Minerals Engineering 17 (2004) 437–445

This article is also available online at:

www.elsevier.com/locate/mineng

A new autogenous and semi-autogenous mill model

for scale-up, design and optimisation
S. Morrell *

SMCC Pty Ltd., 26 Mermaid Place, Sinnamon Park, Qld., 4073 Australia
Received 25 August 2003; accepted 26 October 2003

Abstract
The successful application of modelling and simulation in design and optimisation rests heavily on having a model which is
lifelike. In the case of autogenous (AG) and semi-autogenous (SAG) mills this has provided a significant challenge to modellers, due
in part to the interaction between feed characteristics and mill performance and also due to the difficulty in obtaining high quality
relevant data from full-scale circuits. As a result the literature contains numerous papers on AG/SAG mill modelling, the majority of
which contain little or no data to indicate how well the models predict full-scale mill performance. In an attempt to address this
problem a new AG/SAG mill model has been developed which incorporates the capability to respond to changes in most design and
operational variables. The model can be used as a stand-alone design tool, to scale-up from pilot data and to optimise existing
circuits. It predicts throughput, power draw and product size distribution as well as load size distributions and slurry hold-up. The
overall structure of the model and the sub-processes that it contains is described. The model is successfully validated using data from
over 21 different full-scale circuits.

2003 Elsevier Ltd. All rights reserved.

Keywords: Autogenous grinding; Grinding; SAG milling

1. Introduction circuit and the ease with which combinations of circuit

designs can be evaluated on a throughput, power draw
Modelling and simulation of autogenous (AG) and and product size basis. Perhaps its most important
semi-autogenous (SAG) mills are valuable tools for advantage over the more traditional power-based tech-
helping to design the best circuit for a given application niques is its ability to interface with blast fragmentation
and subsequently to optimise its performance. This can models so as to assess and optimise the impact of blast
be achieved by narrowing the choice of the most likely design and ROM feed size distribution. Practical expe-
successful processing routes at the pre-feasibility stage, rience has shown that feed size variation can be signifi-
as well as augmenting a pilot test programme with cant and this can have a major impact on AG and SAG
simulations using breakage characterisation data from mill circuit performance.
drill core samples. In cases where pilot-scale testwork is The successful application of modelling and simula-
not possible, simulations using models which are based tion in design and optimisation rests heavily on having a
on data from a wide range of full-scale circuits can model which is lifelike. In the case of AG and SAG mills
significantly reduce the risk of not piloting. Even where this has provided a significant challenge to modellers,
pilot data are available simulations using data from a due in part to the interaction between feed characteris-
database, where a good match can be made in terms of tics and mill performance and also due to the difficulty
ore hardness and circuit design, provides an additional in obtaining high quality relevant data from full-scale
and very valuable crosscheck against the pilot results. circuits. As a result the literature contains numerous
Added advantages of the simulation route are its ver- papers on AG/SAG mill modelling, the majority of
satility, its ability to simulate the entire comminution which contain little or no data to indicate how well they
predict full-scale mill performance (Austin et al., 1976,
1985, 1987; Barahona, 1984; Nordell et al., 2001; Wey-
*
Tel.: +61-7-3715-7778; fax: +61-7-3376-2637. mont, 1979). Stanley’s work (1974) and that of Leung
E-mail address: [email protected] (S. Morrell). over a decade later (1987) stand out in that both
0892-6875/$ - see front matter
2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mineng.2003.10.013
438 S. Morrell / Minerals Engineering 17 (2004) 437–445

