Академический Документы
Профессиональный Документы
Культура Документы
ABSTRACT
Ausgrind is Ausencos in-house program used for power-based comminution calculations. Proprietary
calculations have been developed and reconciled with unit process operation and benchmarked with
plant performance.
Ore breakage data from characterization and variability testwork, equipment geometry, flowsheet
configuration, efficiency factors and operating conditions are inputs. Ausgrind contains a database of
operating data and circuit types from many operations and over a wide range of ore types and
throughputs. These are referenced when developing flowsheets and circuit efficiency factors which
correct the specific energy required to grind from the nominated feed size to the specified product size. As
circuit efficiency factors are determined in Ausgrind, correction factors such as those listed by Bond and
Rowland are not employed.
The principal outputs are the total circuit specific energy and the specific energy for each of the
components in the comminution circuit. Ausgrind provides a valuable tool for evaluating comminution
options and optimizing circuit and equipment configurations.
Comminution is a pivotal and critical component in most mineral processing operations for several
reasons: it liberates the valuable mineral rendering it amenable to subsequent processes; it represents one
of the highest capital cost investments for unit processes in the treatment plant; and it contributes a
significant component to operating cost due to its high energy requirements. Therefore, determining the
most cost-effective circuit and predicting the energy requirement for each unit process to achieve the
desired size reduction in the product stream from the comminution circuit are critical tasks for circuit
design.
This paper describes the Ausgrind approach for estimating power and design of comminution circuits
and compares its outcomes to other published approaches. Practical examples are also presented, showing
applications for different circuit configurations for comminution operations, such as multi-stage crushing,
SAG and ball mill-based circuits.
INTRODUCTION
The progressive decrease in ore grades in the past decades has led the minerals industry to process ore at
higher rates to achieve the same metal production, driving significant increases in energy consumption.
Determining the most cost-effective circuit and predicting the energy requirement for comminution
circuits has become an even more critical task in concentrator design.
Von Rittinger (1867 apud Bond, 1961) and Kick (1885 apud Bond, 1961) proposed theoretical comminution
equations based on surface area and particle volume respectively, while Bond (1952) suggested an
empirical relationship that related the power requirement to the length of new cracks formed. Bonds
equation was developed for the purpose of designing conventional crushing and grinding circuits (crushrod-ball mill or crush-ball mill). By the 1970s, the application of conventional circuits tended to be limited
to relatively low capacities. Since then, circuits featuring autogenous (AG) and semi-autogenous (SAG)
mills, often combined with ball mills, have developed a prominent status in the minerals industry. Several
methodologies have been established for designing AG/SAG based grinding circuits and predicting their
energy requirements.
The most frequently used methodologies for calculating energy requirements of comminution circuits
including AG/SAG mills are briefly described in the following section. This paper focuses on the
description of the Ausgrind approach and also compares its outcomes to other published approaches.
Practical examples are presented, showing measured power draw and Ausgrind predictions for different
SAG and ball mill-based circuit configurations.
hardness and establishing comminution energy requirements. Nevertheless, all these methods rely on the
use of efficiency factors to correct the comminution circuit energy requirements. Ausgrind also uses the
Bond work indices, along with JKDWT, SMC, SPI or MacPherson index, as inputs for calculating the
comminution circuit Ecs.
The empirical equations for determining the efficiency factor (f SAG) in Ausgrind calculations were derived
from Ausencos database (Bueno and Lane, 2011). The f SAG relationships are ore-dependent and specific to
the selected circuit configuration. They may also be adjusted when secondary crushing or intensive
blasting practices are in place (i.e. for finer circuit feed sizes, F80). Figure 1 shows the Ausgrind fSAG
relationships for different circuit configurations and feed sizes. Likewise, an f SAG relationship was
calculated for Morrells method (2009) as the ratio between the total Ecs using SMC results and Bonds
formula; and is plotted in Figure 1 for comparison. In contrast to Ausgrind, the published methodology
used by Morrell (2009) assumed the same circuit efficiency (f SAG) for different circuit configurations, but
also varied with ore hardness (DWi).
High Competency
Moderate Competency
One or two stage crush to SAG or
ball mill
<0.5 Mt/a
0.5 to 2 Mt/a
1 to 5 Mt/a
5 to 10 Mt/a
>10 Mt/a
Low Competency
Single stage SAG
mill
Single stage SAG
mill or
SAG and ball mill
Single stage SAG
mill or
SAG and ball mill
SAG and ball mill
SAG and ball mill
This matrix is used as an indication for circuit selection. However, it must be carefully adapted when
problematic ore types present in the deposit can influence the performance of specific stages of the plant.
