Cone Crusher Performance: Research

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/276226567
Cone Crusher Performance
Research · May 2015

DOI: 10.13140/RG.2.1.2212.7526
CITATIONS READS
33 12,281
1 author:
Magnus Evertsson
Chalmers University of Technology
67 PUBLICATIONS 771 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
AMIRA P9Q View project
Real-Time Optimization of Crushing Plants View project
All content following this page was uploaded by Magnus Evertsson on 13 May 2015.
The user has requested enhancement of the downloaded file.

THESIS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
CONE CRUSHER PERFORMANCE
CARL MAGNUS EVERTSSON
: 2 2
20
0 0
20
a r i
u
jan
24
MACHINE AND VEHICLE DESIGN

CHALMERS UNIVERSITY OF TECHNOLOGY
GÖTEBORG, SWEDEN 2000
Cone Crusher Performance
Carl Magnus Evertsson
ISBN 91-7197-856-9
Copyright © 2000 Carl Magnus Evertsson
Doktorsavhandlingar vid Chalmers tekniska högskola

Ny serie nr 1551
ISSN 0346-718X
Machine and Vehicle Design

Chalmers University of Technology
SE-412 96 Göteborg
Sweden
Telephone + 46 (0)31-772 1000
URL www.chalmers.se
Cover:
Vertical cross-section of a cone crusher by Peter Schachinger
Printed in Sweden by
Chalmers Reproservice
Göteborg, 2000
To Maria and Alice
MACHINE AND VEHICLE DESIGN
CHALMERS UNIVERSITY OF TECHNOLOGY
ABSTRACT
Cone crushers are used by both the aggregate producing and the mining industry. Both
industries are interested in increasing the product quality while at the same time lowering the
production costs. Prediction of crusher performance has been focused on, since crushing is a
vital process for both industries.
In this thesis a method for prediction of cone crusher performance is presented. The model is
based on the laws of mechanics and constitutive relations concerning rock breakage
characteristics. There are some crucial assumptions which are of central interest for the model.
The validity of these assumptions has been verified by full-scale tests.
The overall size reduction process is a result of several subsequent crushing events.
Therefore, the process occurring in a cone crusher must also be modelled in the same repetitive
way. Each crushing event is modelled with a selection and a breakage function. Selection
corresponds to the probability of a particle being broken when an aggregate of particles is
compressed. Breakage represents the way a single particle is broken into smaller fragments.
The appearance of the selection and breakage functions is rock material specific and can be
obtained by laboratory tests. The characterization of the fragmentation behaviour for rock
materials is done by form conditioned compression crushing tests. Two modes of breakage are
possible to achieve in a cone crusher. The location of the choke level is the criterion which
determines the breakage mode. Interparticle breakage is achieved above the choke level while
only single particle breakage is achieved below this level.
The crusher model takes the fragmentation behaviour of the rock and feed size distribution
into consideration. Moreover, chamber and machine geometry together with machine
parameters such as closed side setting, stroke and eccentric speed, is accounted for. On all
occasions continuity of mass is preserved.
Three main factors are identified to promote the size reduction process occurring in a cone
crusher. These factors are: breakage modes, number of crushing zones and compression ratio.
The main factors are affected by both design and operating parameters. For a given crusher, the
factors depend on eccentric speed, closed side setting, rock material breakage characteristics
and feed size distribution. The main factors provide a fundamental and detailed understanding
of how a cone crusher operates. Any design consideration should be evaluated against these
main factors.
The model can be used as a simulation tool to assist in the design process of crushers. Any
arbitrary design can then be studied. If a set of simulations is performed for a given crusher, a
Crusher Performance Map is achieved, which in turn can be used when optimizing a given
crushing task or a crushing plant.
Keywords: aggregates, comminution, crushing, mining, modelling,

particle size, simulation, size-reduction.
i
DISSERTATION
This dissertation comprises the following papers:
Paper A: Evertsson, C. M., “Prediction of Size Distributions from Compressing Crusher
Machines”, Proceedings EXPLO 95 Conference, Brisbane, Australia, 173-180, 4-7
September 1995.
Paper B: Evertsson, C. M. and Bearman, R. A., “Investigation of Interparticle Breakage as
Applied to Cone Crushing”, Minerals Engineering, Vol. 10, 199-214, February
1997.
Paper C: Evertsson, C. M., “Output Prediction of Cone Crushers”, Minerals Engineering,
Vol. 11 215-232, March 1998.
Paper D: Evertsson, C. M., “Modelling of Flow in Cone Crushers”, Minerals Engineering,
Vol. 12, 1479-1499, December 1999.
Paper E: Evertsson, C. M., “Size Reduction in Cone Crushers”, Minerals Engineering
Conference '99, Falmouth, England, 22-24 September 1999 (Accepted for
publication in Minerals Engineering).
DIVISION OF WORK BETWEEN AUTHORS

Paper B:
The draft paper was written by Evertsson and revised by Dr. R. A. Bearman.
ii
CONTENTS
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i
Dissertation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii
Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Production of Ballast Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Crushing of Mineral Ores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Basic Comminution Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Operating Principle of Cone Crushers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Scientific Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
5 Model Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.1 Structure of Crusher Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.2 Modular Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
6 Nominal Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
7 Size Reduction Process Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
8 Rock Fragmentation Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
8.1 Breakage Behaviour Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
8.2 Size Distribution Width . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
8.3 Compression Crushing Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
8.4 Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
8.5 Breakage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
9 Flow Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
9.1 Material Flow Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
9.2 Velocity Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
9.3 Capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
10 Interaction between Rock Flow and Size Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
10.1 Mass Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
10.2 Volumetric Filling Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
10.3 Utilized Compression Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
10.4 Relation between bulk density and size distribution . . . . . . . . . . . . . . . . . . . . . . . . . 31
10.5 Confinement and Breakage Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
10.6 Choke Feeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
10.7 A Particle Shape Model for Single Particle Breakage . . . . . . . . . . . . . . . . . . . . . . . 34
iii
11 Simulation Results and Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

11.1 Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
11.2 Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
11.3 Size Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
11.4 Capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
12 Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
13 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
13.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
13.2 Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
13.3 Size Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
13.4 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Paper A: Prediction of Size Distributions from Compressing Crusher Machines
Paper B: Investigation of Interparticle Breakage as Applied to Cone Crushing
Paper C: Output Prediction of Cone Crushers
Paper D: Modelling of Flow in Cone Crushers
Paper E: Size Reduction in Cone Crushers
iv
ACKNOWLEDGEMENTS
The initiative for this research project was taken by producers of aggregates, contractors and
manufactures of crushing equipment in Scandinavia. The work has been carried out at Machine
and Vehicle Design, Chalmers University of Technology.
I wish to express my sincere thanks Professor Göran Gerbert for his excellent supervision.
With his great devotion and general competence, Professor Gerbert has supported, inspired
and encouraged me throughout the project.
The promoters (also financial sponsors of the project) formed a technical reference group to
which the author has regularly reported and presented research results. The people who has
been working in the reference are (in alphabetical order, past and present): Jonas Andersson
(Nordberg Sweden AB), Per Andersson (SveBeFo), Richard Bern (Svedala-Arbrå AB), Jan
Bida (NCC Industri AB), Hans Carlsson (Sabema Material AB), Kjell Carlsson (Sabema
Material AB), Leif Dahlqvist (Skanska AB), Konny Deubler (Råsjö Kross AB), Christer
Hagert (Vägverket), Pär Johnning (Ballast Väst AB), Christian Ottergren (Svedala Industri
AB), Lennart Persson (Skanska AB), Bo Ronge (Chalmers University of Technology), Sigvard
Sundelid (Gatu & Väg AB) and Sven-Olof Wennberg (Nordberg Sweden AB). Each member
of the technical reference group is hereby gratefully acknowledged.
The author is grateful to the following companies and organizations for funding this work:
Development Fund of the Swedish Construction Industry (SBUF), Gatu & Väg AB, National
Swedish Road Administration (Vägverket), National Swedish Sand and Crushed Stone
Association (GMF), NCC Baggermanns AB, Nordberg Sweden AB, Råsjö Kross AB, Sabema
Material AB, Skanska AB and Svedala Industri AB.
I would like to thank all my colleagues (past and present) at Machine and Vehicle Design.
They have all, in some way, contributed to the progress of my work.
Dr. Anders Hedman deserves special thanks for reading the thesis and many of the papers.
During the years Anders has been helpful in increasing my vocabulary and among other things
taught me that Gothenburg is a city in Nebraska.
Finally, I want to express my gratitude to Maria for her love and for having the patience of
an angel with me. I must also thank our daughter Alice for being there with her exultingly
happiness and always ready to jump for joy.
Carl Magnus Evertsson

Göteborg, January 2000
v
vi
1 INTRODUCTION
Cone crushers are used by both aggregate producers and the mining industry. The purpose of
using crushers is to reduce the particle size of rock materials or to liberate valuable minerals
from ores. A tremendous amount of material is crushed yearly by cone crushers. Despite this
fact, the detailed knowledge of how a cone crusher works from a process point of view is quite
poor. In order to achieve significant advances in crushing efficiency, increased knowledge is
needed. High quality knowledge can only be gained through fundamental understanding.
1.1 PRODUCTION OF BALLAST MATERIAL

Crushers are of central importance for the production of aggregates from rock material as it is
in the crushers that the product is actually produced. The important crushing aspects for the
aggregate producers are to increase the product quality and capacity while at the same time
lowering the production costs.
During the last few years the quality requirements on roads and railroads have become more
accentuated, owing to the desire to minimise maintenance costs. In turn, this has led to
increasing requirements in the preparation of ballast materials and pavement aggregates. No
longer can the quarry operator offer either quality or quantity – both are now required. Owing
to lack of availability and to legal restrictions, access to natural gravel in northern Europe is
decreasing. The use of crushed rock material is therefore becoming more and more important.
The total annual consumption of ballast material in Sweden has varied between 65 and 100
million tons during the nineties. The relative amount of crushed rock is increasing steadily and
was approximately around 50% in 1999. The consumption of natural gravel is decreasing.
Production of crushed rock material comprises drilling, blasting, crushing and screening.
Sometimes it is claimed that crushing is the most important production step, because the final
product is formed in the crushers.
Crushed ballast material is produced in two main consecutive steps where the particles are
formed, viz. blasting and crushing. The raw material for the process is solid rock. The purpose
of blasting is primarily to liberate the material from the surrounding rock and secondarily to
reduce its handling size.
After blasting, the rock material is transported to a crushing plant, often situated near the pit.
Crushing plants have very different layouts depending on when and for what purpose they
were built. The properties of the rock material, together with the type of product and capacity
demand, are factors that influence the choice of crushers and crushing plant layout.
An example of a crushing plant is shown in Figure 1. In the first step, feeder and grizzly, of
the process an important quality improvement takes place. Soil, clay and fines are separated
from the rock of higher quality. The fine material originating from weak zones in the rock
mass would otherwise pass through several process steps, with unnecessary and undesirable
waste in capacity. If the weaker material is allowed to be added to the final product, a
significant decrease in product quality appears.
Today, there is no technology available from which the total size reduction of all the
demanded product gradings can be achieved in one single step. Consequently, crushing is
performed as a series of consecutive process steps. By screening, ready-made gradings are
separated at intermediate and final stages. Correctly explored, product quality increases with
the number of crushing steps. Repeated size reduction improves the quality and strength
gradually at every stage. Gentle crushing, i.e. small size reduction in the last stage, is known to
give a product with high strength and a good particle shape, i.e. high cubicity.
1
The primary crusher is normally a jaw crusher which is fed with blasted rock material. As
the size of the blasted material may vary greatly, the primary crusher has to have a large inlet
opening. Secondary crushers are normally of the cone or gyratory type. Tertiary and final
crushers are mostly cone crushers. In situations where very high cubicity is required, the final
crusher may alternatively be an impact crusher if the rock is particularly difficult to shape.
Gyratory crusher
Feeder with grizzly
Screens
Screen
Jaw crusher +64

Screens
Stockpile Cone crusher

0-20
Bins
0-5 5-22 22-32 32-45 45-64
0-200
Bins
0-22 22-32
Cone crusher Impact crusher
Screens
Bins
16-22 11-16 0-2 2-5 5-8 8-11
Figure 1. Layout of a four-stage crushing plant for ballast production.
1.2 CRUSHING OF MINERAL ORES

In the mining industry crushing is the first mechanical stage in the process of comminution, in
which the main objective is the liberation of valuable minerals from the gangue. The most
important crushing aspect is normally the reduction ratio, which is an implicit measure of the
product size distribution.
A layout of a size reduction plant for comminution of a nickel rich ore is shown in Figure 2.
The plant consists of three crushing stages and one single milling stage. Only one product is
required from the crushing plant. That product is then the feed to the milling stage. After
milling the material is fed to a concentration plant.
The primary crusher is normally a gyratory or a jaw crusher, depending on the capacity
requirement. Gyratory crushers have significantly higher throughput than jaw crushers. This is
the reason why gyratory crushers are commonly used as primary crushers by the mining
industry. Primary crushing can be performed underground in the mine or on the surface close
to the subsequent crushing and comminution stages.
Secondary crushers can be of the gyratory or cone type. Tertiary crushers are almost
2
exclusively cone crushers. The configuration of secondary and tertiary crushers together with
screens varies from case to case. Closed circuits are very common where the crushers are
producing a ball mill feed. For closed circuit crushing the finished product is the undersize
from the screens while the oversize material is fed back to the crushers and recrushed. Open
circuit crushing is often used in intermediate crushing stages, or when the secondary crushing
plant is producing a rod mill feed.
+55
Screens
20-55
Gyratory 0-20
crusher Gyratory
crusher
+55 0-70
Stockpile 20-55 Screens

