Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

SME Annual Meeting

Feb. 28-Mar. 03, 2010, Phoenix, AZ

Preprint 10-042

DESIGN OPTIMIZATION OF GRASBERG BLOCK CAVE HAULAGE

D. Marsh, Univ. of Arizona, Tucson, AZ


C. Peppin, Stantec Mining, Tempe, AZ
I. Ross, Freeport-McMoRan Copper and Gold Inc., Phoenix, AZ

ABSTRACT the material to the surface. This system should transport an average
of approximately 160,000 tonnes per day (tpd).
The feasibility study for the Grasberg Block Cave mine, finalized
in May 2008, indicated a production capacity of 160,000 tonnes per
day. A rail haulage system was an integral part of the mine design.
This paper describes the simulation process used in optimizing the
design and discusses the key factors and variables that impact the
capacity of the system. The final recommendations for the haulage
level layout and associated equipment are also included.
INTRODUCTION
Freeport-McMoRan Copper and Gold Inc (FCX) is an international
mining company with large long-lived, geographically diverse assets
and significant reserves of copper, gold and molybdenum.
PT Freeport Indonesia (PTFI) is a subsidiary of FCX. Its principal
asset is the world-class Grasberg open pit mine which was discovered
in 1988. The mine is located in the Sudirman Mountain range of
Papua, Indonesia. These mountains contain the Grasberg/Ertsberg
mineral district which forms one of the world’s largest copper and gold
reserves. Current PTFI Reserve estimates show about 2.7 billion
tonnes of ore with approximately 75% of them in undeveloped
underground reserves.
There are six main areas within the Reserves: Grasberg Open Pit,
Grasberg Block Cave (GBC), Kucing Liar, Big Gossan, Deep Ore Zone
(DOZ) and the Deep Mill Level Zone (DMLZ). These are depicted in
Figure 2. GBC Rail Haulage Layout.
Figure 1.
OBJECTIVES

N Grasberg
The aim of the work described in this paper is to optimize the ore
handling system design for the GBC mine and thus determine the
open pit
Grasberg appropriate ore haulage elevation.

Kucing Liar The feasibility study had demonstrated, by simulation, that


160,000 tpd was achievable from the planned rail haulage system. A
DMLZ
DOZ
decision to move secondary crushing units from the underground
Big Gossan Grasberg layout (see Figure 3) to a location on surface meant that 25 metres of
Block Cave height was available to move components of the ore handling system,
Kucing
Liar
without impacting on the elevation of either the extraction level or the
Mill
N Amole tail pulleys of the incline conveyors. Figure 4 depicts the flexibility
Portals
2,900 m elev DOZ resulting from the plan to remove the secondary crusher.
Plan View
Big
Kucing Liar Spur
Gossan This allowed elevation changes to be considered for both the rail
Grasberg
MLA
BC Spur DMLZ
haulage level itself as well as the primary crushers. Elevation change
Big Gossan Spur DMLZ Spur impacts the available surge capacities within the orepasses, between
the rail dump and the crusher or between the crusher and the
Portals Common Infrastructure
(at Ridge Camp) 2,500 m elev conveyor.
Figure 1. Grasberg/Ertsberg Ore Bodies. The optimization process should determine the optimal capacities
The GBC ore body contains 1,007 Mt at 1.02% copper and 0.81 for all of the main components of the haulage system and hence
g/t gold. The ore will be extracted by block caving utilizing a “straight determine the appropriate haulage level elevation.
through” (or El-Teniente style) drawpoint layout containing 2,356 METHODOLOGY
drawpoints. The haulage level will consist of railed equipment loading
from ore chutes with the capability of loading from two chutes Simulation
simultaneously. The rail haulage layout is depicted in Figure 2. The optimization process was to be performed using a simulation
Bottom discharge rail cars will dump into gyratory crushers which feed approach. A model of the ore handling system was developed using
®
crushed material onto two 1.82 m wide (72”) conveyer belts that bring Arena simulation software and followed a modular approach to ensure
that future enhancements and system add-ons could be
1 Copyright © 2010 by SME
SME Annual Meeting
Feb. 28-Mar. 03, 2010, Phoenix, AZ

