Seventh Large Open Pit Mining Conference 2010 PDF

SEVENTH LARGE OPEN PIT
MINING CONFERENCE 2010
27 - 28 JULY 2010
PERTH, WESTERN AUSTRALIA
The Australasian Institute of Mining and Metallurgy

Publication Series No 5/2010
msirnrn
THE MINERALS INSTITUTE
Published by:
Level 3, 15 - 31 Pelham Street, Carlton Victoria 3053, Australia
i
OThe Australasian Institute of Mining and Metallurgy 2010
All papers published in this volume were refereed prior to publication.
The Institute i s not responsibleas a body for the facts and opinions advanced
in any of its publications.
ISBN 978 1 921522 22 2
Desktop published by:

Kylie McShane and Olivia Tet Fong
Printed by:
BPA Digital
11 Evans Street, BurwoodVictoria 3125, Australia
ii
FOREWORD
The AuslMM's Seventh Large Open Pit Mining Conference is the open pit operators conference
of choice and is in keeping with one of the aims of The Institute - to provide a professional
forum for members and industry experts to present and discuss those issues, innovations and
achievements that are a t the core of their professional activities.The AuslMM is recognised as the
leader in mining industry conferences and the Seventh Large Open Pit Mining Conference is no
exception. We welcome you to Perth and to this conference and appreciate your contribution to
the success of this event.
The mining industry has in the time since our last Open Pit Mining Conference seen boom
times, the global financial crisis (GFC), depressed commodity prices, redundancies and tight
time, an apparent emerging return to boom times, significant innovations and advancements in
technology and application within the industry and a significant threat to the Australian mining
industry through the lack of understandingof the financing of mining projects and the real effect of
a proposed taxation change. Western Australia and the mining industry appear to be being accused
of recovering and advancing too fast for some other areas (of Australia) and in some people's eyes.
All of these issues promote active discussion and the papers at this conference will address some of
these and beyond and promote active discussion and ideas for future improvements.
In order forThe AuslMM to deliver to you a world leading conference such as the Seventh Large
Open Pit Mining Conference, we are indeed indebted to the authors who have prepared abstracts
and papers, and the presenterswho have contributed significantlyto this conference. Without their
contribution we may still have a very enjoyable social event and some networking and discussion,
but we would not have the stimulus and discussionthat will result from their papers.The AuslMM
is indeed appreciative of their contribution and we look forward to their presentations. One effect
of the GFC resulted in some people having abundant time on their hands. We saw significant
redundancies and four day weeks and as we have begun to return to some normality, and even
as some would say another emerging boom time, we have noticed those of us still in the industry
with spare time on our hands have shrunk rapidly. This has been an influence on the number of
potential papers that have been able to be delivered. This also emphasises the contribution by
those who have been able to deliver and we thank them for that. We do acknowledge those who
had intended to contribute to the conference but were unable to do so due to work pressures and
commitments. We encourage them to keep in touch.
The AuslMM is in the privileged position of having the quantity and quality of members and
friends of The Institute who volunteer their time to assist the Conference and Events Department
in planning and running conferences.Their effort contributes to the high standard achieved in such
conferences and maintains the very significant cost effectiveness to members and non-member
attendees. We are indeed very appreciative of the contribution of those members and friends. The
Institutewhich is one ofthe few mining conference organisers that maintainthe peer review process,
ensures the standard of AuslMM conferences and we are very grateful to the effort and time they
have provided to this conference. Your committee and peer reviewers are listed separately in this
volume and you are welcome to peruse that list and acknowledge their contributionsas appropriate.
vii
Please take the time to review the list of sponsors and exhibitors listed separately in this volume and
visibly represented throughout the conference. Without their support and assistance we would not be
able to bring you this quality and cost-effective conference.
The position and reputation that The AuslMM enjoys as the leader in mining conferences is due to the
expertise and experience ofThe InstitutesConference and Events Department within AuslMM Services.The
conference committee lost our original conference coordinator, Katy Andrews, but in keeping with being
a learned society we were pleased to lose her to university studies. We thank her and the Conference and
Events team for their contributions and welcome Belinda Martin to our committee as Katys replacement.
We appreciate her contribution and look forward to this and future conferences with her.
Toyou,thedelegatesattending thisconference,weagain say welcomeand appreciate your contribution
to making this the success it will be. To those who were not able to enjoy the privilege of attending the
conference but are now reading this volume we trust you enjoy it and are motivated to join with us in
the extras we have in networking and discussion associated with the presentationsat future conferences.
Richard Flanagan
Conference Chair
viii
CONTENTS
Keynote Address
Innovation in an Uncertain World G Danckert 3
Papers
Enhanced Drilling in Rio Tinto CMcHugh,A Stokes, 7

F Oppolzer, R Hennessy and
S Smith
Practical Management of Progressive Large Failures B D Hendersonhall, D Lucas, 15
at Minerals and Metals Group Century Mine N Kerr and C Pennisi
The Development of a Novel Eye Tracking Based D Zhu, T Gedeon and 31
Remote Camera Control for Mining Teleoperation K Taylor
Sydvaranger Magnetite Mine, Norway - Owner S Cruickshanks 39

Mining Tenders and Implementation
Slope Monitoring in Open Pit Mining - A Case S Narendranathan, A J Beer 45

Study and M E Heap
Minimising Crest Loss by Eliminating Surface A Marton 57

Dilation Due to Presplitting
Ensuring Professional Engineering Capability - E Taylor and C Murray 65

Queensland Professional Engineers Act 2002
Three-Dimensional Hydromechanical Simulation D Beck, B Fillery and 71

for Large, Faulted Pit Slopes F Reusch
Linking Operator Competency with Performance P Davis 79
Capacity Increase of the Worlds Strongest Belt M Guerrero, H Reynaud, 81

Conveyor - Pushing the Limits in Chilean Copper S Tappeiner and M Schmid
Ore by RWE
Incorporating Economic Risk in Pit Slope Designs F J Lai, W E Bamford, 83

S T S Yuen and T Li
Mine Planning and Risk Management Associated L Cerny, J Dixon, 93

with the Breakthrough of the Telfer Sublevel Cave C McArthur and U Singh
into the Active Telfer Open Pit
Proximity Cameras and Global Positioning Systems N Ferreira and C McElman 103
- An Integrated Approach
Payload Variance Effects on Truck Bunching P F Knights and S Paton 111
Productivity Enhancements to Increase Ore A Renstrom 115

Production from 18 to 36 Mt at Aitik
Mine Control - Why do we do it, and how do we R Ballantyne 121

Achieve Sustainable Success?
Long-Term Mine Planning at the Largest Phosphate R de Lemos Peroni, 127

Mine in Brazil L N Capponi, J FCosta,
E Gontijo and D C Veiga
Author Index 135
xvi
Keynote Address
Innovation in an Uncertain World
G Danckert'
ABSTRACT
As Pliny commented two millennia ago, The only certainty is that nothing is certain.
Despite the many breakthroughs, advances and promises, the way we each live and work remains awash with technological,
economic and political uncertainty. The resources sector is no exception in this regard.
Mining itself is much more than a complicated process or recipe; it is a complex adaptive system. The way we adopt and adapt
new technologies, the way we evolve data into corporate wisdom, the way we make decisions and even the way we view our concept
of rationality should each be the subject of intense review.
Translating these sentiments into sustainable company growth is indeed the challenge. This is unlikely to be achieved within
convention, but through innovation.
The game is changing!
1. MAuslMM, General Manager, Resource Development - Iron ore, RioTinto, Central Park, 152-158 St Georges Terrace, Perth WA 6O00. Email: gerard.danckert@riotinto.com
SEVENTH LARGE OPEN PIT MINING CONFERENCE / PERTH, WA, 27 - 28 JULY 2010 3
Papers
Enhanced Drilling in RioTinto
C McHugh, A Stokes*, F Oppolzer3,R Hennessf and S SmithS
ABSTRACT
Rio Tinto has embarked on a journey to use automation to improve the performance of its surface
mines. The driver for automation is to improve the accuracy and consistency of mining processes
and less to do with cost savings through labour reduction.
A key task in most surface mines is blasthole drilling. The degree to how well this is done cascades
through the downstream mining process. It impacts scheduling, excavation, slope stability,
material handling, beneficiation and final product blending. Improved blasthole drilling was
identified as an objective that has a significant business case in most of Rio Tintos surface mines.
This paper describes Rio Tintos drill enhancement project that has developed a robust system
for automating most of the functions on a Terex SKSS16 drill rig. Automated drilling is currently
being used in production at the West Angelas iron ore mine in the Pilbara region of Western
Australia. A description of the essential elements of the drill automation system is given together
with an overview of how automated drilling is performed remotely from the mine office.
A review of some of the benefits of the enhanced drill system is given together with the future
plans to extend the functionality of the automated drill and the expected benefits.
INTRODUCTION
Rio Tinto first evaluated the potential of automating surface under blasting can create large blocky material which lowers
mining equipment and associated mining processes in a study productivity and increases maintenance costs.
conducted back in the mid-1990s. The conclusion at this RTIO drills over 600 ooo blastholes every year and therefore
time was that there were numerous benefits to introducing giving consistency in this process has the potential to deliver
automation into a surface mining operations and indeed some significant value to the Company. Figure 2 shows an aerial
forms of mining automation was imminent. Some early forms view of a typical blasthole drill bench at RTIO Pilbara mines.
of automated haul trucks had already been demonstrated and These are generally large and of regular shape making them
displayed at mining shows. Drill monitors were becoming ideal for the deployment of an unmanned automated drill.
commonplace and papers had been written citing experiences
The business case for this project was predicated on
of automating functions on blasthole drills. A conclusion improving the accuracy of the drilling process in RTIO Pilbara
of the study was that most of the requisite technology for mines. Reduction in labour was not seen as a significant driver
automating blasthole drills existed and OEMs would begin although it was recognised that there was some potential
marketing automated surface drills within the next few years. advantage if a single operator were able to operate multiple
This was identified as a good starting point at which to begin drills. Improving blasthole drilling accuracy and quality
automation within a surface mine due to the limited number of has several advantages namely, hole placement, improve
interactions between the drill rig and other manned activities. verticality of the hole, more accurate hole depth and reduced
By 2005 little had changed. Despite the fact that all the redrills through improved hole quality and stability. Within
essential elements required to create an automated surface RTIO Pilbara mines over 2 0 drills were identified as being
drill had been available for some time, no automated rig potential targets for automation. It was also recognised that if
was offered by any of the major mining drill manufacturers. successful the same technology would be applicable at many
Rio Tinto decided that there was sufficient value in drill of Rio Tintos other surface mines in a variety of commodities.
automation to develop its own. All of the development and field testing of the automated
Effective blasting is a key business driver for Rio Tinto drilling has been conducted at RTIO West Angelas mine.
Iron Ore (RTIO) Pilbara mines located in north of Western
Australia (Figure 1). Poor drilling and blasting has the potential APPROACH
to destroy value by pulverising lump ore (that attracts a price The drill enhancement project has had a phased approach.
premium), causing excessive dilution and ore loss. Conversely, The initial phase was to prove the viability of automating a
1. MAuslMM, Manager - Automation, RioTinto, 1 Research Avenue, Bundoora Vic 3083. Email: Charles.McHugh@RioTinto.com
2. General Manager - Automation, RioTinto, 1 Research Avenue, BundooraVic 3083. Email: Andrew.Stokes@RioTinto.com
3. Senior Advisor - Automation, RioTinto, 1 Research Avenue, BundooraVic 3083. Email: Florian.Oppolzer@RioTinto.com
4. lead Technical Engineer, RioTinto Centre for Mine Automation, Australian Centrefor Field Robotics, Rose Street Building, University of Sydney, NSW 2006. Email: r.hennessy@adr.usyd.edu.au
5. Product Support Manager, RemoteControl Technologies Pty Ltd, Unit 3-5/51 1 Abernethy Road, Kewdale WA 6105. Email: shane@rctnet.au
C McHUGH eta/
FIG 1-location of RioTinto Iron Ore Pilbara operations.

The automated functions included tramming the drill
accurately over designed hole location, levelling, collaring,
drilling of the hole and lowering of the drill ready for the next
hole in the pattern.
The automation of the drill progressed with a number of
discrete development steps as follows:
develop drill automation algorithm and demonstrate on
laboratory drill,
selection and conversion of a standard blasthole
production drill to electro-hydraulic control,
install remote control system which allows auxiliary
control by automation computer,
develop safe systems of work for the drill,
commission line of sight tele-remote control,
commission automated levelling function,
commission automated drilling function,
commission automated tramming function, and
I monitor performance against identical manually operated
FIG 2 -Aerial view of a typical RioTinto Iron Ore drill bench.
production drills.
drill with tele-remote operation from a nearby location. This
line of sight phase allowed a high integrity communications Drill automation algorithms
link to be maintained between the driller and drill. Following The diverse uniaxial compressive rock strengths encountered
the successful completion of the line of sight remote operation in the RTIO Pilbara mine sites can range from 1 MPa in
the following phases were designed to remove the driller to shale to 150 MPa in the banded iron formations. The main
non-line of sight locations and finally for the driller to control production drills in the RTIO fleet are capable of changing
multiple drills simultaneously. between rotary drilling in the softer rock and down the hole
hammer (DTH) in the harder rock. It is possible for a drill to
Phase one - Line of sight operation encounter a wide variety of rock strengths within a single hole.
The objective of this phase of the program was to retrofit an The experimental rotary drill at the University of Queensland
existing Terex- SKSS 16 diesel drill so that it could be operated was used to develop a set of algorithms capable of exerting
remotely by an operator located in a cabin within line of sight precise control over the drilling parameters. Samples of the
of the drill. In addition to tele-remote control the system was common rock types encountered in the Pilbara were shipped
designed so it could perform most of the routine functions to the University of Queensland where they were drilled using
automatically with minimal intervention from the operator. an automation interface that monitored the drill sensors. The
8 SEVENTH LARGE OPEN PIT MINING CONFERENCE / PERTH, WA, 27 - 28 JULY 2010
ENHANCEDDRILLING IN RIOTINTO
trials demonstrated that the algorithms could rapidly adapt shutdown system into the console. The independent shutdown
the pull down and rotation pressures when drilling two rock system operates with minimal software and is designed as
units of different strengths superimposed on each other. Once a redundant shutdown in case of emergency. RCT installed
the basic algorithms had been developed in the lab the next flashing beacons that indicate the drill is being remotely
stage was to customise it with the input of site production operated. Seven fixed cameras were installed to allow drillers
drillers. The inputs of site production drillers help develop adequate vision during operation. The remote control system
the suite of collaring algorithms and the key steps to achieving was installed on drill 21R03 when it was in Perth during the
quality production holes. electro-hydrauliccontrol conversion.
The line of sight communication systems installed by RCT
Selection and conversion of a drill to electro- were the following:
hydraulic control remote control used a licensed 900 MHz radio frequency
The Terex SKSS16 blasthole production drill was selected as band;
the drill for the conversion because it is the most common
automation, vision, audio and data used a standard public
model in the RTIO fleet. The drill weighs approximately
85tonnes and is capable of drilling a 16 m by 229 mm diameter 2.4 GHz 802.11 b/g standard; and
hole in a single pass. It can be configured to operate in rotary independent shutdown used an encrypted public 430 MHz
or DTH mode. After discussions with the Terex it was decided radio frequency.
to completely convert drill 21R03 electro-hydraulic control.
The drill was transported from the West Angelas mine to Safe systems of work
Perth for conversion. The conversion required the removal of RCT designed two remote control drill training modules based
the pilot hydraulic system and replacement with a series of on operation and maintenance of the system. Drillers and
cables, electric valves, solenoids and joysticks. maintenance personnel were required to pass these modules
before using the remote control system. Rio Tinto then
Installation of the remote control system provided one week production training for drillers under the
Remote Control Technologies (RCT) were contracted to supervision of a Rio Tinto trainer.
provide a drill remote control system complete with mobile Significant consideration was given to the safety systems to
operator tele-cabin (shown in Figure 3). The remote control be used with the automated drill. The safety interlocks of the
receiver is designed with an auxiliary input which was later original drill were maintained for redundancy such as pipe
configured to take automation commands. The RCT system in the hole to prevent machine damage. Aside from potential
allowed remote control of all drill functions although not equipment damage to the drill itself, two key additional risks
all functions were later automated. After consultation with were identified from not having an operator on board the drill:
the production drillers, the remote control console was
1. collision of the drill with vehicles or personnel on the drill
constructed in a similar layout to the existing drill console
(shown in Figure 4). RCT also integrated the independent bench, and
2. tramming of the drill over an edge of the drill bench.
Consideration was given to developing collision detection
systems to mitigate collisions on the drill bench, but a risk
assessment with site personnel concluded that the existing
drilling safety protocols were sufficient for remote operation.
The drill benches are isolated through the use of berms
and barricades and restricted access is standard operating
procedure. The only extra requirement was for signage
indicating the drill may be remotely operated.
Ensuring the drill remains within the permitted area on the
bench was achieved through a combination of the navigation
system and video cameras mounted on the drill. The seven
video cameras installed on the drill provided the operator
with a comprehensiveview of the drill and its surrounds. This
included a below deck camera so that the hole collar and the
position of the drill with respect to obstacles and bench edges
FIG 3 -Typical setup of drill tele-cabin with automated drill in background. could be seen.
Tramming to all edge holes was performed under tele-
remote control to allow the operator to inspect the condition
of the edge of the pattern and to ensure that tramming control
near to the edge was maintained.
Commissioning of remote control

The commissioning of the remote control system on-site was
relatively smooth. The ergonomics and quietness of the drill
tele-cabin had a major impact on the acceptanceof the system.
The vision system provided a view of the drill not possible
from the drill cabin. The cameras proved reliable and did not
require constant cleaning.The audio system was discontinued
due to the interference of engine noise.
The drillers progressed very quickly from their first remote
control hole to the first remote control pattern. The line
I of sight communications proved extremely reliable with
FIG 4 - layout of the operators control in the tele-cabin. operational distances exceeding 800 m. The vision system was
C McHUGH eta/
a configured for high fidelity consuming approximately seven algorithm was unable to drill in certain conditions the drillers
megabits per second of bandwidth. The latency of the vision were able to swap to tele-remote control. The automated
system was approximately 20 milliseconds and operators drilling was extremely accurate in drilling depth with accuracy
learnt to allow for the small delay. Drillers experimented with within one centimetre of design depth.
more complex tasks such as rod handling to reduce the time
required on the drill. Field tests demonstrated the instantaneous automated
drilling penetration rate was comparable with expert manual
During the commissioning of the remote control system a drilling however overall drilling rate was slower due to the
third party high precision global positioning system (HPGPS)
repeatable hole cleaning routine. Redrilling dropped from five
was installed on the mine drill fleet. Remote access to the third per cent to less than one per cent with automated drilling.
party system was installed in the tele-cabin which allowed
drillers to drill using the remote control with the HPGPS Commissioningof automated tramming
guidance. The accuracy of the collar position improved
The automated tramming required the installation of an extra
although there was significant latency in the position update. HPGPS system with GLONASS capability and installation of
More than 50000 m of production drilling was achieved track encoders. The automated tramming algorithms were
using line of sight remote control operation. The performance based on systems developed at the Australian Centre for Field
of the drill under remote control was comparable with Robotics. The HPGPS antennas were installed at the top of the
manual production. This confirmed that reliable remote mast to ensure good satellite coverage. During commissioning
control operation could be used as a fall back strategy if the the HPGPS coverage proved extremely reliable with an average
automation functions did not perform satisfactorily. The drill of 12 satellites visible. The drillers were able to download a
was used preferentially in areas were hazardous fibres were pattern to the automation touch screen, select the target hole
present. and confirm the path. It was possible for drillers to plan a path
that is not hole specific. The automated tramming speed was
Commissioning of automated levelling function slightly slower than tele-remote operation due to the update
The automation computer that connected to the remote rate of GPS corrections but was more accurate and required
control receiver was installed on the drill and an automation less effort by the operator. If the drill was aligned with the
touch screen computer installed in the drill tele-cabin. The general direction of holes then the speed and accuracy were
automated levelling system uses a feedback from a tilt sensor improved. The automated tramming function was also setup
installed on the drill. After discussion with Terex on the to tram between patterns. Only single hole tramming was
correct levelling procedure an algorithm was developed. The enabled for production so that the drillers confirmed each
automated levelling algorithm was tested several hundred time the path was clear. The automation tramming reduced
times in conjunction with remote control tramming and the collar location error by almost 50 per cent compared to
drilling. The levelling algorithm had a 98 per cent success rate manual tramming with HPGPS.
with a mean error of 0.2 degree off vertical.
Phase two - Conversionto a non-line of sight
Commissioningof the automated drilling function communication system
The automated drill function was installed on the drill using Line of sight communications were chosen to ensure high
the algorithms developed using the experimental drill at the reliability of the remote operation of the drill. There were
University of Queensland with expert driller input. The drill operational challenges regarding the continual movement
automation used feedback from the drills programable logic of the drill tele-cabin from one pattern to another and
control for pressures and rotation speed. A draw string encoder maintenance of the small onboard generator. After 12 months
was installed on the machine to accurately measure depth of field trials and production it was decided to move all the
and penetration rate. Initially the draw string mechanism communications to one system that allowed non-line of sight
in the encoder proved unreliable lasting only two weeks of wireless communications. A communication system based on
the Motorola Motomesh Solo with the Motorola Canopy@
production before catastrophic failure. After investigation by
broadband backhaul was installed on a series of communication
the manufacturer and modification of the design, the draw
trailers. The mesh works in the 2.4 GHz spectrum and co-
string encoder has become extremely reliable with mean time exists with existing 802.11 b/g wireless network at site. The
between failures of several months. broadband backhaul operates in the 5.4 GHz spectrum. The
The automated drill function strategy was to minimise the mesh network had a maximum throughput of 1.2 megabits
percentage of redrilling whilst keeping within the drill bit per second per access point. The remote control commands
manufacturers pressure and rotation speed specifications. were converted from serial to Ethernet packages and the
Based on expert driller advice and engineering tests the vision system was adjusted so that the maximum throughput
drilling algorithm was customised for the mine ground was approximately 550 kilobits per second. The reduced
conditions taking into account the following: fidelity of the vision did not reduce the productivity of the
remotely operated drill. The drill tele-cabin was moved semi-
pull down and rotation pressures,
permanently to the mine production office and connected to
rotation speed, the mains power. This reduced the maintenance of the drill
collaring, tele-cabin, eliminated time for setup and reduced the noise
water injection, and associated with the on board generator. The purpose built
hole cleaning. communication trailers were located at various points around
It became evident during field trials that the automation the mine and the drill has been operated over 800 m from an
required some operator input. With their prior knowledge access point. The access points have been located over 2000 m
of the ground conditions they could maximise hole stability. from the drill tele-cabin.
The automated drilling function allows drillers a degree of From the experience gained in phase one, the management
flexibility to change the input parameters. If the automation of West Angelas had sufficient confidence in the safety
ENHANCED DRILLING IN RIOTINTO
procedures used to control access into the drill area were both a technical and operational/implementation aspect.
sufficient for non-line of sight operations. The main difference Specific achievements of the project are as follows.
from phase one operating procedures was access to the bench
for functions interacting with the drill itself (eg refuelling Summary of major milestones of the project
and rewatering of the drill). For activities such as these the The development path was a logical progression of technical
drill was trammed tele-remotely off the pattern, secured and achievements towards building a robust drill automation
shutdown. system. The first stage first requires the remote control
functionalityfollowed by the staged introduction of automated
Phase three - Development of a multi-drill functions. Table 1 shows the chronological order of major
system achievements of the drill enhancement project.
After the successful production deployment of the first
generation automated drill a decision was made to build the Comparison of manual and automated accuracy
next generation system capable of remotely operating two and speed of drill functions
automated drills. A purpose built mobile automation control Aseries offield trials were conductedto determine the accuracy
vehicle (ACV) was designed and constructed in Perth with two of the automation system against the baseline data of manual
drill operator stations and a work bench (Figures 5 and 6). drilling and manual drilling using HPGPS guidance. During
the trials it became evident that there was an immediate trade
off between speed and accuracy in the automated levelling and
tramming functions. Through a series of tests the algorithms
were tuned to give the best accuracy within a productive time.
Table 2 shows the accuracy and speed of drill 21R03 functions
against four other manually operated SKSS16 drills during
production trials. Automation increases the levelling and
tramming time by approximately 15 minutes per 12 hour shift.
Production performance of the automated drill

The West Angelas Mine drill fleet consisted of five Terex
SKSS16 drills providing a good comparison of automation
against manual performance. Drills 21R03,21R04 and 21R05
are more than four years old. Drills 21R11 and Drills 21R12
I were delivered in May 2009.
FIG 5 -The automationcontrol vehicle. Table 3 shows the production and maintenance key
performance indicators of Drill 21R03 against four other
manually operated SKSS16 drills with HPGPS guidance. Drill
I 21R03 started single shift automation operation in May 2009
moving to double shift operation in September 2009 after
a major rebuild. When used in the double shift drill 21R03
has drilled approximately 80 per cent of the production in
automated mode. The main factor limiting higher automation
utilisation was the lack of full site coverage of the non-line
of sight communication. Therefore generally patterns were
either completely drilled on automation or manual depending
on the wireless communication coverage.
CONCLUSIONS
The development of automated blasthole drilling for use in
the Pilbara has been very successful. Automated drilling has
been working with various levels of autonomy for almost two
FIG 6 -The multi-drill operator's console in the automationcontrol vehicle. years. The reliability has of the automation system has been
outstanding and the feedback from operators on the use of
The ACV was designed for rapid deployment and complete the system excellent. The automated drill has successfully
mine mobility including an onboard generator to power the demonstrated that it can match or exceed the average
system electronics. Considerable work was undertaken by production performance of a manned drill. In addition
RCT to reduce the vision bandwidth. With the new vision to matching the productivity levels of a manned drill the
on demand system it may be possible to remotely supervise automated drill system has demonstrated numerous other
an automated drill with less than 250 kilobits per second additional benefits which when added together produce a
allowing up to three drills per access point. Considerable value in excess of the result produced by a manned drill.
work has been undertaken to develop an Ethernet based
independent shutdown system. This system will transmit a On average the automated drill:
broadcast message capable of shutting down all drills in case provides a safer operating environment for the driller,
of a remote or automation system failure. produces a more stable hole,
produces a more accurate hole collar position,
RES ULTS drills a more accurate hole angle,
Overallthe drill enhancement project has been very successful. drills a more accurate hole depth,
The concept of automating a drill appears simple enough at is more productive, and
first glance, however there are many hidden complexities from has a longer rotary bit life.
C McHUGH eta/
TABLE 1
Drill enhancement major milestones.
%
j Date I Description 1
I 23/01/2006 I Drill enhancement proiect approved. 1

I 23/05/2007 1 Firstautomated rotarv hole drilled with laboratory ria at the Queensland ExDerimentalmine. I
[06/06/2007 --^_^_---I., ~
First automated rotary hole drilled with laboratory rig through two different rock types
_ _ , ~ _ _ _ _ . ______-________---,-^ _."
at the QueenslandExperimental mine. 1
21/11/2007 21R03 to electro-hydraulic control.
_____l____l__ "~ l
_ll
-_ I__^_----- "."...I"
1 02/07/2008 Remote control commissioninqof drill 21R03 at the West Anqelas mine. I
1 17/07/2008 1 Firstfully tele-remote percussion productionhole. 1

I 06/08/2008 [ First automated levellingon a productiondrill hole. _- -- 1
1 First automated rotary productionhole.
- - ~ _ l l l l I____ _I-
126/08/2008 I
~~ _I--_ _-__ _- __ _ _ - ~ -i
' 13/09/2008 -- - I First automated DTH productionhole. -_____II^-_l.-___l__
- -~ -
I
I
1
I_ I II_ II_
1 14/10/2008 ?First remotely supervised fully automated tram, level, drill, productionhole.
1 23/11/2008 I First remotely supervisedfully automated rotary drill pattern consistinq of 77 holes.
i25/11/2008 I First demonstration of automated trammina between rows of drill holes.
1r----"-----"
24/05/2009 1 Automated productionof drill 21R03 dayshift only.
'I- II ___I
~
Automated productiondrilling of drill 21R03 24 hours seven days per week.

First demonstration of non-line of siqht automated drill operation 2.2 kin away from the mine productionoffice. i
I
/07/03/2010 1 Drill 21R03 achieves 100000 m of automated production drillinq. I
I 19/03/2010 1 Constructionof the automated control vehicle to control multiple automated drills. i
TABLE 2
Comparison of accuracy and speed between manual and automated drilling.
1 Description I Manual drilling Manual drilling with high precision

global positioningsystem
I Drill 21R03 automation
L __ I
TABLE 3
Comparisonof production and maintenance key performance indicators of the automated drill against a fleet of identical production drills at the West Angelas mine.
1
1j Drill Drill
1 I Drill Drill Drill

1, Drill Drill
1
i
1 Description 21R03 21 R03 21R03 21R04 I 21R05 21Rll 21R12
automation ~ manual total manual i manual i manual manual
1 Totalproductionmeters18/05/2009to07/03/2010 i 99834 1 48852 1 148686 207489 163004 1 146156 1 109372
I Total~roductionmeters06/09/2009to07/03/2010 1 87863 1 25099 i 112962 118649 i 96605 1 102454 i 106172
! Total~roductionhours06/09/2009to07/03/2010 i 1588 1 389 I 1977 I 1850 i 1698 I 1773 I 2119
Penetration rate (m/h) 18/05/2009 to 07/03/2010
--_I ----.+--_-----
Average rotary bit life (m) 21/05/2009 to 17/11/2009 750

~
~
~
4 __ ~- - 4 ~ ' - 0 - ~ ~- ~ ~ -
_I
Availability (%) 06/09/2009 to 07/03/2010 l___l_

88 I
Utilisation(%) 06/09/2009 to 07/03/2010 49 ~ 47 48 57
Potential future benefits of drill automation reduce training time to become expert operators,
Some of the potential additional benefits of the automated smoother and more curate levelling,
drill include the following: increased availability of drill,
single operator controls multiple rigs, increase mean time between failures,
flexible operator assignment between operations, reduce maintenance costs,
ENHANCED DRILLING IN RIOTINTO
improve blast outcomes, Work is currently underway with Atlas Copco to integrate
reduced infrastructure costs associated with labour, and their automated drills into the Rio Tinto system. Other
reduced overall labour costs. improvements and future plans includes modifying the
design of the Terex blasthole drill to take advantage of remote
Next steps operation.
Rio Tinto is now evaluating the potential of automated
drill operation with operators supervising and controlling ACKNOWLEDGEMENTS
multiple drill rigs. This requires a re-evaluation of the
workflows associated with the blasthole drilling process and The authorswould liketo acknowledgeRio Tintofor permission
the implementation of a fully integrated control and safety to publish this paper. They would also like to acknowledgethe
system. These will then be incorporated into a standard contributions of several groups who provided the expertise to
operating procedure for the use of automated drills in Rio make the automated drill reality. These groups include Terex
Tinto operations. Wide scale deployment of the automated Mining, Remote Control Technologies Limited, the Australian
drills will then begin. Centre for Field robotics and CRC Mining.
Practical Management of Progressive
Large Failures at Minerals and Metals
Group Century Mine
B D Hendersonhall, D Lucas2, N Kerr3and C Pennisi4
ABSTRACT
Minerals and Metals Group (MMG) Century mine (Century) is a large open-cut mine in northwest
Queensland approximately 2.5 km long, 1.5km wide and 240 m deep, with an ultimate pit depth of
320 m. In 2009,the mine experiencedthree failures which began as small failures, and progressively
developed into significant multi-batter failures. The incidents were managed by a combination of
analysis, redesign and practical management to achieve positive outcomes for the mine.
On the south wall of the mine, a series of batter-scale wedge failures occurred in December
2007.Analysis of structure orientations and design changes aimed to minimise the risk of failure
extending over multiple batters, while accepting a risk of further small-scale failures. Design
changes were constrained by a culturally sensitive area (CSA) located south of the wall. As mining
progressed, shallow wedge failures developed over multiple batters and cracks were observed 60m
back from the pit crest, indicating that a large-scale wedge failure was developing. Movement
increased in January 2009 as a result of significant rainfall. After considering several design
options, the area was stabilised by a cut-back of the wall.
Two multi-batterfailures alsooccurred on the north wall as mining progressed through a structurally
complex area. Both north wall failures began as small failures that progressively developed, and
required a redesign of the wall below as well as practical management of safety aspects. The complex
structural geology was difficult to predict ahead of mining, and required geotechnical input into
design changes as mining progressed.
All three failures potentially had significant impact for ore recovery, and were managed by a
combination of analysis, redesign and operational management have resulted to achieve positive
outcomes in a challenging mining environment.
INTRODUCTION
Century mine is a lead-zinc-silver open pit mine owned and
operated by MMG (Minerals and Metals Group). Century is
the worlds second largest open pit zinc mine producing some
500000 tonnes of zinc annually. Century Mine is situated
at Lawn Hill in the remote lower gulf region of northwest
Queensland, Australia, as shown in Figure 1.
Description of the mine

Century mine is a truck, excavator and shovel mine. The
ultimate pit is approximately 2.5 km long by 1.5 km wide at
the surface with a final depth of approximately320 m. The pit
is mined in a series of stages with older stages being filled with
waste material via staged in-pit dump system and is currently
mining stages 7 and 8, the penultimate cut. Drilling and
blasting is performed from 12 m and 24 m benches with blast I
drill holes of 165 mm and 251 mm diameters. Blasted material FIG 1 - location of Centurys operations: the mine at lawn Hill and the
is excavated in 12 m benches or in smaller increments (3 m, associated dewateringand ship-loading facilities at Karumba.
1. MAuslMM, Senior GeotechnicalEngineer, 4/84 Mitchell Street, North Ward Qld 4810. Email: bdhendersonhall@optusnet.com.au
2. MAuslMM, PrincipalGeotechnicalEngineer, Mining One Pty Ltd, level 2,455 Bourke Street, Melbourne Vic 3000. Email dlucas@miningone.com.au
3. GeotechnicalEngineer, Mining One Pty Ltd, level 2,455 BourkeStreet, MelbourneVic 3000. Email: nkerr@miningone.com.au
4. GeotechnicalEngineer, Mineralsand Metals Group Century, PO Box 8016, Garbutt BC Qld 4814. Email: Chris.Pennisi@mmgroupltd.com
B D HENDERSONHALLet a/
4 m or 6 m flitchs). Blastholes of 165 mm diameter are used maintain ore continuity. Mine plans and pit designs were
for ore blasts, presplitting and limit wall blasting. adapted to conditions as mining progressed. Geotechnical
The orebody consists of stratabound lead-zinc-silver in engineers coordinated remediation efforts, and produced
bedded shales overlain by Cambrian limestone and sandstone final wall designs essentially bench by bench in response
to the progression of failures and exposure of geology that
cover. Pit wall designs match the overarching geology and
would maximise ore recovery with adequate minimisation of
bedding dip and dip direction. As a result of the widely spaced
geotechnical risk.
exploration drill holes, which are in places spaced as much as
250 m, the complex structural geology of the north wall was This paper outlines the geotechnical considerations and
poorly understood. technical aspects of the failures as well as the monitoring
regime giving the technical background to the failures.
Failure description and management The primary focus is the chronology of the failures and the
practical management of the failures as they progressed.
Three major wall failures occurred progressively from
late 2007 to early 2009 as detailed in Tables 2, 3 and 4.
Figure 2 shows the location of the failures in the southwest GEOLOGY AND GEOTECHNICAL
wall, north wall and northwest wall. All had significant impacts CON5 IDERAT10NS
on future mine designs and potentially for ore recovery. Century mine is a stratiform orebody, hosted mainly in shales
Centurys geotechnical engineers play an important role with minor sandstone. The strata generally dip gently toward
in geotechnical hazard management by putting in place the west and steepen in the west of the mine to dip steeply
management plans and systems for in pit activities that east, with significant localised variability. The ore-hosting
ensured geotechnical issues were anticipated and managed. strata are unconformably overlain by limestone, which forms
Century was able to respond quickly to the failures to a capping to the entire area.
FIG 2 -Century mine stage 7 approved pit shell as of October 2009 showing the three failure areas; southwest wall (SW), northwest wall (NW) and north wall (N).
PRACTICAL MANAGEMENT OF PROGRESSIVE LARGE FAILURES AT MINERALS AND METALS GROUP CENTURY MINE
The mine is transected by several major faults including compressive strength (UCS) values. The lithologies as listed in
Pandoras Fault to the South, and Nikkis Fault to the north Table 1 are intersected by the stage 7 approved pit shell and
which define the margins of the orebody (see Figure 3, plan, shown in Figure and ~i~~~~4.
and Figure 4, cross-section).
Structural geology
Lithology The Century pit is dominated by three major structures. These
Lithologiesrelevant to the stage 7 and stage 8 pit designs in the are Pandoras Fault in the south, Nikkis Fault in the north,
three failure areas are detailed in Table 1 with their unconfined and an unconformity which is flat lying on a more regional
Black Carbonaceous Shale (BCS) Hangingwall Sandstone (HWS)
Black Laminated Shale (BLS) Cambrian Limestone (CLS)
Upper Footwall Shale-Siltstone (UFW) n Soils and Sands (extremely weathered CLS)
Hanging wall Siltstone-Shale (HWD) Clayey Gravels
FIG 3 -Century mine geology (plan view) displayed on the stage 7 and 8 approved pit shell as of October 2009. Geologicalcontacts are based on the 2008 reserve
geology surfaces as well as field observations and as such are indicativerather than exact boundaries. Thick lines show Pandoras Fault in the south, Nikkis Fault in the
north and the cross-section line for the cross-sectionshown in Figure 4.
TABLE 1
Lithology codes used in this paper and their unconfined compressive strength values. Units are listed stratigraphically from oldest to youngest. Not all Century units are
listed, only those relevant to this paper. Unconfinedcompressive strength values have been taken from Cadman (2005) and Lucas (2007).
I Unit I Abbreviation I UCS (MPa) I

I Black laminated shale I BLS I 30 I
I Black carbonaceousshale I BCS r 30 I
I Umer footwall shale-siltstone I UFW I 50 I
1 Hanainawall siltstone-shale I HWD I 50 I
Hangingwall sandstone HW5 75
Cambrian limestone CLS 75
B D HENDERSONHALLeta/
Black Carbonaceous Shale (BCS) Hangingwall Sandstone (HWS)
Black Laminated Shale (BLS) Cambrian Limestone (CLS)
Upper Footwall Shale-Siltstone (UFW) Soils and Sands (extremely weathered CLS)
Hanging wall Siltstone-Shale (HWD) Clayey Gravels
FIG 4 - Century mine geology northsouth cross-sectionthrough the stage 7 and 8 approved pit shell as of October 2009. Geologicalcontacts are based on the 2008
reserve geology surface as well as field observations and as such are indicative rather than exact boundaries.Thick lines show Nikkis Fault in the north, Pandoras Fault
in the south and the relatively flat lying unconformity.
TABLE 2
Chronology of failures and managementfrom November 2007 to, and including, June 2008.
1 Month 1
Southwest wall
1
I November 2007 1 - Batter scale failures 1104-1120 RL initiate redesian of SW wall. i

December 2007 1 - 17/12/2007- Redesignof the SW wall finalised and revisedstage 7 design approved
1 27/12/2007 Wedae failure below culturally sensitive area 1104-1120 RL.
-
January 2008 4-. Batter

----scale wedge failures as mining progresses. __ -_
- Batter scale~-wedge failures as mining progresses.
~ I
- I ~
3
I_____ ----t*
February 2008
March 2008
I
Batter scale wedge failures as mining progresses.