researchers went to great lengths to obtain pilot and full- p i ¼ d i si ð2Þ

scale mill data with which to indicate how well their
models predicted observed data. However, the range of where
mills they studied was limited and although their data
are very useful they cover a relatively small part of the fi tph of particles of size i in the feed
spectrum of designs currently seen in operating plants pi tph of particles of size i in the product
around the world. In an attempt to address this problem ri breakage rate of particles of size i
a new AG/SAG mill model has been developed which si mass of particles in the charge of size i
incorporates the capability to respond to changes in di discharge rate of particles of size i
most design and operational variables and which has aij breakage distribution function
been validated using a wide range of industrial data.
The simplicity of this model is both the source of its
greatest strengths and greatest weaknesses. Its greatest
2. Model development strengths are its ease of use and great versatility, whilst
its greatest weakness is the lack of any physical
2.1. Introduction description of the sub-processes on which it depends.
For it to be used successfully, therefore, a series of
In grinding mills the energy that the rotation of the supplementary models must be developed and linked to
mill shell imparts to the grinding media is subsequently it. These are described in the following sections.
transferred to ore particles with which they are in con-
tact, so causing breakage. The process is a discrete one, 2.3. Transport
as it comprises many individual collisions or breakage
‘‘events’’. The products of these breakage events may Transport of slurry out of the mill is described using a
then leave the mill via the discharge grates and pulp discharge function, which by rearranging equation (2)
lifters due to entrainment by slurry, or remain to undergo can be defined as
further breakage. The specific energy associated with pi
di ¼ ð3Þ
particle breakage is related to the energy of the grinding si
media and the mass of the target rocks, whilst the fre-
For a given mill the values of di can be calculated from
quency of these events is associated with the number of
measurements of the mass and size distribution of rocks
grinding media and the rate at which they circulate
in the mill (si ), and the throughput and the size distri-
within the mill. Hence for modelling purposes it is nec-
bution of rocks exiting the mill (pi ). However, measuring
essary to have a description of at least the following:
the size distribution and mass of rocks in a full-scale mill
is extremely difficult and hence much of the data that are
• frequency of breakage events for each size fraction;
available on the shape of the discharge function relate to
• energy associated with each breakage event;
the pilot-scale (Stanley, 1974; Leung, 1987). Very few
• size distribution of the products from each breakage
data have been reported on full-scale mills (Stanley,
event;
1974; Morrell, 1989) but these confirm the shape of the
• transport of slurry through the grates/pulp lifter.
function seen at the pilot-scale. By way of example, raw
data from a 7 m SAG mill are given in Fig. 1. The
The above sub-processes are all related to the size
function is characterised by two regions. The first ex-
reduction, throughput and power response of the mill
tends up to a particle size of about 1 mm (xm ), where the
and hence, by suitably configuring the model, all three
characteristics can be simultaneously predicted from the
one overall model.

2.2. Size reduction and throughput

The most useful framework for encapsulating the size

reduction process in a grinding mill is one based on the
so-called population balance model (Epstein, 1947), the
Whiten variant of this approach (1974) being the most
valuable for AG/SAG mill modelling. The model con-
tains two equations:
X
i
0 ¼ fi þ pi þ rj sj aij
ri sj ð1Þ
j¼1 Fig. 1. Discharge rate distribution from a 7 m diameter mill.
 S. Morrell / Minerals Engineering 17 (2004) 437–445 439