Sample Selection
Once the preliminary project concepts are defined, the next step is to establish the attributes of the
testwork program. The testwork program must include appropriate ore characterization tests that are
chosen to describe the orebody breakage properties. A detailed evaluation of geology through the
inspection of drill cores and geological sections is required when defining the comminution testwork
program. These inspections allow identification of relationships between the variability of ore
competency/hardness and the lithologies, rock types and mineralization.
The number and type of samples is dependent on the size and status of the project, project risk profile,
acceptable contingencies in defining equipment sizes, capital cost and operating cost. Typically, the
testwork program is divided into three phases. The first phase occurs during the scoping and
prefeasibility studies and samples are selected on the basis of major known rock types. The second and
third phases occur during the feasibility study. The second phase focusses on variability analysis across
the orebody where discrete samples are tested and compared with geometallurgical indicators.
Throughput and operating costs are forecasted more accurately with a relatively higher amount of
samples. For the third phase, samples are composited on the basis of spatial representativity, mine period
production, or hard wired ore typing. Table 2 shows the recommended number of samples for the last
phase of the testwork.
Table 2 Required number of samples during feasibility studies
Basis
spatial representativity
mine period production
hard wired ore typing
Number of samples
>100
5 to 8
10 to 30
Comments
includes samples from Phase 1 and 2
excludes samples from Phase 1 and 2
includes samples from Phase 1 and 2
product size (P80) is different to 150 microns, the energy difference is calculated as per the Bond
relationship. The Ausgrind total Ecs calculation is described in the equation below.
Total Ecs = [(Bond Ecs to 150m) * (fSAG F80_effect)] [Bond Ecs to final P80]
The base case SAG mill Ecs is calculated as function of ore competency (in this case as measured by the
DWi), as shown in Figure 4.
A series of factors are then applied to adjust the base case SAG Ecs for the operating conditions, mill
dimensions (e.g. aspect ratio) and others. Some of these adjusting factors are shown in Figure 5 and the
SAG Ecs calculation is described in the equation below.
SAG Ecs = [Base Case SAG Ecs] * [adjusting factors]
The pebble crusher Ecs is calculated separately and vendor data is used for equipment selection.
The ball mill specific energy is calculated as the residual using the following equation:
Ball mill Ecs = Total grinding circuit Ecs SAG mill Ecs
For circuits without a SAG mill, the ball mill specific energy equals to the total grinding circuit specific
energy. Similarly, for circuits without ball mills, the SAG mill specific energy equals to the total grinding
circuit specific energy. Mill shell and motor size selection considers the specific energy requirements and
the specified plant throughput, together with the expected operating envelope of each mill.
BENCHMARKING AUSGRIND
As several tests have been developed to suit the requirements of AG/SAG circuit design, the different
testwork methods and interpretation of their results can produce very different outcomes. Differences in
SAG mill Ecs determination and throughput predictions, particularly for competent ores (Bailey et al.,
2009), motivated Ausenco to benchmark the Ausgrind model. This section includes comparisons of the
Ausgrind outcomes with operational data of the SABC circuit from Newcrests Cadia concentrator and
also other published approaches.
Cadia
Ausgrind was used to calculate the energy requirements of the Cadia SABC circuit using information
available in the literature (Dunne et al., 1999, Dunne et al., 2001, Hart et al., 2001). The grinding circuit
data input used for the calculation are listed in Table 3. The information on ore characteristics in this table
refers to the monzonite ore. The SAG mill operating conditions were obtained from Dunne et al. (2001).
