IMPROVEMENTS IN SAG MILL THROUGHPUT FROM

FINER FEED SIZE AT THE NEWMONT AHAFO OPERATION

*Adrian Dance1, Sonny Mwansa2, Walter Valery2, George Amonoo3 and Bryon Bisiaux 3

1
SRK Consulting (Canada) Inc
22nd Floor, 1066 West Hastings Street
Vancouver BC V6E 3X2 Canada
(*corresponding author: adance@srk.com)
2
Metso Process Technology & Innovation
Unit 1, 8-10 Chapman Place
Eagle Farm Qld 4009 Australia
3
Newmont Ghana Gold Limited
C825/26 Lagos Avenue
Ghana E Legon, Accra, Ghana

1
 IMPROVEMENTS IN SAG MILL THROUGHPUT FROM
FINER FEED SIZE AT THE NEWMONT AHAFO OPERATION

ABSTRACT

Newmont Ghana Gold Limited’s Ahafo operation have been working with Metso Process
Technology & Innovation (PTI) since early 2009 to increase mill throughput from a combination of
controlled blasting practices and optimised primary crusher and grinding circuit conditions. For such Mine-
to-Mill projects, PTI’s methodology is termed Process Integration & Optimisation (PIO).

Ahafo operate a SAG mill, ball mill and pebble crusher (SABC) circuit that processes a range of
primary ore of varying hardness from different pits which is blended with soft oxide material. The PTI
study focussed on ball mill limited material from the Apensu open pit.

Following a benchmarking study of all blasting, crushing and grinding practices in January 2009,
PTI recommended changes in both blasting and crushing conditions in order to generate a finer feed size to
the SAG mill. Additional operating changes in the grinding circuit were recommended.

A series of trials were conducted in November 2009 where three blasts were fired using a
different blast design (hole diameter and explosive type) in the same area of the Apensu pit as the
benchmarked material in January 2009. Top and bottom benches (or flitches) from each blast were sent
exclusively to the grinding circuit and the material was tracked using PTI’s SmartTag™ system.

The results from the six trials showed a significant increase in primary ore throughput at a reduced
blend of oxide material. Ahafo have since standardised on these modified blasting conditions for all
Apensu ore zones and have been able to sustain higher throughput compared to historical performance on
similar material.

This paper summarises the phases of the PIO project, the changes made in operating practices and
the outcomes of the blasting trials that validated the predicted increase in mill throughput.

KEYWORDS

SAG milling, feed optimisation, simulation, modelling

BACKGROUND

Newmont Ghana Gold Limited’s Ahafo operation is located in the Brong-Ahafo Region of Ghana,
approximately 290 kilometers northwest of the capital Accra. Ahafo poured its first gold on July 18, 2006
and commenced commercial production in August 2006. Ahafo are currently mining a number of open pits
with the majority of ore coming from the Apensu and Awonsu working areas. A blend of both primary and
oxide ore is fed to the mill to balance the head grade and recovery in the leach circuit.

Ahafo have been working with Metso Process Technology & Innovation (PTI) since early 2009 to
increase mill throughput from a combination of controlled blasting practices and optimised primary crusher
and grinding circuit conditions. For such Mine-to-Mill projects, PTI’s methodology is termed Process
Integration & Optimisation (PIO).

2
Ahafo Grinding Circuit

Ahafo operate a 54 x 75 in gyratory crusher for primary ore and a toothed MMD sizer for oxide
ore. This is followed by an SABC circuit comprising a 34 x 16.4 ft, 13 MW semi autogenous grinding
(SAG) mill in closed circuit with two, 600 kW pebble crushers. SAG milling is followed by a 24 x 39 ft, 13
MW ball mill in closed circuit with a cluster of 26 inch (660 mm) cyclones. After classification, the milling
product is thickened and pumped as feed to carbon-in-leach tanks. The SABC circuit flowsheet is shown
below in Figure 1.