discharge rate is largely constant and equal to that of The value of c is a weighted radial position, which is
water. Therefore particles smaller than 1 mm plus their expressed as a fraction of the mill radius and is calcu-
associated water are assumed to behave as an homoge- lated using the formula:
neous fluid. For particles above about 1 mm the dis- P
r i ai
charge rate progressively reduces and reflects their c¼ P ð6Þ
rm ai
increasing tendency to be influenced by the classifying
action of the grate. As a result the function reaches zero where
at a particle size near to the grate/pebble port aperture,
which in this case was 18 mm. ai open area of all holes at a radial position ri
To model the discharge function advantage was taken rm radius of mill inside the liners
of the two distinct regions illustrated in Fig. 1. The re-
gion with a constant discharge rate reflects the purely The parameter values n1
n6 and g were found to be
fluid characteristics of the discharge stream. This con- functions of pulp lifter size and were all modelled using
stant rate (dm ) is directly related to the slurry flow rate the general equation proposed by Latchireddi (2002)
out of the mill and the quantity held up in the mill itself. ni ¼ ng
ki eð
kj kÞ ð7Þ
This is seen from Eq. (3), which for water and particles
smaller than xm can be re-written as where
P
pi ng parameter values for grate-only condition
dm ¼ Pm ð4Þ
m si ki , kj constants
k depth of the pulp lifter expressed as a fraction
The summations in the numerator and denominator in of mill diameter
Eq. (4) are the slurry discharge and hold-up, respec-
tively, where slurry is defined as water plus all particles The parameters in the above equations were fitted to
smaller than xm . Hence, to determine dm an equation over 760 laboratory and pilot mill data sets and 19 full-
relating slurry hold-up in a mill to the rate at which it scale mill data sets. The fit of the equations to the full-
flows out is required. Such a model was constructed scale mill data sets are shown in Fig. 2.
using the approach originally adopted by Morrell and When analysing the full-scale mill data the discharge
Stephenson (1996). In this approach two equations were slurry flowrate was calculated from the sum of the water
developed, one of which related to flow via the grinding plus particles smaller than xm . As a result Eq. (5) can be
charge interstices and the other to flow via the slurry directly related to Eq. (4), which can now be expressed
pool, in cases where this was present. Morrell and Ste- as
phenson incorporated the effects of grate design, mill
Q
speed and charge volume. However, the range of grate dm ¼ ð8Þ
designs they used for their experimental programme was Js
very limited and, more seriously, did not take into ac- Eq. (5) is therefore used initially to predict Js from a
count the effect of pulp lifters. Latchireddi (2002) recti- given slurry flowrate (Q), both Q and Js being subse-
fied this deficiency by conducting a detailed laboratory quently used in equation (8) to predict dm . For most
and pilot study of the influence of grate design and pulp
lifters. His results were then used to revise the original
equations and incorporate the influence of pulp lifter
depth (k). The result was the following general equation:

Js ¼ gcn1 An2 Jtn3 /n4 Qn5 Dn5 ð5Þ

where

Js the net fractional slurry hold-up inside the mill

A fractional open area
Jt fractional grinding media volume
/ fraction of critical speed
Q slurry discharge flowrate
c mean relative radial position of the grate holes
g coefficient of resistance which varied depending
on whether flow was via the grinding media
interstices or the slurry pool (where present) Fig. 2. Predicted slurry hold-up vs. observed in 19 full-scale AG and
n1
n6 model parameters SAG mills.
440 S. Morrell / Minerals Engineering 17 (2004) 437–445

practical purposes xm can be taken to be 1 mm and

hence for particle sizes up to this size di can be consid-
ered to be equal to dm .
Having estimated di for particle sizes up to 1 mm the
discharge function values for particle sizes greater than 1
mm then need to be predicted. This is done using the
Leung approach (1987) of using a log-linear relation-
ship, which goes to zero at the grate/pebble port size (see
Fig. 1).

