See discussions, stats, and author profiles for this publication at: [Link]

net/publication/267695053

Optimization of the Design of AG Mill Shell Liners at the Gol-E-Gohar

Concentration Plant

Conference Paper · October 2014

CITATIONS READS

0 720

3 authors:

Mohsen Yahyaei Mostafa Maleki

The University of Queensland Vali-e-Asr University Of Rafsanjan
54 PUBLICATIONS   340 CITATIONS    13 PUBLICATIONS   54 CITATIONS   

SEE PROFILE SEE PROFILE

Samad Banisi
Shahid Bahonar University of Kerman
68 PUBLICATIONS   621 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

JKMRC Dynamic Simulation Platform View project

Model-informed Process Prediction and Control (MiP&C) View project

All content following this page was uploaded by Mohsen Yahyaei on 15 July 2015.

The user has requested enhancement of the downloaded file.

Optimization of the design of AG Mill shell liners at the
Gol-E-Gohar concentration plant
M. Maleki-Moghaddam1, M. Yahyaei 2, A. Haji-Zadeh 3, H. Asadi 4 and S. Banisi 5
1- University of Rafsanjan, Iran
2- Kashigar Mineral Processing Research Center, Shahid Bahonar University of Kerman, Iran
3- Gol-E-Gohar Mining and Industrial Company, Iran
4- Gohar-Ravesh Sirjan Company, Iran
5- Mineral Processing Group, Shahid Bahonar University of Kerman, Kerman, Iran

ABSTRACT

One area in AG/SAG milling field which has received a great attention is the optimization of liner design
for any application to increase the availability of mills. The Gol-E-Gohar iron ore concentration plant uses
three 9m × 2.05m autogenous mills (AG) in parallel in a dry operation. The AG mill shell is lined with
three series (57 cm each) of 36-row liners. Due to large variations in feed characteristics and inadequate
blending, the performance of AG mills has been lower than the target value. Therefore, to improve the
comminution efficiency, the optimization of the design of AG mill shell liners was applied in this plant.
By simulation it was found that increasing the liner lifter face angle from 7 to 30° while maintaining the
original lifter height (i.e., 22.5 cm) could provide an appropriate charge trajectory. Measurement of the
wear profiles of shell liners indicated that the wear along the liner length was not uniform and the highest
wear occurred between 0.5 m and 1.2 m of the mill length. A new liner design was proposed by
redistributing the metal from parts with a lower wear rate to the parts with a higher wear rate to arrive at
a uniform wear profile along the liner length at the liner removal time. The non-uniform shell liner was
constructed and installed in one of the three parallel dry AG mills. It was found that the first and second
liner designs increased the mill throughput by 2.6 and 7.5%, respectively. After conversion to SAG
milling on the account of liner change, 18% (from 413 to 489 t/h) increase in throughput was realized. It
was predicted that the non-uniform liner designs will increase the liner life time by 10% and will reduce
the scrap metal from the current value of 59% to 50%.