accommodated. A Microsoft Excel interface enabled entry of input that allowed the variables for a number of scenarios to be input and
parameters and reporting of results from alternative simulation run as a batch. Scenarios did not have to be run individually.
scenarios. Model results are presented and updated by Arena in Excel
worksheets. The model and the associated logic were developed by The logic in the simulation allows the program to simulate many
an external consultant. “what if” scenarios that occur in a block cave mining operation. For
example, the program takes into account delays such as random
equipment breakdowns or hung-up drawpoints. On completion of the
simulation run, a results sheet is created for the specific scenario. The
results sheet contains productivity data for equipment and all
production data along with other user-defined results.
Variables
The variables used for this series of scenarios included:
• Number of trains
• Number of cars per train
• Orepass length
• Rail dump pocket capacity
• Crushed ore bin capacity
If each of the variables above only had 5 reasonable values, that
would require 3,125 simulations to test all the possible combinations
5
(5 ). To reduce the number, many of the non-plausible variable
combinations could be eliminated based on inspection of scenario
Figure 3. Feasibility Study Layout of Rail Dump, Crusher and Ore results that did not come close to meeting the 160,000 tpd
Handling System. requirement.
For example, if a scenario was simulated using 4 trains with 24
cars and the result was significantly short of 160,000 tpd, simple
reasoning will eliminate all 4 train scenarios with less than 24 cars.
However, even after initially eliminating the unlikely scenarios,
hundreds of possible combinations remained.
Some input variables such as the LHD loading and tramming
time, drawpoint hang-ups, and time required to break oversized
material on the grizzlies were kept constant. Early scenario results
showed that sufficient associated resources were available to ensure
that they were not limiting factors in the system.
The base cases from the feasibility study determined that the ore
handling system would be capable of producing 160,000 tpd and these
were used as general starting points. A broad range of scenarios were
run and any scenarios that yielded results 5% lower than the 160,000
tpd requirement were discarded.
Figure 4. Proposed Removal of Secondary Crusher.
Smaller variations to each scenario were conducted with a focus
The model simulates and evaluates the ore handling system of on holding multiple variables constant and only changing one variable
the 160,000 tpd GBC mining operation. The model is utilized to at a time. Exceptions to this were where two variables were linked by
evaluate changes in ore handling productivity. Such changes may a common constraint (one can only increase if the other decreases).
result from fluctuations in ore handling and process equipment This approach allowed better understanding of the effect each variable
capacities, variations in the number of available drawpoints, deviations had on the production rate.
in ore fragmentation, and modifications to designed surge capacities.
Trains and Cars
Operations simulated on the extraction level include the following: The number of trains operating on the haulage level was varied
along with the number of ore cars per train. The feasibility base case
• Loading from production drawpoints by LHDs and tramming used 6 trains of 24 cars with a spare train. Scenarios ranging from 4
to orepasses. trains to 12 trains were run with a range of cars and rail dump surge
• Drilling and blasting of drawpoint oversize and hang-ups capacities in order to determine the most robust combination.
(frequency estimated using a Geotechnical model).
• Breaking of oversize rocks on orepass grizzlies as predicted The combination of 6 trains with 20 cars appeared to be the
by modeling completed by Stantec. combination which was the most suitable. This was then used in the
further scenarios where ore pass and surge capacities were varied.
Simulated operations on the haulage level and ore handling
system include the following: Orepass Length and Dump Capacity
The orepass capacities (volumes) were varied by changing the
• Loading of trains orepass lengths from the extraction level to the ore haulage level. The
• Hauling from orepasses to crushers and dumping into length of the ore passes and crusher dumps were varied in correlation
pockets with each other in 5-m increments. For example if the ore passes were
• Crushing and conveying of ore to the surface stockpiles at lengthened by 20 m the crusher dump would only be lengthened by 5
the mill. m.