2.4. Breakage distribution function

The breakage distribution function is the progeny size

Fig. 3. Measured breakage rate distribution from a 7 m SAG mill.
distribution that results from breakage of the parent
rock. This distribution will vary depending on the rock
strength and the energy with which it is broken. The general shape for a range of mill sizes and conditions
function therefore needs to be ore specific and related to (Stanley, 1974).
the breakage energy. The JKMRC’s approach is ideal It is convenient to represent distributions such as that
for this purpose. This uses a two-parameter description given in Fig. 3 using cubic splines (Ahlberg et al., 1967).
of rock breakage that is obtained from a drop-weight Hence five so-called spline ‘‘knots’’ were chosen at
test (Napier-Munn et al., 1996). The two parameters (A particle sizes of 0.25, 4, 16, 45 and 128 mm, giving rise to
and b) are ore specific and relate the t10 (a size distri- associated breakage rate values R1–R5.
bution index) to the applied specific energy (Ecs ). The t10 Although the general shape shown in Fig. 3 appears
in turn can be used to generate a size distribution (Na- to be maintained for a wide range of conditions it varies
rayanan and Whiten, 1988). The equation used for in a systematic manner with factors such as ball load
describing the relationship between the t10 and Ecs is: (Morrell and Morrison, 1996) and total charge volume
(Morrell et al., 2001). However, there is a major problem
t10 ¼ Að1
e
bEcs Þ ð9Þ
with determining what the exact variation is, which is
The specific comminution energy (Ecs ) has the units related to the difficulty in obtaining load size distribu-
kW h/t and is the energy applied during impact break- tions. From Eq. (1) it can be seen that without a load
age. As the impact energy is varied, so does the t10 . size distribution the breakage rate cannot be calculated
Higher impact energies produce higher values of t10 , directly. The data used to construct Fig. 3 were obtained
which are reflected in products with finer size distribu- from an audit of a 7 m SAG mill whose entire rock,
tions. slurry and ball load of 150 tonnes was removed, sepa-
Eq. (9) provides the means to predict the product size rated and sized (Morrell, 1989). This task was extremely
distribution but it requires the specific energy used for time consuming and costly and is not practicable for the
breakage before it can predict the t10 . Leung’s approach frequencies necessary to obtain the required mill design
is used to do so (Leung, 1987). Leung assumed that, and operational dependencies of the breakage rate dis-
although a wide range of energies are experienced by tribution. However, it is possible using data fitting
rocks in a mill, on average it should be linearly related to techniques to estimate the breakage rate distribution
the mill diameter. He therefore developed an algorithm without a load size distribution. The feed, product and
that estimated the mean grinding media size and calcu- breakage distribution function must still be known, but
lated its kinetic energy on impact when dropped from a in this case a non-linear fitting routine is used to
height equal to the mill diameter. The specific energy simultaneously estimate both the breakage rate and load
experienced by each particle size was then estimated size distribution. This can be done by choosing the load
using this mean kinetic energy. size and breakage rate distribution that best predicts the
observed product size distribution. The criterion of
2.5. Breakage rate minimising the sum of the squared deviations between
the observed and predicted sizings of the product stream
From equation (1) it can be seen that, providing the is normally used for this purpose. The problem with this
feed and product flowrates and size distributions, the approach is that there is often a range of solutions that
mill rock charge mass and size distribution, as well as give very similar results in terms of the goodness of fit.
the breakage distribution function are measured, the This is illustrated in Fig. 4, which shows two different
breakage rate distribution can be back-calculated. An breakage rate distribution that were fitted to a single a
example of the resultant distribution is shown in Fig. 3. set of data in which the load sizing was not available.
This distribution has been shown to follow the same The breakage rate distributions gave almost identical fits
 S. Morrell / Minerals Engineering 17 (2004) 437–445 441

data). The reason for attributing the differences to the

rate of rotation (rpm) is that the energy calculation
algorithm should have accounted for the influence of
mill diameter via its relationship to breakage energy. As
the breakage rate is related to breakage frequency (it has
the same units) it is considered to be more likely that the
differences are related to the frequency with which
the mill rotates. For the same fraction of critical speed
the rotational rate of a pilot mill is much higher than a
full-scale mill and hence it is believed that the higher
rates at coarser sizes exhibited in Fig. 5 are due to this.
The observed result at finer sizes is thought to be related
to the much higher numbers of grinding media layers
that are present in the bulk of the charge in full-scale
mills. It is between these layers that much of the
Fig. 4. Variation in the fitted breakage rate distribution from a full-
scale SAG mill. grinding of finer particles is thought to take place.
Given the effect of scale and the need to translate
trends from the pilot-scale to those at full-scale it was
to the product sizing. Together they define an envelope, decided to use the 11 data sets with their matching pilot
within which the true distribution is expected to fall. The and full-scale conditions to develop the breakage rate
magnitude of envelopes such as these makes the exper- relationships. This was done through a series of equa-
imental determination of trends in the breakage rate tions that related the rates R1–R5 to operational con-
distribution difficult to define with precision. ditions (including rpm). This follows a similar approach
Another solution to the problem is to use pilot mill to that originally developed by Mutambo (1993) and
data, as it is much easier to obtain load sizings. How- Morrell and Morrison (1996) in the JKMRC’s so-called
ever, the question then arises as to whether the trends ‘‘variable rates model’’. Different equations were ob-
seen in pilot mill breakage rates are repeated in full-scale tained to these earlier works, the general form being as
mills. To help answer this question, data from 11 dif- follows:
ferent pilot-scale tests were analysed and compared to
LnðRi Þ ¼ ki1 þ ki2 Jb Db þ ki3 x þ ki4 ðJt Þ ð10Þ
the results obtained from the full-scale mills that were
subsequently built to treat the same ores. The results are where
given in Fig. 5 and shown the average pilot-scale dis-
tributions for the 11 pilot tests compared to the average Ri breakage rate values for rates i ¼ 1–5
from the 11 full-scale mills. A distinct difference between Jb % by volume of balls
the two distributions is seen. This is principally attrib- Jt % by volume of grinding media (ball plus rocks)
uted to the influence of the rate of rotation (not diam- Db make-up ball size
eter, which might be the first conclusion drawn from the x mill rotational rate
ki1
i4 constants for rates i ¼ 1–5

