDWD SUHVHQWHG LQ )LJXUHV  DQG  FOHDUO\ GHPRQVWUDWHV WKLV
LPSURYHPHQW DQG WKH EHQHÀWV RI FRQWUROOLQJ SURFHVVLQJ RSHUDWLRQV

Conclusions
3URFHVV FRQWURO DFURVV DOO IDFHWV RI SURGXFWLRQ RSHUDWLRQV KDV EHHQ UHFRJQLVHG WR
PD[LPLVH FDSLWDO HIÀFLHQF\ SURGXFWLYLW\ SURGXFW TXDOLW\ DQG DVVHW XWLOLVDWLRQ 3URFHVV
FRQWURO VWUDWHJLHV FRQWURO PHWKRG GHYHORSPHQW DQG VWDQGDUGV IRU SURGXFWLRQ FRQWURO
DQG UHSRUWLQJ DUH DSSOLHG WR H[LVWLQJ VLWHV DQG QHZ SURMHFWV UHJDUGOHVV RI SURFHVV DQG
VLWHVSHFLÀF GLIIHUHQFHV LQ FRQWURO KDUGZDUH DQG VRIWZDUH
7KH GHYHORSPHQW RI RSHUDWLRQDO FRQWURO VWUDWHJLHV DQG FRQWURO VWDQGDUGV KDV IROORZHG
D WRSGRZQ DSSURDFK EDVHG RQ DXWRPDWLRQ DQG FRQWURO RI DOO NH\ LQSXWV DQG RXWSXWV
WKURXJK WKH SURGXFWLRQ FKDLQ WR PHHW ÀQDO SURGXFW TXDQWLW\ DQG TXDOLW\ WDUJHWV
'HYHORSLQJ VWDQGDUG SURFHVV FRQWURO VWUDWHJLHV DQG VSHFLÀFDWLRQ RI WKH HTXLSPHQW DQG
LQVWUXPHQWDWLRQ UHTXLUHG IRU FRQWURO GXULQJ GHVLJQ KDV SURYHQ WR UHDOLVH SURGXFWLRQ
EHQHÀWV DOPRVW DV VRRQ DV FRPPLVVLRQLQJ KDV EHHQ FRPSOHWHG
7KH LPSURYHPHQW DQG GHYHORSPHQW RI SURFHVV FRQWURO KDV IROORZHG D ERWWRPXS
DSSURDFK RI JHWWLQJ WKH ULJKW HTXLSPHQW DQG EDVLF FRQWURO LQ SODFH DQG RSHUDWLQJ UHOLDEO\
EHIRUH LQFUHDVLQJ DXWRPDWLRQ ZLWKLQ DQG EHWZHHQ SURGXFWLRQ XQLWV %\ HIIHFWLYHO\
DSSO\LQJ DGYDQFHG FRQWURO PRGHOV WR LQGLYLGXDO XQLWV RI SODQW WKH YDULDELOLW\ LQ WKHLU
PHDVXUDEOH RXWSXWV KDV EHHQ UHGXFHG 7KLV DOORZV HDFK XQLW WR PD[LPLVH SURGXFWLYLW\
DQG UXQ FRQVLVWHQWO\ WR WKH SK\VLFDO DQG TXDOLW\ FRQVWUDLQWV RI WKH SURGXFWLRQ XQLWV
7KH UHGXFHG YDULDELOLW\ LQ HDFK SURFHVVLQJ XQLW KDV EHHQ GHPRQVWUDWHG WR LPSURYH WKH
SHUIRUPDQFH RI GRZQVWUHDP XQLWV

Acknowledgements
7KH DXWKRUV DFNQRZOHGJH WKH VXSSRUW RI 3DQ$XVW PDQDJHPHQW DQG WKH FRQWULEXWLRQ
RI RSHUDWRUV DQG PHWDOOXUJLVWV DW 3KX .KDP DQG %DQ +RXD\[DL ZKR SURYLGHG YDOXDEOH
IHHGEDFN GXULQJ GHYHORSPHQW RI WKH DGYDQFHG FRQWURO PRGHOV LPSOHPHQWHG :LWKRXW
WKH SRVLWLYH GDLO\ LQSXW DQG LQVLJKW LQWR WKH SURFHVV WKH FRQWURO PRGHOV FRXOG QRW KDYH
SURYLGHG WKH UREXVW OHYHO RI FRQWURO DFKLHYHG

256 Comminution Handbook | Spectrum Series 21

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

GRINDING CIRCUIT CONTROL AT NEWMONT’S AHAFO GOLD MINE

IN GHANA
7KH $KDIR PLQH LV ORFDWHG VRXWK RI WKH FLW\ RI 6X\DQL LQ WKH FHQWUDO ZHVW SDUW RI
*KDQD )LJXUH  1HZPRQW 0LQLQJ &RUSRUDWLRQ RSHUDWHV IRXU RSHQ SLWV DW $KDIR
ZLWK UHVHUYHV FRQWDLQHG LQ  SLWV ,Q  $KDIR SURGXFHG   R] RI JROG DQG DW
 'HFHPEHU  UHSRUWHG  0R] RI JROG UHVHUYHV
:KHQ SURFHVVLQJ VWDUWHG LQ  FRQWURO GHYLFHV ZHUH FRQYHQWLRQDO KRZHYHU LQ 
1HZPRQW GHFLGHG WR LQVWDOO DQ DGYDQFHG SURFHVV FRQWURO $3& V\VWHP DQG HQWHUHG LQWR
DQ DJUHHPHQW ZLWK 0HWVR IRU LWV LPSOHPHQWDWLRQ 7KH FRQWURO VWUDWHJ\ LV GHVLJQHG WR
DGGUHVV VSHFLÀF GLVWXUEDQFHV LQFOXGLQJ VWDUWVWRS RI SHEEOH FUXVKHUV XQDYDLODELOLW\ RI
R[LGHV FKDQJHV LQ RUH W\SH DQG KDUGQHVV DQG FKDQJHV LQ URFN VL]H GLVWULEXWLRQ UHVXOWLQJ
IURP VHJUHJDWLRQ LQ WKH VWRFNSLOH 7KH SURSRVHG V\VWHP FRQVLVWHG RI DQ H[SHUW V\VWHP
ZLWK H[WHQVLYH XVH RI IX]]\ ORJLF VXSSOHPHQWHG E\ DQ RQOLQH G\QDPLF PRGHO RI WKH
SURFHVV XVHG DV D VRIW VHQVRU D 9LVLR5RFNΠURFN VL]H PHDVXUHPHQW VHW RYHU WKH 6$*
PLOO IHHG EHIRUH DGGLWLRQ RI UHF\FOHG SHEEOHV DQG DQ $XGLR0LOOΠV\VWHP WR DQDO\VH WKH
6$* PLOO VRXQG DQG GHWHFW LPSDFWV RQ OLQHUV 7KHVH FRPSRQHQWV DUH HPEHGGHG LQ WKH
2&6© VRIWZDUH SODWIRUP ZKLFK PDNHV LW HDV\ WR VXSSRUW WKH HQWLUH V\VWHP UHPRWHO\

FIG 15.9 – Map of Ghana (image courtesy GhanaWeb).

The grinding circuit at Ahafo

6HSDUDWH RUH VWDWLRQV UHFHLYH R[LGH DQG SULPDU\ RUHV IURP GLIIHUHQW RSHQ SLWV
7KH FRPPLQXWLRQ FLUFXLW LV VKRZQ LQ )LJXUH  2[LGH RUH IHHGV D  PP 00'
6L]HU GLVFKDUJLQJ GLUHFWO\ RQWR WKH 6$* IHHG FRQYH\RU 3ULPDU\ RUH LV FUXVKHG E\ D
 ð  P  ð  LQ J\UDWRU\ FUXVKHU DQG JRHV WR D VWRFNSLOH ZLWK D OLYH FDSDFLW\
RI DSSUR[LPDWHO\  KRXUV RI SODQW RSHUDWLRQ 7KH JULQGLQJ FLUFXLW LV D 6$* PLOO ² EDOO
PLOO ² FUXVKHU 6$%& FLUFXLW 7KUHH DSURQ IHHGHUV UHFRYHU RUH IURP WKH VWRFNSLOH DQG

Comminution Handbook | Spectrum Series 21 257

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

Priŵary
Kre

Kdžide
Kre

FIG 15.10 – Ahafo comminution circuit (MMD Sizer, gyratory/pebble crushers and mills) (image courtesy Hakan Dundar).

GLVFKDUJH RQWR WKH 6$* IHHG FRQYH\RU 7KH JULQGLQJ FLUFXLW FRQVLVWV RI D  P ¡ ð
 P (*/  IW ¡ ð  IW > IW (*/@ 0HWVR 6$* PLOO ZLWK D   N: GXDO SLQLRQ
GULYH DQG D  IW ¡ ð  IW 0HWVR EDOO PLOO DOVR ZLWK D   N: GULYH LQ FORVHG FLUFXLW
ZLWK D FOXVWHU RI  ð  PP ¡  LQ ¡ F\FORQHV 3HEEOHV DUH UHF\FOHG WR EXIIHU ELQV DQG
FUXVKHG E\ D KHDY\GXW\ 0HWVR 03 FUXVKHU
7KH SURSRUWLRQ RI RUH H[WUDFWHG IURP GLIIHUHQW SLWV YDULHV RYHU WLPH 7KH $SHQVX
RUH LV ÀQHU DQG VRIWHU DQG WKH SODQW WHQGV WR EH PRUH RIWHQ EDOO PLOO OLPLWHG ZKHQ WKH
SURSRUWLRQ RI $SHQVX RUH LV KLJKHU 7KH 6XELND RUH LV JHQHUDOO\ KDUGHU LW UHTXLUHV
PRUH 6$* PLOOLQJ DQG LV NQRZQ WR \LHOG KLJKHU JROG UHFRYHULHV 7KH R[LGH RUH FRPHV
IURP WKH XSSHU OHYHOV RI WKH GHSRVLW ,W LV FRQYHQLHQW WR SURFHVV PRUH R[LGH RUH ZKHQ
WKH FLUFXLW EHFRPHV 6$* PLOO OLPLWHG *ULQGLQJ LV D FULWLFDO SDUW RI WKH SURFHVV VLQFH
JROG UHFRYHU\ LV YHU\ VHQVLWLYH WR JULQG VL]H ,Q RUGHU WR DFKLHYH D JRRG JROG UHFRYHU\
LW LV QHFHVVDU\ WR SURGXFH D SDUWLFOH VL]H GLVWULEXWLRQ 36' ZLWK  SHU FHQW RU PRUH
SDVVLQJ  ǍP

Conventional controls
7KH LQVWUXPHQWDWLRQ RI WKH JULQGLQJ FLUFXLW LV FRPSUHKHQVLYH DQG ZHOO PDLQWDLQHG
7KHUH DUH IRXU ZHLJKWRPHWHUV PHDVXULQJ WRQQDJH RI R[LGH RUH WRWDO IUHVK IHHG R[LGH
SOXV SULPDU\ RUH WRWDO PLOO IHHG LQFOXGLQJ UHF\FOHG SHEEOHV OLPH DQG JULQGLQJ EDOOV
DQG SHEEOHV SURGXFHG E\ WKH 6$* PLOO 0RVW ZDWHU DGGLWLRQV DUH PHDVXUHG DV ZHOO DV
WKH SRZHU ORDG FDOFXODWHG IURP EHDULQJ SUHVVXUH DQG URWDWLRQDO VSHHG RI WKH 6$*
PLOO 6$* VSHHG LV FRQWUROOHG ZLWK D OLTXLG UHVLVWDQFH VWDUWHU /56 EXW DV IUHTXHQWO\
UHTXLUHG E\ /56 WKH PLOO VSHHG VKRXOG QRW EH PRGLÀHG FRQWLQXRXVO\ RU IUHTXHQWO\

258 Comminution Handbook | Spectrum Series 21

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

7KH VXPS LQWR ZKLFK ERWK PLOOV GLVFKDUJH LV HTXLSSHG ZLWK D YDULDEOH VSHHG SXPS
DQG D OHYHO SUREH +\GURF\FORQH IHHG UDWH GHQVLW\ DQG SUHVVXUH DUH PHDVXUHG DQG WKH
K\GURF\FORQHV DUH DXWRPDWHG
7KLV LQLWLDO LQVWUXPHQWDWLRQ ZDV VXSSOHPHQWHG DW WKH VWDUW RI WKH DGYDQFHG FRQWURO
SURMHFW ZLWK 9LVLR5RFN FDPHUDV WR PHDVXUH RUH VL]H GLVWULEXWLRQ RQOLQH DQG DQ
$XGLR0LOO PLFURSKRQH WR DQDO\VH WKH VRXQG JHQHUDWHG E\ WKH 6$* PLOO DQG GHWHFW
GLUHFW LPSDFWV RI PDWHULDO RQ WKH OLQHUV $Q 2XWRWHF 36,® VHQVRU PHDVXUHV WKH 36' RI
F\FORQHV RYHUÁRZ DQG WKH UHVXOW LV LQWHJUDWHG LQWR WKH FRQWURO ORJLF 7KH FRQWURO URRP
LV ORFDWHG QHDU WKH PLOOV

Disturbances
7KH W\SLFDO IXQFWLRQ RI DQ $3& V\VWHP LV WR UHDFW WR GLVWXUEDQFHV VR WKDW PD[LPXP
WKURXJKSXW UHDVRQDEO\ VWDEOH RSHUDWLRQ DQG VXLWDEOH SURGXFW TXDOLW\ DUH FRQWLQXRXVO\
DFKLHYHG 7KH PDLQ GLVWXUEDQFHV DW $KDIR DUH
x Changes in ore type, hardness and size distribution 7KLV UHVXOWV IURP WKH KHWHURJHQHLW\
RI WKH GHSRVLW DQG WKH LQHYLWDEOH GLVFRQWLQXLWLHV DGGHG E\ WKH RSHUDWLRQ RI WKH PLQH
EODVWLQJ DQG WUXFN WUDQVSRUWDWLRQ DUH GLVFRQWLQXRXV E\ WKHLU QDWXUH
x Segregation in the stockpile 7KH FRDUVHVW URFNV WHQG WR UROO GRZQ WR WKH SHULSKHUDO
SDUWV RI WKH VWRFNSLOH $OWKRXJK VRRQHU RU ODWHU DOO SDUWLFOHV RI RUH DUH SURFHVVHG
WKH SURSRUWLRQ RI URFNV DQG ÀQH SDUWLFOHV LQ WKH PDWHULDO H[WUDFWHG E\ WKH DSURQ
IHHGHUV FKDQJH RYHU WLPH DQG DUH PRUH YDULDEOH WKDQ WKH SURSRUWLRQ SURGXFHG E\
WKH SULPDU\ FUXVKHU
x Availability of oxide ore 7KH DYDLODELOLW\ RI R[LGH RUH LV OLPLWHG E\ WKH RSHUDWLRQ RI WKH
PLQH $IWHU KHDY\ UDLQV LW FDQ EH WHPSRUDULO\ LPSRVVLEOH WR IHHG R[LGH RUH QRUPDOO\
x On/off operation of the pebble crushers 7KH LQVWDQWDQHRXV FDSDFLW\ RI DQ 03 SHEEOH
FUXVKHU H[FHHGV WKH W\SLFDO RSHUDWLQJ IHHG UDWH ZKLFK OHDGV WR RSHUDWLQJ WKH FUXVKHUV
GLVFRQWLQXRXVO\ 7KLV PDMRU LQFRQYHQLHQFH LV DWWHQXDWHG E\ FUXVKLQJ DV ÀQH DV
SRVVLEOH EXW LQ SUDFWLFH WKH RSHUDWLRQ UHPDLQV F\FOLF ZLWK RQH FUXVKHU RQ IRU WHQ
PLQXWHV DQG ERWK FUXVKHUV RII IRU ÀYH PLQXWHV DQG VR RQ
7KHVH GLVWXUEDQFHV PDNH LW GLIÀFXOW WR RSHUDWH WKH SODQW VDWLVIDFWRULO\ HLWKHU PDQXDOO\
RU ZLWK D VLPSOH FRQWURO ORJLF DQG WKXV MXVWLI\ WKH GHFLVLRQ WR LPSOHPHQW DQ $3&
V\VWHP

Components of the advanced process control system

$3& DW $KDIR LV EDVHG RQ 0HWVR 2&6 9LVLR5RFN DQG $XGLR0LOO WHFKQRORJ\ 7KH $3&
FRQVLVWV RI
x OCS software ² LPLWDWHV WKH GHFLVLRQPDNLQJ UHJLPHQ RI D V\VWHP·V EHVW RSHUDWRU RQ
KLVKHU EHVW GD\ EXW GRHV VR RQ D  EDVLV
x the VisioRock system ² HVWLPDWHV 6$* PLOO IHHG 36' DQG RWKHU SURSHUWLHV RI WKH IHHG
x the AudioMill wireless system ² PHDVXUHV DQG DQDO\VHV 6$* PLOO VRXQG
x an expert system ² ZLWK IX]]\ ORJLF DQG DQ RSWLPLVLQJ FRQWURO VWUDWHJ\ IRU WKH HQWLUH
circuit
x a soft sensor ² HVtiPDtHV NH\ SrRcHVV VtDtH YDriDEOHV tKDt cDQQRt EH PHDVurHG RQOiQH
7KiV VHOHctiRQ RI cRPSRQHQtV PDNHV it SRVViEOH tR KDYH D ZHOOiQIRrPHG $3& V\VtHP

Comminution Handbook | Spectrum Series 21 259

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

VisioRock, AudioMill and the soft sensor generate information that complements
data collected by conventional sensors. VisioRock provides information to the soft
sensor and directly to the expert system, while AudioMill and the soft sensor provide
information to the expert system.

VisioRock system
The main physical components of the VisioRock system (seen in Figure 15.11) are:
x a camera, which is installed above the SAG feed conveyor belt (before any addition
of recycled pebbles, lime or grinding media) and connected to a LAN (local area
network)
x a shielding and lighting system to protect the area of the conveyor observed by the
camera from the light of the day and to light it optimally
x DÀHOGFDELQHWORFDWHGQHDUWKHFDPHUDDUHD
VisioRock analyses several images of the belt per second, estimates particle size
distribution for each image and delivers an average estimate every 20 seconds. In the
case of Ahafo, VisioRock estimates 13 size classes from -6 mm to +225 mm and computes
special points of the PSD; that is, P20, P40, P60, P80 and P100 (top size) and proportions of
PPÀQHV²PPDQG!PPURFNV6RPHRIWKHVHGDWDDUHXVHGLQWKH
control logic and others are stored in the information system.

FIG 15.11 – VisioRock screen shot at Ahafo (image courtesy Metso).

AudioMill system
The physical components of the AudioMill system are:
x an industrially hardened microphone installed at a distance of a few tens of
centimetres from the mill shell, near the area of the toe of the charge
x a junction box with a local antenna and a source of power, to power the microphone
and emit the signal; mill sound is transmitted from the mill area to the control room
wirelessly

260 Comminution Handbook | Spectrum Series 21

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

x D rHcHStiRQ DQtHQQD DQG D rHcHiYHr iQ tKH cRQtrRO rRRP OiQNHG tR D VuitDEOH DuGiR
cDrG iQ tKH cRPSutHr
6SHDNHrSKRQHV iQ tKH cRQtrRO rRRP PDNH it SRVViEOH IRr RSHrDtRrV tR OiVtHQ tR tKH
RriJiQDO QRQÀOtHrHG QRiVH RI tKH PiOO IrRP tKH cRQtrRO rRRP $uGiR0iOO cRQtiQuRuVO\
PHDVurHV tKH VRuQG iQtHQVit\ IRr PDQ\ IrHTuHQciHV DQG cDOcuODtHV DQ iPSDct OHYHO
$YHrDJH YDOuHV DrH VHQt tR tKH cRQtrRO ORJic HYHr\ tHQ VHcRQGV ,Q RrGHr tR GHtHrPiQH tKH
iPSDct OHYHO $uGiR0iOO ¶OiVtHQV· tR KRZ tKH PDtHriDO tKrRZQ E\ HDcK OiItHr IDOOV ,I it IDOOV
RYHr tKH cKDrJH tKH VRuQG iV rDtKHr cRQtiQuRuV ,I it IDOOV RYHr tKH OiQHrV it IDOOV DV ZDYHV
² RQH ZDYH SHr OiItHr $t $KDIR tKHrH DrH  KiJK OiItHr EDrV DQG  ORZ OiItHr EDrV DQG
$uGiR0iOO GiVcriPiQDtHV EHtZHHQ tKH iPSDctV JHQHrDtHG E\ HDcK t\SH

Soft sensor in OCS

7KH VRIt VHQVRr iV DQ RQOiQH G\QDPic SKHQRPHQRORJicDO PRGHO RI tKH SrRcHVV IrRP tKH
DSrRQ IHHGHrV tR tKH c\cORQH RYHrÁRZV 3rHGictHG YDOuHV RI PiOO SRZHr PiOO PDVV DQG
tRQQDJH RI rHc\cOHG SHEEOHV DrH cDOcuODtHG uViQJ DQ ([tHQGHG .DOPDQ ÀOtHr DQG DrH
cRPSDrHG ZitK PHDVurHG YDOuHV &RrrHctiRQV RI PRGHO SDrDPHtHrV DrH tKHQ PDGH 7KH
VRIt VHQVRr SrRYiGHV HVtiPDtHV RI DOO VtrHDPV RI VROiGV DQG ZDtHr tKrRuJKRut tKH SrRcHVV
iQcOuGiQJ 36' DQG SHr cHQt VROiGV iQViGH tKH PiOOV 7KiV iV uVHG tR cRQtrRO tKH SHr cHQt
VROiGV iQViGH tKH EDOO PiOO Dt D YDOuH tKDt PD[iPiVHV JriQGiQJ HIÀciHQc\

Control strategy

General
7KH cRQtrRO VtrDtHJ\ uVHV iQIRrPDtiRQ JHQHrDtHG E\ cRQYHQtiRQDO iQVtruPHQtDtiRQ tKH
9iViR5RcN DQG $uGiR0iOO V\VtHPV DQG tKH 2&6 VRIt VHQVRr tR DQDO\VH tKH SrRcHVV
DQG ZritH VHt SRiQtV iQ tKH SrRJrDPPDEOH ORJic cRQtrROOHrV 3/&V HYHr\ tHQ VHcRQGV
7KH cRQtrRO VtrDtHJ\ iV cRQÀJurHG DV VHtV RI Iu]]\ ORJic DQG criVS ruOHV iQ tKH 2&6
H[SHrt V\VtHP PRGuOH 7KH VtrDtHJ\ cKHcNV tKH crHGiEiOit\ RI tKH GDtD rHcHiYHG IrRP
tKH VHQVRrV DQG DQDO\VHV tKH SrRcHVV tR GHtHct HPHrJHQc\ VituDtiRQV tKDt rHTuirH
iPPHGiDtH rHDctiRQV tR DYRiG PDMRr SrREOHPV Rr SODQt VtRSV VucK DV 6$* PiOO RYHrORDG
tKH SRZHr EHiQJ tRR KiJK IRr VRPH PRtRrV DQG tKH SDcNiQJ RI cOD\ PDtHriDO EHtZHHQ
tKH OiItHr EDrV 7KH rHVSRQVHV tR HPHrJHQc\ VituDtiRQV cRQViVt RI ViJQiÀcDQt cKDQJHV iQ
ZDtHr DGGitiRQV DQG iQ IHHG rDtH RI SriPDr\ RrH Rr R[iGH RrH
7KH HTuiYDOHQt SriPDr\ IHHG (3) rDtH iV DQ iPSRrtDQt cRQcHSt iQ tKH $3& VtrDtHJ\
Dt $KDIR ZKHrH it cRPEiQHV IHHGIRrZDrG DQG IHHGEDcN DctiRQV tR DQticiSDtH tKH
cRQVHTuHQcHV RI GiVturEDQcHV VucK DV cKDQJHV iQ DYDiODEiOit\ RI R[iGH RrH 7KH IHHG
rDtH RStiPiViQJ VtrDtHJ\ GRHV QRt PDQiSuODtH tKH SriPDr\ RrH IHHG rDtH GirHctO\ it
PDQiSuODtHV tKH (3) ZKicK iV DQ iQtHrQDO VHt SRiQt 7KH VHt SRiQt IRr tKH SriPDr\ RrH
IHHG rDtH iV cRPSutHG DV D IuQctiRQ RI tKH (3) VHt SRiQt tKH tRQQDJH RI rHc\cOHG SHEEOHV
tKH 36' RI SriPDr\ IHHG PHDVurHG E\ 9iViR5RcN DQG tKH tRQQDJH RI R[iGHV 7KH ORJic
tDNHV iQtR DccRuQt tKH QuPHrRuV GHDG tiPHV tKDt rHVuOt IrRP tKH OHQJtK RI tKH EHOtV )Rr
iQVtDQcH tKH GHDG tiPH EHtZHHQ tKH tKirG IHHGHr DQG 6$* iQOHt iV DrRuQG tKrHH PiQutHV
ZKicK iV ViJQiÀcDQt cRPSDrHG tR D IRurPiQutH PHDQ rHViGHQcH tiPH iQ tKH 6$* PiOO
7KiV iQtHrQDO VHt SRiQt KHOSV tR DGGrHVV GiVturEDQcHV VucK DV cKDQJH iQ IHHG 36' ODcN RI
DYDiODEiOit\ RI R[iGH RrH DQG GiVcRQtiQuRuV rHc\cOiQJ RI SHEEOHV