1
INTRODUCTION
The use of large AG/SAG mills in the mineral industry has been on rise due to a significant reduction in
capital and operating costs and increase in the plants throughputs. A notable advantage of SAG over AG
mills is their flexibility in encountering feed hardness variations. The availability of these mills plays an
important role in the economics of the operation. One of the main reasons of mills down time is the time
required to change worn or broken liners (Maleki, Yahyaei and Banisi, 2012). Worn liners not only
decrease the mill operating time but they also affect the charge motion, which has a significant effect on
the grinding efficiency. It is then reasonable to assume that modelling of liner wear profile could assist
optimization of the liner design for longer life and proper charge behaviour (Kalala, Breetzke and Moys,
2008; Yahyaei, Banisi and Hadizadeh, 2009).
There are different criteria to evaluate the performance of liners. Liner life time (in working hours or
product weight), and liner consumption (in working hours or weight of ground material) are the two
most applied criteria. The liner performance not only should be considered in terms of cost of liner
materials and change of worn liners, but it should also be studied from mill capacity and performance
aspects (Parks, 1989; Yahyaei, Banisi and Hadizadeh, 2009).
Aims of new liner designs are to improve the effective grinding through changes in shell lifter face angles
and the spacing of the lifters (Powell et al., 2006; Rajamani, 2006; Royston, 2006). Sequential
developments of liner design have been performed with the aim of decreasing mill down time, resulting
from the replacement of worn liners. Liner material and design play key roles on the liner life time but
experiences have shown that the latter is more critical. To arrive at an optimized liner design, it is
customary to install one or two series of conventional liners and study their wear profile over time. The
initial design could then be modified via redistribution of metal from the parts with lower wearing rate to
higher wearing rate (Parks, 1989; Yahyaei, Banisi and Hadizadeh, 2009).
It has been known for many years that a change in the face angle on SAG mill shell lifters results in a
change in the trajectory of the thrown grinding balls. With the development of trajectory-generating
computer programs, the effects of face angle, packing and lifter height have been incorporated in the shell
lifter design. Increasing shell lifter face angles (for the same mill speed) does reduce the impact point of
thrown balls and can reduce shell liner damage (Powell et al., 2006; Rajamani, 2006; Yahyaei and Banisi,
2010). Charge trajectory modelling has been used to present and interpret aggregate motion of a number
of balls critical to the performance of the mill (Royston, 2003).
The availability of large-scale cost-effective computing power in recent years has allowed the
development of Discrete Element Modelling (DEM) computer programs which model charge trajectory in
mills. DEM helps to understand charge motion in SAG mills for given liner designs and lifters, ball and
rock properties and mill operating conditions. Based on the success of the 2-D model of the SAG mill the
3-D model in DEM was developed. These type of models could help to study charge motion and energy
draft more accurately (Mishra and Rajamani, 1993; Cleary, 2001; Kalala et al., 2008; Powell et al., 2011).
The mill trajectory is the main concern in liner designing issues because the direct impact of balls to the
liner shell breaks them gradually. The charge trajectory can be optimized by the modification of liner
design.

2
Yahyaei and Banisi (2010) designed a spreadsheet-based software (GMT) to model charge trajectory in
tumbling mills. Charge shape and impact points predicted by the GMT (Grinding Media Trajectory)
program was based only on a single ball trajectory. Maleki, Yahyaei and Banisi (2013) proposed new
relationships to modify the GMT results to take into consideration the effect of charge. By applying the
corrections to the charge shape and impact points, the GMT software outputs became more realistic.

The Gol-E-Gohar iron ore concentration plant

This study was carried out at the Gol-E-Gohar mining and industrial company located southeast of Iran.
Three 9 m × 2.05 m autogenous mills (AG) are used in parallel in a dry operation to grind a feed all under
32 cm which is the product of a gyratory crusher. The SAG mills use 4023 hp motors and work with
constant 12 rpm rotational speed (i.e., 85% of critical speed) in one direction. The mill shell is lined with
three series (57 cm each) of 36-row liners. Specifications of liners used are shown in Table 1. According to
the liner manufacturer recommendation, the liners should be changed when the lifter height reaches 100
mm which is one third of its initial height. Due to large variations in feed characteristics and inadequate
blending, the performance of AG mills has been lower than the target value.
In this research, liners with new design were installed and AG mills were converted to SAG mills to
improve the grinding efficiency at the Gol-E-Gohar concentration plant. Based on wear profile
measurements, a modified liner was proposed to arrive at a uniform wear profile at the end of liners life
time. The liner profile was measured over the period of liner life time and the effect of the modifications
on the comminution efficiency was determined.

Table 1 Specifications of liners at the Gol-E-Gohar concentration plant used in GMT software

Plate thickness Lifter Lifters in Number of

(mm) Face angle (degrees) Height (mm) Width (mm) row rows
75 7-30 225 250 3 36

METHODOLOGY

Charge trajectory predication software

In order to determine the charge trajectory, a new version of GMT (Yahyaei and Banisi, 2010) was used. In
the new GMT, Maleki, Yahyaei and Banisi (2013) corrected the charge shape and single ball trajectory to
take into account the effect of charge by applying new relationships. In comparison with other software
packages GMT has the advantage of using all MS Excel © readily available functions and capabilities. The
main feature of GMT is the ability to show the crescent-like shape of the charge along with the trajectory,
which has not been incorporated in similar software packages. Its data entry section includes five parts to
be completed by the user: general information, mill, grinding media and liner data, and parameters.