Gyratory
Crusher

MMD Sizer
Primary Ore

to Leach
Oxide Ore
Tanks

Pebble
Crushers
Cyclones
Stockpile

Screen
SAG Mill
Ball Mill

Figure 1: Ahafo Grinding Circuit Flowsheet

Mill Throughput Optimisation Project

Following initial site observations, the project objectives were established to increase mill
throughput while maintaining or achieving a finer grind size. This was to be achieved through a
combination of blasting, crushing and grinding operating changes. An initial benchmarking study would
identify where the opportunities were greatest and what effect the combined changes would have on mill
performance.

Initially, the focus was to reduce the SAG mill feed size while processing the lower tonnage
Subika ore and improve the grinding circuit product size while processing the ball mill limiting Apensu ore.
Subika ore was not the main ore processed during benchmarking, leaving Apensu as the principal focus for
optimisation.

The benchmarking study of all comminution practices was conducted by PTI in January 2009,
with calibrated models of blast fragmentation, crushing and grinding circuits being used to generate a list
of prioritised recommendations. These were implemented by Ahafo including a change in blast pattern
design (hole diameter and explosive type) to create additional fines and more consistent fragmentation
between the top and bottom parts of the blasted bench.

Ahafo blast two benches at once and then mine separately the top and bottom flitches, with the top

3
flitch having a significantly coarser fragmentation compared to the bottom flitch.

Ahafo implemented the modified blasting conditions over a number of months to monitor the
effects of dilution, vibration and excavator digging rate. A series of validation trials were then conducted in
November 2009 with the new pattern design in the same area of the Apensu pit as the benchmarked
material trialed in January 2009. Top and bottom flitches from each blast were sent exclusively to the
grinding circuit and tracked using PTI’s SmartTag™ system.

PROCESS INTEGRATION & OPTIMISATION METHODOLOGY

The methodology involves a number of steps: benchmarking, rock characterization, measurements,

modeling/simulation and where required, material tracking. A PIO project is normally comprised of a
number of site visits spaced over a few months. The first site visit is to establish current operating practice,
initiate rock characterization and collect measurements of blast fragmentation and mill performance. This
is followed by modeling and simulation studies to determine how to best exploit the hidden inefficiencies.
These recommendations are then followed by further site visits to implement the changes, monitor the
results and ensure the improvements are maintained over time.

Ore Characterisation

Table 1 below compares ore hardness testwork results for the current and future Ahafo pits.
During the PIO study, the Subika and Apensu pits were predominantly mill feed material. The JKMRC
Drop Weight A*b values in Table 1 compared to Bond Work Index (BWi) values show that Subika ores
are typically SAG mill limited while Apensu ores are ball mill limited in the Ahafo SABC circuit.

Table 1: Ahafo Ore Properties

Ore Source Drop Weight Bond Ball Work Circuit
A*b Index, kWh/t# Limitation
Subika 22 to 28 15 to 17 SAG mill
Apensu 33 to 38 15 to 22 ball mill
Amoma 26 to 57 11 to 21 variable
Awonsu 36 18 ball mill
$
Oxide 76 10
# $
BWi at closing screen size of 106 m A*b value derived from BWi by SMCC Pty Ltd.

Ahafo also blend in oxide ore to balance the head grade and leach circuit recovery which also has
an impact on mill tonnage.

Benchmarking Study

The first step of a PIO project is to benchmark the current practices by auditing the operation and
control of the blasting, crushing, grinding and flotation processes.

The quality of blast pattern implementation is assessed and the resulting ROM fragmentation
measured using image analysis. The crushing, grinding and separation circuits are surveyed and process
control strategies reviewed. All of these measurements allow mathematical models to be developed for the
complete process chain. These models are later used to simulate the impact of operational changes in the
mine or concentrator on the entire process.

4
 A particular issue at Ahafo is the practice of blasting double benches (or flitches) which are mined
separately. With the lower explosive loading in the top half of the blast holes, the top flitch generally has a
much coarser fragmentation than the bottom flitch (see Figure 2 for example photos).