0-180
0-20
20-55 Storage
Cone crushers
0-30
20-30
Screens
0-20
SAG ball mill
Hydro-
0-20 Storage Pump cyclone
0-0.125
to concentration
plant
Figure 2. Layout of a size reduction circuit for a nickel rich ore.
1.3 BASIC COMMINUTION PRINCIPLES

There are three comminution principles of industrial interest. Other possible principles exist
(electrical etc.) but to date these have not reached the level required for commercial
applications. The basic comminution principles described here are:
• Compressive crushing
• Impact crushing
• Attrition
All important minerals and rock materials are crystalline. In a crystalline structure the atoms
are ordered in a regular three-dimensional pattern. To reduce the particle size of a rock
material we have to divide each particle into smaller particles. For a given volume of material,
the total surface area must increase as the number of particles increases. The increase of the
overall surface area is a process which is energy consuming.
During compression, energy is transformed from potential energy into elastic energy, which
is stored in the crystal lattice. When the stress level somewhere in a crystal reaches a critical
value, the interatomic bonds will be broken and a crack will propagate. The stress
concentration depends on the mechanical properties of the individual materials, but most
important is normally the presence of cracks or flaws in the matrix. These cracks and flaws
will act as sites for the stress concentration and from these sites the cracks will propagate.
When a crack propagates, a new surface will be created. As an atom on a free surface has a
higher potential compared to an atom in the matrix, the creation of a free surface will be an
3
energy consuming process. A part of the elastically stored energy will transform into surface
energy as the crack propagates. The rest of the elastic energy will dissociate as the crystal
relaxes. This behaviour is known as brittle fracture. In contrast, a tough and ductile material, as
for example steel, can relax the elastic energy by the mechanism of plastic flow. The energy is
then consumed in distorting the shape of the material.
1.3.1 COMPRESSIVE CRUSHING

In compressing crushers the rock material is crushed by squeezing the particles between two
surfaces. By compression of the material, contact forces between the particles give rise to a
tensile stress field within the rock particles. Cracks are formed when the stress field at some
local point exceeds a critical value. When the cracks propagate or link together by brittle
rupture, a macroscopic breakage is achieved which results in smaller fragments.
Compressive crushing can be either force or form conditioned. For the force conditioned
case the displacement, and thus the size reduction, is a function of the applied force. By
controlling the force, and implicitly also the energy input, the size reduction can be controlled.
In the form conditioned case, the size reduction and the applied force are a function of the
displacement. Size reduction can be controlled by controlling the compression. Energy input
and the resulting force are secondary parameters in the form conditioned case.
Jaw, gyratory, cone and roller crushers are all examples of crusher machines that utilize
form conditioned compression to achieve size reduction. There is an essential and important
difference between roll crushers and the other crusher types. In a roll crusher the material is
only compressed once while, for example, in a cone crusher the size reduction is a repetitive
process.
A particle which is broken by compressive crushing generates daughter fragments that fall
into two main size classes. A discrete number of larger fragments are formed by cracking as a
consequence of the tensile stress state. Some amount of fine material is generated at the contact
points due to shear stresses.
1.3.2 IMPACT CRUSHING

Another commonly used fragmentation principle is impact breakage. Impact breakage is an
energy controlled form of breakage. The energy can be transmitted to the particles by a rotor
which accelerates the particles to a velocity close to the peripherical speed of the rotor. The
particles are then let to impact with a wall of other particles where the fragmentation occurs,
see Figure 3. The amount of size reduction obtained is controlled by the amount of energy
transmitted to the particles. The kinetic energy is in turn dependent on the particle velocity.
The size reduction can therefore be controlled by the speed of the rotor.
s
P = P(s) mv
2
E part = ---------
2
P v
P
Figure 3. The main difference between form conditioned (left) and energy conditioned (right)
breakage is the principle for energy input.
Crushing machines utilizing energy controlled breakage are called impact crushers. These
4
types of crusher are known to produce a product with a good particle shape, i.e. high cubicity.
Depending on the energy level, the crusher can be used to primarily improve shape or to give
significant size reduction. A drawback of impact crushers is that they produce a significant
amount of fines. This property is a direct consequence of the utilized comminution principle.
From strength tests utilizing a modified Hopkinson Pressure Bar, Briggs and Bearman [4]
concluded that impact crushers produce material that contains less variation in strength than
the feed because every particle is exposed to the breakage mechanism. Cone crushers do not
break material as consistently as impact crushers, as there is variation in both the number of
particles exposed to the breakage mechanism and the degree to which the particles are crushed.
1.3.3 ATTRITION
Attrition is breakage by shear failure and produces mostly fine material. Generation of fines is
often undesirable. Attrition occurs mainly in practice due to interaction between particles and
is thus to be considered as interparticle breakage.
1.4 OPERATING PRINCIPLE OF CONE CRUSHERS

In a cone crusher the rock material is crushed between rigid surfaces. The motion of the
moving surface is independent of the loading of the crusher. Therefore, the explored
comminution principle is form conditioned compressive crushing.
There is no significant difference between cone and gyratory crushers. In both types, the
crushing action is achieved by an eccentric gyratory movement of the main shaft axis, as
shown in Figure 4. Particles are nipped, compressed and crushed between mantle and concave.
Breakage of particles can be both single particle (particles are broken between cone and
mantle) and interparticle (between other particles). The mantle and the concave must be
replaced regularly as they are continuously subjected to wear.
Crushing takes place continuously and is interparticle to a great extent. Interparticle
crushing is achieved when a particle is stressed and broken between other particles.
Interparticle breakage is desirable from the point of view of wear and service life. Interparticle
breakage is also considered to give the best shape of the produced particles. To achieve
interparticle breakage choke fed conditions are normally recommended. Choke fed conditions
imply that the inlet of the crusher is covered with feed material. The feed material is then fed
into the crusher by the influence of gravity.
The basic operating principle of a cone crusher is shown in Figure 4. The properties, i.e. size
distribution and quality parameters, of the product leaving the crusher are a result of the
interaction between crusher and rock material. In turn, the interaction is dependent on chamber
geometry, crusher dynamics and rock material characteristics.
Concave, mantle, eccentricity and location of the pivot point together form the chamber
geometry of the cone crusher. Like a crank, the eccentricity turns the main shaft with constant
speed. At the top the main shaft is journalled in a pivot point. For some designs, the pivot point
is just virtual. The mantle is fixed to the main shaft and the concave to the crusher frame. The
resulting motion of the mantle and main shaft will be a nutating motion.
If a vertical cross-section of the crusher is considered, the mantle will move cyclically
forwards and backwards relatively to the fixed concave. During the squeezing motion the
intervening rock material will be subjected to a compressive stress field and thereby
comminuted. During the releasing motion the material is able to flow down through the
chamber. The motion is driven by the influence of gravity and is governed by the motion of the
mantle.
5
Feed
Concave Concave
Mantle CSS
s
Product n Main shaft axis
Figure 4. Principle of a cone or gyratory crusher.
At a fixed point of location somewhere in the crusher chamber, the distance between mantle
and concave will vary between open and closed position, as shown in Figure 5. The motion can
be described with a cyclic function of the eccentric angle α . The eccentric angle corresponds
to the position of the eccentricity and therefore also to the position of the main shaft.
The distance between the mantle and concave at the outlet of the crushing chamber is called
the Closed Side Setting when the mantle is in its closed position. The closed side setting is
commonly referred to as the CSS.
Nominal
distance 1st cycle 2nd cycle
OSS open
s nom
neutral
CSS closed
opening closing opening closing
α = ωn t
π 2π 3π 4π
Figure 5. For a fixed point in the chamber the distance between mantle and concave varies
cyclically with time.
The breakage of material in cone type gyratory crushers is traditionally regarded as relying
upon single particle breakage. In the last ten years the emphasis has shifted with manufacturers
trying to generate higher degrees of interparticle breakage. Increasing the degree of
interparticle crushing is claimed to improve crushing efficiency and product shape.
When the material passes the crusher it will be subjected to several repeated compressions
due to the cyclic oscillation. During every compression the material will be partly crushed.
Due to the downward flow and the repeated compressions of the crusher chamber, the material
will follow a step-wise motion.
6
1.5 PROBLEMS
At a first glance a cone crusher may seem to be a quite simple piece of equipment – rocks are
reduced in size by passing a machine with moving surfaces. This opinion might be true if no
requirements whatsoever are put on either the crushed product or the operation efficiency of
the crusher. By applying strictly defined requirements on product quality in combination with a
desire to optimize important production aspects, the need for a detailed understanding of
crusher operation arises.
The size reduction process occurring in a cone crusher has not yet been fully understood.
However, the possibility to simulate and predict cone crusher performance is of great interest
for the development of crushers as well as for the design and optimization of crushing plants.
A model giving manufacturers the possibility to tailor make a rock crusher to specific
customer needs would undoubtedly be a major breakthrough for the quarrying and mining
industry.
The problems related to cone crushers can be divided into three different groups. These
groups are:
• Process design of crushing plants
• Crusher design
• Crusher operation
1.5.1 PROCESS DESIGN OF CRUSHING PLANTS

When designing comminution processes it is of course of great interest to have knowledge
about the different components and machines involved.
The performance of existing crushers can be mapped by full-scale tests. These multi-
dimensional maps can then be used for flowsheet simulations to evaluate different process
designs. By applying suitable optimization methods it would be possible to obtain the optimal
choice of operational parameters for the crushers involved. The disadvantage of this approach
is obvious, as full-scale testing can be very time consuming.
1.5.2 CRUSHER DESIGN

The size reduction process occurring in a cone crusher is complex by nature. There are no
direct correlations between input and output parameters. If no models are available the
possibility to predict the performance of an arbitrary cone crusher is very limited. Therefore,
developments and advances in crusher design will be restricted to incremental improvements
of existing equipment.
The parameters related to the crusher can be divided into independent and dependent
parameters. Independent machine parameters are eccentric speed, closed side setting ( CSS ),
stroke and crushing chamber design. The inherent material properties such as strength and
attrition resistance are also to be considered as independent. Feed size, shape and strength are
results from previous process stages and therefore treated as independent. The dependent
parameters correspond to the output of the crusher. Crusher related parameters are capacity,
power requirement and hydraulic pressure (resultant vertical load). Dependent parameters
characterizing the product are product size, shape and strength. The dependent parameters are,
in one way or another, dependent on at least one independent parameter.
A way to illustrate the dependencies between the independent input parameters and the
crushing result is to formulate a so-called interdependency matrix. This matrix relates the input
to the output parameters, as shown in Eq. (1).
7
Speed
CSS
Capacity X X X X 0 0 X 0 0
Stroke
Power requirement X X X X X X X X X
Chamber design
Hydraulic pressure = X X X X X X X X X (1)
Material strength
Product size X X X X X X X X X
Attrition resistance
Product shape X X X X 0 0 X X X
Feed size
Product strength X X X X X X 0 0 X
Feed shape
Feed strength
An X in the interdependency matrix symbolizes a dependency between two parameters. If the

interdependency matrix had been quadratic and diagonal, the crushing process would have
been easy to analyse. In that case one input parameter would only affect one single output
parameter. The process would have been easy to tune if, for example, speed had been the only
parameter affecting capacity.
A process with a quadratic and triangular interdependency matrix would have been possible
to tune by adjusting the input process parameters in correct order. In the case of crushing with
existing crusher equipment, such as jaw and cone crushers, the interdependency matrix will be
different from triangular. This is the fact behind the complexity of the crushing process. As an
example it can be seen from Eq. (1) that the product size distribution is affected by all the
independent parameters.
1.5.3 CRUSHER OPERATION