Comminution Handbook | Spectrum Series 21 261

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

Feeders control with VisioRock

,Q DQ 6$%& circuit DQ iQcrHDVH iQ tKH SrRSRrtiRQ RI ÀQH PDtHriDO iQ tKH IrHVK IHHG
iPPHGiDtHO\ PDNHV it SRVViEOH tR SrRcHVV D KiJKHr tRQQDJH SrRYiGHG tKH EDOO PiOOV cDQ
DEVRrE tKH H[trD tRQQDJH 7KiV iV ZK\ it iV iPSRrtDQt tR RStiPiVH DQG ZKHQ SRVViEOH
cRQtrRO IHHG 36' ,Q JHQHrDO KRZHYHr tKH EHQHÀtV RI PDQiSuODtiQJ IHHGHrV iQ 6$*
PiOOiQJ circuitV iV GHEDtDEOH ViQcH DOO PDtHriDO iQ tKH VtRcNSiOH KDV tR EH SrRcHVVHG VRRQHr
Rr ODtHr 6trDtHJiHV tKDt DiP tR iQcrHDVH tRQQDJH E\ PDQiSuODtiQJ IHHGHrV PD\ RQO\
JHQHrDtH D tHPSRrDr\ EHQHÀt DQG DrH QRt JHQHrDOO\ rHcRPPHQGHG
$t $KDIR KRZHYHr D VuEVHt RI tKH cRQtrRO VtrDtHJ\ GirHctO\ PDQiSuODtHV tKH RutSutV
RI tKH cRQtrROOHrV RI tKH IHHGHrV tKrRuJK 9iViR5RcN ,t ZDV IRuQG tKDt PDQiSuODtiQJ
IHHGHr QuPEHr  ZKicK tHQGV tR GHOiYHr cRDrVHr PDtHriDO KHOSV tR PD[iPiVH tKH DYHrDJH
tRQQDJH :KHQ 6$* PiOO PDVV iV KiJK 2&6 GHcrHDVHV tKH SrRGuctiRQ RI IHHGHr QuPEHr 
tR ORDG tKH EDOO PiOO circuit ZitKRut IurtKHr ORDGiQJ tKH 6$* PiOO DQG ZKHQ 6$* PiOO
PDVV iV ORZ 2&6 iQcrHDVHV tKH SrRGuctiRQ RI IHHGHr QuPEHr 
7KH DYDiODEiOit\ RI 9iViR5RcN iV DEVROutHO\ criticDO IRr iPSOHPHQtiQJ VucK D VtrDtHJ\
SrRSHrO\ 9iViR5RcN cRQtiQuRuVO\ cKHcNV tKDt D cKDQJH iQ tKH RutSut RI IHHGHr QuPEHr 
JHQHrDtHV tKH H[SHctHG cKDQJH iQ 36' :KHQ tKHrH iV QR cKDQJH tKH SRtHQtiDO ÁH[iEiOit\
rHVuOtiQJ IrRP VHJrHJDtiRQ iQ tKH VtRcNSiOH KDV EHHQ tHPSRrDriO\ H[KDuVtHG DQG IHHGHrV
cRPH EDcN tR D VtDQGDrG VHttiQJ IRr D GHÀQitH DPRuQt RI tiPH ,t iV DQ H[DPSOH RI D
VtrDtHJ\ tKDt PDQiSuODtHV tKH IHHGHrV ZitK tKH KHOS RI 9iViR5RcN DOtKRuJK tKHrH iV QRt D
9iViR5RcN cDPHrD RQ HDcK IHHGHr

Mill sound control with AudioMill

$t tKH HDrO\ VtDJH RI tKH $3& SrRMHct it ZDV iQtHQGHG tR cRQtrRO tKH 6$* PiOO E\
RSHrDtiQJ ZitKiQ D rDQJH RI iPSDct OHYHOV :KHQ $uGiR0iOO ZDV Sut iQ VHrYicH cORVH tR
tKH RutViGH RI tKH PiOO it iPPHGiDtHO\ GHPRQVtrDtHG itV DEiOit\ tR TuDQtiI\ tKH iPSDctV
RQ tKH OiQHrV DQG GiVcriPiQDtH EHtZHHQ iPSDctV rHVuOtiQJ IrRP PDtHriDO tKrRZQ RQtR
tKH OiQHrV E\ KiJK DQG ORZ OiItHr EDrV ,PSDctV cDQ EH cKDQJHG E\ YDr\iQJ PiOO VSHHG
Eut EHcDuVH tKH circuit iV rHODtiYHO\ RItHQ EDOO PiOO OiPitHG it iV QRt DOZD\V SRVViEOH
tR ORDG tKH 6$* PiOO DV PucK DV GHVirHG ZKicK PHDQV tKDt tKH PiOO VSHHG VKRuOG EH
YDriHG rHODtiYHO\ IrHTuHQtO\ 7KH rHVtrictiRQV iPSRVHG E\ tKH /56 PHDQ tKDt it iV QRt
SRVViEOH tR PDQiSuODtH PiOO VSHHG IrHTuHQtO\ DV D cRQVHTuHQcH tKH cRQcHSt iV RQO\
SDrtO\ iPSOHPHQtHG DOtKRuJK tKH $uGiR0iOO ViJQDOV DrH uVHG iQ cRQtrRO ORJic

Ball mill control

7KH EDOO PiOO cRQtrRO VtrDtHJ\ DiPV tR PD[iPiVH tRQQDJH ZKHQ tKH circuit iV EDOO PiOO
OiPitHG DQG tR RStiPiVH tKH 36' RI K\GrRc\cORQH RYHrÁRZ ZKHQ tKH circuit iV 6$* PiOO
OiPitHG ,t PDQiSuODtHV ZDtHr DGGitiRQV tR tKH VuPS DQG tR tKH EDOO PiOO DQG it cORVHV
DQG RSHQV K\GrRc\cORQHV ,t PDQiSuODtHV tKH (3) iQtHrQDO VHt SRiQt ZKHQ tKH circuit
iV EDOO PiOO OiPitHG *riQGiQJ HIÀciHQc\ iV cRQtiQuRuVO\ PD[iPiVHG E\ cRQtrROOiQJ tKH
VOurr\ VROiGV GHQVit\ iQViGH tKH EDOO PiOO (PSiricDOO\ tKH RStiPuP VOurr\ VROiGV GHQVit\
iQ tKH PiOO iV HVtiPDtHG tR EH DrRuQG  SHr cHQt PRVt RI tKH tiPH 3ODQt PHtDOOurJiVtV
cDQ PRGiI\ tKH VOurr\ VROiGV GHQVit\ VHt SRiQt iI RrH t\SH cKDQJHV 7KH DctuDO VOurr\
VROiGV GHQVit\ iQViGH tKH PiOO iV cRQtiQuRuVO\ HVtiPDtHG E\ tKH 2&6 VRIt VHQVRr DQG
ZDtHr DGGitiRQ tR tKH PiOO iV PDQiSuODtHG E\ tKH 2&6 H[SHrt V\VtHP tR rHDcK tKH tDrJHt
7KH DEiOit\ tR PD[iPiVH JriQGiQJ HIÀciHQc\ tKrRuJK JRRG cRQtrRO RI tKH SHr cHQt VROiGV

262 Comminution Handbook | Spectrum Series 21

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

iQViGH tKH PiOO iV RQH RI tKH EHQHÀtV RI D VRIt VHQVRr 7KH VOurr\ VROiGV GHQVit\ IHHGiQJ
tKH K\GrRc\cORQHV iV cRQtrROOHG Dt D À[HG VHt SRiQt ZKHQ tKH circuit iV EDOO PiOO OiPitHG
DQG iV DutRPDticDOO\ GHcrHDVHG ZKHQ tKH circuit iV 6$* PiOO OiPitHG iQ RrGHr tR DcKiHYH
D VtHHSHr DQG ÀQHr cut ZitK tKH K\GrRc\cORQHV

Project implementation
7KH SrRMHct ZDV VuccHVVIuO EHcDuVH RI tKH VtrRQJ iQYROYHPHQt E\ VtDII Dt 1HZPRQt DQG
0HtVR 1HZPRQt·V DctiYH iQYROYHPHQt cRYHrHG DVSHctV DV GiYHrVH DV iQVtDOOiQJ YiViRQ
DQG DcRuVticV KDrGZDrH PDNiQJ VurH NH\ 1HZPRQt rHVRurcHV ZHrH DYDiODEOH Dt tKH riJKt
tiPH cRQtriEutiQJ tR tKH GHViJQ RI tKH cRQtrRO VtrDtHJ\ DOORZiQJ rHPRtH DccHVV tR tKH
2&6 cRPSutHr tKrRuJK tKH VHcurit\ V\VtHP DQG HQcRurDJiQJ RSHrDtRrV tR uQGHrVtDQG
DQG uOtiPDtHO\ truVt tKH 2&6 V\VtHP
$Q iQitiDO cRQtrRO VtrDtHJ\ ZDV GHÀQHG MRiQtO\ E\ 1HZPRQt VtDII DQG 0HtVR VSHciDOiVtV
DQG tKHQ cRQÀJurHG iQ 2&6 ,Q SDrDOOHO 0HtVR cRQÀJurHG DQG tuQHG tKH VRIt VHQVRr
7KH cRQtrRO VtrDtHJ\ ZDV tKHQ tHVtHG DQG rHÀQHG EDVHG RQ SODQt REVHrYDtiRQV DQG tKH
V\VtHP ZDV RSHrDtiRQDO OHVV tKDQ Vi[ PRQtKV DItHr ViJQDturH RI tKH cRQtrDct 5HPRtH
DccHVV ZDV HVtDEOiVKHG HDrO\ iQ tKH SrRMHct 7KH cROODERrDtiRQ ViJQiÀcDQtO\ cRQtriEutHG tR
tKH JRRG DccHStDQcH RI tKH $3& V\VtHP E\ tKH RSHrDtRrV

System performance
$ \HDr DItHr tKH $3& V\VtHP ZDV Sut iQ RSHrDtiRQ itV DYDiODEiOit\ iV QHDr  SHr cHQt IRr
H[DPSOH EHtZHHQ 'HcHPEHr  DQG 0D\  2&6 cRQtrROOHG tKH SODQt  SHr cHQt
RI tKH tiPH 7KH YHr\ IHZ tiPHV tKDt tKH $3& V\VtHP ZDV QRt DYDiODEOH ZDV cDuVHG E\
SrREOHPV ZitK tKH 23& VHrYHr OicHQcH iQGHSHQGHQt RI 2&6 DQG ZKHQ 9iViR5RcN ZDV
tHPSRrDriO\ uQDYDiODEOH DItHr D SrREOHP ZitK tKH SrHVVurH DQG TuDOit\ RI tKH cRPSrHVVHG
Dir IRr tKH Dir EDrriHr tKDt NHHSV tKH cDPHrD ZiQGRZ cOHDQ SHrPDQHQtO\
1HZPRQt·V HYDOuDtiRQ iV tKDt tKH $3& V\VtHP DORQH JHQHrDtHV D  SHr cHQt iQcrHDVH
iQ tRQQDJH SOuV D  SHr cHQt iQcrHDVH iQ rHcRYHr\ 7KH tZR IDctRrV cRPEiQHG JiYH D
 SHr cHQt iQcrHDVH iQ RuQcHV

Present status of the advanced process control system

7KH $3& V\VtHP Dt 1HZPRQt $KDIR iV ZHOO DccHStHG E\ tKH RSHrDtRrV DQG SODQt
PHtDOOurJiVtV DQG it SrRYiGHV EHQHÀtV DV H[SHctHG 2&6 cRQtrROV tKH 6$%& circuit
DSSrR[iPDtHO\  SHr cHQt RI tKH tiPH DQG tKiV circuit ruQV  SHr cHQt RI tKH tiPH Dt Rr
YHr\ QHDr D SK\VicDO cRQVtrDiQt rHÁHctiQJ tKDt tKH $3& V\VtHP IuOO\ uVHV tKH JriQGiQJ
cDSDcit\ 9iViR5RcN HVtiPDtHV tKH 36' RI 6$* PiOO IrHVK IHHG 7KiV iQIRrPDtiRQ iV uVHG
E\ tKH $3& tR DQticiSDtH tKH RSSRrtuQitiHV tR PRGiI\ tRQQDJH ZKHQ IHHG Vi]H YDriHV
,t iV DOVR uVHG E\ D cRQtrRO ORJic tR PDQiSuODtH tKH DSrRQ IHHGHrV DQG tDNH DGYDQtDJH RI
VHJrHJDtiRQ iQ tKH VtRcNSiOH tR ORDG tKH 6$* PiOO DQG EDOO PiOO circuitV JiYiQJ RStiPuP
SHrIRrPDQcH
$uGiR0iOO GHtHctV iPSDctV RQ OiQHrV DQG GiVcriPiQDtHV RQ tKH HIIHctV RI tKH KiJK DQG
ORZ OiItHr EDrV 7KH VRIt VHQVRr NHHSV tKH H[SHrt V\VtHP iQIRrPHG RI PDtHriDO PDVV DQG
36' tKrRuJKRut tKH SrRcHVV ,t PDNHV it SRVViEOH tR cRQtrRO SHr cHQt VROiGV iQViGH tKH EDOO
PiOO Dt D YDOuH tKDt PD[iPiVHV JriQGiQJ HIÀciHQc\

Comminution Handbook | Spectrum Series 21 263

 CHAPTER 15 t CASE STUDIES OF CONTROL SYSTEMS

BIBLIOGRAPHY

Phu Kham case study

Baas, ' +iOOH 6 DQG .DrDJHRrJRV -  ,PSrRYHG ÁRtDtiRQ SrRcHVV cRQtrRO Dt 1HZcrHVt·V 7HOIHr
2SHrDtiRQ iQ Proceedings Ninth Mill Operators’ Conference, SS ² 7KH $uVtrDODViDQ ,QVtitutH
RI 0iQiQJ DQG 0HtDOOurJ\ 0HOERurQH
Bennett, ' &rQNRYic , DQG :DONHr 3  5HcHQt SrRcHVV GHYHORSPHQtV Dt tKH 3Ku .KDP
cRSSHrJROG cRQcHQtrDtRr iQ Proceedings 11th Mill Operators’ Conference, SS ² 7KH
$uVtrDODViDQ ,QVtitutH RI 0iQiQJ DQG 0HtDOOurJ\ 0HOERurQH
Bullemer, 3 5HiViQJ ' %urQV & +DMGuNiHZic] - DQG $QGr]HMHZVNi -  Effective Operator
Display Design: ASM Consortium Guideline 2008 &rHDtH6SDcH ,QGHSHQGHQt 3uEOiVKiQJ
3ODtIRrP
Crnkovic, , *HRrJiHY 7 +DrERrt * DQG 3KiOOiSV 0  &RPPiVViRQiQJ DQG RStiPiVDtiRQ
RI tKH 3Ku .KDP cRSSHrJROG cRQcHQtrDtRr iQ Proceedings Tenth Mill Operators’ Conference,
SS ² 7KH $uVtrDODViDQ ,QVtitutH RI 0iQiQJ DQG 0HtDOOurJ\ 0HOERurQH
International Organization for Standardization,  ,(&  (QtHrSriVHcRQtrRO V\VtHP
iQtHJrDtiRQ ² 3Drt  0RGHOV DQG tHrPiQRORJ\ HGitiRQ 
Runge, . 0c0DVtHr - /D 5RVD ' :RrtOH\ 0 DQG *u\Rt 2  $ cRrrHODtiRQ EHtZHHQ
9iViR)rRtKŒ PHDVurHPHQtV DQG tKH SHrIRrPDQcH RI D ÁRtDtiRQ cHOO iQ Proceedings Ninth Mill
Operators’ Conference, SS ² 7KH $uVtrDODViDQ ,QVtitutH RI 0iQiQJ DQG 0HtDOOurJ\ 0HOERurQH
Zhu, <  0XOWLYDULDEOH6\VWHP,GHQWLÀFDWLRQIRU3URFHVV&RQWURO S  (OVHYiHr 6ciHQcH 2[IRrG

Ahafo case study

Broussaud, $ /HJrDQG * .RN ' 5Ru[ ( *u\Rt 2 DQG 5HYDORr 0  ,QtHJrDtHG DGYDQcHG
JriQGiQJ cRQtrRO V\VtHP Dt 1HZPRQt $KDIR iQ SAG2011, Metsco Proceedings 2012 &DQDGiDQ
,QVtitutH RI 0iQiQJ 0HtDOOurJ\ DQG 3HtrROHuP 0RQtrHDO
GhanaWeb,  $YDiODEOH IrRP KttSZZZJKDQDZHEcRP! >$ccHVVHG  0D\ @

264 Comminution Handbook | Spectrum Series 21

 HOME

Chapter 16

CIRCUIT DESIGN
Bianca Foggiatto, Marcos Bueno,
Eddie McLean, Greg Lane and Chris Morley

PROJECT DEFINITION
Determining the project concept precedes all other tasks in the design of a comminution
circuit. Factors to be taken into account include the:
x location of the orebody
x nature of the ore (geological and mineral characteristics)
x local environmental restrictions
x SURGXFWVSHFLÀFDWLRQVZKHUHWKHVHLQÁXHQFHSURFHVVLQJVWUDWHJ\
x ZDWHUDQGHQHUJ\DYDLODELOLW\DQGFRVWV
7KHSURFHVVWRREWDLQWKHÀQDOSURGXFWPXVWEHHVWDEOLVKHGRIZKLFKFRPPLQXWLRQLV
DQLPSRUWDQWSDUW$OWHUQDWLYHVIRUFRPPLQXWLRQFLUFXLWGHVLJQDUHLQWHJUDOWRÁRZVKHHW
development.

PROCESS DEVELOPMENT AND CIRCUIT DESIGN

7KHUHODWLRQVKLSEHWZHHQWHFKQLFDOGHYHORSPHQWRIWKHSURFHVVOHIWWRULJKWDQGFLUFXLW
GHVLJQWRSWRERWWRPLVVXPPDULVHGLQ)LJXUH
Risk is inherent in every circuit design and process development should focus on
managing risk through:
x WHVWZRUN
x HPSLULFDORUWKHRUHWLFDOVFDOHXSIRUXQLWSURFHVVHV
x experience or benchmarking
x conservative design
x modelling/simulations
x expenditure of additional capital.

Comminution Handbook | Spectrum Series 21 265

266 Comminution Handbook | Spectrum Series 21

CHAPTER 16 t CIRCUIT DESIGN

FIG 16.1 – Circuit design logic flow: technical (left-to-right) and design (top-to-bottom) (image courtesy Ausenco).
 CHAPTER 16 t CIRCUIT DESIGN

SAMPLE SELECTION AND TEST WORK

7KH WHVW ZRUN SURJUDP PXVW LQFOXGH RUH FKDUDFWHULVDWLRQ WHVWV DQG LQVSHFWLRQV RI
JHRORJLFDOVHFWLRQVWRGHWHUPLQHUHODWLRQVKLSVEHWZHHQWKHYDULDELOLWLHVRIRUHFRPSHWHQF\
DQG KDUGQHVV OLWKRORJLHV URFN W\SHV DQG PLQHUDOLVDWLRQ 6DPSOLQJ PXVW EH FDUHIXOO\
SODQQHG7KHQXPEHUDQGW\SHRIVDPSOHVGHSHQGRQWKHVL]HDQGVWDWXVRIWKHSURMHFW
SURMHFWULVNSURÀOHFRQWLQJHQFLHVLQGHÀQLQJHTXLSPHQWVL]HVDQGFDSLWDODQGRSHUDWLQJ
FRVWV7KHH[WHQWRIWKHWHVWZRUNSURJUDPQHHGVWREHEDODQFHGDJDLQVWWKHVFKHGXOH
WKHWHVWZRUNFRVWVDQGDGHWHUPLQDWLRQRIWKHWHFKQLFDOULVNV7\SLFDOO\WKHWHVWZRUN
program is divided into three phases:
x 3KDVH  ² VFRSLQJ DQG SUHIHDVLELOLW\ VWXGLHV LQ ZKLFK VDPSOHV DUH VHOHFWHG RQ WKH
EDVLVRIPDMRUNQRZQDQGFKDUDFWHULVWLFURFNW\SHV
x 3KDVH²DIHDVLELOLW\VWXG\LQZKLFKWKHIRFXVLVRQYDULDELOLWLHVDFURVVWKHRUHERG\
7KURXJKSXWDQGRSHUDWLQJFRVWVDUHIRUHFDVWPRUHDFFXUDWHO\ZLWKDUHODWLYHO\KLJKHU
number of samples.
x 3KDVH  ²DIHDVLELOLW\ DQGGHVLJQVWXG\ IRU ZKLFK VDPSOHV DUHFRPSRVLWHG RQ WKH
EDVLVRIVSDWLDOUHSUHVHQWDWLYLW\PLQHSHULRGSURGXFWLRQRUKDUGZLUHGRUHW\SHV
7DEOHVKRZVWKHUHFRPPHQGHGQXPEHURIVDPSOHVIRUWHVWZRUNLQDPRGHUDWHWR
ODUJHSURMHFWDQGZLWKPRGHUDWHJHRORJLFDODQGPLQHUDOLVDWLRQFRPSOH[LW\
6HYHUDOPHWKRGVDUHDYDLODEOHIRUWHVWLQJEUHDNDJHFKDUDFWHULVWLFVRIZKLFKWKRVHXVHG
ZLGHO\ZHUHGHVFULEHGLQ&KDSWHU7KHIROORZLQJLVWKHJHQHUDOVXLWHRIWHVWVXVHGIRU
comminution circuit design in various combinations depending on the objectives and
DVVHVVPHQWRIULVN%RQG:RUN,QGH[WHVWVFUXVKLQJ²&:LURGPLOO²5:LEDOOPLOO
²%:L60&7HVW®-.':7-.'URS:HLJKW7HVWDQG63,®6$*3RZHU,QGH[WHVW
ZKHUH6$*UHIHUVWRVHPLDXWRJHQRXVJULQGLQJ0DF3KHUVRQ·VDXWRJHQRXVPLOOLQJWHVW
data is not typically used unless the data is historical.

TABLE 16.1
Required number of samples during feasibility studies (modified from Lane, Foggiatto and Bueno, 2013).
Basis Number of samples Comments
Spatial representativity >100 Includes samples from Phase 1 and 2
Hard-wired ore types 10 to 30 Includes samples from Phase 1 and 2
Mine period production 5 to 8 Includes samples from Phase 3

DESIGN CRITERIA
In the early stages of project development only limited data are available and
FDOFXODWLRQVRIHTXLSPHQWVL]HXVHWKHPD[LPXPRUDYHUDJHRIWKHUDQJH7UHQGVLQRUH
hardness are assessed by making incremental increases in data values and repeating
WKHFDOFXODWLRQV&RPSDUDEOHLQGXVWU\GDWDRULQKRXVHGDWDIURPVLPLODURUHW\SHVDUH
XVHGIRUEHQFKPDUNLQJ$VWKHSURMHFWGHYHORSVWKHYROXPHDQGFRPSOH[LW\RIGDWD
WKDW GHVFULEH RUH FKDUDFWHULVWLFV DQG YDULDELOLW\ LQFUHDVH DQG SODQW IHHG VFKHGXOHV ²
ZKLFKLQFOXGHYDU\LQJWRQQDJHJHRWHFKQLFDOGDWDJUDGHOLWKRORJ\DQGPLQHUDOLVDWLRQ
DQG RUH H[SORLWDWLRQ VHTXHQFH ² FDQ EH FRQVLGHUHG ,QSXW GDWD DQG LQIRUPDWLRQ IRU
VWXGLHVIURPFRQFHSWWRÀQDOIHDVLELOLW\FRPHIURPVHYHUDOVRXUFHVDQGFDQEHJURXSHG
DVIROORZV

Comminution Handbook | Spectrum Series 21 267

 CHAPTER 16 t CIRCUIT DESIGN

x 3URMHFWRSHUDWLQJDQGSURGXFWLRQSDUDPHWHUVZKLFKDUHW\SLFDOO\VHWE\WKHSURMHFW
RZQHU7KHVHIRUPSDUWRIWKHGHVLJQEDVLVVXFKDVRSHUDWLQJVFKHGXOHDQGDYDLODELOLWLHV
x 7HVW ZRUN E\ ODERUDWRULHV YHQGRUV DQG FRQVXOWDQWV UHVRXUFH PRGHOV DQG
FRQVXOWDQWV·UHSRUWV7KHVHDUHWKHPDLQVRXUFHVRIGDWDIRUWKHSURMHFWDVWKH\FRQWDLQ
GHWDLOHGWHVWZRUNDQGDQDO\VHVIRUWKHUHVRXUFHLWVJHRORJ\JHRWHFKQLFDOGDWDPLQH
SODQDOOXQLWSURFHVVHVWKHZDWHUTXDOLW\DQGHQYLURQPHQWDOFRQWUROV
x 2SHUDWLQJSUDFWLFHDQGLQGXVWU\VWDQGDUGV6RXUFHVRIGDWDDUHLQKRXVHH[SHULHQFH
RSHUDWLQJKLVWRU\RQVLPLODUWUHDWPHQWSODQWVDQGSXEOLVKHGRUFRQÀUPHGRSHUDWLQJ
information.
x 9HQGRUGDWDKDQGERRNDQGUHJXODWRU\DQGHQYLURQPHQWDOVWDQGDUGV
x ,QKRXVHGHVLJQDQGRSHUDWLQJDQGFRPPLVVLRQLQJGDWDEDVHV

Purpose
The design criteria value:
x LVDQLQGLYLGXDOUDWHRUXQLWXVHGIRUVL]LQJHTXLSPHQW
x is attainable at the stated operating conditions
x H[FOXGHV DGGLWLRQDO GHVLJQ DOORZDQFHV E\ WKH HQJLQHHU RU YHQGRU WR HQVXUH
attainment.
The design criteria values are used for four main purposes:
 input data for process calculations
2. LQSXWGDWDIRUPDVVZDWHUDQGHQHUJ\EDODQFHV
3. XQLWGHVLJQSDUDPHWHUVIRUHTXLSPHQWVL]LQJ
 data essential for design or to prevent ambiguity.