Figure 2: Example Blast Fragmentation from Top vs. Bottom Flitch (Double Benching)

To quantify the effect of changes in blasting conditions, the PIO study trialed the top and bottom
flitch material separately into the grinding circuit.

Ore Tracking

In any Process Integration and Optimisation project, it is crucial that the material being fed to the
plant during the survey is the same material that has been characterized in the mine from the audited blast.
PTI have developed a material tracking system, called SmartTag™, that allows parcels of ore to be
followed from the mine, through the crusher and intermediate stockpiles or ROM pads, and finally into the
grinding mills. The SmartTags™ are built around robust Radio Frequency ID (RFID) transponders. Figure
3 shows the hardened tag and temporary antenna installation under a conveyor belt. The tags require no
internal power source and therefore can remain in stockpiles and ROM pads for extended periods before
being detected by the system hardware. The current design allows tags to be detected from a distance of up
to one metre that is ideal for conveyor belt applications.

Figure 3: SmartTag™ RFID Ore Tracking Tag & Antenna Installation

5
 Following detonation of the audited blast, the top flitch material was mined and placed in a
segregated stockpile on the ROM pad ahead of the primary crusher. The bottom flitch material was later
directly dumped into the primary crusher. SmartTags™ were placed on the muckpile surface prior to
mining as well as added to the ROM stockpiles. A total of 200 tags were placed in each flitch with two
antennae installations to detect the tags, one on the primary crusher product belt and one after the stockpile
on the SAG mill feed belt.

The use of SmartTags™ provided clear evidence that the audited material was entering the
grinding circuit and the frequency of tag detections was used to indicate whether any other ore polygons
were inadvertently being sent to the primary crusher. Placing antenna both ahead and after the coarse
stockpile allowed live retention time to be estimated.

Modeling/Simulation Exercise

The benchmarking trial on Apensu top and bottom flitch material provided detailed data on blast
fragmentation, primary crusher and grinding circuit performance that was used to developed models of the
entire system. Image analysis of photos taken from the muckpile and trucks dumping at the primary
crusher were used to calibrate a blast fragmentation model. Samples of the primary crusher product
material were used to confirm the amount of fines being generate in the blast. Complete grinding circuit
surveys conducted while processing the top and bottom flitches were used to calibrate a JKSimMet model.

A simulation study was then completed to identify opportunities for higher mill throughput. A
range of recommendations were made including changes in blasting (hole diameter and explosive type),
primary crusher operation and grinding circuit operating conditions. Simulation results on similar Apensu
material (combined with PTI experience) suggested that changes in blasting alone could increase mill
tonnage by 4% while grinding circuit changes alone could increase tonnage by 6%. The combination of
blasting and grinding circuit changes should improve throughput by even more than the sum of the two
changes.

Ahafo received these recommendations in May 2009 and implemented the new blasting
conditions immediately for all ore zones in the Apensu pit. By late 2009, Ahafo were ready to conduct the
validation trials.

BLAST VALIDATION TRIALS

The objective of the blast validation trials is to confirm (as accurately as possible) the expected
benefit of finer feed predicted in the simulation study. If possible, the same or similar material should be
trialed so that the effect of only recognised changes are shown in the results. It is important to complete
such validation studies as, in the future, changing feed or circuit conditions may mask the benefits of finer
fragmentation. If higher drill & blast costs (e.g. powder factor) are an outcome of a Mine-to-Mill or PIO
exercise, documented evidence needs to be available to justify the additional operating cost.

Trial Program

Ahafo set up three blast patterns using the new conditions in the same area of the Apensu pit as
was trialed in the benchmarking study earlier in 2009. These blast patterns were mined separately in top
and bottom flitches, resulting in six mill trials that were carefully monitored by PTI along with samples
collected for sizing and hardness testwork.