I - - - ~ - - ~ -- _ ___I-_ -- - -- I -- -*
1
p- ___ I I_-xxII- -_I I - - - -
15/04/2008 - Signs of failure - trickling and dusting, small failures (1080 - 1104 RL).
1
a
16/04/2008 - Main failure -wedge from 1080 - 1104 RL. Fort)-eight hour exclusion put in place, regular visual inspection,JSA for clean-up and
adjustmentsto dnlling to remove pre-split from failure area. Identification of a potential macro wedge daylighting between 1044- 1056 RL.
j -17/04/2008 - Action plan for remediation developed. 1
I
I ;17/04/2008 - Remediation undertaken ----from 17/04/2008.
_--Ix_--_ --- - - ___-
/ May 2008
__I _I -I I I -x _ I
I 10/05/2008 -Wedge failure 1080 - 1092 RL at 5.30 pm after blasting below at 11.OO am.
I 10/05/2008 - Remediation commences. I
a
____
- Remediation- - - - ^ I I _ ~ ~ _ _ _ _ _
--_I__ ^ - 4
-
progresses and mining completed down to
-- -- -- 1056 RL. _I ~ ~ I -- I - --~-
scale with local variation (refer Figure 5). The units below the Pandoras fault
unconformity are sedimentary (predominantly shales) and are Pandoras Fault trends eastwest and turns to trend northwest
dominated by bedding, which dips at variable orientations but along the SW wall. It dips to the north, displaces the ore
generally into the pit. Other structures of note are Rayners (north side down) and exposes the older BCS and BLS on the
Fault which trends northsouth across the centre of the pit and south side of the fault. Pandoras Fault is described as being
a series of eastwest trending structures which appear to be a scissor fault with the displacement of strata varying along
associated with the overall trend of Nikkis Fault. These are strike of the fault. The fault is associated with some smaller
also shown in Figure 5. structures which are seen mainly in the south wall of stage 6.
The structures in the BCS, which combined with bedding led
There are several un-named structures north of Nikkis Fault to the SW wall progressive failure, could be a conjugate set of
that could not be accurately identified, but that impacted on discontinuities associated with Pandoras Fault.
the N wall failure and the subsequent redesign. Disparities
in the presentation in pit of the unconformity in comparison Nikkis fault
with its interpretation are most likely due to these and other Nikkis Fault is nearly vertical and strikes approximately
un-named structures. Reinvestigation of the drill hole data eastwest across the north of the pit. While Nikkis Fault does
could not add any further insight into the nature of the not have a primary role in the N W and N failure mechanisms,
lithology or structural characteristics of the N wall. it is of particular interest as it truncates the ore in the north of
TABLE 3
Chronologyof failures and management from July 2008 to, and including, February 2009.
Month Southwest wall North wall

-
luly 2008 14/07/2008 - Multi-batter wedge failure 02.00 am 1056 - 1128 RL 01/07/2008 -Two hour, 20 m blanket stand-off from north wall
(72 m). Failurewindrowed off, signs of groundwater on lower bench. after blasting is put in place after frequent reports of minor rock
1 16/07/2008 -Seepage observed at the 1056 - 1068 batter in falls after blasting by pit personnel.
absence of recent rainfall. Numerous single batter wedge failures on 08/07/2008 - JSA for drilling and loading of the 1004 - 785 ps, first
1056 - 1068 RL. Action plan developed for slip remediation. of what becomes routine JSA for any activities involving personnel
19/07/2008 - Prism 23 - 921 alarms and 1137 RL was inspected and on foot within 20 m and or demarcated zone on the north wall.
cracks observed.
1 21/07/2008 - 1032 - 758, Blast is fired - after blast exclusion put in
place along full length of SW wall.
25/07/2008 - Remediation commences.
4ugust 2008 14/08/2008 - Remediation assessed as complete.
23/08/2008 - Major blast below the area, Prism movement shows
acceleration subsequentto this blast.
ieptember 2008 1 01/09/2008 - 1137 RL further cracking observed. 22/09/2009 - Communicationof north wall stability and
B 13/09/2008 - Further prism installation operational restrictions likely to be ongoing and worsen to
B 26-28/09/2008 -Cracks picked up by survey, crackmeters installed key stakeholdersas a result of unconformityand Nikki's fault
and further prisms installed. interactions.
Ictober 2008 1 07/10/2008 - 1020 - 1032 RL planadwedgefailure on BLSIBCS October geotechnical review by Mining One Pty Ltd - management
contact. of north wall issues with blanket jsa's and after blast exclusions
v 24/10/2008 - 1008 RL Dotential failure Dlane identified. endorsed.
lovember 2008 8 09/11/2008 - Exclusion zones below failure established, restricted Parallel jointing and potential for whole batter slab style failures
zone and JSA put in placefor restricted access to noted.
1137RL.
t 10/11/2008 - Extensometerand alarms in place.
t 29/11/2008 - 1008 - 1020 RL single batter wedge failure. I_ .-
letember 2008 I 04/12/2008 - Southwest hazard plan issued. a 14-21/12/2008 - 1008 - 1020 RL, Exposure of the unconformity
in the face. Moisture in the face and unconformitysurface dipping
into the pit. Initial thoughts that the unconformitycould be a
faulted contact.
22/12/2008 - 1008 - 1020 RL, Wedge failure in the shale created a
small undercut in the face.
22-29/12/2008 - Open cracks in the CLS between 1020 and 1044
RL and above lead to emplacement of exclusion zones along
sections of north wall as rainfall intensifies.
anuary 2009 I 01/01/2009 - 1137 RL, extensometer alarm goes off approx 02.00 02/01/2009 - 1020 - 1044 RL, Slab style failure onto pattern
am, signs failure is accelerating. contained by geotechnical demarcated exclusion zone.
I 02/01/2009 - 1032 - 1132 RL, Major multi-batter failure of SW wall
following 20 mm in one hour rain event, face bulge observed over
- 11/01/2009 - 996 - 1008 RL, CLS crest fails above the exposed
unconformitydamaging excavator 283, other joints and cracks
almost entire length of SW wall prior to failure. opening up in the CLS.
' 07/01/2009 -Southwest failure experiences further major failure. 13/01/2009 - 1020 - 1044 RL, Remediation of 1020 - 1044 failure
20/01/2009 - Dave Lucas site visit to assess options going forward. commences.
.
I
I 21/01/2009 - Commencement of design work for a fill buttress to 30/01/2009 - 996 - 1008 RL, As more of the unconformityis
prevent further failure of the SW wall. exposed with BCS presenting below the CLS further deteriorationof
' 27/01/2009 - Lucas, 2009a report received. the CLS occurs. Exclusion zone in dace for Dersonnel on foot.
:ebruary 2009 ' 12/02/2009 - Formal presentation on options going forward by 01/02/2009 - 1008 RL, Cracks observed in the berm from look-out.
century geotechnical personnel -cutback, fill buttress, rock buttress. 02/02/2009 - 996 - 1008 RL, Remediationof overhanging CLS on
Ongoing work defining potential ore loss associated with different 1008 crest commencesand ramp built up to 1008 RL.
options going forward and design concepts. 14/02/2009 - 996 - 1008 RL, Cracked berm blasted (996 - 769
Minimum required fill buttress determined and then design blast).
produced by planning engineers (minimum mining widths, ramp
grades).
- 17/02/2009 - 996 - 1008 RL, Remediation of blasted 996 - 769.
21/02/2009 - 996 - 1008 RL, Ramp constructed up to 1008 RL,
more cracks identified in the 1008 berm.
28/02/2009 - 996 -1008 RL, Second phase of blasting on the 1008
RL berm.
the pit. The area north of Nikki's Fault has limited exploration by the fault. Structures in the CLS above the unconformity
drill hole data, with holes spaced up to 250 m apart. Structures strike nearly parallel to the fault, and are therefore nearly
in the hanging wall shale (HWD) below the unconformity parallel to the N wall. Figure 5 shows both the 2008 reserve
and north of Nikki's Fault appear to have been deformed interpretation of Nikki's Fault and the revised interpretation
Next Page
6 D HENDERSONHALLeta/
from 2009 mapping and reinterpretation. Nikki's Fault is to both the NW and N wall failures. Initial data used for pit
perhaps better interpreted as a structural zone for the purpose design predicted that the unconformity was relatively flat
of pit wall stability assessment. lying and typically dipping into the pit wall. The unconformity
was also interpreted as continuous along the north wall and
Unconformity uninterrupted by nokhsouth striking structures. Lithologies
The contact between the CLS and shales (HWD and BCS) is an below the unconformity were anticipated to be similar to
unconformity which is the dominant structure contributing those exposed on the north wall of stage 5, an earlier mine
TABLE 4
Chronologyof failures and management from February 2009 to, and including, October 2009.
lonth Southwest wal North wall Northwest wall

larch 2009 Fill buttress 14/03/2009 - Remediation of 996 - 1008 RL completed and t 11/03/2009 - Seepage observed on the 996 RL along
construction ramp mined off. the unconformity.
commences 30/03/2009 - 996 RL, Western edge of unconformity p 14/03/2009 - 984 RL, Structure exposed that defines
exposed with BCS underneath and significant clay filled western edge of the impendingfailure.
structure above sub vertical and near parallelto face. Grey b 15/03/2009 - 984 RL, Wet, weak, structurallydeformed
CLS defined as western limit of instability on the 1020 RL. shales exposed dipping steeply into the pit from
underneath the unconformity.
ipril 2009 12/04/2009 - Heavy rainfall. B 03/04/2009 - 972 RL, Batter scale unravelling
13/04/2009 - 984 - 1004 RL, Small fretting observed in the and superlicial shedding of shales from below the
BCS below the unconformitysoon after material mined off in unconformityon structurally relatedjointing.
front of it. Sericite alteration with weakly bonded bedding plains.
16/04/2009 (day shift) - 984- 1008 RL, Failure in haematitic Beddingdipping steeply into the pit
1 Significant flows from drain holes and through wall
shales. Geotechnicalexclusionzone implemented.
observed.
16/04/2009 (night shift) - Failureof CLS above haematitic
shales resulting in further failures to the west in the grey
limestone which is bounded by the black shale and the fault
in the CLS observedearlier in the month forming a detached
CLS slab approx 36 m high.
18/04/2009 - CLS slab rotates off the high wall after a
nearby shot from 1044 RL to the base of the unconformity.
19/04/2009 - 984 RL bund wall establishedto physically
delineate the geotechnical exclusion zone.
20/04/2009 - CLS slab further rotatesand breaks into two.
nay 2009 10/05/2009 - 1020 RL and 1044 RL, Cracks observed in berm 26/05/2009 - 972 RL, Drain hole drilled to eastern
to the west of and above the detached slabs extent shows extremely high flow rates.
11/05/2009 - Commencement of fill ramp to access the base
of the unstable CLS slabs from the east.
22/05/2009 -Scaling of loose material to the east of the
slab. Platform created for D&B to access the eastern side of
theslab on 1020 RL.
une 2009 07/06/2009 - Successful blasting of the far eastern slab. . 03/06/2009 - 960 - 984 RL, Batter scale failure on
08/06/2009 -Successful blasting of remaining slab. bedding which was previously undercut bedding during
' 15/06/2009 - Ramp being constructedacross the 1020 RL wall scaling. This failure defines the eastern extent of
to access the western berm to enable further drilling and the main NW wall failure.
blasting of remaining unstable grey limestone section. . 06/06/2009 - 960 - 972 RL, Mined to design.
10/06/2009 - 960 - 972 RL, Batter scale failure.
11/06/2009 - 960 - 972 RL, Further batter scale failures
and cracking observedabove 972 RL following blasting
of 948 - 759 trim.
- 12/06/2009 - 948 - 984 RL, Significant seepage
observed across face.
3/06/2009 - 960 - 984 RL, Further batter scale failures
in the location of where the main NW wall failure wouI(
occur.
16/06/2009 - 948 - 984 RL, Crack beganto separate from
wall and on day shift 16June 2009 the eastem section
began to breakaway. Exclusion implementedon NS.
- 17/06/2009 - 948 - 984 RL, Frettingand unravelling of
wall continued and 50 m GRT exclusionzone perimeter
was established. 948 RL, Extentsof failure defined by
east and west faults.
22/06/2009 - Safety bund constructedto demarcate
exclusion zone.
Previous Page
PRACTICAL MANAGEMENT O F PROGRESSIVE LARGE FAILURES AT MINERALS A N D METALS GROUP CENTURY MINE
TABLE 4 CONT ...

Chronologyof failures and management from February 2009 to, and including, October 2009.
Month Southwest wall North wall Northwest wall

luly 2009 Southwest report on
stabilisationoptions and
- 06/07/2~9 Grey limestone blasted from the 1020 RL, - Failure continuesto progress showingfurther
-
coming away from structure surface cleanly. undercutting of CLS along the unconformity.
associated design work
finalised. Commitment to
- behind
12/07/2009 Remediation of blasted material and wall
-
commences.
- 21/07/2009 Decisionto redesign and
-
leave failure without attempting further
continue constructing fill
buttress pending decision
- start
13/07/2009 -Signs of instability on the 1020 1044 RL
appear with cracking progressing on the 1044
to
- remediation.
on way forward. RL behind where the previous slab had peeled off. All
personnel removed from work area.
-- 14/07/2009 - 1020 - 1044 RL, Ongoing failure of the CLS.
19/07/2009 - 1020 RL, Clean up of the failed berm.
. 20/07/2009 - Further cracking observed in CLS face,
personnel removed from failure area.
21/07/2009 - Assessment by GRT concludes that further
remediationwould cause further instability and is not
sufficiently safe to continue any remediationefforts.
22/07/2009 - Redesign commences.
31/07/2009 - Redesign is complete accommodating
ability to do geotechnical design as you go.
\ugust 2009 - Fillrecommences
buttress construction 06/08/2009 - North wall and northwest wall workflow
communicated.
- failures.
03/08/2009 - 936 - 948 RL, Batter scale
Various batter scale failures as anticipated in the design 06/08/2009 - North wall and northwest wall
and design methodology. worMow communicated.
Failure continues to progress showing further
undercuttinqof CLS alonq the unconformity.
ieptember 2009 Ongoing mapping and data collection.
Various batter scale failures as anticipated in the design
- undercutting
Failure continues to progressshowingfurther
of along the unconformity.
CLS
and design methodology.
ktober 2009 - 05/10/2009 Design reviewed below 948 RL based on
-
new data and interpretation and allowing for adjustment
Failure continues to progress showing further
undercuttingof CLS along the unconformity.
based on new data as faces are exposed. Various batter 10/10/2009 - 900 - 924 RL, Batter scale
scale failures as anticipated in the design and design failure.
methodology.
. 26/10/2009 - 936 - 948 RL, Batter scale planar failure
along Nikkis fault contact.
cut-back to the east of the area of interest. The unconformity which is a more competent rock mass with distinct laminated
in the areas of the NW and N wall failures proved to be quite bedding. Overlying both the BLS and BCS, in the SW wall, is
different and more variable than first anticipated. extremely weathered Cambrian limestone sediments (CLS)
which are characterised by their soil like properties. These
Bedding sediments comprise the upper 17 m of the SW wall including
the first 12 m bench.
The orebody and surrounding host rocks are sedimentary
(predominantly shales) and bedding is a major structure BCS and BLS were grouped as one unit for the purpose of
below the unconformity. Bedding generally dips gently to the initial geomechanical characterisation, but it has become
evident that the two should be treated as separate lithologies,
west over much of the pit, steepening in the west and dipping because of their differing behaviour in response to water. BCS
steeply east. Bedding orientations are variable, and generally is observed to absorblretain water and does not drain freely
into the pit, in the south and north walls. Figure 6 shows whereas the BLS is able to drain more freely along bedding.
bedding orientations in the SW wall. BLS also tends to be more brittle than the BCS, which is more
In the NW wall north of Nikkis Fault, bedding is deformed ductile.
by intense faulting and folding as can be seen in Figure 7.
Generally dipping into the pit at approximately 50, it is Northwest watt (NW wall)
however quite variable and prediction of bedding orientation The NW wall geology is characterised by weak altered shales
at a batter scale is difficult. below the unconformityoverlainby 124 m of CLS. Shales below
the unconformity are tightly folded and faulted throughout
Geotechnical aspects of the failure areas the area. The amplitude and intensity of the folding increases
near the unconformity, and this has created a very weak rock
Southwest wall (SW wall) mass zone directly beneath the unconformity, which is prone
The SW wall lies within the black carbonaceous shale (BCS) to to unravelling.
the south of the north dipping Pandoras Fault. BCS is typically
a tight and fairly massive rock mass with distinct bedding but North wall (N wall)
indistinct other structures. To the southeast of the area of The N wall failure occurred in CLS above the unconformity.
interest, the BCS meets the black laminated shale (BLS) unit The CLS is in itself a competent rock however structures
B D HENDERSONHALL et ol
FIG 5 - Century mine major structures shown with the stage 7 and 8 approved pit shell as of October 2009. Faults are based the 2008 reserve geology surfaces and on
short-term geology model.
within the CLS are characterised by weak clays and sands. 2009 have distinctly different failure mechanisms. Figures
In the main production areas of the pit this variation in the as referred to throughout the following sections have been
CLS rock mass resulted in large natural cavities and zones of selected to best depict the failure mechanism and the
extremely weak rock mass containing oversize CLS boulders. resultant scale of the failures as at October 2009.
This limestone consisting of weak rock mass and oversize
boulders is at times referred to as near surface limestone Water (groundwater and surface water) was a contributing
(NSL). This presented issues for both drilling and blasting as factor in accelerating the failures which were all initially
well as excavation. In batter faces this resulted in a variation triggered through interaction of the geology and pit geometry.
from well presented faces to extremely lumpy and difficult to
form faces. Southwest wall (SW wall)
Rock mass and structural conditions in the north and The SW wall failure is a complex three-dimensional
northwest wall failure areas were difficult to predict with progressive failure mechanism through the interaction of
certainty ahead of mining though the potential for issues bedding, subvertical structures in varying orientations and
was identified. In particular, the weak and friable nature the nature of the BCS rock mass. The failure is essentially a
of the rock mass beneath the unconformity, the structural
series of wedge failures.
(bedding) orientations, unidentified structural failure
surfaces and the local dip of the unconformity, all contributed Progressive cracking was observed and monitored on the
to the development of the failures and each factor could not 1137 RL over several months, and defined the extents of the
have been foreseen without significant amounts of forward potential wedge (refer Figure 8). It was able to gauge the
investigations such as drilling. progression of the failure by this progressive cracking.
Bedding rotates (refer Figure 6) around a potential hinge
Failure mechanisms point which is the location of the multi-batter failure that
All three large wall failures that progressively occurred at occurred 2 January 2009 (refer Figure 9). It is thought that
the Century mine between November 2007 and November this hinge may be associated with Pandoras Fault and extend
PRACTICAL MANAGEMENT OF PROGRESSIVE LARGE FAILURES AT MINERALS A N D METALS GROUP CENTURY MINE
FIG 6 - Bedding measurementsin the southwest wall area taken in November 2007.
FIG 7 - Bedding deformed by intense folding and faulting betweenthe northwest and north wall failures 924 to 936 RL.
down to meet the fault. The subvertical structures which retention within the unit is believed to have led to opening
define the other wedge forming failure planes are conjugate of discontinuities within the rock mass, creating extended
structures to bedding. failure planes. Considerable face bulging was observed across
The failure is confined to the BCS unit. This can be attributed the entire wall after 219 mm of rain from the 31 December
2008 to 2 January 2009.
to two factors: the change in bedding orientation through
this area, and the differing behaviour of BCS in response to
groundwater and drainage. BCS does not drain freely, and Northwest wall (N W wall)
water was observed seeping from the lower batter prior to the The N W wall failure occurred as a result of relatively minor
first multi-batter failure. This was the first part of the pit which undercutting of weak rock mass, below the unconformity that
exposed a wall completely comprised of BCS without another forms the base of the CLS. Removal of material in front of the
unit in contact with it that would assist drainage. Water area was the key cause, and groundwater had a contributing
FIG 8 - Southwest wall as per the stage 7 design (post-November 2007 review) showing the cracks on the 1137 RL and the potential failure.
FIG 9 -Southwest wall as photographed on 23 February 2009 showing the BLS/BCS contact on the left, the main failure right of centre and on the far right the
Pandoras Fault intersection.
effect (Kerr and Lucas, 2009). The failure progressed from up into the limestone. The unconformity at this location is
several batter scale failures and conditions as shown in observed to be flat lying.
Figure 10 into a multi-batter failure as shown in Figure 11.
Substantial water seepage was observed in the shale batters North wall (N wall)
where the failure initiated and in the area between the NW The N wall failure was a series of progressive failures that
wall and N wall failures (Figure 10). Geotechnical engineers initiated by toppling. Toppling (in the fundamental sense)
had commenced a program of drainhole installation as soon requires two structures: one steeply dipping or near vertical
as seepage was observed in these batters. Some sections were creating high slabs of rock, and the second set dipping
missed due to operational constraints, and many of the holes gently out of the wall to create a basal release surface. Both
collapsed because of the weak and friable nature of the rock conditions were present in the north wall, in the form of near-
mass. The few holes that survived had consistently high flow vertical joints in the CIS, and the unconformity at the base
rates. of the CLS dipping at low and variable angles out of the wall
The NW wall failure has remained confined to the weak (Kerr and Lucas, 2009). While the near-vertical structures
friable rock mass in shales below the unconformity, and has could be seen, they are widely spaced and are only parallel to
not developed into progressive failures (such as toppling) one part of the curved pit wall and toppling was not identified
FIG 10 - Photographofthe northwestwall prior to failure, taken 12 June 2009.

Steps were taken to remediate the initial toppling failure,
shown in Figure 13, but were ceased when it became evident
that as further slabs detached from the wall that the same
steep joints further into the wall were relaxing into the pit and
opening up, causing increasing operational risks to personnel
and equipment. Figure 14 shows the failure as at November
2009.
Monitoring
Monitoring of excavations is a key component of in the
identification of potential stability issues and the ongoing
assessment of areas with known issues. Monitoring of the SW
wall included visual inspections, crack monitoring, and prism
monitoring.
The area south of the southwest wall (the 1137 RL, or
natural surface) includes a culturally sensitive area (CSA),
and Century personnel were initially not permitted to access
this area for visual inspections. Once the cultural issues were
resolved, access to the area was granted to a limited number
of personnel, who were then able to install monitoring
FIG 11- Photograph of the northwest wall failure, taken 13 November 2009. equipment and conduct routine visual inspections.
as a potential mechanism, because the nature of the basal Radar monitoring was considered for all three failures but
release structure (in this case, the unconformity) was not well was decided against. The nature of the failures' progression to
understood. that time suggested they could continue to be managed safely
by other monitoring and management strategies. Given that
The unconformity surface has a low shear strength decision, implementing radar monitoring did not warrant the
(similar to a fault), and the initial failure occurred when the cost and the level of resourcing that would be required.
unconformity was exposed as mining progressed to the west
on the 1008 mining level. Figure 12 shows the N wall prior to Prism monitoring
the main failure and the remediated section on the 1008 RL.
Century mine maintains a comprehensive prism monitoring
Toppling occurred after the relatively small failure events regime consisting initially of two automatic total stations,
in the underlying shales created localised undercutting which was increased to three in mid 2009.The active prism
of the CLS, and allowed movement on the basal structure network consists of 250 - 300 prisms, which are monitored
(the unconformity). The localised undercutting of the CLS and analysed using Softrock's Autoslope and Quikslope
occurred as a result of the variable (and locally unfavourable)
software. Prism monitoring assisted in the interpretation and
orientation of the unconformity. This basal release occurred
in one unfavourable location, which happened to coincide management of the failures, but did not predict them as initial
with the unfavourable orientation of the pit wall relative to movement was identified visually.
the joint direction. Toppling of the steep joints could not (and In the SW wall, there had been limited opportunities to
did not) occur in the upper part of the wall in CLS. install prisms and coverage prior to the failure was relatively
FIG 12 -North wall prior to main failure, photograph taken 26 March 2009.
FIG 13 - North wall after first main failure, photograph taken 18 April 2009, just after Cambrian limestone sediments slab has rotatedoff the wall.
poor. Elevated work platforms were used to install prisms in polypipe crack meters, steel crack pins and wire extensometers
locations that personnel on foot were unable to access. Once connected to a visual and audible alarm that was also connected
prisms were lost through the progression of the failure there to the pit radio system.
was no ability to reinstate prisms to monitor the failures Crack meters proved to have little value after the onset of
except in a few circumstanceson the N wall. failure. The main issue limiting the success of the crack meters
as a monitoring tool was the erosion and displacement of the
Crack monitoring soil in which they were installed.
Once cracks were observed on the 1137 F U above the SW Regular observations and surveying of cracks proved to
wall, they were surveyed and monitored. Methods included be valuable in characterising the progression of the SW wall
PRACTICAL MANAGEMENT OF PROGRESSIVE LARGE FAILURES AT MINERALS A N D METALS GROUP CENTURY MINE
FIG 14 - Photographof the north wall failure, taken 13 November 2009.

failure. Regular inspections were carried out to identify the Ongoing and ad hoc exclusion zones were put in place
extent of cracking. throughout the progression of the failures as required. In most
The wire extensometer system was connected to the pit cases exclusionzones were in place for the majority of the failure
radio system, and an alarm and warning light. It activated on progressions and were not limited to the specific references to
New Years Eve 2008 as a result of the rainfall and subsequent exclusion zones as detailed in Tables 2 through 4.
movement of the SW wall failure. An exclusion zone was also At the same time as the SW wall multi-batter failures began
in place in the pit along the length of this wall as well as the to present in April 2008 (refer Table I), Century geotechnical
restricted area on the 1137 RL above. After initial movement, engineers were also developing the systems and processes
the wire extensometer system was retrieved to prevent its loss, to implement batter-to bedding across the west wall of the
prior to the main failure on 2 January 2009. pit. This process involves design as you go by geotechnical
Crack monitoring was not a viable option for the N W and N engineers; wall designs were created for each 1 2 m bench in
wall failures as access to berms immediately above the failures response to the exposed geology to create a wall where the
was either not possible or not safe. Polypipe crack meters were bedding planes form the wall. This batter to bedding can be
used successfullyfor remediation works on the 1 0 0 8 RL berm seen in the far right of Figure 9 and the left of Figure 11. This
prior to the major failure. is noted to highlight that as well as managing the progression
of the three wall failures detailed in throughout this paper and
Visual inspections in Tables 2 to 4 there were a number of other major projects
being managed by Century geotechnical engineers.
Visual inspections are a fundamental part geotechnical hazard
management. Century geotechnical personnel have a very
strong operational focus and maintained a solid relationship REDESIGN OPTIONS AND FINAL OUTCOMES
with operations and other technical services teams. Effective Given the distinct differences between each of the failures
and efficient communication, in conjunction with other there were then different options for moving forward.
monitoring tools ensured that adequate management strategies
were in place and no personnel or equipment were placed at Southwest wall
risk. Analysis of the SW wall failure and its likely further progression
indicated three potential solutions (Lucas, 2009a):
CHRONOLOGY Buttressing with a fill buttress and redesigning the wall
below to accommodate the fill buttress (ie a rock buttress
The chronology of these failures highlights the progression below). A substantial rock buttress would be required to
of the failures and the associated management strategies. It control stability with respect to Pandoras Fault, which
should be noted that the chronology detailed in Tables 2 to would otherwise have been mined out in the original pit
4 has been summarised to cover key events and decisions design.
to highlight the failures progression and management. The Buttressing with a more sophisticated fill buttress retained
three failures have been compared side by side to enable an by an engineered structure, to mine out Pandoras Fault
appreciation of the progression of the failures as a whole, and and avoid the need for the rock buttress. This option was
some operational issues have been omitted. assessed and ruled out, as being non-economic.
B D HENDERSONHALLet ol
3. A cut-back of the SW wall. While the buttress options essentially meant that remediation of the failure was not an
preserved the integrity of the culturally sensitive area option, because of access difficulties and dangers to personnel.
(CSA), a cut-back would mine out the CSA, an option The redesign options were:
that required negotiation and the approval of the area's
1. A cut-back from surface. This was considered and ruled
traditional owners.
out, as being non-economic.
The cut-back was the chosen option and commenced in
2. Redesign of the wall below the 948 RL (Kerr and Lucas,
December 2009 after successful negotiations.
2009).
fill buttress and rock buttress Redesigning below the 948 RL results in ore sterilisation,
but the amount of ore lost did not prove to be of sufficient
To preserve the CSA, a fill buttress would have been required
economic value to warrant proceeding with the cut-back
in the pit to buttress the failure and prevent its progression.
(option 1 above).
Analysis determined the minimum size requirements of a fill
buttress required above the 1020 RL floor, up to 1104 RL, The wall below the 948 RL was designed with an initial batter
constructed against the southwest wall. To accommodate the at 936 - 948 RL of 65" and below 936 RL batter face angles
footprint of the fill buttress, the SW wall below 1020 RL would of 75". Two single 12 m benches were designed followed by
have to be redesigned and moved into the pit from the original two 24 m benches. All berms are 9 m and all walls below the
stage 8 position. The design below 1020 RL is termed the rock failure would be presplit. Shovels were given a no-dig stand-
buttress (Lucas, 2009b). off to ensure no undercutting of walls would occur. Wall work
conducted by excavators occurs in two or more passes.
The fill buttress height of 1104 RL was determined by
assessing kinematics of potential wedge failures. Structural North Wall
integrity was required at a distance of more than 28 m from the
crest to prevent damage to the CSA, although ground within 28 The N wall had onlytwo options, similar to the NW wall: a cut-
m of the crest would still experience movement and cracking. back from the surface or a redesign below the 984 RL.
The buttress thickness was determined by considering the The safety catch bund was constructed to a height of
SW wall to be in a state of equilibrium (factor of safety (FOS) approximately 10 m as close to the failure as possible to still
approximately equal to LO), and then adding a fill buttress to provide sufficient catch capacity for any further failure of
show the increase in FOS that would be achieved. Analysis the north wall. A 45" batter was then designed for the 972 to
showed that a 25 m wide buttress would give a FOS = 1.39. 984 RL batter. Initially a design comprising 24 m benches was
However, mining constraints required that the fill buttress be proposed to maintain the ability to mine behind Nikki's fault
40 m wide at 1104 RL. and maximise ore recovery. This design concept was assessed
as having too high a geotechnical risk on the basis that the
Based on this design the footprint of the fill buttress would geology behind Nikki's fault was uncertain. An alternative
encroach on Pandora's Fault, and the wall below 1020 RL was proposed comprising 12 m benches, 9 m berms and 85"
(termed the rock buttress) would need to accommodate the presplit batters. This design allowed the flexibility to respond
fault, which would otherwise have been mined out. Stability of bench by bench to the geology as it was exposed.
the rock buttress would be governed by geological structures: As further information was obtained on the attitude of
Pandora's Fault and bedding. Nikki's Fault it was possible to conduct a further redesign
The size and geometry of the rock buttress was determined below 948 RL that took into account the orientation of Nikki's
by analysing several cross-sections throughout the wall, while Fault with the intersection of the footwall. Nikki's fault
also considering 3D influences. A 3D numerical model could terminates the ore on the N wall and so a design that mined
not be carried out in the required timeframe. to Nikki's Fault and the intersection of Nikki's Fault with the
footwall would essentially recover all the ore from the N wall.
CUt-back The section of wall between the N and NW failures proved
The cut-back option aims to redesign the final pit wall above more challenging and it was decided to try and achieve the
the 1020 RL floor, removing the failing and cracked part shallowest wall angle possible while still aiming to achieve the
of the slope, but also removing the CSA. While this option ore toe position. This angle was determined to be 50". Ongoing
mines extra waste, it achieves full stage 8 ore recovery. The batter scale failures in weak shales along this section of wall
orientation of the wall had always been constrained by the led to acceptance of essentially a battered rill slope at 50" with
requirement to preserve the CSA. The cut-back is relatively buttressing catch bunds emplaced on each mined bench.
straightforward from a geotechnical point of view, compared The redesign is still quite aggressive and batter scale failures
to the issues with the buttress option. It effectively mines out were anticipated. These were expected to be superficial and to
the failing wall back to behind the cracked zone, and mines to not lead to larger or deeper-seated failure.
the shape of the failure which is controlled by bedding and a It was also recognised that all ore would need to be
joint set. recovered from the N wall prior to any significant rainfall
The design mines the entire SW wall behind (south of) to keep geotechnical risk to an acceptable level. In the event
Pandora's Fault, and given the existing data available behind of significant rainfall, further operational controls and
the SW wall further major structures within the BCS with the restrictions would need to be implemented.
exception of bedding are not anticipated. The redesign also had to take into account changing geology
and geotechnical issues on exposure, due to the structurally
Northwest wall (NW) complex nature of the area that could not be interpreted by
Redesign options for the NW wall were relatively other means in the timeframe required.
straightforward. The nature of the failure undercutting the
unconformity and subsequently a column of approximately CONCLUSlONS
124 m of CLS meant that the failure needed to remain Century mine personnel successfully applied practical
buttressed in order to maintain the integrity of the CLS. This management strategies to geotechnical hazard and risk
management. Three large progressive wall failures each had The case studies highlight the issues that can arise in a
significant implications for safety and ore recovery, and each structurally complex mining environment where geotechnical
had different and complex failure mechanisms. They were analysis applications are limited and practical management is
successfully managed by the application of practical controls, the key to a successful outcome.
including exclusion zones and controlled working areas,
monitoring for ongoing assessment and redesigns of each pit ACKNOWLEDGEMENTS
wall. The authors wish to acknowledge Century mine for approval
The SW wall had issues with the bedding orientation causing to publish this paper.
batter-scale wedge failures, and the potential for ongoing The authors also wish to acknowledgethe Technical Services
failures was identified although the shape of the wall was and Mining Operationsteams at Centurymine who assisted the
constrained by a culturally sensitive area (CSA) and initially authors of this paper to manage the challenging geotechnical
could not be redesigned to a geotechnically optimum design. issues detailed in this paper as well as several others along the
Small batter failures eventually progressed to large deep- way in a very difficult economic environment with exemplary
safety results and successful business outcomes.
seated cracking, which worsened after significant rainfall in
early 2009. Redesign options included buttressing to stabilise REFERENCES
the wall and preserve the CSA, but eventually the preferred
Cadman, J, 2005. Materials Database Review for Zinifex Century
option, a cut-back, was chosen after successful negotiations. Mine May 2005, BFP Consultants Pty Ltd, internal report,job no
The N and NW wall failures occurred in a structurally 1803130.
complex area which has limited exploration drilling data, K e n , Nand Lucas, D, 2009. Century Mine North and North West wall
due to the wall being north of a fault that truncates the Assessment, Mining One Pty Ltd, internal report, 134OG-2012.
orebody. Localised variability of an unconformity at the Lucas, D, 2007. Century West wall 2007 Drilling Investigations
base of limestone and presence of unpredicted structures Data Report, Coffey Mining Pty Ltd, internal report,
made prediction of failure conditions difficult. Each of these MINENHILooo41AA REog-VooA.
failures commenced as small failures at the unconformity, Lucas, D, noo9a. Century Mine - Southwest wall Geotechnical
and developed progressively into larger failures. The limited Assessment Jan 2009, Mining One Pty Ltd, internal report,
ability to interpret or predict complex structural conditions 1235G-1761.
below the unconformity added to the difficulty of designing Lucas, D, 2009b. Century Mine South West wall Geotechnical
a solution, and required that the wall design be modified as Review of Stabilisation Options, Mining One Pty Ltd, internal
mining progressed deeper. report, 1235G-1894.
SEVENTH LARGE OPEN PIT MINING CONFERENCE / PERTH, WA, 27 - 28 JULY 2010
The Development of a Novel Eye
Tracking Based Remote Camera
Controrfor Mining Teleoperation
D Zhul,T Gedeon2and KTaylor3
ABSTRACT
The use of teleoperation in mining is increasing due to requirements to improve safety and
reduce the number of people required to work in remote and difficult environments. We recently
completed a project to develop a teleoperated control system for a rock breaker at an iron ore
mine in the Pilbara region of Western Australia, with the objective of demonstrating the feasibility
of remote rock breaking over long distances. This required controlling both the rock breaker
and an array of pan-tilt-zoom (FTZ) cameras using joysticks. Controlling the cameras distracted
operators from the primary task of controlling the rock breaker to position and fire the hammer.
This lengthened task times as both the operators attention and their hands have to swap between
different control interfaces.
In this paper, we present a novel design using human eye gaze as an interactive input for the
remote camera control. It follows a simple and natural design principle: Whatever you look a t the
screen, it goes to the centre. A prototype system has been implemented by integrating computer
vision based eye tracking technology, with the advantage of being real-time, robust and marker-
less.
We conducted a user evaluation, where users undertook a control task while obscuring the
direct view of the working area and using a remote camera to transfer the live video stream as the
visual feedback. Subjects were required to participate in the experiment using eye tracking camera
control and traditional joystick control respectively. The statistical analysis of objective measures
indicated that our novel eye tracking control significantly outperformed the joystick control. The
results of the post-experimental subjective measures also revealed much higher user preference
for using eye tracking control, which provides clear evidence that this design is a better interface
for remote camera control in mining teleoperation settings.
INTRODUCTION
Safety is an important issue in the mining industry. In and rescue activities, etc. Due to the improving broadband
order to prevent people from working under hazardous networks, users are able to be situated thousands of kilometres
conditions or in difficult environments, teleoperation has away to accomplish control tasks, allowing them to work under
been being increasingly adopted as an effective solution in a more convenient condition or to use their skills at many
mining industries. Compared to line-of-sight control on mine different sites around the world. Particular examples like:
sites, such a solution allows people to remotely control or web-based robot control (Taylor, Dalton and Trevelyan, 1999),
manipulate complex mining machinery over long distances, internet-based remote control for mining machinery (Kizil and
with the merits of providing replaceable surrogates, reducing Hancock, 2008),remote surgery and simulated collaborative
number of staff and being cost-effective. Although there has surgical training using haptic devices (Gunn et al, 2005),etc.
been a continuous long-term effort spent on researching and Within the teleoperation model, the user interface is a
developing autonomous or semi-autonomous systems for fundamental and essential component that affects teleoperation
a variety of mining tasks, human observation, intervention, performance significantly. Observations indicate that directly
and supervision are still integrally involved in these systems controlling a robot while watching a video feed from the
(Hughes and Lewis, 2005). remote camera(s) remains the most common interaction form
in teleoperation (Fong and Thorpe, 2001). Therefore, the
The benefits of using teleoperation are already apparent in basic perceptual link between the user end and the remote
numerous areas, ranging from space exploration, inspection, environment is most often through a live video stream from a
robotic navigation, surveillance, underwater operations remote camera or a set of cameras.
1. PhD Candidate, CSlRO ICT Centre and School of Computer Science, CSand IT Building, Australian National University, North Road, Acton ACTO200. Email: dingyun.zhu@csiro.au
2. Professor,School of Computer Science, CS and IT Building,Australian National University, North Road, Acton ACT 0200. Email: tom.gedeon@anu.edu.au
3. ResearchScientist, CSlRO ICT Centre, CS and IT Building, AustralianNational University, North Road, Acton ACT 0200. Email: ken.taylor@csiro.au
DZHU,TGEDEON AND KTAYLOR
The requirements for user interfaces for teleoperation PROJECT BACKGROUNDAND RELATED WORK
of mining vehicles and systems has been briefly discussed
in (Hainsworth, 2001), with the demonstrations of two Remote rock breaking in mining teleoperation
teleoperated mining systems. In this paper, it clearly indicates The telerobotic rock breaker (Duff et al, 2009) was a recent
that conventional user interfaces such as joysticks, switches, project for CSIRO to develop a remote control system for a
wheels, mice and keyboards are still the major control rock breaker at an iron ore mine in the Pilbara region of
elements used in mining teleoperation. They are relatively Western Australia.
simple, allowing teleoperation to be a viable and profitable The rock breaker on the mine site is a serial link manipulator
technique, which satisfies the basic client requirements for
arm with a large hydraulichammer at the tip to break oversized
mining systems of robustness and reliability. However, in most
rocks. The arm is installed at a run-of-mine (ROM) bin, where
teleoperation settings, the operator often has multiple devices
to control simultaneously, for example, controlling a robot a number of horizontal bars (referred to as a grizzly) are
and a remote camera at the same time. Using conventional fitted at the bottom in order to prevent oversized rocks from
interfaceswill lead to frequent switching of attention and hands entering the crusher below. Figure 1 shows (a) the overview of
between different control tasks and interfaces. This distracts the rock breaker; (b) and the grizzly in the ROM bin.
the operator from concentrating on the primary control task, On the remote mine site, a number of haul trucks with ore
reduces productivity, and increases both workload and the from a nearby quarry will be delivering their load into the
number of avoidable operational mistakes. bin (see Figure 2). The operator is required to break those
oversized rocks stuck on the grizzly by operating a two-
This paper addresses this attention and hands switch
handed joystick controller. The operator has limited time to
problem in a situation where an operator is controlling one break the rocks, as trucks arrive at short intervals (about 90
or more remote cameras while carrying out other mining seconds). Since dumping a load raises a large could of dust,
teleoperation tasks. Instead of using conventional control a water spray is used to settle the dust, which requires about
interfaces to overload the operators control ability, we present 30 seconds before the operator has a clear view of the bin.
a novel design that uses human eye gaze as an alternative Then, the operator only has about 60 secondsto move the arm
input for the remote camera control using computer vision from its rest position, place it carefully onto a rock, break it by
based eye-tracking technology. With the user evaluation of a firing the jackhammer, and return the arm to the rest before
modelled experiment for an implemented prototype system, next truck arrives. The actual remote rock breaking process
experimental data was gathered through both objective is shown in Figure 3, which allows the operator to have a
(performance) measures and subjective (user preference) desktop based teleoperation environment and live videos as
measures. the visual feedback.
FIG 2 - Haul truck dumping a load into the run-of-mine bin.
THE DEVELOPMENTOF A NOVEL EYE TRACKING BASED REMOTE CAMERA CONTROL FOR MINING TELEOPERATION
Joystick Controller
for Rock Breaker
FIG 3 - Multi-control situation in remote rock breaking process.