Owing to variations in the rock material, such as type of mineral, grain size, distribution of
cracks etc., quality problems often occur.
Process control is desirable for maintaining high product quality, but today there are no
systems available for this purpose owing to the absence of on-line detectors for product size
distribution and shape. Another drawback is that very few models describing crusher
characteristics are available.
A proposed system for closed loop process control is illustrated in Figure 6. To achieve
closed loop process control, development of both detectors and crusher models is needed.
Without detectors the system will be of the open loop type and thus very sensitive to feed
variations and wear of the crusher chamber.
on-line on-line
detector detector
Feed Product
Crusher
CSS Speed
Crusher
model
Product
target
Figure 6. Possible system for process control of a single crushing stage. On-line detectors could be
of the optical, acoustic or microwave type. Control signals from the crusher model are
closed side setting (CSS) and eccentric speed.
8
In every crushing stage an undesirable amount of fine material is produced. The definition of
fines varies, but a common definition is particles with a size of less than 2 mm. In some
geographical regions the disposal of fines can be difficult. Therefore, it is of great interest to
the ballast producing industry to reduce the production of fines, while still preserving the
overall reduction ratio.
2 SCIENTIFIC APPROACH
This work was carried out at the Machine Element Group, Machine and Vehicle Design,
Chalmers University of Technology. The focus of research within the group is on machine
components and machines. The main activity is the systematic search for new knowledge
concerning the principles of operation, characteristic behaviour and performance of the objects
studied. It is believed that significant improvements in machine design can be achieved by
fundamental understanding of the component or machine.
The research is problem oriented. This means that the problem itself is in focus and not the
method or tools to solve it. The different steps in the problem oriented research method put
into practice by the Machine Element Group are shown in Figure 7. The problem oriented
research method is thoroughly described by Hultén [22].
Phenomenon
Problem Breakdowns
Malfunction
Lack of knowledge
Research Literature
Observations Data acquisition
Guiding experiments
Physical principles
Modelling Assumptions
Fundamental experiments
Verification Experiments
Characteristics Simulations
and Performance
Implementation Design considerations
Solution New or redesigned product
Figure 7. Structure of a problem oriented research process.
Research is defined as the activities undertaken to create new knowledge, which in turn is
used to find a solution to the original problem.
Most research projects start with a problem. The character of the problem can be very
different. Machines or components can reveal unexplained breakdowns or there are
malfunctions. Unexpected phenomena might occur. Sometimes there is a lack of knowledge
which originates from a need of increasing the performance of the object.
9
The first step in the research process is to identify a knowledge gap. Different kinds of
observations are used for this purpose. Literature reviews are used to outline the current status
of the knowledge base related to the problem. Data acquisition or guiding experiments can be
performed in order to clarify the nature of the problem. When the knowledge gap is identified
the objectives of the research project can be formulated.
Modelling is probably the most central part of the research process. The aim of modelling is
to describe the phenomenon in mathematical terms. Already known physical effects are
normally used when the models are formed.
Very often, the phenomenon is complex in nature. If the complete behaviour, including all
possible aspects, is to be described, the mathematical model would be large and complex and
thus difficult to use in simulations. In order to make the model possible to handle
simplifications are needed. Assumptions and simplifications should be done on such a level
that no restrictions are obtained at the level where the problem is formulated. Which
assumptions and simplifications that should be made depend directly on the degree of
explanation or resolution required of the final model.
The model structure can be described with a theory tree, see Figure 8. The theory tree is a
hierarchical structure of submodels, which together form the hypothesis. The top level of the
tree corresponds to the final model describing the problem. The different sub levels correspond
to models which are often based on accepted theories taken from a common knowledge base.
Cone Crusher Performance Top level
Wear Energy Size Distribution Capacity Sub level 1
Sub level 2
Sub level 3
Figure 8. Theory tree.
If a phenomenon cannot be explained by already existing theories, fundamental experiments

can be performed to obtain relations between different parameters. Typically, constitutive
relationships are found in this way.
Verifications are carried out by experiments in order to verify the accuracy of the final
model. The requirements set on the level of agreement differ depending on the aim of the
model. To achieve a high qualitative correlation is often more important than to achieve
quantitative agreement. If the verification fails, the model has to be improved, and in some
cases new observations are needed.
Once a verified model is achieved, simulations can be carried out in order to obtain
information about the characteristics and performance of the object studied. Through the
physical effects used in the model, fundamental understanding is achieved. Compared to full-
scale testing, mathematical simulations are usually far more cost effective. From the
simulations, conclusions can be drawn concerning important effects and to which degree
performance can be improved.
The new knowledge gained can be formulated in so-called design considerations. These
considerations can be implemented by redesigning the original product or in a completely new
design. Properly done the new product is a solution to the original problem.
10
3 OBJECTIVES
The overall purpose of this work is to achieve a fundamental understanding of how a cone
crusher works from a size reduction point of view.
The objective of this thesis is to develop a theoretical model, which can be used to predict
product size distributions and capacities from cone crushers. An important property of the
process model is that it must be closely physically defined. Prediction of the performance of a
cone crusher already at the design stage would enable improvements without full-scale testing.
Crushers already existing in crushing plants can be investigated and the best operating
parameters chosen.
Crushing is quite a complex process and the results are affected by both machine properties
and rock properties. Therefore, machine design and operating conditions, together with the
breakage characteristics of the rock material, should be taken into consideration.
Prediction of energy consumption and power requirement is not an objective of this work, as
the operating principle of cone crushers is form conditioned. Energy consumption is
considered to be a secondary effect of the size distribution process. Liner wear prediction is not
an objective of work.
4 LITERATURE REVIEW
Historically, the process of size reduction has been studied in terms of the energy consumed
during the operation of a comminution plant (Lynch [27]). This was a logical starting point
because size reduction is responsible for a large proportion of the costs of ore treatment and the
energy consumed is the major cost in size reduction. This basis of investigation was influenced
more by the economics of the operation than by any other factor.
Relatively few models describing the size reduction process occurring in cone crushers have
been presented. Bond’s model [3], Third Theory of Comminution, presented in the early fifties
is still frequently used today by manufacturers when selecting crusher equipment. The success
of the model is probably explained by its ease of use and not by the quality of the predicted
results. Before Bond presented his model, both Rittinger [30] and Kick [23] had worked with
similar types of approach. The theories of Rittinger, Kick and Bond are described in details in
Paper A.
For single fracture of a brittle solid, the distribution function for fragment size after fracture
has been derived by Gilvarry [19]. The physical point of departure is the Griffith theory of
brittle fracture [20]. The derivation is based on a closely defined physical model and deduced
by the laws of probability. Single fracture of a brittle solid is defined as fracture by an external
stress system, which is instantly applied and permanently removed when the first Griffith
flaws begin to propagate. Subsequent flaws are activated by stress waves produced by
propagation of prior ones. Comminution of a specimen by repeated fracture of specimens is
excluded by the definition, since there is no assurance that the flaw distributions in the
fragments exposed to the repeated breakage are the same as in the original specimen.
Rosin and Rammler [32] suggested an empirical equation describing the size distribution
after breakage. Broadbent and Callcott [7][8] suggested a modification of the Rosin-Rammler
equation for use in their investigations of coal breakage processes. The distributions given by
Rosin-Rammler and Broadbent-Callcott are examples of some well-known and frequently
used size distribution functions. During the last three decades a number of both analytical and
empirical functions have been suggested by different authors. The mathematical form of the
functions vary from simple to rather complex. Klimpel and Austin [24] derived an expression
11
in which they tried to link physical properties of the particles to the breakage function. Barnard
and Bull [1] determined an empirical breakage function for the cumulative size distribution by
performing drop-shatter tests with brown-coal.
Drop weight tests have been used by several authors. Examples of such work have been
carried out by Rose [31] and Briggs [5].
Models describing interparticle breakage behaviour are relatively rare. Fewer tests have
been conducted and reported with particle beds than with single particles. Prasher [29]
concludes that this fact represents a research gap, since most industrial comminution involves
beds, and single-particle tests cannot adequately represent the interaction of particles with
different size, shape and orientation. Buss et al. [9] worked with the aspects of energy
utilization in particle beds. Liu and Schönert [26] concluded that it should be possible to model
closed circuit comminution with high pressure roller mills originating from models describing
interparticle breakage. The problem with roller mills is in a way similar to cone crushers but
there are also great differences. In a roller mill the material is only subjected to one single
crushing event, in contrast to repetitive crushing in cone crushers. Interparticle breakage
behaviour was modelled from an energy point of view. Size reduction in a cone crusher is form
conditioned. Therefore energy is considered to be a secondary effect, i.e. instead of energy
input there is energy consumption. Further, Schönert [33] has shown that the probability of a
particle breaking depends upon the number of forces acting on it.
The well-known flow model developed by Gauldie [17][18] is only capable of predicting
crusher capacity and does not embrace any ideas concerning the size reduction process.
Gauldie’s model has been used by several authors for prediction of the throughput from
crushers [10][14]. The model has a serious weakness since it lacks a genuine constitutive law
coupling size distribution to bulk density.
Empirical models, such as the one of Whiten [34][35], provide excellent guides for
operational purposes but these models are insufficient for describing the going on size
reduction and flow mechanisms in the crusher. These types of model cannot be used for design
purposes.
Bearman [2] presented mathematical relationships which are capable of predicting energy
and eighty per cent passing size of product based on identified controlling factors. These
factors were certain rock strength parameters and discharge setting. Especially good
predictions were achieved for the rock strength parameter fracture toughness. This is a strong
indication that tensile failure is the dominating breaking mechanism in both jaw and cone
crushers.
Crusher models which are based on laws of mechanics are very rare. Briggs [5] proposed a
model based on equations of motion, but which was only capable of treating single particle
breakage.
A large number of authors have worked with quality aspects of aggregate production
(Eloranta [10], Heikkilä [21], Kojovic [25]). As long as no fundamental model has been
accepted describing the crushing mechanisms in cone crushers, the conclusions are sometimes
contradicting and may be dependent on the specific application.
12
5 MODEL DEVELOPMENT
The overall model for predicting cone crusher performance presented in this thesis has
successively been developed in the following four parts:
• Size reduction process modelling
• Characterization of rock material breakage behaviour
• Flow modelling
• Interaction between flow and size reduction
Process modelling comprises identification of the comminution principle utilized together
with the basic operating principle of cone crushers (Paper A). The overall size reduction
process is modelled as a series of repeated crushing events. Form conditioned crushing tests
are performed to characterize the fragmentation behaviour of different rock materials (Paper
B). Equations of motion are used to describe the flow of material in the crushing chamber
(Paper D). A geometry module capable of handling arbitrary chamber geometries delivers
geometrical data to the flow model. To guarantee that mass continuity is preserved and that no
other requirements are violated the interaction between material flow and size reduction is
given some extra attention (Paper E).
5.1 STRUCTURE OF CRUSHER ANALYSIS

In Figure 9 a structure for analysis of a cone crusher is shown (Paper D). To calculate the
product size distribution of the output from a cone crusher according to the suggested
structure, two main calculation models are needed. These models are:
• Size reduction model
• Flow model
The size reduction model predicts the size distribution after compressing the rock material.
Every single reduction step is described by a selection and a breakage function. The flow
model is important since it describes how the rock material moves down through the crusher
chamber. Thereby the flow model provides input to the size reduction model which is used to
determine the values of the selection and breakage functions. Two models for flow and size
reduction respectively have been presented earlier (Papers A, B, C, D and E). The interaction
between these two models is quite complex and intricate, as the overall size reduction in a cone
crusher is a result of a repeated consecutive comminution process.
Cone Crusher Performance Model
Crusher Flow Model

Geometry Crusher
Performance
Rock
Material Size
Reduction
Figure 9. Structure of the crusher performance model.

The flow and size reduction models will be closely tied together as the input to the size
reduction model will depend on the result of the flow model. In turn, the flow model will
depend on the size distribution, as there will be several repeated consecutive reduction steps.
13
Further information needed to carry out the calculations is:

• Fragmentation behaviour
• Crusher geometry
• Density model (constitutive relation between size distribution and density)
The fragmentation behaviour or the breakage characteristics are achieved from form
conditioned compressive bed tests where an aggregate of particles is comminuted under
controlled conditions (Paper B). Any existing and non-existing profile design must be able to
be analysed by the crusher geometry module. Mantle and concave geometries together with
other important crusher parameters are given as input to a geometry module (Paper D). This
module provides the data needed for the flow model. A constitutive relationship which relates
size distribution to uncompressed bulk density is needed as the density varies with particle size
distribution (Paper D).
5.2 MODULAR APPROACH

Based on the structure in Figure 9, a calculation method has been developed which has been
implemented in a computer program. A modular approach is suitable and natural according to
the structure of the calculation method. Each module is self sustained and can be upgraded or
replaced, if needed. The information flow follows the direction of the arrows in Figure 10.
Input Calculation Output

Intermediate Final
Machine
parameters
CSS, speed, stroke Flow model Velocity Capacity [tph]
Zone boundaries
Machine and s/b-ratios
chamber geometry
Feed-size Filling ratio, η v

distribution
Size reduction
process model
Material breakage Product size “Crusher
characteristics distribution Performance Map”
S-, B-functions Density
Figure 10. The calculation method has a modular structure. Every single module can be replaced if
necessary.
Input to the calculation method are machine parameters such as closed side setting (CSS),
speed, stroke, chamber geometry, feed size distribution and material breakage characteristics.
Input is divided into four modules. Prediction of material flow and size reduction is performed
in two separate modules. Output from the method are capacity and product size distribution.
Zone boundaries and compression ratios ( s ⁄ b -ratios) are calculated as intermediate results.
5.2.1 CRUSHER PERFORMANCE MAP