The resultant equations gave a set of responses to

changes in ball load, ball size, total load and speed
which are illustrated in Figs. 6–9 as well as matching the
differences between pilot and full-scale shown in Fig. 5.

2.6. Power draw

The size reduction part of the model predicts what the

mass of rock and the quantity of slurry will be in the mill
for a given feed rate. The power model uses this infor-
mation to predict what the volume, density and position
of the charge will be. The position of the grinding media
portion of the charge is described using the conceptual
view shown in Fig. 10 (Morrell, 1996a,b).
The slurry portion of the charge occupies the grinding
Fig. 5. Comparison between pilot and full-scale AG/SAG mill media interstices. However, it is not uniformly dispersed
breakage rate distributions. but fills the interstices from the shoulder down towards
442 S. Morrell / Minerals Engineering 17 (2004) 437–445

Fig. 6. Predicted effect on the breakage rate distributions as ball load Fig. 9. Predicted effect on the breakage rate distributions as total load
is changed. is changed.

Fig. 7. Predicted effect on the breakage rate distributions as speed is

changed.

Fig. 10. Schematic of simplified charge shape.

effectively extends beyond the grinding media toe. This

situation occurs in all overflow discharge ball mills and
gives rise to the lower power draw which is usually seen
in overflow ball mills compared to grate discharge units
with the same size, ball charge and speed. It also causes
the drop in power seen in AG/SAG mills that are fed at
a rate that is too high for the grate/pulp lifters to handle
(Morrell and Kojovic, 1996).
Using an energy balance approach, where power is
Fig. 8. Predicted effect on the breakage rate distributions as ball size is taken to be the rate at which potential and kinetic en-
changed. ergy is imparted to the charge, Eq. (11) was developed
for the cylindrical section of the mill. Some mills have
conical end-sections and, as they also draw power, have
the toe of the grinding media charge (Latchireddi and
to be included if accurate predictions are to be obtained.
Morrell, 2003). This it does so in response to changes in
Eq. (12) was therefore developed and describes the
the slurry flowrate. The toe of the slurry portion of the
power draw of a single conical end-section
charge is therefore not necessarily coincident with that Z rm
of the grinding media. This is illustrated schematically in Pcylinder ¼ Vr Lrgðqc ðsin hs
sin ht ÞÞ
Fig. 11, which shows the movement of the slurry toe as ri
flowrate increases. If the flowrate is high enough a slurry
pool forms (Fig. 11d) in which case the slurry toe þ qp ðsin hs
sin htp Þ dr ð11Þ
 S. Morrell / Minerals Engineering 17 (2004) 437–445 443

Fig. 11. Schematic of observed trends in slurry hold-up.