When the operator is trying to break a rock, it is necessary is called head flicking control which is mostly like a switch.
for them to have a close view of the target so that detailed It uses a users quick head movements to control the camera
information can be obtained to specify the spot on the target motion (eg if the user quickly rotates the head to either left
for positioning the tip and firing the hammer. It is not practical or right direction then moves back to the original position,
to mount the remote camera on the arm to couple the camera which appropriately turns on the camera to start panning
motion to the control of the remote robot, as the camera would along this direction and flick to the opposite direction to
be damaged when the hammer is fired to break a rock. It is also make the camera stop). In general these did not improve
a confusing viewpoint for gross arm movements. The remote significantlyon joystick control. We further extended the work
camera is installed on the side of the bin with a zoomed-in view by importing more natural interactive forms into the control
transferring the live video back to the operator. Tip tracking interface design for teleoperation, a few user studies have also
can be engaged (Duff et al, 2009) as a method for freeing the been conducted with the results reported in Zhu, Gedeon and
Mendis (2010).
operator from camera control, whereby the remote calibrated
camera follows the tip of the hammer as its position in space
is known from sensors measuring the joint angles. This works
Eye tracking in user interface
well in the part of the cycle where the operator is interacting Eye tracking has been used as an augmented input medium or
with a rock but does not work well during gross movements to control modality in user interface previously. It has a number
and from the rest position and where the operator is planning of compelling advantages, one being that eye movement can
their next move as the area of interest can be far from the tip. be a background task for humans, which does not require
In these cases, tip tracking is disengaged and the operator conscious attention. Zhai, Morimoto and Ihde (1999) suggest
has to use another joystick controller to control the camera several other advantages:
motion for adjusting the view or observe locations they it can be an effective solution for situations that prohibit
intend to move to. This turns out to be a typical multi-control the use of the hands, for example, when the users hands
problem that requires operators to switch hands and attention are disabled (quadriplegic) or continuously occupied with
quite often between different control interfaces and tasks. other tasks (such as the hands-busy problem in the rock
breaking task);
Remote camera control in teleoperation increasing the speed of user input, as the eye can move
Apart from those conventional remote camera control more quickly in comparison to other input mediums; and
methods, several alternative approaches or designs have been reducing workload, repetitive stress, fatigue (nearly
developed or discussed. For example, (Cohen, Conway and fatigue-free interaction (Saito, 1992) and potential injury
Kiditscheck, 1996) proposed the possibility of using a set of caused by physically operating other devices.
circular oscillatoryhand gestures to control a remote cameras Numerous approaches,techniques, applicationsand systems
pan and tilt motion. In addition, another pan-tilt-zoom (PTZ) using eye tracking have been proposed and developed, ranging
camera control system using a Wii remote and a set of infrared from traditional pointing, virtual environment interaction,
sensors has also been described in (Goh et al, 2008), as the real-world device control, etc. Back to the late i99os, (Yanco,
wide popularity of the Nintendo Wii and its advantages of 1998) introduced a prototype robotic wheelchair system with
being low-cost and easy to use. These approaches are able an eye tracking control interface. Recently, (Tall et al, 2009)
to provide more interactive or natural ways for the remote constructed another experimental robotic vehicle which
camera control, but they still require users to pay attention could be remotely driven by a gaze-controlled interface.
and use their hands to operate. Furthermore, eye tracking has also been used as an alternative
type of user intention for virtual gaming (Gedeon, Zhu and
For the early stage of this project, we explored the possibility Mendis, 2008), leading a group of robots to accomplish
of using head tracking as another alternative input for the cooperativetasks (Zhu, Gedeon and Mendis, 2010).
remote camera control in order to solve the hands-busy
problem in teleoperation (Zhu, Gedeon and Taylor, 2009).
Two different types of head tracking control techniques have SYSTEM DESIGN AND IMPLEMENTATION
been described and evaluated, one is using a users continuous Our design of using eye tracking for camera control is based
head movements, called head motion control; the other one on the gaze coordinates on the screen. Figure 4 shows details
D ZHU,T GEDEON AND KTAYLOR
FIG 4 - Eye tracking for camera control.

of the design. We can break down the entire approach into the video stream, the camera will always bring that region to
following three steps: the centre of the screen. In this process usually the user does
1. Processing the raw gaze data to filter noisy points,
not feel they are performing deliberate control of the camera
recognise gaze fixation by calculating the centroid of movements.
grouped non-noisy points. The prototype system was implemented in two major parts:
2. Calculating the distance and the angle between the current the user end and the remote camera site, in between can be a
fixation position and the centre of the screen. standard network connection. The overall system architecture
3. If the fixation is in the central area (distance < the radius is illustrated in Figure 5.
of the central area), the camera will remain at the current At the user end, we integrated the FaceLAB@ 4.5 eye
position. On the other hand, if the fixation is out of the tracking system (laptop-based version) into our prototype,
central area, the camera will start moving with the angle which provides the real-time gaze tracking at a 60 Hz
calculated between the current fixation and the centre of frequency without the use of markers. This avoids the need to
the screen. make the user wear any specialised devices, offering comfort
Since human raw gaze points are quite noisy, and are and flexibility. The eye tracker was connected to a main PC
not suitable for direct application (Jacob, 1991), we used a through a local network for transferring the real-time raw
modified version of the Velocity-Threshold Identijication gaze data. The FaceLAB@Client Tools SDK was installed on
fixation detection algorithm for filtering the raw gaze points
the main PC, called by the eye tracking camera control code
into fixations, as this method is straightforward to implement,
runs very efficiently, and can easily run in real time (Salvucci for receiving the raw data from the local network. The control
and Goldberg, 2000).The camera motion will keep following code translates the raw gaze data into corresponding camera
the users current fixation direction, if its position is not in the control commands and sends them to the remote camera
centre area of the screen. The overview of the entire process through the external network. The laptop-based eye tracker
is that wherever the user focuses their visual attention in the shared the user screen for eye tracking on the main PC,
FIG 5 - System architecture for the eye tracking remote camera control prototype.
THE DEVELOPMENTOF A NOVEL EYETRACKING BASED REMOTE CAMERA CONTROL FOR MINING TELEOPERATION
as the user would only be seeing the video stream from the
remote camera on the user screen. The eye tracking camera
control code and other relevant software integrations were all
implemented in Visual C++.
On the other site, we used the Pelco@ES3oC as the remote
camera to be controlled in the prototype system with the
capability to perform pan and tilt functions simultaneously
(maximum pan speed = iooo/s,maximum tilt speed = 30/s).
It was connected to the user end through an external network,
transferring the live video stream back to the user and also
receiving the control commands to carry out the relevant
camera movements.
USER EVALUATION
A lab-based user study was conducted to evaluate how well the
eye tracking camera control could perform in a model of a real- FIG 7 - Experimentalsetting: (1) participant; (2) remote camera; (3) screen
world hands-busy setting, in comparison to a conventional view of video stream; (4) eye tracker; (5) gamepad as joystick control;
joystick control. The rock breaker operating in a production (6) redesigned foosball table; (7) covers to obscure participants direct view of
environment was not available for user studies, so we modelled the foosball table.
the original setting by a functional physical model (Gedeon
and Zhu, 2010), using a physical game analogue: a redesigned statistical analyses on the collected data clearly tell us that
foosball game with two handles (see Figure 6). the eye tracking control performed significantly better results
than the joystick control (see Figure 8).
I User feedback from the questionnaires remains consistent
with the objective results. The eye tracking control significantly
outperformed the joystick control through all the criteria we
selected (see Figure 9).
In addition, most of the participants directly commented
that the eye tracking control was quite effective for resolving
the hands and attention switch problem in the experiment.
Compared to the joystick control, it was more convenient and
flexible, required less physical movements and consciousness,
which resulted in saving time and effort from the alternative
of switching hands, obtaining significantlymore opportunities
to kick the ball and score more goals.
CONCLUSIONS AND FUTURE DIRECTIONS

The work presented addresses a user interface problem that
existed in our remote rock breaking project, which requires
operators to frequently switch hands and attention between
FIG 6 - Redesignedfoosball table for the user experiment. different mining control interfaces. We introduce a novel
We recruited university students as experimental subjects. design of using eye tracking for remote camera control as an
The reason we chose to construct our functional physical effectiveapproach for this particular problem with a developed
model primarily on such a game based task was because of the prototype system. We conducted a lab-based user evaluation
competitiveness and engagement we observed among the real on the prototype system comparing eye tracking control and
joystick control, using a functional physical model, assessing
operators performing operations in the original setting. We
the results through both objective and subjective measures.
believed that our new design could be an appropriate model The objective results demonstrate out-performance of using
for the original rock breaking setting with the advantage of eye tracking control and the user preference and comments
being more compelling and interesting to our student based also indicate that eye tracking control can solve the switching
subjects. problem effectively.
Figure 7 shows the experimental setting of the evaluation. Future directions would be exploring more human
For each camera control method (eye tracking and joystick), natural interaction based design prototypes, or multi-model
participants had five minutes to play the foosball game. designs by combining gaze and other types of interaction
No pretraining time was offered to get used to either of the techniques particularly for mining teleoperation, and also the
control method. Both goals and kicks were recorded as each investigation of improved user evaluation.
participants objective performance, and following with a
questionnaire to collect participants subjective preference, ACKNOWLEDGEMENTS
regarding naturalness, required consciousness, distraction
The authors would like to express their appreciations to all the
and time to get used to each camera control. volunteers that participated in the user study, also thank Chris
Gunn, Matt Adcock and Bodhi Philpot from the Immersive
EXPER IMENTA L RES ULTS Environments Team, ICT Centre, CSIRO for their valuable
The major objective measures are according to the analysis of suggestions and helps on the design and implementation
the number of goals and the number of kicks each participant of the system. Thanks to Eleonora Widzyk-Capehart for
achieved in the corresponding camera control trail. The reviewing the work.
D ZHLTGEDEON AND KTAYLOR
FIG 8 -Objective performancecomparison-goals and kicks.
FIG 9 -Mean results for questionnaires:(5-point Likert scale: 1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 =agree, 5 = strongly agree).
This work was supported by the Transforming the Future Gedeon, T D and Zhu, D, 2010. Developing a natural interface for
Mine theme under CSIRO National Minerals Down Under a complex task using a physical model, in Proceedings Second
(MDU) Research Flagship. IEEE International Conference on Intelligent Human Computer
Interaction (IIICI 2010), Allahabad, India.
Gedeon, T D, Zhu, D and Mendis, B S U, 2008. Eye gaze assistance for
REFERENCES a game-like interactive task, International Journal of Computer
Cohen,J C, Conway, L and Kiditscheck, D, 1996. Dynamical system Games Technology, (623725):i-10.
representation, generation, and recognition of basic oscillatory
Goh,A H W, Yong, Y S, Chan, C H, Then, S J, Chu, P L., Chau, S Wand
motion gestures, in Proceedings Second International Conference
Hon, 1 W, 2008. Interactive PTZ camera control system using Wii
on Face Gesture Recognition, pp 60-65.
remote and infrared sensor bar, in Proceedings World Academy of
Duff, E, Caris, C, Bonchis, A, Taylor, K, Gunn, C and Adcock, M, 2009. Science, Engineering and Technology, pp 127-132.
The development of a telerobotic rock breaker, in Proceedings Gunn, C, Hutchins, M, Stevenson, D, Adcok, M and Youngblood, P,
Seventh International Conference on Field and Service Robots 2005. Using collaborative haptics in remote surgical training, in
(FSR 2009), pp 1-10.
Proceedings First Joint EuroHaptics Conference andSymposium
Fong, T and Thorpe, C, 2001. Vehicle teleoperation interfaces, on Haptic Interfaces for Virtual Environments and Teleoperator
Autonomous Robots, 11(1):9-18. Systems (WorldHaptics 2005), pp 481-482.
THE DEVELOPMENTOF A NOVEL EYETRACKING BASED REMOTE CAMERA CONTROL FOR MINING TELEOPERATION
Hainsworth, D W, 2001. Teleoperation user interfaces for mining Taylor, K, Dalton, B and Trevelyan, J, 1999. Web based telerobotics,
robotics, Autonomous Robots, ii(i):19-28. Robotica, 17(1):49-57.
Hughes, S and Lewis, M, 2005. Task-driven camera operations for Yanco, H, 1998. Wheelesley, a robotic wheelchair system: Indoor
robotic exploration, IEEE Transactions on Systems, Man, and navigation and user interface, Assistive Technology andArtiJicia1
Cybernetics - Part A: Systems and Humans, 35(4):513-522. Intelligence, Applications in Robotics, User Interfaces and
Jacob, R J K, 1991. The use of eye movements in human-computer Natural Language Processing, pp 256-268.
interaction techniques: What you look at is what you get, ACM
Zhai, S, Morimoto, C and Ihde, S, 1999. Manual and gaze input
Transactions on Information Systems, 9(3):152-169.
cascaded (magic) pointing, in Proceedings i p h ACM SIGCHI
Kizil, M S and Hancock, W R, 2008. Internet-based remote Conference on Human Factors in Computing Systems (CHI
machinery control, in Proceedings First International Future 19991, PP 246-253,
Mining Conference and Exhibition 2008, pp 151-154 (The
Australasian Institute of Mining and Metallurgy: Melbourne). Zhu, D, Gedeon, T and Mendis, B S U, 2010. Fuzzy methods and eye
gaze for cooperative robot communication, International Journal
Saito, S, 1992. Does fatigue exist in quantitative measurement of eye
of Intelligent Information and Database Systems, 4(1):43-59.
movements?, Ergonomics, 35(5/6):607-615.
Zhu, D, Gedeon, T and Taylor, K, 2009. Keyboard before head
Salvucci, D D and Goldberg, J H, 2000. Identifying fixations
and saccades in eye tracking protocols, in Proceedings 2000 tracking depresses user success in remote camera control, in
Symposium on Eye Tracking Research and Applications (ETRA Proceedings 12th IFIP TC13 Conference on Human-Computer
2000),pp 71-78. Interaction (INTERACT 2009), Uppsala, Sweden, Lecture Notes
in Computer Science, 5727:319-331.
Tall, M, Alapetite, A, Agustin, J S, Skovsgaard, H H, Hansen, J
P, Hansen, D W and Mollenbach, E, 2009. Gaze-controlled Zhu,D, Gedeon, T and Taylor, K, 2010. Natural interaction enhanced
driving, in Proceedings 27th ACM SIGCHI Conference Extended remote camera control for teleoperation, in Proceedings 28th
Abstracts on Human Factors in Computing Systems (CHI 2010), ACMSIGCHI ConferenceExtended Abstracts on Human Factors
PP 4387-4392. in Computing Systems (CHI 2010), Atlanta, USA.
Sydvaranger Magnetite Mine,
Norway -0wnerMining Tenders and
Implementation
S Cruickshanks'
ABSTRACT
Northern Iron Ltd (NIL) is an Australian listed company that recently brought the historic
Sydvaranger open pit Magnetite Mine (the Project) in Northern Norway successfully back into
production.
NIL made the decision to operate the Project on an owner mining basis, due to limited large
scale mining expertise amongst Norwegian mining contractors, a long mine life, a limited range in
planned mining rate and to minimise costs.
As a new mine operator in a country where labour practices are strictly regulated, there was
the desire to reduce the risk associated employing and managing skilled personnel. This was
achieved by coupling the purchase of the required mining equipment from the original equipment
manufacturer's (OEM) regional dealers with OEM supported Maintenance and Repair Contracts
(MARC).
This paper examines the process that was followed to implement an owner mining operation
in a short time frame; approximately 12 months from the initial requests for quotation to actual
delivery of mining equipment to site, and some of the challenges faced.
By following a formal and well-tested process, not only was the implementation achieved in the
set time frame, but significant value was added through the cost savings achieved in the equipment
purchase price and MARC rates that were negotiated through this process.
INTRODUCTION
Project setting and history
The Sydvaranger Iron Project is located at latitude 70"N, During the Second World War, the German army occupied
longitude 30"E in the municipality of Serr-Varanger in the area to provide a staging post for their invasion of Russia.
Finnmark, northern Norway, north of the Arctic Circle, as The Russian nickel mines immediately across the border
shown in Figure 1. and the iron ore mines at Bjerrnevatn were considered to be
The concentrator and port facilities are located in the town of strategic importance to the German war effort. The Allies
of Kirkenes, and the open pit mine is located approximately bombed and destroyed the Kirkenes facilities to prevent the
8 km to the south near the town of Bjerrnevatn. Germans from producing iron ore.
The mining project covers an area of approximately 35 kmz The plant was rebuilt in the early 1950s using funds provided
and is bounded by two major bodies of water, namely in part by the Marshall Plan, as it was deemed to be of regional
Langfjorden to the west and the Pasvik River to the east, as strategic importance given the proximity to the Russian
shown in Figure 2. The Pasvik River forms the border between border. Government of Norway ownership of the mine in the
Norway and Russia. post Second World War period varied between 87 per cent
and 100 per cent.
Prior to the original development of the magnetite iron ore
mine in 1910, the surrounding area was largely uninhabited, In 1969, Sydvaranger AS commenced production of acid
except for the Sami reindeer herders and a small fishing pellets for use in steel mills. This was followedby the progressive
village at Kirkenes. production of high grade magnetite concentrate in 1989, a hard
ferrite product in 1994 and a pigment product in 1995.
The mine operator Sydvaranger AS was founded in 1860
and the plant was commissioned in 1910. Production from the Ore production has been from a number of open pits, as
mine was carried out continuously from 1910 until 1997, apart shown by shading in Figure 2.
from interruptions during the two World Wars, and in exess In 1985, a decision was taken by the Norwegian government
of 200 Mt of magnetite ore has been mined. to close the mine by 1995.At this time, the mine and associated
1. MAuslMM, Specialist Consultant,Coffey Mining, 1162 Hay Street, West PerthWA 6005. Email: Stuart-cruickshanks@coffey.com
SEVENTH LARGE OPEN PIT MINING CONFERENCE / PERTH, WA, 27 - 28 JULY 2 0 1 0 39

S CRUICKSHANKS
FIG 1 -Project location.

companies employed 1400 people and was run as a social DEFINITION OF OWNER MINING
producer, with maximising employment being a major factor
The mining process consists of a number or inter-related
in the operation of the mine.
functions, including:
The mine management attempted to extend the mine life by
mine design,
investigating the potential to produce higher value products
and developing an underground mine beneath the Bj~rnevatn scheduling and budgeting,
open pit. This resulted in a limited extension to operations, drilling and blasting,
but mining was completed in 1996 and processing ceased in loading and hauling of ore and waste, and
April 1997. equipment maintenance.
In 2007 NIL was formed to acquire all the mineral assets As discussed by Kirk (2000) most of these functions have a
and associated production facilities of the Sydvaranger Iron supply and/or service component, which is typically defined
Project held by SydvarangerGruve AS. NIL acquired a 100 per by a formal contract. Supply contracts are normally based on
cent interest in Sydvaranger Gruve AS immediately prior to its a specific material or product where the focus is on quality,
listing on the Australian Stock Exchange in December 2007. product specification, delivery and price. Service contracts
NILS development plan was predicated on refurbishing the are typically more complex and which party is responsible for
existing infrastructure with the necessary plant and equipment what specific function is important. How risk is shared and
to process approximately 7 Mt of ore a year to produce in the the duration of the service may also be significant issues.
order of 3 Mt of iron concentrate a year over an estimated
mine life of 28 years. At an average strip ratio of 1.8:1(waste There are often a number of choices as to who has the prime
tonnes : ore tonnes) the total material movement requirement responsibility for each function or part of a function, ranging
for the mining fleets is approximately 20 Mt a year. from most functions being under the direct control of one
SYDVARANGER MAGNETITE MINE, NORWAY - OWNER MININGTENDERS AND IMPLEMENTATION
FIG 2 - Deposits and existing pit outlines.
contracting company (Contract Mining) to the other extreme DECISION BETWEEN OWNER AND CONTRACT
of the mine owner controlling the majority of functions
(Owner Mining).
MINING
Kirk (2000) concluded that where direct mining costs are a
In most mines the actual situation is somewhere in between. significant proportion of total operating costs, a mine has a
For the purposes of this paper the distinction between Contract life of more than five years, there is the ability to maintain
and Owner mining is based on which party has direct control a reasonably constant mining rate and there are existing or
over the mining equipment and the people who operate that available experienced mining personnel then owner mining
equipment, which is Sydvaranger Gruve in this case. should be considered along with the option of contract mining.
S CRUICKSHANKS
The use of contract mining offers mine owners a number Tender documentation was accompanied with a cover
of advantages for projects that have short mine lives, widely letter to include details of the tender schedule and process,
varying mining rates, limited company borrowing or credit such as procedures to be followed, date and details for a
capacity, complicated joint venture arrangements, a lack of site visit and closing date for the tenders.
mining experience, a policy of out-sourcing or a rigid labour Formal site visits jointly hosted by Coffey Mining and NIL
market. senior personnel. Coffey Mining responsible for managing
For the Sydvaranger project, the long mine life and the itinerary,taking minutes, answeringtechnicalquestions.
a relatively uniform mining production requirement Following the site visits a formal addendum was issued
lend themselves to an owner miner operation. However, for each tender answering questions raised during the site
Northern Iron is a single mine company that did not have visit and detailing agreed changes to the tender documents
an existing pool of experienced and trained operating or or process. Further addenda issued in response to further
maintenance personnel to draw on. This coupled with rigid questions received up until tender close date.
labour conditions in Norway were factors that could favour a
contractor carrying out the mining operations. This was the All tenders received and initial evaluation carried out.
reason contract mining tenders were issued in parallel with Tender clarificationletters drafted and sent to all tenderers.
mining equipment supply and maintenance tenders, plus key It is during the clarification process that the first aspects
consumable supply tenders, to enable both the contract and of negotiation and compromise were commenced. Once
owner mining options to be evaluated. clarification responses were received a cost analysis and
During the contract mining tender process it became comparison was carried out.
apparent that there was little large-scale open pit mining Short lists of preferred suppliers were prepared. This
experience among the available contracting companies, resulted in:
or sufficient interest from them to obtain this experience. short list of two equipment suppliers (including
This negated any advantage the mining contractors had in maintenance),
having experienced personnel to call on. As any potential shortlist of two explosive suppliers,
mining contractor would be recruiting from the same pool of short list of two fuel and lubricant suppliers, and
people available to NIL under the same labour conditions it preferred tyre supply and management contractor
was decided that there would be little advantage in using a selected.
mining contractor. Labour conditions different to Australian Formal meetings with the short listed suppliers were
conditions outlined in the 73 page Working Environment held in Oslo and at the Sydvaranger site. These meetings
Act of 2007 includes tight restrictions on working hours, involved discussions and negotiations on technical,
including shift duration, overtime restrictions and hours commercial and contractual issues. For the equipment
worked per week, generous leave provisions and difficulty and tender the actual support promised by the OEM to the
cost of terminations due to poor performance. Pay levels are suppliers was important.
also high but the range in pay from unskilled to management The suppliers in-country facilities were inspected to help
levels is relatively small and doesnt encourage acceptance of assess their capability to support the contracts.
greater responsibility. After the aforementioned meetings formal responses
However in order to minimise the risk associated with were received from the suppliers and further evaluation
the ability to directly employ skilled personnel, to cap the completed, resulting in formal recommendations being
maintenance cost, to share the risk of productivity and made to NIL.
availabilityand to reduce capital needs for supporting tools and Once a preferred supplier was identified the first draft of the
equipment it was decided to have all equipment maintenance
relevant contracts were compiled. A formal letter advising
carried out on a contract basis. This was achieved by coupling
the purchase of the required mining equipment from the the preferred supplier of their status, and any special or
OEMs regional dealers with OEM supported Maintenance contingent conditions, accompanied this contract draft.
and Repair Contracts (MARC). Another round of formal meetings were held with the
suppliers to finalise a second draft of the contract in which
TENDER AND SUPPLIER SELECTION PROCESS most key technical, commercial and contractual matters
were agreed. Coffeys approach in these meetings was
The mining equipment supply and maintenance tender to actually make the agreed changes to the electronic
process included the following steps: documents at the meetings using a data projector and
Expressions of interest (EOI) sought from equipment having people on both sides, ie both NIL and the supplier,
dealers and main consumable suppliers. being able to make decisionsat the meeting. This approach
Tender list prepared based on responses to expressions of generally leads to fewer instances of non-agreement
interest. between both parties, as well as reducing the negotiation
Tender documents drafted for the following items: time frame.
equipmentsupply(full cost, with all specified and agreed With this draft finalised it was possible to finalise a letter
extras for the climate or site needs, commissioned on- of acceptance of their tender with the supplier. It is normal
site); that a letter of acceptanceis sufficient for the supplier to be
equipment maintenance and repair (based on hourly able to place orders on the factory with deposits from the
maintenance rates plus monthly fee for overheads and purchaser to closely follow.
including availability and performance guarantees); Final contract negotiations were instigated after the letter
explosive supply (including all infrastructure and a of acceptance was issued. For the equipment supply
complete down hole service); contracts another three drafts were required (five drafts
fuel and lubricant supply and infrastructure in total). For the equipment maintenance and repair
construction; and contract five more drafts required (seven drafts in total)
tyre supply and tyre maintenance and management before a final contract was reached reflecting the greater
services. complexity and commercial value of this contract.
SYDVARANGER MAGNETITE MINE, NORWAY - OWNER MININGTENDERS AND IMPLEMENTATION
TABLE 1
Owner mining implementation schedule.
Coffey were then involved in ongoing support of NIL month contract and support to both NIL and the maintenance
personnel in the implementation of these contracts. providers through the services of subconsultant Sam
The overall time frame for the complete tender process, from Adamson of MMS Mining.
EOI to final contracts being in place, is summarised below for Of course no start up of a mining operation, and especially
the two mining equipment supply and maintenance contracts: in a remote location above the Arctic Circle in winter, goes
0 EOIissued: January 2008 without a few speed bumps. Delays with the process plant
First tender documents issued: February 2008 start up did allow for a slower ramp up to full mining
Site visit: March 2008 production than was originally planned. This was fortunate
as the lack of large scale mining experience and maintenance
Negotiations to final draft: June 2008
contracts in Norway led to a greater training and supervision
Acceptance letters issued and June 2008 requirement than had been originally envisaged.
equipment ordered:
The Project also experienced issues as a result of cultural
Final contracts signed: September/October 2008 differences. For example, there is a difference in construction
First mining fleet delivered: March 2009 culture between Norwegians and Australians. It was observed
During the tender process significant cost savings of that the Norwegian construction mentality was towards
approximatelyUS$2.5million in fleet supply and maintenance highly engineered, long lasting structures as opposed to the
cost, compared to the submitted tenders, were achieved. Australian mining industry tendency to engineer for the life of
the project and to a more utilitarian standard.
Other technical issues included problems with reverse
IMPLEMENTATION OF OWNER MINING circulation (RC) grade control fitted to the blasthole drills,
A major benefit of having comprehensive contracts in place drill consumable life and costs, dust suppression, load and
prior to project start up was that it allowed newly recruited haul productivity, establishing adequate quality control in
NIL site management to effectively hit the ground running drill and blast to consistently develop flat mining benches.
and concentrate their efforts on project implementation,
including recruitment and training of operations personnel. ACKNOWLEDGEMENTS
It also provided a secure framework for the suppliers to The author wishes to thank Northern Iron Limited for their
commence construction of the required facilities with a permission to use their information in this paper. The author
reduced risk of disputes. also wishes to thank the following individuals for their
The implementation schedule plan versus actual is shown assistance in preparing this paper:
in Table 1. Linton Kirk, Senior Principal Mining Engineer, Coffey
By following the tested formal tender process the mining Mining;
fleet, maintenance support and most major consumable Andrew Beuchner, Production Engineer, Sydvaranger
suppliers were on site on time to meet the clients ambitious Gruve - Northern Iron Limited; and
time schedule. It is also of note that the contracts and the Sam Adamson, Director, MMS Mining.
subsequent maintenance procedures and systems put in place NIL has given approval to publish this paper.
by the MARC contractors are being used as a benchmark by
the OEMs for subsequent projects. REFERENCES
In a tight labour market Coffey Mining was also able to Kirk, L J, 2000. Owner versus contract mining, in Proceedings
support this process by providing a production engineer Mine Planning and Equipment Selection Conference, Athens,
on a contract basis, a temporary mine manager for a three November.
Slope Monitoring in Open Pit Mining -
A Case Study
S Narendranathan, A J Beer2and M E Heap3
ABSTRACT
The complexities of mining in current economic conditions, where ore extraction schedules are
varied on a weekly basis so as to ensure constant supply to the process plant, can result in a
number of scenarios that pose geotechnical challenges. This paper presents a case study where
alterations in the mining schedule of a particular mine resulted in the necessity to mine under a
large uncompacted embankment, ie integrated tailings and mine waste land form. To effectively
ensure that ore recovery was maximised, a well-designed slope monitoring system to evaluate
design performance and to provide timely warning of impending failures was imperative.
In order for an appropriate slope monitoring strategy to be developed,a thorough understanding
of the potential slope failure mechanisms, capabilities and limitations of available monitoring
techniques is required. Communication of the monitoring results and response strategy to alarms
generated by monitoring systems must also be carefully considered to maximise effectiveness.The
results of preliminary stability modelling can be used to guide the design of a monitoring system,
which in turn can then generate data which can be used to iteratively refine the geotechnical
model. This paper illustrates the development and implementation of a slope monitoring system
and associated communication system, using the Daydream embankment at Fortescue Metals
Groups (FMGs) Cloudbreak Operations as a case study.
INTRODUCTION
Importance of appropriate slope monitoring
Slope monitoring is becoming widely used in most open pit occur following very small deformations. Such failures may be
mines, large embankments and waste dumps. The monitoring highly localised and, unless the area is being very closely and
can range from simple methods, such as visual observations accurately monitored, failures may occur apparently without
and simple crack dilation measurement, to more complex warning. In such situations, accurate, well-targeted and
methods such as micro-seismic monitoring and the use of the possibly telemetered monitoring systems are required, which
slope stability radar. The overall monitoring design has to be are discussed in more detail later on in this paper.
tailored to meet the requirements of the individual slope or
Conversely, waste dumps, large tailings embankments and
structure. To facilitate this, an understanding of the potential
slope instability mechanism is required and the appropriate integrated waste landforms (ie combined tailings storage
monitoring techniques selected. facilities and waste dumps) are constructed of non-brittle
material; hence they are capable of experiencingdeformations
Failure to adequately monitor a slope may result in:
of much greater magnitude than those shown by brittle rock
0 loss of life, masses. These deformations may be recorded over a large
severe equipment damage, area, and still not lead to slope failure. Such structures are
production delays, and more challenging to assess when compared to those excavated
possibly the loss of the mine. in brittle ground conditions.
The effectiveness of monitoring depends on the extent to
This paper outlines a case study which demonstrates the
which the slope gives adequate advanced warning before
requirement for scrutinised geotechnical monitoring of
failure, and on the ability of the monitoring system to detect
an embankment, due to proximity of personnel exposure.
such a precursor. There is considerableevidence (for example,
The process by which the existing stability scenario of the
Ding eta!, 1998)to indicate that slopes do give ample warning, embankment was calibrated from an early deterministic
which means that it is well worth spending time on the design model (as opposed to a finite element model as is usually
and implementation of appropriate monitoring systems. the case) using the Newmark Analysis is discussed in detail;
It is important to note, however, that in open pits currently along with the iterative process of back analysis used to obtain
being excavated in hard and brittle rock, failures can often the as built strength parameters of the embankment. The
1. MAuslMM(CP),Associate GeotechnicalEngineer, Coffey Mining Pty Ltd, 1162 Hay Street, West Perth WA 6005.Email: Sanjive-Narendranathan@coffey.com
2. PrincipalGeotechnicalEngineer,Coffey Mining Pty Ltd, 1162 Hay Street, West PerthWA 6005. Email: Adam-Beer@coffey.com
3. Associate GeotechnicalEngineer, Coffey Mining Pty Ltd, 1162 Hay Street, West PerthWA 6005.Email: Mark-Heap@coffey.com
S NARENDRANATHAN.A J BEER AND M E HEAP
general importance of slope monitoring and feed back for the by calibrating the observed deformations, as recorded by the
purposes of stability model calibration is discussed below in monitoring instrumentation.
greater detail.
DAYDREAM CASE STUDY BACKGROUND
Slope monitoring and its role in thedesign
loop Embankment design
Most slope design engineers are familiar with the processes In June 2009, FMG completed the construction of an
involved in the design of an engineered slope, as illustrated embankment to the south of the tailings storage facility (TSF)
in the flow chart (Figure I). The importance of the role of at the Daydream Pit, Cloudbreak Operations, East Pilbara,
Western Australia. This embankment was intended to retain
slope instrumentation and monitoring is demonstrated in tailings and allow drill and blast and surface miner operations
steps 4 and 5, whereby the preliminary geotechnical model (SMO) downstream of the TSF (Figure 2). The area to be
(as assessed during the initial investigation) is limited, as mined was approximately 180 m from the downstream toe of
it is usually based on limited drill-hole information. This the embankment and the duration of mining was anticipated
model needs to be refined during the construction process to be in the order of three to four months.
1) Preliminary Geotechnical Investigation involving:

. Investigative Geotechnical Drilling Campaign
. Core Logging and Sampling
. Laboratory Strength Testing
2) Formulation of a Preliminary Geotechnical Model:

. Determination of rock mass and defect plane strengths
. Definition of geotechnical design domains
. Identification of potential stability controls within the
respective geotechnical domains
--
I
The Design Loop-
4---------.
T
I
I I
I
3) Undertaking a Slope Stability Analysis:
. Assigning material parameters for the rock mass I
I and defect planes within the individual geotechnical I
domains
I . Determination of the critical failure geometry I
. Formulation of achievable slope geometry for the I
I respective domains
I I
I
I
I
I I
I I
I 4) Mine Development and Design Execution: I
. Mine development commences based on I
I preliminary geometry as per step 3
. Slope instrumentation and monitoring should be I
I undertaken to assess its performance based on the
I
I initial design and the analysed failure geometry
I
I
I I
I I
I I
I
5) Assessment of Slope Monitoring Data and
Design Optimisation:
I
I . Regularly assess slope monitoring data to I
I
determine zones of potential instability
. Appropriately modify slope geometry to reflect I
I
slope monitoring data within zones of concern
. Refine geotechnical model i.e rock mass strengths
I
I to reflect deformation noted within slopes I
I I
_
l_____,,
-
,
--
- I
FIG 1- Flow chart of slope design process.
SLOPE MONITORING IN OPEN PIT MINING - A CASE STUDY
0 As construction continued, safety bunds were created at the

sides of the embankment (in line with safe work practices)
that were approximately 3 m - 4 m wide and 1.5 m high
It was not possible to compact the placed embankment
material below these bunds, which resulted in a zone of
3 m - 4 m width of uncompacted materials on both the
upstream and downstream faces of the embankment.
The final embankment was approximately 32 m wide at
the top and 29 m high, measured from the downstream
side.
Proposed mine extraction schedule
Access to high-grade ore, from the area downstream of the
Daydream TSF, had been discussed amongst the FMG Mine
FIG 2 -As built embankment. Planners for a number of months prior to eventual execution.
The delay in access was due to speculation concerning the
The embankment was designed as follows: stability of the Daydream embankment, given the limited
The downstream batter of the embankment, particularly information that was available for input into its design.
the lower lifts, were to be constructed of coarse, relatively
The extraction of high grade material was to commence
free draining mine waste to reduce the potential for a
mid-February 2009. There was some debate whether a
raised phreatic surface within it. It was proposed that
the horizontal width of this layer should not be less than surface miner (SMO) was to be used or conventional drill and
20 per cent of the total embankment width at the level blast methods. This debate, to some degree, hinged on the
where the materials were to be placed. embankments capacity to tolerate close-proximity blasting,
Alluvialmaterials (dark red in colour) were considered to be even though blast designs were restricted to a maximum
unsuitable for the downstream embankment construction. peak particle velocity (PPV) of 25 mm/sec. One of the major
Previous studies involving such waste materials indicated limitations of using the SMO was the limited space available
that they are sensitive to water inundation; sodic soils and to manoeuvre. Ultimately, combinations of both methods
Pindan soils collapse when wet. were utilised. It was anticipated that all ore downstream of
The construction sequence was to comprise: the Daydream TSF would be removed within three months.
1. Paddock dumping of the waste over the full embankment
width in each lift, which was not to exceed 4 m in height. MONITORING SYSTEM DESIGN AND
2. A dozer was to flatten the tops of the paddock dumped INSTA LLATION
stockpiles to provide a horizontal running surface for the
placement of the next lift. Aims
3. Paddock dumping of the waste over the full embankment The aims ofthe monitoring system were determined to be:
width, on the flattened surface, over the formerly
To evaluate the need for ongoing monitoring (existing
dumped materials, would then be undertaken. The haul
cracks and inspection of embankment for evidence of
trucks placing new mine waste were to traffic the full
seepage or slumping, particularly after blasting).
embankment width, with the exception of the outer 4 m
on each lift. To evaluate the need to review and/or modify the mining
schedule, should embankment seepage or evidence of
4. A dozer was to then flatten the tops of the paddock dumped
additional embankment distress be observed.
stockpiles to provide a horizontal running surface for the
placement of the next lift. To facilitate economic extraction of the remaining ore, as
per the mine plan, ie to minimise the possibility of
Steps 3 and 4 were to be repeated until the embankment
reached its full height. Clean up costs - The cost of mobilising and operating
equipment, which would otherwise be used for mine
Actual embankment construction production to clean up failure masses on the pit floor,
Due to on-site constraints and to expedite construction, so that regular mining activities could be resumed.
the design had to be modified. Accordingly, the following Slope reformation - Depending on the volume of
construction methodology was adopted: failure, battering back (flattening) of the slope may be
required, so as not to initiate further failures. This is
The bottom 5 m (approximately) of the embankment was
sometimes carried out in parallel with regular pit floor
constructed of uncompacted material, normally used as
blast hole stemming, which was incorporated in the design cleanup activities.
to function as a drainage layer. As the existing ground Unrecoverable ore - In an extreme case, there is the
surface sloped towards the south, this coarse layer was potential for the slope failure to bury the orebody,
keyed into the existing slope. The lowermost 8 m of the or at least a critical blend, required for processing
downstream face of the embankment was also constructed purposes. The consequences of such a failure become
of free draining materials to mitigate against build up of significant, as production ceases entirely, until the area
pore water pressure. is remediated and the ore recovered.
0 Above the basal uncompacted layer, the embankment was Damage to equipment, personnel and infrastructure
built up in 1.5 m - 2.0 m layers. These layers were watered - Where key pieces of infrastructure, equipment or
and rolled, using dump trucks, at the time of construction, personnel are damaged or injured as the result of a
to assist compaction. slope failure, the consequences are significant.
S NARENDRANATHAN.A J BEER AND M E HEAP
Design The V W P sensors were installed along the centreline of

An integrated system was designed to monitor slope the Daydream embankment, spaced evenly. In a symmetric
deformation on the embankment which consisted of embankment, this would be the location of the highest
anticipated water levels, and thus would measure the highest
visual monitoring (inspection to identify for tension water pressure to give conservative pore pressures for stability
cracks, etc); calculations.
survey prisms (ATS); The location of borehole inclinometers with respect to the
crack extensometers; expected failure surface is critical. It has been demonstrated
borehole inclinometers; on many occasions (eg Beer and Narendranathan, 2009)that
vibrating wire piezometers (VWP); and without an understanding of the expected failure mechanism
slope stability radar (SSR). and the likely size of any possible failure, it is very easy to
install inclinometers in a location where they will provide no
The visual monitoring and radar can be regarded as blanket useful information or warning of impending failure. In the
monitoring techniques; that is, they are not directed at one case of the Daydream embankment, experience, backed up
particular target. The location of the radar itself is usually by stability modelling, indicated that the most likely failure
controlled more by operational factors (not obstructing moving surface lies relatively close to the downstream face of the
plant) than geotechnical considerations,as there will usually be embankment. Thus the optimal location for the inclinometers
a number of suitable locations where the radar can be placed so was determined to be as close as practicable (from an
as to be facing the monitored slope at a suitable angle. installation viewpoint) to the downstream edge of the top of
the embankment.
The survey prisms installed on the Daydream embankment
were actually installed by FMG before the desirability of The arrangement of the various elements of the monitoring
installing a comprehensive monitoring system was realised. system is shown in the plan in Figure 3.
Thus their positioning did not form part of the system design
process, although the assessment of prism monitoring data Installation
has been successfully integrated into the monitoring system. Initially, a detailed risk assessment was undertaken to address
The location of the crack extensometers was determined in issues associated with the access of drilling equipment onto
this case by the existing cracks on the top of the Daydream the embankment for the installation of instrumentation. An
embankment. Crack extensometers were installed on the evacuation and rescue plan was developed using the results of
cracks which had been determined by visual observation this risk assessment.
to be the most active. In other situations, where surface Once the instrumentation was installed, a supplementary
cracks have not yet developed, it can be envisaged that crack risk assessment was undertaken to address all issues associated
extensometers would be placed over the predicted backscarp with working downstream of the Daydream embankment.
location derived from modelling. In this case, the backscarp of The conclusions of this risk assessment and the subsequent
the optimised surfacewith the lowest modelled factor of safety mitigation measures were communicatedto all personnel that
would be used. would be working in the vicinity of the embankment.
FIG 3 -Layout of Daydream instrumentation.
SLOPE RESPONSETO MINING AND BACK difference, twice, to obtain the total displacement. However,
as this process is relatively time consuming and complex,
ANA LYS IS without the aid of software packages, Newmark put forward
Premining factor of safety of Daydream the following a simple approximation, which has been used by
the authors in this paper:
embankment
Drill and blast operations commenced on the downstream Dn = ppv2/ 2ac x d t
side of the embankment on 18 November 2009. At that stage where:
it was anticipatedthat the mining operations in the area would
continue for approximately three months. One of the main Dn is the total slope displacement
intended outcomes of the geotechnical monitoring program ppv is the peak particle velocity
was to effectively assess future slope response to the mining t is the duration of the motion
activities on the embankment, in particular:
vibrations as a result of mining equipment, and Slope response and model calibration
vibrations as a result of blasting activities. The first stage in calibrating a stability model would be to
However, prior to being able to effectively assess future assess the asbuilt FOS. As mentioned previously,the authors
slope response, the existing factor of safety (FOS) of the slope have chosen to utilise the Newmark analysis for this purpose.
needed to be quantified. The typical approach adopted by The first blast that was undertaken, post embankment
most geotechnical engineers is to assume that a slope can be construction, was on 24 November 2009. The vibration
considered to be stable if it has a minimum FOS of between resulting from this blast was picked up by the various
1.2 and 1.3. monitoring instruments that were installed on, and within,
the embankment.
The Newmark analysis The resulting deformation, as depicted below for
In the setting of the Daydream embankment the process inclinometer A, is an example which shows approximately
of back analysing an as built FOS constituted a complex 2.5 mm (Figure 4) of displacement occurring as a result of the
problem. As mentioned previously, the main reason for this blast, which registered a PPV of 19.5 mm/s.
was that there had been very little control of the construction
It should be noted that this displacement occurred over a
process to ensure:
period of five days and could be consideredto be creep induced
Adequate compaction of the placed fill material, which has a by the blast for the purposes of the Newmark analysis.
direct bearing on the shear strength of the overall slope.
Note: A similar amount of displacement was observed in
Adequate drainage to alleviate the build up of excess the other instruments, apart from the crackmeters, which
destabilising pore pressures. measure surface, as opposed to in ground deformation.
As a result of these multiple unknowns, which dont usually
occur in most open pit slopes, as these are, essentially,
excavations within in situ material that can easily (within
reason) be quantified; an alternate method of calculating the
standing FOS of the slope was embarked on by the authors;
the Newmark Analysis.
In essence, the Newmark Analysis puts forward a
methodology that correlates the magnitude of slope
deformation to the peak particle velocity induced as a result of
a mining induced vibrations, say from blasting.
The following description of the Newmark analysis is given
by Newmark (1965):
The Newmark Analysis transcends the limit
equilibrium approach to provide a n estimate of the
displacement of a landslide block subjected to seismic FIG 4 - Blasting related deformation.
motion. This is done by assessing the portion of the
vibration accelerogram above the critical acceleration Critical acceleration
ac, which is integrated twice to obtain the critical Based on the measured magnitude of deformation of 2.5 mm
displacement. (Dn), the critical accelerationfor the embankment can be back
calculated thus:
The critical acceleration is defined as follows, after Newmark
ac = (FS - i)g x Sinp pp3 / Dn = 2ap
where: ac= 76.05 mm/P

ac = critical acceleration in metres per second squared
b/S2) As built factor of safety
Fs = static factor of safety The as built FOS can be back calculated thus:
P = slope angle
FS = {(ac / SW) / 9) + 1
A rigorous Newmark Analysis involves subtracting the
critical acceleration from an accelerogram and integrating the FS = 1.118
5 NARENDRANATHAN,A J BEER AND M E HEAP
Back calculation of material parameters and friction angle have equal influence on the overall value of
Sincethe asbuiltFOS has been determined the in situ material the calculated FOS.
strengths can be calculated through a process of iterative back
calculation. This was undertaken in the following manner:
1. The originally assumed material parameters and stability
model was revisited to identify the most critical shear
strength component, ie the friction angle or the cohesion.
2. The inferred phreatic surface was revised based on
measurements from the installed W s .
3. The critical shear strength component was iteratively
varied, to assess the implication of the overall calculation
to the embankments FOS; this process is repeated till a
FOS of 1.118 (or as close as possible) is achieved.
4. The resulting shear strength components, after completing Percentage of Range (mean = 50%)
step 3, are considered to be the as built FOS for the
embankment. These parameters will be used for future FIG 6 - Initial limit equilibrium model spider diagram.
(predictive) modelling purposes, ie to establish standoff
distances and blasting proximity, etc. iterative back calibration of stability model
Initial limit equilibrium modelling The iterative process of back calculation was undertaken as
described previously. Application of this process suggested
The initial stability modelling was undertaken using the
Rocscience program slide (Figures 5 and 6). The following that the following shear strength parameters, would be
material parameters were utilised: appropriate, given a FOS of 1.118 (Figure 7):
Friction angle 30 Friction angle 21.5O
Cohesion 25 kPa Cohesion 25 kPa

Based on the stability modelling it was identified that
the cohesion and the friction angle were equally critical in Operational outcomes
relation to the calculated FOS as can be viewed from the Now that the as built FOS has been determined, and material
spiderdiagram, which basically assesses the influence of the shear strength properties modified in the light of the
respective parameters on the calculated FOS. iterative modelling, have been ascertained, it was possible to
It can be seen that the gradients of the cohesion and the undertake modelling to evaluate the anticipated influence on
friction angle are very similar. indicating that the cohesion embankment stabilitv of future changes or events such as:
FIG 5 - Initial limit equilibrium model for embankment.
FIG 7 - Back calculated limit equilibriummodel for embankment.