By combining the total capacity with the product size distribution for every eccentric speed one
will achieve a set of curves corresponding to the net capacity for different product gradings. The
final result achieved is called a Crusher Performance Map (CPM). The CPM can be multi-
14
dimensional and shows how the final product depends on variations in machine parameters for
a given crushing chamber and feed. A Crusher Performance Map for a tentative cone crusher is
shown in Figure 11 (Paper C). This particular CPM is two-dimensional as it was calculated for
only one value of the CSS. It can be seen that the eccentric speed has a great influence on the
final product i.e. size distribution and net capacities of different product gradings are affected
by the eccentric speed.
The crusher performance map describes the combined response from the interaction between
crusher and rock material. Only the total capacity is independent of material choice and thus
valid for all rock materials. The map in the general case is multidimensional and has as many
dimensions as there are machine parameters. In the case where eccentric speed and closed side
setting are the varying parameters the crusher performance map will have three dimensions.
250
200
packing
+0-24
Total Capacity
+0-16
Throughput, [tph]
150
+0-11
+0-8
100
+0-2 +4-8
50
+2-4 +11-16
Net Capacities +8-11
+16-24
0
0 100 200 300 400 500 600 700 800
Eccentric Speed, [rpm]
Figure 11. Crusher Performance Map for a tentative cone crusher showing the capacities for
different product gradings. Only interparticle breakage is considered (Paper C).
From the crusher performance map it is obvious that total capacity and all net capacities
varies quite substantially with eccentric speed. For eccentric speed higher than 400 rpm packing
occurs if the feed to the crusher is not restricted. Net capacities for different product gradings
does not have maxima at the same eccentric speed. This fact makes optimization very
interesting and challenging. It is up to the aggregate producer to formulate criteria to which the
optimal eccentric speed can be chosen.
15
6 NOMINAL GEOMETRY
Before any simulations can be carried out, the nominal crusher chamber geometry has to be
encoded numerically. For this reason an input module for geometry was designed (Paper D). By
means of the generally designed input module, every possible chamber geometry can be treated.
The calculations in the flow model are therefore based on true geometry. Every existing
chamber geometry, independent of brand and model, can be described with a set of nodes which
are connected with arcs or straight lines. If desired, any arbitrary continuous function can be
used to describe the profile of the chamber geometry between two given nodes.
The nominal geometry for a Svedala Hydrocone crusher is shown in Figure 12. Already at
this stage it is possible to calculate a set of nominal parameters that characterize the crusher
chamber. Nominal stroke and bed thickness are determined at every point of a cross-section of
the chamber. The ratio between stroke and bed thickness is called the nominal compression
ratio and is denoted ( s ⁄ b ) nom . This parameter is central for later predictions of size reduction.
The horizontal cross-sectional area is calculated at all vertical levels. The minimum of the
cross-sectional area defines the so-called choke level of the crusher chamber. The location of
the choke level is important since it will control the different types of breakage modes in the
crushing chamber.
ID08, H3000F,
Crushing CSS=15,
Chamber e =14.5
Profiles Nominal Data
0
0 −0.3
s [mm]
Y−coordinate [m]
Y−coordinate [m]
−0.4 . . b [mm]
−0.2 (s/b)x100
−0.5 β [deg]
−0.4 CHOKE LEVEL −0.6
−0.6 −0.7
−0.8
−0.2 0 0.2 0.4 0.6 0 20 40 60 80
X−coordinate [m]
Crushing and Bed Vectors Nominal Cross−Sectional Area
−0.3
−0.3
Y−coordinate [m]
Y−coordinate [m]
−0.4
CHOKE LEVEL
−0.4
−0.5
−0.5
−0.6
−0.6
−0.7
−0.7
−0.8
0.2 0.4 0.6 0.06 0.08 0.1 0.12
X−coordinate [m] 2
Area [m ]
Figure 12. Nominal geometry for a Svedala Hydrocone crusher (Paper D). β is the angle between
the crushing and bed vectors.
16
7 SIZE REDUCTION PROCESS MODEL

In a cone crusher the incoming feed material is transformed to product by a repeated
comminution process. The number of reduction steps varies between different crushers, but it
has been shown that it can be ten or more (Paper D). An analytical model describing the
process should therefore include all these reduction steps.
In Figure 13, a model is presented using two basic mechanisms for describing each single
reduction cycle, viz. selection and breakage (Paper A). The product from one crushing event is
the feed to the following one. The size reduction occurring in every single crushing event is
described by a selection and a breakage function. It is believed that these two mechanisms can
describe all aspects of importance of the breakage behaviour.
f p1 p2 pN – 1 pN
S1 B1 S2 B2 SN BN
final
feed
product
( I – S 1 )f ( I – S 2 )p 1 ( I – S N )p N – 1
 --s- Material  --s- Material  --s- Material

 b u, 1 characteristics  b u, 2 characteristics  b u, N characteristics
Figure 13. Process model of a cyclically operating crusher with N consecutive crushing events. The
utilized compression ratio (s/b)u is determined from the location of each crushing event,
crusher dynamics and rock material density (Paper E). S=selection, B=breakage.
Selection S corresponds to the probability of a particle being broken. Breakage B reflects

the way a particle is broken, i.e. the size distribution of the daughter fragments. By introducing
matrix algebra, every single reduction step can then be described as
p i = [ B i S i + ( I – S i ) ]p i – 1 (2)
where p i represents the product from and p i – 1 the feed to step i . Furthermore, B i and S i are
matrix operators corresponding to breakage and selection, respectively. The total number of
crushing events occurring within the modelled crusher is denoted N . The value of the elements
in the breakage and selection matrices B i and S i will vary depending on the location in the
crusher chamber. The dimensionless machine parameter ( s ⁄ b ) u is called the compression ratio.
This parameter describes how much the rock material is compressed at a certain location in
crushing event i . The compression ratio is therefore used to determine the values of the
elements in B i and S i . The value of ( s ⁄ b ) u is obviously a key parameter for size reduction and
should therefore be delivered by the flow model.
Two different main types of breakage can occur in a cone crusher. The different kinds of
breakage are called breakage modes. The two breakage modes considered are interparticle and
single particle breakage. The breakage modes are not mutually exclusive, but can occur
simultaneously in a breakage zone depending on the surrounding conditions.
By introducing a mode matrix M i to Eq. (2), a general expression including both types of
breakage is achieved (Paper A). A crushing event where both interparticle and single breakage
occur is shown in Figure 14. The corresponding mathematical description is given in Eq. (3).
The mode matrices are diagonal by definition, and the only values the elements can take are
one or zero. The sum of the mode matrices is identical with the identity matrix I , in which the
elements are one in the main diagonal and zero in every other position.
17
inter
Mi Si Bi
pi
pi – 1 ( I – S i )p i
single
Mi Si
 --s-  --s- Material

 b nom, i  b u, i characteristics
Figure 14. A single crushing event comprising both interparticle and single particle breakage.
Depending on particle size and the location in the crushing chamber a particle will
either be crushed in the interparticle or the single particle breakage mode.
inter inter single single

pi = [ [ Bi S i + ( I – S i ) ]M i + Bi Mi ]p i – 1 (3)
inter single
Mi + Mi = I (4)
The conditions that control and activate the different breakage modes are the degree of
confinement and the local closed side setting (Paper E). It is assumed that interparticle
breakage occurs above the choke level where confined particle beds are obtained.
Furthermore, single particle breakage occurs when a particle is nipped between mantle and
concave. It is assumed that this is the only breakage mode obtainable below the choke level.
8 ROCK FRAGMENTATION CHARACTERISTICS

Jaw, cone and gyratory crushers all utilize form-conditioned compression to achieve size
reduction. For this reason geometry controlled (form conditioned) compression crushing tests
are performed to achieve the breakage characteristics of the rock material (Papers B and E).
The breakage behaviour is described mathematically by a selection and a breakage function.
From the form conditioned compression tests, the breakage behaviour is obtained as a primary
effect of the compression ratio. Force and energy are secondary but important parameters and
can be obtained if desired.
8.1 BREAKAGE BEHAVIOUR ASSUMPTIONS

The two functions S and B constitute a description of the breakage behaviour for the rock
material of interest. Some crucial assumptions are needed to make it possible to obtain S and
B by laboratory tests.
The value of the selection function S varies both with the compression ratio and with the
size distribution of the material which is to be crushed. The value of S varies with the size
distribution width but is independent of particle size within a given size distribution. The
particle size distribution is characterized by its width or spread and is assumed to be reflected
by a normalized standard deviation σ ℵ of the distribution.
The breakage behaviour is described with a breakage function B. The breakage function is a
cumulative function and the value of B describes the relative amount (by mass) of particles
passing a given relative size. For a given original particle size, the breakage behaviour is
18
assumed only to be dependent on the compression ratio. This means that a large particle gives
rise to a size distribution of the daughter fragments similar to a small particle, if the size scale
is normalized by the original size.
8.2 SIZE DISTRIBUTION WIDTH

It is desirable to describe a particle size distribution with a single number. Here, the size
distribution is assumed to be characterized by the normalized standard deviation σ ℵ of the
frequency function f corresponding to the size distribution (Paper E). The normalized standard
deviation is achieved by dividing the standard deviation σ by the mean value x of the
distribution
σ
σ ℵ = --- (5)
x
where σ and x are defined by

n n n
1 1
σ2 = ------------- ∑ f i ( x i – x ) 2 x = ---- ∑ f i x i N = ∑ fi (6)
N–1 N
i=1 i=1 i=1
The advantage of using the normalized standard deviation as a measure of the spread of the
size distribution is that this number will be independent of particle size. Further, the
normalized standard deviation will always take values in the interval zero to one.
8.3 COMPRESSION CRUSHING TESTS

Knowledge about the appearance of the S and B functions is needed before any simulations
can be carried out using the size reduction model. To quantify the S and B functions,
laboratory-scale compression crushing tests were performed (Paper E). In the tests, the rock
material is crushed in a steel cylinder. The equipment is designed to imitate the conditions, in
which a volume of material is compressed in a real crushing chamber (Paper B and [12]). The
laboratory test corresponds to the compressing part of the machine cycle when the liners move
towards each other. The material is then locked between the chamber walls and can only
deform elastically, plastically and/or break into smaller particles. The compression is form
conditioned. For the rock materials of interest, brittle fracture is the only fracture mechanism
of importance. At normal operating conditions, the maximum radial velocity of the mantle
relative to the concave is below 0.5 m/s. It is assumed that the breakage and selection
appearance are independent of the strain rate at this level.
For laboratory compression tests, the normalized compression ratio is denoted s ℵ. The same
variable is called the utilized compression ratio ( s ⁄ b ) u if it is achieved from calculations
originating from the crusher model. The numerical values have the same meaning regarding
these two variables.
The tests were performed in six different series, each with a constant compression ratio s ℵ.
The material chosen for the tests was gneiss taken from a quarry in Dalby in southern Sweden
where Svedala has a test-plant facility. The same material was used in the simulations and the
full-scale tests. The first test in each series were done on a mono-sized aggregate of fraction
+16-19 mm. The remaining tests were done on the material originating from the preceding
tests. After every compression test the material was sieved. The resulting size distributions
from each test are shown in Figure 15.
19
1 1
s ℵ = 0.10 s ℵ = 0.15
0.9 0.9
tests tests
0.8 simulations 0.8 simulations
Cumulated Weight Fraction

0.7 0.7
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0.5 1 2 4 8 16 32 64 0.5 1 2 4 8 16 32 64
Particle Size, [mm] Particle Size, [mm]
1 1
0.9
s ℵ = 0.20 0.9
s ℵ = 0.25
tests tests
0.7 0.7
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0.5 1 2 4 8 16 32 64 0.5 1 2 4 8 16 32 64
1 1
s ℵ = 0.30 s ℵ = 0.35
0.9 0.9
tests tests
0.7 0.7
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0.5 1 2 4 8 16 32 64 0.5 1 2 4 8 16 32 64
Figure 15. Test and simulated results from the six series of consecutive compression crushing tests.
20
8.4 SELECTION
The value of selection is obtained directly from the compression tests. For the first test in each
series with the narrow size distribution, defined by +x 1 -x 0 (+16-19 mm), the value of S is
simply obtained as the weight fraction passing size x 1 (16 mm). For the following tests the
value of S is achieved as
x x
W i 1 – W i 1– 1
S i = ----------------------------
x
- (7)
1 – W i 1– 1
1 x
where W i is the weight fraction passing size x 1 after test i and S i – 1 is the value of the
selection function from the preceding test. The achieved values of S are plotted in Figure 16a
against the compression ratio. It is clear that the breakage probability S varies both with the
compression ratio s ℵ and with the feed size distribution σ ℵ. The value of S increases with
increasing s ℵ which is an obvious and natural behaviour. The size distribution width
influences the value of S , so that for short fractions higher values of the breakage probability
are obtained compared to wide fractions.
A two-variable selection function S ( s ℵ, σ ℵ ) given in Eq. (8) can satisfactorily describe the
breakage probability (Paper E). The appearance of the selection function achieved after fitting
Eq. (8) to experimental data is shown in Figure 16. The good fit with the experiments confirms
the assumption about the behaviour of the selection function.
2 2 2 2
S ( s ℵ, σ ℵ ) = a 1 s ℵ σ ℵ + a 2 s ℵ σ ℵ + a 3 s ℵ (8)
2 2
+ a4 sℵ σℵ + a5 sℵ σℵ + a6 sℵ + a7 σℵ + a8 σℵ + a9
The values of the fitted constants a i for the tested material are
a 1 = – 17.20202 a 2 = 1.64983 a 3 = – 4.61729
a 4 = 6.26490 a 5 = 1.41795 a 6 = 4.21650
a 7 = – 0.025766 a 8 = – 1.22568 a 9 = 0.055819
1
0.9
0.8 1
0.7 0.8
σ ℵ = 0.06
0.6 0.6
σ ℵ = 0.86
S
0.5
SS
0.4
σℵ
0.4
0.2
0.3
0
1
0.2
0
0.1 0.5 0.4
0.3
σℵ 0.2
0.1 sℵ
0
0 0.1 0.2 0.3 0.4 0.5 0.6 σ0 0 0
ss/b s/b
ℵ
a) b)
Figure 16. a) Two-variable selection function fitted to experimental data.
b) In 3D the function appears as a surface in space.
21
8.5 BREAKAGE
The breakage behaviour is assumed only to be dependent of the compression ratio. Therefore,
to calibrate a breakage function it is sufficient to use the result from the first test in each series
shown in Figure 15. Normalized size distributions resulting from the compression tests with
different s ℵ-ratios are shown in Figure 17. These breakage functions are obtained by
normalizing the size distribution curve below size x 1 (16 mm) with the value of S in the
corresponding case.
1 1
0.9
Dalby Gneiss 0.9 Diabase
tests Gneiss
0.8 simulations 0.8
0.7 0.7