Z Li Z rc ht angular displacement of the grinding media toe

Pcone ¼ Vr rgðqc ðsin hs
sin ht ÞÞ position at the mill shell
0 ri
htp angular displacement of the slurry toe position
þ qp ðsin hs
sin htp Þdr dLc ð12Þ at the mill shell (see Fig. 11b)
qc density of rock/ball charge (excluding pulp)
where qp density of pulp phase

Lc length of cone-end, measured from the cylin- The variation in the angles ðhs ; ht Þ, as well as the
drical section, at a radius of rc velocity variation with radial position ðVr Þ, were mod-
Li length of charge surface within cone ends elled empirically from observations of the charge in a
L length of cylindrical section of the mill inside glass laboratory mill (Morrell, 1993).
liners Eqs. (11) and (12) describe the net power draw and do
P power delivered to the charge (net power) not include electrical and mechanical losses. To predict
r radial position the total (gross) power draw these losses must be added
ri radial position of charge inner surface to the net power. A further equation is therefore re-
rm radius of mill inside liners quired which relates these losses to the mill’ s dimen-
rc radius of cone-end at a distance Lc from the sions and speed. The no-load power equation proposed
cylindrical section (m) by Morrell (1996a) was used for this purpose.
Vr tangential velocity of a particle at radial dis- The resultant model should be applicable to all AG
tance r and SAG mills as well as ball mills. The accuracy of the
hs angular displacement of shoulder position at model is demonstrated in Fig. 12 using 140 ball, AG and
the mill shell SAG mill data sets.
444 S. Morrell / Minerals Engineering 17 (2004) 437–445

The simulations were run by first inputting the design

and operational conditions as well as the JK rock
breakage parameters, A and b. The breakage rate dis-
tribution was then predicted. Then, using the observed
feed size distribution the throughput was increased until
the observed total charge level was reached. The resul-
tant power draw, throughput and product size distri-
bution predictions of each circuit were then compared to
those measured in the field. The results are given in Figs.
13–16. The throughput and power draw were found to

Fig. 12. Observed vs. predicted power draw of AG, SAG and ball
mills.

3. Model validation

To test the validity of the model data from 21 dif-

ferent full-scale AG/SAG mill circuits were used. The
data comprised mill dimensions, grate design, mill
Fig. 14. Observed vs. predicted power draw.
speed, throughput, feed size distribution, product size
distribution, power draw, ball load, ball size, normal
operating load level and rock breakage characteristics
(JK A and b values). The range in the major variables
covered by the data sets is given below.

ball load: 0–17% by volume

ball size: 100–150 mm
total load: 20–40% by volume
speed: 68–80% of critical
diameter: 1.7–12 m
length: 0.5–12 m
aspect ratio ðL=DÞ: 0.33–1.85

Fig. 15. Observed vs. predicted product size ()150 l).

Fig. 13. Observed vs. predicted throughput. Fig. 16. Observed vs. predicted product size (P80).
 S. Morrell / Minerals Engineering 17 (2004) 437–445 445