close proximity blasting; and REVIEW OF DATA RECORDING AND
an elevated phreatic surface, as a result of sustained TELEMETRY
tailings discharge into the TSF and surface water inflows
(ie recharge caused by rain). Data recording systems at daydream
Blasting induced vibrations The data from the respective monitoring techniques has been
collected as follows.
It was initially thought that blast stand-off distances with
respect to the embankment could be optimised so as to The visual monitoring is recorded manually and any changes
facilitate economic ore extraction. However due to a change noted are reported in the daily and weekly reports.
in mine production strategy (ie use of free-digging rather than The inclinometer survey results are recorded by a data-
drill and blast), this proved to be unnecessary. logger and used to manually update the inclinometer database
in the site office.
Tailings discharge
During the initial stages of monitoring, the prism data was
One of the main unknowns at the start of the project was the recorded by the mine surveyors and the results presented in a
effect of continuing tailings discharge into the embankment; spreadsheet format for review and analysis. As more specific
as this and recharge due to seasonal rainfalls have the potential instrumentation was installed, monitoring of the temporary
to elevate the phreatic surface within the embankment. This prism locations was discontinued.
was managed reactively by allowing tailings discharge to The slope stability radar (SSR) recorded deformation data
continue until increasing displacement trends were noted in to an on-board computer. This data was transmitted to the
the installed instruments, subsequent to which discharge was mine office computer, using a UHF radio modem. The system
arrested temporarily till the displacements settled. was also capable of transmitting data via the mines wireless
However as a calibrated model had been formulated it was network system although this is not in place at Daydream.
now possible to model the sensitivity of the slope stability The office-based computer could be accessed via the mines
network so that, all personnel who have the access password,
to the level of the phreatic surface within the slope, so as
can examine the radar data and associated information.
assess to what water levels could be considered acceptable. The radar system had an alarm capability, which could be
The modelling using the calibration-derived shear strength triggered by a number of user-definable criteria, for example,
parameters indicated that up to 5 m of additional head could a particular displacement threshold over a fixed period of
be accommodated within the slope, whilst still maintaining an time, indicative of an accelerating movement profile.
acceptable FOS. This knowledge was used by the engineers on The crackmeter and VWP data were continuously
site to determine predictively as opposed to reactively when recorded by a data-logger sited at one end of the Daydream
tailings discharge had to be halted, so as to allow the phreatic Embankment. The data was currently recorded every two
surface to subside prior to reinstating discharge again. hours. The recorded data was downloaded daily and added to
S NARENDRANATHAN,A J BEER AND M E HEAP
the monitoringdatabase. The system also sent a daily SMS text The great advantage of such a system is that it utilises an
message to registered users of the system,which was also used existing communication network so the data transmission
to update an internet database. An alarm facility also used the path is effectively taken care of by a third party, ie the
SMS system to warn registered users when monitored data telecommunicationssupplier. For transmitting large amounts
exceeded predetermined values, ie potentially indicative of of data, such as produced by the SSR, data charges alone
the onset of failure. would soon make such a system non-viable. The sporadic
delays experienced by all users of text messaging systems
REVIEW OF DAYDREAM DATA RECORDING must also be borne in mind when deciding whether to adopt
this telemetry system for alarming on safety critical slopes.
SYSTEMS
An alternative to these two systems was to use a mines
Automated systems wireless network system, where such exists for dispatch
purposes; it is the authors understanding that no such system
The degree to which data was obtained manually or otherwise
is currently available at Cloudbreak. The wireless network
varied with the type of monitoring being undertaken. Some
system does away with the need for line-of-sight between the
of the measurements, such as the VWP and crackmeter data
instrumentation and the data receiving point, and also has a
were very suitable for recording with a data-logger without
potentially higher bandwidth than the other two systems. In
the requirement for human intervention. The SSR was also
the case of the SSR, it also permits the end user to proxy in
configured to record and process a large amount of data,
directly to the radar computer and set up new scans remotely.
continuously, without human intervention.
However, the safety implications of monitoring information
taking up a significant fraction of a wireless networks
Manual systems bandwidth must be remembered, especially when such a
At the other end of the scale, the inclinometer surveys had system is being used for tele-remote operation of machinery.
to be undertaken manually, which could be a laborious and
time-consuming. This meant that data points were limited to a The varied nature of the data being recorded, often using
few readings per day. The introduction of the human element suppliers proprietary logging systems means that the data is
also gave rise to the possibility of human error affecting the not currently incorporated into a single monitoring database.
values measured. Although in-place inclinometer systems are A desirable development would be that of a database into
available, which would be suitable for continuous monitoring, which all monitoring data could be downloaded, and which
these are expensive and can only provide indications of could be accessed graphically. Whereas such systems do exist,
movement at the fixed levels within the borehole that they are they can be complex to set up given the disparate nature of
installed. the data being incorporated within them. The cost effort taken
to set up a single user-friendly monitoring database needs to
The initial prism monitoring system was also manual, be weighed carefully against the benefits to the end users of
although a number of stages of automation to reduce human the data. Care must also be given to determining appropriate
involvement are possible. The first of these would be to use levels of access to such a user-friendly monitoring system; if
a motorised theodolite with proprietary software, which people who do not understand the implicationsof monitoring
could be mounted on a survey pillar and automatically record results are given free access to all site data, the possibility of
a round of data, which is then downloaded manually into a unnecessaryor false alarms and despondency being generated
database in the office. The ultimate development of this is to is created.
use a robotic theodolite, with data telemetry, so that data is
continuously recorded, and errors introduced by repeated INTERFACE WITH MINING OPERATIONS
set-ups are eliminated. Several proprietary systems currently Monitoring results are communicated to mining operations
exist which have this capability. through a combination of alarms and monitoring reports as
follows.
Advantages and disadvantages of telemetry
systems Alarms
The data telemetry systems that existed on site were the The radar, crack extensometers and W s monitoring the
radio modem system used by the SSR and the SMS system Daydream embankment are all suitable for connection to
to transmit the crackmeter and VWP data. The radio modem an alarm system, due to the nature of the information being
system worked well when there was line of site between collected and the data telemetry already described. In the case
the relevant monitoring instrumentation (ie the SSR) and of the Daydream embankment, only the radar and the crack
the receiving station. One of the limitations was that data extensometers were actually configured to trigger alarms.
transmission rates can be relatively slow, which could lead Examination of the VWP data indicated a tendency to respond
to delays of ten minutes or so, in data being received at the to water levels in the tailings pond behind the embankment,
office computer, particularly when atmospheric conditions
rather than respond to the rainfall events which experience
are poor. For the purposes of recording data, this delay should
not normally present a problem, however, when using the had indicated had a causal connection with slope movements.
radar for monitoring safety critical slopes, consideration must Given that W response to water level changes and rainfall
be given by the engineer as to whether this amount of delay (in events was also gradual in nature, it was considered that
addition to the several minutes that it can take to complete a setting alarm triggers on W data values would not be as
large scan) is acceptable. The authors have attempted to use a meaningful as utilising the radar and crackmeter data.
repeater station with the radio modem system at another site,
but the data transmission rate was reduced to such an extent Radar
that the system was effectively unworkable. In situations Six different categories of alarm are generated by the radar:
where very small amounts of data are being transmitted, such red, orange, yellow, grey, green and blue. Of these, only the
as with an automated theodolite system, the radio modems red and orange alarms are triggered by slope movement;
can prove reliable and effective. the others are triggered by problems either with the radar
The text messaging SMS system is useful for transmitting a computer or data telemetry systems. The yellow, green and
daily summary of data received, and also for sending alarms. grey alarms effectively indicate that real-time monitoring of
the slope is no longer occurring. The blue alarm indicates a The results of this monitoring are submitted to the FMG
malfunction, but the radar continues to operate, and orange Mine Manager in the form of a daily report. This report also
and red alarms may still be triggered in this case. The alarm is comments on the observed response of the embankment (if
generated by a computer situated in Dispatch, which gives an any) to events such as blasting adjacent to the embankment,
audible signal. The computer can also be configured to display or rainfall events.
a message corresponding to each alarm, informing Dispatch The FMG Mine Manager communicated the relevant
as to who to contact and what actions should be undertaken information, if any, at the preshift handover meetings. Typical
in response to the alarm. The red and orange alarms are communications would inform of any issues such as alarms
considered to be the most serious, as they are triggered by triggered and resulting access restrictions.
real or apparent slope movements. These were initially set to
trigger when movements over a four pixel area exceeded 5 mm Weekly monitoring and detailed report
and 3 mm respectivelyover five hours. These initial threshold On a weekly basis, a detailed monitoring report was produced.
levels were assigned on the basis of prior experience at other This summarised the overall trends observed over the
mine sites with similar slope stability issues. preceding week and included graphical representations of
False alarms occurred during episodes of high winds and the monitoring results. Interpretation of the results was also
rains which caused localised rilling. To accommodate such undertaken on a weekly basis by the geotechnical engineer.
events without triggering an alarm the threshold levels were The interpretation attempted to link observations from the
reset to 20 mm and 15 mm over a five hour period for a different monitoring techniques to events which may have
ten pixel area. affected the embankment, and explain the observations in a
way which provided guidance to the mine management.
So that personnel were aware of the correct response to each
type of radar alarm, a Daydream Embankment Monitoring At the weekly monitoring report level, ongoing trends in
Radar Alarm Procedure was written, and distributed to monitoring data was assessed and, if there were any warning
signs of developing instability; these would be discussed in the
all personnel working in the Daydream embankment area.
report and recommendations for continuing to work safely
This procedure defined the mandated actions subsequent would be made. The appropriateness of the alarm levels for
to the triggering of the SSR alarms. Each category of alarm the SSR, the crackmeters and the W s were also reassessed
is associated with a different action; the red to grey alarms on a weekly basis. An example of the type of information
initiating immediate evacuation of all personnel from the area transmitted in a weekly report is the weekly report issued on
in the immediate vicinity of the embankment. The level of 30 December 2009.
response required was communicated by UHF radio, channel
4, from FMG dispatch to all site personnel. 30 December weekly report
A heavy rainfall event occurred during this week resulting
Crack extensometers in 24 mm of rainfall over a five hour period. The instrument
Data is recorded from each crack extensometer every two responses that were reported as follows.
hours and recorded on the installed datalogger. If a threshold
value for recorded movement is exceeded, then an alarm Slope stability radar (SSR)
is sent in the form of an SMS text message from a modem Radar displacement graph from 21 December 2009 to
connected to the datalogger to a nominated recipient list. 29 December 2009 (Figures 8 and 9).
The text message lists the identity of the crackmeters whose
movements have exceeded the threshold. The threshold
levels, to trigger an alarm, have been set at 10 mm on the
-; 12r. I
Deformation (Average)
upstream, and 5 mm on the downstream over a 24 hour v
period. The nominated recipients include mine dispatch

and the site geotechnical engineer. In the event of an alarm
being received, the site geotechnical engineer would inspect
the embankment and attempt to cross-referencethe reported
movements with those recorded by the SSR. If the movements
were consistently recorded by both systems, the procedure for
the relevant radar alarm is followed.
Time
Additionally, the modem returns a status SMS report at
06:15 every morning listing the absolute values of crackmeter I- Deformation (Average) I
movements.
FIG 8 - Slope stability radar displacement plot.
Daily monitoring report
An area of the embankment crest, to the east, showed an
The following tasks are completed on a daily basis by the on- increased rate of movement in comparison to the remainder
site geotechnical engineer at Daydream: of the slope. This movement was due to the development of
Visual inspection of the embankment for any changes, a local rainwater erosion gully, and associated sloughing,
particularly in those areas not covered by monitoring related to the rainfall event.
instrumentation.
Inclinometer surveys of two of the four functioning Inclinometers
inclinometer installations. This results in each of the As the result of the rainfall event the probe could not pass
installations being read every second day. beyond 14 m downhole depth within inclinometer C. It is
Downloading of crackmeter and VWP data from the thought that the deluge of water had caused a local collapse
embankment data-logger. within a localised zone. No changes in displacement were
Review of radar data for the preceding 24 hour period. noted in other inclinometers.
5 NARENDRANATHAN, A J BEER AND M E HEAP
id
FIG 9 - Radar screen capture from 21 December 2009 to 29 December 2009.
Vibrating wire piezometers (VWP) 25 December. As a result of these noted increases, mining
VWPA reacted in response to discharge of tailings and ponding
activity was suspended below the embankment. The radar
of water next to the embankment in the week preceding and VWP data were closely monitored over the following
the rainfall event. Even so, the plot shows that the rainfall few days until the radar plot indicated that movement had
event resulted in a phreatic peak occurring on 25 December ceased and the VWP plot showed pore pressure levels were
(Figure 10). This is particularly significant as instrument returning to their prerainfall event levels. Mining activity
reaction, as well as modelling exercises, have indicated was recommenced on night shift on 27 December, when the
that the embankment stability was particularly sensitive to radar indicated that movement had ceased. This response was
variations in phreatic surface. documented in the 30 December weekly report.
Outcomes arising CONCLUSIONS AND COMMENTS

The slope movements and increase in pore pressure at VWPA 0 The Daydream study demonstrated the potential of the
had already been reported in the relevant daily report for Newmark analysis to be used to determine the factor of
1
FIG 10 -Vibrating wire piezometers plot.
safety for a slope where there is great uncertainty as to ACKNOWLEDGEMENTS

what values should be adopted for the shear strength
The authors would like to thank FMG, in particular Mr Martin
parameters of the materials constituting the slope. In
Drage Operations Manager for Cloudbreak Operations, for
this case, the uncertainty arose because of the non- the opportunity to utilise the findings from the Daydream
conformance of construction process to the designed
Embankment for the presentation of this paper. The authors
process. also acknowledge the input provided by Mr Kweku Ayemin
Once the factor of safety had been determined, the and Mr Hamid Bagheri of Coffey Mining who provided site
material parameters could be back-calculated iteratively geotechnical engineering services.
to generate a geotechnical model for the slope.
The modelling undertaken by the back analysis also REFERENCES
provided information to guide the installation of
monitoring instrumentation by identifying the critical Beer, A J and Narendranathan, S, 2009. Slope monitoring and
management aspects, Presentation given at ACG Seminar:
slope failure surface. Knowing the location of this,
Geotechnical Engineering for Open Pit Mines, Brisbane,
instrumentation such as inclinometers could be located to Australia, June.
optimise the usefulness of data obtained.
Monitoring data provided feedback to mining operations Ding, X, Montgomery, S B, Tsakiri, M, Swindells, C F and Jewell,
R J , 1998. Integrated monitoring systems for open pit wall
personnel to ensure safe working. The level of monitoring
deformation, MERIWA Project No M236, MERIWA Report No
and development of an effective communications 186, June.
procedure meant that interruptions to mining were
minimised in length and optimum production rates Lilly, P A, 1982. A general definition of excavation reliability and its
maintained. application in mining engineering, university notes (unpublished),
Curtin University of Technology.
The monitoring data also indicated that generally the
slope was stable (except for limited movement during Lilly, P A, 2000. The minimum total cost approach to optimise pit
periods when the phreatic surface was elevated). This slope design, university notes (unpublished), Curtin University of
corroborated the assumptions made in developing the Technology.
stability model, and thus provided the feedback part of the Newmark, N M, 1965. Effects of earthquakes on dams and
design loop. embankments, Geotechnique, 15(2):139-160.
To date, the slope continues to be stable and monitoring Zavodni, Z M, 2007. Time-dependent movements of open-pit
has not indicated any further movements or changes in slopes, Slope Stability in Surface Mining (eds: W A Hustrulid,
water pressure since. M K McCarter and D Van Zyl), pp 81-87 (SME: Colorado).
Minimising Crest Loss by Eliminating
Surface Dilation Due to Presplitting
A Marton
ABSTRACT
In most mines using presplitting the final excavated wall shows a zone of damage just below the
crest and well defined presplit half barrels. The crest damage (surface dilation) zone varies and is
added to by subsequent adjacent perimeter blasting operations. Keeping the preconditioning to
a minimum is possible using the methods described in this paper, developed on several diverse
and geographically distant mining operations to control crest dilation and preconditioning by
presplitting in advance - not just laterally but vertically in advance.
The first documented case (Delbridge, Marton and McSweeney, 2004) in the development of
the concept was undertaken at AngloGold Australia Pty Ltds Sunrise Dam Gold Mine (SDGM),
located beside Lake Carey, some 730 km northeast of Perth, Western Australia. The mine had been
developed in a series of cutbacks to a depth of over 300 m and began underground mining during
the latter stages of open cut operations. This mine is fairly unique in that it utilised a single pass
vertical presplit which initially extended over three bench heights and after the failure mechanisms
were understood, over four benches. The presplit drilling for the three stacked benches started on
the floor of the second bench so that any dilation was confined to the surface one bench above the
crest and modified perimeter blasting was then developed to curtail damage from blasting adjacent
to the new crest.
The second case is based on a South African coal mine in the Witbank Coal Measures, 140 km
northeast of Johannesburg, Republic of South Africa (MA). It was important that blasting for
optimal wall stability and minimal scaling was developed in advance. A dragline is scheduled to
excavate the #2 seam midburden which is the lowest of the economically viable coal seams in the
five seam Witbank Coal Measures. Mining of the overburden, coal and partings above the dragline
midburden pass are by excavator and truck and each mining layer is individuallyblasted. The results
of presplitting the midburden through the overlying coal seams and partings have provided a presplit
with no surface dilation of the midburden crest and ensures any subsequent crest loss is solely due
to blasting adjacent to the vertical presplit. Typical of all coal mines in South Africa only vertical
presplits are used.
Wall control is most commonly delivered by presplit blasting and one common sight after the
dust clears is a ridge of broken blocks along the presplit row, usually with a main crack running
between the presplit blastholes and often parallel cracks either side of the perimeter line. The
surface rock strata or block structures have been dilated by the venting gases and shock waves.
Decoupled, fully coupled, single or double deck charges are used depending on a variety of factors,
anecdotal practices and perceived benefits. Preconditioning of the crest rock by blasting from
above may not be visible to the naked eye (or non-existent in overburden in strip coal mines) but
the high speed venting of explosives gases (and often the water in the presplits) preferentially
follow existing cracks and structures because there is no developed presplit crack near the surface.
This nearly always lifts, shifts and separates the surface structures and geological layers.
Stopping the dilation by sacrificialor in advance presplitting has been shown to reduce the
crest loss at both metal and coal mines. Drilling and presplit blasting through an upper level of
bench or layer provides a degree of restraint to the target layer where the crest must be maintained
and any damage caused by the venting gases will not cause concern as it is removed as planned,
leaving a stable crest.
INTRODUCTION
While coal mines and metal mines have vastly different lost crests have a major impact on drilling, blasting and pit
geological origins and structures the use of presplitting is limits, and on the safety of anyone or anything working below
prevalent in wall control for safety and mining economics. such walls. The importance of crest control in coal mines with
All blasting practitioners therefore strive to meet the criteria walls up to 60 m high is self evident. Small rocks falling from
of clean, stable walls, crests and toes to design, and catch that height can be even more damaging than the large blocks
benches (where applicable) that effectively work. Loose or sliding out of a 15 m face in a hard rock mine.
1. MAuslMM, Consulting Mining Engineer, Marton Mining EngineeringServicesPty Ltd, 318 The Esplanade, Speers Point NSW 2284. Email: alxmrtn@bigpond.net.au
A MARTON
Understanding the mechanisms of block movement and

managing the preconditioning of the crests at Sunrise Dam
Gold Mine (SDGM) was critical in developing a solution. Once
the blocks vibrate and shift the cohesive strength is markedly
reduced and very little can be done to maintain the stability of
the crests. The size and weight of the overlying matrix of blocks
above the crest area plays an important part in stopping the
movement even though the vibrations from blasting may still
be happening. Blasting adjacent to blocks that already have
a very low factor of safety due to the preconditioning from
presplitting simply frees the blocks or sets the factor of safety
(FoS) to less than one.
Understanding the mechanism of crest loss in laminated
deposits in coal mines is necessary to reach the objective
of safe and stable high walls. In coal mines the strata has
relatively low cohesion between the layers so blasting will
cause dilation and loss of cohesion due to shock loading and
gas penetration of the weaker bands.
I
FIG 3 -Shows the presplit wall after the main blast but also shows evidence of
This is quite noticeable under high speed filming and in
many post blast results at the surface after presplitting. The crest loss.
following set of photos highlight the starting points in the
development of the two crest retention programs at their
respective mines.
Figures 1 - 4 are of the coal mine and Figures 5 - 8 are of
the gold mine. Figure 1 shows typical presplitting surface
expression at the coal mine. Figure 2 is a close-up. The crack
between holes cannot be distinguished. Figures 3 and 4 show
the presplit wall after the main blast however the crest loss
is also evident. Figure 5 shows the typical post blast surface
expression at the gold mine. Figure 6 shows a close-up of the
I
FIG 4 - Shows the presplit wall after the main blast but also shows evidence of
crest loss.
I
FIG 1 - Shows typical surface expression at the coal mine.
FIG 5 -Shows typical post blast surface expression at the gold mine.
crest and the unacceptable back cracking. Figure 7 shows the

wall condition after trenching while Figure 8 shows the final
wall after all the benches were removed. These wall conditions
are not acceptable. The aim is to find a method of blasting to
stop the crest loss.
As rock is blasted the volume increases, so for a presplit to
I be formed there must be a degree of expansion of the rock.
Generally the only free face when presplits are blasted is
FIG 2 -Shows a close-up but the crack between holes cannot be distinguished. the surface and therefore there must be expansion in that
MINIMISING CREST LOSS BY ELIMINATING SURFACE DILATION DUET0 PRESPLITTING
FIG 6 -Shows a close-up showing clear crack formation but unacceptable back FIG 9 -Showing a full bench rotated along an existing structure.
cracking.
I
FIG 7 -Shows the wall post trenching to expose the crest.
A
FIG 10 -Showing the Sunrise Dam Gold Mine triple bench configurationand
the potential sources of crest damage (including from the bench above).
(1) Holes directly over or into the new crest; (2) holes within one burden of the
new crest; (3) holes intersectingunfavourably orientedjoint planes or forming
connecting cracks; (4) presplit holes venting into joints or the preconditioned
crest area; (5) batter and buffer holes directly opposite the crest.
controlled hard rock operation. In each case the resulting
dilation of the surface leads to loss of crest. The SDGM triple
bench configuration is shown with the potential sources of
crest damage from blasting in and against the catch bench.
In the case of a laminated deposit as shown in Figure 11,
the pressures from the explosives gases and any moisture in
the blasthole will tend to separate the layers. Each layer will
snap off at some point and the layers will separate. The more
I weight holding the layers together or thicker the layer the less
the dilation of those layers. In Figure 11 the blasthole on the
FIG 8 -Shows the typical final results of multi-bench presplits.
left is shown as drilled from the top of the interburden while
direction. In some cases where there is basal plane with low that on the right is drilled from the top of the upper coal seam.
resistance, such as a coal seam, or open joint structures, it has The dilation of the upper layers is going to be dependent on
been noted that the full block moves. Figure 9 is an example of the thickness, strength and weight of material in the crest area
the full bench rotated along such an existing structure. and the 5.6 m coal seam will be quite well broken however
When such planes do not exist or the weakest link is the uppermost shale and sandstone interburden layers will be
near the presplits, the damage is concentrated around the cracked but hopefully constrained. Being a flat lying deposit
presplit line. The separation is determined by the explosive there is little chance of sliding movement.
energy, the degree of loss of cohesion of individual layers, the Geotechnically speaking, block movement results if the
degree of movement/rotation of blocks and the weight of the cohesion and frictional resistance is reduced below a FoS of
blocks. Figure 10 shows a stylised expansion in a structurally one. If the cohesion between blocks is lost but rotation does
A MARTON
FIG 11-Shows the stylised expansion of a laminated deposit. The blasthole on the left is shown as drilled from the top of the interburden while that on the right is
drilled from the top of the upper coal seam.
not occur the blocks may still remain in place due to the In coal mines the laminated geology consisting of varying
resistance to sliding from the surface asperity. rock types and strengths tends to exhibit macro and micro
The objective therefore is to minimise dilation and rotation jointing due to regional shifts and stresses. The interburden
surface is less likely to be affected by subdrill but the larger
of the surface blocks or layers.
volume of explosives gases (and water) are more likely to
dilate the much weaker laminated structures and break up
Determining controls for dilation the surface layers into slabs which have to be cleaned off to
Trials with smaller diameter holes, smaller charges, decking make the wall safe. (Blasting the coal may cause some minor
and/or decoupled charges, explosives types, initiation sub surface damage.) Blasting in such geology usually results
timing, reduced and increased spacings were conducted in distinctly slabby block fragmentation. The rock strength
along conventional presplit improvement lines. In all cases is orders of magnitude higher than the cohesive strength of
the post blast analysis found that the presplit quality was the layers leading to delamination before fragmentation.
poor when the charges were reduced to eliminate the surface Overlying material will tend to resist the delamination of such
dilation. Decking sometimes produced a better result but not layers due to its weight in much the same way as was found in
if the charges were reduced. Reduced spacing and reduced the SDGM case. If the dilation does not occur, nor rotation of
charges lead to drilling, charging (and cost) issues and did the blocks which reduce the sliding resistance, then the crest
not markedly show much improvement in crack quality. blocks will remain in situ.
The lowest charges failed to form cracks visible in the sides Based on the previous work a set of presplitting trials were
although the surface was not disrupted. set up in the coal mine to gauge the effect of blasting with the
The use of video cameras at SDGM showed the block coal bench acting as a deadweight to prevent dilation of the
movement resulting from the presplit blasting over three surface layers.
benches. The solution found in drilling through an additional The trials also compared total charge against decked charges
bench provided a presplit without block movement ... and against placement of decks. Figures 12,13and 14 show the
eliminatingtwo of the mechanisms allowing crest loss through various combinations. As the sandstone above and below the
dilation. The extra bench provided substantial weight, and #3 seam is generally much harder than the shale/mudstone
dilation of the surface was confined to that bench. Its surface the placement of charges against the harder sections was
had been preconditioned by subdrilling and blasting from implemented as shown in Figures 12,13 and 14. Geologging
the bench above but the explosives gases venting along of the interburden enables charge placement design based
the extended presplit were not likely to see any subdrilling on rock strength and coal seam positions. Density logs were
preconditioning more than one bench height below the collar. used to underpin design. The traces have been superimposed
By reducing the subdrill over the new crest and having the on the photos of the exposed highwall. Charges were AN/
presplit already in place when the bench sitting over the crest CN emulsions with and without AN prill in 127 mm diameter
was being blasted provided a reflective plane acting to protect sleeves in 251 mm (generally wet) blastholes, primed with
the new crest. This left only the damage from trim blast holes 400 g boosters and 10 g/m detonating cord downlines, fired
adjacent to the crest to be controlled. in sets of 20 holes.
MINIMISING CREST LOSS BY ELIMINATING SURFACE DILATION DUET0 PRESPLITTING
FIG 12 - Showing design criteria for trials blasting through #4 seam -variable placement, constant charge.
FIG 13 -Showing design criteria for blasting interburden - placement and charge variations.
The split factor (charge/m*) was the basic rule. The and continuing on with one design can be dangerous. To
interburden vanes from 18 to 22 m and the blastholes are enable adaptation based on results the criteria adopted was
always drilled through to the bottom of # 2 coal seam (-6 m based on the presence of cracks and the surface expression.
thick) to ensure the coal is able to be cleaned back to a straight Subsequently it has been possible to review the results after
highwall line. the interburden removal by the dragline. Figures i,15 and 16
show the difference in surface expression between presplitting
Monitoring blast results for the coal mine case through coal or not, but not the effect of charge placement nor
from different charges. (No difference has been noticed for
The initial data capture consisted of video, surface expression the 10 kg difference in charge nor from the placement.) The
measurements (where applicable) and observations with removal of the four seam coal also provided an insight into the
respect to crack visibility along the presplit hole. success (or otherwise) of blasting through the coal to keep the
Presplits are usually only deemed to be successful after crest intact. The surface after the coal was removed was swept
full exposure of the wall. However the adjacent blasts usually with a mobile broom and the holes surveyed to determine the
have a very significant impact on the appearance of the position of the presplits and then correlate back to the charges
exposed walls. The exposure is often months in the future used. The surface cracks are similar for both single charges
A MARTON
FIG 14 - Showing design criteria for blasting interburden - charge/placement variation.
FIG 15 - Shows the surfacecracks left post blast through the coal seam. FIG 16 -As in Figure 15, this figure shows the surface cracks left post blast
through the coal seam.
and split decks of the same total charge. However the surface
expression/dilation is markedly different! Where there was no
coal the surface disruption is up to 2.5 m (see Figure I). As can
be seen in Figure 17 the surface is intact and a crack is clearly
visible when the presplits were blasted through the coal.
To ensure the effect of adjacent holes (interburden infill/
main pattern) were not masking the result, the trial was
extended to cover the variability of the various interburden
designs, eg a section of 55 kg/hole, single charge was followed
by a 15 + 30 kg/hole double deck charge versus a 60 kg single
deck, etc. Each section was straddled by the same interburden
design. The results of the interburden blasting post presplitting
is entirely dependent on drilling accuracy and charging. All
other aspects could be controlled and managed.
Subsequent blasting results for the main blast were seriously
affected by the accuracy of the drilling and severe rainfall
events. The mispositioning of batter row blastholes can ruin
a final wall. Too much burden is arguably worse than too little m
as not being able to clean the wall back to the presplit can FIG 17 - Post-excavationof the coal.
MINIMISING CREST LOSS BY ELIMINATING SURFACE DILATION DUE TO PRESPLITTING
I
FIG 18 - Blasting sequence for final wall control at Sunrise Dam Gold Mine.
leave dangerous hang-ups. Nor can the quality of the presplit CONCLUSlONS
be evident if the wall cannot be cleaned off. This occurred
In both the hard rock and coal mines the damage from
in many parts of the trial zone. Redrilling due to backfilling
presplitting were shown to be detrimental to the crest and
from heavy rains and poor drilling practices masked several
sections of the trials. The presplit results where the main blast a similar solution improved the condition of the crests. The
blastholes were unaffected showed sound and stable walls and weight of rock or coal resting over the crest contributed to
sharp crests. confining structures or crest blocks and thereby stopping loss
of cohesion and rotation which reduces sliding resistance.
Blast results for the hard rock case Drilling and blasting through the four seam coal clearly
After analysing the results from the range of single bench produced the best results; no dilation yet a visible crack
and multi-bench preplits and blasts plus modifications to along the length of the presplit hole plus a discernable crack
designs of the trim blasts the best results occurred when between blastholes. The results are the first step to a stable
broken material was left in situ on the catch bench. The wall. The impact of the batter row and other blastholes on
three stackedbenches are presplit in one 23.5 m pass and the wall are still to be fully controlled. The dilation of the
each 7.5 m bench is then blasted individually. Further trials interburden surface was not evident. Cracks between the
were then undertaken to drill the presplits in advance using presplit blastholes were clearly discernable after the coal was
two 16 m passes. The sequence shown in Figure 18 involved removed and could be traced down the blastholes.
drilling and blasting the presplit for the new bench from one
The blastholes where the top deck was placed above the
bench above the crest (Figure 18a). This meant the presplit number three seam showed better cracking to the surface.
length decreased from 23.5 m to 16 m but the top 7.5 m of The larger single deck charged blastholes tended to show
presplit was wasted as it would be blasted out with the final more lateral displacement of the upper layers while the lesser
bench blast - a sacrificial presplit. Each 7.5 m bench was charges did not show clear crack formation between the
then blasted individually. Each blast had certain criteria to presplits at the surface.
control subgrade damage over the new crest (Figure 18b), or Not all the presplitting can be carried out through the coal
backbreak damage against the new crest (Figure 1812).Finally due to coal mining requirements but the benefits of drilling
the material over the catch bench was excavated standing on through the coal where possible have been shown in terms of
the lower bench adjacent to the new crest (Figure i8d). The better drilling surfaces, less blast delays, less mining delays
sacrificial presplitting proved to be the most important factor. (waiting for the interburden to be exposed for presplitting
The second important factor was leaving the blasted bench then cleaning up the damaged surface and drilling the
over the new catch bench unexcavated. If this deadweight was main pattern) and most importantly the surface dilation is
removed the crest block movement recurred when blasting eliminated.
the bench adjacent to the crest.
In the hard rock situation the sacrificial cost of drilling an
Findings from previous work to reduce backbreak and
vibration at a mine in northern Queensland were also extra 7.5 m of presplit hole far outweighed the cost of secondary
introduced to reduce damage to the crests when blasting the breakage of large wedges which were pushed out of the crest
bench adjacent to the crest. Seismograph measurements at and the cost of moving the blocks aside while cleaning the
the wall with a broken rockmass between the blast and wall walls. Drilling and blasting presplits in advance (on the upper
show a low sonic velocity and reflection of the shock waves bench) also had operational benefits as there were no trim
in the horizontal and traverse directions. By masking the wall pattern blastholes near the damaged area around the top of
from the larger charges with a narrow zone of broken ground the presplits. The damage zone around the presplits adjacent
between the presplit and the production charges it is possible to the crest was much smaller and the batter row holes could
to further reduce wall damage. This method has been adapted, be effectively collared and drilled.
initiating the trim patterns from the third row off the wall,
with a delay offset on the fourth row. This effectivelyformed a REFERENCES
zone of broken ground in advance of the large holes.
Delbridge, D, Marton, A and McSweeney, K, 2004. Final pit wall
Comparison blasts using free face and masked firing stability - Understanding crest loss mechanisms and possible
techniquesproved the choked blasts did not produce increased solutions to block movement, in Proceedings Explo 2004,
backbreak while vibration levels at the wall are lower than for pp 181-186 (The Australasian Institute of Mining and Metallurgy:
unshielded firing sequences. Melbourne).
Ensuring Professional Engineering
Capability - Queensland Professional
Engineers Act 2002
ETaylor and C Murray2
ABSTRACT
The provision of professional engineering services in Queensland is regulated through the
Professional EngineersAct 2002. The Act requires that anyone providing professional engineering
services must be registered and is subject to continuing professional development requirements. The
engineering profession is actively requesting the Queensland Government to expand registration
to include associate and technologist engineering practitioners. In view of the strong moves at the
Federal level towards national registration of professional groups the lessons of the Queensland
experience are very relevant across Australia. This paper describes the rationale behind the Act
and outlines its operation. It discusses some misconceptions regarding the applicability of the
Act to various engineering activities. There is much that the Act has to offer individual engineers
and industry generally in establishing, articulating and maintaining standards for professional
engineering work. The paper considers legal actions taken to date under the provisions of the Act
and discusses the way in which this parallels industry moves towards quality assurance of work
force capability.
INTRODUCTION
Building a better industry includes consideration regarding need to be maintained to ensure the protection of the public,
who should regulate the professionals who work within and the professionalism of engineers. On the other hand,
that industry. The regulation of professional groups such over regulation, or inappropriately restrictive standards, can
as doctors, lawyers and nurses through legislation is well impact negatively on workplace productivity. Within any
established and professionals in these disciplines do not think regulation scheme it is important to establish formal dialogue
twice about ensuring that their registration, their licence to between the professional and industry organisations and the
practice, is in place before they walk into the ward or the regulation body. As the Queensland scheme is undertaken
courtroom. The situation for professional engineers is not so within a co-regulatory arrangement with the profession it is
clear. In all states except Queensland regulation is variable, agile in responding to changing professional norms regarding
practice. The profession encompasses individual practitioners,
often focusing on engineers working within the building
professional associations, such as The AusIMM, and industry
sector and/or requirements for chartered membership within organisations, such as the Queensland Resources Council.
Engineers Australia.
Queensland has had a comprehensive legislation based THE PROFESSIONAL REGULATION
regulation system for professional engineering services since
1929. This has had significant benefits for the engineering ENVIRONMENT IN AUSTRALIA
profession and industry as it has provided an independent, The regulation of professionals and professional services is one
robust regime to uphold the standards of engineering dimension of the complex relationship between professionals
practice and maintain public confidence in the professional and the community. The expectation that professionals will
engineering services being provided. put community/client interest first is stated explicitly through
the Codes of Ethics of professional organisations and the
Scarcity of skilled personnel, particularly in highly technical Hippocratic Oath, and is implicitly assumed by the public to
areas such as engineering, continues to be an issue confronting be the basis of professional action. Community and business
industry. Pressure to reduce professional standards to allow expectations regarding the competence and accountability of
greater access for less credentialed personnel can build. professionals are very high and this is increasingly reflected
While reducing standards can appear to be an attractive in regulation schemes and the larger role being played by
option, it poses a significant threat to industry capacity to government in this arena.
successfully manage the complex operational, environmental
and health and safety challenges facing it. Disasters caused by The regulation of any professional group consists of the
technical failure are expensive, in human, environmental and following elements:
commercial terms. However regulation must also be about registration,
finding an appropriate balance. On the one hand standards compliance,
1. Chair, Board of Professional Engineers of Queensland,TheMansions: 40 George Street, Brisbane Qld 4OOO. Email: Admin@bped.qld.gov.au
2. Registrar, Board of Professional Engineers of Queensland,The Mansions: 40 George Street, Brisbane Qld 4OOO. Email: tlare.murray@bpeq.qld.gov.au
SEVENTH LARGE OPEN PIT M I N I N G CONFERENCE / PERTH, WA, 27 - 28 JULY 2010 65

ETAYLOR AND C MURRAY
0 disciplinary regime, and of registration to create a single national registration and