Function
7KH VFKHPDWLF ÁRZ FKDUW RI DFWLYLWLHV LQ )LJXUH  LOOXVWUDWHV WKH SLYRWDO UROH RI WKH
design criteria in collating data from several sources and channelling this data for mass
EDODQFHZRUNDQGIRURWKHUFDOFXODWLRQV'HULYHGYDOXHVIURPPDVVEDODQFHVFDOFXODWLRQV
DQGHTXLSPHQWVL]LQJDUHQRWLQFOXGHGLQWKHGHVLJQFULWHULDH[FHSWWRSURYLGHFODULW\
and avoid ambiguity.

Codes
7KHGHVLJQFULWHULDDQGYDOXHVDUHSURMHFWVSHFLÀFDQGFRPHIURPVHYHUDOVRXUFHV7KHGDWD
DQGUHIHUHQFHVDUHFRGHGWRUHDGLO\LGHQWLI\WKHLUVRXUFHW\SLFDOO\DVVKRZQLQ7DEOH

FIG 16.2 – Schematic logic showing role of design criteria (image courtesy Ausenco).

268 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

TABLE 16.2
Coded sources of data for design criteria (provided by Ausenco).
Code Source of data or information for design criteria
1 Data provided by client (client or owner documents that nominate project parameters; and written instructions or formal
responses that contain decisions for the project)
2 Test work data
3 Reports and supporting studies (consultants’ reports; and analyses and studies, the outcomes from which are validated
data and defined parameters)
4 Standard industry practice, acceptable operating practice
5 Vendor data that nominates specific information relating to the design
6 Engineering handbook, textbook data
7 Regulatory standards, codes and compliance – operational and environmental, nominated in project documents or
governed by conditions of contract
8 Recommendation, standard procedures or in-house data from engineer
9 Not available – to be provided from test work and/or others when available; assumed value, to be confirmed by client

Comminution design criteria

An example of the process design criteria required for design of a comminution circuit,
in this case a crushing and SAG milling circuit, is presented in Appendix 16.1.

FLOW SHEET OPTIONS

There are many options for comminution circuits. The more common circuits involve
autogenous, semi-autogenous, ball and rod milling applications, as well as the use of
crushing and high pressure rolls on feed and internal recycle streams. Primary crushing
(and secondary crushing in the case of three crushing stages) are common to the circuits
described and are excluded from these Áow sheets for clarity. A regrind circuit treating
a Áotation concentrate stream is provided.
The basic guidelines for selecting the appropriate Áow sheet are presented as follows

Stage crush, rod and ball mill circuit

(2CRB or 3CRB) (3CRB means
3 crushers, 1 rod mill, 1 ball mill)
This circuit has low-tonnage applications
and has become less common since SAG
milling increased its applicability.

Stage crush and ball mill circuit

(2CB or 3CB)
This circuit has low-tonnage applications
and has become less common since SAG
milling increased its applicability.

Comminution Handbook | Spectrum Series 21 269

 CHAPTER 16 t CIRCUIT DESIGN

Single stage SAG (or AG) mill circuit

(SS-SAG)
This is a low-capital circuit applicable
to low-to-moderate throughput and
moderate ore competency. Autogenous
milling typically requires piloting.

SAG and ball mill circuit (SAB)

This circuit is typically selected for
plants of moderate ore competency and
hardness.

Stage crush, SAG and ball mill circuit

(2C-SAB)
This circuit is usually selected for plants
treating harder ores.

SAG and ball mill with pebble crusher

circuit (SABC)
This circuit is typically selected for plants
above 2 Mt/a treating moderate-to-high
competency ores.

AG and ball mill with pebble crusher

circuit (ABC)
Fully autogenous grinding has the
potential to reduce operating costs.

AG and pebble mill circuit (AP)

Fully autogenous grinding has the
potential to provide the lowest operating
costs. Success depends on control of feed
size and ore competency. May require
pebble crushing to control pebble volume.

270 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

Stage crush, HPGR and ball mill circuit

(2C-HPGRB)
This circuit is considered for large
throughput plants treating competent
ores to improve energy efÀciency
(when compared with SABC circuits).

Stage crush, HPGR, SAG and ball mill

with pebble crusher circuit
(HPGR – SABC)
This circuit has been used for large
throughput plants treating competent ores
to debottleneck the SAG milling circuit.

Regrind
This circuit is used to grind the rougher
concentrate Àner to enhance mineral
liberation and improve grades in
cleaner Áotation stages. Ball mills can
be followed by tertiary milling using
vertical mills.

&ircuit selection may be inÁuenced by plant capacity and ore competency, as shown
in Table 16.3.
2ther factors considered as part of circuit selection include nature of the ore, test
work and sample requirements (for example, pilot trials if required), client preference,
location and terrain (layout issues), local and international practice based on scale of
operations and ore characteristics, energy cost and circuit energy efÀciency.

TABLE 16.3
Selection of circuit type by ore competency and throughput (adapted from Lane, Foggiatto and Bueno, 2013).
Throughput High competency Moderate competency Low competency
<0.5 Mt/a 2CB or 3CB 1C-SAG or 2C-SAG or 2CB or 3CB SS-SAG
0.5–2 Mt/a 2CB or 3CB or SABC SAB SS-SAG or SAB
1–5 Mt/a SABC SAB or ABC SS-SAG or SAB
5–10 Mt/a SABC SAB or ABC SAB
>10 Mt/a SABC or 2C-HPGRB SABC SAB
2CB – 2 crushers, 1 ball mill SAG – semi-autogenous grinding
SAB – SAG and ball mill circuit SABC – SAG and ball mill with pebble crusher circuit
ABC – AG and ball mill with pebble crusher circuit HPGR – high-pressure grinding rolls
SS – single stage

Comminution Handbook | Spectrum Series 21 271

 CHAPTER 16 t CIRCUIT DESIGN

COMMINUTION ENERGY CALCULATIONS

7KHSDUDPHWHUWKDWPXVWEHNQRZQIRUFRPPLQXWLRQFLUFXLWGHVLJQLVWKHHQHUJ\UHTXLUHG
SHUWRQQHWRJULQGWKHPDWHULDOIURPDNQRZQIHHGVL]HWRDNQRZQSURGXFWVL]HVSHFLÀF
HQHUJ\ 6HYHUDO WHFKQLTXHV DUH XVHG WR SUHGLFW WKLV YDOXH WKH\ DUH DOO EDVHG RQ WKH
ZRUNRI)UHG%RQGGLVFXVVHGLQ&KDSWHUVDQG%RQGGHYHORSHGWKHUHODWLRQVKLSV
EHWZHHQRUHKDUGQHVVWRQQDJHSURFHVVHGVL]HUHGXFWLRQDFKLHYHGDQGSRZHUUHTXLUHG
DQG%RQG·VHTXDWLRQVDUHWKHEDVLVRIPLOOVHOHFWLRQWRGD\DOWKRXJKPRGLÀFDWLRQVKDYH
EHHQPDGHWRKDQGOHWKHODUJHLQFUHDVHVWKDWKDYHRFFXUUHGLQPLOOVL]HVIHHGVL]HVDQG
feed rates.
%RQG·V HTXDWLRQV ZHUH HPSLULFDO EDVHG RQ GDWD DYDLODEOH IURP JULQGLQJ PLOOV XVHG
LQWKHV7KHFDOFXODWLRQVJLYHQLQWKLVVHFWLRQDUHUHOHYDQWWRWKHPXFKODUJHUPLOOV
XVHG  \HDUV ODWHU DQG WKH\ VKRZ KRZ %RQG·V UHODWLRQVKLSV KDYH EHHQ PRGLÀHG WR
FRQIRUPZLWKDSSOLFDWLRQVZKHUHPDFKLQHVDUHRSHUDWHGRXWVLGHWKHUDQJHRIHIÀFLHQF\
FRQVLGHUHGE\%RQG([DPSOHVDUHWKHHIÀFLHQF\IDFWRUVIRUURGDQGEDOOPLOOVSURSRVHG
E\5RZODQGDQG%DUUDWW·VIDFWRUIRU6$*PLOOV%DUUDWWDQG$OODQ2WKHU
PHWKRGVIRUSUHGLFWLQJSRZHUDQGSRZHUVSOLWVLQFRPPLQXWLRQFLUFXLWVIHDWXULQJ$*
6$*PLOOVDUHGHVFULEHGLQ&KDSWHU

Ausgrind worked examples

7KHIROORZLQJZRUNHGH[DPSOHVZHUHGHYHORSHGXVLQJHPSLULFDOIRUPXODHDYDLODEOHLQ
$XVJULQG$XVHQFR·VLQKRXVHFRPPLQXWLRQSURJUDP/DQH)RJJLDWWRDQG%XHQR
7KHEDVLFVWUXFWXUHLVVLPLODUWRWKDWGHYHORSHGE\%DUUDWWEXWKDVEHHQIXUWKHU
HQKDQFHGWRHQDEOHWKHXVHRIWKH-XOLXV.UXWWVFKQLWW0LQHUDO5HVHDUFK&HQWUH-.05&
-.'URS:HLJKW7HVWSDUDPHWHUV63,WHVWDQGRWKHUPHDVXUHVRIRUHFRPSHWHQF\7KH
H[DPSOHVSUHVHQWHGKHUHDUHEDVHGRQWKH-.'URS:HLJKW7HVWDQG%RQG:RUN,QGH[
test results. Input parameters are:
x %RQG&UXVKHU:RUN,QGH[ N:KW
x %RQG5RG0LOO:RUN,QGH[ N:KW
x %RQG%DOO0LOO:RUN,QGH[ N:KW
x 'URS:HLJKW7HVW$ðE 
7KH$XVJULQGPHWKRGFDOFXODWHVWKHWRWDOVSHFLÀFHQHUJ\(FVIRUWKHJULQGLQJFLUFXLW
DVDSURGXFWRIWKHFDOFXODWHG%RQG(FVDQGWKHFLUFXLWHIÀFLHQF\IDFWRUVFLUFXLWHIÀFLHQF\
IDFWRUfSAGDQG6$*IHHGVL]HHIIHFWIDFWRU F80 ²UHIHUWR&KDSWHUIRUGHWDLOV
effect
7KH$XVJULQGWRWDO(FVFDOFXODWLRQLVGHVFULEHGE\WKHIROORZLQJHTXDWLRQ

Total Ecs = 9]Bond Ecs to 150 nmg # a fSAG - F80effect kC ! 7Bond Ecs to final P80A

7KHRYHUDOOVSHFLÀFJULQGLQJHQHUJ\WRUHGXFHDSULPDU\FUXVKHUSURGXFWZLWKP of
PPWRDÀQDOSURGXFWPRI—PLVHVWLPDWHGIRUWKUHHFLUFXLWFRQÀJXUDWLRQVLQ
WKHIROORZLQJH[DPSOHV

SABC circuit
%RQGSRZHUFDOFXODWLRQVDUHFRQGXFWHGIRUHDFK:RUN,QGH[
x &UXVKLQJVSHFLÀFHQHUJ\IURPFRIPPWRPRIPP

272 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

10 - 10 10 10
-
C Ecs = CWi # e P80 F80 o = 15.0 # c 10 000 100 000
m = 1.0 kWh/t

x 5RGPLOOVSHFLÀFHQHUJ\IURPFRIPPWRP of 2 mm
10 - 10 10 - 10
R Ecs = RWi # e P80 F80 o = 16.5 # c 2000 10 000
m = 2.0 kWh/t

x %DOOPLOOVSHFLÀFHQHUJ\IURPF of 2 mm to PRIPP
10 - 10 10 - 10
B Ecs = BWi # e P80 F80 o = 15.1 # c 150 2000
m = 9.0 kWh/t

x %DOOPLOOVSHFLÀFHQHUJ\IURPFRIPPWRÀQDOPRIPP
10 - 10 10 - 10
B Ecs = BWi # e P80 F80 o = 15.1 # c 106 150
m = 2.3 kWh/t

Circuit efficiency factor

The fSAGUHODWLRQVKLSVDUHRUHGHSHQGHQWDQGVSHFLÀFWRWKHVHOHFWHGFLUFXLWFRQÀJXUDWLRQ
7KH\PD\DOVREHDGMXVWHGZKHQVHFRQGDU\FUXVKLQJRULQWHQVLYHEODVWLQJSUDFWLFHVDUH
LQSODFHWKDWLVIRUÀQHUFLUFXLWIHHGVL]HVF).
)RUWKLV6$%&FLUFXLWfSAG DQG F80 is not applicable in this case as there are no
effect
secondary/tertiary crushing stages.

Total grinding specific energy

Total Ecs = 7]Bond Ecs to 150 nmg # fSAGA + Bond Ecs to final P80
= ]1.0 + 2.0 + 9.0g # ]1.32g + 2.3 = 18.1 kWh/t
7KH6$*PLOODQGEDOOPLOOVSHFLÀFHQHUJLHVDUHVXEVHTXHQWO\FDOFXODWHG)RUWKH6$*
PLOODSURSULHWDU\HTXDWLRQEDVHGRQWKH':LYDOXHLVXWLOLVHG)LJXUHDQGDVHULHV
RIIDFWRUVDUHWKHQDSSOLHGWRDGMXVWWKHEDVHFDVH6$*(FVIRUWKHRSHUDWLQJFRQGLWLRQV
DQGPLOOGLPHQVLRQVIRUH[DPSOHDVSHFWUDWLRDQGRWKHUV²UHIHUWR)LJXUHVDQG
LQ&KDSWHU)RUWKHEDOOPLOO(FVLVFDOFXODWHGDVWKHGLIIHUHQFHEHWZHHQWRWDO
(FVDQG6$*(FV

FIG 16.3 – SAG Ecs relationship with ore hardness – DWi (image courtesy Ausenco).

Comminution Handbook | Spectrum Series 21 273

 CHAPTER 16 t CIRCUIT DESIGN

SAG mill specific energy

Base case SAG Ecs = f ]DWig = 8.1 kWh/t

The main variables affecting the adjusting factors selected for the Ausgrind base case
DUHDVIROORZV
x $VSHFWUDWLR ńfaspect ratio 
7KHDVSHFWUDWLRIDFWRULVVHWWRDQGLQFUHDVHVDVWKHDVSHFWUDWLRLQFUHDVHV
x %DOOORDG WHQSHUFHQWńfball load 
The ball load factor changes as a function of ball load. Higher ball loads result in
ORZHU6$*PLOOVSHFLÀFHQHUJ\
x Feed size (F PPń fF 80
The feed size factor is set at unity as a function of ore competency. It changes
EDVHGRQPLQHEODVWGHVLJQJHRWHFKQLFDOIDFWRUVSUHFUXVKHUVHWWLQJVDQGRWKHU
DVSHFWVRIGHVLJQWKDWLPSDFWRQ6$*PLOOIHHGVL]H*HQHUDOO\DPPUHGXFWLRQ
in FUHVXOWVLQDWKUHHWRÀYHSHUFHQWUHGXFWLRQLQ6$*PLOOVSHFLÀFHQHUJ\
x 3HEEOHUHF\FOH SHUFHQWńfpebble recycle 
The pebble recycle factor is set to unity for a standard pebble recycle condition.
7KHVWDQGDUGFRQGLWLRQLVDIXQFWLRQRIRUHFRPSHWHQF\DQG6$*PLOOEDOOORDG
,QWKLVFDVHDOOWKHFRUUHFWLRQIDFWRUVDUHHTXDOWRRQHEHFDXVHVWDQGDUGFRQGLWLRQVZHUH
VHOHFWHGKRZHYHULIWKHDIRUHPHQWLRQHGYDOXHVGHVFULEHGYDU\IURPWKHVHOHFWHGUDQJH
WKHVH IDFWRUV ZRXOG YDU\ DFFRUGLQJ WR )LJXUHV  DQG  SUHVHQWHG LQ &KDSWHU 
7KHZRUNHGH[DPSOHVWKDWIROORZVKRZKRZWKHVHYDOXHVPD\YDU\

SAG mill ECS = 5Base Case SAG Ecs? # 6adjusting factors@ = 8.1 # 1.0 = 8.1 kWh/t

Ball mill specific energy

,QWKLVFDVHWKHEDOOPLOOVSHFLÀFHQHUJ\LVHTXDOWRWKHGLIIHUHQFHEHWZHHQWRWDODQG6$*
PLOOVSHFLÀFHQHUJ\&UXVKHUFDOFXODWLRQVDUHFRQGXFWHGVHSDUDWHO\

Ball mill ECS = Total Ecs - SAG mill Ecs = 18.1 - 8.1 = 10.0 kWh/t

High-pressure grinding rolls and ball mill circuit

,QLWLDOO\%RQGSRZHUFDOFXODWLRQVDUHFRQGXFWHGIRUHDFK:RUN,QGH[LQDVLPLODUPDQQHU
DVWKH6$%&FLUFXLWKRZHYHUIRUWKLVFLUFXLWFRQÀJXUDWLRQF is the secondary crushed
VL]H VHOHFWHG DV  PP 3ULPDU\ DQG VHFRQGDU\ FUXVKHU FDOFXODWLRQV DUH FRQGXFWHG
separately:
x &UXVKLQJVSHFLÀFHQHUJ\IURPFRIPPWRPRIPP
10 - 10 10 10
-
C Ecs = CWi # e P80 F80 o = 15.0 # c 10 000 50 000
m = 0.8 kWh/t

x 5RGPLOOVSHFLÀFHQHUJ\IURPFRIPPWRP of 2 mm

R Ecs = 2.0 kWh/t

274 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

x %DOOPLOOVSHFLÀFHQHUJ\IURPF of 2 mm to PRIPP

B Ecs = 9.0 kWh/t

x %DOOPLOOVSHFLÀFHQHUJ\IURPFRIPPWRÀQDOPRIPP

B Ecs = 2.3 kWh/t

Circuit efficiency factor

)RUWKLV+3*5%FLUFXLWfSAG DQG F80 LVQRWDSSOLFDEOHLQWKLVFDVHDVWKHUHLVQR
effect
6$*PLOOLQWKHFLUFXLW

Total grinding specific energy

Total Ecs = 7]Bond Ecs to 150 nmg # fSAGA ! 7Bond Ecs to final P80A
= ]0.8 + 2.0 + 9.0g # ]0.95g + 2.3 = 13.5 kWh/t

HPGR specific energy

$VVXPLQJWKDWWKH+3*5IHHGLVQRWWUXQFDWHGWKDWLVWKH+3*5IHHGLVQRWSUHVFUHHQHG
DQG WKDW WKH +3*5 FLUFXLW LV FORVHG E\ VFUHHQV ZLWK D  SHU FHQW UHF\FOH WKH +3*5
VSHFLÀFHQHUJ\LVFDOFXODWHG5DQJHVIRUUHF\FOHVLQ$XVJULQGDUHWHQWRSHUFHQWIRU
RSHQFLUFXLWVDQGWRSHUFHQWZKHQFORVLQJVFUHHQVDUHLQVWDOOHG

HPGR Ecs = 2.5 kWh/t

Ball mill specific energy

,Q WKLV FDVH WKH EDOO PLOO VSHFLÀF HQHUJ\ LV HTXDO WR WKH GLIIHUHQFH EHWZHHQ WRWDO DQG
+3*5VSHFLÀFHQHUJ\

Ball mill Ecs = Total Ecs - HPGR Ecs = 13.5 - 2.5 = 11.0 kWh/t

Conventional stage crush and ball mill circuit

,QLWLDOO\ %RQG SRZHU FDOFXODWLRQV DUH FRQGXFWHG IRU HDFK :RUN ,QGH[ LQ D VLPLODU
PDQQHUWRWKH6$%&FLUFXLW)RUWKLVFLUFXLWFRQÀJXUDWLRQKRZHYHUF is the tertiary
FUXVKHGVL]HVHOHFWHGDVPP3ULPDU\VHFRQGDU\DQGWHUWLDU\FUXVKHUFDOFXODWLRQV
are conducted separately:
x &UXVKLQJVSHFLÀFHQHUJ\IURPFRIPPWRPRIPP

10 - 10 10 10
-
C Ecs = CWi # e P80 F80 o = 15.0 # c 10 000 20 000
m = 0.4 kWh/t

x 5RGPLOOVSHFLÀFHQHUJ\IURPFRIPPWRP of 2 mm

R Ecs = 2.0 kWh/t

x %DOOPLOOVSHFLÀFHQHUJ\IURPF of 2 mm to PRIPP

B Ecs = 9.0 kWh/t

Comminution Handbook | Spectrum Series 21 275

 CHAPTER 16 t CIRCUIT DESIGN

x %DOOPLOOVSHFLÀFHQHUJ\IURPFRIPPWRÀQDOPRIPP
B Ecs = 2.3 kWh/t

Circuit efficiency factor

)RUWKLV&%FLUFXLWfSAG DQG F80 LVQRWDSSOLFDEOHLQWKLVFDVHDVWKHUHLVQR6$*
effect
mill in the circuit.