With a set of input variables established and entered in the Excel The calculated capacity for the crusher dump was 1,000 tonnes
file, a user can run a simulation that covers a chosen amount of time. per 5-m increment. It was assumed that all the orepasses have the
All the scenarios were set up to simulate a 12 month period in order to same constant diameter and the capacity was calculated based on
gain statistical reliability of the outcomes. An input macro was utilized overall length. The extraction level (top of the orepass) was kept
constant and the haulage level elevation was varied.
2 Copyright © 2010 by SME
SME Annual Meeting
Feb. 28-Mar. 03, 2010, Phoenix, AZ

Determining the optimal relationship between ore pass volume, The number of runs used in this process was 815. This number
rail dump pocket and crushed ore bin capacities, will enable the of scenarios might appear to be excessive. However, when the life of
optimal orepass length to be determined and consequently dictate the the mine and the ore to be handled is considered (almost 20 years and
elevation of the haulage level. 1 billion tonnes), this is a relatively small investment in time for such a
significant undertaking.
SIMULATION RESULTS
The optimization process has allowed management to make the
Trains and Cars decision on the appropriate elevation for the rail haulage.
Initial results from the scenarios varying the number of trains
indicated that 6 trains with 20 cars each (plus a spare train) appeared REFERENCES
to be the optimal combination because it met the 160,000 tpd most
reliably when other capacities were changed. This combination was Hewitt, S., Sudjatmoko, Casten, T. and Brannon C. 2008. Grasberg
then used in the next series of scenarios to determine the optimal Block Cave Access and Logistic Support Systems. In
combination of pass capacities and available surge capacity. Proceedings MassMin 2008, Lulea, Sweden, June 9-11. Lulea,
Sweden: Lulea University of Technology Press.
Model Revision
During the process of running all the scenarios some problems Botha, J., Watson, S., Arkadius, T. and Samosir, E. 2008.
were encountered. Where repeated instances of apparently illogical Simulation Applications at PT Freeport Indonesia’s DOZ/ESZ
results or trends were observed, the root causes were sought. After Block Cave Mine. In Proceedings MassMin 2008, Lulea, Sweden,
investigation, amendments were made to the train dispatch logic that June 9-11. Lulea, Sweden: Lulea University of Technology Press.
solved the ambiguous results. Once the new version of the Arena ACKNOWLEDGEMENTS
model was created, this effectively made all the previous scenario runs
obsolete. The previous runs needed to be re-run with the new Arena The authors wish to thank to Freeport-McMoRan Copper and
model to ensure that the results could be properly compared. Gold for permission to publish this paper. Thanks also go to Stantec
Mining and Jaco Botha for their assistance during the simulation
Pass Capacities process.
Using 6 trains with 20 cars, the results indicated in Figure 5 show
the improvement in productivity of the system as Rail Dump capacity
was increased. It can be seen in Figure 5 that the outcomes are non-
linear. The general trend is that higher production tonnage is achieved
with more rail dump pocket capacity but there are diminishing
increases above 2,200 tonnes.

Figure 5. Rail Dump Capacity versus Average Daily Production.


The system output was not sensitive at all to the volume of the
crushed ore bin. This is ascribed to the fact that the conveyors have a
higher capacity than the crushers. The amount of surge available in
the crushed ore bin is insignificant when compared to the crusher
throughput or the conveyor capacity.
Results also indicated that in general terms, greater orepass
capacity improved total system productivity. The gains were relatively
small, probably due to the smoothing effect in the system with
approximately 120 passes and the scenario being measured over 365
days of production. The 2,200 t Rail Dump pocket capacity requires an
increased of 10 m in height. This means that a 15 m increase in the
ore pass length is still possible whilst remaining within the 25m of
available height. This was seen to be an optimal solution and the
recommendation is for the ore haulage level to be lowered 15 m from
the elevation used in the feasibility study.
CONCLUSIONS
The simulation approach is an appropriate tool to model the
®
complex ore handling system planned for the GBC. The Arena
software was flexible enough to allow all the relevant variables to be
tested to find the optimal combination.

3 Copyright © 2010 by SME

You might also like