0 an appeal process. accreditation system for ten health professions.
All elements are required for the regulation to be effective The new arrangement will help health professionals
and are the basic components of all regulation systems, both move around the country more easily, reduce red
of professional associations and those based on legislation. tape, provide greater safeguards for the public and
Professional regulation ranges from: promote a more flexible, responsive and sustainable
use of professional associations regulation, for example, health workforce. For example, the new scheme will
engineering in most states; maintain a public national register for each health
profession that will ensure that a professional who
partial legislative regulation, for example, engineering
has been banned from practising in one place is
services provided within the building industry in Victoria;
unable to practice elsewhere in Australia (Australias
and
Health Workforce Online, 2010).
full legislative regulation for example, medical
practitioners, lawyers and nurses across Australia and In late 2009, the Australian Health Workforce Ministerial
professional engineering services in Queensland. Council announced the appointment of the 108 inaugural
members of the ten national boards for the new National
Within these systems, legislation can be used to a greater Registration and Accreditation Scheme for the Health
or lesser extent in combination with existing professional Professions under the Health Practitioner Regulation
association regimes. So, for example, tenders for government (AdministrativeArrangements) National Law Act 2008. The
infrastructure developments and local government work often appointments enable the national boards to begin preparatory
require engineering work to be signed off by a person with work for the commencement of the new national scheme on
Engineers Australia Chartered Professional Engineer status. 1July 2010.
There is no legislative imperative underpinning such a tender
condition. It is used as a mechanism to ensure professional Similar moves are occurring with law as part of the COAG
competence. National Legal Profession Reform agenda (Australian
Government, 2010).
While the discussion paper Reforming the Regulation of
the Professions issued by the National Competition Council
in May 2001 recognised that the primary objective of PROFESSIONAL REGULATION AND
regulating a profession is to protect the welfare of consumers ENGINEERING
of professional services and to protect the public... (The At present there is no COAG agenda for the Engineering
State of Queensland, 2002, pi), there are always questions Profession. However,there are moves, particularlyfrom within
regarding the efficacy of professional regulation. Regulation, the engineering profession, to put national regulation on the
and particularly self-regulation, of a professional group brings agenda. A number of submissions on registration, such as that
with it negative possibilities such as professional monopoly, from Consult Australia (formerlyACEA) to Minister Tanner in
restraint of trade, and cronyism. groupthink. These can March 2008, have been made to relevant Ministers, both State
have serious negative impacts on public safety and effective and Federal. Despite the active engagement of professional
workforce capability planning within industry. engineering associationsin discussions on regulation it would
To address these concerns, professionals are not entirely seem that, unlike the health and legal profession, there is little
free to self regulate and the extent of the control a profession engagement by practicing engineering professionals with the
is allowed to exercise over its own activities is at least partly topic. For engineering,there are still many issues surrounding
a political question, decided through a political process. regulation and the appropriateness of possible regulation
(Johnston, Gostelow and Jones, 1999, p 539). In Australia, models to be debated and clarified.
the issue is more complex as the history of separate state The word public appears to be misunderstood, and
development followed by Federation has led to a diverse array to be read literally as meaning people generally in the
of professional regulatory systems. community and that therefore regulation is only applicable
A submission to COAG from the Australian Council of for those who deal directly with the public and/or whose
Design Professionals notes: work is commissioned directly by members of the public.
Consequently many engineers, particularly those working in
Currently the systems of idenhfying those qualijied areas such as electrical and mining where often there is no
to provide these professional services range from direct relationship with what they have previously perceived
statutory systems, administered by State and to be the public, assume that protection of the public and the
Territory governments agencies, to association requirement for regulation is not applicable to them. However,
maintained selfregulatory registers. legal advice from Queensland Crown Law is clear that the
I n other words there is a plethora of registration public encompasses work colleagues, clients and anyone
systems that are confusing to consumers, costly to the who can be impacted by the professional engineering services
professions and inefficient to administer. In terms of being provided. The work of all professional engineers impact
global markets these systems do not readily facilitate on public health, wellbeing and safety. As a 2002 Report of
the identification of Australiandesign professionals the Royal Academy of Engineering notes:
(BEDP, 2010). Engineering is first and foremost an activity that
interacts deeply with society. It requires a profound
Overall the general consensus is that there is public benefit
understanding of societys needs and aspirations
to professional regulation, that regulation is most effective
and an ability to communicate and debate with the
and efficient if driven by the profession, and that it needs to community how best to meet those needs. It is an
be developed at the national level. Government moves in this economic activity concerned with the optimum use of
direction have already commenced in a number of professions. scarce resources. It is a cultural activity that impacts
For the health workforce, the Council of Australian on lifestyle and behaviour. It does all this within a
Governments (COAG) has already moved to implementation disciplined framework able to deliver the devices
of a national system. In March 2008, COAG signed an and artefacts best suited to meet societys needs (The
Intergovernmental Agreement setting up a new system Warren Centre Report 2009, p 6).
ENSURING PROFESSIONAL ENGINEERING CAPABILITY - QUEENSLAND PROFESSlONAL ENGlNEERSACT2002
So the publicencompasses everyone, in an inclusive sense, perspective to discussions regarding engineering. Together
who will be impacted by the provision of the professional they have built a very comprehensive understanding of
engineering services. Thus it is important for professional professional engineering and its work practices and standards
engineers to recognise and accept that the definition of the requirements to the national and international debates.
public encompasses work colleagues on site as well as the Within a co-regulatory model of regulation they each play an
residents in the town located nearby. important role.
Yet when discussingthe application and value of professional Within the profession, there appears to be gathering
regulation, the misunderstanding of definition of publicdoes recognition of the value of regulation to support the public
lead some to suggest that engineering regulation should be identity of and standing of professional groups. Western
limited to areas of direct public engagement, for example Australia is in the process of considering implementation
including civil engineering within building and construction. of regulation and recent surveys undertaken by Engineers
Another suggestion has been that regulation could be Australia suggest that there is strong support for regulation
structured along industry sector lines, rather than keeping among engineering practitioners.
all engineering within a single portfolio. Both scenarios pose
significant risk to the engineering profession. PROFESSIONAL REGULATION OF
Due to its complex team and corporation based work ENGINEERING IN QUEENSLAND
arrangements, professional engineering already struggles Through the Professional Engineers Act, the state of
with the visibility of its professional persona within the Queensland has registered professional engineering services,
community. If professional engineering regulation schemes across all areas of engineering, since 1929. Because the
are scattered among different industry sectors, the profession term engineer was widely used it was decided that it would
will find it progressively more difficult to articulate a single impossible to restrict its and so the provision of professional
professional identity and retain professional decision-making engineering services is the subject of restriction.
autonomy differentiated from the trade and technical sphere.
During the 1990s there was extensive dialogue with the
One example can be found within the building sector. profession and community regarding continuing regulation.
Professional engineering practitioners, such as fire safety In 1992 the Heads of Government were considering
engineers,are increasinglyrequiredto operate undertechnician mutual recognition principles for goods and occupations
licences with their expertise determined by non-professionals. throughout Australia. In 1993 a Federal Vocational Education
A subtle loss of autonomy in peer assessment of professional Employment and Training Working Party recommended,
competence is underway. Within fire engineering practice, among other things, that the:
prescriptive codes and standards drive decision-making
engineering profession be deregistered in Queensland
and if professional judgement informs design the process is
on the basis that Queensland was the only State in
characterised as alternative solutions. A parallel scenario in
Australia that provided a formal registration system
medicine would be for all medical discussions and procedures
for professional engineers and deregistration did not
to be documented as prescriptive codes and for any deviation
pose a risk to public health and safety (Queensland
to be viewed as an alternative solution which the doctor is
Government, 2005,p 11).
required to justify. This characterisation of professional
decision-making has major implications for quality assurance The Queensland response to this recommendation was
for professional engineering expertise, for innovation and risk contained in a May 2005 discussion paper Registration
management in design and construction, and for social and of Professional Engineers released by the Administrative
economic development. For instance, one consequence would Services Department. The discussion paper also included the
be a further increase in the documentation and compliance outcomes of a concurrent Queensland Government review
load for organisations as they replaced professional decision of all legislation with a significant effect on the business
making with prescriptive code. It requires further discussion community. The discussion paper concluded that, despite the
within the profession and industry. Working Party recommendation, registration should continue
as:
Medical and legal practitioners have not been splintered
along industry lines and have therefore been able to retain it was essential to have some form of regulation of
a strong, readily identifiable professional persona. Their professional engineers by statutory means, to provide
strong professional identity continues despite the fact a higher degree of consumer protection than that
that, increasingly, these professions are found in corporate currently available ... (p 2)
environments similar to those in which engineering is It was further noted that other States were introducing
practiced. various forms of registration of engineering subsequent to the
However this is not to suggest that there should be only one Working Party recommendations.
organisationor organisational model representing engineering. The review period included extensive dialogue between
Presently in Australia, the profession is covered by a number the profession, community and government, analysis of
of complementary professional organisations. Engineers the implications of National Competition Policy and public
Australia (EA) is a national forum for the advancement of benefit assessment. It was further extended to encompass
engineering and the professional development of members, the 2001 National Competition Council paper Reforming the
The Australasian Institute for Mining and Metallurgy (The Regulation of the Professions noted earlier, and ensure that
AusIMM) represents the interests of professionals associated
the Queensland Governments obligations under National
in all facets of the mining, exploration and minerals
Competition Policy (NCP) were met.
processing industries, Consult Australia (previously known
as ACEA), is the peak industry body representing consulting The outcome of the extended review was to reaffirm the
companies that provide professional services to the built and decision to continue with regulation of the profession. It
natural environment, and the Association of Professional recommended some major changes to the existing format
Engineers, Scientists and Managers Australia (APESMA) and removal of anti-competitive elements which could not be
represents professional employees. Each brings a different justified on public interest grounds.
ETAYLOR A N D C MURRAY
The rationale underpinning the 2001 decision to continue option 4 - small net cost - would be likely to result in a
government regulation of the profession is detailed on reduced risk for consumers within the building industry,
page 2 of the Professional Engineers Bill 2002 Explanatory but this would occur at the expense of consumers in
Notes. The decision other industries and may result in confusion over the
recognised a need for regulation (or continuing appropriate complaints process (p 4).
regulation) where information asymmetry exists Option 3 was chosen. The Professional Engineers Act 2002
- that is, where consumers are not reasonably able commenced operation on 1 January 2003.
to inform themselves suflciently about the skills of
diflerent practitioners. OUTLINE OF THE OPERATION OF THE
The provision of professional engineeringservices is an area PROFESSlONAL ENGlNEERS ACT2002 (THE
where information asymmetry exists as a key characteristic ACT)
for purchasers
The objects of the Act are set out in section 3:
due to the technical nature of services provided
relative to the general level of knowledge of a. to protect the public by ensuring professional engineering
consumers, and an adverse timing dimension where services are provided by a registered professional
potential problems may not emerge until well after engineer in a professional and competent way,
the service is provided... b. to maintain public confidence in the standard of services
provided by registered professional engineers, and
Regulation provides a means of distinguishing c. to uphold the standards of practice of registered
those persons who have achieved competency in the professional engineers.
provision of professional engineering services. The Act achieves the objects primarily by:
Meeting obligations under the National Competition Policy
providing for the registration of individuals as registered
was achieved
professional engineers under this Act,
by removing existing legislative requirements for providing for the monitoring and enforcement of
the regisiration of engineering companies and compliance with this Act,
engineering units, thereby freeing u p the manner by imposing obligations on persons about the practice of
which the business of delivery of engineering services engineering, and
may be constructed(Professiona1Engineers Bill 2002
establishing the Board of Professional Engineers of
Explanatory Notes p 3).
Queensland (s4).
A number of alternatives were considered the structure of Implementation of the co-regulatory approach to (a)
regulation and each was subject to the public benefit test. The
the registration of individuals was difficult under the
alternativeswere:
original structure of the 2002 Act. It required the Board
option 1 - deregulation; of Professional Engineers of Queensland to develop the
option 2 - State Government regulation of engineers only professional competency standards against which the regimes
in the building industry; of professional organisations would be judged for suitability
option 3 - co-regulation, with accredited professional as Assessment Entities. This format took ownership for
engineering organisations taking responsibility for establishing what constitutes appropriate professional
assessment, and government accrediting professional standards away from the profession. This was unintended and
organisations and conducting any disciplinary action; and inconsistent with the co-regulatory approach. In addition the
option 4 - co-regulation as in option 3, but involving a cost to register members of developing and maintaining such
board structure that governs all professions involved in competency standards apart from professional organisations
the building design and planning industry. would have been prohibitive.
The rationale behind option 2 and the argument often used
The Act was amended in 2008 to address this issue. The
to limit the breadth of engineering regulation was that the
Professional Engineers and Other Amendments Legislation
asymmetry of expertise is most acute in the building and,
2008 Act provided for professional engineering organisations
particularly domestic building, arena and therefore it should
to apply to be Assessment Entities for Registration. The
be the only area subject to government backed regulation.
process is similar to that used to determine the suitability
At that point the option of State Government of schemes submitted for approval under Professional
regulation of engineers only in the building Standards Acts, now in place in New South Wales, Western
industry was determined to be unsuitable as it did Australia and Queensland. A set of criteria in s112 of the
not achieve the policy objective ... in regard to the Act concentrates on the capability and standing of the
practice of engineering in industries other than the professional organisation submitting an application, rather
building industry (Professional Engineers Bill 2002
than determining the value of the professional standards that
Explanatory Notes p 3).
will be used by that professional organisation to determine
This is consistent with the inclusive definition of the individual professional capacity. The latter is seen to be the
term the public. jurisdiction of the professionals within that professional
The conclusions reached were that: organisation. This change means that the professional bodies,
option 1 - large net cost and does not meet the objectives rather than the Board, determine the competency standards
of the legislation; required of individuals seeks registration. At present EA,
option 3 - a small to moderate net benefit; largely meeting The AusIMM and the Institution of Fire Engineers (IFE),
objectives through similar mechanised to base case, Queensland Group are approved Assessment Entities.
but with improved assessment of competency for initial The Act is based on prescribing who can undertake
and ongoing registration, thereby enhancing consumer professional engineering services which are defined in
protection; and schedule 2 as meaning:
ENSURING PROFESSIONAL ENGINEERING CAPABILITY - QUEENSLANDPROFESSlONALENGlNEERSACT2002
an engineering service that requires, or is based a. conduct that is of a lesser standard than that which might
on, the application of engineering principles and reasonably be expected of the registered professional
data to a design, or to a construction or production engineer by the public or the engineers professional peers;
activity, relating to engineering, and does not include b. conduct that demonstrates incompetence, or a lack
a n engineering service that is provided only in of adequate knowledge, skill, judgment or care, in the
accordance with a prescriptive standard. practice of engineering;
A prescriptive standard means a document that states c. misconduct in a professional respect;
procedures or criteria: d. fraudulent or dishonest behaviour in the practice of
for carrying out a design, or a construction or production engineering; and
activity, relating to engineering; and e. other improper or unethical conduct.
the application of which, to the carrying out of the design, Disciplinary proceedings were held in the Commercial and
or the construction or production activity,does not require Consumer Tribunal (CCT) until 2009, when it was replaced
advanced scientificallybased calculations. by the Queensland Civil and Administrative Tribunal
Only Registered Professional Engineers can provide (QCAT). Providing professional engineering services while
professional engineering services s115 (I). A person must unregistered or unsupervised is an offence under the Act and
provide professional engineering services within their area the matter is heard before a Magistrate. Some examples of
of competence s115 (3). However, a person does not commit disciplinary proceedings are given in Table 1.
an offence if they provide professional engineering services or The final operational tool of the Act is the Code of Practice.
services outside their area of competence if they are under the The objectives of the Code are to:
direct supervision of a Registered Professional Engineer s115 fulfil the requirements of section 108(1)of the Professional
( 2 ) and (4). S115 (5) states:
Engineers Act 2002, which states that the Board of
a person carries out professional engineering Professional Engineers of Queensland must make a code
services under the direct supervision of a registered of practice to provide guidance to registered professional
professional engineer only if the engineer directs the engineers as to appropriate professional conduct or
person in the carrying out of the services and oversees practice;
and evaluates the carrying out of the services by the set out the minimum levels of professional conduct
person. required to be met by registered professional engineers
The operational reality of the Act is that the main drivers are in Queensland so that they can carry out all their roles
the objects of the Act. When a disciplinaryproceeding is being within a framework of integrity, care for the public, and
considered the specific facts of the case and the professional competency; and
behaviour of the individualas presented are consideredwithin assist the Commercial and Consumer Tribunal in
the context of the objects. In alignment with the objects of the determining whether the registered professional engineer
Act, disciplinary proceedings against Registered Professional has behaved in a way that constitutes unsatisfactory
Engineers of Queensland are: professional conduct, or practice (Professional Engineers
a. unsatisfactory professional conduct, Act 2002, s129).
b. failed to comply with a provision of this Act, The approved Code is admissible as evidence in disciplinary
convicted of an offence related to the practice of proceeding brought by the Board against a registered
professional engineer. It was developed collaboratively with
engineering,
all the major professional engineering bodies and a number
c. contravened an undertaking entered into by the engineer of interested professional engineers through a series of
and the board under section 73(z)(b), and workshops and meetings. It must be reviewed every couple
d. contravened a condition of the engineers registration. of years with the involvement of the profession. Even when
Unsatisfactory professional conduct is defined as: the code is not due for review, the Board maintains frequent
TABLE 1
Sample of prosecutions under the Act.
Year of decision Actions leadingto disciplinary proceedings Outcome

Magistrate
2004 I Holdinq out that RPEQ I Fine $75 OOO,Costs $21 000
Commercialand Consumer Tribunal (or aqreement)
2003 Inadequatedesign and inspection -collapse of a large retainingwall Reprimand
ED003-06 Penalty $1500
Boards costs - District Court Scale (DCS)
Deregistration
Penalty $3000, Boards costs - DCS
Certification without inspection Reprimand

Penaltv $2000, Boards costs - DCS
ETAYLOR A N D C MURRAY
dialogue with the profession and particularly with the engineering work. The most recent discussion on the benefits
professional organisations. Professional competency, codes of a Registration Scheme with legislative force can be found in
of ethics, guidelines for conduct and other instruments the PPIR Project Report released in 2009 (p 48). The report
are considered and reviewed regularly. The Board and notes the benefits as:
professional organisations recognise that is important to
enhanced professional standards,
ensure the practising professionals are supported with ethical,
competency and performance expectations based on shared assured legitimacy,
understanding and consistency across the profession. better assessment and management of risks,
streamlined public administration,
BENEFITS OF PROFESSIONAL REGULATION stronger competition,
AND WORKFORCE CAPABILITY higher public profile for the engineering industry and
profession, and
DEVELOPMENT enhanced international mobility and trade.
Organisations across industries and sectors, from mining Professional regulation regimes provide an independent,
to biomedical consulting, are implementing strong human robust mechanism through which companies, individual
resources management of their technical and professional professionals and clients can vet the professional
capability.Graduate and ContinuingProfessionalDevelopment competency of, and the knowledge and expertise currency
Programs, for example, are prevalent as companies seek out of the professional engineers from whom they are receiving
value-adding tools for capability development and retention professional engineering services.
of expertise. Capability management is viewed as a critical
component of appropriate due diligence, risk management and Disciplinary actions taken under the provisions of the Act
strategic positioning in the constantly changing, increasingly are instructive. Analysis provides valuable information on
complex national and international marketplace. risks associated with and quality assurance of professional
engineering work and issues related to workforce capability. It
The Board is in constant dialogue with professionals could be used to inform engineering education and Graduate
regarding questions raised about legal liability, implications and Continuing Professional Development Programs and
of regulation on the responsibility and autonomy of so support workforce capability and professional risk
various members of the engineering team, including other management.
professions, and how responsibility for engineering work
completed outside the state can be accommodated within
the jurisdictionally bound regulatory framework. These CONCLUSlON
issues are not limited to a regulated environment. They are Within Queensland regulation of the profession is a
symptomatic of the significant issues facing the practice of living process. There is constant engagement between the
engineering and the capacity of professionals and engineering legislative instrument, the Board of Professional Engineers
companies to manage increasingly complex, diverse, but of Queensland, and the profession, both at the individual and
tightly interconnected and interdependent, professional association level, to co-regulate the profession. This ensures
activities. The situation has been recognised within the significant professional input into what constitutes appropriate
profession and work is being undertaken to develop tools to professional standards. It allows for flexibility to meet change
assist professionals and companies. yet provide consistency and rigour around professional
Over the past five years, the Warren Centre for Advanced engineering standards and provides an independent source of
Engineering based at Sydney University has been undertaking information to assist organisations ensure their professional
a Professional Performance, Innovation and Risk (PPIR) engineering capability.
Project. The project grew from discussions among leading
engineering professionals regarding the present status of
engineering and risk. It aims to:
REFERENCES
Australias Health Workforce Online, 2010. National registration
define professional performance in the practice of and accreditation scheme [online]. Available from: <http://www.
engineering; nhwt.gov.au/natreg.asp>
0 propose a set of initiative to bring about change in the
Australian Government, 2010.Council of Australian Governments
professional, liability and legal frameworks that govern (COAG) national legal profession reform [online]. Available
everyday engineering; and from: <http://www.ema.gov.au/w/agd/agd.nsf/Page/Consu
ensure that the outcomes of these changes offer the ltationsreformsandreviews_CouncilofAustralianGovernments(C
material benefits to everyone involved in buying, selling OAG)NationalLegalProfessionReform>
and using engineering products and services (The Warren Built Environment Design Professions (BEDP), 2010.The Australian
Centre, 2009, p 8 ) . Council of Built Environment Design Professions [online].
The PPIR Project analysed the contemporary realities for Available from: <http://www.bedp.asn.au/news/og/december.
the professional engineers (pii) and the issues mirror those htm>.
being discussed between the Board and the Profession. While Johnston, S, Gostelow, P and Jones, E, 1999. Engineering and
the recent work of the Warren Centre relates to a generally Society, An Australian Perspective (Addison-Wesley Longman:
unregulated engineering environment and is pragmatically Australia).
predicated on an opt-in scheme, the whole concept of an Queensland Government, 2005. Registration of Professional
Engineering Profession Act should probably be viewed as Engineers, discussion paper, Administrative ServicesDepartment,
a final step of enshrining in law what will already be largely Queensland Government, May.
happening in practice (p 50). The State of Queensland, 2002. Professional Engineers Bill 2002,
There is much that the regulation of professional engineering explanatory notes (Government Printer).
through the Professional Engineers Act 2002 has to offer The Warren Centre, 2009. Professional Performance, Innovation
individual engineers and industry generally in establishing, and Risk, The Warren Centre for Advanced Engineering (Sydney
articulating and maintaining standards for professional University: Sydney).
70 SEVENTH LARGE OPEN PIT MINING CONFERENCE / PERTH. WA, 27 - 28 JULY 2010
Three-Dimensional Hydromechanical
Simulation for Large, Faulted Pit Slopes
D Beck, B Fillery and F Reusch3
ABSTRACT
Owing to improvements in computer capacity, it is now possible to undertake fully coupled or
partially coupled, three-dimensional (3D) hydromechanical Finite Element (FE) simulation of
mines, including explicit representations of large numbers of discontinuities.The analysis requires
fewer assumptions than uncoupled analysis or two-dimensional (2D) analysis, but may require a
change in how hydrological data is collected. The way hydrological data is incorporated in stability
assessment and the approach to uncertainty in hydrological information also requires discussion.
The current procedures for this type of analysisare discussed,along with some practicaldifficulties
and teachings from early applications. A work flow for integration of 3D hydromechanical
modelling with pit planning and some considerations for the management of uncertain data are
also proposed.
INTRODUCTION
Frequent simplifylngassumptions used in slopemodels include explicitlyfor the particular geometry of the problem and there
that the rock mass is fully drained or that discontinuities can are fewer assumptions about the governing physics. Values for
be ignored. Naturally, models which assume things that are parameters need to be estimated but a range analysis can be
definitely untrue - such as that strength is infinite (elastic used to bracket the expected outcome.
models) or that structure has no effect (continuum models) The conflict between the simple approaches and more
cannot have high similitude with the real problem, but the sophisticated analyses is a complex technical, and possibly
hope is that trends in behaviour might be correlated. The risk philosophical dilemma. Ignoring issues caused by faulting and
of relying on these assumptions is that water or structure or pore water pressure, for example, will result in a potentially
rock mass failure are causes of instability, and ignoring them wrong answer, while it is certainly true that at many mines
introduces unquantifiable errors. The usual justification there is often insufficient data for a model with a high-
for simplifying assumptions is lack of information, resolution, explicit representation of these factors.
understanding or a goal of reducing analysis costs. Either The problem needs debate; overly simplistic analysis is
the available data or the available tools do not allow a more potentially dangerous and mines need long term stability
sophisticated accounting for water or structure, or the budget forecasting to design and engineer their slopes.
only accommodatessimpler analysis.
These are not often valid arguments. The standards for NUMERICAL EXPERIMENT
sufficiencyof design should be that the levels of risk resulting
from uncertainty are below an acceptable level and that the To demonstrate the reasonableness of certain simplifylng
essential requirements of the design are met. The difficulty of assumptions for pit models, a series of numerical experiments
a particular type of analysis needed to compute risk, or the has been conducted on a model of a fictitious pit which is
incremental cost of information to properly calculate it do not faulted and is mined in a rock mass with a shallow phreatic
usually excuse poor design. surface (ie a high water table). Technical details of the
experiment are highlighted in Beck, Fillery and Reusch (2010)
The most sophisticated slope stability analysis is coupled,
gD, discontinuum, non-linear hydro-mechanical analysis. and the analysis is discussed in more detailed there.
This form of analysis has not been common owing to historic The general procedure involved testing the same 3D
computational limitations, but is now employed by some of geometry with various faultedlunfaulted and drained/fully
the worlds largest mines. These models incorporate very drained assumptions so that the results could be compared.
large numbers of discontinuities, a non-linear constitutive The intent of the experiment was not to prove or disprove
model for the rock mass and the consideration of pore water that modellers should never undertake one or another type
pressure is governed by theories of poroelasticity. The benefit of analysis. The intent was to show that if water, strength
of this analysis is that the physical laws governing interactions and structure matter for a particular problem, the effect of
between solids, structures and water are accounted for certain simplifying assumptions can be to create an incorrect
1. Principal Engineer - Mining Group, Beck Arndt EngineeringPty Ltd, 9 Reid Drive, ChatswoodWest NSW 2067. Email: dbeck@beckarndt.com.au
2. Senior Engineer - Applied Mechanics, Beck Arndt EngineeringPty Ltd, 33 Cedric Street, Stirling WA 6021. Email: bfillery@beckarndt.com.au
3. Principal Engineer - Applied Mechanics, Beck Arndt EngineeringPty Ltd, Roennebergstr 15, Berlin 12161, Germany. Email: Freusch@beckarndt.com.au
D BECK, B FILLERY AND F REUSCH
impression of the likely outcomes that cannot be accounted are turned off. This is essentially an elastic continuum
for. model. Few practitioners would consider this an elastic
This matters very much in the typical mine planning and continuum approach for pit slope modelling.
modelling process that is followed in many mines. In the Case 2: A 2D, SSD FE model, coupled hydro-mechanical
example simulated for the experiment, modelling at the mine simulation, faults turned on, but having similar hydraulic
may start with a much simplified approach similar to Case z conductivity to the rock mass. This is the probably the
most typical modelling approach, usually considered
or 3 of Figure 1. The intent of the simplified model in a typical
sufficient for preliminary numerical investigations.
modelling project might be to aid in identifymg vulnerabilities
or to bracket certain design parameters, such as the slope Case 3: A 3D, SSD FE model, fully drained rock mass,
faults turned off.
angle or timing of pushbacks while minimising cost, or
Case 4: A 3D, SSD FE model, fully drained, faults turned
minimising the turn-around time for results. If vulnerabilities
on in the model, but having similar hydraulic conductivity
are identified at any stage, the analysis would be escalated to the rock mass.
to a more sophisticated approach, possibly a discontinuum, Case 5: A 3D, SSD FE model, coupled hydro-mechanical
non-linear 3D model for example, with fewer assumptions but simulation, faults turned on, but having similar hydraulic
increased cost. conductivity to the rock mass.
A problem with this work flow would occur if the original In order, the five cases represent increasing model size
simple stage was unable to capture the mechanisms of a real but varying assumptions. Cases 1 through 4 have the most
problem, or to offer some hint of it, so that the identification assumptions and Case 5 the least. Each additional feature also
of the vulnerability was entirely up to the imagination of the represents a small upgrade in model costs, as the transition
modeller. There are many reasons that a modeller may not from ZD to 3D to faulted to coupled models frequently
identify a risk for which the model shows no sign. Another involves purchasing additional software or licenses and
requires additional engineering skill. Naturally, even more
problem occurs if the effect of the assumptions on the results realistic models are possible than Case 5, which in this case
is so severe that there is no correlation between the model makes the least assumptions of all those in the experiment;the
and the future slope behaviour. In this circumstance it might intent here is to test the most common and basic simplifying
be incorrectly concluded that no more detailed analysis assumptions.
is necessary, or else the initial design might be grossly
insufficient. Material properties and assumptions
There should be general concern that most of the simplifying In each experiment, the material properties for the rock
assumptions discussed so far (elasticity, continuum or fully and for the faults (where applicable) were the same. The
drained) are likely to be non-conservative, so that design constitutive model for faults and the continuum components
parameters derived from them will over estimate the factor of are detailed by Reusch and Levkovitch (in press) and the
safety of the slope. material properties are listed in the Appendix, Tables A1 to
The numerical experiment consisted of five separate models, A3.
each using the model geometry shown in Figure 1, and each In summary, after Beck, Fillery and Reusch (2010):
representing a modelling approach employed with varying 0 The continuum parts (ie the unfaulted rock) are modelled
frequency for similar problems by engineers at mines around as a strain softening, dilatant Hoek Brown material.
the world: This means that as strain increases the material softens,
Case 1: A zD, Strain Softening Dilatant (SSD), Finite weakens and dilates.
Element (FE) model. The rock mass is fully drained (okPa 0 Discrete structures that are explicitly represented in
pore water pressure, or P W P throughout) and the faults the models are represented by cohesive elements. In
m
FIG 1 - Pore water pressure distribution for Cases 4 and 5 at end of mining (Beck, Fillery and Reusch, 2010).
THREE-DIMENSIONAL HYDROMECHANICAL SIMULATION FOR LARGE, FAULTED PIT SLOPES
FE simulations cohesive elements allow simulation of Take for example a case of a pit at an early stage of extraction.
the discrete behaviour associated with faults or shears The effective radius of curvature of the walls (a positive effect)
and can be used to construct a rock mass compromising is smaller than it will be at a later stage of extraction.
discrete rock blocks separated by discontinuities. Using At this stage, a 2D model may be able to be adjusted to
this technique, faults and shear zones are free to dislocate replicate instability or stability as observed in the pit, but the
and dilate and the fault surfaces themselves can dilate and material properties that are back-calculated will be unrealistic
because the real walls are affected by the confinement and
degrade. favorable geometry afforded by the compactness of the pit and
0 In the hydro-mechanical models, conventional equations proximity of the other walls and this must be compensated for.
governing fluid flow and PWP (Darcy, 1856) are solved Now imagine the outcome at a later stage of extraction, when
simultaneously with the equations for deformation and these same properties are reused. The model was never truly
damage inside 3D, strain softening,dilatant, discontinuum calibrated in the first place, and now the material properties
models. A typical PWP cross section is shown for Case 5 in will produce an incorrect result.
Figure 1, after Beck, Fillery and Reusch (2010).
Case 2 -Two-dimensional, strain softening
There are several other key assumptions:
dilatant, finite element, faulted, with pore
0 The stress field was: u,, = 1.3 0,.PWP at boundaries is
steady state gravimetrichydraulic head.
water
The pits were extracted in horizontal slices approximating This case represents the scenario with the greatest similarity
the height of a single bench to the base case in terms of water, structure and strength
assumptions, but assumes the problem is 2D. This is a common
Only a very small number of discontinuities were case that would be employed by engineers on mine sites. The
simulated. The spacing of structures in the model relative full hydro-mechanical coupling is an advantage over the use
to the pit size means that only global scale (ie wall scale) of flow nets, or one-way coupling (results of hydrological
interpretation is possible. This does not reflect current
simulation imposed on the mechanical simulation), but the
computational capacity; these models were designed to
benefits of full coupling in a small pit such as this are probably
run very quickly on a personal computer. not as significant as for larger more complex excavations.
In hydromechanically coupled cases the faults are
modelled with the same hydraulic conductivity as the rock The results show some similarities to the base case in terms
mass. of failure mechanisms; water and structure drive the failure,
but the extent and magnitude of the damage is entirely
EFFECTS 0F VA R I0US HYDRO-MECHA NICA L different and the conclusions and description of instability
would be different.
ASSUMPTIONS
The model simplifications are omissions that introduce error; Case 3 -Three-dimensional, strain softening
the pit is not 2D, the rock mass is not fully drained and it is dilatant, unfaulted, fully drained
faulted. This means that the faulted, hydro mechanically Assuming that the rock mass is unfaulted and fully drained
coupled Case 5 - which makes the fewest assumptions - results in a significant underestimation of the potential for
would have the least error, all else being equal and this is the instability for this scenario.
base case against the other cases can be compared.
For the base case (Case 5), instability is clear as a zone of
Figure 2 compares the cases in terms of rock mass damage
(plastic strain) and horizontal and total displacement. yielded, significantly displaced material in the lower half
Figure 3 shows the physical interpretation of plastic strain of the left hand slope, but for Case 2 neither the damage or
based on a discontinuum FE simulation of rock mass scale displacement are significant. The shape of the deformed
numerical specimens (Beck, Reusch and h d t , 2009). This zone is similar to the shape of the failed zone in the base
figure explains intuitively how rock mass damage is related to case. However displacement and plastic strain have physical
hydraulic connectivity in the model or in a rock mass. meanings that cannot be scaled and the levels of movement
for Case 2 and the nature of the deformation is not consistent
Case 1 -Two-dimensional, strain softening with the instability predicted by the base case.
dilatant, finite element, unfaulted, fully As damage is of course cumulative, or path dependent, the
drained lack of damage at this stage of the model also means that
The two-dimensional case has the most deformation for this future forecasts of damage, at later stages of the pit would
small pit and significantly more deformation than the base probably represent an even greater underestimation.
case (Case 5) despite the lack of faults and PWP. In contrast, It can be concluded that a 3D model with unfaulted and
this model also has low wall damage. The reason for this is fully drained assumptions offers a significant improvement
that the circular pit shape is clearly not able to be represented over the 2D case with similar assumptions but its uses as a
with high similitudeby a 2D model. The effect of the 2D model predictive tool are limited. The improved representation of
is the same as assuming the pit is very long compared to its the geometry (3D versus 2D) cannot add sufficient value to
depth, and with the assumed properties a long slope of this compensate for the loss of the effects of structure or water.
steep gradient would indeed fail. The stress path predicted by
the 2D model is simply incorrect. Case 4 -Three-dimensional, strain softening
The modelled scenario may well be an extreme case of the dilatant, faulted, fully drained
unsuitability of 2D modelling, but in any situation the only This case is faulted, but without the effects of PWP the depth
way to properly quantify the effects of the 2D assumption is to of damage in the walls is still too small and the instabilitydoes
simulate the correct 3D geometry. In some documented cases, not occur.
attempts have been made to adapt material properties in 2D
simulationsto account for the omission of a spatial dimension, If as is assumed in this case, PWP does contribute to eventual
but this akin to assuming properties are dependent on the failure, a model with these assumptions would have proven
geometry of the slope. insufficient despite being closer to the base case than the other
Case 1:
2 0 , Dry,
No
Fault
Case 2
2D,
Wet,
Faulted
Case 3:
3D, Dry,
No
Fault
Case 4:
30, Dry,
Faulted
Case 5:
30,
Wet,
With
Fault
FIG 2 -Simulated plastic strain, horizontal displacementand displacement magnitudes for all four cases after Beck, Fillery and Reusch (2010). In the no fault cases, the
fault traces are visible even though they are turned off.
FIG 3 - Simulated rock mass scale specimen depicting an interpretation of the physical meaning of plastic strain in SSD FE models
(Beck, Fillery and Reusch 2010).
THREE-DIMENSIONAL HYDROMECHANICALSIMULATION FOR LARGE, FAULTED PIT SLOPES
scenarios. With increasing depth, as PWP becomes more approximate a situation with range of potentially correct
important for stability, the model with these assumptions scenarios than to simulate a single scenario which makes
would have been even less useful. specific assumptions. The adage causes confusion. At
times it has been taken as implying that a simplistic model
Case 5 -Three-dimensional, strain softening will somehow produce a more realistic result or is a better
dilatant, faulted, coupled hydromechanical approach than a model that captures more of the physics
model of the problem, that approach being more complex.
This case represents the real base case; it makes the fewest This incorrect interpretation confuses the meaning of
geotechnical assumptions. The hypothesis that was tested complexityand assumptions.
by the experiment is whether an experienced engineer might 0 A good example of where the statement may apply, is
reasonably be able to derive similar conclusions from much when considering whether to use a non-linear model or
simplified models as could be made from this model that an elastic model; the mathematically more complex non-
includes more features but makes fewer assumptions. linear model requires an estimation of rock strengths and
The result of the experiment is clear: none of the simplified other parameters that are hard to measure, but the general
models can be interpreted to match the results of Case 5 - behaviour of the model should approximate real rock
the 3D coupled faulted model. The simplified models showed masses better. Range analysis will allow understanding of
deformation in similar places but did not show conditions that an approximate range of possibilities that can potentially
were unambiguously able to indicate instability in places that bracket the real result. The simple elastic model can
were clearly unstable in the coupled discontinuum model. however never be correct in a slope instability problem
as instability of the elastic slope is impossible. The elastic
The modelling also shows how each additional simplification
model, in this case is always precisely, or exactly wrong as
introduces error that is not easy to account for using intuition
a result of its assumptions.
and experience alone. This makes using the simplified models
very difficult unless their assumptions can be proven by field The same concepts apply to faulted versus unfaulted
observations. analysis, or hydromechanical versus full drained
simulation. If faulting or pore water pressure contribute to
A possible conclusion is that without a sufficient deformation, they need to be accounted for.
representation of structure, water and geometry model
Specificallyrelated to hydromechanical problems:
predictions can be wrong and do not add understanding. The
effect of ignoring the effects of PWP in the case of this pit, was 0 Pit models that ignore, or over-simplify poroelastic effects
perhaps as significant to the overall model error as ignoring when they may be relevant are hard to justify.
faults and neither the 2D,unfaulted or fully drained cases In the absence of data excluding the importance of PWP,
would have been sufficient in a real world scenario. stability analysis without consideration of poroelastic
There are some considerations for similar problems. effects may not be sufficient for planning purposes.
A common approach to the modelling of open pits is to
Generally:
stage the analysis, progressing from 2D,to 3D,adding
There is not more uncertainty in a 3D model than a 2D faults then water. However, in this experiment the
model. A 3D model can use exactly the same data as a 2D progression showed that such an approach would have
model to improve understanding and certainty in the area cost time and run the risk of inappropriate conclusions at
covered by the data. the early stages.
There is more uncertainty in a 2D model than a 3D model Mines should carefully consider whether the staged
using the same data, as the effect of the lost dimensionality approach adds value as in some instances, it may not.
on results in the well sampled volume cannot easily be In many cases it may enhance the understanding for
accounted for using experience and intuition alone. The the modelling process, but the conclusions may not be
effects of making assumptions outside a well sampled sufficient for mine planning purposes.
volume are usually easily accounted for (by discounting 0 Most slope models use a representation of PWP derived
the precision of results outside the well sampled volume from 2D analysis. Given the sensitivity of some slopes to
and conducting sensitivity analysis on the effects in the PWP, this approach requires great care and needs to be
area of interest). validated whenever it is applied.
There are not fewer assumptions in a simpler model, 0 3D analysis is not significantly more complex or costly, so
such as an elastic, unfaulted or undrained model and 2D analysis may be hard to justify.
there is not automatically less error compared to coupled 0 Unfaulted (continuum) simulations of open pits are
hydromechanical analysis. The hydromechanical models common, even though the spacing and persistence of the
may better capture the known governing physics of a slope smallest structures which are modelled is a limitation on
problem, and require fewer assumptions even though the resolution of the analysis.
input values will need to be estimated. Quantifying the 0 The numerical experiment shows that continuum analyses
effects of assumed input values requires a range analysis results in an underestimate of damage and deformation
to appreciate the uncertainty that is introduced, but this even in locations where instability is not occurring due
does not introduce additional error compared to simpler to dislocation on faults alone. This is an obvious and
analysiswhich ignores the full range of potential outcomes. predictable outcome, but it suggests that it is hard to
There is an adage in modelling that it is better to be justify persisting with unfaulted analyses except in very
approximately correct than exactly wrong, but this occasional circumstances.
statement is not meaningful for considering what The conclusions regarding unfaulted and fully drained
mechanisms or physics need to be incorporated into a analysis also extend to surface subsidence simulations for
model. The statement is meant to imply that it is better to caving mines.
IMPLEMENTING REALISTIC The structural model can be varied systematically: faults

can be turned on or off or various possibilities regarding
HYDROMECHANICALASSUMPTIONS IN REAL-
fault intersections or locations can be trialled.
WORLD SITUATIONS In some cases, smaller scale structure may be represented
The experiment shows that ignoring faulting and PWP will using a Discrete Fracture Network (DFN). DFNs are
result in a potentially wrong answer for pit slopes but this statistically generated networks of discontinuities used
is expected. A potential dilemma for decision making is that to approximate the fractures in a rock mass, and there
fault and PWP data may not be available. is usually a range of solutions that can match the sample
The problem should not be intractable. Mine planning data. Distinguishable scenarios, representing different
assumptions the DFN can be trialled.
and design in adverse environments is always a data limited
Ranges or possibilities for the hydromechanical properties
problem and it not just hydrological parameters which are
are tested in each of these approaches. Additional
uncertain. As a result, mines have developed many procedures
possibilities for range analysis related to poroelastic
for managing uncertainty.
effects include the testing of drainage programs, the
Common to these approaches is some sort of trigger and nature of inflows of the consequencesof unforeseen water
response philosophy: multiple courses of action are identified sources (perched water). Risk assessment and a proper
ahead of time, and the eventual decision of the correct way to appreciation of the problem can be used to frame the
proceed will be made prior to some predetermined last safe scenarios that need to be tested.
moment to act, based on the best available information at the These are just examples of sensitivity, or range analysis
time. The time between planning and design and this last safe that can be undertaken. Modern computing allows this
moment is used to collect information, either as a trigger or comprehensive analysis to be undertaken in a reasonable
to assist in making a better decision, while the likelihood and time. The end result will be a list of possible scenarios and
consequences of risks are appropriately considered. an understanding of the environment that would need to be
An example is used below to demonstrate how to manage
present for any of the scenarios to occur in practice.
a lack of hydrological data to facilitate hydromechanical The main effort now is to incorporate this understanding
modelling early in a project using one such approach. of causal factors in potential instability into the design, and
this requires a nuanced interaction between the modeller, the
Example application - Open pit concept designer and management.
analysis At this point in the example, the mine planners and the
management mission is to maximise the return on investment
An open pit is in a prefeasibility stage of assessment. There is
in the pit. This requires that the pit be stable and optimally
limited hydrological data other than a simple interpretation
designed. It would not be reasonable to design to the worst
of the phreatic surface and some limited permeability data
case, as this will not in most cases lead to the best return.
which is inconclusive. The mine has a first pass design based
on cost/recovery optimisation and now wishes to refine the In any case, there will be key variables identified by the
design to account for geotechnical constraints. range analysis that need to be quantified before one design
scenario can be selected over another. For example in this
A typical approach may be to simulate pore water pressure case, imagine that the scenario parallels the situation of the
changes using a 2D model, on a section where the phreatic numerical experiment:
surface has been measured. The next stage in such a process Case 3: the unfaulted, fully drained case allows a steep wall
would be to map the results of such analysisinto a simplified2D
angle (most economic, if the geotechnical scenario is true);
model with structure or to map a idealised 3D approximation
and
of it into a 3D model. In many cases such a model may ignore
Case 5: the faulted, high water level case allows a much
the presence of structure.
lower wall angle (least economic, unless the high water
Based on the results of the numerical experiment above, this level proves true, in which case it is the only technically
process would be deficient: feasible design).
the 2D model would not have sufficient similitude, as At this point the situation becomes clearer. There is some
shown in the experiment above (it simply produces the point in the future at which one design case or the other needs
wrong answers); and to be selected. This can be termed a decision point. There will
the PWP assumptions in the 3D mechanical model would be a last safe moment to make this decision.
not sufficiently capture poroelastic effects, and would not Between now and the last safe moment, or decision point,
adequately represent the PWP distribution. the environment can be sampled via a measurement program
A preferred approach would be to simulate the best available designed to test these critical variables; the degree of
structural and hydrological information in 3D, noting exactly faultedness and the true water level.
what assumptions are made. It may be the case that the This allows the mine to proceed. Either there is an economic
measurements have to be extrapolated some distance to allow imperative to collect the additional information required to
this, but the result will be an assumption that can be tested - commit to design Case 2, or there is not and design Case 4 can
future measurement will confirm or disconfirm the validity of be adopted. It may sometimes be the case that the required
the assumptions. information for a decision will only be gained by mining,
and this too can be accommodated provided the data for an
With knowledge of the assumptions that represent optimised decision will be available before the decision point
significant extrapolations, a series of experiments can then be (the last safe moment to make decision).
devised:
This of course requires some planning. The mine will need
Range analysis on the values of model parameters can be to ensure that they are properly orientated to follow the
conducted. optimised course of action when the decision point passes,
The boundary assumptions can be varied. and this may require some compromise.
THREE-DIMENSIONAL HYDROMECHANICAL SIMULATION FOR LARGE, FAULTED PIT SLOPES
Over time, the plan must continue to be revised and the effects of structure and poroelastic effects, but which
diligent, continuous review to identify unforeseen situations considers a range of possible scenarios, and to integrate the
is essential, but the adoption of the most resilient plan should understanding of causes of slope failure that results into the
place the mine in the best position to take advantage of mines design management procedures. Modem computers
additional opportunities and manage problems as they arise. allow such analysis to occur in a timely manner, so it is difficult
to justify lesser analysiswhich ignores these important effects.
CONCLUSIONS
A numerical experiment has been used to show that some REFERENCES
common model simplifications, undertaken to reduce cost or Beck, D A, Fillery, B and Reusch, F, 2010. 3D Hydro-mechanical
model difficulty can introduce gross errors. If the results of simulation of faulted open pit slopes, American Rock Mechanics
over-simplified models which exclude structural or poroelastic Association Symposium.
effects are used as basis for design, an insufficient design Beck, D A, Reusch, F and Amdt, S, 2009.A numerical investigation
would have resulted in the example scenario. of scale effects on the behavior of discontinuous rock, American
The uncertainty arising from the over-simplified analysis is Rock Mechanics Association Symposium.
difficult to overcome if left to an operating stage to deal with, Darcy, H, 1856.Les Fointaines de la Ville de Dijon. Victor Dalmont,
as the error is non-linear; the behaviour of the modelled pits Pans.
is unlike the real pit. It is also very difficult to use a trigger and Reusch, F and Levkovitch, V, in press. Application of a non-
response plan to overcome these effects. linear confinement sensitive constitutive model to mine scale
The solution is to conduct analysis that captures the simulations subject to varying levels of confining stress, Eurock
2010.
fundamental mechanisms of slope deformation, including
APPENDIX
TABLE A1 TABLE A2
Peak strength parameters for rock. Residual strength parameters for rock, applied progressively between
0.7 per cent and three per cent strain.
Hoek Brownclassification
Intact unconfinedcompressive strenqth (UCS) I 35 MPa Hoek Brown classification
Geoloqical strenqth index (GSI) I 36 Disturbancefactor 0.4
I 7 I Intact modulus I 12000MPa
Disturbancefactor 0
Intact modulus 12 OOO MPa
Hoek Brown criterion
I 0.71 la 0.52
I 8.2 x lo4 I Failureenvelope range

a 0.51 I(J3max I 8.75 MPa
Failure envelope range Rock mass parameters
03max 8.75 MPa Tensile strength -0.024
Rock mass parameters I ucs 0.51 MPa
Tensile strenqth I -0.04MPa I Deformation modulus I 830 MPa
ucs 0.90 MPa