be predicted with a reasonable high degree of accuracy. Latchireddi, S.R., 2002. Modelling the performance of grates and pulp
The predicted product sizes followed the observed val- lifters in autogenous and semi-autogenous mills. Ph.D. Thesis,
University of Queensland, Australia.
ues well though more scatter in the data were apparent Latchireddi, S.R., Morrell, S., 2003. Slurry flow in mills: grate-only
in comparison to the predicted throughput and power discharge mechanism (Part 1). Miner. Eng. 16 (7).
draw results. Morrell, S., 1989. Simulation of bauxite grinding in a semi-autogenous
mill and DSM screen circuit. M. Eng. Thesis, University of
Queensland, Australia.
Morrell, S., 1993. The prediction of power draw in wet tumbling mills.
4. Conclusions Ph.D. Thesis, University of Queensland, Australia.
Morrell, S., 1996a. Power draw of wet tumbling mills and its
A new AG/SAG mill model has been developed relationship to charge dynamics––Part 1: a continuum approach
to mathematical modelling of mill power draw. Trans. Inst. Miner.
which incorporates some of the tried and proven ele- Metall Sect. C 105, C43–C53.
ments of a number of previous models as well intro- Morrell, S., 1996b. Power draw of wet tumbling mills and its
ducing a number of recent developments, particularly in relationship to charge dynamics––Part 2: an empirical approach
the field of slurry transport. The model was refined using to modelling of mill power draw. Trans. Inst. Miner. Metall. Sect.
a suite of full-scale mill data that had complementary C 105, C54–C62.
Morrell, S., Kojovic, T., 1996. The influence of slurry transport on the
pilot-scale data. Validation of the resultant model was power draw of autogenous and semi-autogenous mills. In:
carried out using results from 21 full-scale mill circuits. Proceedings of 2nd International Conference on Autogenous and
It was shown that the model provided good predictive Semi-autogenous Grinding Technology, Vancouver, Canada, pp.
capability in terms of throughput and power draw as 378–389.
well as being able to follow product size trends well. The Morrell, S., Morrison, R.D., 1996. AG and SAG mill circuit selection
and design by simulation. In: Proceedings of SAG’96, Vancouver,
model should find application as a stand-alone design
vol. 2, pp. 769–790.
tool, to scale-up from pilot data and to optimise existing Morrell, S., Stephenson, I., 1996. Slurry discharge capacity of
circuits. autogenous and semi-autogenous mills and the effect of grate
design. Int. J. Miner. Process. 46 (1-2), 53.
Morrell, S., Valery, W., Banini, G., Latchireddi, S., 2001. Develop-
ments in AG/SAG mill modelling. In: Proceedings of International
References Conference on Autogenous and Semi-autogenous Grinding Tech-
nology, vol. IV, pp. 1–84.
Ahlberg, J.H., Nilson, E.N., Walsh, J.L., 1967. The Theory of Spline Mutambo, J., 1993. Further Development of an Autogenous and
Functions and Their Applications. Academic Press, New York. Semi-autogenous Mill Model. M.Eng. Sci. Thesis, University of
Austin, L.G., Weymont, N.P., Prisbrey, K.A., Hoover, M., 1976. Queensland, Australia.
Preliminary results on the modelling of autogenous grinding. In: Napier-Munn, T.J., Morrell, S., Morrison, R.D., Kojovic, T., 1996.
14th International APCOM Conference, The Pennsylvania State Mineral comminution circuits their operation and optimisation.
University, pp. 207–226. JKMRC Monograph Series.
Austin, L.G., Weymont, N.P., Barahona, C.A., Suryanarayanan, K., Narayanan, S.S., Whiten, W.J., 1988. Determination of Comminution
1985. An improved simulation model for semi-autogenous grind- characteristics from single particle breakage tests and its applica-
ing. SME-AIME 85, 378. tion to ball mill scale-up. Trans. Inst. Miner. Metall. 97, C115–
Austin, L.G., Menacho, J.M., Pearcy, F., 1987. A general model for C124.
semi-autogenous and autogenous milling. In: Proceedings of the Nordell, L.K., Potapov, V., Herbst, J.A., 2001. Comminution simu-
20th International Symposium of Applications of Computers and lation using discrete element method (DEM) approach––from
Maths in Mineral Industries. Metallurgy, vol. 2. SAIMM, pp. 107– single particle breakage to full-scale SAG mill operation. In:
126. Proceedings of SAG 2001, Vancouver, vol. IV, pp. 235–251.
Barahona, C.A., 1984. Modelling and simulation of semi-autogenous Stanley, D.D., 1974. The autogenous mill—a mathematical model
grinding systems. M.Sc. Thesis, The Pennsylvania State University. derived from pilot and industrial scale experiment. Ph.D. Thesis
Epstein, B., 1947. The material description of certain breakage (Unpublished), University of Queensland, Australia.
mechanisms leading to the logarithmic-normal distribution. J. Weymont, N.P., 1979. The analysis and simulation of autogenous
Franklin Inst., 244–471. grinding systems. Ph.D. Thesis, The Pennsylvania State University.
Leung, K., 1987. An energy based ore specific model for autogenous Whiten, W.J., 1974. A matrix theory of comminution machines. Chem.
grinding mills. Ph.D. Thesis. University of Queensland, Australia. Eng. Sci. 29, 588–599.