0.6 0.6
0.5 0.5
0.4 0.4 sℵ
sℵ
0.3 s ℵ = 0.35 0.3 s ℵ = 0.35
0.2 0.2
0.1 0.1
s ℵ = 0.10 s ℵ = 0.10
0 0
0.5 1 2 4 8 16 32 64 0.5 1 2 4 8 16 32 64
a) b)
Figure 17. a) Normalized breakage behaviour of the tested gneiss material (Paper E).
b) Comparison of the breakage behaviour for diabase (Paper B) and gneiss (Paper E).
It is obvious that the overall reduction increases when the s ℵ -ratio increases. A function
given in Eq. (9) with s ℵ as parameter and four constants can be fitted to the experimental data
(Papers B and E). Here x ℵ is a particle size relative to the initial particle size x 0 . The particle
size x s is defined in Eq. (10) in which x min is a small reference particle size ( x min = 0.008
mm).
α + α2 sℵ
B ( x ℵ, s ℵ ) = ( 1 – ( α 3 + α 4 s ℵ ) )x s 1 + ( α 3 + α 4 s ℵ )x s (9)
log 2 ( x ⁄ x min )
x ℵ = ----------------------------------
- (10)
log 2 ( x 0 ⁄ x min )
After fitting Eq. (9) to the data for gneiss in Figure 17 the numerical values of the constants
are (Paper E)
α 1 = 18.9539 α 2 = – 36.2309 α 3 = – 0.0095 α 4 = 0.5657
For comparison data for diabase are also shown in Figure 17 (Paper B). It can be noted that
the two different types of rock materials have a similar breakage behaviour. Though, for a
given compression ratio, gneiss generates more fines than diabase.
Finally, the obtained selection and breakage functions were applied to each test in Figure 15.
The simulated results are in good agreement with the tests, which confirms the overall
breakage behaviour approach and assumptions. It is assumed that the obtained breakage
behaviour can be used in the simulations of repeated size reductions in a cone crusher.
22
9 FLOW MODEL
The crushing process is modelled as a series of repeated crushing events. The flow model must
therefore predict where and under which conditions (how) the material is compressed during
each size reduction cycle. The overall size reduction process is described in Eq. (2). For a
given rock material, the operators S i and B i depend on the dimensionless machine parameter
( s ⁄ b ) u . The flow model should therefore predict the compression ratio at those points where a
material volume is subjected to the crushing events (Papers B and C).
The flow model must be able to handle variations in both design and operational parameters.
The design parameters are all the geometric parameters needed for describing different
chamber profiles or crusher designs. For a given design or crusher, the operational parameters
are closed side setting CSS , eccentric speed n and nominal stroke s nom .
9.1 MATERIAL FLOW MECHANISMS

The flow model developed is based on equations of motion (Paper D). Three different
mechanisms are assumed to describe the material flow, namely sliding, free fall and squeezing.
Sliding occurs when a rock particle is in contact with the mantle and slides downwards. If the
mantle accelerates away rapidly enough, the corresponding particle will fall freely. The free
falling particle can catch up and impinge on the mantle surface. If so, the impact is modelled as
fully plastic. If the mantle is still moving away from the concave, the particle will continue its
motion by sliding.
The motion of particles relative to the mantle surface is assumed to cease and stop instantly
when the particle is in contact with the mantle surface and when the mantle is moving towards
the concave. This assumption implies inherently that bridging of particles occurs for the stated
conditions. Bridging is a self-locking mechanism which forces the particle motion to stop
relative to the mantle surface.
Once the particles have stopped moving, squeezing will start. During squeezing, particles
will be compressed and thereby crushed. Fragmentation by brittle fracture will occur when the
induced tensile stresses exceed a critical value. Particles which are smaller than the local
closed side setting will be exposed to interparticle breakage. These particles will be in contact
with both mantle and concave and thus they will be exposed to single particle breakage during
squeezing.
Depending on the material transport mechanism, three different speed regions have been
identified (Paper D). How the particle is transported in these three speed regions is shown in
Figure 18. During squeezing the velocity of the particle is determined by the mantle motion.
The velocity of the particle relative to the mantle surface will be zero. Shear stresses might
occur depending on the chamber geometry.
At low speeds the motion of the particle is purely determined by sliding. Sliding will occur
until the particle is nipped between the mantle and the concave. As long as the mantle surface
is moving away from the concave, sliding will occur. This is defined as speed region I.
In speed region II the eccentric speed is high enough to accelerate the mantle surface away
from the particle. The particle will then fall freely. At some point the particle will catch up
with the mantle surface. If the mantle is still moving away from the concave the particle will
continue its motion with sliding.
At high speeds the particle will be in free fall and impinge on the mantle at an eccentric
angle greater than π . This is speed region III.
23
Speed Region I Speed Region II Speed Region III

free fall free fall
Iii
hi
slide slide α>π
II
hi
I
hi
α = π
α = π
closed closed
open open
Figure 18. Three speed regions are defined depending on the active mechanism for material
transport (Paper D).
Depending on the speed region the effective stroke that will be used for compression and
crushing will vary. The effective stroke can be equal or less than the local nominal stroke. How
the nominal stroke is utilized, depending on the transport mechanism, is shown in Figure 19. It
is concluded that geometry controls the motion of the particle at low eccentric speeds while
crusher dynamics controls the motion at higher speeds. In speed region I and II the local
nominal stroke is fully utilized and therefore equal to the effective stroke. The effective stroke
is less than the nominal one in speed region III.
Speed Region I Speed Region II Speed Region III

Nominal free
distance slide squeeze fall slide squeeze free fall squeeze
OSS
s eff s eff s eff
CSS
ω = ω1 ω = ω2 ω = ω3 αc
αc = π αc = π αc > π
ωn t
π 2π π 2π π 2π
Figure 19. The effective stroke varies depending on the eccentric speed ( ω 1 < ω 2 < ω 3 ).
By performing repeated strokes it is possible to show how a particle is transported down

through the crusher chamber. As the predominant flow mechanism is free fall, this simulation
will give us the fastest possible way a particle can take. Examples of the resulting paths from
repeated strokes are shown in Figure 20.
24
269 rpm 360 rpm 579 rpm

−0.2 −0.2 −0.2
−0.3 −0.3 −0.3

Y−coordinate [m]
−0.4 −0.4 −0.4
−0.5 −0.5 −0.5
−0.6 −0.6 −0.6
−0.7 −0.7 −0.7
9 steps 11 steps 21 steps

−0.8 −0.8 −0.8
0.25 0.3 0.35 0.4 0.45 0.25 0.3 0.35 0.4 0.45 0.25 0.3 0.35 0.4 0.45
X−coordinate [m]
Figure 20. Repeated strokes show how a particle is transported through the crusher chamber. When
the eccentric speed increases, the number of crushing zones increases. The mantle profile
is shown in its neutral position (Paper D).
From the simulation the so-called crushing zones are defined. A crushing zone corresponds
to a volume of material which is crushed when the eccentricity of the main shaft is turned one
lap. The boundaries of a crushing zone are set by the mantle and concave surfaces together
with lines perpendicular to the mantle surface. These lines intersect the starting and finishing
points of the simulated path for each crushing step. When the locations in space and time are
known for these zones, the geometry module will provide all geometrical data needed for the
size reduction calculations.
9.2 VELOCITY DISTRIBUTIONS

Putting the flow mechanisms and other assumptions together will result in a velocity distribution
which has the eccentric angle (or time) as independent parameter. The velocity distribution will
depend on the eccentric speed. This will in turn reflect the active flow mechanisms. Some
theoretical velocity distributions corresponding to different eccentric speeds is shown in Figure
21.
25
v(α)
Effect of t d Compression
ω3 t d α = ωn t
π 2π
Slide
Free fall
vc ( ω1 ) v c ( ω 3, t d )
vc ( ω3 )
α c ( ω 3, t d )
vc ( ω2 )
αc ( ω1 ) αc ( ω2 ) αc ( ω3 )
Figure 21. Velocity distributions as a function of eccentricity angle α , with eccentric speed as
parameter ( ω 1 < ω 2 < ω 3 ) (Paper D).
Let us introduce a short time delay t d . During this time period the particle does not move
relative to the mantle surface. The time delay provides the possibility to handle feed material
with varying moisture content.
9.3 CAPACITY
Once the velocities of all particles in a horizontal cross-section are known it is possible to
calculate the capacity. In the simulations we will simulate a full crushing cycle of a particle
which starts from rest at the choke-level. The choke-level is defined as the level where the area
of a horizontal cross-section has its minimum. Then we will approximate the velocity field at
the choke level with the velocity distribution of the particle. This simplification is not likely to
give any serious impact since the total falling distance per cycle is quite small (less than 60 mm
at low speeds).
The capacity is calculated by integrating the mass-flow field over a horizontal cross-section
of the crushing chamber. In theory, the vertical level where the integration is carried out can be
chosen arbitrarily. A natural choice is the level where the cross-sectional area has its
minimum. For most crushers this level will be identified as the choke-point which controls the
capacity of the crusher. A horizontal cross-section through the crushing chamber is shown in
Figure 22.
v(α)
v(α)
α Ro
α = 0 Ri ( α )
ωn dr
Concave r
α = αc Mantle
dα
α
Figure 22. Horizontal cross-section of the crushing chamber (Paper D).
26
The mass flow in the downward direction will be

αc Ro αc
1
∫ ∫ ρv ( α )r dr dα = --- ∫ ρ ( R o – R i (α) )v ( α ) dα
2 2
Q down = (11)
2
0 Ri ( α ) 0
R i, min ≤ R i ≤ R i, max (12)
The radius R i ( α ) is the radius from which a particle with speed v ( α ) originates. The bulk
density of the aggregate is denoted ρ.
In the upward direction we will have to divide the mass flow by two as particles in contact
with the concave surface will have zero vertical velocity.
2π R o 2π
1 1
= --- ∫ ∫ ρv ( α )r dr dα = --- ∫ ρ ( R o – R i (α) )v ( α ) dα
2 2
Q up (13)
2 4
αc Ri ( α ) αc
The overall capacity is then calculated as
Q = η v ( Q down – Q up ) (14)
We here introduce a volumetric filling ratio η v . This parameter is equal or less than one
depending on how well filling of the crushing zone is achieved.
So far, we have assumed that the velocity is equal for all points in the radial direction during
the downward motion, see Figure 23. This is a first-order approximation. A more general
approach would be to assume some distribution of the velocity in the radial direction.
R i, max Ro
Ri
R i, min
Choke
level
Concave
Mantle
v(α)
open closed
Figure 23. The assumed velocity distribution for a vertical cross-section (Paper D).
27
10 INTERACTION BETWEEN ROCK FLOW AND SIZE REDUCTION

The flow model predicts the locations and the effective compression ratios of the crushing
zones. So far, no interaction between crusher and rock material has taken place with respect to
size reduction.
The mathematical interaction between the flow and size reduction model comprises the
problem to find a solution, which predicts the size distribution after every crushing step, and
which also fulfils stated conditions (Paper E).
10.1 MASS CONTINUITY

Mass continuity must be preserved between all crushing zones and at any moment in time. The
capacity in the downward direction Q down can be achieved by multiplying the rotational speed
n of the eccentricity by the specific mass flow m i . The specific mass flow is equal to the bulk
density of the aggregate ρ i multiplied by the zone volume V i . This is basically Gauldie’s
approach and is a slight approximation. To preserve continuity means that the value of the
specific mass flow is constant for all zones. It has been found necessary to introduce a
volumetric filling ratio η v, i (Papers D and E). The value of this parameter reflects to what
extent an accessible zone volume is utilized.
Q down = m i n = η v, i ρ i V i n (15)
Though the expression for capacity is rather simple, it is still quite complicated to determine
the value of the specific mass flow for the reason that we do not know the value of the
volumetric filling ratio. The specific mass flow cannot be directly calculated due to the
dependency between bulk density and size distribution. Therefore, an iterative and recursive
calculation procedure is needed.
10.2 VOLUMETRIC FILLING RATIO

The volumetric filling ratio η v, i describes to what extent the accessible zone volume is filled.
The value of η v, i varies from zone to zone. If the filling ratio of the inlet zone could be
determined, the specific mass flow could also be calculated. It would then be possible to
calculate the corresponding value of η v for the other zones. The specific mass flow of course
remains constant between the crushing zones. Otherwise, mass continuity would not be
preserved.
The volumetric filling ratio depends on several factors. It is affected by the design of the
crusher inlet, rheology of the feed material, size distribution (especially in relation to the inlet
opening), moisture content, etc. The value of the filling ratio is a measure of how much air
there is in the rock material aggregate to be crushed. For the application studied, the value of
η v has been found to be significantly below one. The possibility to decrease the closed side
setting of the crusher is determined by the value η v . A low value of η v corresponds to a large
proportion of air, which in turn allows a relatively small value of the closed side setting CSS .
If the feeding conditions were altered such that the filling ratio was increased, the capacity
of the crusher would increase in theory. On the other hand, if the feeding of the inlet zone was
made more efficient, the packing density might be exceeded at some stage. Thereby the closed
side setting would have to be increased, leading to a decreased overall size reduction. This
phenomenon has been observed in operating crushers [28] and is not always a desired effect.
It is assumed that small changes in closed side setting do not affect the filling ratio for the
inlet zone. Test results for a crusher operating on the limit to packing can therefore be used to
28
calibrate the filling ratio for the inlet zone. During the simulations with different closed side
settings the value of η v was kept constant independently of the closed side setting. If no full-
scale tests are available for calculating the volumetric filling ratio, the inlet of the crushing
chamber needs to be modelled in detail.
The direction of interdependence between the parameters involved in the size reduction
predictions is shown in Figure 24. The density of the material entering the first zone is known,
as it is equal to the uncompressed bulk density ρ 1, u of the incoming feed material. Of all
crushing zones, only one single zone can control the final capacity. A recursive type of
calculation scheme is needed to find this zone.
Geometry Crusher dynamics