Total grinding specific energy

Total Ecs = 7]Bond Ecs to 150 nmg # fSAGA ! 7Bond Ecs to final P80A
= ]0.4 + 2.0 + 9.0g # ]1.11g + 2.3 = 15.0 kWh/t

,Q WKLV FDVH WKH EDOO PLOO VSHFLÀF HQHUJ\ LV HTXDO WR WKH FRUUHFWHG JULQGLQJ VSHFLÀF
HQHUJ\7KHWKUHHFUXVKHUV·VSHFLÀFHQHUJLHVDUHFDOFXODWHGVHSDUDWHO\

SMC worked examples

7KHVHZRUNHGH[DPSOHVZHUHSUHVHQWHGE\0RUUHOO7KH\DUHEDVHGRQWKH60&
7HVWDQG%RQG%DOO0LOO:RUN,QGH[WHVWUHVXOWV7KHIROORZLQJUHVXOWVZHUHREWDLQHG

SMC Test Bond Ball Mill Work Index test

0LD N:KW FDUULHGRXWZLWKD—PFORVLQJVFUHHQ
0LF N:KW 0LE N:KW
0LK N:KW

7KUHH FLUFXLWV DUH HYDOXDWHG 6$%& +3*5EDOO PLOO DQG FRQYHQWLRQDO VWDJH FUXVK
IROORZHGE\EDOOPLOO7KHRYHUDOOVSHFLÀFJULQGLQJHQHUJ\WRUHGXFHDSULPDU\FUXVKHU
SURGXFWDVWKHIHHGWRJULQGLQJZLWKDPRIPPWRDÀQDOSURGXFWPRIPP
LVGHWHUPLQHGDVIROORZV
7KHIROORZLQJJHQHUDOVL]HUHGXFWLRQHTXDWLRQSURSRVHGE\0RUUHOOLVXWLOLVHGWR
FDOFXODWHWKHVSHFLÀFHQHUJ\
x i x
Wi = 4 # Mi # ` x2f_ 2 - x1f_ 1ij

ZKHUH
Wi LVWKHVSHFLÀFHQHUJ\N:KW
x and x2 DUHWKHSHUFHQWSDVVLQJVL]HIRUWKHSURGXFWDQGIHHG—PUHVSHFWLYHO\
Ki is a constant for open and closed circuits

SABC circuit
Coarse particle tumbling mill specific energy
Wa = 4 # 0.95 # 19.4 # `750-]0.295 + 750/1000 000g - 100 000-]0.295 + 100000/1000 000gj = 9.6 kWh/t

Fine particle tumbling mill specific energy

Wb = 4 # 1.0 # 18.0 # `106-]0.295 + 106/1000 000g - 750-]0.295 + 750/1000 000gj = 8.4 kWh/t

276 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

Pebble crusher specific energy

In this circuit it is assumed that the pebble crusher feed P LV  PP $V D UXOH RI
WKXPEWKLVYDOXHFDQEHHVWLPDWHGE\DVVXPLQJWKDWLWLVRIWKHQRPLQDOSHEEOHSRUW
DSHUWXUH,QWKLVFDVHWKHSHEEOHSRUWDSHUWXUHLVPP7KHSHEEOHFUXVKHULVVHWWRJLYH
a product PRIPP

Wc = 4 # 1.19 # 7.2 # `12 000-]0.295 + 12 000/1000 000g - 52 500-]0.295 + 52 500/1000 000gj = 1.12 kWh/t

7KH SHEEOH FUXVKHU IHHG UDWH LV H[SHFWHG WR EH  SHU FHQW RI QHZ IHHG WRQQDJH
,IH[SUHVVHGLQWHUPVRIWKH6$%&FLUFXLWQHZIHHGWKHQ

Wc = 1.12 # 0.25 = 0.3 kWh/t

Total net comminution specific energy

WT = / Wi = 18.3 kWh/t

High-pressure grinding rolls and ball mill circuit

,QWKLVFLUFXLWSULPDU\FUXVKHUSURGXFWLVUHGXFHGWRDQ+3*5FLUFXLWIHHGFRIPP
E\FORVHGFLUFXLWVHFRQGDU\FUXVKLQJ7KH+3*5LVDOVRLQFORVHGFLUFXLWDQGUHGXFHVWKH
PPIHHGWRDFLUFXLWSURGXFWPRIPP7KLVLVWKHQIHGWRDFORVHGFLUFXLWEDOOPLOO
that reduces the grind to a PRIPP

Secondary crushing specific energy

W c= #
4 1.0 # 7.2 # ^55 # ]35 000 # 100 000g-0.2h
# ]35 000 -]0.295 + 35000/1000 000g
g - 100 000-]0.295 + 100 000/1000 000g = 0.4 kWh/t

High-pressure grinding rolls specific energy

Wc = 4 # 1.0 # 13.9 # ^35 # ]4000 # 35 000g-0.2h
# ]4000 g - 35 000-]0.295 + 35000/1000 000g = 2.4 kWh/t
-]0.295 + 4000/1000 000g

Coarse particle tumbling mill specific energy

Wa = 4 # 1.0 # 19.4 # ]750-]0.295 + 750/1000 000g - 4000-]0.295 + 4000/1000 000gg = 4.5 kWh/t

Fine particle tumbling mill specific energy

Wb = 4 # 1.0 # 18.8 # ]106-]0.295 + 106/1000 000g - 750-]0.295 + 750/1000 000gg = 8.4 kWh/t

Total net comminution specific energy

WT = / Wi = 15.7 kWh/t

Comminution Handbook | Spectrum Series 21 277

 CHAPTER 16 t CIRCUIT DESIGN

Conventional stage crush and ball mill circuit

,QWKLVFLUFXLWSULPDU\FUXVKHUSURGXFWLVUHGXFHGWRDPLOOIHHGVL]HPRIPPYLDD
VHFRQGDU\WHUWLDU\FUXVKLQJFLUFXLWFORVHG7KLVLVWKHQIHGWRDFORVHGFLUFXLWEDOOPLOO
that grinds to a PRIPP

Secondary/tertiary crushing specific energy

Wc = 4 # 1.0 # 7.2 # ]6500-]0.295 + 6500/1000 000g - 100 000-]0.295 + 100 000/1000 000gg = 1.7 kWh/t

Coarse particle tumbling mill specific energy

Wa = 4 # 1.0 # 19.4 # ]750-]0.295 + 750/1000 000g - 6500-]0.295 + 6500/1000 000gg = 5.5 kWh/t

Fine particle tumbling mill specific energy

Wb = 4 # 1.0 # 18.8 # ]106-]0.295 + 106/1000 000g - 750-]0.295 + 750/1000 000gg = 8.4 kWh/t

Tumbling mill size distribution correction

,PSOLFLWLQWKH60&DSSURDFKLVWKDWWKHIHHGDQGSURGXFWVL]HGLVWULEXWLRQVDUHSDUDOOHO
DQG OLQHDU LQ ORJORJ VSDFH KRZHYHU WKLV LV QRW WUXH ZKHQ FORVHG VHFRQGDU\WHUWLDU\
FUXVKLQJLVIROORZHGE\EDOOPLOOLQJLQZKLFKFDVHWKHVSHFLÀFHQHUJ\FRUUHFWLRQIRUVL]H
distribution (WsQHHGVWREHPDGH,WLVDVVXPHGWKDWWKHDGGLWLRQDOVSHFLÀFHQHUJ\IRU
EDOOPLOOLQJLVHTXDOWRWKHGLIIHUHQFHEHWZHHQRSHQDQGFORVHGFLUFXLWFUXVKLQJVSHFLÀF
energies to account for the nominated ball mill feed size.
x i x
Ws = 4 # 0.19 # Mic # a x2f_ 2 - x f_ 1ik
1
Ws = 4 # 0.19 # Mic # ]6500-]0.295 + 6500/1000 000g - 100 000 -]0.295 + 100000/1000 000g
g = 0.9 kWh/t

Total net comminution specific energy

WT = / Wi = 16.5 kWh/t

EQUIPMENT SIZING AND SELECTION

Crushing and screening equipment

Crushers
7KHVHOHFWLRQRIFUXVKHUVLVLQÁXHQFHGE\URFNFKDUDFWHULVWLFVDQGFDSDFLW\7KHPDLQ
URFNFKDUDFWHULVWLFVWKDWDUHWDNHQLQWRDFFRXQWDUHIHHGWRSVL]HPD[LPXPHQGSURGXFW
VL]HKDUGQHVVDEUDVLYHQHVVSUHVHQFHRIFOD\PLQHUDOVDQGPRLVWXUH
$FFRUGLQJ WR FUXVKHU YHQGRUV WKH FUXVKHG SURGXFW FKDUDFWHULVWLFV GHSHQG RQ WKH
combined impact of the rock characteristics and the crusher operational parameters.
7DEOH  VKRZV D EDVLF JXLGHOLQH IRU FUXVKHU DSSOLFDWLRQV SURSRVHG E\ 0HWVR
0LQHUDOV

278 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

TABLE 16.4
Guidelines for crusher application (adapted from Metso Minerals, 2008).
Major crusher type Typical process Maximum Typical maximum Typical Amount
stage feed size end product size capacities of fines
(mm) (mm) (t/h) produced
Gyratory crusher Primary 1500 200–300 Over 1200 Low
Jaw crusher Primary 1400 200–300 Up to 1600a Low
Horizontal impact crusher Primary/secondary 1300 200–300 Up to 1800 Medium/high
Cone crusher Secondary 450 60–80 Up to 1200 Low
Cone crusher Tertiary 300 10–30 Up to 1000 Low/medium
Vertical-shaft impactors Secondary/tertiary 150 5–20 Up to 600 High
a. When installed with a preceding grizzly to remove fines.

7KHÀUVWVWHSLQVHOHFWLQJDFUXVKHULVWRGHÀQHWKHDVUHFHLYHGRUHWRSVL]HUHODWHGWR
RUHSURSHUWLHVDQGEODVWLQJDQGWKHUHTXLUHGFUXVKHUJDSHIHHGRSHQLQJ7KHJDSHLV
related to the maximum particle size in the feed to the crusher by:

Gape = Max. particle size/ ]0.8 to 0.9g

7KHQH[WVWHSLVWRGHÀQHWKHGHVLUHGSURGXFWFKDUDFWHULVWLFV%DVHGRQWKHUHGXFWLRQ
ratio and feed size F WKH GHVLUHG SURGXFW P is calculated. Reduction ratios for
FUXVKLQJDUHORZDQGYDU\DFFRUGLQJWRWKHFUXVKLQJVWDJH7KHIROORZLQJUDQJHVDUH
FRPPRQO\XVHGDGDSWHGIURP&KDYHVDQG3HUHV0XFNHUPDQQ
x SULPDU\FUXVKLQJ²W\SLFDOO\DURXQGEXWXSWRLQVSHFLÀFFDVHV
x VHFRQGDU\FUXVKLQJ²W\SLFDOO\EXWXSWRZKHQSURFHVVLQJVPDOOWKURXJKSXWV
x WHUWLDU\FUXVKLQJ²EHWZHHQDQG
x TXDWHUQDU\FUXVKLQJ²
$VFUXVKHUFDSDFLW\GHSHQGVRQRSHUDWLRQDOFRQGLWLRQVLWLVQHFHVVDU\WRGHÀQHWKH
FORVHG VLGH VHWWLQJ &66 9HQGRUV· WDEOHV DUH XVHG IRU WKLV SXUSRVH 7KH\ LQGLFDWH
FUXVKHG SURGXFW VL]H GLVWULEXWLRQV REWDLQHG IRU GLIIHUHQW &66 RIWHQ FDOOHG JUDGDWLRQ
curves). The crusher selection is based on capacity tables provided by vendors. These
FDSDFLWLHVDUHXVXDOO\IRUPDWHULDOVZLWKDEXONGHQVLW\RIWP3. The effective crusher
FDSDFLW\LVFDOFXODWHGE\FRUUHFWLQJWKHVWDQGDUGYDOXHV&RUUHFWLRQIDFWRUVIRUGHQVLW\
KDUGQHVV &UXVKLQJ :RUN ,QGH[ IHHG VL]H GLVWULEXWLRQ DQG PRLVWXUH DUH QRUPDOO\
provided by vendors according to crusher type.

Screens
7KH VFUHHQ W\SH LV VHOHFWHG DFFRUGLQJ WR WKH GXW\ LW ZLOO SHUIRUP $ JXLGHOLQH LV
SUHVHQWHGLQ7DEOH7KHVHOHFWLRQRIDVFUHHQVL]HLVEDVHGRQWKHUHTXLUHGVFUHHQLQJ
DUHD DQG WKH EHG WKLFNQHVV FRPSDWLEOH ZLWK WKH VFUHHQ DSHUWXUH WR JXDUDQWHH WKH
PDWHULDOVWUDWLÀFDWLRQ$OWKRXJKLWYDULHVZLWKWKHRUHGHQVLW\WKHPD[LPXPIHHGVL]H
YDOXHLVXVXDOO\IRXUWLPHVWKHVFUHHQDSHUWXUH,QWKHOLWHUDWXUHVHYHUDOHTXDWLRQVIRU
VFUHHQDUHDFDOFXODWLRQWKDWZHUHGHYHORSHGE\VFUHHQYHQGRUVPD\EHIRXQGWKHVHDUH

1. Assume a prescreened feed. Adjustment is required for direct feed with no prescreening.

Comminution Handbook | Spectrum Series 21 279

 CHAPTER 16 t CIRCUIT DESIGN

TABLE 16.5
Guidelines for screen application (Metso Minerals, 2008).
Service Maximum Separation Applicable screentype Capacity
feed size size range
(mm) (mm) (m3/h)
Large material intermediate classification 400 100–200 Primary screens 300–1500
Medium material classification 250 50–100 Inclined and banana screens 100–1300
Fine material classification 200 2–50 Horizontal, inclined and banana screens 50–400
Extra-fine classification 25 0.2–6 High-frequency screens 10–40
Dewatering 13 0.5 Counter-slope screen with linear motion 100–250

FRQVLGHUHGFRQVHUYDWLYH9HQGRUVVXFKDV0HWVR0LQHUDOVDQG7HOVPLWK

KDYH SXEOLVKHG WKHLU IRUPXODH ZKLFK DUH DYDLODEOH LQ WKHLU KDQGERRNV ,Q JHQHUDO
these formulae are similar because the screen capacity is corrected by a similar range
RIIDFWRUV,QVSHFLÀFDSSOLFDWLRQVKRZHYHUYHQGRUVPD\KDYHSURSULHWDU\HTXDWLRQV
7KHVL]LQJWHFKQLTXHSUHVHQWHGKHUHLVEDVHGRQWKHXQGHUÁRZWKURXJKSXW0HWVR
0LQHUDOV7KHVFUHHQDUHDLVJHQHULFDOO\FDOFXODWHGXVLQJWKHIROORZLQJHTXDWLRQ
Qu
Area
Qspec

Qspec Qbasic # CFs

ZKHUH
Qu is the amount of undersize in the feed (t/h)
Qspec LVWKHVSHFLÀFVFUHHQLQJFDSDFLW\WKP2)
Qbasic is the basic capacity (t/h/m2 GHWHUPLQHG JUDSKLFDOO\ EDVHG RQ WKH GHVLUHG
separation size
CFs are correction factors
6RPHH[DPSOHVRIIDFWRUVFRQVLGHUHGZKHQFDOFXODWLQJVFUHHQLQJDUHDVDUHSHUFHQW
RI RYHUVL]H IUDFWLRQV SHU FHQW RI PDWHULDO LQ WKH IHHG VPDOOHU WKDQ WKH VHSDUDWLRQ
VL]H GHFN SRVLWLRQ ZHW RU GU\ VFUHHQLQJ VROLGV GHQVLW\ VXUIDFH RSHQ DUHD VKDSH
RI VXUIDFH RSHQLQJ SDUWLFOH VKDSH VHSDUDWLRQ HIÀFLHQF\ VFUHHQ W\SH DQG PRLVWXUH
content.
7KLV WHFKQLTXH LPSOLHV WKDW WKH DFWLYH VFUHHQ DUHD LV LQLWLDOO\ FDOFXODWHG DQG WKH
VFUHHQ VL]H LV WKHQ FKRVHQ EDVHG RQ PLQLPXP VFUHHQ ZLGWK IRU DQ DGHTXDWH IHHG
OD\HU WKLFNQHVV 7KH VFUHHQ ZLGWK LV FDOFXODWHG XVLQJ WKH IROORZLQJ HTXDWLRQ 0HWVR
0LQHUDOV
Q
B= + 0.15
3.6 # v # d
ZKHUH
Q LVRYHUVL]HYROXPHWULFÁRZUDWHP3/h)
v is the material transport speed (m/s)
d LVWKHPDWHULDOOD\HUWKLFNQHVVPPZKLFKLVGHWHUPLQHGJUDSKLFDOO\DFFRUGLQJ
WRWKHYHQGRU·VUHFRPPHQGDWLRQV

280 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

7KH PDWHULDO WUDQVSRUW VSHHG LV XVXDOO\ GHÀQHG E\ YHQGRUV DFFRUGLQJ WR WKH VFUHHQ
W\SHDQGLWVVORSH&KDYHVDQG3HUHVVXJJHVWHGWKDWWKHPD[LPXPPDWHULDOOD\HU
WKLFNQHVVLVSURSRUWLRQDOWRWKHDSSDUHQWGHQVLW\DVSUHVHQWHGLQ7DEOH

TABLE 16.6
Maximum material layer thicknesses (Chaves and Peres, 2009).
Apparent density (t/m3) Maximum layer thickness (mm)
>1.6 4 × screen aperture
1.6–0.8 3 × screen aperture
<0.8 2.5 × screen aperture

Grinding and classification equipment

Autogenous and semi-autogenous grinding mills
7KHWHFKQLFDODQGHFRQRPLFGULYHUVWKDWVKRXOGEHREVHUYHGZKHQVL]LQJDQGVHOHFWLQJ
DQ$*6$*PLOODUHOLVWHGLQRUGHURIVLJQLÀFDQFH
 UHTXLUHGSLQLRQVKHOODQGPRWRUSRZHU
2. mill dimensions
3. SRZHUEDODQFHEHWZHHQ$*6$*PLOODQGEDOOPLOO
 motor drive technology
 plant layout.
7KH UHTXLUHG JULQGLQJ SRZHU LV REWDLQHG E\ PXOWLSO\LQJ WKH FDOFXODWHG VSHFLÀF
HQHUJ\ (FV E\ WKH PLOO WKURXJKSXW ZKLOH WKH PLOO VKHOO GLPHQVLRQV DUH DGMXVWHG WR
VXLWWKHGHVLJQRSHUDWLQJFRQGLWLRQV7KHUHDUHHPSLULFDOPRGHOVIRUH[DPSOH$XVWLQ
/RYHGD\IRUFDOFXODWLQJWKHPRWRUSRZHUGUDZIRUDJLYHQPLOOVKHOODQG
RSHUDWLQJFRQGLWLRQ7KH0RUUHOOSRZHUPRGHOLVZLGHO\DFFHSWHGDQGXVHGIRU
WKLV SXUSRVH ,W LV UHFRPPHQGHG WKDW D IDFWRU RI  WR  SHU FHQW RYHU WKH UHTXLUHG
SLQLRQVKHOOSRZHUEHDSSOLHGZKHQVHOHFWLQJWKH6$*PLOOPRWRULQRUGHUWRDOORZIRU
WKHSRWHQWLDORSHUDWLQJHQYHORSHRIWKHPLOOKRZHYHUWKHPDJQLWXGHRIWKLVDOORZDQFH
GHSHQGVRQWKHPHWKRGXVHGWRGHWHUPLQHWKHVSHFLÀFHQHUJ\UHTXLUHPHQWDQGWKHULVN
DVVRFLDWHGZLWKWKHFDOFXODWLRQ
7RDFKLHYHGHVLJQWKURXJKSXWODUJH6$*PLOOVLQ6$%&FLUFXLWVW\SLFDOO\FRQVXPHÀYH
WRWHQSHUFHQWPRUHSRZHURQDQDYHUDJHPRQWKO\EDVLVWKDQLQGLFDWHGE\WKH¶GHVLJQ·
FRQGLWLRQ7KHW\SLFDOSDUDPHWHUVDQGRSHUDWLQJFRQGLWLRQVDGRSWHGZKHQGHVLJQLQJD
0:6$*PLOODUHVKRZLQ7DEOH7KHEDODQFHEHWZHHQ6$*PLOODQGEDOOPLOO
VSHFLÀFHQHUJ\FDQEHPDQLSXODWHGE\FKDQJLQJWKHPLOODVSHFWUDWLRIHHGVL]HPLOOORDG
DQGRUEDOOORDG7KLVLVXVXDOO\GRQHLQSURMHFWVZKHUHWKLVGLIIHUHQFHLVUHODWLYHO\VPDOO
DQGVRPHHFRQRPLFEHQHÀWVDUHDFKLHYHGE\PDWFKLQJPLOOPRWRUVL]HV
7UDGLWLRQDOO\ WKH PLOO GULYH RI FKRLFH ZDV WKH SLQLRQ DQG JLUWK JHDU ,W KDV EHHQ
SUHIHUUHG RYHU WKH JHDUOHVV PLOO GULYH *0' GXH WR LWV ORZHU FDSLWDO FRVW HDVH RI
LQVWDOODWLRQ VKRUWHU FRQVWUXFWLRQ WLPH DQG VLPSOLFLW\ WR RSHUDWH DQG PDLQWDLQ 0LOO
YHQGRUVZLOOVXSSO\XSWR0:WZLQSLQLRQGULYHVZLWKVRPHKDYLQJ0:GHVLJQV
LQSODFH7KHFXUUHQWWKLQNLQJLVWKDW0:LVWKHPD[LPXPDFFHSWDEOHSRZHUDJHDU
GULYHQPLOOFDQDFFRPPRGDWHWKURXJKDWZLQSLQLRQGULYHDUUDQJHPHQW*0'VDUHD
SURYHQGHVLJQVROXWLRQIRUJULQGLQJDSSOLFDWLRQVIRUODUJHSRZHUHG6$*0:DQG

Comminution Handbook | Spectrum Series 21 281

 CHAPTER 16 t CIRCUIT DESIGN

TABLE 16.7
Typical parameters and operating conditions for designing a 22 MW semi-autogenous (SAG) mill.
Mill data Value Unit Comments
Diameter (inside shell) 12.2 m 40 ft
Liner thickness 0.15 m Suggest 0.10–0.15 m
Aspect ratio (L/D) 0.55 - Typically 0.5 to 0.6 for a high aspect mill
Effective grinding length (EGL) 7.5 m 22 ft
Feed cone angle 15 degree Typically 15
Discharge cone angle 15 degree Typically 15
Trunnion diameter 2.7 m
Fraction of critical speed 0.74 - Suggest 0.72 to 0.78 (0.80 maximum)
Ball volume nominal design 12 % v/v Suggest 10–15% for design
Ball volume maximum operating 16 % v/v Typically 0–20%
Ball volume structural design 20 % v/v Suggest 2–5% above maximum operating
Total filling nominal design 26 % v/v Suggest 26% for SAG, 28% for AG
Total filling maximum operating 35 % v/v Suggest 30–35%
Discharge slurry pulp density 70 % w/w Suggest 65% to 78%, typically 70%
Discharge mechanism Grate - Grate discharge for SAG or AG mill
Discharge screen Trommel - Trommel may be required for large flows and/or pebble cleaning

EDOO0:PLOOVLQWKHPLQHUDOSURFHVVLQJLQGXVWU\*0'VKDYHEHHQWKHSUHIHUUHG

GULYHVROXWLRQIRUODUJHGLDPHWHUDQGKLJKSRZHUHGPLOOVGXHWRWKHOLPLWDWLRQVRIJHDU
driven technology.
7KHVHOHFWLRQRIPLOOVL]HVGULYHWHFKQRORJ\DQGQXPEHURIPLOOVKDVDGLUHFWLPSDFW
LQWKHSODQWOD\RXWKRZHYHUWKHUHDUHSURMHFWVZKHUHDSDUWLFXODUOD\RXWLVGHVLUDEOH
and can drive the mill selection. A discussion on plant layout is presented in the plant
layout section.
*UDWH GHVLJQ DQG WKHLU DSHUWXUHV DOVR KDYH D GLUHFW LQÁXHQFH RQ 6$* PLOO
SHUIRUPDQFH /DUJH JUDWH RSHQ DUHDV DQG DSHUWXUHV DUH XVXDOO\ VHOHFWHG WR SURPRWH
WKHGLVFKDUJHRISHEEOHVIRUSHEEOHFUXVKLQJ2QWKHRWKHUKDQGLQRSHUDWLRQVZKHUH
LWLVGHVLUHGWRSURPRWHÀQHJULQGLQJWKHJUDWHRSHQLQJVDUHUHVWULFWHGWKXVOLPLWLQJ
WKHURFNRXWÁRZ

Rod and ball mills

%RQG·VPHWKRGLVZLGHO\XVHGIRUURGDQGEDOOPLOOVL]LQJ$VWKHWHVWFRQGLWLRQVGLIIHU
IURP WKH LQGXVWULDO RSHUDWLQJ FRQGLWLRQV HIÀFLHQF\ IDFWRUV DUH XWLOLVHG WR FRUUHFW WKH
VSHFLÀFHQHUJ\7KHVHIDFWRUVZHUHGHVFULEHGLQ&KDSWHU%DVLFDVVXPSWLRQVWKDWQHHG
WREHPDGHEHIRUHVL]LQJWKHPLOODUHGHVFULEHGLQWKHJXLGHOLQHVLQ7DEOH
7KHFRUUHFWHGVSHFLÀFHQHUJ\LVFDOFXODWHGDQGJULQGLQJFRQGLWLRQVDUHDVVXPHGEDVHG
RQWKHGHÀQHGJULQGLQJFLUFXLWWKURXJKSXW7KHUHTXLUHGJULQGLQJSRZHULVREWDLQHGE\
PXOWLSO\LQJWKHFRUUHFWHGVSHFLÀFHQHUJ\E\WKHPLOOWKURXJKSXW
0LOOVHOHFWLRQLVEDVHGRQYHQGRUV·WDEOHVWKDWUHODWHPLOOGLDPHWHUDQGSRZHU8VXDOO\
FKDUWHGSRZHUYDOXHVDQGFDOFXODWHGSRZHUDUHGLIIHUHQWDQGWKHUHIRUHQHHGFRUUHFWLQJ

282 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

TABLE 16.8
Guidelines for rod / ball mill application (modified from Chaves and Peres, 2009).
Parameter Rod mills Ball mills
Length/diameter (L/D) ratio 1.4 to 1.6 1 to 2
Diameter (m/ft) <3.81/12.5 <8.53/28
Length (m/ft) >6.1/20 See note
Mill filling (% v/v) 35–40 28–45
Rotation speed (% of critical speed) 65–75 65–80
Mill power (kW) <2000 <22 000
Note: Ball mill length is normally limited by structural issues. The largest ball mill currently in operation (2015) has a diameter of 8.5 m (28 ft) and length of
13.4 m (44 ft).

0LOOOHQJWKLVLQFUHDVHGSURSRUWLRQDOO\WRWKHSRZHUGLIIHUHQFHWRDFKLHYHWKHUHTXLUHG
SRZHUXVLQJWKHIROORZLQJHTXDWLRQ
Pcorrected
Lcorrected # Ltable
Ptable

(TXDWLRQVUHODWLQJURGDQGEDOOPLOOSRZHUGUDZZLWKGLDPHWHUPLOOÀOOLQJDQGURWDWLRQ
VSHHGZHUHSURSRVHGE\5RZODQGDQG.MRV

Rod mills P = 1.752D0.34 _ 6.3 - 5.4Vp i f cs kW/t of rods charged

0.1
Ball mills P = 4.879D0.3 _ 3.2 - 3Vp i f cs b 1 - l S kW/t of balls charged
2]9 - 10f csg + s
ZKHUH
P LVWKHSRZHUGUDZSHUPHWULFWRQRIPHGLDN:W
D is the mill diameter (m)
Vp LVWKHIUDFWLRQRIPLOOYROXPHÀOOHGZLWKJULQGLQJPHGLD
fcs is the fraction of critical speed
Ss is a correction factor (Ss is used only if the mill diameter is larger than 3.3 m and if
the maximum ball diameter (BPPLVVPDOOHUWKDQD 
]B - 12.5Dg
Ss = 1.103
50.8

5RGPLOORSHUDWLRQFDQEHFODVVLÀHGDVÀQHFUXVKLQJRUFRDUVHJULQGLQJDVLWSURGXFHVD
relatively coarse product. Rod mills have limits on their shell length as rods tend to bend
if too long. The discharge of rod mill product occurs through openings in a discharge
WUXQQLRQRUWKURXJKSRUWVLQWKHSHULSKHU\RIWKHPLOOVKHOODVLOOXVWUDWHGLQ)LJXUH
The types of discharge are:
x FHQWUHSHULSKHUDOGLVFKDUJHZKLFKQRUPDOO\JLYHVPLQLPXPÀQHVFRQWHQW
x HQGSHULSKHUDOGLVFKDUJHZKLFKLVPDLQO\IRUGU\FRDUVHJULQGLQJ
x RYHUÁRZ GLVFKDUJH ZKLFK LV WKH PRVW ZLGHO\ XVHG DQG XVXDOO\ ZLWK D WURPPHO
screen attached to the trunnion.
%DOO PLOOV DUH DOVR FODVVLÀHG DFFRUGLQJ WR WZR W\SHV RI GLVFKDUJH PHFKDQLVP JUDWH
GLVFKDUJH RU RYHUÁRZ GLVFKDUJH $ PLOO ÀWWHG ZLWK WKH GLVFKDUJH JUDWH EHWZHHQ

Comminution Handbook | Spectrum Series 21 283

 CHAPTER 16 t CIRCUIT DESIGN

A B C
FIG 16.4 – Rod and ball mill discharge arrangements (image courtesy Metso).
(A) Centre peripheral discharge; (B) end peripheral discharge; (C) overflow discharge.