Deformationmodulus 1500 MPa
TABLE A3
Fault strengthsfor faulted cases.
Mohr-Coulombpeak
C I 1.25 MPa
phi I 23.5"
Mohr-Coulomb residual
C 0.1 MPa
ohi 21"
Linking Operator Competency with Performance
P Davis
ABSTRACT
With the launch of Immersive Technologies next generation of Advanced Equipment Simulators for the mining industry,
correlating and influencing safety, productivity and maintenance with operator competency improvements is now made even
more effective.
In early 2008, Immersive assembled a project team to investigate the potential linkages between operator competency and
performance. Advanced Equipment Simulator Users with a range of sophistication were interviewed and consulted. The result
was new and enhanced valuerealisationprograms to directly assist mine sites to deliver the expected return on investment (ROI)
from simulators and operator competency development. In just over one year, 29 mine sites around the world have benefited from
34 completed programs. Case studies have been compiled to demonstrate the impact of the implementation of this approach,
including at the following operations:
Anglo Coal - Callide (Australia),
Freeport McMoRAN - Grasburg (Indonesia),
Equinox Minerals - Lumwana (Zambia),
Peabody Energy - Cabllo (USA), and
Xstrata - Lomas Bayas (Chile).
Significant performance improvements have been achieved including:
reductions of 47 per cent in gear shifting and 37 per cent in over speed errors,
improvements of 68 per cent in park break errors and 12 per cent in spotting time, and
increase of 54 per cent in passing emergency fire handling competency tests.
This presentation draws on these case studies to clearly demonstrate the linkage between improved operator competency
through the implementation of training programs tailored around Advanced Equipment Simulators and production performace
measured by a range of standard metrics at mines around the world.
I
FIG 1 -Strategic value of operator competencyprograms.
1. MAuslMM, SeniorVice President, BusinessDevelopment, ImmersiveTechnologies,4/40 Hasler Road, Osborne Park WA 6017. Email: pdavis@lmmersiveTechnologies.com
Capacity Increase of the Worlds Strongest Belt
Conveyor - Pushing the Limits in Chilean Copper
Ore by RWE
M Guerrero, H Reynaud2,S Tappeiner3and M Schmid4
ABSTRACT
Mining of copper ore is still the most important business in Chile. In order to mine ore with a copper content of less than one
percent at Minera Los Pelambres in an economical way, opencast mining is mandatory. Due to the prevalence of avalanches at
the mine, the concentrator was located in a stable area, some distance from the mine. As a consequence, the material has to be
transported between mine and concentrator by conveyor.
A downhill conveyor system has been implemented for more than ten years as the most suitable and most economical ore
transport alternative. Due to the length (12.6km) and inclination (1300 m drop) of the conveyor system, the stresses in the
steel cord belt are the dominant design criterion, resulting in a belt quality of ST 7800.Additional dynamic belt stresses during
transient operating conditions are avoided by the installation of high capacity adjustable frequency drives. The rating of a single
drive unit is 2500 kW. During ten years of operation, several capacity increases have been implemented, from an initial 5800 t/h
up to the current capacity of 9900 t/h.
This presentation describes the engineering challenges of the implementation of the worlds strongest belt conveyor system.
Capacity increase measures like the development of new control and breaking strategy, proved by dynamic simulation of belt
forces and belt speed scenarios are explained. During this long-term partnership of RWE Power International, the former
Rheinbraun Engineering, with Minera Los Pelambres, RWE has been engaged continuouslyfor development work, safety checks,
expert opinions and maintenance audits, and giving new impulse to conveyor technology.
REFERENCES
Dithich, Wand Einenkel, 0,2000. Downhill tunnel conveyor on the Chilean side of the Andes, Surface Mining, 52(3):235-244.
1. Managing Director, Head of Crushingand Conveying, Minera 10sPelambres, Av. Apoquindo 4001, Piso 18, Las Condes, Santiago CP 7550162, Chile. Email mguerrero@pelambres.cI
2. Head of Special Mechanical Projects, Minera 10s Pelambres, Av. Apoquindo 4001, Piso 18, Las Condes, Santiago CP 7550162, Chile. Email: hreynaud@pelambres.cI
3. Head of Mining and Conveying Technology, RWE Power International,RE GmbH, Stuettgenweg2, Cologne 50935, Germany. Email: stefan.tappeiner@rwe.com
4. Deputy Head Business Development,RWE Power International, RE GmbH, Stuettgenweg2, Cologne 50935, Germany. Email: martin.schmid@rwe.com
SEVENTH LARGE OPEN PIT MINING CONFERENCE / PERTH, WA, 27 - 28 JULY 2 0 1 0 81

Incorporating Economic Risk in Pit
Slope Designs
F J La?, W E Barnford*, S T S Yuen3and T Li4
ABSTRACT
Advances in technology have led to larger and deeper open pit mines and, consequently, also
increased the potential economic impact associated with slope failures. If the ultimate aim of
mining operations is to maximise project profit, it is surprising to see related research activitieshave
mostly been focusing on technical areas such as slope stability modelling, but with limited studies
conducted to date to assess the relationship between geotechnical uncertainties and economic risk.
Given most major decisions in the mining industry nowadays are made by management staff with
financial background, any design approach that would help to link slope stability analysis results
with monetary values would no doubt enhance communicationbetween geotechnical practitioners
and decision makers. This paper presents a study which assesses economic impact of slope failures
to assist in mine planning. The study also estimates the impact of slope failures on project value,
which would enable economic risks associated with pit slope designs to be closely examined.
The proposed risk assessment techniques, used by financial institutions for many years to assess
portfolio risks, can assist geotechnical practitioners in quantifying risks of slope failures. Risk-
adjusted concepts are also suggested to assist management in selecting a slope design with optimal
risk and return ratio.
INTRODUCTION IDEAL PIT SLOPE DESIGN PRACTICES

As the concept of sustainable development attracts more Pit design is an iterative process between the geotechnical
attention in recent years, increasingly geotechnical and mine planning disciplines. Continuous communication
practitioners working in the mining sector have to also between the two disciplines is essential to create an optimal
consider potential social, environmental and economic design. At the initial stage, it is common to divide the site into
impacts in their designs. While factor of safety and multiple domains, where each domain consists of materials
probability of failure are the two terms commonly used with similar geotechnical properties. The geotechnical
by geotechnical practitioners to represent the stability of engineers would then perform analysis and recommend
open pit slopes, these terms are not necessarily familiar to preliminary design slope angles for each domain. Once the pit
management staff mostly with financial training. In order to limit is determined, a more detailed pit and phase design is
effectively communicate their recommendations to decision then produced by the planners and passed on to the engineers
makers, geotechnical practitioners need to present results to assess for slope failure risks. A production plan is then
that are easy to understand in a financial context. Since many drawn up to enable the prediction of cash flows. During the
of the major decisions are made with profits in mind, it would construction phase, a better understanding of slope stabilities
be logical to link geotechnical analysis results with potential can be obtained by on-going site observation and analysis.
economic implications. This may lead to changes in the slope designs and hence
modificationsto the original mine plan as improved knowledge
This paper focuses on the economic considerations of the on the orebody, geotechnical properties and effects of other
geotechnical engineering aspect in open pit slope designs. Risk factors such as rainfall and water table become available. The
management tools used in the financial sector are discussed mine plan may also be changed as a result of commodity price
in the paper and recommended to geotechnical practitioners fluctuations.
for slope risk assessment. These tools have been used in the
The slope analysis results are often represented as factor of
financial sector for many years to manage financial risks and to safety (FoS), probability of failure (PoF) and potential failure
ensure any decisionsmade are consistent with a companys risk volume. FoS measures the ratio between the driving force and
policies. Risk-adjusted return, which measures the relationship resisting force in a system:
between risk and return, assists decision makers in ranking
design options. This makes certain that the risk associated with R
FoS = -
the selected design is compensated with the appropriate return. D
1. SAuslMM, The Universityof Melbourne, Departmentof Civil and EnvironmentalEngineering, The University of Melbourne, Vic 3010. Email: tjkpl@unimelb.edu.au
2. MAuslMM, Principal Fellow, Departmentof Civil and EnvironmentalEngineering, The Universityof Melbourne, Vic 3010. Email: wbamford@unimelb.edu.au
3. Senior lecturer, Departmentof Civil and EnvironmentalEngineering, The Universityof Melbourne, Vic 3010. Email: stsy@unimelb.edu.au
4. Principa1,TL Geotechnics and Mining. Email: tao-lil23@yahoo.com.au
SEVENTH LARGE OPEN PIT M I N I N G CONFERENCE / PERTH, WA, 27 - 28 JULY 2010 83

F J LA1 eta/
where: The selection of pit slope angles has a significant effect on

D = driving forces the overall value of the mine operation. Singh (1986) and
Stewart and Kennedy (1971) demonstrated that a much higher
R = resisting forces
net present value (NPV) can be achieved by steepening the
In theory, stabilityof the slope is achieved when FoS is equal slope angles, but their economic analyses did not consider the
to 1. In reality, due to uncertainties in the system and changes associated probable increase in slope failure risk. Lilly (2000)
to site conditions over time, the design FoS is often required and Call (1992)applied a cost-benefit approach to incorporate
to be greater than 1. the cost of failure and slope formation cost to establish the
Due to the variability of geotechnical parameters, total cost. Under their method, the slope design with the
probabilistic approaches have been introduced to run slope lowest expected overall cost is considered optimal. However,
analysis. Probabilisticanalysis treats FoS as a random variable the approach only considers the expected slope failure cost
and calculates the probability of FoS being less than or equal and ignores possible variation in failure cost.
to 1. The PoF is defined as:
BENEFITS OF INCORPORATING ECONOMIC
PoF = Probability (FoS I1) RISK IN PIT SLOPE DESIGNS
Slope stability modellingtechniques are likely to become more
The acceptable FoS and PoF values are dependent on
sophisticated as technology advances. Examples of recent
several factors, including slope scale, consequences of
research efforts include the modelling of discontinuities
failure, assumptions made, type of slope analysis performed
and reliability of data. Geotechnical practitioners make distribution (Rogers et al, 2009) and their fracture
recommendations on slope designs based on these results. propagation (Cundall and Damjanac, 2009; Elmo et al,
FoS values that are deemed acceptable in the past vary from 2009). To take advantage of further improvements on slope
1.1 to 1.5, while acceptable PoF values vary from five per cent modelling, geotechnical practitioners should spend more
to 50 per cent (Wesseloo and Read, 2009). effort to ensure that their slope analysis results are presented
in a context that can be easily understood by decision makers.
CHALLENGES OF EXISTING SLOPE DESIGN With many managers coming from a finance background,
management will likely pay more attention on slope analysis
PRACTICES results if their impact on profit is highlighted. By incorporating
Ideally, mine planning and geotechnical professionals economic risk into slope designs, geotechnical practitioners
should work closely together to produce the optimal slope may find it easier to justify how the extra spending will lead to
design. In reality, however, Hustrulid, McCarter and Van Zyl greater economic benefits for the mine overall. Furthermore,
(2001) observed that there is a tendency for each discipline
the determination of economic risk requires geotechnical
to specialise in its own area with limited awareness of the practitioners to work closely with mine planners, and would
requirements of other disciplines. They also commented
encourage more interaction and understanding between the
on the limited influence the geotechnical group may have
two disciplines.
on the final decisions. This is probably due to the differing
concerns of the geotechnical and mining departments. Mining Ultimately,mining operationsneed to maximiseproject value,
engineers are focusedon maximisingprofits and tend to favour as such many of the major decisions are undoubtedly driven by
aggressive slope designs where some failures are considered economics. The managers, mine planners, bankers and clients
acceptable. Geotechnicalpractitioners, on the other hand, are are not interested in geotechnical terms and principles. Rather,
traditionally trained to be conservative, as many had worked the onus is on the geotechnical practitioners to express their
on huge civil infrastructures that required a long design slope analysis results in economic terms and to report them in
life with failures leading to severe outcomes. As such, the a way that is easily understood by all stakeholders.
application of geotechnical principles alone may not lead to an
optimal mine design. Instead, an inter-disciplinary approach C HALLENGES 0F INCO R PORAT1NG
may assist geotechnical practitioners to take other factors into ECONOMIC RISK ANALYSIS IN SLOPE DESIGNS
consideration, and to put forward their analysis results to
Risk assessment of slope designs require slope models to be
other mining professionals for reference and adoption.
analysed probabilistically. Risk is defined as:
Design recommendations from geotechnical practitioners
are often based on FoS and PoF as mentioned above. Risk = PoF x consequencesof failure.
While these values are important for assessing the risk of
design on safety considerations, they do little to justify the As a result, an inaccurate PoFestimate leads to an inaccurate
recommendations from an economic point of view. This risk estimate. This concern is exacerbated in large-scale
potentially leads to recommendations being neglected when slopes where the acceptable PoF is often small. Errors in slope
economic decisions are being considered. modelling process may lead to significant percentage changes
The cost of failure is often used to represent the in the estimated PoF and consequently considerable error in
consequences of slope failures. The potential failure surface, the estimated risk. In addition to this, the consequences of
required to estimate the cost of failure, is commonly predicted major slope failures are often severe, therefore the estimated
by slope modelling analysis. The figure quoted is often the risk would be quite sensitive to any change in PoF.
failure surface with the lowest FoS, that is, the surface that As the project progresses, more geotechnical data are
is most likely to fail. The method will theoretically provide an collected and the slope models become a better representation
estimate of the most likely failure cost should a slope failure of the site. Given that the economic risk assessmentsaccuracy
occur. However, as slopes do not always fail as predicted by is heavily dependent on the slope model, it is more beneficial
the slope analysis, it is therefore unreliable to calculate the for the assessment to be performed at a later stage of the project
cost of failure based solely on the most likely failure surface. development.
a4 SEVENTH LARGE OPEN PIT MINING CONFERENCE / PERTH, WA, 27 - 28 JULY 2010
INCORPORATINGECONOMIC RISK IN PIT SLOPE DESIGNS
While the estimated PoF from slope models may differ from Location of failure - Slope failures located close to an
the actual PoF, it is still worthwhile to perform risk analysis. operating ramp may cause damage to nearby mining
The exercise is not expected to provide a definitive solution; equipment. Mining operations may also be disrupted
rather, it would assist decision makers by reducing the bounds leading to deferred processing of ore. However, if slope
of uncertainty. In going through the risk assessment process, failure occurs in an area that does not affect current or
it enables the economic risk associated with slope designs to planned mining operations, it will incur a much lower cost.
be assessed. This would encourage decision makers to deal The type of slope remediation required varies depending
with the uncertainty and be better informed when managing on the location of failure. Furthermore, if failure occurs on
geotechnical risk. top of or within an orebody, it may prevent further mining
of the ore.
In addition to this, more sophisticated numerical tools such Interim or ultimate slope - Slopefailures would disrupt the
as FLACQD(HCItasca, 2008b) and 3DEC (HCItasca, 2oo8a) planned mining sequence of the pit. If failure occurred at
are often required to model slopes located in more complex an interim slope, cleaning up the failed tonnage effectively
situations. Although computing technology has improved in brings forward material that is intended for mining later,
recent years, the running of probabilistic analysis is still not hence causing disruption to the original mining schedule.
practical in these modelling tools due to the intensive power If failure occurred at the ultimate slope, failed material
required. equates to extra unplanned material, increasing the total
tonnage of material being mined (CANMET, 1976).
ESTIMATING ECONOMIC RISK OF PIT SLOPE Grade differentiation between stockpiles and the ore being
DESIGNS mined - If slope failure prevents ore material from leaving
Discounted cash flow, which recognises the time value of the pit, stockpiles would instead be fed to the processing
money, is a commonlyused method for valuing projects. When plant. In this instance, the cost of failure may include the
all future cash flows are discounted to their present value and delay in production of the high-grade ore. The cost is the
summed, the NPV is calculated. NPV is widely used to assess time value of money, caused by the immediate processing
long-term projects, such as those in the mining industry. of the lower grade stockpile rather than the production of
high-grade ore from the pit.
A few risk management conceptswill be briefly introduced in Once the consequences of slope failure are identified, they
this paper to assess the impact of slope designs on the project can be incorporated into the mining schedule to simulate
NPV. The steps recommended for applying these concepts are the effects of slope failures on the project NPV. The mining
summarised in Figure 1. schedule contains information such as the date and location
of mining activities, quantity of ore and waste tonnage, ore
Estimating the economic consequencesof grade and the intended destination of the mined material.
slope failure events All this information can be used to predict future cash flows
One of the ways to assess the economic risk associated with of the mining operation. The steps below are proposed to
geotechnicaldesigns is to estimate the impact of slope failures incorporate slope failure costs into the project NPV
on NPV. To do this, the PoF and the consequence of failure 1. The open pit is divided into several geotechnical domains.
need to be estimated. The PoF and failure tonnage can be Each domain is allocated to represent sections of the pit
establishedby examining results from slope stability analyses. that share similar geotechnical properties.
2. Numerical models are employedfor slope stabilityanalysis.
It is difficult to predict the consequenceof failure as it depends
Probabilisticanalyses are carried out to determine the PoF
on a number of factors. Some of the more important factors
of each slope.
are briefly described below ( h i et al, 2009):
3. Based on the slope analysis results, mine operators are
Failure tonnage - Higher failure tonnage leads to higher provided with the recommended slope designs.
direct cleanup cost. Lilly (2000) suggested applying an 4. The economic consequences of slope failures are then
efficiency factor of 70 - 80 per cent to the normal mining estimated.
cost due to the extra difficulty in excavatingfailed material. 5. Once the long-term mine plan is produced, the effects
In addition to this, higher failure tonnage also leads to of slope failures can be incorporated into the mining
longer cleanup time, therefore delaying equipments from schedule. This is carried out by simulating slope failures
returning to mining activities. throughout the entire mine life and including the failure
Availability of mining equipment for clean up - If mining costs in the mining schedule. The economic returns of the
equipment is running at capacity, they would need to be project can then be estimated.
mobilised from production activities. This will slow the 6. Step 5 is performed repeatedly to derive a distribution
mines overall production. curve for economic returns.
J
FIG 1 - Steps in implementing the proposed risk management concepts.
FJLAIetal
Incorporatingslope failures into the mining Sampling is first performed on the largest scale slopes
schedule followed by the smaller slopes. This is because if a segment of
the overall slope fails, the smaller slopes within this segment
Once a mining schedule is obtained from the long-term mine would also fail and there is no reason to further sample the
planners, it can be used to estimate the operations projected bench and inter-ramp slopes.
cash flows until the end of the mine life. Typical information
of the mining schedule is listed below: In order to estimate the impact of slope failures, each unit
mining location within the pit; cell is designated a slope type based on the potential impact of
date of mining activities; failure. The three properties suggested to determine the unit
estimation of tonnes of ore and waste being mined; cells slope type are listed below:
predicted ore grade; and 1. geotechnical properties of the material - they determine
destination of the mined material, for example to the the probability of failure and failure tonnage;
mineral processing plant or stockpiles. 2. scale of slope - this is often classified into three categories,
The changes in slope geometry over time are also required bench, inter-ramp and overall; and
for inputs into slope modelling. The slope analysis results 3. location of slope - this factor was discussed earlier and it
are used to determine the probability of failure of slopes plays a major role in determining the economic impact of
throughout the mine life. slope failures.
Besides estimating the probability of slope failure, it is also When assessing the consequence of bench failures, it is
important to determine the approximate failure tonnage for sometimes acceptable to use the expected failure tonnage
risk analysis. The usage of mean failure tonnage may result to estimate the cost of failure. For bench slopes with small
in a lack of understanding of potential catastrophic failures. number of unit cells, the failure costs are relatively low and
Lilly ( 2 0 0 0 ) suggested using the maximum failure tonnage slight errors in estimating these costs are acceptable. For
to estimate the consequence of slope failure. In most cases, slope types with large number of units, while there would be
the maximum failure tonnage only has a small chance of variation in failure tonnages, performing Latin Hypercube
occurrence. Thus, using the maximum failure tonnage for sampling on such a high quantity of benches would lead to
risk assessment may lead to over-conservative designs. An results similar to the expected cost of failure anyway.
alternative is to quantify the possible failure tonnage with a Overall and inter-ramp slopes are usually present in much
distribution curve. lower number of units than bench slopes. Failure in the
large-scale slopes would result in major disruptions. The
The Latin Hypercube sampling method can be used to run expected cost of failure may not accurately reflect the actual
the slope analysis and determine the slope failure tonnage. consequence of failure. Therefore for large scale slopes,
Variable model inputs are represented by distribution this paper proposes representing the failure tonnage with a
curves. Sets of random numbers are then drawn from these probability distribution curve, hence including a wider range
distribution curves and entered into the model to simulate one of scenarios for risk assessment. Figure 2 shows an example of
outcome. The sampling is repeated many times to determine the possible rock mass failure tonnage of a slope predicted by
the likely range of the final result. Slide (Rocscience, 2005). It is clear that the potential failure
The unit cell concept (CANMET, 1976) has been applied to tonnage varies and a more comprehensive risk assessment
unitise the slope walls. A unit cell is defined as a section of slope will be required to account for this uncertainty.
wall with equal width and height. The height of the slope wall, Most areas of the pit are subjected to multiple failure
together with the equivalent distance along the wall, forms the mechanisms. If simulation results in multiple failures in a unit
boundary of a unit cell. Since the height of the wall is fixed, the cell in the same time period, the failure mode with the higher
number of unit cells depends on the horizontal distance along tonnage would be used to calculate the failure cost.
the wall. The number of unit cells increases as the distance along After all the costs and delays caused by slope failures are
the wall increases. The unit cell concept is important for Latin incorporated into a mining schedule, the economic returns of
Hypercube method as it quantifies the slopes for sampling. the project can be calculated.
FIG 2 - A distribution curve showing the potentialfailure tonnage of a rock mass failure.
INCORPORATING ECONOMIC RISK IN PIT SLOPE DESIGNS
Each simulated run represents a possible scenario for the VaR indicates how likely loss would exceed a particular
mine. If the simulations are repeated many times, the results value, however it does not provide estimates for excess
can be used to plot a histogram of the expected returns losses. Two designs with identical VaR values may have very
associated with all the runs to obtain the returns distribution different risk exposures. They may be assigned with the same
graph. value under the VaR method because tail events are ignored.
Consequently, the concept of Conditional Value at Risk, which
RISK MANAGEMENTTOOLS FOR PIT SLOPE will be discussed in the following paragraph, is introduced to
DESIGNS focus on events with losses that exceed VaR.
Once the simulated NPV distribution curves are established, Conditional value at risk (CVaR)
they can be treated with risk management tools. As mentioned CVaR, defined as the expected losses above the VaR value,
above, many such tools exist in the financial sector and the is calculated by taking the weighted average of the losses
two commonly-used concepts are briefly outlined below. exceedingVaR. The calculation of CVaR provides management
with an indication of what to expect if an event exceeding VaR
Value at risk (VaR) occurs. Artzner et a1 (1997) provides further discussion on the
In simple terms, VaR is the maximum likely loss incurred advantages of CVaR over VaR. Figure 4 shows an example
over a specified period of time at a given confidence level. of CVaR, it concentrates on scenarios with losses exceeding
Figure3showsatypicalVaRdiagram.In thisexample,theVaRat the VaR in Figure 3. The CVaR in Figure 4 means the average
95 per cent confidence level is -$loo x lo6.This means that outcomes on the worst five per cent scenarios is -$226 x lo6.
loss is expected to exceed $100 x lo6 in five per cent of the
time. This method can be used to assess risk associated with RISK-ADJUSTED RETURNS
the position of an asset, a portfolio of assets or an entire Management are often presented with several design options.
company (Lai and Stange, 2009). From an economic point of view, the optimal design has the
FIG 4 - CVaR with 95 per cent confidence level.

i
F J LA1 eta/
highest return with minimal risk. However, the decision is penalises both upside and downside volatility whereas the
often not clear-cut as the design that generates the highest Sortino ratio only penalises volatility below the required rate
return does not always have the lowest risk. Risk-adjusted of return. Generally, most managers place greater emphasis
return, which measures the return in relation to the amount on avoiding downside risk (Sortino and Van Der Meer, i99i),
of risk undertaken, can assist managers to rank design options the Sortino ratio is therefore more consistent with their
and select the one with the optimal balance between risk and attitude to risk.
return. Several ratios that measures risk-adjusted return are
briefly discussed below. Others
Other commonly used methods for assessing risk adjusted
Risk adjusted return on capital (RAROC) return in the financial sector include the Information ratio,
The Risk Adjusted Return on Capital rule (Dowd, 1998) is Treynor-Black ratio, Capital Asset Pricing Model and
defined as: Arbitrage Pricing Theory. Some of these approaches may also
be useful in assessing slope designs.
RAROC = RP ~
It should be pointed out that the ratios will highlight the
VaR
design options with favourable risk and return balance but
where: does not consider the companys attitude to risk. Management
R, = expected return should still consider the companysrisk appetite when making
VaR = value at risk a decision.
The RAROC increases as return increases, and decreases
as risk (represented by VaR) increases. However it has CASE STUDY -TELFER GOLD MINE
the drawback of heavily favouring low-risk strategies. For Telfer Gold Mine, owned by Newcrest Mining Limited, is
instance a riskless strategyhas zero VaR and therefore RAROC located in the East Pilbara region of Western Australia.
equates to infinity. For this reason, the RAROC is not the most The open pit contains three major geological units - Outer
appropriate means for assessing low-risk strategies. Siltstone Member, Middle Units and Malu Quartzite Member.
Gold and copper are produced in this mine.
Traditional sharpe ratio The operation started as an underground mine in 1977 and
One of the quick and simple methods to estimate risk adjusted was suspended in 2000 due to concerns in rising costs. Later,
return is the traditional Sharpe ratio (Dowd, 1998): a feasibility study carried out in 2002 concluded that mining
where: was economically viable, leading to the commencement of
R,, = expected return open pit operations in mid-2003.
R, = benchmark return Telfer is comprised of both open pit and underground
Rp-% = expected differential return operations. The open pit operation includes two open pits, the
Main Dome and the West Dome. This case study focuses on
ued = predicted standard deviation of differential return the footwall wall of the Main Dome.
The traditional Sharpe ratio is the expected differential
return per unit of risk associated with expected differential Slope stability analysis
return. It is clear from the equation that a higher R or a lower
Discontinuities in the Main Dome were classified into five
oedleads to a higher traditional Sharpe ratio and, therefore
joint sets, Ji, 52, J3, J 4 and J5. The defect sets were then
a higher risk adjusted return. The ratio can be used to rank
plotted onto a stereonet in Dips (Rocscience, 2oo8a) to
different strategies based on their risk-adjusted return.
investigate the potential failure mechanisms in the footwall.
Sortino ratio Kinematic analyses were then carried out on the joint sets.
Unlike the Sharpe ratio, which penalises both upside and Based on the analysis and the continuity data of joint sets, it
downside volatility, the Sortino ratio only penalises volatility was established that structural failures involving the joint sets
below a specified rate of return. The Sortino ratio is defined are highly unlikely on the inter-ramp and overall slopes.
as: A section of the footwall also contains continuous clay-filled
bedding defect dipping 40 out of the wall. The bench slopes
RP -_
Sortino ratio = _ RP adjacent to the haul ramp in this section of the wall have been
DR designed at 40 to decrease the chances of planar failure
where: through undercutting of bedding. The inter-ramp slopes at
R,, = expected return
locations with clay-filled steeply dipping bedding were also
designed at 40 to increase stability. The potential failure
% = benchmark return mechanisms for the footwall are summarised in Table 1.
DR = downside risk, defined as:
m
Wedge failures
DR = ( - i ( T - x)f(x)dx
i Swedge (Rocscience, 2008c) was used to analyse wedge
failures. Slope analysis was carried out probabilistically
with the Barton-Bandis criterion (Barton and Bandis, 1990).
where: Persistence analysis was also performed, taking into account
f(x) is the probability density function of the return the joint continuities when estimating the probability of
wedge failure.
The Sortino ratio is similar to the Sharpe ratio. The only
difference is in the denominator where the Sharpe ratio uses Probabilistic analysis was then carried out to assess the
standard deviation, and the Sortino ratio only considers frequency in which Ji and 52 were likely to intersect. Joint
downside deviation. This effectively means the Sharpe ratio spacing data were collected and used to simulate J i and
All slopes in the pit were classified into different slope types
TABLE 1 according to their potential impact of failure. Each slope
Major failure mechanismsin the footwall. type has its own cost model for cost of failure calculations.
The number of unit cells associated with each slope type was
Slope type Failure mechanism Discontinuities estimated for each time period. Probabilistic software, @Risk
I Bench I Wedge I J1 andJ2 1 (Palisade Corporation, 2008), was then used to run Latin
Hypercube sampling on the large-scale slopes unit cells. If
I Bench I Rockmass 1 nla I simulation resulted in slope failures, the failure costs would
Bench Planar Bedding be added to the total cost of the period in which the failure
Inter-ramp Rock mass nla occurred. Simulation was repeated 1000 times to obtain a
distribution curve for the projects return.
Overall Rock mass nla
It was assumed that the geotechnical parameters would
remain relevant for one year. Therefore, if a slope was
calculated to have a probability of failure of ten per cent,
there would be a ten per cent, chance of failure within the
next one year. If the slope was sampled quarterly, there
would be a 2.5 per cent chance of failure within the next three
months. Sampling of slope failures and cost calculationswere
performed quarterly in the first three years and annually
POFwetige failure -
- POFwetigefailure x width of unit cell (m) from the fourth year onwards. Accuracies of the cash flows
unit cell meter in the later years are less important because they are heavily
discounted and therefore have a lower impact on the project
Planar failure value. Also, analysis in later year cash flows are only rough
Rocplane (Rocscience, 2008b) was used to assess the estimates as data from those periods are more likely to
probability of failure in the bedding plane. Spacing data contain significant uncertainties. The companys opportunity
were used to estimate the expected number of bedding plane cost of funds was taken as the benchmark return for NPV
in a unit cell. The probability of failure per unit cell can be calculations.
represented as follows: The internal rate of return (IRR) results from a run of 1000
times are summarised in Figure 5. Since large-scale slopes
POFplanar failure -
- POFplanar failure expected number of defect in Telfer have low probability of failure, most runs do not
X
unit cell defect unit cell result in major slope failures. As shown in the figure, IRR
is approximately 30 per cent when there are no major slope
Rock mass failure failures. The variation in IRR shown in the graph is due to
Rock mass failure was assessed with the Generalised Hoek- slope failures. The VaR with 95% confidence level is 17 per
Brown failure criterion (Hoek, Carranza-Torres and Corkum, cent, suggesting there is a five per cent chance that slope
2002). Rock mass failure analysis was performed with Slide, failures would lead to a project IRR of less than 17 per cent.
a 2D limit-equilibrium slope analysis program. In Latin Similarly, for VaR with 99 per cent confidence level, there is a
Hypercube sampling, it has been assumed that if a rock mass one per cent chance that major slope failures will lead to the
failure is simulated, the entire unit cell would fail. Therefore, project making an IRR of less than two per cent.
the total failure tonnage for a unit cell is calculated as follows:
Interpretation of results
Failure - Failure tonnage x width of unit cell (m) A hypothetical alternative slope design was created to
unit cell meter demonstrate the VaR concepts. It contains steeper inter-
ramp slopes and overall slopes. This leads to a lower slope
However, more accurate estimates can be achieved by using formation cost but a higher failure cost. The mining schedule
a 3D slope stability model. was also modified to reflect any changes in mining activities in
different time periods. Generally, waste excavation is delayed
Incorporating slope failures into the mining while the mining of ore is brought forward. The hypothetical
schedule mining schedule contains a large number of assumptions and
A mining schedule in spreadsheet format, containing would require improvements before it can be applied to the
information on mining activities planned for each period, was mine. It was simply created to demonstrate how the VaR,
obtained from the long-term mine planners. Together with CVaR and Sharpe ratio can be used to compare alternative
the mining and processing costs, the mining schedule was slope designs.
used to predict the operations costs and revenue over the life A summary of the results for the hypothetical and existing
of the mine. slope design in Telfer is shown in Table 2.
TABLE 2
Summary of slope risk analysis results.
Slope design
Existing
I I VaR9596
17%
CVaR 95%
7%
I ratio
3.85
27% 5%
Hypothetical 29% 6% 18% 9% 3.59
(steeperslopes)
F J LA1 eta/
FIG 5 - Internal rate of return results of the existing mine plan from lo00 simulation runs.
Traditionally, engineers are mainly concerned with the and return ratio compared to the hypothetical mine plan.
mean and standard deviation of returns associated with As a result, the mine operator should keep the existing mine
each slope angle. The mean provides the expected rate of plan. Only the Sharpe ratio is presented in this case study, but
return. The standard deviation is a probability measure that the RAROC and Sortino ratio are also useful references for
represents the variability of returns. The standard deviation is determining the optimal risk and return balance.
not quantified in monetary terms, and this makes it difficult to
report to management the potential loss associated with slope CONCLUSIONS
failures. The VaR and CVaR values provide management with An increasing number of risk sources are being quantified in
an estimated value of losses should large scale slope failures monetary terms in the mining industry. While the variable
occur. nature of geotechnical properties creates a difficultly to
The proposed VaR and CVaR concepts were used to assess quantify risk, it is important that geotechnical engineering
the risk associated with the two design options. As mentioned does not fall behind in translating data into monetary terms in
earlier, the existing mine plan has a VaR of 17 per cent at line with the rest of the industry. One of the main difficulties
95 per cent confidence level, implying that slope failures would in quantifymg geotechnical risk is the lack of data. Since
cause the return to drop lower than 17 per cent in five per cent it is impossible to capture site conditions completely, the
of the time. In Table 2, the existing mine plan has CVaR of actual slope behaviour may vary from that predicted by slope
seven per cent at 95 per cent confidence level, meaning that models. However, the aim of risk assessment exercise is not
the expected return is seven per cent for the worst five per cent to produce a definitive solution, rather, the process enables
of the cases. The VaR and CVaR values would be submitted uncertainties to be identified for examination and reduces the
to management to assess if the risk levels are consistent with range of likely scenarios.
the companys overall risk strategy. If the design exceeds the The slope design process requires close cooperation between
companys risk tolerance, the design is too risky and should geotechnical and mining staff. The differing focus of the two
be modified. professions potentially leads to a difference in opinions.
One of the challenges here is the selection of an economically Geotechnical practitioners tend to favour low-risk designs,
optimal slope angle in conjunction with conflicting risk and whereas mining engineers usually prefers more aggressive
return indicators. The risk-adjusted ratios partly solve this approaches to maximise profits. In view of this, it is suggested
problem by incorporating both risk and return into one that geotechnical recommendations, supported by economic
value. When presented with various design options, the slope justification, will provide additional information for the
angle with the highest risk-adjusted ratio would be the one mining engineers.
with the highest risk and return ratio. This is demonstrated A method to incorporate the cost of slope failures into
in Table 2. In this case, the hypothetical design has a higher the mining schedule has been suggested. This enables the
expected return compared to the existing design. However, economic impact of slope risk on the project value to be
the hypothetical design also has a higher standard deviation predicted. Several risk management tools from the financial
on returns, which represents a design with more risk. In the sector may be employed on pit slope designs. VaR predicts
absence of more information, the decision to balance the the likely maximum expected loss over a specified period of
risk and return would have to be made based on personal time at a given confidence level. CVaR estimates the expected
judgement and experience with limited quantitative analysis. loss for events that exceed VaR. Risk adjusted return ratios
The risk-adjusted ratios should be used in conjunction measure the risk-return trade-off. It is proposed that RAROC,
with VaR and CVaR to assess options with conflicting risk traditional Sharpe ratio and Sortino ratio may be employed
and return parameters. In Table 2, the existing mine plan to assist managers in ranking slope design options with
has a higher Sharpe ratio, therefore it has a favourable risk conflicting risk and return values.
The Telfer Gold Mine has been used as a case study to Hoek, E, Carranza-Torres, C and Corkum, B, 2002. Hoek-Brown
demonstrate how the techniques proposed in this paper could failure criterion, 2002 edition, in Proceedings NARMS-TAC
be applied to an open pit mine. Conference, pp 267-273 (University of Toronto: Toronto).
Hustrulid, W A, McCarter, M K and Van Zyl, D J A, 2001. Slope
ACKNOWLEDGEMENTS Stability in Surface Mining, 456 p (SME: Littleton).
The geotechnical and mining engineers at the Newcrest h i , F J, Bamford, W E, Yuen, S T S and Li, T, 2009. Implementing
value at risk in slope risk evaluation, in Proceedings Slope
Mining Perth office and Telfer site office have provided data Stability 2009 (Universidad de 10s Andes: Santiago).
for use in the paper. Specialthanks to Mr Charles Mkandawire
h i , F J and Stange, W, 2009. Using value at risk for integrated
of Newcrest Mining Limited for his valuable suggestions. The
project risk evaluation, in Proceedings Project Evaluation 2009,
management of Newcrest Mining Limited is acknowledged for pp 223-232 (The Australasian Institute of Mining and Metallurgy:
permission to publish this paper. Melbourne Australia).
Lilly, P A, 2000. The minimum total cost approach to optimum pit
REFERENCES slope design, in Proceedings International Symposium on Mine
Artzner, P, Delbaeu, F, Eber, J M and Heath, D, 1997. Thinking Planning and Equipment Selection, pp 77-82 (AA Balkema:
coherently, Risk, 10:68-71. Rotterdam).
Barton, N and Bandis, S, 1990. Review of predictive capabilities Palisade Corporation, 2008. @RISK: Palisade Corporation.
of JRC-JCS model in engineering practice, in Proceedings Rocscience, 2005. Slide 5.0, Rocscience Inc.
International Symposium on Rock Joints (eds: N Barton and
Rocscience, 2oo8a. Dips 5.0, Rocscience Inc.
0 Stephansson), pp 603-610 (AA Balkema: Rotterdam)
Rocscience, 2008b. RocPlane 2.0, Rocscience Inc.
Call, R D, 1992. Slope stability, SME Mining Engineering
Handbook (ed: H L Hartman), second edition, volume 1, Rocscience, 2008c. Swedge 5.0, Rocscience Inc.
pp 881-896. Rogers, S, Elmo, D, Beddoes, R and Dershowitz, W, 2009. Mine
Canada Centre for Mineral and Energy Technology(CANMET), 1976. scale DFN modelling and rapid upscaling in geomechanical
Pit slope manual supplement 5-3 Financial computer programs, simulations of large open pits, in Proceedings Slope Stability
Otawwa, Pit Slope Project of the Mining Research Laboratories, 2009 (Universidad de 10s Andes: Santiago).
Minister of Supply and Services. Singh, S P, 1986. Economical and geotechnical aspects of steeper
Cundall, P A and Damjanac, B, 2009. A comprehensive 3D model slopes, in Proceedings International Symposium on Geotechnical
for rock slopes based on micromechanics, in Proceedings Slope Stability in Surface Mining, pp 23-29 (AA Balkema: Rotterdam).
Stability zoo9 (Universidad de 10s Andes: Santiago). Sortino, F and Van Der Meer, R, 1991. Downside risk, Journal of
Dowd, K, 1998. Beyond Value a t Risk: The New Science of Risk Portfolio Management, 17(4):27-31.
Management (Wiley: New York). Stewart, R M and Kennedy, B A, 1971. The role of slope stability
in the economics, design and operation of open pit mines, in
Elmo, D, Moffitt, K, DAmbra, S and Stead, D, 2009. Characterisation
Proceedings First International Conference on Stability in
of Brittle Rock fracture mechanisms in rock slope failures, in
Open Pit Mining, pp 5-22 (The American Institute of Mining,
Proceedings Slope Stability 2009 (Universidad de 10s Andes:
Metallurgical, and Petroleum Engineers: Littleton).
Santiago).
Wesseloo, J and Read, J , 2009. Acceptance criteria, in Guidelines
HCItasca, 2oo8a. 3DEC HCItasca.
for Open Pit Slope Design (eds: J Read and P Stacey), pp 221-236
HCItasca, 2008b. FL4CsD HCItasca. (CSIRO Publishing: Melbourne).
Mine Planning and Risk Management
Associated with the Breakthrough of
the Telfer Sublevel Cave into theActive
Telfer Open Pit
L Cerny, J Dixon2,C McArthur3and U Singh4
ABSTRACT
The Telfer main dome open pit is operated by Newcrest Mining Limited at a current production rate
of 50 Mt/a from a pit dimension compromising 2 km strike length, 1 km width and 250 m depth.
As part of the refeasibility of the Telfer operation undertaken early this decade, Newcrest
identified that an underground sublevel cave (SLC) operation along with an increased scale of open
pit would support a new plant targeting the processing of gold and copper ore at 17 Mt/a rates.
The SLC was initiated in late 2006 and broke into the open pit in late 2009.This situation is
unique across the mining industry where a cave breakthrough was planned and occurred within
an active large open pit operation.
The SLC broke though into the west side of the Main Dome open pit into a specially prepared bench
that was left at the 5384 RL of cut-back stage 3.A main haulage ramp is within 100 m from the edge
of the cave breakthrough zone and this ramp passes down to the active levels which are progressing
toward the 5200 RL.
This paper presents insight into the mine planning and risk management associated with continued
operation of the open pit operation while an active SLC breaks through. The cave breakthrough
presents unique areas for risk management as there is potential for rapid propagation and subsidence.
To support the undertaking, mine design and major hazard management planning has been carried
out. Focus areas include; dual ramp system design, preloading the breakthrough area, monitoring
of cave propagation and air gap using a seismic system and deep hole extensometers, wall stability
and subsidence monitoring using radar and prism systems, and key trigger and response planning.
INTRODUCTION
The Telfer gold mine owned by Newcrest Mining Limited is sublevel caving operation. The SLC was initiated in late 2006
located in the Great Sandy Desert some 400 km east-south- and the SLC extraction areas are some 800 m below the
east of Port Hedland. western side of the open pit.
The Telfer deposit was discovered in 1971 when anomalous The open pit is currently mining at rates in the order of
gold and copper values were returned from outcrops in what 50 Mt/a of ore and waste, while the underground is hoisting
is now known as Main Dome. The Telfer mining operation in the order of 6 Mt/a. The ore sources combine to feed a mill
started in 1977 and used open pit methods to mine oxide that treats approximately 20 Mt/a.
material and narrow vein underground methods to access From the Telfer open pit and underground ore sources
high-grade reefs. In late 2000the operation was put onto care for the 12 months to 30 June 2009,the operation produced
and maintenance after producing approximately 6 Moz of 629108 oz gold and 32905 tonnes of copper (Newcrest
gold (Newcrest, 2002a). Mining Limited, 2009).
A feasibility study to re-establish mining operations at Telfer As of the end of 2009 the sublevel cave has broken through
was completed in 2002.This study concluded that concurrent to the western side of the open pit surface.
large scale open pit operation and sublevel cave underground The cave interaction with the western side of the open pit
operation was a viable proposition. topography will continue over the coming years and the longer
Telfer is currently operating one open pit mine, the Main term mine strategy suggests mining of the cave zone may be
Dome open pit, and one underground mine, the Telfer Deeps required to access further open pit ore at depth.
1. Senior Mine Planning Engineer, NewcrestMining limited, Telfer Gold Mine,Telfer WA 6762. Email: luke.cerny@newcrest.com.au
2. MAuslMM, GeotechnicalEngineer, Newcrest Mining limited Telfer Gold Mine, Telfer WA 6762. Email:Justin.dixon@newcrest.com.au
3. MAuslMM, Planning Superintendent, Newcrest Mining limited, Telfer Gold Mine,Telfer WA 6762. Email: Christian.Mcarthur@newcrest.com.au
4. PrincipalGeotechnical Engineer, NewcrestMining limited, Telfer Gold Mine, Telfer WA 6762. Email: Uday.singh@newcrest.com.au
L CERNY eta/
There have been numerous areas of interaction and risks difficulties may be experienced in the west wall of the Main
that have required management and mine planning. Some of Dome pit. The area would need careful management, however
the design, system and risk management methods used in this the measures outlined above indicate mining should be able
unique scenario are detailed in this paper. to continue without major interruptions (Newcrest, 2002b).
The study outlined a sequence where the open pit would mine
FEASlBILlTY STUDY through the cave breakthrough zone on a number of occasions
The feasibility study recommended that the Main Dome and through the life of mine. The first phase was outlined as a west
West Dome orebodies be mined as conventional open pit wall cut-back approximately two years after breakthrough. As
operations. An underground sub level cave operation was cave extraction was continuous through this time the broken
recommended for the Telfer Deeps orebody which is located cave column would empty and would require open pit mine
some goo m below what was on the western side of the Main waste placed as fill covering the void and supporting the walls.
A diagram confirming the methodology is shown following in
Dome open pit design at the time.
Figure 2.
Cave proximity and location is shown in Figure i with the
image depicting the M3o pit design with a cave shape outline. EVOLUTION OF PIT DESIGN
The cave shape was expected to be predominantly vertical.
The open pit operation at Telfer restarted in 2003 and the
The feasibility study anticipated that there would be cave
open pit sublevel cave was initiated in late 2006.
breakthrough five to six years after the commencement of
underground production (Newcrest, 2002a). As the operations continued the sequence, timings and
designs evolved in response to actual conditions and
A number of areas and mine planning requirements requirements. The philosophy was to monitor the cave
to mitigate risks associated with the impact of the cave performance and refine decisions such as pit ramp location in
breakthrough were suggested including: response to operating conditions (Newcrest, 2002a).
loading waste material into the breakthrough area to Timing changes meant the open pit had the ability to extract
provide support and prevent breakthrough area expanding, high-grade ore from the M i o and M i 2 reefs in the area where
the placement of an exclusion zone at some point as the the cave would breakthrough. This ore was previously left
cave approaches breakthrough removing personnel and for later stages post breakthrough or consumed in the cave
equipment from the area, subsidence column.
an open pit mining sequence which endeavours to avoid Incorporated in the designs associated with accessing the
scheduling staged material in the breakthrough area Mio ore was the introduction of a significant flat area on
around the time of breakthrough, the western side of the pit design around the cave extraction
alternative pit designs that provide alternative access by limits. This flat area became known as the cave breakthrough
positioning ramps away from the cave breakthrough zone bench.
and walls above the cave breakthrough zone, and The rationale of providing a large flat breakthrough bench
monitoring and management techniques developed at the included producing a geometry that would be more favourable
Ridgeway SLC to be utilised for the Telfer Deeps SLC.
~~
to breakthrough, would allow surface monitoring systems
The independent technical review of the Telfer feasibility to be easily installed and would allow backfill to be readily
project suggested that when the cave reaches the surface, introduced to the breakthrough area.
FIG 1 -Various views of theTelfer Main Dome M30 and SLC (Newcrest,2OO2a).
MINE PLANNING AND RISK MANAGEMENT ASSOCIATED WITH THE BREAKTHROUGH OF THE TELFER SUBLEVEL CAVE
FIG 2 -Telfer Main Dome cut-back schematic showing SLC (Newcrest, 2OO2a).
By 2007 the main dome stage 3 pit design had evolved to that RISK MANAGEMENT
shown following in Figure 3. The features of the mdstg3-ef
The risks associated with the interaction between the Telfer
design relevant to cave breakthrough were the breakthrough
open pit and underground SLC are managed by the Cave-Pit
bench, the dual access ramps to this breakthrough bench
and the main south ramp haulage some 70 m from the cave Interaction Major Hazard Management Plan (MHMP) and
breakthrough zone associated Trigger, Action and Response Plan (TARP). The
MHMP and the TARF are summarised from Singh, Dixon and
The location of the main south haulage ramp meant it was
McArthur (2010).
at risk from changes in cave apex location. The main ramp
was also reliant on wall stability both below the breakthrough
Major hazard management plan
bench and the high wall above the breakthrough bench.
The Newcrest safety and health management system required
Introduction of dual ramp access a formal major hazard management plan to be developed
for the cave-pit interaction. The MHMP for the cave-pit
During 2007 the reliance on a single pit floor access ramp that
interaction detailed the management requirements for those
had potential exposure from cave breakthrough was a risk for
hazards assessed as high or extreme with respect to the
which an engineering solution was sought. The stage 3 pit at
impact of the sublevel cave (SLC) on the open pit.
the time was also constrained from the as mined topography
so an additional ramp system was assessed to provide The MHMP summarised the critical elements of the:
production benefits. significant potential incidents (SPI) related to the
A number of ramp options were considered and the interaction that had the potential to cause multiple
evaluation narrowed on the design for a ramp that accessed fatalities;
the pit from the north thereby avoiding the cave zone. A contributing factors for the SPIs;
northern ramp for Main Dome stage 3 faced the major issue accountabilities for the monitoring, application and
of fitting significant haul length into a restricted pit area. To effectiveness of the controls to prevent an interaction
do so the design was aggressive with a number of switch-backs related SPI; and
resulting from fitting over 1200 m of ramp length into 500 m trigger and response plans (TARP) for the SPIs in order to
of pit circumference. detect signs and symptoms and respond proactively before
The intention was to develop the north ramp as a contingency they occur.
that supported the main southern ramp. The southern ramp The specific areas which were dealt with in this plan were:
provides clear waste haul cycle benefits, while evaluation cave propagation up to the initial 300 m exclusion pillar,
showed minor ore haul cycle upside from the north ramp. cave propagation into the initial 300 m exclusion pillar,
Overall the north ramp material compared well against and
the existing stage 3 material inventory with a stripping ratio cave breakthrough to the pit and subsidence crater
comparable to the design without the north ramp. The north forming in the pit.
ramp was eventually endorsed and construction began in Areas specific to backfilling activity were not addressed
2008 (Cerny, 2008). in the MHMP as these required separate risk assessments
The features of the Main Dome stage 3 pit design including involving issues such as working on a high wall and working
the north ramp are shown following in Figure 4. near the pit edge.
L CERNY eta/
FIG 3 - Telfer Main Dome stage 3 design mdstg3-ef.
Built into the system was the requirement to review the Trigger, action and response plan
MHMP at set trigger points to ensure that the plan remained The TARP identified the trigger points assigned to lead
valid, appropriate and practical as circumstances changed. indicators of potential hazards. These triggers provided
In early 2004, a risk assessment was conducted on the warning that an escalation of an unwanted event may occur
impact of the Telfer SLC on the open pit. Existing controls if action is not taken to manage the hazard. The triggers were
related to the potential loss scenarios were identified, and linked to prescribed actions to be undertaken by personnel in
recommendations for additional control procedures and designated roles and reported to those in accountable roles.
measures were drafted where the residual risks were high or The triggers were based on monitoring data and hence the
the existing controls could be improved. monitoring systems played an integral role in managing the
The existing controls were listed in a control management cave breakthrough.
plan to ensure that they were clearly defined and monitored Some examples of the types of triggers featured in the plan
for ongoing appropriateness and effectiveness. The control include:
management plan also listed the personnel responsible and
Cave back cannot be estimated and cave propagation
accountable for the actions, and a requirement for regular
review of the control. rate cannot be determined through extensometer, micro
seismic or open monitoring hole data.
The risk reduction plan was drafted to capture all Average propagation rate for a week exceeds 1.5 mmfday
outstanding actions that were required to reduce the risk or drops below 0.2 m/day. This was determined through
down to an acceptable level. These residual risks were closed extensometer, micro seismic or open monitoring hole
out if completed actions were sufficient to remove the risk or data.
if effective controls were implemented. Where controls were Cave stalled and/or air gap is sufficient to create a
implemented, these were then transferred to the control potential instantaneous collapse. Determined through
management plan. extensometer, micro-seismic or open hole data.
FIG 4 -Telfer Main Dome stage 3 design including north ramp.
Subsidence or Cracking noted on the south ramp or batters part of the meeting agenda. This board included the current
or berms above the ramp, from visual Observations, prisms TARP, geotechnical hazard map, the weekly cave status
or movement and surveying radar (MSR). report, diagrams showing the current cave proximity to the
Subsidence or slumping of the backfill tip-head, from pit and photographs showing the areas of the western wall to
visual observations. be aware of in relation to the cave breakthrough.
Evidence of rock noise, ground vibration, cracking or A weekly cave status report was issued, detailing any triggers
rock-fall events from walls above or below the south ramp, which may have occurred in the week as well as a summary of
determined through visual observations and seismic data. monitoring data and an example is shown in Figure 5.
Irregular movement in deep hole extensometers detected, At the end of each month a review of the TARP was conducted
ie movement in anchors closer to surface detected or a and all triggers were detailed in a report, which also featured
large increase in movement rate. a summary of all caving data for the month.
Wall movement indicated by MSR or Prism
This report was reviewed and signed off by the principal
instrumentation.
geotechnical engineer, production superintendent, open pit
Fill levels in vent rises slump by more than 5 m in a single mining manager and general manager. This ensured that all
event. This was determined through the alarm system on levels of management were aware of the cave performance for
VR8 and visual observations of the other rises. the month and could review triggered events.
Significant rainfall events. At each revision of the TARP,a training program was rolled
Geotechnical monitoring systems not operating. out where the workforce was made aware of the identified
Communication hazards, new controls and response actions.
Cave related information was communicated to supervisors The open pit production crews were presented with regular
through the daily production meeting. A cave interaction updates so that past triggers, changes in cave, subsidence and
board was set-up in the meeting room and was reviewed as response actions could be discussed and feedback given.
L CERNY et ol
FIG 5 -Weekly cave statusreport providing a quick overview.