Table 3 Cadia grinding circuit data input for Ausgrind calculations
Ore
JK or SMC parameters (Ab)
SG
DWi
Bond crusher work index
Bond rod mill work index
Bond ball mill work index
Value
36
2.7
7.5
16.5
19.0
17.1
Unit
kWh/t
kWh/t
kWh/t
Crusher Circuit
Throughput - Primary crusher
Primary crusher circuit F80
Primary crusher circuit CSS
Crushing circuit P80
Value
4000
335
120
98
Unit
t/h
mm
mm
mm
Grinding Circuit
Throughput - Grinding
Cyclone overflow P80
Select circuit type
Value
2065
170
SABC
Unit
t/h
micron
t/m3
Value
1
12.2
0.5
6.1
0.72
12
25
70
Unit
Value
2
6.7
1.68
11.3
0.72
30
70
Unit
m
m
%
%
%
m
m
%
%
The specific energy measured during the survey (Dunne et al., 2001) and that calculated using Ausgrind
Total
17.6
16.6
(kWh/t)
- 1.0
Circuit
configuration
SMC worked
example
SABC
Kubaca circuit
SAB
Similco circuit
SACB
Confluencia
circuit
SACB
DBC worked
example - soft
ore
DBC worked
example medium ore
Sossego circuit
SSAG
SSAG
SABC
Method
Ausgrind
SMC
Ausgrind
MacPherson
Starkey
Ausgrind
FLSmidth
Ausgrind
MillPower
JKSimMet
Aminpro
Ausgrind
DBC
SMC
Ausgrind
DBC
SMC
Ausgrind
Delboni
Total
18.1
18.3
25.9
24.9
29.8
26.6
24.3
16.1
15.7
15.3
14.9
16.6
17.4
16.8
18.0
17.0
17.2
16.3
19.5
(kWh/t)
+ 0.2
- 1.0
+ 4.1
- 2.3
- 0.4
- 0.8
- 1.2
+ 0.8
+ 0.2
- 1.0
- 0.8
+ 3.2
The delta () in Table 5 is the difference in the total grinding specific energy between calculations with
Ausgrind and the other methods. According to Table 5, the lowest delta values were obtained when
comparing Ausgrind and the SMC method (+ 0.2 and 0.8 kWh/t). The highest delta values were obtained
when comparing Ausgrind and Starkeys method for the Kubaca circuit (+ 4.1 kW/t, a 15% difference) and
Delbonis method for the Sossego circuit (+ 3.2 kWh/t, a 20% difference). The specific energy predictions
for the other examples resulted in delta values lower than 2.3 kWh/t, meaning that the relative difference
was within a range from 2 to 9%.
CONCLUSIONS
A number of well-established methodologies used for calculating comminution energy requirements have
been presented in this paper. The published results from these methodologies were used to benchmark
the Ausgrind program. Ausgrind reproduced the specific energy requirements calculated by other
methods for various projects within 10% accuracy, with only a few exceptions, and predicted within 5%
the survey data from four different operations. Therefore, Ausgrind has demonstrated that its
methodology is a suitable and reliable method for designing AG/SAG milling circuits.
Every approach discussed in this paper relies on the use of efficiency factors (directly or indirectly) to
correct the Bond comminution circuit energy requirements for an AG/SAG circuit. Some methods are
more conservative than others, resulting in different efficiency factors and estimated energy requirements.
Notwithstanding the merits of comminution methodologies, the representativity of sample selection, the
applicability of the testwork and the quality of testwork data are critical for the success of any circuit
design. In light of this, the Ausgrind method includes a guideline for best practice in obtaining quality
input data.
ACKNOLEDGMENTS
The authors would like to thank Ausenco Services for the permission to publish this paper. Eddie McLean
is also acknowledged for his valuable comments and revision on the paper.
REFERENCES
Bailey, C., Lane, G., Morrell, S. and Staples, P. (2009) What can go wrong in comminution circuit design?,
In Proceedings of the 10th Mill Operators Conference, pp. 143149, Mill Ops 2009. AusIMM:
Melbourne.
Barratt, D.J. and Allan, M.J. (1986) Testing for autogenous and semiautogenous grinding: a designers
point of view, Minerals and Metallurgical Processing, pp. 65 74.
Barratt, D. J. (1989) An update on testing, scale-up and sizing equipment for autogenous and semiautogenous grinding circuits, In Mular, A. L. and Agar, G. E. (Eds.) Proceeding of Advances in
Autogenous and Semi-Autogenous Grinding Technology, pp. 2546, SAG 1989. UBC: Vancouver.
Bond, F.C. (1952) The third theory of comminution, Transaction of the AIME, 193, pp. 484 494.
Bond, F. C. (1961) Crushing & grinding calculations part I, British Chemical Engineering, 6:378385.
Bueno, M. and Lane, G. (2001) A Review of 10 Years of AG/SAG Pilot Trials, In Major, K., Flintoff, B. C.,
Klein, B. and McLeod, K. (Eds.), Proceedings of the International Autogenous Grinding Semiautogenous Grinding and High Pressure Grinding Roll Technology, SAG 2011. UBC: Vancouver.
Becerra, M. and Amelunxen, P. (2012) A comparative analysis of grinding circuit design methodologies,
In Proceedings of the 9th International Mineral Processing Conference, Procemin 2012, pp. 468476,
Santiago.