 --s- 
 b nom  --s- 
 b eff 

 Q  Size reduction process
---- 
ηv  n  --s-
Feeding conditions    b u
ρ 1, u  Rock breakage behaviour

ρ i, u Size
Distribution
CSS
Figure 24. Direction of interdependence for a size reduction cycle.
A condition which must be fulfilled to avoid packing is that the density after every
compression step must be lower than the solid density ρ s .
ρ i, compressed < ρ s (16)
The maximal possible compression ratio is achieved when all air is compressed out of the
aggregate system. Packing will be avoided as long as the following condition is fulfilled:
ρi
 --s- <  1 – ----
- (17)
 b u  ρ s
The packing condition must be checked for every crushing zone. If the condition is fulfilled,
the calculation may proceed with size reduction and then continue with the next zone. If the
condition is violated, the value of the volumetric filling ratio has to be decreased, and we have
to start it all over again with zone number one. The calculation procedure is shown in
Figure 25.
29
Feed
From zone n-1
ηv = 1
Zone i
ρi
 --s- >  1 – ---- ρi
 --s- <  1 – ----
Decrease η v - -
Zone 1  b u  ρ s  b u  ρ s
Packing Compression
Zone i
Size Reduction
Zone N
To zone 1
To zone n+1
a) Product b)
Figure 25. The iterative and recursive calculation procedure used for determination of the
volumetric filling ratio.
10.3 UTILIZED COMPRESSION RATIO

Due to the volumetric filling ratio, only a fraction of the total nominal stroke will be utilized for
size reduction. A part of the effective stroke will act as an idling stroke. Size reduction cannot
begin until the uncompressed density of the aggregate is reached.
The nominal compression ratio ( s ⁄ b ) nom is given from the geometric analysis of the
crushing chamber based on design drawings. The effective compression ratio ( s ⁄ b ) eff
corresponds to the ratio which is possible to utilize after considering crusher dynamics. The
utilized compression ratio ( s ⁄ b ) u is less than the effective ratio and is a result of mass
continuity and requirements such as packing. The relation between nominal, effective and
utilized compression ratio is therefore:
 --s- <  --s- <  --s- (18)

 b u  b eff  b nom
In Figure 26 a schematic representation of a crushing zone is shown. After taken the idling
stroke s t into account, the utilized compression ratio ( s ⁄ b ) u can be calculated.
30
b nom
s nom
oi
hi
st su
bu
s eff
b eff
Figure 26. A schematic representation of a crushing zone.
The utilized bed thickness can be calculated from the specific mass flow Q ⁄ n , density ρ i ,
crushing zone height h i and zone circumference o i
Q 1 1 η v, 1 ρ 1 V 1 1
b u = ---- ⋅ ---- ⋅ --------- = ----------------------- ⋅ --------------------- (19)
n ρi hi oi ρi h i ( 2πR i )
The utilized stroke can then be calculated:
s u = s eff – b eff + b u (20)
Finally, the utilized compression ratio is obtained as the ratio between utilized stroke and
bed thickness:
su
 --s- = ----
- (21)
 b u bu
10.4 RELATION BETWEEN BULK DENSITY AND SIZE DISTRIBUTION

A model which relates size distribution width to uncompressed bulk density of aggregates is
needed for the calculations, as the density will vary between the crushing zones. The density
model developed was calibrated using results achieved from full-scale crushing tests, see Figure
27a (Paper D). To make the size distribution model independent of the specific rock type, a
normalized density ρ ℵ was introduced. The normalised density is achieved by dividing the
uncompressed bulk density of the aggregate ρ achieved from the tests by the solid density of
the rock material ρ s . The measured relationship between normalized size distribution and
normalized density is shown in Figure 27b.
31
ρ
ρ ℵ = ----- (22)
ρs
, y
100 1
90 0.9
80 0.8
theoretical
70 0.7 2
3
Cumulated Weight [%]
60 0.6 7
6 5 4
ρℵ
50 0.5
1
40 0.4
2 3
30 0.3
4 5
20 0.2
6
10 7 0.1
0 0
0.5 1 2 4 8 16 32 0 0.2 0.4 0.6 0.8 1
Particle Size [mm] σℵ
a) b)
Figure 27. a) Product size distributions used to calibrate the density model for gneiss.
b) Normalized density as a function of normalized size distribution width.
The use of typical size distributions originating from full-scale tests implies that the model is
valid in the region of practical interest. However, it was found during the size reduction
simulations that wider size distributions could be achieved than the ones achieved from full-
scale tests. For this reason, a theoretical point with σ ℵ = 1 was introduced, corresponding to
the widest possible size distribution.
The relation between ρ ℵ and σ ℵ is assumed to follow the expression given in Eq. (23),
which was fitted to the experimental data.
3 2
ρ ℵ = D 1 σ ℵ + D 2 σ ℵ + ρ ℵ, mono (23)
The constant ρ ℵ, mono corresponds to the normalized density of an aggregate with a very
narrow size distribution, i.e. a mono-size distribution with σ ℵ = 0 . After fitting Eq. (23) to
the test results, the following values of the parameters were achieved (Paper E)
ρ ℵ, mono = 0.54284 D 1 = – 0.062277 D 2 = 0.24758
The normalized test results from the density measurements together with the adapted data
are shown in Figure 27b.
10.5 CONFINEMENT AND BREAKAGE MODES

Confined conditions are required if interparticle breakage is to occur when a bed of particles is
compressed. The particle bed could either be fully or partly confined. If no confinement at all is
achieved then only single particle breakage can occur.
For choke-fed conditions, the inlet zone together with the zones down to and including the
choke zone will be choke fed and therefore also confined. This implies that interparticle
breakage occurs. Single particle breakage will also occur for particles with a particle size
larger than the local closed size setting. For the crushing zones below the choke zone the
volume increase makes the forming of confined particle beds unlikely. Therefore, it is assumed
32
that only single particle breakage will occur in these zones.
Feed material level
Inlet
1st zone
2nd
Confined crushing zones
3rd -interparticle breakage can occur
4th
Choke level
5th
6th
7th A IP
No confinement
8th -only single particle breakage Choke level
9th A tot
10th
11th b)
Outlet
Y a)
Figure 28. a) General appearance of the horizontal cross-sectional area of a crusher.

b) A Crushing zone which is cut by the choke level.
Considerable proportions of interparticle breakage in the lowest zones would only be

possible under the conditions that the material flows in rills. This type of flow seems very
unlikely to occur, which implies that forming of particle beds is rare and thus very little
interparticle breakage occurs.
The number of crushing zones does not necessarily need to be an integer number. If the
choke level cuts through a crushing zone, interparticle breakage is only achieved in the upper
part of the crushing zone and single particle breakage in the lower part, see Figure 28b. The
amount of interparticle breakage ξ IP is assumed to be equal to the ratio between the area of the
upper part and the total area of the crushing zone.
A IP
ξ IP = --------
- (24)
A tot
The product size distribution descending from the partly confined zone will be a mix of
particles originating from interparticle breakage and single particle breakage:
p = ξ IP p IP + ( 1 – ξ IP )p SP (25)
10.6 CHOKE FEEDING

Over the years the idiomatic expression choke feed has been established to describe the feeding
conditions for a cone crusher. A choke fed crusher is a crusher with an inlet totally covered with
feed material. The common opinion is that choke fed conditions are vital for achieving
interparticle breakage, which in turn is a breakage mode that governs good particle shape.
A common solution for achieving choke feeding is to use a small bin on top of the crusher.
33
The level of the feed material is kept between an upper and a lower limit by some control
device. This type of solution inherently assumes that the feeding conditions are not affected by
the pressure exerted by the material volume above the inlet (cf. pressure in fluids). There are
examples where the level of the material surface in the bin affects the crushing parameters
leading to variations in the final product [28].
The author agrees with the statement that choke feeding is necessary for achieving a stable
feed (stationary) to the inlet of the crusher. A stable feed guarantees minimal variations in the
final product. However, the currently used definition of choke fed conditions might not be a
sufficient requirement to achieve proper feeding conditions. The pressure exerted by the bulk
material in the bin on top of the crusher will affect the volumetric filling ratio as well as the
confinement of the first crushing zone. Requirements on the feed material level may also be
necessary.
10.7 A PARTICLE SHAPE MODEL FOR SINGLE PARTICLE BREAKAGE

From a simulation point of view, there is a great difference between interparticle and single
particle breakage concerning particle size. In the calculations, all particles will be treated
independently of particle shape, as if they were one-dimensional. To handle the shape problem
a shape model describing the distribution of the variation in geometry is needed.
A problem arises when simulations are compared with results from sieving. Up until now
the crushing simulations are one dimensional while sieving is a two-dimensional method.
Obviously, to compare simulations with sieving requires some extra care.
For interparticle breakage, the local closed side setting, i.e. the distance between mantle and
concave, is always larger than the particle size. The particles originating from interparticle
breakage will have no geometric calibration in terms of particle size. By definition, in the case
of single particle breakage, the local closed side setting is always smaller than the particle size.
The particle will be in contact with both mantle and concave. The resulting product is thereby
calibrated by the crusher.
In reality, there is always some amount of over-sized material compared to the closed side
setting at the crusher outlet. A minor amount of this material might originate from a crushing
zone located a distance upstream the crushing chamber. The major part of the over-size is
explained by the fact that the single particle breakage mode generates miss-shaped particles.
Such a particle can have a smallest size t which is calibrated by the local closed side setting of
the crusher, while it also has a width d which can be significantly larger, see Figure 29. At
sieving, such a particle will appear as a particle of size d , while it was treated as a particle of
size t during the simulations.
Figure 29. An elongated particle where t<d<l.
One way to eliminate the problem would be to perform single particle breakage tests. The
34
breakage function describing single particle breakage would then be achieved by laboratory
sieving. Miss-shaped over-sized particles would then inherently be included in the breakage
function. In this work the breakage function achieved from the interparticle compressing
crushing tests is intended to be used. As these tests are purely interparticle, a model capable of
estimating the over-sized material is needed. Here it is assumed that a fraction γ of the
particles broken in the single particle mode is elongated. Further, it is assumed that the
geometry of these flaky particles can be approximated with a ratio between particle width d
and thickness t , which is denoted λ .
d
λ = --- (26)
t
Single particle breakage will occur if the thickness of a particle is greater than the local
closed side setting.
d
t = --- > CSS local = b – s (27)
λ
This condition is valid for nominal, effective and utilized bed thickness and stroke, see Figure
26.
Further, the compression ratio, which a particle of size d is exerted to, is
λ(b – s)
s ℵ, single = 1 – ------------------- (28)
d
The values of the parameters in the particle shape model chosen are λ = 2 and γ = 0.5 .
The expression describing the size reduction is achieved by modifying Eq. (3). The expression
now consists of two parts, one for the flaky and one for the non-flaky cubical particles.
inter inter,flaky single,flaky single,flaky
pi + 1 = γ { [ Bi S i + ( I – S i ) ]M i + Bi Mi }p i + (29)
inter inter single single
( 1 – γ ) { [ Bi S i + ( I – S i ) ]M i + Bi Mi }p i
35
11 SIMULATION RESULTS AND EXPERIMENTS

To validate the crusher flow and size reduction models full-scale tests were performed at
realistic operating conditions. During the tests both the eccentric speed and the closed side
setting were altered. The crusher chosen for these tests was a Svedala Hydrocone equipped
with an fine-crushing chamber.
11.1 ACCURACY
Computed results are rarely exact but if the difference between the computed result and the
real behaviour of a system is small enough, the error is of no importance. If there is an
unacceptable difference there is also a need to identify the origin of the error.
Modelling errors occur if there is a difference between the physical system and the
mathematical model.
Feed and product in the size reduction process are described by size distributions. These
distributions are cumulative functions of a linear size of the particles. The use of a linear size is
perhaps an easy and convenient description but it does not tell anything about the shape of the
particles. The linear size is defined by a sieve analysis. The size of a particle is then smaller
than the width of a square aperture, which the particle can pass through, and it is larger than a
width, which it cannot pass through. Single particles, which are assigned a certain size from
the sieve analysis, can have completely different shapes and will therefore also be likely to
behave differently during breakage.
There is also a loss in accuracy in the transformation from the sieve analysis data to the
frequency vector, which is used to characterize the feed material coming to the reduction
process. The approximation of a continuous distribution by a discrete vector will therefore give
rise to a discretization error which has been obvious when working with this project.
11.2 FLOW
Simulations of particle flow are conducted in Paper D. In the simulations the time delay t d was
set to 0.01 second. The simulated vertical position of a particle starting from the choke-level is
shown in Figure 30. The predominant flow mechanism is free fall. Sliding occurs only at low
eccentric speed. The total covered distance in the vertical direction decreases with increasing
eccentric speed. One notice that if the nominal stroke is to be fully utilized the eccentric speed
chosen should be just below 300 rpm.
The velocity distributions are shown in Figure 31. The influence of the constant time delay
on velocity is clearly seen. The velocities in the upward direction are substantial compared
with the ones in the downward direction. However, the upward velocities should be divided by
a factor of two before comparison, as the velocities on the concave surface have zero velocity.
In Figure 32 the velocities are shown in a polar plot. Here it is clearly demonstrated how the
eccentric angles are utilized. For the lowest speed the nominal stroke is fully utilized while for
the highest only half of the nominal stroke is utilized.
By repeated simulations it is possible to show how a particle is transported through the
crusher chamber. As the predominant flow mechanism is free fall, this simulation will give us
the fastest possible way a particle can take. The resulting path of a repeated simulation is
shown in Figure 33. For a given eccentric speed the height of the crushing zones does not seem
to change significantly. With increasing eccentric speed the number of crushing zones
increases. The effective bed thickness and stroke decreases when the eccentric speed increases.
36
−0.43
Effect of t d
Start position
Fully open
−0.44
Particle trace
Eccentric speed
−0.45 600 rpm
Vertical position, rY, [m]
Free fall 550