3
The model mill
In this research a model mill with the diameter of 100 cm and a special design made possible to increase
the mill length in steps of 3.6 cm up to 21.6 cm (Figure 1-a). In other words, the mill length could be
increased by 6-fold. The transparent end of the mill made accurate trajectory determination possible by
taking films and photographs. A 2.5 kW motor with a variable speed drive was used which provided a
sufficient flexibility to test various operating conditions. The scale down ratio of 9 to 1 was used to
construct the model mill using the plant mill dimensions. In order to exactly duplicate the plant liners
arrangement in the mill, a polyurethane ring was accurately machined to create 36 lifters in a row (Figure
1-b). By this method any number of liners in a row with different designs could be tested. Between the
lifters 36 plates were also included consistent with the plant liner design.

(a) (b)

Figure 1 a) The model mill, b)The polyurethane ring machined to scale down the liners arrangement of the plant mill

3-D linear wear profile measurement device

The usual wear rate determination method is to assume uniform wear profile for a liner and measure the
profile at one cross section. However it has been shown that the liner wear profile is not uniform (Banisi
and Hadizadeh, 2006). Thus, a new method was used which measures the wear profile by building a 3-D
model of liners after any given operating time. For this, a device made of aluminium with the dimensions
of 30×50×500 mm was constructed to measure the cross section accurately. The main part of the device
contained equally spaced holes (Figure 2).

Figure 2 The wear profile measurement device

4
To build the 3-D liner model, a software called Solidsworks© was used (Figure 3). At this stage, the
weight and volume of the liner could be easily determined. To determine the wear rate, measurements
were made at four equally distanced cross sections. Scheduled shut-down periods of the SAG mill at the
plant made regular measurements and monitoring of the liners possible.

Figure 3 Typical 3-D model of a worn liner after 3000 h of operation at the Gol-E-Gohar concentration plant

RESULTS AND DISCUSSIONS

Charge trajectory prediction

The charge trajectory, when using 125 mm balls and assuming 18% total filling, was predicted by the
GMT software (Figure 4). The ball impact points clearly indicated that, with the original liner design, ball
addition could result in direct impacts on the liners, causing severe damage. The low throughput of the
mill could also be attributed to the inappropriate trajectory, where the toe of the charge does not receive
any appreciable direct impacts from the falling load. The effect of different fillings on the position of the
ball impacts was also studied. Impact points were measured in degrees, starting from the horizontal line
passing the mill centre (i.e., 3 o’clock position) and moving counter-clockwise. The distance between
impact points and the toe was measured in degrees. Even at 24% filling which was considered high for
the operation, there was 32º difference between the ball impact points and the toe position.

Impact point

Predicated charge trajectory

Toe Predicted mill load

Figure 4 Simulation of 125mm ball (18% filling) trajectory for the old liner by the GMT software

5
New liner design
The new liner not only should provide an appropriate load trajectory, but it should also be as similar as
possible to the original liner design, to facilitate its removal and installation. By simulation of the charge
trajectory at various liner shapes, it was found that the lifter face angle was the most effective parameter
influencing the charge trajectory. As a result, for the new design the lifter face angle increased, while lifter
height was kept constant. To obtain a lifter face angle providing an appropriate charge trajectory, the face
angle was increased from 7 to 30° and the ball trajectories were obtained by the GMT software (Figure 5).
The results indicated that when the face angle increased from 7 to 30°, the distance between the impact
point and the toe decreased from 32 to 8° for 24% filling and decreased from 38 to 12° for 18% filling.

Lifter face angle (degrees)

Mill load
15
20
25
30

Impact point

Toe

Figure 5 Simulation of 125 mm balls trajectories with different lifter face angles (24% filling)

Since ball impact points were within the safe operating windows (i.e., -6.4 to 12˚ (Maleki, Yahyaei and
Banisi (2013)), the lifter face angle of 30° was selected for the new design. To ensure the appropriate load
trajectory, the laboratory-scaled models of the original and new liners were constructed and tested in the
1m-diameter laboratory mill. Steel balls within the size range of 4-12 mm were added to provide the
desired fillings. For both the original and new liners, the tests were performed at 14, 16, 18, 20 and 24%
filling. Since the rotation speed of Gol-E-Gohar mills was fixed at 85% of critical speed, this rotation speed
was selected for all tests. Figure 6 illustrates the charge shape and trajectory for the original and new
liners at 20% filling.