Burgess, D. (2012) A method of calculating autogenous/semi-autogenous grinding mill specific energies
using a combination of Bond work indices and Julius Kruttschnitt parameters, then applying
efficiency factors, In Proceedings of the 11th Mill Operators Conference, MillOps 2012. AusIMM:
Hobart, pp. 3744.
Delboni, H., Rosa, M. A. N., Bergerman, M. G. and Nardi, R. P. (2006) Optimisation of the Sossego SAG
mill, In Allan, M. J., Major, K., Flintoff, B. C., Klein, B. and Mular, A. L. (Eds.) Proceedings of the
International Conference on Autogenous and Semi-Autogenous Grinding Technology, pp. 3950, SAG
2006. UBC: Vancouver.
Starkey J. and Dobby, G. (1996) Application of the Minnovex SAG Power Index at five Canadian SAG
plants, In Mular, A. L., Barratt, D. J. and Knight, D. A. (Eds.) Proceedings of the International
Conference on Autogenous and Semi-Autogenous Grinding Technology, pp. 345360, SAG 1996. UBC:
Vancouver.
Dunne, R. Chittenden, R., Lane, G. and Morrell, S. (1999) The Cadia gold copper project exploration to
start up, In SME Annual Meeting. Preprint 90-160. SME: Denver.
Dunne, R., Morrell, S., Lane, G., Valery, W. and Hart, S. (2001), Design of the 40 foot diameter SAG mill
installed at the Cadia Gold Copper Mine, In Barratt, D. J., Allan, M. J. and Mular, A. L. (Eds.)
Proceedings of the International Conference on Autogenous and Semi-Autogenous Grinding Technology,
pp. 4358, SAG 2001. UBC: Vancouver.
Hart, S., Valery, W., Clements, B., Reed, M., Song, M. & Dunne, R. (2001) Optimisation of the Cadia Hill
SAG mill circuit, In Barratt, D. J., Allan, M. J. and Mular, A. L. (Eds.) Proceedings of the International
Conference on Autogenous and Semi-Autogenous Grinding Technology, pp. 1130, SAG 2001. UBC:
Vancouver.
Lane, G.S., Fleay, J., Reynolds, K., and La Brooy, S., (2002) Selection of comminution circuits for improved
efficiency, In Proceedings of the Crushing and Grinding Conference. Kalgoorlie.
MacPherson, A. R. (1978) A simple method to predict the autogenous grinding mill requirements for
processing ore from a new deposit, Transaction of the AIME, 262, pp. 236240.
Marks, A., Sams, C. and Major, K. (2001) Grinding circuit design for Similco mines, In Major, K., Flintoff,
B. C., Klein, B. and McLeod, K. (Eds.) Proceedings of the International Autogenous Grinding Semiautogenous Grinding and High Pressure Grinding Roll Technology, SAG 2011. UBC: Vancouver.
Morrell, S. (2004) Predicting the specific energy of autogenous and semi-autogenous mills from small
diameter drill core samples, Minerals Engineering, 17 (3): pp. 447451.
Morrell, S. (2009) Predicting the overall specific energy requirement of crushing, high pressure grinding
roll and tumbling mill circuits, Minerals Engineering, 22 (6): pp. 544549.
Napier-Munn, T. J., Morrell, S., Morrison, R. D. and Kojovic, T. (2005) Mineral comminution circuits: their
operation and optimisation, Julius Kruttschnitt Mineral Research Centre: Brisbane.
Rowland, C. A. (1972) Grinding calculations related to the application of large rod and ball mills,
Canadian Mining Journal, v. 93 (6).
Siddall, B.; Putland, B. (2007) Process Design And Implementation Techniques For Secondary Crushing
To Increase Milling Capacity, In SME Annual Meeting. Preprint 07-079. SME: Salt Lake City.
Starkey, J. and Holmes, G. (2001) Design of the Kubaca grinding circuit using SPI and Bond, In
Proceedings of the Canadian Mineral Processors Conference. CMP 2001. Ottawa.
Starkey, J., Hindstrom, S. and Nadasdy, G. (2006) SAGDesign testing what it is and why it works, In
Allan, M. J., Major, K., Flintoff, B. C., Klein, B. and Mular, A. L. (Eds.) Proceedings of the
International Conference on Autogenous and Semi-Autogenous Grinding Technology, pp. 240254, SAG
2006. UBC: Vancouver.