500
−0.46 450
Mantle surface 400
350
−0.47 Compression 300
−0.48 250
Slide
−0.49
−0.5
0 1 2 3 4 5 6 7
Eccentricity angle, α , [rad]
Figure 30. Simulated vertical position of a particle at different eccentric angles. For the studied
crusher, free fall is the predominant flow mechanism ( t d = 0.01 s).
0.4
Effect of t d
0.2
Compression
0
Velocity vY, [m/s]
−0.2
−0.4 Free fall
−0.6
600 rpm
Slide 550
500
−0.8 450
400
350
250 300
−1
0 1 2 3 4 5 6 7
Eccentric angle, [rad]
Figure 31. Velocity distributions in the vertical direction at choke-level ( t d = 0.01 s).
37
90°
1 m/s
120° 60°
150° 0.5 30°
250
180° 0°
α c, 300 300
210° 350 330°

600 rpm
400 550
450 500
240° 300°
270°
Figure 32. Polar plot showing the magnitude of the velocity distributions in vertical direction at
choke-level ( t d = 0.01 s).
269 rpm 360 rpm 579 rpm

−0.2 −0.2 −0.2
−0.3 −0.3 −0.3

Y−coordinate [m]
−0.4 −0.4 −0.4
−0.5 −0.5 −0.5
−0.6 −0.6 −0.6
−0.7 −0.7 −0.7
9 steps 11 steps 21 steps

−0.8 −0.8 −0.8
0.25 0.3 0.35 0.4 0.45 0.25 0.3 0.35 0.4 0.45 0.25 0.3 0.35 0.4 0.45
X−coordinate [m]
Figure 33. Repeated simulations show how a particle is transported through the crusher chamber.
When the eccentric speed increases the number of crushing zones increases.
38
11.3 SIZE REDUCTION

Size reduction simulations were conducted in Paper E. Size distribution tests were performed
at the same time as the flow tests. The simulations started with calibration of the volumetric
filling ratio η v according to Figure 25. From the full-scale tests it was found that packing
occurred at a closed side setting of 7.2 mm. The volumetric filling ratio of the inlet zone were
determined to be 0.70 for this case. For the simulations η v was assumed to be constant
independently of the closed side setting.
Simulated results, together with measured feed and product size distributions for three
different eccentric speeds and for two different closed side settings, are shown in Figure 34.
The same cases but in more detail, with the size distributions after each reduction event, are
shown in Figure 35. The corresponding compression ratios are shown in Figure 36.
The model response to changes in closed side setting matches the full-scale tests very well.
The number of interparticle crushing zones (#IP) decreases when the closed side setting
increases. This is a consequence of an upward movement of the choke level with increasing
CSS . The effect is strongly dependent on the specific design and is therefore not necessarily a
general behaviour for cone crushers. The most important effect associated with increasing
CSS is a decreasing compression ratio. When compression ratio decreases, the overall size
reduction also decreases. This effect is obvious, but has never been proven by this type of
simulation before. The combined effect of decreasing number of crushing zones and
compression ratio is an overall decrease in size reduction.
The model response when the eccentric speed is changed is very interesting but difficult to
interpret. Due to the relatively small difference in overall size reduction for both simulations
and for full-scale tests it is difficult to draw quantitative conclusions by only comparing the
size distributions for the resulting product, see Figure 34. However, by studying how the size
distributions evolve from each crushing event, considerable differences are discovered. The
number of interparticle crushing zones increases significantly when the eccentric speed
increases. Simultaneously, the compression ratio decreases as a dynamic effect of increasing
speed (Paper B and E). Under these conditions it is remarkable that such small differences are
achieved for both simulations and full-scale tests. The explanation is probably to be found in
the location of the choke level. If the choke level is placed at the outlet of the crushing
chamber, a more pronounced dependence on eccentric speed is to be expected (Paper D).
The main part of the total size reduction is achieved by interparticle breakage. The number
of interparticle breakage events is relatively small in all of the studied cases. A typical number
is between 2.5 and 4.5.
After the zones of interparticle breakage and down to the outlet, the breakage mode is purely
single particle. Here, the material is mainly calibrated by the crusher, and the resulting size
reduction is relatively small in this region. The number of single particle crushing events is
between 6 and 10.
In terms of reduction efficiency, the part of the crushing chamber above the choke level is
more efficient compared to the lower part. The calibration zone is commonly defined as the
part of the chamber above the outlet where the profile mantle and concave surfaces are nearly
parallel. It is often claimed that the calibration zone governs particle shape. This zone is
located below the choke point and thus the breakage mode is single particle. An inappropriate
design of the calibration zone can be devastating to particle shape [6].
A closer inspection of the agreement between simulated results and full-scale tests shows a
systematic over-prediction of the overall size reduction for the simulation. This is probably an
effect which originates in the assumption that the zones where interparticle breakage occurs
39
are fully confined. If the degree of confinement is lower, the values of the selection and
breakage functions will decrease and thus also the overall size reduction.
If repeated size reduction simulations are conducted with different values of one or several
parameters the performance behaviour of the crusher is obtained. In Figure 37a size
distributions are shown for varying closed side setting. The corresponding Crusher
Performance Map is shown in Figure 37b. For comparison, full-scale test results are shown in
Figure 38. The similarities between simulations and test results are convincing. The qualitative
agreement is especially good in terms of the locations of maximal relative capacities together
with the trends of the different fractions.
1 1
0.9 0.9
0.8 0.8

0.7 0.7
CSS=8 CSS=8
0.6 0.6
CSS=12 CSS=12
0.5 0.5
Feed Feed
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0.5 1 2 4 8 16 32 64 0.5 1 2 4 8 16 32 64
Particle Size [mm] Particle Size, [mm]
a) b)
Figure 34. a) Simulated results for three different eccentric speeds and two different CSS.
b) Full-scale tests.
40
1 1
0.9 n=269 rpm 0.9 n=269 rpm

css=8 mm css=12 mm
0.8 #IP=2.99 0.8 #IP=2.55

0.7 0.7
simulation simulation
0.6 feed 0.6
feed
product product
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0.5 1 2 4 8 16 32 64 0.5 1 2 4 8 16 32 64
1 1
0.9 n=360 rpm 0.9 n=360 rpm

css=8 mm css=12 mm
0.8 #IP=3.89 0.8 #IP=3.29
0.7 0.7
simulation simulation
0.6
feed 0.6
feed
product product
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0.5 1 2 4 8 16 32 64 0.5 1 2 4 8 16 32 64
0 d 0 d
1 1
0.9
n=429 rpm 0.9
n=429 rpm
css=8 mm css=12 mm
0.8 #IP=4.74 0.8 #IP=4.02
0.7 simulation 0.7 simulation

feed feed
0.6 0.6
product product
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0.5 1 2 4 8 16 32 64 0.5 1 2 4 8 16 32 64
Figure 35. Simulated size distributions after each reduction step for a Svedala Hydrocone.
41
0 d 0 d
−0.3 −0.3
n=269 rpm n=269 rpm

−0.35 −0.35
css=8 mm css=12 mm
−0.4
#IP=2.99 −0.4
#IP=2.55
−0.45 −0.45
Y−coordinate [m]
Y−coordinate [m]
−0.5 nominal −0.5 nominal
effective effective
−0.55 utilized −0.55 utilized
−0.6 −0.6
−0.65 −0.65
−0.7 −0.7
−0.75 −0.75
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80
(s/b)−ratio (s/b)−ratio
−0.3 −0.3
n=360 rpm n=360 rpm

−0.35 −0.35
css=8 mm css=12 mm
−0.4 #IP=3.89 −0.4 #IP=3.29
−0.45 −0.45
Y−coordinate [m]
Y−coordinate [m]

effective effective
−0.6 −0.6
−0.65 −0.65
−0.7 −0.7
−0.75 −0.75
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80
0 d 0 d
−0.3 −0.3
−0.35 n=429 rpm −0.35 n=429 rpm

css=8 mm css=12 mm
−0.4 #IP=4.74 −0.4 #IP=4.02
−0.45 −0.45
Y−coordinate [m]
Y−coordinate [m]

effective effective
−0.6 −0.6
−0.65 −0.65
−0.7 −0.7
−0.75 −0.75
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80
Figure 36. Compression ratios at different settings and speeds originating from simulations for a
Svedala Hydrocone crusher.
42
p p
1 0.45
0.9 0.4
0.8
0.35
0.7
0.3 11-16
Relative Capacity
4-8
0.6 0-2
0.25
0.5 8-11
8
S=
0.2 16-32
CS
0.4 Feed
0.15
0.3 2-4
16
S=
0.1
0.2
CS
0.1 0.05
0 0
0.5 1 2 4 8 16 32 64 8 9 10 11 12 13 14 15 16
Particle Size [mm] CSS [mm]
a) b)
Figure 37. a) Simulated product size distributions at eccentric speed 269 rpm.
CSS varies with 1 mm steps between 8 and 16 mm.
b) Simulated Crusher Performance Map for eccentric speed 269 rpm.
1 0.45
0.9 0.4
0.8
0.35
11-16
0.7
0.3
Relative Capacity
0.6 8-11 16-32

0.25
4-8
0.5
0-2
8
0.2
S=
Feed
CS
0.4
0.15
0.3
16
0.1 2-4
S=
0.2
CS
0.1 0.05
0 0
0.5 1 2 4 8 16 32 64 8 9 10 11 12 13 14 15 16
Particle Size [mm] CSS [mm]
a) b)
Figure 38. a) Product size distributions from full-scale tests at eccentric speed 269 rpm.
CSS varies with 1 mm steps between 8 and 16 mm.
b) Crusher Performance Map from full-scale tests at eccentric speed 269 rpm.
43
11.4 CAPACITY
By integration of the velocity distributions multiplied with the volumetric filling ratio and the
density, the total capacities for the different eccentric speeds are obtained. The simulated
capacities together with measured are shown in Figure 39. The simulation was conducted with
time delay set to 0.01 second and with a filling ratio set to 0.70 as in the size-reduction
simulations.
140
120 t d = 0.01
η v = 0.70
100
Capacity, Q [tph]
80
CSS=15 mm
60 CSS=12 mm
CSS= 8 mm
40
20
0
250 300 350 400 450 500 550 600
Eccentric speed, n [rpm]
Figure 39. Simulated capacities (lines) compared with full-scale tests (x-marks). The simulated
capacities are calculated for three different CSS at the same eccentric speeds as is the
full-scale tests (269, 360 429, 514 and 579 rpm).
The capacities were evaluated at the choke level. The volumetric filling ratio of the inlet
zone were held constant between the simulations. This implies that the filling ratio at the choke
level will vary as the density varies with the size distribution of the material entering this level.
44
12 DESIGN CONSIDERATIONS
crusher, viz. breakage modes, number of crushing zones and compression ratio. These main
factors provide a possibility for a fundamental and detailed understanding of how a cone
crusher operates. The factors are affected by both design and operating parameters. Any design
consideration should be evaluated against the three main factors.
Interparticle breakage governs particle shape whilst single particle breakage is most energy
efficient [6]. If a number of interparticle crushing zones is followed by a calibration zone it is
strongly recommended that this zone is parallel or slightly divergent. In this way the
probability to produce miss-shaped elongated particles close to the crusher outlet is minimized.
The number of crushing zones is controlled by crusher height and eccentric speed. A crusher
which has the choke zone relatively high up the chamber could probably be shortened from the
bottom with a couple of crushing zones without any significant loss in overall size reduction.
For a crusher with the choke zone near the outlet of the chamber, the first interparticle
crushing zones will have very little affect on the size reduction. In these zones the ( s ⁄ b ) -ratios
are so low that the value of the selection function will approach zero and thus not contribute to
any size reduction.
The compression ratio is mainly controlled by chamber profiles, crusher dynamics and
volumetric filling ratio. Together these factors results in the utilized compression ratio ( s ⁄ b ) u .
In turn the utilized compression ratio determines the values of selection and breakage.
The breakage behaviour of rock materials seems to be strongly non-linear given the
approach in this work. For the gneiss and diabase material studied in Paper B and E the
combined behaviour of selection and breakage is obviously non-linear, see Figure 16 and 17. It
is likely that there exists a “best way” to achieve a desired size reduction. This way could be
formulated as a size reduction strategy. The size reduction strategy comprises the optimal
choice of ( s ⁄ b ) -ratios, number of crushing events and breakage modes.
The optimal size reduction could be found by iterative calculations in which the selection
and breakage functions are utilized. The size reduction strategy is achieved by the crusher
machine as shown in Figure 40. For a given crusher concept, e.g. a cone crusher, there will be
theoretical limitations in that is possible to achieve. In a case where the crusher concept limits
the size reduction strategy there will be a potential for innovations. New crusher types could be
synthesized originating from the strategy.
Size Reduction
Strategy
Crusher
Feed Product
Figure 40. A size reduction strategy can be applied to the transformation of the feed to the product.
45
13 CONCLUSIONS
13.1 GENERAL
A calculation method with a modular structure based on a deductive crusher model has been
developed. The presented model is a comprehensive analytical tool for predicting cone crusher
performance and provides detailed understanding of how and where size reduction is achieved
in a cone crusher. Full-scale tests has been performed to validate the model. The agreement
between simulated results and experimental data is very good. The model is capable of
establishing a direct correlation between crusher design parameters, rock fragmentation
behaviour and crusher performance. The model can be used as a simulation tool to assist in the
design process of cone crushers. Any arbitrary design can then be studied.
In the crusher model there are only two parameters which are unknown and has to be tuned.
The time delay parameter t d is used to compensate for varying moisture content. The
volumetric filling ratio η v is determined by iterative calculations in which the packing criteria
is checked.
Data governing the detailed understanding of how a cone crusher works is achieved from the
simulations. Detailed information about how different machine parameters affect the size
reduction process in every crushing zone is provided by the model. Different crusher
behaviours can be explained depending on crusher layout. The identity of a crusher brand or
model can be characterized by a couple of crusher related parameters. Different machine
layouts can be compared and evaluated early at the design stage. Improvements in crusher
design can be done without excessive testing.
crusher. These factors are: breakage modes, number of crushing zones, and compression ratio.
The main factors are affected by both design, rock material and operating parameters. For a
given crusher application, the factors depend on eccentric speed, closed side setting and feed
size distribution. The main factors provide the possibility for fundamental and detailed
understanding of how a cone crusher operates. Any design consideration should be evaluated
against these three main factors.
If a set of simulations is performed for a given crusher, a Crusher Performance Map is
achieved, which in turn can be used when optimizing a given crushing task or a crushing plant.
The Crusher Performance Map describes how the product output from the crusher varies with
machine parameters. By applying optimization criteria to the Crusher Performance Map the
best operating conditions can be found. It appears that the output from a tentative cone crusher
strongly depends on the eccentric speed. For the studied crusher the dependence with eccentric
speed was vague
The cone crusher performance model gives manufacturers the possibility to tailor make
customized crushing chambers. With the three main factors as design considerations new
crushing chamber geometries can be synthesized to fulfil the desired behaviour of the crusher.
Desired values of the main factors are in turn achieved by choosing correct values of the
design and operational parameters. The potential for improvements is high and therefore also
of great interest for the industry. The improved understanding of the size-reduction process
also governs the possibility for synthesis of knew crusher concepts.
46
13.2 FLOW
From the derivation of the flow model together with simulated results, a number of interesting
conclusions can be drawn. The validity of the flow model is achieved through the agreement
between simulated results and full-scale test data. Especially, the response of the model to
changes in closed side setting is very promising. In turn, this fact verifies that the cross-
sectional area at the choke-level is the controlling factor of the overall capacity.
The flow model provides a fundamental understanding of the flow mechanisms occurring in
a cone crusher. These mechanisms have earlier been described only briefly.Free fall is the
predominant flow mechanism at typical eccentric speeds. Standard eccentric speeds chosen by
crusher manufacturers seem not to fully utilize the nominal stroke at the choke point. The
choice of a specific eccentric speed is probably empirically found and reflects the complex
situation between crusher capacity and overall size reduction.
The simulations show that there is a substantial flow of material upwards due to the uplift-
effect. This upward flow of material will reduce the overall capacity and cannot be ignored.
The same effect is necessary for the horizontal (radial) transport of the material from the inlet
to the outlet of the crushing chamber.
By simulating the fastest way through the crusher it is shown that the number of crushing
zones is about twice as high than previously believed.
13.3 SIZE REDUCTION