SUMMARY OF MONITORING SYSTEMS surveyed. One of the monitoring holes drilled as part of the
extensometer program was left open so that this could also be
The risk management plan was supported by geotechnical used for camera surveys.
monitoring systems. The implementation of both subsurface
monitoring systems to measure cave propagation and surface Pit wall monitoring
cave influence monitoring and associated monitoring triggers Three main areas of monitoring were used to identify surface
at Telfer has been summarised below from Dixon, Singh and displacement of the rock mass. These were a movement and
McArthur (2010). surveying radar, a prism system and visual observations.
Subsurface monitoring Movement and surveying radar (MSR)
Three main areas of monitoring were used within the The movement and surveying radar (MSR), manufactured by
rockmass to determine cave location and shape. These were a Reutech Radar Systems has been used to monitor the western
micro-seismic array, a series of deep hole extensometers and wall cave influence area since February 2009. An output from
the use of camera surveying of open holes. the radar is shown in Figure 7.
The radar system was introduced to monitor the highwall
Micro-seismic array above the 5384 mRL bench and was also used to monitor
An real time micro-seismic system was installed underground the western wall below the breakthrough bench immediately
to monitor seismicty around the underground workings and above the south ramp.
to track the cave over its 800 m propagation to surface.
Prisms
Deep bole extensometers The Main Dome prism array consists of 158 prisms, with 55
A series of six deep hole extensometers as shown in of these located around the cave influence area, as illustrated
Figure 6 were installed from the 5384 mRL bench in August in Figure 6. All prisms are measured automatically from two
2008.The six holes were drilled to depths ranging from 180 - monitoringstations,with email alerts sent if prism displacement
305 m. The holes were installed prior to the cave breaching the exceeds defined velocity thresholds. Prior to the MSR arriving
300 m pillar between the cave and pit. In each hole 20 anchors onsite prisms were the primary form of surface monitoring.
were installed at equal spacing and connected to stainless
steel wires which were in-turn connected to spring loaded Visual observations
pulleys in the headframe. Displacement is measured using
Regular geotechnical inspections were undertaken whilst
potentiometers.
working within the cave breakthrough area, this included
Camera surveys of open holes routine photography.
Vent Rise 8 (VRS) was a 4.5 m diameter ventilation rise
associated with previous underground operations and was Cave apex tracking
located at the northern end of the cave breakthrough bench. A cave model was produced at the end of each month based
VR8 had been backfilled however the backfill material on all monitoring data and an example is shown in Figure 8.
emptied into the cave. This allowed the vent rise to be camera Initially the cave was modelled using micro-seismic data only,
FIG 8 - Cave model and micro-seismic event contouring at end of May 2oO9
(view looking east).
FIG 6 -Monitoring plan of cave influence area.
however once the cave was within 300 m of the 5384 mRL
bench data from the deep hole extensometers was used in
conjunction with seismic data to determine the cave shape.
From December 2008, the highest rate of cave growth
occurred at the north-eastern extents of the cave. One
additional deep hole extensometer was installed in response
to this trend of cave progression east towards the primary
haulage southern ramp.
Figure 9 shows the cave progression from December 2008
up until breakthrough in October 2009.
The global caving rate for the SLC was estimated 0.6 m/day
and the highest monthly rates recorded were up to 1.2 m/day
in the months leading up to breakthrough.
From August 2009 when the cave was estimated to be within
50 m of breakthrough, surface influence monitoring was used
as the primary means of determining cave proximity to the pit.
CAVE LOADING
The idea of loading material onto the cave zone was discussed
in the feasibility study in a number of areas because:
the extent of breakthrough zone could be limited if the
-t I area was supported or confined by backfill, and
after breakthrough cave material will continued to
FIG 7 - Radar monitoring: section looking west indicating wall movement areas be drawn so continuous backfill would prevent a void
relative to cave, October 2009 (view looking west). developing.
L CERNY eta/
The full preloading option was selected based on safety

aspects of the operation and the ability for all the benefits of
preloading to be in place earlier.
Design and implementation

The preloading was planned to take place in two phases.
Phase one involved dumping activities on the 5384 mRL
breakthrough bench level. Phase two involved dumping
activities from the top of the western highwall at the
5480 mRL.
Phase 7
Phase 1 commenced with the paddock dumping of the entire
5384 mRL bench. Following this a tip head was established
18 m above the 5384 mRL and this area was dumped out to
the limits of the bench. This was followed by the establishment
of a second tip head 32 m above the original breakthrough
bench RL.The final activity during phase one was the paddock
dumping out of this final 5416 mRL.
This phase was finished in May 2009 and approximately
i.iMt of waste material had been brought onto the
breakthrough bench.
The extensometers had been installed in locations across
this breakthrough bench. The extensometer headframes were
1
FIG 9 -Cave apex tracking December 2008 to October 2009 (view looking
protected by concrete culverts and covered with medium grade
rock material before backfilling. The cabling systems were
also protected. During the preloading activities the locations
were delineated to minimise haulage over the extensometer
north). headframes.
Loading of broken waste rock material onto the cave Afurther challengewas that the breakthrough bench contained
breakthrough area developed into an integral part of the two redundant ventilation rises that required exclusion zones
open pit cave management interaction. The open pit would be as determined by the site void management procedure. This
active at breakthrough and the SLC would be operational for reduced the working width on the preload bench and was
an estimated five years following cave breakthrough. compounded by these exclusion zones expanding with the
The rationale of conducting the material loading was that increased height as preload was introduced.
the material would limit rock mass damage and hence crater
extension through confinement on the crater limits. This Phase 2
would lessen the impact of the crater on both the stability Following the completion of phase 1, preloading activities
of the western highwall above the breakthrough bench and moved to the western highwall 5480 mRL. This phase
damage to the southern haulage ramp. The loading material composed of creation of a safe tip face so that waste material
would also provide a buttress effect on the western highwall. could be used to fill 65 m down to the phase 1 preload.
Three preloading options were considered: Phase 2 was completed by a combination of dump truck
1. full preloading - dumping as much waste as possible over
tipping, cast blasting and excavator side-casting.
the entire breakthrough bench, This phase was completed over the five months following
2. partial preloading - dump as much waste as possible only the completion of phase 1. Phase 2 material movement was
within predicted initial breakthrough area, and estimated at 1.0 Mt bringing the total material preloaded
to 2.1 Mt. Photos of the preload progression are shown in
3. post-loading - start dumping waste in the subsidence
Figure 10.
crater after cave breakthrough.
Each option had different timeframes for commencement Management of preloading risks
and completion, equipment selection, risk management
procedures and cave propagation monitoring strategy. The preloading of material into the cave breakthrough area
was managed under the MHMP and associated TARP. Area
The loading of waste in the cave area offered a reduction specific Job Safety and Environment Analysis (JSEAs) and
in waste haulage times so was viewed as an opportunity that risk assessments were used as required.
would benefit the operation.
Evacuation and barricading points were placed at the
Preloading offered the ability to complete the task without various entry points to the 5384 RL cave breakthrough bench
the use of remotely operated machinery and minimised the and the 5480 RL highwall. As per the TARP these barricade
risk of backfilling an empty crater. points were enacted following a trigger event.
The monitoring strategy for tracking cave progression One of the major risks associated with both phases of the
however would be more challengingfor the preloading options. preloading was the potential for air gap development in the
It would mean a requirement for all instruments and cabling SLC and the then sudden collapse of the pillar between the
to be adequately protected from dumping of the material and a cave back and 5384 mRL breakthrough bench. During both
requirement to allow for potential redundancy (Dixon, 2010). phases 1and 2 this area was monitored by the deep hole wire
MINE PLANNING AND RISK MANAGEMENT ASSOCIATED WITH THE BREAKTHROUGH OFTHETELFER SUBLEVEL CAVE
FIG 10 -Completed cave loading.

MAJOR CAVE BREAKTHROUGHAREAS
AND EVENTS
Continued use of the south ramp during cave propagation
had only been possible through the implementation and
compliance with the MHMP and TARP.
The major risk to the south ramp during cave propagation
was the impact the SLC would have on the slope stability and
increased potential for rock-fall or slope failure.
North-western failure
In October 2009 a slope failure of approximately 200 ooo
tonnes occurred in the western wall some 250 m north
of the cave breakthrough area. This failure has not been
directly attributed to the cave breakthrough however indirect
relationships to the breakthrough have been suggested.
1 The failure event was at the base of the south ramp. The
area of movement was identified by the radar system before
FIG 11-Snapshot of radar detected slope displacement- February 2010. failure and was being tracked. While the failure was assessed
extensometers, the Telfer underground microseismic system the south ramp was closed forcing all haulage via the north
ramp system.
and camera surveys down the open vent rise on the 5384 mRL
bench. This was the first significant radar detected failure event at
Telfer and as such provided useful information on calibrating
Preloading activities ceased and risk assessment reviews radar alarm thresholds. The information from the analysis
undertaken when critical crown pillar distances were of this failure event was applied to the trigger levels for
breached, namely 300,200and 150 m. radar alarms, especially the slope influenced by the cave
The 5384 mRL bench was permanently closed with breakthrough.
completion of phase 1 preloading in mid May 2009 at which At the time of the north western failure, the radar recorded
point the cave proximity was 162 m. Permanent closure of the movement in the slope influenced by the cave breakthrough
breakthrough bench was set to occur once the 150 m pillar had above the south ramp was approximately 0.6 mm/day. This
been breached. rate was below that required for exclusion of the area.
Phase 2preloading from the highwall above the breakthrough A short period of time after the failure, while the failure
mechanisms were being investigated, the movement rate in
bench occurred while the pillar thickness reduced.
the cave influenced slope above the south ramp went above
Consequently the risk of sudden breakthrough and drawdown trigger levels. This would have closed the south ramp if it had
of preload material had to be managed. An exclusion area was not already been closed.
applied using permanent delineation to keep personnel and
equipment behind the historic stable crest. This was aided Cave breakthrough
by the use of high precision GPS capable equipment and by Significant subsidence of the preload material occurred in
employing the technique of excavator side casting. mid November 2009 and represented SLC breakthrough.
L CERNY eta/
Subsidence was confined to the originally planned 5384 mRL

breakthrough bench.
The radar system detected an instantaneous collapse and
the slope movement rate after breakthrough was in the
vicinity of 0.8 mm/h. Total slope detected movement by the
radar system at this time was 650 mm. The prism monitoring
system suggested subsidence of 1.8 m and 25 mm/day in a
north easterly direction.
Current status
As of the end of February 2010,the western wall above the
south ramp had moved a total of 1.6 m as detected by the
radar. Wall movement rates had increased to 1 mm/h on
seven occasions since October 2009.The south ramp has not
been impacted by rock-falls and has not shown any significant
movement but will remain closed for the wet season.
Since early November the radar began to detect a small
section of slope in front of the cave apex as moving towards FIG 12 -Current slope condition.
the cave. As of the end of February 2010 total displacement The mine planning strategies enacted have meant continued
towards the cave had been in the vicinity of 225 mm. performance of the Telfer open pit. Having dual access ramps
The prism monitoring system is able to give an indication of to the base of the pit proved crucial as cave breakthrough
movement direction and subsidence that the MSR could not. affected the main southern haulage ramp. The proximity of a
This was north east and also a measurement of subsidence. main haulage ramp as less than 100 m from the breakthrough
Prisms installed in a berm 12 m below the breakthrough area was not ideal.
bench directly east of the cave breakthrough have shown
displacement up to 5.0 m. The planned large breakthrough bench area has so far
proved effective in providing a geometry that has limited the
Significant cracking and fretting of the pit wall above
breakthrough influence on the open pit operation.
the south ramp is visible with evidence of movement along
bedding planes and faults as shown in Figure 12. This is Longer term benefits of preloading are yet to be realised
confirmed by the degree of displacement detected by the however it is expected to assist in limiting the size of the crater
monitoring systems. and provide stability to the high wall above the cave.
Long-term cave subsidence ACKNOWLEDGEMENTS

A long-term strategy for the introduction of waste rock to the
cave subsidence column for the remainder of the life of the The Telfer cave interaction could not be managed without
SLC was required. strong commitment and teamwork from both the Telfer
underground and open pit operations, along with the
The methodology that was determined was to introduce
supporting planning teams.
waste rock mined from the open pit from the high wall at
the 5480 mRL. The backfill introduced will act as a buttress We would also like to acknowledge Newcrest Mining Limited
against the western wall slope however this will render for the opportunity to publish this paper.
monitoring techniques such as radar and prisms ineffective.
A monitoring strategy has been developed to manage
REFERENCES
the long term stability of this slope utilising seven In-Place Cerny, L, 2008. Main Dome Open Pit Stage 3 North Ramp Design
Change Report, Newcrest Mining Limited Internal company
Inclinometers (IPIs) with data loggers and remote connection report.
and an additional three manual inclinometers.
Dixon, R, Singh, U and McArthur, C, 2010. Interaction between
Phase one involved the installation on four 100 m deep IPIs a propagating cave and an active pit at Telfer Mine - Part 11:
along the 5480 mRL highwall behind the preloading tip-head. Monitoring Interaction, in Proceedings Cauing 2010, pp 307-320.
Phase two will initiate if movement is identified in the phase 1 Newcrest Mining Limited, 20028. Newcrest Mining Limited Telfer
inclinometers and will involve the installation of a second and Project feasibility study, November 2002, Telfer Project, Newcrest
third row of inclinometers further behind the pit crest. Mining Limited.
All backfilling operations are then planned to move to behind Newcrest Mining Limited, 2002b. Independent technical review -
the second row of instrumentation with access between this Telfer Development proposal, Behre Dolbear Australia FTy Ltd,
row and the pit edge to only be with remote machinery. December 2002.
Newcrest Mining Limited, 2009. Annual results twelve months
CONCLUSION ended 30 June 2009 [online]. Available from: <http://www.asx.
com.au/asx/statistics/announcements.do?by=asxCode&asxCo
The breakthrough of the Telfer sublevel cave into the active de=NCM&timeframe=Y&year=2oog,3lkgqdbw22kfgr[l].pdf>
Telfer open pit has been a novel experience. Up to writing [Accessed: 7 March 20011.
in February 2010 the cave breakthrough had been managed Singh, U, Dixon, R and McArthur, C, 2010. Interaction between
successfully under the major hazard management plan. a propagating cave and an active pit at Telfer Mine -
Because there was very little history of similar situations a Part I: Interaction management, in Proceedings Cauing 2010,
conservative risk management strategy was adopted. PP 321-332.
Proximity Cameras and Global
Positioning Systems - An Integrated
Approach -
N Ferreira and C McElman
ABSTRACT
Safe, efficient production is a common objectivefor all large open pit mines. Despite comprehensive
policies, procedures and regulations aimed at protecting workers, several people die each year
in mining equipment-related collisions. The limited sightlines of heavy equipment operators
coupled with difficult operating conditions are major contributing factors to the majority of mining
equipment collisions. Off road haul trucks are the most common type of machinery involved in
surface accidents. These vehicles represent the primary target for improving safety performance
by mine operators. Although several commercial products have been introduced to address these
issues, none have achieved significantmarket acceptance. The current collision avoidance solutions
lack integration with onboard fleet management and equipment health systems. To achieve
acceptance in industry, any collision avoidance system must offer high reliability, minimal false
alarms, and value for money. With these criteria in mind, this paper will examine the pros and cons
of the various technological approaches to collision avoidance in an open pit mining environment.
The unique global positioning systems (GPS) and camera-based proximity detection solution
on offer from Modular Mining Systems will be described. This solution uses intelligent filters to
minimise false alarms. Tight integration with fleet management provides context information to
enhance warnings and enable back-office near-miss analysis. Available today, this combination
creates a practical solution to improve the situational awareness of heavy equipment operators, at
a compelling price point. Future directions for maximising heavy equipment operator situational
awareness will also be examined.
INTRODUCTION
The United States Department of Labors Mine Safety and factors such as rain, snow, fog, dust and operating at night
Health Administration (MSHA) reported in January 2010 (poor illumination) that also impair the operators visibility
that in the USA, on average, 40 incidents with five to six around the vehicle.
fatalities involving powered haulage equipment occur each Research based on data collected from the MSHA fatality
year at surface mining operations (MSHA, 2010). reports show the equipment types with the highest number
Research in mining related accidents in Queensland and of fatal accidents were dump trucks (38 per cent) followed by
front-end loaders (33 per cent). The majority (67 per cent) of
New South Wales, Australia has shown that there were 147
accidents involved obstacles or workers in the rear blind area
accidents involving heavy vehicles resulting in two fatalities of reversing equipment (Ruff, 2007).
during the past ten years (Mark and Verhoef, 1999).
Collision warning systems utilise sensors for object detection
With in excess of 32 ooo off-highway haul trucks operating and an operator interfaceto provide suitableaudioand/or visual
worldwide (The Parker Bay Company, 2010) the potential warning to indicate the proximity of a fixed or mobile hazard to
for haul truck collision related fatalities is very high. By the operator. Typical sensor technologies include radar (pulse
simple extrapolation of the annual powered haulage unit or Doppler), ultrasonic, infrared, radio transceiver tags, GPS
fatalities in the USA (-6400 units), the worldwide powered or cameras. Each of these proximity sensing technologies have
haulage fatality rate can be estimated at 30 per annum. This inherent strengths and weakness that will be explored.
translates into a staggering estimated cost of US$31.8 million
per annum, based on the National Institute for Occupational Limitations of existing systems, include frequent false
Safety and Health median cost of mining related fatalities of alarms, limited detection range, and a lack of specific
US$i.06 million per incident (NIOSH, 2006). information on an obstacle, eg location, identity (Ruff, 2004).
Limited visibility is the major factor attributed to collisions
involving off-highway trucks (Mark and Verhoef, 1999). In TECHNOLOGY
addition to the inherent blind spot problems present in large There are various technologies available that can be used
haul trucks, as shown in Figure 1, there are environmental to assist with the detection of obstacles that are not always
1. Senior Product Specialist, Modular Mining Systems Inc, 403-2071 Kingsway Avenue, Port Coquitlam BCV3C 6N2, Canada. Email: ferreira@mmsi.com
2. Product Manager, Open Pit, Modular Mining Systems Inc, 403-2071 KingswayAvenue, Port Coquitlam BCV3C 6N2 Canada. Email: mcelman@mmsi.com
N FERREIRA A N D C McELMAN
FIG 1- Blind area diagram for a Euclid EH4500 280-ton-capacity haul truck - 1.5 m plane (source: Ruff, 2007).
visible from the operators cab and provide a suitable warning Cameras are passive in nature and therefore can only provide
to the operator. Numerous commercial off-the-shelf devices suitable coverage of operator blind spots and cannot provide
are available, however, the majority of these were developed warning alarms or distances to potential hazards around the
for the on-highway automotive industry and are not suitable vehicle. Due to this passive nature, equipment operators need
for use in the harsh, off-highway mining environment. Some to be diligent in the use of cameras when they are deployed as
of this technology been ruggedised and commercialisedfor the
a standalone proximity detection solution.
mining industry but the market acceptance has been relatively
low. This is due in part to the high capital expenditure required
to retrofit the technology to the equipment. Radar
Some of the more popular technologies available are A radar system emits electromagneticwaves from a directional
described in this paper. It should be noted that there are also antenna which are reflected off objects in the path of the
other technologies being used for proximity detection. radio beam. Radar can be used to identify the range, altitude,
direction, or speed of both moving and fixed objects. There
Cameras are two main types of radar suitable for proximity detection
High image quality video cameras housed in ruggedised systems.
enclosuredesignedto withstand the harsh mining environment
can be mounted in strategic locations on heavy equipment. Pulse radar
Cameras have been in use for some time to provide blind spot
The distance to an object is calculated by transmitting a short
coverage around haul trucks. These cameras are typically
hardwired to a liquid crystal display (LCD) mounted in the pulse of radio wave and measuring the time it takes for the
operators cab where the video feed is displayed. The operator reflected signal to return to the antenna. A s the receiver cannot
can select which camera to view or the switching process can detect return signals whilst transmitting, the devices must
be automated based on specific actions such as dynamically switch between transmit and receive mode at a predetermined
switching to forward or rear facing cameras based on gear rate. This makes it difficult to combine short range and long
selection. range radar in one device.
PROXIMITY CAMERAS AND GLOBAL POSITIONING SYSTEMS -AN INTEGRATED APPROACH
Doppler radar to indicate the type (haul truck, services vehicle, cable tower,
Doppler radar makes use of the Doppler effect to produce personnel, etc) and ID of the tag, ie 'Light Vehicle 010'.
range data about objects at a distance. This is achieved by Radio frequency tag based systems typically do not produce
transmitting a signal and analysing how the original signal false alarms. However, they cannot provide location or
has been altered by the object(,) that reflected it. Variations travel direction specific information about the hazard to
in the frequency of the signal give direct and highly accurate the operator. All objects requiring protection need to be
measurements of a target's velocity relative to the radar source equipped with WID tags.
and the direction of the transmission. Doppler radar detection
requires either movement of the obstacle or the vehicle. In Global positioning systems
order to provide 360"coverage around a haul truck, requires Differential global positioning systems can provide the exact
up to eight antennae be mounted at know distances (between location of all GPS equipped vehicles in real time. GPS, on its
three and 6 m) around the periphery of the vehicle (Glynn, own, cannot warn on the proximity of other vehicles or fixed
2005). hazards. However, a solution that enables GPS equipped
vehicles to transmit their position via a radio network so that
Passive systems like radar depend on signal reflections vehicles are aware of the locations of all the other equipment
from nearby objects and will alarm more often than active units relative to their own position will be able to provide the
tag-based systems (Ruff, 2007).The generation of these false required proximity warnings to equipment operators. When a
alarms lessens the effectiveness of the solution as equipment vehicle detects that there is a hazard within a predetermined
operators become desensitised by the false alarms. area surrounding the vehicle, an alarm is generated to warn
the operator. The location and hazard zones of stationary
Radio frequency identity detection (RFID) obstacles can be uploaded to the equipment via the radio
RFIDs generally fall into one of three broad categories: network in order to provide suitable protection for these
I. passive transponders (tags) which have no battery and objects. Stationary obstacles do not need to be equipped with
rely on the external signal as a power source, GPS receivers or radios as their locations are already known
2. active transponders which have a battery and therefore
to each of the mobile units. GPS based systems are suitable
can transmit signals autonomously,and for both close range(<so m), slow speed proximity detection
(<lokm/h) and long range ( X O m), high speed (10 km/h to
3. battery assisted passive transponders which have a 50 km/h) proximity detection.
battery to boost signaltransmission but require an external
Currently GPS based proximity detection systems cannot
signal to activate the device.
easily provide protection to pedestrians due to the lack of a
RFID tags are further classified as either low frequency suitable portable radio to package with a miniaturised GPS
(<500kHz), suitable for short range detection or high receiver and CPU.
frequency, suitable for long range detection.
Preliminarytests at a surface mine showed that a GPS-based
WID tag based systems consist of two components, a proximity warning system has the potential to significantly
transceiver and a transponder. The transponder tags are reduce accidents involving collisions or driving over an edge
attached to objects requiring protection, such as light vehicles, at surface mining operations (Ruff, 2004).
stationary objects, personnel, etc and remain passive until
energised by a transceiver signal. One or more transceivers
are attached to all of the large equipment. When a transceiver CONCLUSION
detects a transponder within a predefined range, an alarm is There are many factors to consider when implementing a
generated. Two-way communication between the transceiver proximity detection solution and a careful analysisof operation
and transponder is possible, enabling the alarm to be specificrisk factors should be conducted. The appropriateness
triggered at both units. In order to provide the operator with of each of the technologies can be evaluated using Table 1,
context based information each tag needs to be preprogramed which shows the kinematic calculationsfor stopping distances
TABLE 1
Technology comparison of stopping distancesto prevent a collision (abbreviated). Source: Faul eta/, 2007.
Will the technolow be able to warn the operator in time to allow the machine to be brouqht to a full stop or avoid a potential collision?
Speed Stopping distance Short-range Low frequency High frequency GPS long-range
(km/h) (m) radar RFID RFID radar
5 0.9 Yes Yes Yes Yes Yes
35 I 64.4 I No I No I No I Yes I Yes
45 99.4 No No No Yes Yes

and times for a loaded haul truck driving downhill on a ten per alarms. Based on careful analysis of the specific scenarios
cent gradient (Faul et al, 2007). being addressed, non-threatening proximity scenarios can be
It is apparentfrom the review abovethat there are technologies filtered out by setting speed thresholds, above which alarms
availablethat can be used to improve the situational awareness are ignored, or equipment type exceptions. For example, a
of heavy equipment operators. On their own, each of the truck can be configured to generate a proximity alarm when
technologies can provide a reasonable level of proximity approaching a shovel and the shovel can be configured to
detection,however, due to the sensor specifictechnological and ignore trucks or to alarm when a service vehicle is within the
operational limitations proximity detection systems should warning zone.
employ two or more complementing technologies in order to The onboard field computers update each other with position
be effective. information, in a peer-to-peer fashion, at configurable
refresh intervals. The Proximity Detection Module uses the
HUMAN FACTORS continuously updated, differentially corrected positions of
each GPS equipped vehicle as it moves around the mine to
Overview assess the proximity risk level. Fixed hazard locations are
The successful implementation of a new tool or technology created on the central server and are dynamicallyuploaded to
requires judicious application of hardware, software, each field computer via the radio network.
processes and leadership. If all of these dimensions are not
addressed, an otherwise viable technology solution will fail The Minecompass navigation map is used to augment
due to misuse, rejection, or a lack of confidence. Success the distance based proximity detection by displaying and
requires clear communication of system capabilities and continuously updating the location of all equipment on site
limitations, along with post-installation follow up to ensure relative to the current unit. The current mobile unit is always
benefits are sustained. displayed in the centre of the graphical console and the map
Proximity detection systems represent an additional scrolls as the unit travels along the route. All equipment icons
demand on a mining equipment operators fixed attention are displayed in the actual geographic location and with the
correct orientation based on the direction of travel of the
budget. Detection technology may initiate a timely warning
vehicle. Context information, such as equipment ID and load
of an unseen hazard, but the structure of this warning
status (empty or full), about all the equipment visible on the
will strongly influence comprehension. Warnings must be map is also displayed by Minecompass.
structured to integrate into an operators existing workflow.
Dashboard-style at-a-glance indicators can be processed When the on-board field computer determines that another
more quickly than text messages. Audible warnings convey vehicle or a fixed hazard is within the configured warning
zone (distance), an alarm is generated by the on-board field
information without requiring an operator to divert their gaze
computer and displayed on the graphical console alerting
to a computer screen.
the operator to the hazard, as shown in Figure 2. In addition
Multiple alarm thresholds provide a dual function. Early to real-time notification to equipment operators, proximity
warnings can trigger an operators alertness level, reducing closecallsare relayed to the dispatcher as exceptions, and are
response times in the event a critical alarm event should logged in the central database on the control room server.These
occur. Warning alarms also provide a positive validation that exception logs can later be used for analysis and reporting and
the system is functioning, improving operator confidence in are an invaluable tool in accident investigations. Analysis of
the system. the logs can be done to identify safety risks, such as operators
requiring additional training or high risk operational areas.
INTELLIMINE PROXIMITY DETECTION
Warning levels
Overview The Proximity Detection Module offers two severity levels, the
Modular Mining Systems Inc has developed a proximity first alarm level is classified as a warningwhilst the second
detection system that is fully integrated with its production alarm level is classified as danger. Each severity level is based
monitoring and fleet management solutions. The IntelliMine on a configurable distance based zone (radius) around the
NextGen Proximity Detection module is specifically designed mobile unit, the distances for the two levels of alarm can be set
to improve the situational awareness of mobile equipment globally or specified per equipment type. When a hazardous
operators by providing visual and audible alarms of proximity object or other mobile equipment is detected within the
to other enabled mobile equipment or fixed hazards in the area. outer safety circle an audible tone sounds in the cab and a
Distance-based proximity detection is particularly beneficial visual warning flashes on the onboard display indicating the
in heavily congested areas such as loading and dumping severity of the alarm and the relative location of the hazardous
locations, which often require low-speed maneuvering in obstacle, ie front, left side, etc. On the mobile console, a yellow
large-equipment blind spots, reversing, and driving around
warning circle is shown around the vehicle and a type-specific
obstacles. GPS distance-based proximity detection has been
icon representing the hazard is shown in the exact location
used for many years at Provision sites around the world and
along with other information such as unit ID. When the
was released to DISPATCH Classic customers in 2009. With
the release of DISPATCH NextGen, proximity alarms are obstacle is detected within the inner safety circle a second
augmented via a map-based display of relative equipment alarm is generated and a red danger circle is shown on the
positions, along with optional blind spot video camera mobile console, as shown in Figure 3. The relative position
integration. and severity status is also updated on the warning icon.
The mobile console will continue to display the warning
Solution state by showing the safety circles, the updated locations of
Using existing on-board hardware, wireless communications the hazards or other mobile equipment and the warning icon
infrastructure and integration with the central control server, until such time as the on-board computer determines that
the systems context-awarenessis designed to minimise false there are no longer any risks in the warning zone.
PROXIMITY CAMERAS AND GLOBAL POSITIONING SYSTEMS - AN INTEGRATED APPROACH
FIG 2 - Initial alarm on mobile console indicating a warning -object in front.
FIG 3 - Mobile console displays warning state, positionand ID of the othervehicles.

RoadMap
Although haul-truck to haul-truck collisions are a concern, over the radio network an alarm is generated on the mobile
collisions between haul trucks and light vehicles represent console, Figure 4, to alert the operator that proximity
a significant hazard as well. An upcoming release of the detection is not functioning. The proximity failure alarm will
RoadMap solution, for light vehicles, is fully compatible with repeat every ten seconds until such time as the system health
the Proximity Detection Module in DISPATCH. This will allow is restored to an acceptable operational level.
mine operators to improve mine site safety by equipping light
vehicles with an inexpensive solution that uses the existing On-board cameras
network and control room infrastructure. RoadMap will
inform haul trucks working in congested areas, such as shovel The Proximity Detection Module can be configured solely on
pads and crusher areas, that a light vehicle is in the area and GPS, but the integration with on-board cameras, a concept
will also provide heavy vehicle proximity alerts to the drivers that was shown at MINExpo in 2008,is also available. This
of the light vehicles. Proximity Detection in RoadMap provides solution integrates the GPS based proximity alarm notifications
the same level of configurability and alarm generation as in with up to four on-board video cameras managed through an
DISPATCH NextGen. onboard video server. When a proximity alert is generated the
system will automatically switch the mobile console display to
System health status show the video stream from the camera facing the direction
The on-board computer continuously monitors the health of the hazard. When multiple hazards are present the video
status of both the GPS signal and the radio network to ensure display area will automatically split to show the video stream
that proximity detection can function correctly. When either from all affected cameras, as shown in Figure 5. This provides
the positional dilution of precision (PDOP) falls below a an additional level of information to the equipment operators
Moderate rating or the there is a break in communication to use when dealing with proximity alarms.
FIG 4 - Proximity health warning.
FIG 5 - Mobile console showing camera view and Minecompass.