6
 Figure 4: Detonation of One of Three Validation Trial Blasts

Due to limited material in each blast pattern, mill trials were conducted for eight hours before
material was depleted.

Trial Results

The results of the validation trials showed a significant increase in mill throughput as a result of
the finer fragmentation. Additional photographs of the muckpile and trucks dumping at the primary crusher
confirmed an increase in fines content as well as greater similarity between the top and bottom flitch
fragmentation. Primary crusher performance improved with higher throughput at the same power draw
compared to the January 2009 benchmarking study.

Table 2 summarizes the six validation trials compared to the benchmarking results for top and
bottom flitch material separately. The expected increase in mill throughput predicted in the simulation
study from the blasting and mill operating changes was clearly outdone by the validation trials.

Overall, a significantly higher feedrate was observed on both flitch material types, with a lower
blend of oxide. Tonnages on both top and bottom were 1,100 to 1,200 tph indicating the change in blast
design likely had a greater effect on top flitch fragmentation (as expected).

The finer material reduced the SAG mill specific energy to around 8.5 kWh/t from the
benchmarking results of 10.3 to 10.8 kWh/t.

Table 2: Summary of Blasting Trial Results on SAG Mill Performance

Feed Properties
Survey Flitch DWi, BWi, Feedrate, % Oxide SAG Grind P80,
kWh/m3 kWh/t tph kWh/t microns
Benchmark Top 7.9 15.2 971 23.8 10.3 156
Trial 1 Top 8.3 20.8 1,167 21.9 8.4 177
Trial 2 Top 7.4 22.0 1,187 21.7 8.5 161
Trial 3 Top 7.3 18.7 1,132 13.8 8.2 148
Benchmark Bottom 10.1 16.0 1,051 22.0 10.8 125
Trial 4 Bottom 7.5 18.1 1,117 12.3 8.7 214
Trial 5 Bottom 7.3 21.8 1,180 17.9 8.1 221
Trial 6 Bottom 8.5 18.0 1,149 11.9 9.0 221

7
 The lower % oxide in the trials compared to benchmark feed was evident in the coarser final grind
80% passing (P80) sizes – particularly for the bottom flitch trials. In addition, samples of the trial feed
showed a higher BWi value than was reported in the benchmark study. Further recommendations on ball
mill circuit operating changes are being reviewed by Ahafo (including the option of recycling cyclone
underflow back to the SAG mill). Grind size improvement is an ongoing project being investigated by
Ahafo personnel since early 2010.

Blasting vs. SAG Mill Specific Energy

One of the aspects of a PIO exercise is to recognize that blasting is the first stage of comminution.
The operational gains to be made are from shifting the overall comminution energy from the less efficient,
higher cost stage (grinding) to the more efficient, lower cost stage (blasting). In addition, there are benefits
in dealing with better fragmented/comminuted material early on in the process (e.g. excavator dig rates,
truck load factors, primary crusher productivity, etc.).

The two figures below illustrate the shift in energy measured during the Ahafo mill feed
optimisation project. The first figure relates to the increase in blasting energy and the second, the SAG mill
response to the finer feed with a reduction in specific energy.

Other Operations
Other Newmont Operation
Ahafo
Ahafo Post-PIO
Blast Pattern Powder Factor, kg/t

Ahafo Pre-PIO

Unconfined Compressive Strength, MPa

Figure 5: Blasting Powder Factor vs. Rock Strength

Figure 5 shows Ahafo blasting powder factors (kilogram per tonne of rock) compared to other
operations against rock strength or Unconfined Compressive Strength (UCS). Included in the operations is
another Newmont site that has routinely applied differential powder factors for their range of rock
properties for a number of years. The blue diamond icons cover a range of operations and different rock
strengths and reveal that operations are adjusting their blasting energy for harder rock strengths even
without a PIO exercise being undertaken.

The Newmont site (orange circles) participated in a PIO project and identified blasting domains
that demanded different blasting conditions or energy levels to achieve a consistent mill feed size
distribution.