WKH F\OLQGULFDO VKHOO DQG WKH GLVFKDUJH WUXQQLRQ XVXDOO\ UHVXOWV LQ OHVV RYHUJULQGLQJ
KRZHYHUWKLVGLVFKDUJHVDODUJHUIUDFWLRQRIFRDUVHSDUWLFOHV7KHUHIRUHWKLVW\SHRIEDOO
PLOOXVXDOO\RSHUDWHVLQFORVHGFLUFXLW7KHRYHUÁRZGLVFKDUJHEDOOPLOOLVVLPLODUWRWKH
URGPLOOZLWKDGYDQWDJHVRIEHLQJVLPSOHWRRSHUDWHDQGUHTXLULQJOHVVHQHUJ\WKDQD
grate discharge mill.

Classification
&ODVVLÀFDWLRQ LV WKH VHSDUDWLRQ RI SDUWLFOHV DFFRUGLQJ WR WKHLU VHWWOLQJ UDWH XVXDOO\ LQ
ZDWHU7ZRSURGXFWVDUHREWDLQHGRQHUHODWLYHO\FRDUVHUDQGWKHRWKHUUHODWLYHO\ÀQH
$JXLGHOLQHIRUVHOHFWLQJFODVVLÀFDWLRQHTXLSPHQWLVVKRZQLQ7DEOH
$VF\FORQHVDUHWKHPDLQFODVVLÀFDWLRQHTXLSPHQWXVHGLQPLQHUDOSURFHVVLQJSODQWV
WKLVVHFWLRQZLOOGHVFULEHWKHVHOHFWLRQDQGVL]LQJRIK\GURF\FORQHVRQO\

TABLE 16.9
Guidelines for classification (adapted from Kelly and Spottiswood, 1995).
Major classifier types Maximum feed size (mm) Power (kW) Capacity range (t/h)
Spiral, rake, drag 25 0.4–110 5–850
Log washer 100 7.5–60 40–450
Hydraulic bowl 12 2.2–15 5–225
Cone classifier 6 None 2–100
Hydrocyclone 12.5 (pumps) Up to 50 m3/min per hydrocyclone
Elutriator 7.5 (valves) 4–120

Hydrocyclones
$UWHUEXUQ   SXEOLVKHG D VLPSOH PHWKRG IRU HVWLPDWLQJ K\GURF\FORQH
GLPHQVLRQVQXPEHURIK\GURF\FORQHVDQGRSHUDWLQJFRQGLWLRQV7KH$UWHUEXUQPHWKRG
LV DSSOLHG WR D ¶W\SLFDO K\GURF\FORQH· ZKLFK KDV WKH SURSHU JHRPHWULFDO UHODWLRQVKLS
EHWZHHQWKHK\GURF\FORQHGLDPHWHUDQGRWKHUGLPHQVLRQV
7KH K\GURF\FORQH GLDPHWHU LV EDVHG RQ WKH UHTXLUHG SURGXFW VL]H RYHUÁRZ
7KHSURGXFWVL]HLVQRUPDOO\GHÀQHGDVDJLYHQSHUFHQWSDVVLQJDVSHFLÀHGPLFURQ
VL]H ZKLFK FDQ EH UHODWHG WR WKH dc UHTXLUHG WR SURGXFH WKH VSHFLÀHG VHSDUDWLRQ
7DEOH
Three correction factors that take into account operating conditions are applied
to dc(application) to calculate dc(base). The dc(application LV WKH FXW VL]H UHTXLUHG
WR SURGXFH WKH VSHFLÀHG VHSDUDWLRQ ZKLOVW WKH dc(base LV WKH VL]H LQ PLFURQV WKDW D
¶VWDQGDUGK\GURF\FORQH·FDQDFKLHYHXQGHUWKHVHOHFWHGFRQGLWLRQV7KHHTXDWLRQDQG
FRUUHFWLRQIDFWRUVDUHDVGHVFULEHGLQWKHHTXDWLRQVIROORZLQJ

284 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

TABLE 16.10
Relationship of d50c to overflow size distribution (Arterburn, 1982).
Per cent passing a specified micron size Multiplier
98.8 0.54
95.0 0.73
90.0 0.91
80.0 1.25
70.0 1.67
60.0 2.08
50.0 2.78

d50c ^ applicationh
d50c ]baseg
C1 # C2 # C3

 &RQFHQWUDWLRQRIVROLGVLQWKHIHHGVOXUU\

53 - V -1.43
C1 = b
53 l

ZKHUH
C is the correction factor
V is the per cent solids by volume of hydrocyclone feed
2. Pressure drop across the hydrocyclone:

C2 = 3.27 ]TPg-0.28
ZKHUH
C2 is the correction factor
ƦP is the pressure drop in kPa
3. 6SHFLÀFJUDYLW\RIWKHVROLGV

1.65 0.5
C3 = d g - g n
S L
ZKHUH
C3 is the correction factor
gs LVWKHVROLGVVSHFLÀFJUDYLW\
gL LVWKHOLTXLGVSHFLÀFJUDYLW\QRUPDOO\
The hydrocyclone diameter operated in the base conditions is then calculated using
WKHIROORZLQJUHODWLRQVKLS

d50c ]baseg = 2.84D0.66

ZKHUH
D is the hydrocyclone diameter in centimetres (cm)
7KH XQLW ÁRZ UDWH Qu) of a hydrocyclone is based on hydrocyclone diameter and
pressure drop. These parameters are typically found in vendor charts (some charts are
SUHVHQWHGLQWKHZRUNHGH[DPSOHWKDWIROORZV

Comminution Handbook | Spectrum Series 21 285

 CHAPTER 16 t CIRCUIT DESIGN

7KHRSHUDWLRQYROXPHWULFVOXUU\ÁRZUDWHQtLVGLYLGHGE\WKHXQLWÁRZUDWHQu)

RI D K\GURF\FORQH ZLWK WKH VHOHFWHG GLDPHWHU WR GHWHUPLQH WKH QXPEHU RI XQLWV 1
UHTXLUHG
Qt
N
Qu

7KHDSH[GLDPHWHULVGHWHUPLQHGEDVHGRQWKHXQGHUÁRZÁRZUDWHZKLOHWKHLQOHWDQG
YRUWH[VL]HVDUHVHOHFWHGEDVHGRQYHQGRUK\GURF\FORQHFRQÀJXUDWLRQV
7KH ZRUNHG H[DPSOH LQ 7DEOH  ZDV SUHVHQWHG E\ $UWHUEXUQ LQ  ,Q WKLV
H[DPSOHWKHK\GURF\FORQHVL]HDQGQXPEHURIK\GURF\FORQHVZHUHVHOHFWHGIRUDURGPLOO
EDOOPLOOFLUFXLWZKHUHQHZIHHGWRWKHURGPLOOZDVWKVROLGV%RWKPLOOGLVFKDUJHV
MRLQHG WRJHWKHU DW K\GURF\FORQH IHHG VXPS DQG ZHUH SXPSHG WR K\GURF\FORQHV
8QGHUÁRZEHFDPHEDOOPLOOIHHG7KHIROORZLQJLQIRUPDWLRQZDVSURYLGHG
x VSHFLÀFJUDYLW\RIVROLGV 
x RYHUÁRZVL]HGLVWULEXWLRQ SHUFHQWSDVVLQJPP
x RYHUÁRZSHUFHQWVROLGVE\ZHLJKW SHUFHQW
x XQGHUÁRZSHUFHQWVROLGVE\ZHLJKW SHUFHQW
x HVWLPDWHGFLUFXODWLQJORDG SHUFHQW
7KHÀUVWVWHSZDVWRFDOFXODWHPDVVEDODQFHIURPWKHLQIRUPDWLRQSURYLGHG
1H[WWKHUHTXLUHGdcZDVFDOFXODWHGIRUWKHVSHFLÀHGRYHUÁRZ)URP7DEOHWKH
LQWHUSRODWHGPXOWLSOLHUDWSHUFHQWSDVVLQJZRXOGEH

d50c = 2.08 # 0.074 = 0.154 mm

,QLWLDOO\ WKH FRUUHFWLRQ IDFWRUV ZHUH FDOFXODWHG DQG WKHQ WKH K\GURF\FORQH GLDPHWHU
ZDVGHWHUPLQHGDVIROORZV
 &RQFHQWUDWLRQRIVROLGVLQWKHIHHGVOXUU\

53 - V -1.43 53 - 33.2 -1.43

C1 = b
53 l = b 53 l = 4.09

2. Pressure drop across the hydrocyclone:

3UHVVXUH GURS DVVXPHG DW  N3D ZKLFK ZRXOG XVH OHVV HQHUJ\ DQG UHTXLUH OHVV
regular maintenance.
C2 = 3.27 ]TPg-0.28 = 3.27 ] 50g-0.28 = 1.1

TABLE 16.11
Mass balance for the hydrocyclone worked example (adapted from Arterburn, 1982).
Parameter Feed Overflow Underflow
Solid rate (t/h) 812 250 562
Water rate (t/h) 562 375 187
Slurry rate (t/h) 1374 625 749
Per cent solids by weight (%) 59.1 40.0 75.0
Slurry density 1.632 1.355 1.966
Slurry volumetric rate (L/s) 234 128 106

286 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

3. 6SHFLÀFJUDYLW\RIWKHVROLGV
1.65 0.5 1.65 0.5
C3 = d g - g n = b 0.93
S L 2 .9 - 1.0 l =

 Hydrocyclone diameter and number of hydrocyclones:

d50c ^ applicationh 0.154
d50c ]baseg = = = 0.037
C1 # C2 # C3 4.09 # 1.1 # 0.93

)URP)LJXUHVHOHFWHGGLDPHWHULVLQFKHVFP
7KH QXPEHU RI XQLWV UHTXLUHG FDQ EH REWDLQHG XWLOLVLQJ WKH FKDUW SUHVHQWHG LQ
)LJXUH
7KHVWDQGDUGK\GURF\FORQHYROXPHWULFFDSDFLW\LVSUHVHQWHGLQ86JDOORQVSHUPLQXWH
LQ)LJXUH86JDOPLQ /V,QWKLVÀJXUHWKHYROXPHWULFÁRZUDWHIRU
DLQFKGLDPHWHUK\GURF\FORQHDWN3DZRXOGEH/V7KHUHIRUHWKHQXPEHURI
units is:
Q
N = t = 234 = 5.85
Qu 40

6L[K\GURF\FORQHVLQRSHUDWLRQZRXOGEHUHTXLUHG
 &DOFXODWHDSH[VL]H
7KH SURSHU DSH[ VL]H LV REWDLQHG IURP DSH[ FDSDFLW\ FXUYHV VXFK DV WKH RQH
SUHVHQWHGLQ)LJXUH)RUVL[K\GURF\FORQHVWKHXQGHUÁRZYROXPHWULFÁRZSHU
XQLWZRXOGEH
QUF = 106 = 18 L/s
6
)URP)LJXUHHDFKDSH[VKRXOGEHFPöLQ

FIG 16.5 – Typical cyclone diameters (image courtesy Society of Mining, Metallurgy, and Exploration).

Comminution Handbook | Spectrum Series 21 287

 CHAPTER 16 t CIRCUIT DESIGN

FIG 16.6 – Volumetric capacity for standard cyclones (image courtesy Society of Mining, Metallurgy, and Exploration).

FIG 16.7 – Apex capacity curves (image courtesy Society of Mining, Metallurgy, and Exploration).

288 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

High-pressure grinding rolls

7KHVL]LQJRI+3*5XQLWVLQYROYHVWKHVHOHFWLRQRIUROOJHRPHWU\DQGVSHHGWRZKLFK
WKHUHTXLUHGPDFKLQHWKURXJKSXWLVGLUHFWO\UHODWHG7KHDSSURDFKGHVFULEHGLVEDVHG
RQWKHVSHFLÀFWKURXJKSXW mo IRUPXODXVHGWRVL]HDQ+3*5XQLWWRDFKLHYHWKHGHVLUHG
WKURXJKSXW7KLVSDUDPHWHULVWKHFDSDFLW\RIDQ+3*5ZLWKDGLDPHWHURIPDZLGWK
RI  P DQG D VSHHG RI  PV DQG LW LV XVXDOO\ REWDLQHG IURP +3*5 EHQFK RU SLORW
VFDOHWHVWZRUN,QDGGLWLRQDVSHFWVRIRUHVXLWDELOLW\WR+3*5SURFHVVLQJDUHREVHUYHG
GXULQJWKLVWHVWZRUNLQFOXGLQJWKHLQÁXHQFHRIPRLVWXUHOHYHOVLQVSHFLÀFFDSDFLW\RU
WKHVXVFHSWLELOLW\RIDQH[FHVVLYHIRUPDWLRQRIFRPSHWHQWÁDNHVWKHODWWHURIZKLFKPD\
QHHGLQWHQVLYHGHDJJORPHUDWLRQEHIRUHIXUWKHUWUHDWPHQW
7DEOH  SUHVHQWV W\SLFDO YDOXH UDQJHV IRU HDFK SDUDPHWHU XVHG LQ WKH HTXDWLRQV
OLVWHGLQWKHIROORZLQJVHFWLRQVWRVL]HDQ+3*5XQLW

TABLE 16.12
Typical value ranges of high-pressure grinding rolls parameters (Ausenco, 2014).
Parameter Symbol Typical value range
Specific throughput (ts/hm )3
mo 150–300
Throughput (t/h) M 50–3000
Roll diameter (m) D 0.95–2.6
Roll width (m) L 0.4–2.3
Roll peripheral speed (m/s) μ 0.95–2.6
Roll angular velocity (rev/min) 19–21
Operating gap (mm) s 20–80
Flake density (kg/L) ρg 2–3
Specific energy (kWh/t) 1.2–2.0
Motor power per roll (kW) 280–3300

Specific throughput
7KH VSHFLÀF WKURXJKSXW  mo ) is a normalised parameter representing the capacity of an
+3*5XQLWZLWKDUROOGLDPHWHUDRIPDUROOZLGWKLRIPDQGDUROOFLUFXPIHUHQWLDO
VSHHG—RIPV,WLVH[SUHVVHGE\WKHIROORZLQJHTXDWLRQYRQ6HHEDFK

mo M/ ]DLng

Angular velocity
7KH VSHFLÀF WKURXJKSXW  mo  LV REWDLQHG IURP ODERUDWRU\ RU SLORWVFDOH +3*5 WHVW
ZRUN%DVHGRQWKHYDOXHRI mo DQGWKHWKURXJKSXWUHTXLUHPHQWWKHUROOJHRPHWU\
FDQEHFKRVHQIURPVWDQGDUG+3*5YHQGRUVL]HV7KHUHTXLUHGUROOSHULSKHUDOVSHHG
is calculated and the resulting roll angular velocity determined. Roll geometry and
VSHHGDUHLWHUDWHGXQWLODQDQJXODUYHORFLW\LQWKHUDQJHRI²UHYPLQLVREWDLQHG
The angular velocity is calculated using the selected roll diameter and peripheral
VSHHGDVIROORZV
60n
Angular velocity =
rD

Comminution Handbook | Spectrum Series 21 289

 CHAPTER 16 t CIRCUIT DESIGN

Operating gap
The operating gap is determined predominantly by the roll diameter and roll surface
WH[WXUHIRUH[DPSOHVPRRWK1L+DUGSURÀOHGVWXGGHGWKHSDUWLFOHVL]LQJRIWKHRUH
EHLQJ SURFHVVHG DQG PRVW LPSRUWDQWO\ ZKHWKHU WKH IHHG LV IXOOÀQHV RU WUXQFDWHG ²
WKDW LV ZLWK ÀQH PDWHULDO VFUHHQHG RXW LQ DQ XSVWUHDP RSHUDWLRQ 7KH PRVW FRPPRQ
roll surface used in hard rock applications comprises hard alloy studs in a cast steel
VXEVWUDWH:LWKWKLVVXUIDFHW\SLFDOYDOXHVIRUWKHRSHUDWLQJJDSDUH²SHUFHQWDQG
²SHUFHQWRIWKHUROOGLDPHWHUIRUIXOOÀQHVDQGWUXQFDWHGIHHGVUHVSHFWLYHO\
$VDFKHFNRQWKHUHTXLUHGFDSDFLW\DQGVHOHFWHGUROOJHRPHWU\DQGVSHHGWKHRSHUDWLQJ
JDS LV FDOFXODWHG XVLQJ WKH FRQWLQXLW\ HTXDWLRQ :KLWH  ZLWK DQ +3*5 SURGXFW
ÁDNHGHQVLW\DVGHWHUPLQHGLQWHVWZRUNXVXDOO\LQWKHRUGHURISHUFHQWRIWKHRUH
SDUWLFOHVSHFLÀFJUDYLW\6*
M = 3.6Lsnt g

$YDOXHVLJQLÀFDQWO\RXWVLGHWKHUDQJHVQRWHGLQGLFDWHVDQHHGWRUHYLHZWKHVHOHFWLRQ

Edge effect
7KH¶HGJHHIIHFW·LVDWHUPXVHGWRGHVFULEHWKHLPSDLUHGFRPPLQXWLRQSHUIRUPDQFHLQ
WKH]RQHDWWKHHGJHRIWKHUROOVGXHWRWKHUHGXFHGFRPSUHVVLYHIRUFHVDWZRUNLQWKLV
]RQH7KH¶HGJH·IUDFWLRQLVFRUUHVSRQGLQJO\FRDUVHUWKDQWKHFHQWUHPDWHULDOGXHWRWKH
reduced energy input.
The edge effect is related to the operating gap (and therefore by extension to the roll
GLDPHWHU DQG WKH HGJH IUDFWLRQ DV D SURSRUWLRQ RI WRWDO SURGXFW LV GHWHUPLQHG DV
IROORZV
f = 2.4s
1000L
7KHHGJHPDWHULDOPD\EHKDQGOHGLQRQHRIVHYHUDOZD\VGHSHQGLQJRQWKHÁRZVKHHW
design:
x IRU VLQJOHSDVV RSHQFLUFXLW +3*5 RSHUDWLRQ LW VLPSO\ DFFRPSDQLHV WKH FHQWUH
material as a part of the total product
x IRUFORVHGFLUFXLWRSHUDWLRQWKHPDMRULW\RIWKHHGJHPDWHULDOZLOOUHSRUWDVVFUHHQ
RYHUVL]HDQGUHWXUQWR+3*5IHHG
x LQHGJHUHF\FOHPRGHWKHHGJHIUDFWLRQLVVHSDUDWHGIURPWKHFHQWUHSURGXFWXVLQJ
VSOLWWHUJDWHVLQWKHGLVFKDUJHFKXWHDQGUHWXUQHGWR+3*5IHHG
)RU RSHQFLUFXLW RSHUDWLRQ VLQJOHSDVV +3*5 WHVWLQJ LV VXIÀFLHQW WR GHWHUPLQH WKH
DSSOLFDEOHVSHFLÀFWKURXJKSXWYDOXHKRZHYHUWKHXVHRIFORVHGFLUFXLWRUHGJHUHF\FOH
PRGHDIIHFWVWKHVSHFLÀFWKURXJKSXWE\FKDQJLQJWKH+3*5IHHGSDUWLFOHVL]LQJDQG
WKHWHVWSURJUDPLQWKHVHFDVHVPXVWLQFOXGHORFNHGF\FOHWHVWVWRUHSOLFDWHWKHIXOOVFDOH
plant design.

Power requirements
:KHQ WKH +3*5 UHTXLUHG FDSDFLW\ VSHFLÀF WKURXJKSXW UROO JHRPHWU\ VSHHG DQG
RSHUDWLQJ SDUDPHWHUV DUH GHÀQHG WKH SRZHU UHTXLUHPHQW FDQ EH GHWHUPLQHG
7KHVSHFLÀFHQHUJ\LVREWDLQHGIURP+3*5WHVWZRUNXVLQJVLQJOHSDVVRUORFNHGF\FOH
WHVWV DV DSSURSULDWH 6SHFLÀF HQHUJ\ LV UHODWHG WR +3*5 QHZ IHHG LQ FRPPLQXWLRQ

290 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

energy calculations and to total feed (including any closed circuit or edge recycle) to
GHWHUPLQHPDFKLQHSRZHUUHTXLUHPHQWV7KHDEVRUEHGSRZHUFDOFXODWHGLVWKHQXVHG
IRUVHOHFWLRQRIDVXLWDEOHPRWRUVL]HWDNLQJLQWRDFFRXQWWKHPRWRUHIÀFLHQF\IDFWRUDQG
GULYHWUDLQORVVHVDQGWKHGHUDWLQJIDFWRUDSSOLFDEOHWRDFFRPPRGDWHWKHYDULDEOHVSHHG
GULYHIDFLOLW\WKDWLVQRUPDOO\XVHGLQPRVWPRGHUQ+3*5LQVWDOODWLRQV

PLANT LAYOUT

Overview and objectives

$OWKRXJK WKHUH LV QR XQLYHUVDO VROXWLRQ IRU SODQW OD\RXWV FRQVLGHUDWLRQV LQYDULDEO\
LQFOXGH RZQHUSUHIHUHQFHVEXGJHWVFKHGXOH DQG VLWHVSHFLÀF DVSHFWV 7KH UROH RI WKH
engineer is to ensure plant design and circuit operation and maintenance is achieved
LQ D IXQFWLRQDO VDIH DQG FRVWHIIHFWLYH PDQQHU WKDW SURYLGHV PD[LPXP YDOXH WR WKH
SURMHFW$FFRUGLQJWR/DQH'DNLQDQG(OZLQW\SLFDOFRQFHQWUDWRUOD\RXWVWHQGWR
EHGHVLJQHGFRQVHUYDWLYHO\IRURQHRUPRUHRIWKHIROORZLQJUHDVRQV
x 'HVLJQHUVGRQRWNQRZKRZ¶FORVHWRWKHZLQG\RXFDQVDLO·ZKDWLVÀWIRUGHVLJQ
RUDUHLQVXIÀFLHQWO\H[SHULHQFHGWRXQGHUVWDQGWKHFRVWULVNEHQHÀWUHODWLRQVKLSV
7KLVFDQFRQWULEXWHWRLQFUHDVLQJWKHKHLJKWWRLQVWDOOHTXLSPHQWUDLVLQJWKHEHQFK
OHYHORISURFHVVDUHDVDQGLQFUHDVLQJWKHGLVWDQFHEHWZHHQHTXLSPHQWDQGSURFHVVLQJ
DUHDV ² DOO RI ZKLFK ZLOO LQFUHDVH WKH FRVW RI EXON PDWHULDOV 7KHVH PD\ EH EDVHG
RQ PLVOHDGLQJ ¶VWDQGDUGV· RU SUHYLRXV SUDFWLFHV IRU H[DPSOH FRQFHQWUDWH ODXQGHU
VORSHVVHWWLQJKHLJKWRIPLOOVDQGÁRWDWLRQFHOOV
x 7KHUHLVDODFNRILQQRYDWLRQLQGHVLJQEHFDXVHWKLVUHTXLUHVH[SHULHQFHDQGHIIHFWLYH
risk management. Examples include: using the topography and gravity instead of
SXPSV DYRLGLQJ WKH XVH RI SLSH UDFNV WKDW GRPLQDWH OD\RXWV DQG EXLOGLQJ FORVH
VWDFNHGYHUWLFDOSODQWDQGHTXLSPHQWVXFKDVSODFLQJWKHPL[LQJWDQNRQWRSRID
VWRUDJHWDQNRUORFDWLQJWKHPLOOFRQWUROFHQWUH0&&XQGHUWKHPLOOIHHGSODWIRUP
x 7KHUHLVDWHQGHQF\WR¶UHLQYHQWWKHZKHHO·LQVWHDGRIUHXVLQJLPSURYLQJRUDGDSWLQJ
SUHYLRXVSURYHQGHVLJQ)XUWKHUPRUHUHIHUHQFHGHVLJQOLEUDULHVDUHQRWZHOONQRZQ
or published.
x 0DQ\ GHVLJQHUV GR QRW KDYH ÀHOG LQVWDOODWLRQ FRPPLVVLRQLQJ DQG VLWH DVEXLOW
H[SHULHQFH SDUWLFXODUO\ RQ SURMHFWV WKH\ KDYH GHVLJQHG 'HVLJQHUV ZLWK WKLV
H[SHULHQFH DUH EHWWHU HTXLSSHG WR XQGHUVWDQG WKH EDVLFV RI OD\RXW DQG RSHUDELOLW\
and then implement these lessons in future designs.
x (QJLQHHUV FRPH IURP RWKHU LQGXVWULHV IRU H[DPSOH SHWURFKHPLFDO RU RLO DQG JDV
LQGXVWULHV 7KHLU GHVLJQ H[SHULHQFH LQ WKHVH ÀHOGV XVXDOO\ KDV OLWWOH RU QR OD\RXW
FRPPLVVLRQLQJRURSHUDWLRQDOUHOHYDQFHWRÁH[LEOHDQGFRPSDFWPLQHUDOSURFHVVLQJ
plant design.
&RVWHIIHFWLYHSODQWGHVLJQJXLGHOLQHVDUHIXUWKHUGHVFULEHGE\/DQH'DNLQDQG(OZLQ
.H\JRDOVDUH
x $OLJQ WKH GHVLJQ WHDP RQ UHTXLUHPHQWV IRU VDIH RSHUDWLQJ DQG PDLQWHQDQFH
practices.
x 2SWLPLVHWKHSODQWIRRWSULQWWRUHGXFHFRQFUHWHVWUXFWXUDOVWHHOSLSLQJDQGHOHFWULFDO
FRQWURO FDEOH DQG UDFHZD\V $UUDQJH DOO DUHDV WR EH FORVH WRJHWKHU IRU H[DPSOH
JULQGLQJFORVHWROHDFKLQJRUÁRWDWLRQGHVRUSWLRQFORVHWRDGVRUSWLRQUHDJHQWVFORVH