Figure 6 Charge shape and trajectory of the original and new liners in the laboratory mill (20% filling)

6
Given the promising results obtained in the laboratory mill, the new liners (Design No. 1) with 30° lifter
face angle and unchanged lifter height (i.e., 22.5 cm) were constructed and installed in AG mill No. 1.
Figure 7 shows the original and new liners. The weight of the liners was kept equal to the original design
to avoid increasing the total weight of the mill.

angle
Face
Old liner design New liner design (Design No.1)

Figure 7 The original and new liners used at the Gol-E-Gohar concentration plant mills

Modeling liner wear profile

The profile of a liner after 6200 h of operation at four equal distances along the liner length is shown in
Figure 8. Unlike the first profile which does not experience much wear, profile 4 which is very close to the
mill centre receives the highest wear. The relative height difference between the second profile and the
fourth profile is 44% which is indicative of a non-uniform wear profile along the liner length. The liners
wear increases from the feed end to the mill centre indicated by the decreasing trend of liner height.

Figure 8 The wear profiles of a liner after 6200 h of operation at four points along the length of the liner

During the liner life time (6200 h), five series of measurements were performed at 2300, 3815, 4350, 5150
and 6200 h of operation, respectively. As a result, the height of lifters at specific locations (i.e., profiles 1–4)
and the corresponding operation time were determined.

Proposed liner profile

Based on the wear profile measurements, the highest wear rate was found to be between 0.5 m and 1.2 m
along the mill length (Figure 9). Therefore, to arrive at a uniform wear profile at the end of liners life time,
a modified liner was proposed. The main idea was to redistribute the metal from the low wear rate
sections to those with a high wear rate. Unlike the old design the proposed design had a curved shape.

7
The lifter height changes along the liner compared with the old version from profile 1 to 6 was -72, -48, -
15, 21, 45, and 63 mm, respectively. This meant 72 mm reduction of lifter height in profile 1 (i.e., the first
profile from the feed end side) and 63 mm increase in the lifter height for profile 6 (i.e., the second profile
at the middle liner) (Figure 9).

Figure 9 Variations of the maximum lifter heights along the liner after 6200 h of operation

It was thought that the curved shape design may cause difficulties in liner production (casting).
Therefore, final design (Design No. 2) was proposed where the first and third liners have sloped surfaces
and the middle one has flat surface (Figure 10). It is worth noting that the weight of the proposed designs
was kept equal to the original design to avoid any problem resulting from higher mill weight.

Figure 10 The 3-D model of the proposed liner (Design No. 2)

Since the dynamic behaviour of the charge is influenced by the lifter height, it was decided to evaluate the
effect of the design change on the charge behaviour. The idea was to check whether the design change
could cause direct ball impact on the shell which in turn could have detrimental effects on the operation.
The study showed that, in the range of lifter height change (150–260 mm), the impact angle of the charge
did not change significantly (i.e., in the order of 1–3°). This provides enough assurance that the new liner
design will not have any adverse effect on the mill as result of direct ball impact on the mill shell.

Investigating the effect of modified liners on mill performance

Given the promising results obtained in the model mill, the new liners (Design No. 1) with 30° lifter face
angle and unchanged lifter height (i.e., 22.5 cm) were constructed and installed in dry AG mill No. 1.
Based on the wear profile measurements, the highest wear rate was found to be between 0.5 m and 1.2 m
along the mill length. A new liner design (Design No. 2) was proposed by redistributing the metal from

8
parts with a lower wear rate to the parts with a higher wear rate to arrive at a uniform wear profile along
the liner length at the liner removal time. The non-uniform shell liners were constructed and installed in
AG mill No.1. Throughputs of the mill before and after changes are summarized in Table 2.

Table 2 Throughputs and product sizes of the mill before and after liner change and after balls addition

Averaging period (months) Throughput (t/h) P80 (µm)

Before change 18 413±56 502±25
Liner design No. 1 13 424±47 501±28
Liner design No. 2 2.5 444±41 546±26
Balls addition 6 489±25 529±26

The first and second liner designs increased the mill throughput by 2.6 and 7.5%, respectively. After
conversion to SAG milling on the account of liner change, 18% (from 413 to 489 t/h) increase in
throughput was obtained. It was predicted that the new liner design will increase the liner life time by
10% and will reduce the scrap metal from the current value of 59% to 50%. The measurement of profiles
of non-uniform liners during 6220 h of operation indicated that this prediction could be achieved. It is
important to note that during the liner design change period the hardness of feed increased which was
observed and verified in other parallel dry mill. In other words, the reported increase is conservative and
the actual increase is higher.