8
 The Ahafo blasting energy level prior to the PIO exercise is shown in Figure 5 compared to the
blast validation trials (post PIO). Measurement of the rock strength for the validation blasts showed that the
UCS was slightly higher than the benchmarked material.

In response, the reduction in SAG mill specific energy for the blasting trials is shown in Figure 6
against the measured Drop Weight Index (DWi) in kWh/m3. The circled area for the six blasting trials
showed an average SAG mill specific energy of 8.5 kWh/t for similar DWi values. This is compared to the
average benchmark result of 10.4 kWh/t.

14
Other Operations
Other Newmont Operation
Ahafo
12
Benchmark
Benchmark (Bottom Flitch)
(Top Flitch)

10
SAG Mill Specific Energy, kWh/t

Apensu
6 Blasting Trials

0
0 2 4 6 8 10 12
Drop Weight Index, kWh/m3

Figure 6: Change in SAG Mill Specific Energy with JKMRC Drop Weight Index

The reduction in SAG mill specific energy by 18% is a significant improvement in efficiency
from the finer feed size generated by value-added blasting. Table 3 below summarizes the throughput
increase at Ahafo over the past few years since the start of the PIO project.

Table 3: Throughput Improvements at Ahafo Over Time

Condition Period Primary tph Oxide tph Total tph % -106µm
Before PIO Oct 08 – Jan 09 722 294 1,016 –
Benchmark Average Jan 09 780 231 1,011 75.7
Validation Trials Nov 09 965 191 1,156 76.6
Ongoing Jan – Mar 10 938 163 1,101 76.3
% Change 30% -45% 8% –

ADDITIONAL GRINDING CIRCUIT IMPROVEMENTS AT AHAFO

Following recommendations by Metso PTI, Ahafo has maintained tighter settings on both the
primary and pebble crushers to maximise the fineness of the product size. Additionally, Ahafo

9
implemented the following in June 2010:

a Metso advanced control package on the SAG mill including AudioMill™, VisioRock™
(primary crusher product and SAG mill feed) and OCS© based expert grinding control software;
installation of an additional cyclone to allow more flexibility and remove the restriction to control
feed density when the pressure is rising;
increased ball mill media charge resulting in higher power draw.

The above have resulted in increased plant stability and improved throughput.

CONCLUSIONS

Newmont Ghana Gold Limited’s Ahafo operation have been working with Metso Process
Technology & Innovation since early 2009 to increase mill throughput from a combination of controlled
blasting practices and optimised primary crusher and grinding circuit conditions.

Following a benchmarking study of all blasting, crushing and grinding practices, PTI
recommended changes in both blasting and crushing conditions in order to generate a finer feed size to the
SAG mill. Additional operating changes in the grinding circuit were recommended.

Blast validation trials conducted in late 2009 showed a significant increase in mill throughput due
to the finer feed conditions. The trial results also showed a general coarsening of the final grind size –
likely from the lower blend of oxide in the feed. Grind size improvement is an ongoing project being
investigated by Ahafo personnel since early 2010.

Ahafo have since standardised on this blast design for all Apensu ore zones and have been able to
sustain higher throughput compared to historical performance on similar material.

The PIO exercise completed at Ahafo demonstrated the importance in considering blasting as the
first stage of comminution. A small increase in well-applied blasting energy resulted in a reduction of SAG
mill specific energy by 18% for similar hardness feed material.

Ahafo continues to monitor the effect of feed size on the grinding circuit using the VisioRock™
image analysis system on the SAG mill conveyor belt and by taking samples occasionally for ore hardness
characterisation. Through a continuous partnership between Ahafo and Metso PTI, significant
improvements in overall process performance have been achieved using the PIO methodology along with
other initiatives implemented by Ahafo.

ACKNOWLEDGEMENTS

The authors would like to thank Newmont Ghana Gold Ltd for their permission to prepare and
present this paper.

10