Comminution Handbook | Spectrum Series 21 291

 CHAPTER 16 t CIRCUIT DESIGN

WROHDFKLQJRUÁRWDWLRQDLUDQGZDWHUVHUYLFHVFORVHWRWKHSODQWELJPRWRUVFORVHWR
WKHVZLWFKURRP
x .HHSHOHYDWLRQRIWKHHTXLSPHQWWRDZRUNDEOHPLQLPXP7KHHOHYDWLRQRIUXQRI
PLQHELQVPLOOVF\FORQHVÁRWDWLRQFHOOVDQGWKLFNHQHUVDUHNH\GHWHUPLQDQWV/DQH
et al
x Avoid use of dominating pipe racks or large platform areas.
x 8VHFRPPRQSODWIRUPVVWDLUVDQGSLSHDQGFDEOHODGGHUVXSSRUWV
x 3ODWHZRUNDQGOLQLQJVKRXOGEHNHSWOHDQIRULQVWDQFHRQO\SXWZHDUOLQHUVLQWKH
FKXWHDUHDVH[SRVHGWRZHDUDQGQRWDOOLQWHUQDOV
x 'HVLJQZLWKDÀWIRUSXUSRVHDWWLWXGHDVWKRXJKLWLV\RXUPRQH\\RXDUHVSHQGLQJ
6RPH¶UXOHVRIWKXPE·DSSO\VXFKDV
x WKHH[HFXWLRQVWUDWHJ\DQGSODQVKRXOGEHVLPSOHZLWK¶EDJJDJH·VLIWHGIURPWKHIDFWV
early and the agreed plan adopted and adhered to
x PLQLPLVLQJ WKH QXPEHU RI LQWHUIDFHV DFURVV DOO SDUWLHV DV HDFK LQWHUIDFH UHTXLUHV
PDQDJLQJDQGLVDSRWHQWLDOVRXUFHRILQHIÀFLHQWFRPPXQLFDWLRQV
x LQYHVWLQJLQSURYHQDQGUHOLDEOHHTXLSPHQWDVLWJHQHUDOO\VDYHVPRQH\LQWKHORQJHU
term
x UHGXFLQJ IRRWSULQW DV FDSLWDO DQG RSHUDWLQJ FRVWV WHQG WR LQFUHDVH SURSRUWLRQDOO\
ZLWKIRRWSULQWDQGSODQWFRPSOH[LW\

SAB/SABC plant layout

7KH WZR PDMRU DUUDQJHPHQWV IRU 6$%6$%& FLUFXLW OD\RXWV DUH HQGWRHQG
FRQÀJXUDWLRQDQGVLGHE\VLGHFRQÀJXUDWLRQUHIHUWR)LJXUH7KHLUDSSOLFDELOLW\
PDLQDGYDQWDJHVDQGGUDZEDFNVDUHVXPPDULVHGLQWKLVVHFWLRQ

End-to-end configuration
7KHIHDWXUHVRIWKLVFRQÀJXUDWLRQDUHDVIROORZV
x UHTXLUHVWZRRYHUKHDGFUDQHVGHSHQGLQJRQWKHPLOOVSHFLÀFV
x DFFHVV LV GLIÀFXOW IRU VHUYLFLQJ WKH SXPSV ZKLFK DUH ORFDWHG DW WKH FHQWUH RI WKH
building
x QHHGV PRUH UHOLQLQJ ÁRRU DUHD EHFDXVH 6$* DQG EDOO PLOOV KDYH LQGHSHQGHQW
UHOLQLQJÁRRUV
x K\GURF\FORQHXQGHUÁRZFDQQRWEHHDVLO\UHWXUQHGWRWKH6$*PLOOLIQHFHVVDU\

Side-by-side configuration
7KHIHDWXUHVRIWKLVFRQÀJXUDWLRQDUHDVIROORZV
x typically produces a smaller building footprint and consolidated overhead crane
span
x DEOHWREHVHUYLFHGZLWKRQHFUDQHLQPRVWDSSOLFDWLRQV
x HQDEOHVHDVLHUDFFHVVIRUVHUYLFLQJWKHSXPSVZKLFKDUHORFDWHGDWWKHSHULSKHU\RI
the building
x KDV D FRPPRQ UHOLQLQJ ÁRRU IRU ERWK 6$* DQG EDOO PLOOV ZKLFK FDQ UHGXFH WKH
EXLOGLQJDUHDKHQFHDVPDOOHUIRRWSULQW

292 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

FIG 16.8 – Semi-autogenous grinding and ball mills: end-to-end (top) and side-by-side (bottom) configurations
(Dufour et al, 2011; published with the permission of the Canadian Institute of Mining, Metallurgy and Petroleum (CIM)).

High-pressure grinding rolls – ball mill circuit layout

,QWKLVVHFWLRQOD\RXWRSWLRQVIRUDQ+3*5²EDOOPLOOFLUFXLWDUHH[DPLQHGIRUWG
/D\RXW$%DQG&FDVHVDQGDWG/D\RXW'FDVH7KHUDWLRQDOHVXSSRUWLQJ
WKH YDULRXV FRQFHSWV LV H[SODLQHG ZLWK WKHLU UHODWLYH PHULWV DQG D UHFRPPHQGDWLRQ
7KLV VHFWLRQ LV EDVHG RQ $XVHQFR·V H[SHULHQFH ZLWK +3*5 FLUFXLWV DQG WKH ZRUN RI
0RUOH\

Layout A – 60 000 t/d

/D\RXW$SUHVHQWHGLQ)LJXUHPLQLPLVHVWKHQXPEHURIEXLOGLQJVUHTXLUHGIRU
WKHFRPPLQXWLRQFLUFXLW,WFRPSULVHVRQO\WZREXLOGLQJV
 VHFRQGDU\FUXVKLQJDQG+3*5
2. screening and grinding.
This arrangement uses a total of ten short transfer conveyors (not including belt
feeders).
8VLQJ D SUHGRPLQDQWO\ V\PPHWULFDO OD\RXW WR DFKLHYH IUHHWUDMHFWRU\ VSOLWWLQJ RI
WKH RUH VWUHDPV WKLV DUUDQJHPHQW REYLDWHV WKH QHHG IRU WULSSHUV VKXWWOHV RU RWKHU

Comminution Handbook | Spectrum Series 21 293

 CHAPTER 16 t CIRCUIT DESIGN

FIG 16.9 – 60 000 t/d layout A (image courtesy Ausenco).

WZRZD\GLVWULEXWLRQGHYLFHVKRZHYHUWKHIHHGWRWKH+3*5IHHGELQLVDV\PPHWULF
DQGWKHQDWXUDOVHJUHJDWLRQRIWKHIUHHWUDMHFWRU\ZLOOFDXVHDFRDUVHÀQHELDVQRWRQO\
EHWZHHQPRGXOHVEXWDOVRDFURVVWKHUROOZLGWKRIHDFK+3*5XQLWFDXVLQJUROOVNHZ
DQGXQHYHQZHDU
6XLWDEOHVXUJHELQVDUHSURYLGHGDKHDGRIWKHVHFRQGDU\FUXVKHUV+3*5VDQG+3*5
ÀQHVFUHHQVEXWQRWDKHDGRIWKHFRQHFUXVKHUFRDUVHVFUHHQV7KHOLNHO\RSHUDWLRQRI
WKHVHFRQGDU\FUXVKHUVLQVWDUW²VWRSPRGH²DVLVW\SLFDOLQDQ\VWDJHFUXVKLQJSODQW²
ZRXOGPHDQWKDWWKHWKHVHVFUHHQVZRXOGRSHUDWHZLWKDF\FOLFIHHGOHDGLQJWRLQHIÀFLHQW
operation and possible overloading.
7KHOD\RXWLQFOXGHVVXEVWDQWLDOVWRUDJHEHWZHHQVHFRQGDU\FUXVKLQJDQG+3*5EXW
QRQHEHWZHHQ+3*5DQGPLOOLQJWKHVPDOOVXUJHELQVQRWSURYLGLQJDQ\PHDVXUDEOH
storage per se7KLVUHVXOWVLQWKH+3*5VÀQHVFUHHQVPLOOLQJDQGHQWLUHGRZQVWUHDP
SODQWDUUDQJHGLQDQXQLQWHUUXSWHGWUDLQ7KLVFORVHOLQNLQJRIWKHFRPPLQXWLRQVWHSV
LQ SDUWLFXODU PHDQV WKDW LPSDLUHG SODQW XWLOLVDWLRQ IDFWRUV ZRXOG DSSO\ 7KH UHGXFHG
DYDLODELOLW\ GXH WR WKH +3*5 PHDQV WKDW ODUJHU HTXLSPHQW ZRXOG EH QHHGHG IRU D
QRPLQDWHGDQQXDOLVHGWKURXJKSXWRUWKDWUHGXFHGWKURXJKSXWZRXOGRFFXUXVLQJWKH
VDPHHTXLSPHQWVL]HV

Layout B – 60 000 t/d

:LWKWKHH[FHSWLRQRIWKHÀQHVFUHHQVWZRPRGXOHVRIHDFKXQLWRSHUDWLRQDUHUHTXLUHG
LQ WKLV GHVLJQ )LJXUH  DV ZLWK OD\RXW $ 7KLV DOORZV WKH V\PPHWULFDO IUHH
WUDMHFWRU\FRQFHSWVIURP/D\RXW$WREHUHWDLQHGEXWZLWKWKHDUUDQJHPHQWH[WHQGHGWR
LQFOXGHWKH+3*5IHHG7KHFRQYH\RUFRQÀJXUDWLRQKDVEHHQUDWLRQDOLVHGUHVXOWLQJLQD
UHGXFWLRQIURPWHQWRHLJKWFRQYH\RUVKRZHYHUWRDFKLHYHWKLVWKHQXPEHURIEXLOGLQJV
has increased to four:
 secondary crushing
2. coarse screening

294 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

FIG 16.10 – 60 000 t/d layout B (image courtesy Ausenco).

3. +3*5
 ÀQHVFUHHQLQJDQGPLOOLQJ
The disadvantage is that operations have been decentralised. The major ore storage
IDFLOLW\ KDV EHHQ UHORFDWHG IURP XSVWUHDP RI +3*5 WR XSVWUHDP RI ÀQH VFUHHQLQJ WR
PD[LPLVHPLOOLQJXWLOLVDWLRQIDFWRUV7KHIUHHWUDMHFWRU\VSOLWRIIHHGLQWRWKHVHFRQGDU\
crusher and coarse screen feed bins is considered acceptable:
x VHFRQGDU\FUXVKHUIHHGLVWUXQFDWHGWKDWLVÀQHVIUHHDQGVHJUHJDWLRQLVOHVVRIDQ
LVVXHWKDQZLWKDIXOOÀQHVIHHG
x WKHFRDUVHVFUHHQLQJVHSDUDWLRQHIÀFLHQF\LVQRWFULWLFDODQGWKHH[SHFWHGVHJUHJDWLRQ
DFURVVWKHZLGWKRIHDFKVFUHHQLVFRQVLGHUHGWROHUDEOH
<HWWKLVLVQRWWKHFDVHIRUWKH+3*5DQGÀQHVFUHHQLQJGXWLHV
x 8QVHJUHJDWHG IHHG DFURVV WKH +3*5 UROO ZLGWK DV QRWHG SUHYLRXVO\ LV HVVHQWLDO
KRZHYHUWKLVOD\RXWZRXOGIDLOWRDFKLHYHWKHREMHFWLYHDVWKHRXWHUHGJHVRIWKHUROOV
UHFHLYHSUHGRPLQDQWO\FRDUVHPDWHULDODQGWKHLQQHUHGJHVWKHÀQHV
x 7KHIRXUÀQHVFUHHQVSURSRVHGZRXOGUHFHLYHELDVHGIHHGZKLOHWKHRXWHUWZRVFUHHQV
ZRXOGUHFHLYHFRDUVHUIHHGDQGWKHLQQHURQHVWKHÀQHV7KHVHVFUHHQVDUHEHGGHSWK
OLPLWHGDQGLWFRXOGEHH[SHFWHGWKDWVXFKDELDVZRXOGOHDGWRRYHUORDGLQJRIWKH
outer screens.
6RPHRIWKHVHGHÀFLHQFLHVDUHDGGUHVVHGLQ/D\RXW&

Layout C – 60 000 t/d

,QWKLVYHUVLRQWKH+3*5DQGÀQHVFUHHQELQVDUHIHGXVLQJDWULSSHURUVKXWWOHVXFK
that each bin receives the full feed stream alternately rather than a continuous feed
RIDSRUWLRQRIWKHWRWDO)LJXUH7KLVDSSURDFKLVJHQHUDOO\XVHGWRRYHUFRPH
WKHVHJUHJDWLRQLVVXHVWKDWW\SLI\DÀ[HGWUDMHFWRU\VLQJOHSRLQWELQIHHGVWUHDP7KH
LQFOXVLRQ RI VKXWWOHV PHDQV WKDW WKH ELQV VKRXOG SUHIHUDEO\ EH LQOLQH ZLWK WKH IHHG
FRQYH\RUWKHDOWHUQDWLYHZRXOGEHDQDGGLWLRQDOFRQYH\RUDQGWUDQVIHUSRLQWDWWKH
bin entry.

Comminution Handbook | Spectrum Series 21 295

 CHAPTER 16 t CIRCUIT DESIGN

FIG 16.11 – 60 000 t/d layout C (image courtesy Ausenco).

7KLVDUUDQJHPHQWDOVRUHTXLUHVIRXUEXLOGLQJVDQGHLJKWFRQYH\RUVDVLQOD\RXW%EXW

LQDVOLJKWO\PRUHFRPSDFWOD\RXW7KHÀQHVFUHHQVDUHUHYHUVHORDGHGWRVXLWWKHUHWXUQ
RIVFUHHQRYHUVL]HWR+3*5IHHG

Layout D – 75 000 t/d

)RU WKH LQFUHDVHG WKURXJKSXW FDVH WKUHH PRGXOHV DUH UHTXLUHG LQ WKH VHFRQGDU\
FUXVKLQJDQGFRDUVHVFUHHQLQJGXWLHV7KLVUHTXLUHVWKHXVHRIVKXWWOHVRUWULSSHUVIRU
IHHG GLVWULEXWLRQ 7KH +3*5 DQG ÀQH VFUHHQLQJJULQGLQJ FRQÀJXUDWLRQ UHPDLQV DW
WZRPRGXOHVHDFKXVLQJODUJHUXQLWV7KXVDOOXQLWRSHUDWLRQVQRZUHTXLUHVKXWWOHIHHG
GLVWULEXWLRQDQGDQDOWHUQDWLYHOD\RXWDVVKRZQLQ)LJXUHEHFRPHVYLDEOH
x conveyor numbers are further reduced (to six) and all but one transfer points are
eliminated
x EXLOGLQJQXPEHUVGHFUHDVHWRWKUHHVHFRQGDU\FUXVKLQJDQG+3*5FRDUVHVFUHHQLQJ
DQGÀQHVFUHHQLQJDQGPLOOLQJ

FIG 16.12 – 70 000 t/d layout D (image courtesy Ausenco).

296 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

,QWKLVDUUDQJHPHQWZLWKDKLJKHUWKURXJKSXWWKDQWKHSUHYLRXVYHUVLRQVDOOJRRG
SUDFWLFH PDWHULDOVKDQGOLQJ UHTXLUHPHQWV DUH PHW ZKLOH DOVR DFKLHYLQJ D VLJQLÀFDQW
UHGXFWLRQLQFRQYH\RUQXPEHUV7KLVUHSUHVHQWVDVLJQLÀFDQWRSHUDELOLW\DQGPDLQWHQDQFH
advantage.

BIBLIOGRAPHY
Arterburn,5$The Sizing of Hydrocyclones.UHEV(QJLQHHUV0HQOR3DUN
Arterburn, 5 $  7KH VL]LQJ DQG VHOHFWLRQ RI K\GURF\FORQHV LQ Design and Installation of
Comminution CircuitVHGV$/0XODUDQG*9-HUJHQVHQFKDSWHUSS²6RFLHW\IRU
0LQLQJ0HWDOOXUJ\DQG([SORUDWLRQ/LWWOHWRQ
Ausenco,$XVJULQGSURJUDPPHLQWHUQDOQRWHV
Austin, /  $ PLOO SRZHU HTXDWLRQ IRU 6$* PLOOV Minerals and Metallurgical Processing
SS²
Barratt,'-$QXSGDWHRQWHVWLQJVFDOHXSDQGVL]LQJHTXLSPHQWIRUDXWRJHQRXVDQGVHPL
DXWRJHQRXV JULQGLQJ FLUFXLWV LQ Proceedings Advances in Autogenous and Semi-Autogenous
Grinding Technology: SAG 1989HGV$/0XODUDQG*($JDUSS²8QLYHUVLW\RI%ULWLVK
&ROXPELD9DQFRXYHU
Barratt, ' - DQG $OODQ 0 -  7HVWLQJ IRU DXWRJHQRXV DQG VHPLDXWRJHQRXV JULQGLQJ D
GHVLJQHU·VSRLQWRIYLHZMinerals and Metallurgical Processing²
Barratt,'-DQG%URGLH07KH¶WHQW·GLDJUDPZKDWLWPHDQVLQProceedings International
Autogenous and Semi-autogenous Grinding Technology (SAG 2001) HGV0DMRUet al&DQDGLDQ
,QVWLWXWHRI0LQLQJ0HWDOOXUJ\DQG3HWUROHXP0RQWUHDO
Chaves,$3DQG3HUHV$(&Teoria e Prática do Tratamento de MinériosYROXPH6LJQXVHGLWRUD
Dufour,/0F9H\.3HUNLQV65RODQGL/DQG<RXQJ(6$*DQGEDOOPLOOFLUFXLWOD\RXW
DQGGHVLJQLQProceedings International Conference on Autogenous and Semiautogenous Grinding
Technology (SAG 2011)HGV.0DMRU%&)OLQWRII%.OHLQDQG.0F/HRG&DQDGLDQ,QVWLWXWH
RI0LQLQJ0HWDOOXUJ\DQG3HWUROHXP0RQWUHDO
Kelly,(*DQG6SRWWLVZRRG'-Introduction to Mineral ProcessingS-RKQ:LOH\DQG6RQV
Lane,*%ULQGOH\6*UHHQ6DQG0F/HRG''HVLJQDQGHQJLQHHULQJRIÁRWDWLRQFLUFXLWV
LQ$XVWUDOLDLQProceedings Centenary of Flotation SymposiumSS²7KH$XVWUDODVLDQ
,QVWLWXWHRI0LQLQJDQG0HWDOOXUJ\0HOERXUQH
Lane,*'DNLQ3DQG(OZLQ'&RVWHIIHFWLYHFRQFHQWUDWRUGHVLJQLQProceedings Metallurgical
Plant Design and Operating Strategies (MetPlant) 2011SS²7KH$XVWUDODVLDQ,QVWLWXWHRI
0LQLQJDQG0HWDOOXUJ\0HOERXUQH
Lane, * )RJJLDWWR % DQG %XHQR 0 3  3RZHUEDVHG FRPPLQXWLRQ FDOFXODWLRQV XVLQJ
$XVJULQGLQProceedings Tenth International Mineral Processing Conference*HFDPLQ
Loveday,%.3UHGLFWLRQRIDXWRJHQRXVPLOOLQJIURPSLORWSODQWWHVWVLQProceedings 11th
Commonwealth Mining and Metallurgical Congress HG 0 - -RQHV ,QVWLWXWLRQ RI 0LQLQJ DQG
0HWDOOXUJ\
McIvor, 5 (  $ PDWHULDO EDODQFH FDOFXODWLRQ SURFHGXUH IRU JULQGLQJ FLUFXLW K\GURF\FORQH
VHOHFWLRQCIM BulletinSS²
McLean,(-:H·UHPHWDOOXUJLVWVQRWPDJLFLDQVLQProceedings Metallurgical Plant Design
and Operating Strategies (MetPlant) 2011SS²7KH$XVWUDODVLDQ,QVWLWXWHRI0LQLQJDQG
0HWDOOXUJ\0HOERXUQH
Metso Minerals,Crushing and Screening HandbookWKLUGHGLWLRQ

Comminution Handbook | Spectrum Series 21 297

 CHAPTER 16 t CIRCUIT DESIGN

Metso Minerals,5RGPLOOVRQOLQHEURFKXUH$YDLODEOHIURPKWWSZZZPHWVRFRP

LQPLQLQJDQGFRQVWUXFWLRQ0&7B,QGLDQVI:HE:,':7%)'()LOH
5RG0LOO(1SGI!>$FFHVVHG-XO\@
Morley,&3ODQWOD\RXWFRQVLGHUDWLRQVIRUKLJKSUHVVXUHJULQGLQJUROOVLQProceedings Fifth
International Mineral Processing ConferenceSS²*HFDPLQ
Morrell,63RZHUGUDZRIZHWWXPEOLQJPLOOVDQGLWVUHODWLRQVKLSWRFKDUJHG\QDPLFV3DUW
D FRQWLQXXP DSSURDFK WR PDWKHPDWLFDO PRGHOOLQJ RI PLOO SRZHU GUDZ Transactions of the
Institution of Mining and MetallurgyMineral Processing and Extractive Metallurgy&²&
Morrell, 6  3UHGLFWLQJ WKH RYHUDOO VSHFLÀF HQHUJ\ UHTXLUHPHQW RI FUXVKLQJ KLJK SUHVVXUH
JULQGLQJUROODQGWXPEOLQJPLOOFLUFXLWV Minerals Engineering²
Morrell, 6  8VLQJ WKH 60& 7HVW® WR SUHGLFW FRPPLQXWLRQ FLUFXLW SHUIRUPDQFH >RQOLQH@
$YDLODEOH IURP KWWSZZZVPFWHVWLQJFRPGRFXPHQWV8VLQJBWKHB60&B7HVWSGI!
>$FFHVVHG0D\@
Muckermann, -  2SWLPL]HG FUXVKHU VHOHFWLRQ IRU WKH FHPHQW LQGXVWU\ >RQOLQH@ SUHVHQWHG
WR$3&$&;;9,7HFKQLFDO&RQJUHVV$YDLODEOHIURPKWWSÀFHPRUJPXOWLPHGLD
WHFQLFRSUHVHQWDFLRQHV&7!>$FFHVVHG0D\@
Rowland,&$6HOHFWLRQRIURGPLOOVEDOOPLOOVSHEEOHPLOOVDQGUHJULQGPLOOVLQDesign and
Installation of Comminution CircuitsFKDSWHUSS²7KH6RFLHW\IRU0LQLQJ0HWDOOXUJ\
DQG([SORUDWLRQ/LWWOHWRQ
Rowland, & $ DQG .MRV ' 0  5RG DQG EDOO PLOOV LQ Mineral Processing Plant Design
FKDSWHU6RFLHW\IRU0LQLQJ0HWDOOXUJ\DQG([SORUDWLRQ/LWWOHWRQ
Telsmith,  Telsmith Mineral Processing Handbook >RQOLQH@ ÀUVW HGLWLRQ $YDLODEOH IURP
KWWSZZZWHOVPLWKFRPSURGXFWVPLQHUDOSURFHVVLQJKDQGERRN!
von Seebach, 0  +LJKSUHVVXUH JULQGLQJ UROOV LQ LQGXVWULDO DSSOLFDWLRQ LQ SME Annual
MeetingSUHSULQW²6RFLHW\IRU0LQLQJ0HWDOOXUJ\DQG([SORUDWLRQ/LWWOHWRQ
White,)0Fluid MechanicsWKLUGHGLWLRQS0F*UDZ+LOO

298 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

APPENDIX 16.1
Example of comminution circuit design criteria (Ausenco, 2015).