The size reduction model is based on the laws of mechanics and constitutive relations
concerning rock breakage characteristics. There are some crucial assumptions which are of
central interest for the model. The validity of these assumptions are verified by full-scale tests.
Early in the development of the crusher model it was claimed that the overall size reduction
process is a result of several subsequent crushing events. Regarding the process model, an
important property is that it must be closely physically defined. This is guaranteed by
modelling size reduction in a cone crusher as a repetitive process.
Each crushing event is modelled with a selection and a breakage function. The appearance
of the selection and breakage functions is rock material specific and can be obtained by
laboratory tests. The fact that laboratory tests can be utilized for prediction of cone crusher
performance is important and will in time reduce the need for full-scale testing.
Two different modes of breakage are possible to achieve in a cone crusher. The criterion
determining which breakage mode acts where is the location of the choke level. The
dominating breakage mode above the choke level is interparticle breakage while mostly single
particle breakage is achieved below this level.
13.4 FUTURE WORK

There are numerous improvements and extensions that can be made in the future. At present
the simulation tool is able to predict both crusher capacity and product size distribution from
cone crushers. There are no inherent limitations to extending the model to also predict other
desired features, such as product quality, crusher wear and power consumption.
Formulation of size reduction strategies will further gain the fundamental understanding of
the size reduction process. The size reduction process does not necessary need to utilized form
conditioned compressive crushing, but could use any comminution principle.
47
48
REFERENCES
[1] Barnard, W. J. and Bull, F. A., “Primary Breakage of Brittle Particles”, Fourth Tewksbury
Symposium, Melbourne, Australia, February 1979.
[2] Bearman, R. A., “The Application of Rock Mechanics Parameters to the Prediction of
Crusher Performance”, Ph.D. Thesis, Camborne School of Mines, England, 1991.
[3] Bond, F. C., “Third Theory of Comminution”, Min. Eng. Trans. AIME, Vol. 193, 484-
494, 1952.
[4] Briggs, C. A. and Bearman, R. A., “An Investigation of Rock Breakage and Damage”,
Minerals Engineering, Vol. 9, 489-497, May 1996.
[5] Briggs, C. A, “Fundamental Model of a Cone Crusher”, Ph.D. Thesis, University of
Queensland, Australia, July 1997.
[6] Briggs, C. A. and Evertsson, C. M., “Shape Potential of Rock”, Minerals Engineering,
Vol. 11, 125-132, February 1998.
[7] Broadbent, S. R. and Callcott, T. G., “Coal Breakage Processes, I. A New Analysis of
Coal Breakage Processes”, J Inst Fuel, 524-528, December 1956.
[8] Broadbent, S. R. and Callcott, T. G., “Coal Breakage Processes, II. A New Analysis of
Coal Breakage Processes”, J Inst Fuel, 528-539, December 1956.
[9] Buss, B., Hanische, J. and Schubert, H., “Über das Zerkleinergungsverhalten seitlich
begrenzter und nicht-begrenzter Kornschichten bei Druckbeanspruchung”, Neue
Bergbautechnik, 12, (5), 277-283, 1982.
[10] Eloranta, J., “Influence of Crushing Process Variables on the Product Quality of Crushed
Rock”, Ph.D. Thesis, Tampere University of Technology, Tampere, Finland, 1995.
[11] Evertsson, C. M., “Prediction of Size Distributions from Compressing Crusher
Machines”, Proceedings EXPLO 95 Conference, Brisbane, 173-180, 4-7 September
1995.
[12] Evertsson, C. M., “Prediction of Cone Crusher Performance”, Lic. Eng. Thesis, Machine
and Vehicle Design, Chalmers University of Technology, Göteborg, Sweden, January
1997.
[13] Evertsson, C. M. and Bearman, R. A., “Investigation of Interparticle Breakage as
Applied to Cone Crushing”, Minerals Engineering, Vol. 10, 199-214, February 1997.
[14] Evertsson, C. M., “Output Prediction of Cone Crushers”, Minerals Engineering, Vol. 11,
215-232, March 1998.
[15] Evertsson, C. M., “Modelling of Flow in Cone Crushers”, Minerals Engineering, Vol.
12, 1479-1499, December 1999.
[16] Evertsson, C. M., “Size Reduction in Cone Crushers”, Minerals Engineering Conference
'99, Falmouth, England, 22-24 September 1999, (Accepted for publication in Minerals
Engineering).
[17] Gauldie, K., “Performance of Jaw Crushers”, Engineering, 456-458, October 9, 485-486,
October 16, 1953.
[18] Gauldie, K., “The Output of Gyratory Crushers”, Engineering, 557-559, April 30, 1954.
[19] Gilvarry, J. J., “Fracture of Brittle Solids, I. Distribution Function for Fragment Size in
Single Fracture (Theoretical)”, J Applied Physics, Vol. 32, 391-399, March 1961.
49
[20] Griffith, A. A., “The Phenomena of Rupture and Flow in Solids”, Phil Trans Royal Soc,
Vol. 7, 29-36, 1933.
[21] Heikkilä, P., “Improving the Quality of Crushed Rock Aggregate”, Ph.D. Thesis,
Helsinki University of Technology, Helsinki, Finland, 1991.
[22] Hultén, J., “Drum Brake Squeal”, Ph.D. Thesis, Machine and Vehicle Design, Chalmers
University of Technology, 1998.
[23] Kick, F., “Contribution to the Knowledge of Brittle Materials”, Dinglers Journal, Vol.
247, 1-5, 1883.
[24] Klimpel, R. R. and Austin, L. G., “The Statistical Theory of Primary Breakage
Distribution for Brittle Materials”, Trans. SME/AIME, 232, 88-94, 1965.
[25] Kojovic, T., “Crushing Aggregate and Shape Control”, Quarry Australia, 24-34, June
1995.
[26] Liu, J. and Schönert, K., “Modelling of Interparticle Breakage”, Preprints of 8th
European Symposium on Comminution, Stockholm, Sweden, 102-115, May 10-12,
1994.
[27] Lynch, A. J., Mineral Crushing and Grinding Circuits, Elsevier Scientific Publishing
Company, 1977.
[28] Nilsson, A., Personal communication, Svedala, 1999.
[29] Prasher, C., L., “Crushing and Grinding Process Handbook”, John Wiley and Sons Ltd.,
1987.
[30] Rittinger, P. R. von, Lerbuch der Aufbereitungskunde, Ernst and Korn, Berlin, 1867.
[31] Rose, H. E., “Drop Weight Tests as the Basis for Calculations of the Performance of Ball
Mills”, Proc. 10th Int. Miner. Process. Congr., London, Publ. IMM (London), 143-147,
April 1973.
[32] Rosin, P. and Rammler, E., “The Laws Governing the Fineness of Powdered Coal”, J Inst
Fuel, Vol. 7, 29-36, 1933.
[33] Schönert, K., “Aspects of the Physics of Breakage Relevant to Comminution”, Proc 4th
Tewksbury Symposium, Melbourne, Australia, February 1979.
[34] Whiten, W. J., “The Simulation of Crushing Plants with Models Developed using
Multiple Spline Regression”, J South African Institute of Mining and Metallurgy, 257-
264, May 1972.
[35] Whiten, W. J., “Models and Control Techniques for Crushing Plants”, Control ’84, Minl.
Metall. Process, AIME Annual Meeting, Los Angeles, USA, 217-225, February 1984.
50
View publication stats

Cone Crusher Performance: Research

Uploaded by

Copyright:

Available Formats

Cone Crusher Performance: Research

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cone Crusher Performance: Research

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Cone Crusher Performance

Research · May 2015

AMIRA P9Q View project

Real-Time Optimization of Crushing Plants View project

The user has requested enhancement of the downloaded file.

CONE CRUSHER PERFORMANCE

CARL MAGNUS EVERTSSON

MACHINE AND VEHICLE DESIGN

Copyright © 2000 Carl Magnus Evertsson

Doktorsavhandlingar vid Chalmers tekniska högskola

Machine and Vehicle Design

Keywords: aggregates, comminution, crushing, mining, modelling,

DIVISION OF WORK BETWEEN AUTHORS

11 Simulation Results and Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Carl Magnus Evertsson

1.1 PRODUCTION OF BALLAST MATERIAL

Jaw crusher +64

Stockpile Cone crusher

Cone crusher Impact crusher

Figure 1. Layout of a four-stage crushing plant for ballast production.

1.2 CRUSHING OF MINERAL ORES

Stockpile 20-55 Screens

Figure 2. Layout of a size reduction circuit for a nickel rich ore.

1.3 BASIC COMMINUTION PRINCIPLES

1.3.1 COMPRESSIVE CRUSHING

1.3.2 IMPACT CRUSHING

1.4 OPERATING PRINCIPLE OF CONE CRUSHERS

Figure 4. Principle of a cone or gyratory crusher.

1.5.1 PROCESS DESIGN OF CRUSHING PLANTS

1.5.2 CRUSHER DESIGN

An X in the interdependency matrix symbolizes a dependency between two parameters. If the

1.5.3 CRUSHER OPERATION

Implementation Design considerations

Solution New or redesigned product

Figure 7. Structure of a problem oriented research process.

Cone Crusher Performance Top level

Wear Energy Size Distribution Capacity Sub level 1

Figure 8. Theory tree.

If a phenomenon cannot be explained by already existing theories, fundamental experiments

5.1 STRUCTURE OF CRUSHER ANALYSIS

Cone Crusher Performance Model

Crusher Flow Model

Figure 9. Structure of the crusher performance model.

Further information needed to carry out the calculations is:

5.2 MODULAR APPROACH

Input Calculation Output

Feed-size Filling ratio, η v

5.2.1 CRUSHER PERFORMANCE MAP

7 SIZE REDUCTION PROCESS MODEL

 --s- Material  --s- Material  --s- Material

Selection S corresponds to the probability of a particle being broken. Breakage B reflects

 --s-  --s- Material

inter inter single single

8 ROCK FRAGMENTATION CHARACTERISTICS

8.1 BREAKAGE BEHAVIOUR ASSUMPTIONS

8.2 SIZE DISTRIBUTION WIDTH

where σ and x are defined by

8.3 COMPRESSION CRUSHING TESTS