CONCLUSIONS
- To improve the comminution efficiency at the Gol-E-Gohar iron ore concentration plant, a change in the
liner design and conversion from AG to SAG milling were practised.
- It was found by simulation that increasing the liner lifter face angle from 7 to 30° while maintaining the
original lifter height (i.e., 22.5 cm), the distance between the impact point and the toe decreased from 32 to
8° for 24% filling and decreased from 38 to 12° for 18% filling. Given the promising results obtained in the
model mill, new liners (Design No. 1) with 30° lifter face angle and unchanged lifter height (i.e., 22.5 cm)
were constructed and installed in AG mill No. 1.
- During the liner life time (6200 h), five series of measurements were performed at 2300, 3815, 4350, 5150
and 6200 h of operation, respectively. Based on wear profile measurements, the highest wear rate was
found to be between 0.5 m and 1.2 m along the mill length. A new liner design (Design No. 2) was
proposed by redistributing the metal from parts with a lower wear rate to the parts with a higher wear
rate to arrive at a uniform wear profile along the liner length at the liner removal time. The non-uniform
shell liners were constructed and installed in the AG mill.
- The first and second liner designs increased the mill throughput by 2.6 and 7.5%, respectively. After
conversion to SAG milling a 18% (from 413 to 489 t/h) increase in throughput was obtained as result of the
liner design change. It was predicted that the new liner design increases the liner life time by 10% and
reduces the scrap metal from the current value of 59% to 50%.

9
 ACKNOWLEDGMENTS
The authors would like to thank Gol-E-Gohar mining & industrial company and Gohar-Raveh Sirjan
company for supporting and implementing the outcome of this research and also the permission to
publish the article. Special appreciation is also extended to the operating, maintenance, metallurgy and
R&D personnel for their continued help.

REFERENCES
Cleary, P.W. (2001) 'Charge behavior and power consumption in ball mills: sensitivity to mill operating conditions,
liner geometry and charge composition' International Journal of Mineral Processing, vol. 63, pp. 79-114.

Kalala, T.J., Breetzke, M. & Moys, M.H. (2008) 'Study of the influence of liner wear on the load behaviour of an
industrial dry tumbling mill using the discrete element method (DEM)', International Journal of Mineral
Processing, vol. 86 (1–4), pp. 33-39.

Maleki-Moghaddam, M., Yahyaei, M. & Banisi, S. (2012) 'Converting AG to SAG mills: The Gol-E-Gohar Iron Ore
Company case' Powder Technology vol. 217, pp. 100–106.

Maleki-Moghaddam, M., Yahyaei, M. & Banisi, S. (2013) 'A method to predict shape and trajectory of charge in
industrial mills' Minerals Engineering vol. 46-47, pp. 157–166.

Mishra, B.K. & Rajamani, R.K. (1993) 'Numerical simulation of chrage motion in ball mills- lifter bar effect' Minerals
and Metallurgical Processing, May, pp. 86-90.

Parks, J.L. (1989) 'Liner design, materials and operating practices for large primary mills' Advances in Autogenous and
Semiautogenous Grinding Technology, Vancouver, vol. 2, pp. 565–580.

Powell, M., Smit, I., Radziszewski, P., Clear, P., Rattray, B., Eriksson, K.G. & Schaeffert, L. (2006) 'Selection and design
of mill liners' In: Proceedings Advances in Comminution, pp. 331-376.

Powell, M.S., Weerasekara, N.S., Cole, S., LaRoche, R.D. & Favier, J. (2011) 'DEM modelling of liner evolution and its
influence on grinding rate in ball mills' Minerals Engineering vol. 24, No. 3-4.

Rajamani, R. (2006) 'Semi-autogenous mill optimization with DEM simulation software' Advances in Comminution,
SME Publication, Part 4, pp. 209–214.

Royston, D. (2003) 'A review of recent experience with large-angle wide-spaced shell lifter liners' In: Proceedings 2003
SME Annual Meeting , SME Preprint, Cincinnati, Ohio, 03-056.

Royston, D. (2006) 'Developments in SAG mill liner design' Advances in Comminution, pp. 399-412.

M. Yahyaei, S. Banisi, & Hadizadeh, M. (2009) 'Modification of SAG mill liner shape based on 3-D liner wear profile
measurements' International Journal of Mineral Processing vol. 91, pp. 111–115.

Yahyaei, M. & Banisi, S. (2010) 'Spreadsheet-based modeling of liner wear impact on charge motion in tumbling mills'
Minerals Engineering, vol. 23, pp. 1213-1219.

10

View publication stats