COMPANY NAME
PROJECT NAME
Project Number: PROCESS DESIGN CRITERIA
Revision:
Date:

Area Description Units Data Source Rev

000 GENERAL

Annual ore treatment Mt/y 4.0 1

Plant throughput per day, average t/d 11,000

001 PRODUCTION SCHEDULES

Primary Crushing
Operating days per year d/y 365 1
Shifts per day 2 1
Hours per shift h 12 1
Operational availability, annual % 68.5 1
Operating hours per year h/y 6,000 1

Grinding and Leaching/Adsorption

Operating days per year d/y 365 1
Shifts per day 2 1
Hours per shift h 12 1
Operational availability, annual % 92.5 1
Operating hours per year h/y 8,100 1

005 ORE CHARACTERISTICS

Type of mine - Open pit 1

Major lithologies and alteration states

Deposit 1 Name 1
Ore type Oxide/Primary 1
Principal lithology Grandiorite 1
Main gold mineralisation assemblages quartz-pyrite-magnetite-hematite 1
Deposit 2 Name 2 1
Ore type Transitional/Primary 1
Principal lithology Grandiorite 1
Main gold mineralisation assemblages minor sulfides (sphalerite, 1
galena, pyrite) in carbonates,
adularia and quartz
Ore Properties
3
Specific gravity of mineralised ore, for design volumetric calculations t/m 2.70 1
3
Specific gravity of mineralised ore, for design mass calculations t/m 2.80 8
3
Bulk density of ROM mineralised ore, for design volume calculations t/m 1.60 1
3
Bulk density of ROM mineralised ore, for design mass calculations t/m 1.70 8
3
Bulk density of crushed mineralised ore, for design volume calculations t/m 1.70 1
3
Bulk density of crushed mineralised ore, for design mass calculations t/m 1.80 8
Ore moisture content, for design % w/w 5 1

Comminution Handbook | Spectrum Series 21 299

 CHAPTER 16 t CIRCUIT DESIGN

Area Description Units Data Source Rev

Comminution Characteristics
Unconfined compressive strength - maximum for design MPa 190 2
Bond Crushing Work Index kWh/t 8.5 2
Bond Impact Work Index, design kWh/t 15.0 2
Bond Rod Mill Work Index, design kWh/t 16.5 2
Bond Ball Mill Work Index, design kWh/t 17.5 2
JK/SMC parameter A 80 2
JK/SMC parameter b 0.8 2
Drop Weight Index 4.5 2
Abrasion Index, range - 0.06 - 0.30 1
Abrasion Index, for design - 0.30 8

020 CRUSHING

Type of circuit Two stage, series crush 4

Crushing Capacity
Throughput - average t/h 670 1
Throughput - maximum for design t/h 775 8

ROM Ore
ROM ore maximum lump size, ) 100 mm 1,000 3
ROM ore typical lump size, ) 65 mm 210 3
Load capacity of haul truck to ROM pad t 92 1
Wheel loader to dump hopper, type - TBA 9
ROM dump pocket live capacity, number of truck loads - 2.0 4
ROM bin grizzly aperture mm x mm 800 4

Primary Vibrating Grizzly

Bar spacing mm 150 4
Material reporting to oversize, average % ROM feed 40 3
Material reporting to oversize, for design % ROM feed 65 8

Primary Crusher
Product size 3 99 mm 220 5
Product size 3 80 mm 130 5
Crusher closed side setting mm 150 5
Maximum capacity at closed side setting t/h 440 8

Secondary Vibrating Grizzly

Bar spacing mm 80 4
Material reporting to oversize, average % ROM feed 35 4
Material reporting to oversize, for design % ROM feed 50 8

Secondary Crusher
Product size 3 99 mm 125 5
Product size 3 80 mm 80 5
Crusher closed side setting mm 70 5
Maximum capacity at closed side setting t/h 340 8

Crushed Ore Stockpile

Live capacity, equivalent milling time h 22 1
Angle of repose, crushed ore degrees 36 1
Angle of drawdown degrees 65 1

Reclaim Feeders
Number of feeders - 2 1
Design capacity, % total SAG feed % 75 8

300 Comminution Handbook | Spectrum Series 21

 CHAPTER 16 t CIRCUIT DESIGN

Area Description Units Data Source Rev

030 GRINDING

Type of circuit
Single stage SAG mill, closed circuit 8
(provision future pebble recycle crusher)
Circuit Capacity
Grinding circuit throughput - nominal t/h 495 1
Grinding circuit throughput - maximum t/h 530 8

Grinding Circuit Product

Mill cyclone overflow 3 80 μm 125 2

SAG Mill
SAG mill feed ) 100 mm 150 4
SAG mill feed ) 80 mm 80 5
SAG mill discharge slurry density % w/w 72 8

SAG mill drive speed - maximum for design % crit 80 8

SAG mill drive speed - minimum % crit 60 8

SAG mill discharge grate aperture mm TBA 9

SAG mill circulating load (cyc U/F : cyc O/F), maximum for design % 700 8
SAG mill ball charge - operating range % v/v 10 - 15 4
SAG mill ball charge - maximum for structural design % v/v 18 8
SAG mill nominal ore and grinding media volumetric loading, operating % v/v 28 4

SAG mill media consumption - oxide/transition ore kg/kWh TBA 9

SAG mill media consumption - primary ore kg/kWh TBA 9

Cyclones
Cyclone underflow pulp density % w/w 75 4
Cyclone overflow pulp density % w/w 42 4

Code Source
1. Client supplied data
2. Testwork - metallurgical, process
3. Consultant report, data
4. Operating practice, industry standard
5. Vendor data
6. Engineering handbook, Regulatory Standards, Codes
7. Environmental
8. Recommended by Engineer
9. Not available. To be provided by Client, test work, others - as available.

Comminution Handbook | Spectrum Series 21 301

 HOME

Glossary

The following are terms common to comminution that may be found in the text, but this
list is by no means exhaustive. Further information can be obtained from sources such
as infomine.com or insidemetals.com and industry brochures such as that produced by
Russell Mineral Equipment or JKSimMet. Many terms may have different meanings
when used in non-mineral processing contexts.

80 per cent This indicates the mesh size of a milling screen of which 80 per cent of the
passing size – particles in a sample will pass through. F80 refers to a feed size distribution
F80 or P80 while P80 refers to product size distribution. See particle size distribution.

A, b, ta Parameters that represent breakage behaviour and can be used to relate

energy input to the amount of breakage for that particular ore. A and b are
generated by the JK Drop Weight Test (p 48), while ta is a measure of the
resistance of ore to abrasion.

air classifier $PDFKLQHLQZKLFKÀQHO\JURXQGSDUWLFOHVDUHVRUWHGE\DLUFXUUHQWVLQWR

fractions that settle at the same rate. Air cyclones are commonly used to
VHSDUDWHWKHFRDUVHDQGÀQHSDUWLFOHVLQWKHSURGXFWIURPDGU\PLOOVXFK
as a cement mill.

apex 7KHFRQHDWWKHEDVHRIDF\FORQHFODVVLÀHUWKURXJKZKLFKFRDUVHVROLGVDUH
discharged in accordance with its minimum cross-section.

appearance A mathematical representation of the size distribution of product particles

function after an impact, sometimes called the breakage pattern. This is a unique
characteristic of an ore or ore-dependent property.

aspect ratio The ratio of mill diameter to mill length.

autogenous A mill where all the grinding is done by the impact of the ore on itself,
grinding (AG) mill rather than introduced grinding media such as steel balls, pellets or rods.

ball milling A method of grinding a substance with or without liquid using media
such as balls or pebbles. It typically takes place in a rotating cylinder or
conical mill.

beneficiate Improve the grade of an ore by removing any impurities; to upgrade.

Bond Work Index An ore hardness index developed by Fred Bond in the 1950s used to
predict the breakage energy required for a certain size reduction. It can
be conducted using traditional rod mill (BRMWi) or ball mill (BBMWi)
circuits closed with hydrocyclones.

Comminution Handbook | Spectrum Series 21 303

 GLOSSARY

breakage pattern See appearance function.

breakage rate A machine-dependent parameter that represents mathematically the

amount of breakage energy available to ore in a particular comminution
device.

cement A substance that binds sedimentary rock to form a solid mass. Common
cements include carbonates, silica and certain iron oxides.

charge The steel balls and ore in a mill at a given time. It may also be the balls and
rocks remaining once the mill has stopped.

circulating load Material returned to a ball mill for further grinding because it is either
oversized or fails to meet the conditions for the next treatment stage.

classification The process of taking particles too small to be screened and using their
VL]HVKDSHRUGHQVLW\FKDUDFWHULVWLFVWRFDXVHGLIIHUHQWLDOVHWWOLQJLQDÁXLG
medium such as air or water.

closed circuit Part of the machine product is recycled to the machine for further
processing because the original processing was incomplete. In a
comminution process this usually refers to the coarse fraction from the
FODVVLÀHUSee open circuit.

coal Rock that contains more than 50 per cent carbonaceous material (weight-
wise) and more than 70 per cent in volume, including moisture. It is
formed from the hardening and condensing of living plant remains
similar to those found in peat.

coking coal Coal with a high carbon content and minimal impurities that can be
converted into coke for fuel. Generally, the best coking coals are 80 per cent
to 90 per cent carbon.

comminution The process of breaking material into smaller particles. In mineral

processing it refers to the action of crushing and grinding.

composite/locked Particles that have two or more minerals.

particles

concentrator An industrial plant where ore is divided into values and rejects, or an
instrument in such a plant where the ores are mechanically cleaned using
either water, air or gravity.

cone crusher A machine that breaks rock by compression between a gyrating cone and
a stationary inverted cone.

Cornish rolls 7KHVWDQGDUGÀQHJULQGLQJPDFKLQHXVHGIRU\HDUVIURPWKHPLGWK

century. It consisted of a pair of counter-rotating horizontal cylinders, one
À[HGDQGWKHRWKHUKHOGE\VSULQJV

crushing Reducing the size of ore into relatively coarse particles by stamps, crushers
or rolls. Crushers can be primary, secondary, tertiary and quaternary, with
primary crushers handling larger and coarser materials, and quaternary
FUXVKHUVKDQGOLQJWKHÀQHUSDUWLFOHVSee Symons crusher.

drilling A comminution process where narrow diameter holes in rock are formed
by high-speed cutting tools.

304 Comminution Handbook | Spectrum Series 21

 GLOSSARY

entrainment 7KHSURFHVVLQEHQHÀFLDWLRQE\ZKLFKSDUWLFOHVHQWHUDSURGXFWZLWKÁXLG
(such as air or water) without being subject to the mechanisms involved
LQUHGXFWLRQFODVVLÀFDWLRQRUFRQFHQWUDWLRQ

explosive A compound mixture capable of a quick chemical reaction that causes it to

explode; for example, gunpowder or dynamite.

F80 See 80 per cent passing size.

fines Small particles for grinding where size depends on its context; for instance,
1 mm is regarded as small for blasting and crushing, while 100 microns is
regarded as small for grinding.

free/liberated Ore fragments that contain a single mineral.

particles

grade The value of an ore sample according to the proportion of mineral it

contains that can be sold.

grain A single piece of mineral that may form part of a particle or larger body
of material.

grindability The ease in which a substance may be ground, assessed by its reduced
size and the power used.

grinding aid A chemical agent added to the charge in a ball mill or rod mill to assist the
grinding process.

grizzly A steel grate placed over a chute or pass to stop large pieces of rock from
passing.

gyratory crusher A machine that compresses and breaks ore using an offset crushing cone
DQGDÀ[HGFUXVKLQJWKURDW

hydraulic A tank that sorts ore pulp using a stream of hydraulic water at a steadily
classifier rising rate. Heavier, coarser particles fall and are discharged at the bottom,
while lighter particles rise up and are removed.

hydrocyclone A device in which particles are suspended in water and centrifugal forces
classifier are applied; coarse fragments are discharged from the vessel’s apex, while
ÀQHUSDUWLFOHVDUHGLVFKDUJHGZLWKWKHZDWHU

jaw crusher $ FUXVKLQJ PDFKLQH FRQVLVWLQJ RI D À[HG SODWH DQG DQ RVFLOODWLQJ SODWH
forming a tapered-jaw effect.

liberated mineral 0LQHUDOV WKDW KDYH KDG WKHLU VXUIDFHV H[SRVHG WR D VXIÀFLHQW GHJUHH WR
allow them to be affected by surface chemistry or other mineral recovery
processes.

lines In comminution this refers to a processing line, which may comprise an

autogenous grinding (AG) mill and ball mills.

Comminution Handbook | Spectrum Series 21 305

 GLOSSARY

mill A processing facility where ore is cleaned and concentrated before being
VHQWWRDUHÀQHU\RUVPHOWHU

mill availability The amount of time over a year when there is no hindrance to a mill’s
production. It can refer to a grinding mill’s availability but is more likely
to be about a concentrator plant’s availability.

milling In a power plant milling is grinding. In an ore plant it may refer to grinding
or to the entire circuit.

mineralisation In geology, mineralisation is the hydrothermal deposition of economically

important metals in the formation of orebodies. In general terms,
mineralisation refers to the ancient reactions of metals in solution,
interacting chemically with a sympathetic rock type to form a concentration
of metal within the rock.

mineralogy A branch of geology that deals in the study of minerals, including how they
DUHIRUPHGWKHLUSK\VLFDOSURSHUWLHVDQGXVHDQGKRZWKH\DUHFODVVLÀHG

open circuit When the entire product from a mineral processing machine moves to
the next downstream process rather than being recycled for further
processing. See closed circuit.

optimisation +DYLQJDVHWRIFRQGLWLRQVWRDFKLHYHPD[LPXPRSHUDWLRQDOHIÀFLHQF\

overgrinding Grinding an ore to a smaller particle size than necessary to liberate the
mineral. Not only does this waste energy, but can result in minerals being
GLIÀFXOWWRSURFHVV

P80 See 80 per cent passing size.

parameters Constants used in model equations that represent properties of the

process. In comminution, there are two groups: unit-dependent and ore-
dependent parameters.

particle A single solid piece of ore of no determinable size or structure.

particle size $OVRNQRZQDVVL]HGLVWULEXWLRQLWGHÀQHVWKHSHUFHQWDJHVW\SLFDOO\E\

distribution (PSD) mass) of particles of different sizes present in an ore sample. The sizes
PHDVXUHGUDQJHIURPFRDUVHWRÀQHDQGWKHSHUFHQWDJHUHSUHVHQWLQJWKH
SDUWLFOHV VPDOOHU WKDQ WKH ÀQHVW VL]H LV  PLQXV WKH VXP RI PHDVXUHG
fractions. The 80 per cent passing size is a measure of the coarseness of
the product.

pebbles Hard and well-rounded small stones used as grinding media.

recovery The amount of value gained from an ore, which in turn measures the
PLOO·VHIÀFLHQF\

reduction ratio Calculates the ratio of F80 to P80, thereby determining how size reduction
has taken place in processing.

rod milling Grinding that employs rods of a small diameter that run the length of the
mill.

306 Comminution Handbook | Spectrum Series 21

 GLOSSARY

rolls crusher A machine comprising one or two rollers made of steel or iron that are
placed at a certain distance apart so that thick substances are crushed as
they pass through.

run-of-mine ore Untreated ore straight from the mine before undergoing processing of
any sort.

scats The particles ejected from a tumbling mill, which may be ore or broken
grinding media.

screens :LUHPHVKSDUWLWLRQVZLWKRSHQLQJVRIDVSHFLÀFVL]HGHVLJQHGIRUJUDGLQJ
particles.

semi-autogenous A combination of AG milling (using rock grinding media) and ball milling,
grinding (SAG) with the aim of improving breakage rates.
mill

size distribution See particle size distribution.

sizer A crusher that uses a multistage system to break the rock down to the
required size. It then separates the undersized rock from the oversized
rock, the latter of which is fed back into the system.

slime 2UHWKDWLVSURFHVVHGWRVXFKDÀQHSRZGHUDVWREHVXVSHQGHGLQZDWHU
forming a kind of thin mud.

stirred milling *ULQGLQJ WKDW WDNHV SODFH ZKHUH WKH F\OLQGHU LV À[HG DQG WKH PHGLD LV
moved by stirrers. These mills may be vertical or horizontal.

survey The process of collecting samples and information around an operating

plant in a systematic way so as to document that process.

Symons crusher $ PRGLÀHG J\UDWRU\ FUXVKHU ZLWK D GRZQZDUG ÁDULQJ ERZO ZLWKLQ
which features a conical crushing head. The main gyrating shaft is driven
by bevel gears. There are two types of Symons crushers: the Standard
Symons is designed to crush coarser than the Shorthead Symons.

texture The physical appearance of a rock, including the size and form of its
mineral grains.

tumbling mill A horizontal cylindrical mill in which contents are tumbled and particles
are broken while it rotates.

unit A single component of a comminution circuit, such as a ball mill or

hydrocyclone.

vortex finder $ WXEH SURMHFWLQJ LQWR D K\GURF\FORQH·V YRUWH[ WKURXJK ZKLFK ÀQHU RU
lighter pulp is removed.

work index See Bond Work Index.

X-ray diffraction A technique for identifying minerals using a multitude of X-rays at

repeated angles. This establishes a unique pattern of diffraction for each
mineral structure.

Comminution Handbook | Spectrum Series 21 307

 HOME

Index

abrasion resistance, 44
abrasiveness, 44
actuators, 230
advanced process control (APC) systems, 241–3, 259–61, 263
DLUFODVVLÀHUV159–64
conical, 160
dynamic, 161–4
static, 159–61
zigzag, 160–1
DLUMHWPLOODQGFODVVLÀHUZKHHOV110–12
Airmet Dusttrak II 8530, 36
$WULWDPLOOVFRPELQHGKDPPHUDQGDWWULWLRQPLOOV103–4
AudioMill
PLOOVRXQGFRQWUROZLWK262
system, 260–1, 263
Ausgrind, 209–12, 272–6
autogenous/semi-autogenous grinding mills, 68–71, 281–2
mill liners, 135
automated mineralogy systems (mineral liberation), 18–19
averaging pitot, 229

EDOOFKDUJHVL]HLQVHPLDXWRJHQRXVJULQGLQJPLOOV68–9
ball mill, 61
for cement, 62–4
circuit, 200, 202–3, 204, 274–6, 277, 278
and circuit design, 282–4
circuit layout, 293–7
ÀWWLQJ219–23
open-circuit single-stage, 200, 203
for ores, 66–8
mill liners, 135–6
VSHFLÀFHQHUJ\204
EDOOPLOO²K\GURF\FORQHFLUFXLWÀWWLQJH[DPSOHIRU218–26
ball milling, single-stage, 173, 175–6
ball/tube mills for coal, 65–6
EHDWHUZKHHOPLOOV104–6

Comminution Handbook | Spectrum Series 21 309

 INDEX

EHGEUHDNDJHFUXVKHUV85–7
%ODLQHPHWKRG41
%RQG%DOO0LOO:RUN,QGH[%%0:L5, 6, 46, 50, 52, 54, 55
%RQG)UHG&KHVWHU44–5, 198
%RQGJULQGDELOLWLHVSRZHUFDOFXODWLRQVXVLQJ197–212
Bond grindability test, 44–6
%RQG5RG0LOO:RUN,QGH[%50:L46, 272
%RQG:RUN,QGH[WHVWUHVXOWV272
%RQG:RUN,QGH[YDOXHV57–8
bore core testing (coal), 179–80
EUHDNDJHGHÀQLWLRQV43–4
EUHDNDJHPHFKDQLVPV7
%UXQDXHU²(PPHWW²7HOOHUWKHRU\40–1
EXONODERUDWRU\WHFKQLTXHVPLQHUDOOLEHUDWLRQ17
burner management system (BMS), 185

cement, ball mills for, 62–4

cement clinker, comminution circuits for, 172–7
FORVHGFLUFXLWSUHJULQGLQJZLWKKLJKSUHVVXUHJULQGLQJUROOVIROORZHGE\VLQJOHVWDJH
ball milling, 175–6
control, 188–9
ÀQLVKJULQGLQJXVLQJKLJKSUHVVXUHJULQGLQJUROOV176
ÀQLVKJULQGLQJXVLQJDYHUWLFDOUROOHUPLOO177
K\EULGJULQGLQJ173–5
RSHQFLUFXLWSUHJULQGLQJZLWKKLJKSUHVVXUHJULQGLQJUROOVIROORZHGE\VLQJOHVWDJH
ball milling, 173
VHPLÀQLVKJULQGLQJ175
single-stage ball milling, 173, 175–6
cement mills, liners in, 137–8
FKXWHVDPSOLQJ28
Cilas Analysette 28 ImageSizer, 38
Cilas products, 33
circuit design, 265–301
$XVJULQGZRUNHGH[DPSOHV272–6
codes, 268–9
comminution design criteria, 269
comminution energy calculations, 272–8
FUXVKLQJDQGVFUHHQLQJHTXLSPHQW278–81
design criteria, 267–9
HTXLSPHQWVL]LQJDQGVHOHFWLRQ278–91
ÁRZVKHHWRSWLRQV269–71
function, 268
JULQGLQJDQGFODVVLÀFDWLRQHTXLSPHQW281–8
KLJKSUHVVXUHJULQGLQJUROOV289–91
KLJKSUHVVXUHJULQGLQJUROOV²EDOOPLOOFLUFXLWOD\RXW293–7
plant layout, 291–8

310 Comminution Handbook | Spectrum Series 21

 INDEX

and process development, 265–6

SURMHFWGHÀQLWLRQ265
purpose, 268
SAB/SABC plant layout, 292–3
VDPSOHVHOHFWLRQDQGWHVWZRUN267
60&ZRUNHGH[DPSOHV276–8
FODVVLÀFDWLRQ147
and circuit design, 284
HTXLSPHQW281–8
FODVVLÀHUFDOFXODWLRQV²PDVVEDODQFLQJDQGFLUFXODWLRQORDG191–7
FDOFXODWLRQRIDFWXDODQGFRUUHFWHGSDUWLWLRQLQJFRHIÀFLHQWV195–6
FRPSDULVRQRIWKHDFWXDODQGLGHDOHIÀFLHQF\FXUYHV196–7
FODVVLÀHUZKHHOV110–12
FODVVLÀHUV145–65
air, 159–63
catalogues, 165
FODVVLÀFDWLRQ147
FRPSDULQJKLJKIUHTXHQF\VFUHHQVDQGK\GURF\FORQHV157–9
HIÀFLHQF\FXUYHPRGHOIRU217–18
ÀQHVFUHHQV153–4
K\GURF\FORQHV150–3
RWKHU150
rake and spiral, 149
size separation in comminution processes, 145–6
trommels, 147–9
vibrating screens, 154–7
closed grinding circuits
FRPSDULQJKLJKIUHTXHQF\VFUHHQVDQGK\GURF\FORQHV157–9
closed-circuit single-stage ball mill, 200, 203
coal
ball/tube mills for, 65–6
FUXVKLQJHTXLSPHQW177–9
JULQGLQJFLUFXLWVLQSRZHUSODQWV181
rotary breakers for, 71–3
FRDOPLOOVLQSRZHUVWDWLRQVFRQWURORI184–7
FRQWUROZLWKLQWKHPLOOLQJV\VWHP186–7
coal preparation plants, comminution for, 177–84
bore core testing, 179–80
FRDOFUXVKLQJHTXLSPHQW177–9
FRDOJULQGLQJFLUFXLWVLQSRZHUSODQWV181
control, 187–8
RYHUYLHZ177
SRZHUSODQWPLOOLQJV\VWHPV181–4
testing of core breakage, 180

Comminution Handbook | Spectrum Series 21 311

 INDEX

comminution, 1–10
background, 2–3
EUHDNDJHPHFKDQLVPV7
FODVVLÀFDWLRQPHFKDQLVPV7–8
FRQWH[W1–2
design criteria, 269
energy calculations, 272
fundamentals, 3–6
grindability, 6
K\GUDXOLFDQGSQHXPDWLFVHSDUDWLRQ8
introduction, 1
PDFKLQHV3, 8–10
principles, 4–5
screening, 7–8
size separation in comminution processes, 145–6
in stirred mills, 108–13
WKHRU\5–6
comminution circuits
cement clinker, 172–7
cement plants, control of, 188–9
coal preparation plants, control of, 187–8
ÁRWDWLRQRIORZJUDGHFRSSHUVXOÀGHRUHV172
ores, 167–72
SRO\PHWDOOLFVXOÀGHÁRWDWLRQ171–2
comminution circuits, modelling, 215–26
EDOOPLOOÀWWLQJ219–23
HIÀFLHQF\FXUYHPRGHOIRUFODVVLÀHUV217–18
ÀWWLQJH[DPSOHIRUEDOOPLOO²K\GURF\FORQHFLUFXLW218–26
K\GURF\FORQHÀWWLQJ223–6
SHUIHFWPL[LQJEDOOPLOOPRGHO215–17
comminution for coal preparation plants, 177–84
comminution processes, modelling mineral liberation in, 19–21
KHXULVWLFPRGHOV19
PDWKHPDWLFDOPRGHOV19–21
Compañía Minera Antamina 170–1
competence, 43–4
FRPSUHVVLRQPDFKLQHV79–98
EHGEUHDNDJHFUXVKHUV85–7
FRQYHQWLRQDOFRQHFUXVKHUV82–4
energy used in different circuits, 96–7
J\UDWRU\FUXVKHU79–80, 81–2
KLJKFDSDFLW\FUXVKLQJFLUFXLWV84–5
KLJKSUHVVXUHJULQGLQJUROOV+3*587, 90–3
+RURPLOOV95–6
MDZFUXVKHU80–2

312 Comminution Handbook | Spectrum Series 21

 INDEX

mineral sizers, 88–9

SULPDU\FUXVKHUV79–82
5KRGD[FUXVKHU85–6
roller mills, 89–90
UROOVFUXVKHU87
vertical roller mills (VRMs), 93–5
9LEURFRQHFUXVKHU86–7
compression test, 56–7
FRQHFUXVKHUVFRQYHQWLRQDO82–4
FRQLFDODLUFODVVLÀHUV160
FRQLQJDQGTXDUWHULQJ29
FRQWURORIDFUXVKHU238–9
control loops, regulatory
ore feed rate, 234–5
ZDWHUIHHGUDWH235–6
control systems, process, 231–3
control systems case studies, 245–63
JULQGLQJFLUFXLWFRQWURO1HZPRQW$KDIRJROGPLQH*KDQD245, 246, 257–63
JULQGLQJFLUFXLWFRQWURO3KX.KDPPLQH/DRV245, 246–56
FRQYHQWLRQDOFRQHFUXVKHUV82–4
FRQYHQWLRQDOVWDJHFUXVK275–6, 278
FRSSHUVXOÀGHRUHVFRPPLQXWLRQFLUFXLWVIRUÁRWDWLRQRIORZJUDGH172
Coriolis meter, 229
Coulter Counter, 36
&RXOWHU/632
&UXVKHU0DSSHU140
FUXVKHUVFUXVKLQJ4
and circuit design, 278–9
control of, 238–9
3KX.KDPPLQH/DRV253
YHUWLFDOVKDIWLPSDFW99–101
FUXVKLQJFLUFXLWV
energy used in different, 96–7
KLJKFDSDFLW\84–5
FUXVKLQJHTXLSPHQWFRDO177–9
cylinder actuators, 230

DBC (Don Burgess Consulting), 207–8

'HUULFN+:155
GLDSKUDJPDFWXDWRUV230
distributed control system (DCS), 232
'URS:HLJKW7HVW48–51
dry screening, 156–7
DSM screens, 153–4
G\QDPLFDLUFODVVLÀHUV161–4

Comminution Handbook | Spectrum Series 21 313

 INDEX

HIÀFLHQF\FXUYHPRGHOIRUFODVVLÀHUV217–18
electrical actuators, 230
electron microscopy, 38–9
energy-intensive stirred agitator (EiSA) mill, 115–16

IHHGJUDGDWLRQFRQWDLQLQJH[FHVVLYHÀQHV200, 204
ÀQHVFUHHQV153–4
ÀQLVKJULQGLQJXVLQJKLJKSUHVVXUHJULQGLQJUROOV176
ÀQLVKJULQGLQJXVLQJDYHUWLFDOUROOHUPLOO177
ÁRZVKHHWRSWLRQVFLUFXLWGHVLJQ269–71
)/6PLGWKJ\UDWRU\FUXVKHU79–80, 82
)ULWVFK$QDO\VHWWHPRGHOV, 33

gear and pinion (tumbling mills), 76

gearless drive (tumbling mills), 76
grindability, 6, 43, 198
grinding, 4, 107–8
cost of, 126
HTXLSPHQW281–8
K\EULG173–5
to nanometre size, 112–13
VHPLÀQLVK175
grinding balls, 74
grinding circuit control, 234–7
logic-based control, 237
1HZPRQW$KDIRJROGPLQH*KDQD245, 246, 257–63
3KX.KDPPLQH/DRV245, 246–56
regulatory control loops – ore feed rate, 234–5
UHJXODWRU\FRQWUROORRSV²ZDWHUIHHGUDWH235–6
JULQGLQJFLUFXLWVLQFRDOSRZHUSODQWV181
grinding circuits, closed
FRPSDULQJKLJKIUHTXHQF\VFUHHQVDQGK\GURF\FORQHV157–9
grinding mills
autogenous/semi-autogenous, 68–70
JULQGLQJUROOVKLJKSUHVVXUH+3*590–3
*ULQG3RZHU206
J\UDWRU\FUXVKHU79–80, 81–2

KDPPHUPLOOV101–3
+DUGJURYH*ULQGDELOLW\,QGH[+*,WHVW46–7
+DUGJURYH5DOSK46–7
KDUGQHVV43
+DYHUDQG%RHFNHU&3$38
+LDF5R\FRSDUWLFOHFRXQWHU35
KLJKFDSDFLW\FUXVKLQJFLUFXLWV84–5
KLJKFKURPLXPZKLWHLURQPLOOOLQHUV127–8

314 Comminution Handbook | Spectrum Series 21

 INDEX

KLJKIUHTXHQF\VFUHHQV
DQGFRPSDULQJK\GURF\FORQHVLQFORVHGJULQGLQJFLUFXLWV157–9
KLJKSUHVVXUHJULQGLQJUROOV+3*587, 90–3, 274–5, 277, 289–91, 293–7
KLJKVSHHGLPSDFWPLOOV99–106
$WULWDPLOOVFRPELQHGKDPPHUDQGDWWULWLRQPLOOV103–4
EHDWHUZKHHOPLOOV104–6
KDPPHUPLOOV101–3
YHUWLFDOVKDIWLPSDFWFUXVKHUV99–101
+,*PLOO114–15
KRUL]RQWDOVWLUUHGPLOOV118–20
KRUL]RQWDOVWLUUHGZHW$+0PLOO121
+RURPLOOV95–6
+RVRNDZD$OSLQHGU\YHUWLFDOVWLUUHG$75PLOO123–4
+RVRNDZD$OSLQHÀQHJULQGLQJPLOOV120–4
KRUL]RQWDOVWLUUHGZHW$+0PLOO121
ZHWYHUWLFDOVWLUUHG$15PLOO121–3
K\EULGJULQGLQJFHPHQW173–5
K\GUDXOLFDQGSQHXPDWLFVHSDUDWLRQ8
K\GURF\FORQHFLUFXLWÀWWLQJH[DPSOHIRU218–26
K\GURF\FORQHÀWWLQJ223–5
K\GURF\FORQHPHWKRGV34–5
K\GURF\FORQHV34, 150–3, 284–8
FRPSDULQJKLJKIUHTXHQF\VFUHHQVLQFORVHGJULQGLQJFLUFXLWV157–9

IsaMill, 118–20

MDZFUXVKHU80–2
-.'URS:HLJKW7HVW-.':748–51, 52, 206, 272
-XOLXV.UXWWVFKQLWW0LQHUDO5HVHDUFK&HQWUH-.05&206–7, 272

laser diffraction particle size analysis, 31–3

basic principles, 31–2
instrument comparison, 32–3
liners see mill liners

0DF3KHUVRQ$5206
PDJQHWLFÁRZPHWHU229
0DOYHUQ,QVWUXPHQWV0RUSKRORJL37–8
Malvern Instruments MS2000, 32
mass balancing, 191–7
Matec Applied Sciences (MAS), 37
metallurgist, process control, 239–40
0HWVRFRQHFUXVKHU82
0HWVR3URFHVV7HFKQRORJ\DQG,QQRYDWLRQ37,168
Micromeritics Saturn Digisizer, 33
0LFURPHULWLFV6HGL*UDSK34
PLOOKHDGVWXPEOLQJPLOOV75

Comminution Handbook | Spectrum Series 21 315

 INDEX

mill liners, 125–42

'WUDMHFWRU\RUWKURZSUHGLFWLRQ134
acoustic monitoring, 140–1
autogenous grinding mills, 135
ball mills, 135–6
broken/damaged, 130
cement mills, 137–8
condition monitoring, 139–41
damaged rubber backing, 130
designing, 133–4
GLUHFWPHDVXUHPHQWRIOLQHUZHDU139–41
GLVFKDUJHOLQHUV133
H[FHVVLYHOLQHUZHDU130
LQVHUWLRQRIQHZOLQHUV142–3
LQVSHFWLRQVZHDUPRQLWRULQJDQGRQJRLQJRSWLPLVDWLRQ138–9
integral liners – Metso Magaliner, 132
lifter bars, 132–3
loose liners, 132
materials of construction, 127–9
for mill types, 135–6
QHZOLQHUVLQVHUWLRQRI142–3
RSHUDWLRQDOFKDQJHVGXULQJOLQHUOLIHWLPH130
SDFNLQJEHWZHHQOLIWHUV136
plate liners, 133
SRWHQWLDOLVVXHVZLWK129–30
relining mills, 141–3
removable lifters, 131
UHPRYDORIZRUQOLQHUV142
rod mills, 136
role in tumbling mills, 125–6
rubber, 128, 129, 132
VHPLDXWRJHQRXVJULQGLQJ6$*PLOOV127, 135
scrubbers, 136
solid or integral, 131
steel, 128, 129, 130–1
steel versus rubber, 129
types of, 130–3
uni-directional and bi-directional liners, 131
variations for ore mills, 126–7
vertical pins, 140
ZDYHOLQHUV131
ZRUQOLQHUVUHPRYDORI142
PLOOSRZHUFDOFXODWLRQVXVLQJWKH%RQGPHWKRGDQGHIÀFLHQF\IDFWRUV200–12
milling circuit calculations, 191–212
DGGLWLRQDOPHWKRGV206–12

316 Comminution Handbook | Spectrum Series 21

 INDEX

FDOFXODWLRQRIDFWXDODQGFRUUHFWHGSDUWLWLRQFRHIÀFLHQWV195–6
FDOFXODWLRQRIPLOOSRZHU205–6
FODVVLÀHUFDOFXODWLRQV²PDVVEDODQFLQJDQGFLUFXODWLRQORDG191–7
FRPSDULVRQRIWKHDFWXDODQGLGHDOHIÀFLHQF\FXUYH196–7
SRZHUFDOFXODWLRQVXVLQJ%RQGJULQGDELOLWLHV197–212
PLOOLQJV\VWHPFRQWUROZLWKLQ186–7
PLOOLQJV\VWHPVSRZHUSODQW181–4
MillMapper, 139–40
Mine to Mill, 168–9
mineral liberation, 11–22
automated mineralogy systems, 18–19
EXONODERUDWRU\WHFKQLTXHV17
GHÀQLWLRQ11
during breakage, 12
KHXULVWLFPRGHOV19
KRZLWKDSSHQV14–17
OLEHUDWLRQVHSDUDWLRQÁRZVKHHWVKRZLQJPXOWLSOHVWDJHV16
PDWKHPDWLFDOPRGHOV19–21
measuring, 17–19
modelling in comminution processes, 19–21
optical microscopy, 17
in separation processes, 21–2
terminology, 12
;UD\WRPRJUDSK\19
ZK\LWLVQHFHVVDU\12–14
mineral sizers, 88–9
PL[LQJEDOOPLOOPRGHOSHUIHFWFRPPLQXWLRQFLUFXLWV215–17
0RUUHOO·VPHWKRG208–9

1HZPRQW$KDIRJROGPLQH*KDQDJULQGLQJFLUFXLWFRQWURO245, 246, 257–63

$XGLR0LOOPLOOVRXQGFRQWUROZLWK262
AudioMill system, 260–1, 263
ball mill control, 262–3
FRPSRQHQWVRIWKHDGYDQFHGSURFHVVFRQWURO$3&V\VWHP259–61
control strategy, 261–3
conventional controls, 258–9
disturbances (advanced process control system), 259
grinding circuit, 257–9
present status of advanced process control (APC) system, 263
project implementation, 263
soft sensor in OCS, 261
system performance, 263
9LVLR5RFNIHHGHUVFRQWUROZLWK262
VisioRock system, 260, 263

Comminution Handbook | Spectrum Series 21 317

 INDEX

20&2UZD\0LQHUDO&RQVXOWDQWV207

open-circuit single-stage ball mill, 200, 203
optical microscopy (mineral liberation), 17
optimising blasting, 169–70
RUHFRPPLQXWLRQPHDVXUHPHQWWHFKQLTXHV43–59
Bond grindability test, 44–6
%RQG:RUN,QGH[YDOXHV57–8
EUHDNDJHGHÀQLWLRQV43–4
compression test, 56–7
'URS:HLJKW7HVW48–51
+DUGJURYH*ULQGDELOLW\,QGH[+*,WHVW46–7
6$*YDULDELOLW\WHVW69755–6
6$*'HVLJQWHVW54–5
SMC Test, 51–3
SPI test, 53–4
6SOLW+RSNLQVRQSUHVVXUHEDU6+3%WHVW58–9
Zeisel test, 47–8
ore feed rate (control loops), 234–5
ore mills, variations in liners for, 126–7
ores, ball mills for, 66–8
ores, comminution circuits for, 167–72
case study – Compañía Minera Antamina, 170–1
ORZJUDGHFRSSHUVXOÀGHRUHV172
Mine to Mill, 168–9
optimising blasting, 169–70
RYHUYLHZ167
SRO\PHWDOOLFVXOÀGHÁRWDWLRQ171–2
RULÀFHSODWH229

3DQ$XVW/LPLWHG246–56
particle counting, 35–6
SDUWLFOHPHDVXUHPHQWWHFKQLTXHV25–41
electron microscopy, 38–9
K\GURF\FORQHPHWKRGV34–5
laser diffraction particle size analysis, 31–3
particle counting, 35–6
SKRWRPLFURVFRS\RSWLFDO37–8
sampling, 25–30
VHGLPHQWDWLRQVHGLJUDSKDQGFROXPQVHWWOLQJ33–4
surface area analysis, 40–1
VXUIDFHFKDUJHWHFKQLTXHV36–7
ZHWGU\VLHYLQJ30–1
X-ray diffraction, 39–40
SHDUOLWLFFKURPLXPPRO\EGHQXP&U0RVWHHOPLOOOLQHUV127
SHUIHFWPL[LQJEDOOPLOOPRGHOFRPPLQXWLRQFLUFXLWV215–17

318 Comminution Handbook | Spectrum Series 21

 INDEX

SKRWRPLFURVFRS\RSWLFDO37–8
EDVLFWHFKQLTXH37
manufacturers, 37–8
3KX.KDPPLQH/DRVJULQGLQJFLUFXLWFRQWURO245, 246–56
control rooms, 252–3
FUXVKLQJ253–4
HTXLSPHQW251–2
ÁHHWFRQWURODQGUHSRUWLQJ253
grinding control, 254–6
introduction, 246–7
operational control design, 251
RSHUDWLRQDOV\VWHPVRYHUYLHZ253
operations management strategy, 250–6
SURFHVVFRQWUROVWUDWHJ\DQGDSSURDFK247–50
reliability, 252
system design, 250–1
plant layout, 291–7
ball mill circuit layout, 293–7
KLJKSUHVVXUHJULQGLQJUROOV293–7
RYHUYLHZDQGREMHFWLYHV291–2
SAB/SABC, 292–3
pneumatic separation, 8
Polysius double rotator mills, 62
SRZHUFDOFXODWLRQVXVLQJ%RQGJULQGDELOLWLHV197–212
HIÀFLHQF\IDFWRUV199–200, 201–2
PLOOSRZHUFDOFXODWLRQVXVLQJWKH%RQGPHWKRGDQGHIÀFLHQF\IDFWRUV200–12
SRZHUSODQWPLOOLQJV\VWHPV181–4
SRZHUVWDWLRQVFRQWURORIFRDOPLOOVLQ184–7
SULPDU\FUXVKHUV79–82
J\UDWRU\FUXVKHU79–80, 81–2
MDZFUXVKHU80–2
process control, 227–43
advanced process control (APC) systems, 241–3, 259–61, 263
control elements, 229–30
FRQWURORIDFUXVKHU238–9
grinding circuit control, 234–7
PHDVXUHPHQWDQGFRQWUROFRVWV²¶UXOHRIWKXPE·233–4
metallurgist, role of, 239–40
model-based control, 243
objectives, 239
SURFHVVFRQWUROWHFKQLTXHVDQGVNLOOV240–1
process measurement, 228–9
UHVSRQVHRIWKH6$*PLOOWRFKDQJHLQIHHGUDWH237–8
UXOHEDVHGH[SHUWV\VWHPVIX]]\DQGFULVS241–3
VWUDWHJ\DQGDSSURDFK3KX.KDPPLQH/DRV247–50

Comminution Handbook | Spectrum Series 21 319

 INDEX

WHFKQLTXHVDQGVNLOOV240–1
transmission, 230–1
process control systems, 231–3
distributed control system (DCS), 232
1HZPRQW$KDIRJROGPLQH*KDQD245, 246, 257–63
3KX.KDPPLQH/DRV245, 246–56
SURJUDPPDEOHORJLFFRQWUROOHU3/&233
process development and circuit design, 265–6
SURFHVVÁRZGLDJUDP3)'227
process measurement, 228–9
ÁRZPHDVXUHPHQW229
SURJUDPPDEOHORJLFFRQWUROOHU3/&233
proportional-integral-derivative (PID) control, 188

UDNHDQGVSLUDOFODVVLÀHUV149
5HÁX[&ODVVLÀHU150
representative subsampling, 27–30
FKXWHVDPSOLQJ28
FRQLQJDQGTXDUWHULQJ29
URWDU\ULIÁHU29–30
scoop sampling, 28
spear sampling, 28–9
5HWVFK&DPVL]HU37
5HWVFK+RULED/$33
5KRGD[FUXVKHU85–6
rod mill circuit, 200, 202–3
rod mill in open-circuit, 200, 201–2
rod mills, 71–2
and circuit design, 282–4
mill liners, 136
vertical roller mills (VRMs), 93–5
roller mills, 89–90
UROOVFUXVKHU87
rotary breakers for coal, 71–3
URWDU\ULIÁHU29–30
5XPSI+DQV108

SAB/SABC circuit, 272–4, 276–7, 292–3

6$*PLOOVseeVHPLDXWRJHQRXVJULQGLQJPLOOV6$*PLOOV
6$*3RZHU,QGH[63,53–4, 207
6$*YDULDELOLW\WHVW69755–6
6$*'HVLJQWHVW54–5, 207
sampling, 25–30
FKXWHVDPSOLQJ28
representative subsampling, 27–30
scoop sampling, 28
spear sampling, 28–9

320 Comminution Handbook | Spectrum Series 21

 INDEX

scoop sampling, 28
screening, 7–8, 154–5
dry, 156–7
screens
FKDUDFWHULVWLFV156–7
and circuit design, 279–81
evolution of, 154–5
ÀQH153–4
KLJKIUHTXHQF\155–6
KLJKIUHTXHQF\FRPSDUHGZLWKK\GURF\FORQHV157–9
Stack Sizer, 155–6
vibrating, 154–7
ZLUHPHVK155
see also sieving
scrubbers, 73–4
6HGL*UDSK34
VHGLPHQWDWLRQVHGLJUDSKDQGFROXPQVHWWOLQJ33–4
basic principles, 33
0LFURPHULWLFV6HGL*UDSK34
VHPLDXWRJHQRXVJULQGLQJPLOOV6$*PLOOV1, 54, 68–71, 281–2
EDOOFKDUJH68–9
ball size, 69
FKDUDFWHULVWLFV69–71
circuits, 200, 204
mill liners, 127, 135
optimised blasting, 169–70
UHVSRQVHWRFKDQJHIHHGUDWH237–8
VSHFLÀFHQHUJ\205
VHPLÀQLVKJULQGLQJ175
separation processes, mineral liberation in, 21–2
settling column, 34
VKHOOVWXPEOLQJPLOOV76
sieve bends, 153–4
sieving
VWDLQOHVVVWHHOYHUVXVEUDVVYHUVXVQ\ORQPHVKVLHYHV31
WHFKQLTXH301
ZHWGU\30–1
see also screens
single-stage ball milling (cement), 173, 175–6
size separation in comminution processes, 145–6
SMC Test, 51–3, 276
60&ZRUNHGH[DPSOHVFLUFXLWGHVLJQ276–8
spear sampling, 28–9
SPI test, 53–4
VSLUDOFODVVLÀHUV149

Comminution Handbook | Spectrum Series 21 321

 INDEX

6SOLW+RSNLQVRQSUHVVXUHEDU6+3%WHVW58–9

Stack Sizer screen, 155–6
VWDWLFDLUFODVVLÀHUV159–61
stirred media detritor, 116–17
stirred mills, 107–24
catalogues, 124
comminution in, 108–13
KRUL]RQWDO118–20
+RVRNDZD$OSLQHÀQHJULQGLQJPLOOV120–4
vertical, 113–17
VWUHVVHQHUJ\VWUHVVUDWHDQGVSHFLÀFHQHUJ\109–10
VXSHUKHDWHGVWUHDPMHWPLOOV112
surface area analysis, 40–1
%ODLQHPHWKRG41
%UXQDXHU²(PPHWW²7HOOHUWKHRU\40–1
VXUIDFHFKDUJHWHFKQLTXHV36–7
Matec Applied Sciences (MAS), 37
Zetasizer, 37

7K\VVHQ.UXSS'RXEOH5RWDWRU62
WRZHUPLOOV113
transmission, 230–1
trommels, 147–9
trunnions (tumbling mills), 76
tumbling mills, 61–76, 107
autogenous/semi-autogenous grinding mills, 68–71
ball mills for cement, 62–4
ball mills for ores, 66–8
ball/tube mills for coal, 65–6
building large, 75–6
components, 75–6
grinding balls, 74
rod mills, 71–2
role of liners in, 125–6
rotary breakers for coal, 71–3
scrubbers, 73–4
Tyler, W S, 155

valves, 230
vertical roller mills (VRMs), 93–5
YHUWLFDOVKDIWLPSDFWFUXVKHUV99–101
vertical stirred mills, 113–17
energy-intensive stirred agitator (EiSA) mill, 115–16
+,*PLOO114–15
+RVRNDZD$OSLQHGU\YHUWLFDOVWLUUHG$75PLOO123–4
stirred media detritor, 116–17

322 Comminution Handbook | Spectrum Series 21

 INDEX

WRZHUPLOOV113
Vertimill, 113–14
ZHWYHUWLFDOVWLUUHG$15PLOO121–3
Vertimill, 113–14
vibrating screens, 154–7
9LEURFRQHFUXVKHU86–7
VisioRock
IHHGHUVFRQWUROZLWK262
system, 260, 263
YRUWH[ÁRZPHWHU229

Warman Cyclosizer, 35
ZDVWHEUHDNDJH4
ZDWHUIHHGUDWHFRQWUROORRSV235–6
ZHWGU\VLHYLQJ30–1

X-ray diffraction, 39–40

;UD\WRPRJUDSK\PLQHUDOOLEHUDWLRQ19

Zeisel test, 47–8

Zetasizer, 37
]LJ]DJDLUFODVVLÀHUV160–1

Comminution Handbook | Spectrum Series 21 323

 Publications of The Australasian Institute of Mining and Metallurgy
SPECTRUM SERIES

# * Spectrum title Authors/Editors Year

1 Making the Mount Isa Mine, 1923 – 1933 Don Berkman 1996
2 History of Drilling Graham McGoggan 1996
3 * The Cobar Mineral Field – A 1996 Perspective Warren Cook, Andrew Ford, 1996
Julian McDermott,
Peter Standish, Craig Stegman
and Therese Stegman
4 Towards 2000 – Resource to Reserve Inputs Seminar, Melbourne, Vic 1997
5 Towards 2000 – National Conference on Ironmaking Resources and 1997
Reserves Estimation, Perth, WA
6 Towards 2000 – The Resource Database Towards 2000, Wollongong, 1997
NSW
7 Towards 2000 – Ore Reserves and Finance, Sydney, NSW 1997
8 Towards 2000 – Assessment of Reserves in Low Rank Coals, Morwell, 1997
Vic
9 Towards 2000 – Ore Reserve Reconciliation Workshop, Darwin, NT 1997
10 Towards 2000 – Gold and Nickel Ore Reserve Estimation Practice 1998
Seminar
11 Towards 2000 – Resource/Reserves Estimation Practice in the Central 1999
West New South Wales Mining Industry, Cobar, NSW
12 Field Guide for Geoscientists and Technicians
First Edition 2004
Second Edition 2007
Third Edition 2010
13 The Extractive Metallurgy of Zinc Roderick J Sinclair 2005
14 Orebody Modelling and Strategic Mine Planning – Uncertainty and
Risk Management Models
First Edition Ed: Roussos Dimitrakopoulos 2005
Second Edition Ed: Roussos Dimitrakopoulos 2007
15 The Extractive Metallurgy of Lead Roderick J Sinclair 2009
16 Flotation Plant Optimisation: A Metallurgical Guide to Identifying Ed: Christopher J Greet 2010
and Solving Problems in Flotation Plants
17 Advances in Orebody Modelling and Strategic Mine Planning I – Ed: Roussos Dimitrakopoulos 2010
Old and New Dimensions in a Changing World
18 History of Flotation A J Lynch, G J Harbort and 2010
M G Nelson
19 The Cadia Valley Mines Ed Malone 2011
20 Cut-off Grades and Optimising the Strategic Mine Plan Brian Hall 2014
21 Comminution Handbook Ed: Alban Lynch 2015

Copies of all publications currently in print may be obtained from:

The AusIMM, Melbourne, Australia | Telephone: +61 (3) 9658 6100 | Email: publications@ausimm.com.au
Key: #: Publication number | Symp: Symposium series number | *: Out of print