The operator also has the ability to switch to camera view miss events, management, supervisors, trainers and safety
manually and can display video streams from one or all four personnel can identify high risk areas, undesirable work
cameras simultaneously, irrespective of whether there is a practices or unsafe operating conditions. When used to review
proximity alarm or not. collisions and near-miss incidents, the playback functionality
provides the means to perform a geospatial graphical analysis
Advantages of events and also provides a unique training support and
The IntelliMine Proximity Detection module utilises existing analysis tool.
onboard hardware and on-site wireless communications
infrastructure. By avoiding additional on-board systems Scenarios
mounted in the operators cabin, capital, maintenance, and Currently, no single technology can perform in all proximity
supply chain costs are reduced. The Proximity Detection detection scenarios. It is essential that mine management
Module is fully integrated with the IntelliMine system, making identify which types of situations are addressed by an installed
use of existing equipment records, operator information, proximity detection system.
equipment operating state and road topology. This enables
The following scenarios are targeted with the initial version
the Proximity Module to provide context rich information
of Proximity Detection:
in order to greatly improve the situational awareness of
equipment operators. heavy-vehicle-to-heavy-vehicle,
The extensive data logging capabilities can be used to heavy-vehicle-to-light-vehicle,
generate site-specificincident reports for safety or Regulatory heavy-vehicle-to-fixed-hazard,and
compliance reporting. By capturing and quantifylng near- light-vehicle-to-fixed-hazard.
PROXIMITY CAMERAS AND GLOBAL POSITIONING SYSTEMS - AN INTEGRATED APPROACH
Future direction is essential. System simplicity, minimisation of false alarms,

and integration into existing screen layouts and workflows
Future development plans include the integration with
will maximise the signal-to-noise ratio of proximity detection
additional proximity-sensor triggers using separate
information. Ongoing validation of operator acceptance,
technologies such as W I D or radar. By building a pluggable
utilisation and satisfactionis required to ensure that proximity
architecture for interfaces, Modular Mining Systems expects detection systems provide a sustainable gain in operator
to be able to support multiple proximity sensing technologies situational awareness and mine-site safety.
in order to support customer preference.
CONCLUSIONS ACKNOWLEDGEMENTS
At a minimum, a proximity warning system should be able to We would like to acknowledgethe followingfor their assistance
detect obstacles in close proximity to a vehicle and alert the with this paper:
operator, with as few false alarms as possible. The ability to a Dr Andree Rottig, Product Manager, Machine Guidance;
perform self diagnostic routines in order to determine the 0 StephanC Cantin, Product Manager, Underground
health of the onboard hardware and provide feedback to the Systems; and
operator on the status of the system is essential. a Modular Mining Systems Inc, for allowing us to present
Additional value can be realised by expanding the solution, this paper.
as shown in Figure 6, where each layer of functionality
provides increasing value. The addition of context rich REFERENCES
information allows equipment operators to make informed F a d , H, Rupial, M, Munoz, 0, Niven, S, Naidoo, A, Janicijevic,
decisionswhen dealing with proximity alarms. Proximity alert D and Nhlapo, V, 2007. Anglo American - Collision avoidance
logging, analyse and reporting is an invaluable management systems technical report [online]. Available from: <http:J/
tool allowing mine operators to enhance existing safety and www.anglochile.cl/en/doc/AFRS/o2%~0Est%C~%Aindar%~0
Equipos%2oM%C~%B~viles%2ode%20Superficie/SME%~0
training process. Appendix%~oo8%2oCAS%2o~Collision%~oAvoidance%~0
System.pdf>.
Glynn, P, 2005. Collision avoidance systems for mine haul trucks
and unambiguous dynamic real time single object detection,
doctorial thesis, pp 28.
Mark, R and Verhoef, H, 1999.Collision avoidance systems for large
mining trucks, in Proceedings Queensland Mining Industry
Health and Safety Conference 1999,179p.
Mine Safety and Health Administration (MSHA), 2010. Accident
Prevention through the use of new technologies [online].
Available from: <http://www.msha.gov/Accident_Prevention/
newtechnologies/CollisionAvoidance/CollisionAvoidanceasp>.
National Institute for Occupational Safety and Health (NIOSH),
2006. Fatal occupation injury cost fact sheet: Mining [online].
FIG 6 -Value adding components to proximity warning solutions. Available from: <http://i~8.2~6.~8.2i/niosh/docs/~006-i~~/
default.html>.
Thus far, unit costs have posed a major hurdle to the Ruff, T, 2004. Advances in proximity detection technologies for
adoption of proximity detection solutions. Modulars GPS surface mining equipment, in Proceedings 34th Annual Institute
on Mining Health, Safety and Research, Salt Lake City, 25 May,
based approach provides an affordable solution by leveraging
8 P.
the mine sites existing investment in on-board computing,
telemetry infrastructure and central reporting applications. Ruff, T,2007. Recommendations for evaluating and implementing
proximity warning systems on surface mining equipment,
Simple adoption does not equate to successful, sustainable National Institute for Occupational Safety and Health, Report of
value in use. In order to be successful in the longer term, a Investigations 9672,Department of Health and Human Services,
proximity detection solution must take human factors into 2007,publication no 2007-146[online]. Available from: <http:/J
account. During implementation, a careful alignment of www.cdc.gov/niosh/mining/pubs/pdfs/zoo~-i~6.pdf~
operator expectations with system capacities and objectives The Parker Bay Company, 2010.Equipment database, 31 March.
PayloadVariance Effects on Truck
Bunching
P F Knights' and S Paton2
ABSTRACT
Data collected from payload management systems at a number of surface mining operations show
significant variance. Heavily loaded trucks travel slower up ramps than lightly loaded trucks.
Faster trucks are slowed by the presence of slower trucks, resulting in 'bunching' and production
losses. A stochastic model was constructed to simulate a truck haulage return cycle of 12.5 km
at a large open pit gold mine in the USA. For a fleet of 35 trucks, the best estimate of the benefit
achieved by a sustained reduction in payload variance of 25 per cent is a throughput increase of
1.6 per cent, within a possible error range of 0.6 to 3.2 per cent. This potential benefit is worthy
of further investigation by Continuous Improvement teams at large open pit mines. The model
also indicates that, under the right circumstances, provision of an overtaking lane can provide
throughput benefits of up to ten per cent, The feasibility of providing unidirectional haulage
circuits within large open pit mines should be examined.
INTRODUCTION
The use of on-board payload measurement systems on mining
trucks has become more prevalent in recent years. Whilst this
has undoubtedly facilitated production management, it has
other uses such as monitoring fleet application so as not to
void original equipment manufacturer (OEM) warranties
on key components. The 10:io:zo rule first advanced by
Caterpillar Inc is a widely quoted empirical rule, whereby ten
per cent of payloads are permitted beyond ten per cent rated
load but no payload should exceed 20 per cent overload. -
"
..
Y) l
Analysis of truck payload data from a number of sites around

Payload (tonner)
the world suggests that this is difficult to achieve in practice.
The variance associated with truck payloads is typically large.
FIG 1 - 930E Payload distribution (after Webb, 2008).
Figure 1 shows a histogram of payloads recorded for a
fleet of Komatsu 930E trucks having a rated payload of benefits of reducing payload variance in a fleet of Komatsu
304 tonnes (320 tons) operating at a large open pit in the 930E haul trucks operating in a large open pit gold operation
United States. The mean load is 293 tonnes with a standard in the USA.
deviation of 20.5 tonnes. Although there are no instances of
payloads exceeding 20 per cent rated load, operating practices METHODOLOGY
at the mine specify payloads to be within a fzo tonne range. It
can be seen that trucks are overloaded approximately 20 per Simulation model
cent of haulage cycles and under-loaded a further 20 per cent
of cycles. Discrete simulation models capable of simulating the
stochastic nature of batch processes have been widely applied
Overloaded trucks are slower up ramp in comparison to to analyse queuing losses in truck/shovel systems. Some
under-loaded trucks. In large open pits having long ramps,
bi-directional traffic and restrictions on haul road widths published examples are by Runciman, Vagenas and Corkal
negate the possibility of overtaking. Thus faster trucks can (1997), Ataeepour and Baafi (1998), and Sturgul(1999).
be delayed behind slower trucks in a phenomenon known as In this study, the simulation language GPSS/H (see Banks,
truck bunching. This is a source of considerable productivity Carson and Ngo Sy, 2003) was used to model the haulage
loss for truck haulage systems in large open pits. circuit of the 930E fleet operating in the large open pit gold
operation. The model was constructed in three sequential
OBJECTIVE sections (Paton, 2009);
This paper aims to quantify the potential throughput benefits 1. a truck generator and loader phase,
as a result of tighter control on truck payload distributions. 2. the haul route phase, and
The paper applies a simulation model to estimate the potential 3. a return and counting phase (see Figure 2).
1. MAuslMM, BMA Chair and Head of Division of Mining Engineering,The Universityof Queensland, Brisbane4072. Email: p.knights@uq.edu.au
2. Graduate Mining Engineer,The Universityof Queensland, Brisbane 4072.
P F KNIGHTS AND S PATON
tm
FIG 2 -Simulation model overview.
The model begins by generating transactions (trucks), which variance occurs only during loaded travel times, cycle time
queue at the loader. Each truck is assigned a random payload data was modified by subtracting loading, dumping and return
from the distribution shown in Figure 1, and a loaded travel times from total cycle time. Loading times were assumed at
time is estimated using empirical and Monte Carlo methods 2.6 minutes, three times the average shovel cycle time.
(see below). This travel time does not include time spent in Dumping was assumed at two minutes and the return leg of
queues. Trucks then progress to the haul route section of the the circuit was estimated at 12.5 minutes, based on average
model provided there is less than the required number of speed of 30 k/h over 6.25 km. Figure 3 presents a scatter
trucks in the circuit. plot of the estimated loaded travel time as a function of truck
In order to simulate truck bunching, the haul route is divided payload (outliers have been excluded).
into 150 segments. Each segment may only contain one truck, The linear regression equation for Figure 3 is:
thereby denying entry of any additional trucks. Additional
trucks will form a queue behind the segment. Once a truck Travel time = 0.1829 x payload + 25.526 (1)
has left an occupied segment, a queued truck can progress
into the segment or advance in the queue if more than one
truck is queued. Since there are 150 segments, trucks traverse
through each segment in one 150th of the haul route travel
time assigned at the loader. This means that each truck is not
allowed to be less than one segments travel time behind the
previous truck, creating a bunching effect.
After traversing the haul route, trucks accumulate a
dump time and return time before leaving the haul circuit.
Cumulative truck payload is counted in this section. Rather
than returning to the beginning, the truck is terminated and
340
a new truck is generated. This is to ensure that the random
number stream continues to progress, and it also enables the
effect of varied fleet size to be studied. The model continues to FIG 3 - Haul route loaded travel time versus truck payload (after Webb, 2008).
run until the timer runs out. The time unit used in the model is
minutes. The model was run for 720 minutes (12hours) for all In order to accurately model bunching, a deterministic
simulations. After each simulation, the counter measurement relationship must be derived between payload and travel time.
indicates the throughput tonnage of the haul circuit. The linear regression equation forms part of this relationship.
A residual plot was constructed by subtracting estimated
Input data travel times from actual travel times (see Figure 4).
Payload and haul cycle data were sourced from an open pit The residual plot does not reveal any underlying relationships
gold operation in the USA, as specified by an undergraduate and a normal distribution was assumed. Standard deviation
thesis developed by Webb (2008).The haulage circuit is a of the residuals is 9.6 minutes. Haul truck travel times were
12.5 km round trip, and P&H 4100 x PB electric rope shovels therefore estimated as a function of the empirical component
are used to load a fleet of Komatsu 930E haul trucks. derived in Equation 1 and a random component derived from
a Monte Carlo simulation from the normal distribution of the
The payload frequency histogram is normally distributed as residuals.
shown in Figure 1, with a mean of 293 tonnes and standard The resulting GPSS/H model was run multiple times to
deviation of 20.5 tonnes. The GPSS/H model was programed simulate throughput tonnages for different haul circuit
to distribute payloads according to this normal distribution. parameters. Payload distribution was altered as a percentage
The standard deviation was later modified to test the effect of of the original standard deviation to determine the effect of
reducing payload variance. different variances.
Haul cycle time data were used to model the effect that truck Since the model uses a loaded cycle time variance that was
payload has on truck speed. Since truck bunching due to load measured for trucks subject to truck bunching, it is reasonable
PAYLOADVARIANCE EFFECTS ON TRUCK BUNCHING
64 216 tonnes, or a 4.2 per cent increase for the fleet of 35

trucks (see Figure 6). Paradoxically, a 50 per cent increase in
payload variance (150 per cent times the base case) will lead
to a reduction of throughput of -4.4 per cent. Payload variance
has a lesser effect on throughput when the haulage circuit is
over or under trucked.
-20 0
25 0
FIG 4 - loaded travel time residual plot.

to assume that the variance of loaded travel time will also -G W
decrease as payload variance decreases. As the nature of
this dependency is unknown, simulations were run for three FIG 6 -Throughput difference relative to the base case
different scenarios; a reduction in loaded travel time variance (051 variance reduction).
equal to the reduction in payload variance (1:i relationship);
a reduction in loaded travel time variance equal to half that of Similar results were found for the scenario in which a i:i
the reduction in payload variance (0.5:irelationship); and an relationship was assumed between the reduction in loaded
uncoupled scenario in which the loaded travel time variance travel time variance and the reduction in payload variance In
remains unchanged as payload variance changes. Because this case, for the fleet of 35 trucks, the model predicts a 3.2 per
the loaded travel time variance contains sources of variance cent throughput gain as a result of a sustained reduction in
other than truck bunching (for example, auxiliary equipment payload variance of 25 per cent. In the uncoupled scenario in
interactions, and truck operator variance), the most likely which loaded travel time variance does not vary as a function
scenario is considered to be the 0.5:1 scenario.
of payload variance, a sustained reduction of 25 per cent in
All results are presented as an average of 50 simulations. payload variance will provide only a 0.6 per cent throughput
Simulationswere also performed where bunching was omitted gain.
from the model, consideringa scenario where faster trucks are
able to overtake slower trucks up ramp. For the fleet of 35 93oE trucks, the best estimate of the
benefit achieved by a sustained reduction in payload variance
RESULTS of 25 per cent is a throughput increase of 1.6 per cent, within a
possible error range of 0.6 to 3.2 per cent.
Truck payload variance reduction
Figure 5 illustrates the relative throughput differences in the Overtaking and no-bunching analysis
case that the loaded cycle time variance reduces at one half
that of payload variance reduction. The model was restructured to allow two trucks to
simultaneously occupy the same haul segment. This
effectively simulates a haul route where trucks are able to
overtake one truck at a time. A model without bunching was
also constructed, resulting in a simulation where any number
of trucks can simultaneously overtake.
Figure 7 shows that, by enabling overtaking, throughput
increased over ten per cent in simulations of fleet sizes of 30 to
40 trucks. The results for a circuit without bunching increase
at a constant rate with the number of trucks. This is logical
since the only queuing experienced by the circuit is at the
loader and is insignificant compared to the size of the circuit.
FIG 5 -Throughput comparison (051 variance reduction).

The base case haul circuit model reaches an optimal peak
throughput of 68 426 tonnes with a fleet of 40 trucks. Only
marginal throughput increases are obtained if the fleet size
is increased beyond 40 trucks, indicating an over-trucked
scenario. Actual fleet size is 35 trucks. At this fleet size, the
model predicts a base case throughput of 61 602 tonnes.
The model estimates the benefit of achieving a sustained
25 per cent reduction in payload variance at a throughput of
62 638 tonnes, a 1.6 per cent advantage. This is equivalent to
3.5 additional truck loads during the 12 hour shift.
If a sustained reduction of payload variance of 50 per
cent can be attained, the model predicts a throughput of FIG 7 -Throughput comparison in which overtaking is permitted.
P F KNIGHTS AND S PATON
However, the no-bunching scenario becomes unrealistic for CONCLUSIONS

large fleet sizes, since it requires unrealistically wide haul
For the case study involving a large open pit gold mine operating
roads.
a fleet of 35 Komatsu 930E trucks, the model predicts that the
DISCUSSION benefit achieved by a sustained reduction in payload variance
of 25 per cent is a throughput increase of 1.6 per cent, within a
Up to this point, this paper has focused on quantifying the possible error range of 0.6 to 3.2 per cent. This potential benefit
potential benefit of sustainably reducing truck payload
is worthy of further investigation by Continuous Improvement
variance. It is useful to briefly examine the mechanisms
teams at large open pit mines.
behind how a mine might achieve this.
Firstly, careful attention to drilling and blasting practices The model also indicates that, under the right circumstances,
are required in order to ensure optimal fragmentation provision of an overtaking lane can provide throughput
distribution. Alarge percentage of fines will tend to over-load benefits of up to ten per cent. The feasibility of providing
shovel dippers and consequently trucks. At the other end of unidirectional haulage circuits within large open pit mines
the scale, a large percentage of coarse particles will tend to should be examined.
under-load truck trays.
Attention should be paid to regularly changing-out ground REFERENCES
engaging tools (GET) on loaders, excavators and shovels.
The current practice of operating GET until worn-out is Ataeepour, M and Baafi, E, 1998. Application of ARENA simulation
questionable from the point of view of productivity and energy system to compare truck-shovel operation in dispatching and
consumption. Bucket fill factors are reduced by worn GET, non-dispatching modes, in Proceedings Mine Planning and
and specific dig energy increases, indicating that worn GET equipment Selection (ed: R Singhal), pp 475-480 (Balkema:
may be an additional cause of payload variance. Rotterdam).
Operator loading practices should be carefully scrutinised. Banks, J, Carson, J S and Ngo Sy, J, 2003. Getting Started with
There is some evidence that shovel operators adjust the GPSS/H, second edition (Wolverine Software Corp: Alexandria).
dipper fill factor on a final pass in order to maintain consistent Paton, S, 2009. Truck bunching due to load variance, undergraduate
payloads (Webb, 2008). Making truck payload information thesis (unpublished), The University of Queensland, Brisbane.
available to loading equipment operators on a pass-by-pass Runciman, N, Vagenas, N and Corkal,T, 1997. Simulation ofhaulage
basis should ensure more consistent truck payloads. This can truck loading techniques in an underground mine using witness,
be done by means of a radio link between the on-board truck Simulation, 68(5):291-299.
payload metre, or simply through a visible payload screen
mounted on the side of the truck. Sturgul, J R, 1999. Mine Design: Examples using Simulation, 367 p
(SME: Littleton).
Finally, every effort should be made to ensure correct dipper-
tray matching. In mines having highly variable geology, it Webb, B, 2008. Effects of bucket load distribution on performance,
may be advantageous to adjust dipper sizes and change-out undergraduate thesis (unpublished), The Universityof Queensland,
dippers in regions where bulk density variation is predictable. Brisbane.
Productivity Enhancementsto Increase
Ore Productionfrom 18 to 36 Mt at
Aitik
A Renstrom
ABSTRACT
Bolidens Aitik open pit mine located in northern Sweden, will expand yearly production from
18 to 36 Mt of copper ore during 2010. The expansion includes a new concentrator located
3 km off the pit, two new apron feeder crushers connected to the ore storages and concentrator
with conveyors at a total length of 7 km.The concentrator has two lines with autogenous grinding
mills and flotation each with a capacity of 18 Mt.
Consequently the mine operation has to be expanded. To achieve this, there was several parallel
processes ongoing; investments in new equipment, reorganisations, improvement projects in
maintenance and productivity. The Aitik 72 project was started to coordinate productivity and
mine planning activities. Initially the long-term strategic mine plan called for a total movement of
ore and waste to increase from 40 to 72 Mt but has today increased to 89 Mt.
Productivity projects have been focused on load and haul activities such as fuel economics
(Renstrom, 2007), effective utilisation of trucks partly by increasing manning levels, improved
dispatch and shovel utilisation focusing on optimal use of loader/truck fleets.
The eco-driving projects previously reported have reduced fuel consumption significantly. This
project is also a productivity project, as part of the fuel savings comes from minimising idle engine
hours due to queues, break-downs, operator shift changes, etc but also from under and overloading.
Effective use of truck fleets, ie transport of material, loading and dumping, thus excluding non-
productive engine hours has been an essential motivator when deciding the size and number of
new trucks required. This has resulted in a reduction of the investment with at least one but maybe
two trucks.
A third major project was to optimise shovels together with truck fleet by improving short-
term, up to two weeks, detailed planning of drilling and shovel activities. Daily scheduled breaks
for maintenance as well as planning of shovel utilisation, down to shift level are included in the
production plans.
KPIs and online production reports, based on real-time monitoring of drill rigs, shovels and
trucks, have been an essential tool used regularly to engage and motivate mine production personal
and information reporting to management.
Several projects have already reached their goals. The projects will continue throughout 2010 to
ensure that productivity targets are met.
INTRODUCTION
Boliden s Aitik open pit mine, located in northern Sweden, The pit expansion is managed by the mine organisation
will expand yearly production from 18 to 36 Mt of copper ore together with a productivity project Aitik 72. The Aitik 72
during 2010. The expansion includes new push-backs and project objectives were to coordinate productivity and mine
satellite pits, a new concentrator located 3 km off the pit, two planning projects starting in 2007, based on a long-term
new crushers connected to the ore storages and concentrator strategic mine plan with a total of ore and waste peaking at
with conveyors at a total length of 7 km. 72 Mt. This has now been increased to 89 Mt.
The expansion at Aitik results in a total ore reserve of The project portfolio includes projects aimed toward
600 Mt and a extended mine life from 2016 to 2029. truck fleet productivity; cost savings in combination with
The concentrator is built with two lines, each with autogenous environmental savings such as: ecodriving, enhancements in
primary mills powered by a 22.5 MW ring motor, flotation mine production planning and maintenance procedures and;
cells and dryers with a total capacity of 2 x 18 Mt/a. several other activities.
1. Project Manager, Boliden Mineral AB, MiningTechnology, Boliden 5-93661, Sweden. Email: arne.renstrom@boliden.com
SEVENTH LARGE OPEN PIT MINING CONFERENCE / PERTH, WA, 27 - 2 8 JULY 2010 115
A RENSTROM
ECO-DRIVING PROJECT Loss of truck hours are mainly caused by queues, short
breaks, operatorfshift changes and trucks off-route trucks on
The project has been ongoing since 2004 with the objective
their way to refuel, change operator, park or maintenance.
to optimise fuel consumption and has been very successful
(Renstrom, 2007). The KPI liter/ekv.kton.km, ie litre Queues and parking trucks is the responsibility of the
per transportation distance compensated for lift height is dispatcher that shall optimise number of trucks to each
established and continuously followed. Eco-driving is today shovel. Trucks not required for production at a specific shift
part of our operator training and influences other activities should be parked as idle truck hours are more expensive then
such as minimising unnecessary truck movements, efficient idle operator hours.
operator shift changes, optimised refuelling, payload, etc Operator related time losses are reduced by hot shift change
which all reduces fuel usage. As can be seen in Figure 1, the and four rotating operators on three trucks. Operator change
consumption levels vary and can be traced back to more or is made at dedicated locations in or close to the truck route.
less intensive management focus. An example is the decrease The four by three system is motivated by the fact that the idle
starting 2009 that is connected to a campaign to increase truck hours are more expensive then the fourth operator and
payload which results in increased ktonnes per cycle without also results in more tonnes per shift.
a corresponding increase in fuel consumption. Travel off-route, is mainly caused by refuelling or travel
Eco-driving is a KPI and a focus area that is dependent to workshop or parking areas. Travel to parking areas has
on operator behaviour, ie how they drive and manoeuvres been reduced by adding parking areas with heating and air
the trucks, how they utilise the trucks by correct payload, connections, in the pit area, close to crushers and waste ramp
minimising extra metres travelled, queuing behaviour, etc. connections, thus being close to ordinary haul routes. Travel
Eco-driving also depends on shift management, for an for refuelling is a recently started activity with the objective
example, the dispatch regime, number of trucks per shovel, to minimise fuelling occurrences, mainly with the help of
operator changes, etc which influence idle and queue time the dispatcher. Every shift will create a refuelling plan for
thus increasing fuel consumption to no use. trucks that require fuel, as opposed to earlier practices where
trucks where refueled every second shift. Travel to and from
TRUCK UTlLlSATlON workshops, for planned maintenance, arent and shouldnt be
At Aitik shovels are generally under-trucked mainly because reduced.
the long-term mining sequence requires mining in several The truck fleet at Aitik with seventeen Cat 793 and from 2011
parallel faces in order to maintain an even ore flow over time. additional nine Cat 795,is expected to manage transportation
This will from time to time set the trucks as the production work at an effective utilisation close to current levels of 77 per
bottleneck and therefore, an important resource to optimise. cent. Compared to earlier levels from 2007 of approx 75 per
cent the difference is approximately one less truck to buy. The
Utilisation might be calculated as engine hours by total
dips in effective utilisation during autumn 2008,are mainly
hours but this doesnt tell the full story. An idle truck engine explained by a period of non-overlapping shift changes and
is not productive, similarly a truck travelling on an off-cycle less operators per truck.
route. Efficient utilisation (Figure 2) is defined as:
Eff.utilisation = ((dumping + travel.empty + loading + travel. PRODUCTION MONITORING AND REPORTING

full + spot) times)/(total cycle time) Many productivity improvement activities are fine-tuning of
already fairly good procedures. Evaluation and management
or in other words, only value adding parts of truck cycle is of such activities requires detailed data from activities in the
efficiently used. mine, which must be collected continuously and stored with a
Truck utilisation is a complex task that requires active shift minimum of manual interaction. For several years Aitik have
management as well as an organisation that has sufficient collected and saved activity and quality data on drills, shovel
number of operators. and truck down to individual equipment activity levels.
FIG 1 - Monthly liters per ekv.kton.km and payload variations 2005 - 2009 (February).
2. Litre/ekv.kton.km =fuel consumption [liter]/(kton x ( horizontaldistance [km] + liftheigt [in kmJx 12.5))
PRODUCTIVITY ENHANCEMENTSTO INCREASE ORE PRODUCTION FROM 18TO 36 MTON AT AlTlK
I
FIG 2 -Monthly variations in truck efficient utilisation 2006 - 2010 (February). Eff.utilisation is the value added use of truck hours versus total truck hours.
Online data (Figure 3) is used for daily production reports, SHORT-TERM PRODUCTION PLANNING
key performance indicator (KPI) reports, updating mine
Expansion from approx 40 Mt/a to a maximum of 83 Mt/a
production schedules to current state and positions, and
while only increasing shovel capacity by 25 per cent and truck
finally stored for activity and productivity analysis. capacity by 50 per cent requires a more intensive planning
Easily accessible KPIs and production reports are essential of daily production and maintenance. To achieve this, the
for productivity projects, all KPIs discussed in this paper and mine organisation has been strengthened with one extra shift
many more are accessible online globally in the mine. foreman per shift, thus giving them more time to work with
Data collection for KPIs and productions are fairly easily planning and productivity. Scheduling of production was
justified but of course there is data collected that might never previously managed as a part of long-term planning with a
be used. The problem is to identify those although throughout resolution down to a weekly updated rolling five week plan.
time there has been several occasions where it has been With a larger production system and a more complex mining
possible to evaluate ideas on historical material instead of sequence there is a need to strengthen the planning capacity
field tests, or that older statistics has been used as a baseline with resources for weekly/monthly planning and daily/weekly
for new projects. planning.
FIG 3 -Online mine report on fuel (Lekv), efficient utilisation, payload and tonnes per shift.
A RENSTROM
The weekly/monthly time frame is managed by an Gannt-Scheme in combination with the CAES-office and real-
additional planning engineer thus enhancing the capacity of time drillmaps gives the planner a very good overview of the
detailed production sequencing for the coming weeks taking production status in the pit.
into consideration everything from drillabilty variations, late
updates of ore limits, shovel capacity and availability, delays The objective for daily/weekly planning is to have a daily
caused by weather, and so on. communication with dispatchers, mine foremen and mine
maintenance as well as with the monthly/weekly planner.
Ongoing is a project to give production planners,
maintenance planners, mine staff, foremen and dispatchers Their prioritised task is to follow production, optimise stops
a real-time overview of the short term mining schedule for maintenance and if necessary adjust the plan.
and status using a Gantt-Scheme (Figures 4 and 5) that is The new organisation includes maintenance planning
updated in real-time with drill holes and truckloads. The and technicians who works close to production planning.
FIG 4 -Snapshot of mine production plan for some days in May 2010 where upper parts shows blasts with blue color for loading and green for drilling. lower part
shows showel plan where blue is loading and orange is maintenance. Red line at left is timeline and shows current status for each blast, ie if coloured to the left of
timeline means delayed and on the right side ahead of schedule.
FIG 5 - Detailsfrom GanttSceduler above shoving that shovel 1142 is loading on blast SLl-150-1-2 and is 18 hours behind schedule (blue area before red line),
currently is estimated to require 5.2 trucks and is scheduledfor 7.00 pm on Tuesday.
PRODUCTIVITY ENHANCEMENTSTO INCREASE ORE PRODUCTION FROM 18TO 36 MTON AT AlTlK
TABLE 1
Estimated cost savings for sample productivity improvements activities in Aitik 2005 to 2009.
Activity
Ecodrivinq (Liter/ekv.kton.km) I 2005 I

Average first year
72 Lekv
1I Resultsachieved
March2010
62 Lekv I
Savings Euro/year
1 600000
Effectivetruck utilisation I 2005 I I I
2007 I (Estimated6896) 1 I I
72%
77%
Payload I 2009 I 201 kton" I 215 kton I 10100000

*Payload restricted due to rebuild program.
A combination of one hour daily shovel stops with service A conclusion made is that to improve productivity in high
and inspections will improve maintenance planning thus performance large open pits, there are no simple solutions.We
minimising unplanned maintenance stops, especially are discussing improvements measured as singe percentages
unscheduled stops in production. over time which can't be verified without extensive real time
monitoring and data management. However as shown is
CONCLUSIONS this paper it is possible to implement changes , keep them
Aitik have many years of experience with productivity work over time and justify the efforts made with the help of those
using activity data to continuouslyfollow improvements over systems.
long periods and use those as motivation for the organisation
to work with and to keep improvements. In the Aitik 72 project
the available information have been used to identify potential REFERENCES
productivity improvements and as a base for investments in Renstrom, A, 2007. Truck fleet utilisation and fuel saving in Aitik,
drills, shovels and trucks. in Proceedings Sixth Large Open Pit Mining Conference 2007,
Some results are indicated in Table 1 but it should also be pp 107-112 (The Australasian Institute of Mining and Metallurgy:
added the value of one-two truck investments never made due Melbourne).
to productivity improvements.
Mine Control -Why do we do it, and
how do we Achieve Sustainable
Success?
R Ballantyne
ABSTRACT
Technology is playing an increasingly large role in the operations of our mines. There is a
temptation to regard the implementation of technological systems as the objective itself when, in
reality, systems are used to support a business process. Only when the process succeeds will the
systems be seen to succeed. Understanding the distinction between system and process is critical.
This understanding, together with the resultant holistic approach, is considered to be critical to
successful implementation and operation of Mine Control process.
This paper examinesthe critical factors essential to success in establishing and sustaining a Mine
Control process that maximises benefits to an operation. These factors have been developed during
the authors experienceof nearly 30 years in Mine Control projects across over 50 sites. Each factor
is discussed in relation to real world examples at well-known mines.
INTRODUCTION Terminology
This paper collates knowledge accumulated over nearly 30 The term Mine Control is used rather than dispatch or
years of experience in Mine Control implementations and j7eet management system (FMS) to emphasise that we are
ongoing process operations at over 50 sites in Australasia, talking about the whole process. Using the latter terminology
Africa and North America. During this time, wide-ranging essentially narrows the scope of the discussion to a system
expectations, levels of understanding and management or type of system which might form a part of a Mine Control
philosophies were encountered. process.
These principle variables can often effectively determine Throughout this paper personnel operating a Mine Control
the probability of success or failure of a Mine Control process are referred to as controllers.
implementation, or indeed its sustainability going forward.
Essentially, the paper attempts to bring clarity to what Mine
Controlis, and highlights some dos and dontsfor those wishing MINE CONTROL - A PHILOSOPHICAL
to either implement or improve a Mine Control process. DISCUSSION
In many cases, a Mine Control project is treated as a system in
The difference between a process and a system its own right that can be launched, forgotten, but still expected
The basic thesis of this paper is that Mine Control is a process, to function effectively in a constantly changing environment
a business process. It comprisesactivities or tasks that produce and with variable resources available.
a specific service and/or product for customers involved in the With this comes the risk of a mind-set such as if1 implement
management and supervision of the industrial process we call this system, how many people can I cut from the workforce
mining. to pay for it?
A system can be considered as an organised set of connected Amore appropriate perspectivemight be a more holistic one,
things or concepts which may be material or immaterial, for
such as what are my goals and if1 implement this process,
instance, a railway system or a political or social system. In
the world of information and computer science, a system what calibre of people will I need to operate, maintain and
could also be a method or an algorithm. Examples include thereby constantly reoptimise the production process?
systems of counting, as with Roman numerals or the decimal The latter philosophical approach does not see the system
system, and various systems for filing papers and various as a silver bullet, but as one of the tools that support the
library systems. business process - namely, improving mining efficiency in
Systems are used to facilitate and support the consistent the constantly changing and imperfect environment that we
execution of a business process. call a mine.
1. PrincipalConsultant, RAK Developments, 25OTomlinson Road, Hovea WA 6071. Email: ray@rakdev.com
R BALLANTYNE
Why dont we always succeed with Mine When doing an audit of a dispatch installation about two
Control? years ago, everybody from the mining manger to the shift
supervisor was questioned about the objectives of Mine
You have probably all heard these and other similar comments Control. When asked our question, 50 per cent responded
many times: with some reference to productivity that was consistent with
those guys in the hut dont know what theyre doing; running the shift better, however, the remaining 50 per cent
I cant believe anything that comes out of that system; felt the system was there to gather data for reporting.
that guy on light duties, he can run Mine Control; In the authors experience, the greatest benefits come from
Mine Control can do that ... theyve got time; improving decision making that occurs in real time. This
the systems useless; becomes the driving objective, the control room collects data
Mine Control dont know what the plan is; to support decision-makingand assist in running the shift. If
the blocks havent been updated for weeks; and we get this right, then a byproduct is that very high quality
I dont know where Mine Control is located. data is available for downstream reporting, analysis and
focused business improvement projects.
Comments of this type are symptomatic of a lack of
alignment or understanding of the role Mine Control plays So, returning to that very basic question: Why do we have
and its importance in a mining operation. Let us ask a very Mine Control?
basic question: Why do we have Mine Control? For this author the answer is simple, it is: To help run the
We are certainly not the first to ask questions of this nature. shift better!
The text below is illustrative of the type of slide sometimes
seen addressing this very question: KEY FACTORS FOR SUSTAINABLE SUCCESS IN
MINE CONTROL
Why have aJleet management system (FMS)?
What do we need to have in place for successful Mine Control
If you cannot answer this, STOP DO NOT PROCEED with that contributes meaningfullyto running the mining operation
the installation. more efficiently?
Fleet management systems are not a silver bullet to fix A hierarchy of key success factors has been developed.
production problems. These follow and are listed below in order of importance:
They can provide invaluable information and control 1. high calibre controllers,
to assist in improving the process, but they will not fix 2. controllersrelationship with field supervision,
production shortfalls.This must be understood before each 3. diligent systems administration,
install and must be accompanied by a change management
4. respected ownerof the process,
plan.
5. committed management support, and
A significant effort is required on behalf of each mine that 6. quality infrastructure - technology must work.
installs FMS. Extra resources are required to maintain and
run the system once implemented, too often we try to let the High calibre controllers
system run itself. This does not work. A senior manager with an open pit mining contractor was
Too often we see a site invest a large amount of finance responsible for Mine Control/dispatch systems from three
to the FMS, only to cut costs by not appointing adequate different suppliers at respective operations. He was asked
resource numbers to manage it on site. which gave the biggest benefits:
One of the major factors influencing poor performance ...it doesnt matter which system we use, the thing that
of a fleet management system is inadequate resources to gives us benefits is the calibre of the person w e put in
implement, run and maintain the system. the control room.
Once the need for high calibre personnel is accepted, then
This document was written from a perspective of what you
the next question is the philosophy regarding organisational
had to do to have a successfulimplementation.
structure. Various personnel manning models have been
The author of this paper takes slight issue with the title and encountered over time, and are listed below where the
thinks it should ask why have Mine Control? However, the controllers are regarded as either:
slide contains a lot of wisdom, undoubtedly learnt the hard
the most senior supervisor on crew,
way.
of equal status to field supervisors,
The message is however very closely aligned with the intent subordinate to a more senior supervisor in the field,
of this paper, albeit from a slightly different perspective. a mandatory step in career path toward a supervisor
The intent is the same - do not throw away good money on position,
a systemto fix problems if you are not going to commit to equivalent status to an equipment operator with no formal
building the required process. Systems are not silver bullets; authority, and
systems will not fix problems and shortfalls. Systems can fulfilling a clerical position to capture data.
however make a very significant contribution to improving A key question is therefore: Where do our controllers fit into
the process. the overall structure of shift management hierarchy?
Our industry invests large sums on infrastructure such
as control rooms, communications equipment and other The most senior supervisor on crew
technology. Some sites erect dedicated buildings to house A client experimented with rotating supervisors through Mine
Mine Control and integrate all the supporting and linked Control for repeated short duty periods. They discovered that
technology such as mine video, dewatering and geotechnical some suited the role and others didnt quite take to it. That
(slope stability) warning outputs. learning may reflect commitment issues or aptitude or indeed
MINE CONTROL - WHY DO WE DO IT, AND HOW DO WE ACHIEVE SUSTAINABLE SUCCESS?
both. They eventually arrived at a situation where the best equipment operator is seconded into the field supervisor
supervisors became permanent in Mine Control and were position.
made the most senior supervisor on the crew. We understand
that a coal mine in the Western Australian southwest also Rotation
used a similar model when they first installed dispatch in the Another approach to the issue of Mine Control being a career
mid 1990s. dead-end is to rotate people through the role. The author
contends that rotation should only be applied when the
Of equal status to field supervisors controllers are full supervisors.
Our second model is where the controllers are supervisors of
Australian open pit Mine Control installations tend to use
equal standing with supervisors in the field. This approach
dedicated controllers but this does not always have to be the
generally doesnt happen in situations where the controller
case. The copper mines of Arizona, New Mexico and Utah and
position is permanent. It does however become a powerful
the gold operations in Nevada tend to have Mine Control (or
option when the concept of rotating controllers is introduced.
dispatch) as part of the supervisorsrole with various forms of
Rotation is discussed later.
rotation used within the supervisors on a crew.
Subordinate to a more senior supervisor in the field There have been some extremely successful cases where
The idea of having controllers with some authority but forms of rotation have been used in Australia. A client
subordinate to field supervisors is a reasonably common underground operation has a regime where there are three
model in Australian open pits. Typically people come to the shift bosses on a crew. They rotate every three days through
role from being equipment operators or, in some cases, a three positions; development, production and Mine Control.
clerical administrative background and largely due to their Being an underground mine this means that the shift boss in
personality characteristics become very good in the role. Mine Control has recent knowledge of the mine. During the
mid 1990s the other of our WA coal operations used statutory
A mandatory step in career path toward a deputies in their site hut. In the pit there were two deputies on
supervisor position each crew. They would rotate between the control room and
A model that worked extremely well, but the author has only the field informally during the shift with at least one manning
seen formally applied at one site, is to make a year in Mine the control room at all times.
Control a mandatory step before becoming a supervisor. On While there are advantages in rotation it also increases the
the other hand there are many instances where the model is difficulty of maintaining consistent practices. This increases
being applied informally and it is common for a controller go both the need for training and the requirement for separate
on to become a supervisor. system administration and supervisory oversight.
Career dead end What not to do

We wish to attract high calibre people into the controller Some examples of what not to do are pertinent at this point
position. They require the same or similar skills, knowledge in the discussion. As you have no doubt gathered, the calibre
and characteristicsthat are required of a supervisor. However of person in Mine Control is extremely important, so to have
we can run into a problem. A permanent position as a no selection criteria is a recipe for failure. Being on light
controller can be regarded as a career dead end. duties does not qualify someone to become a controller. It
may be valuable to have light duties personnel observe in
This was highlighted recently when one of the larger WA
Mine Control to develop their knowledge, but it is not a valid
iron ore mines advertised for a senior controller to supervise
criterion for selection.
the controllers. In a market where skills shortages are
regularly being highlighted by the media there were over 100 For reasons that are not understood clearly, graduate
very good applicants. All were very experienced controllers engineersjust dontseem to work in Mine Control. It is thought
but effectivelyhad no career path out of the role. this may reflect upon a lack of operational credibility and
therefore confidenceor it may be that there is no interpersonal
How then, can this issue be addressed?
relationship with respective crews and supervisorycolleagues.
The fourth of the manning models discussed above, where a
year in Mine Control is a mandatory prerequisite to becoming Possibly the worst example ever encountered involved a
a supervisor, has the solution built in. In this case the period mine owner who had purchased a fleet management system.
in Mine Control is not a career dead-end. The mining contractor was asked to resource a control room
and run the system. The mine was residential close to a town,
An interesting situation came up during a recent assignment and so the contractor contacted Centrelink and asked for the
where the status of the controllers role was under review. three cheapest people they could get.
The site had dedicated controllers that had been drawn
The author is not going to recommend any particular
from the workforce. All incumbent controllers were of a
personnel strategy for Mine Control except to stress the need
high standard however problems occurred whenever one for clear selection criteria, high calibre personnel and the need
of the dedicated controllers was unavailable and a backup to consider organisation structure, authority levels and career
was required. Standard practice had been to second an paths very carefully. Each strategy has its own strengths and
equipment operator to fill in as controller, this was resulting weaknesses and time and effort must be spent to ameliorate
in a serious downgrading of the performance of Mine Control. the effect of those weaknesses, therefore important to select
Further investigation highlighted that a large portion of the the strategy that suits your environment. Once set in motion
current supervisors had done their time in Mine Control this discussion can be guaranteed to generate heated debate
prior to becoming supervisors. A new practice has now within both management and the workforce concerning the
been established whereby, in the absence of a controller, a favoured model. The possibility that defensive behaviours
field supervisor moves in to the controller role and a senior may inhibit honest and open debate should not be discounted.
R BALLANTYNE
Controllers relationship with field supervision Respected owner of the process

Relationships with supervisory colleagues, superintendents The author quite deliberately differentiates between the
and operators appear to be an important factor for sustainable process owner and system administrator roles. Unfortunately
success. The most important of these relationships is between our industry often expects both roles to be fulfilled by the
the controller and shift supervisors. same person.
How much decision making authority can the supervisor In reality, the two roles could not be more different. A
hand over to the controller? Does the supervisor have systems administrator requires a pedantic attention to detail,
confidence in the controllers ability to make decisions? whilst the process owner needs to be able to take a global view
Some time back the author was involved in advising a and understand the ramifications of possible changes.
contractoron settingupa Mine Controlfunction.In preliminary The process owner needs to be someone that can influence
workshops the personality characteristics required in good senior management and at the same time have the authority
Mine Controllers were discussed and emphasis placed upon and skills to achieve behavioural change in the workforce.
the importance of the working relationship between the There have been a very small number of successful cases
controller and the shift supervisor. On the following site visit, whereby the roles of administration and owner are carried by
the controllers had been appointed and commenced work. the same person. In the two cases known to the author, both
Further investigation revealed that each of the appointed involved senior engineers who were absolutely passionate
controllers, who genuinely had been appointed on merit, about Mine Control.
happened to be the partner of the respective shift supervisor.
The controller to supervisor relationships on this site were Sometimes the ownership role can be shared; some
working extremely well and the controllers were very proud sites appoint a senior Mine Control supervisor or even a
of how much their role was enabling the supervisor to run the Superintendent to oversee internal standards, ongoing
shift better. Obviously it wasnt a sustainable situation and training, administration and technical/maintenance issues.
was subsequentlyaddressed. Such a person is usually experienced in Mine Control, widely
respected and passionate about Mine Control. Such a role
This experience highlights the fact that effective also provides representation of the Mine Control workgroup
communication and team alignment is an important but but still requires the involvement of a more senior operations
intangible factor for success. manager in the ownership role.
In that regard, it will be interesting to hear about the current
experiment with remote operations control being operated in Management support
the city of Perth, some two hours flight time away from the In turn, the process owner needs the support of management.
physical mine operations. This can be as simple as recognising situations where other
activities (and particularly changes to the mining process) may
Diligent systems administration impact the integrity and sustainability of the Mine Control
This is critical. When it comes to selection of personnel for subprocess. It might be as simple as facilitating the need for
administrative duties in Mine Control, words like pedantic, consultation and liaison between different workgroups.
rigorous, and consistent will tend to be at the top of the One of the greatest risks to sustaining high levels of success
prerequisites list. in Mine Control is a change in management. A new manager
The systems used in Mine Control require daily who has not seen or been a part of the considerable effort to
administration which obviously raises issues with manning, make Mine Control a success story may not realise how much
particularly in a fly in-fly out environment. hard work and support were needed to get there.
The most critical aspect of administration is managing the The incoming manager will consequently underestimate the
various forms of location that are to be named. At the risk of need for that effort and support to be maintained in perpetuity.
sounding like a real estate agent it is indeed: location location A well documented Mine Control vision statement can be a
location. There are many forms of location information, such valuable tool in managing this situation.
as blasts, blocks, GPS coordinate, etc and if they are not valid
and used with integritythen Mine Control will degrade rapidly Quality infrastructure and technology
with serious consequences for grade control in particular. A fundamental theme of this paper is that we want to attract
Unfortunately in many cases the administration is left to high calibre people to work in the Mine Controller role. To
individual controllers.This often results in multiple names for then expect high calibre people to work in an inadequate
locations, people and equipment as each roster crew recreate environment is a recipe for failure. The Mine Control facility
the same entities in the system, albeit with slightly different should be appropriatelylocated,well equipped, well appointed
nomenclature. and well maintained.
Given that systems administration is essential and that it
should be carried out largely independently of the controllers The control room
who should fill the role? Too often, even when the need for the By far the most important aspect is to have Mine Control
role has been recognised, it is given to the graduate engineer, located where key stakeholders (particularly field supervisors)
to do in their spare time. Unfortunately this practice of can easily drop in to discuss immediate issues with the
treating Mine Control systems administration as a junior role
controller.
has contributed to the perception that, like the controller role,
it can be a career dead end. Entry should be facing the controller; it is extremely
There are various models including a shared role between an disconcerting when working in Mine Control to have visitors
experienced engineer and a senior Mine Control supervisor. entering the room from behind.
The whole situation becomes more complex when the role The table or console where controllers work should not have
of systems administration becomes confused with ownership table legs that intrude when the controller moves position.
of the process. Some additional criteria are listed below for reference:
MINE CONTROL - WHY DO WE DO IT, AND HOW DO WE ACHIEVE SUSTAINABLE SUCCESS?
minimum dimensions 6 m x 5 m; and management support. Regretfully, many sites focus

split system air conditioning with backup power excessively on the sixth factor, infrastructure, and make
(generator); their project one of technology implementation whereas
sustainable success is people driven.
0 kitchenette;
There has been much use of the term sustainable success
airlock door system;
and an obvious question would be how do we determine if
ease of cleaning, cleaning equipment;
Mine Control is really a value add proposition? The author
close proximity to toilet facilities; believes that success is being achieved when there is mutual
able to comfortably accommodateminimum of two people respect demonstrated between Mine Control personnel and
at workstations; their colleagues in management, supervisory staff, operators
comfortable chairs; and technical groups such as geology and grade control.
floor protectors; Outcomes of success include:
multi-channel radio console;
improved safety and incident response;
windows should face south to avoid looking into sun, this
may partially dictate location; greater mining efficiency, although this is often hard to
prove in a constantly changing environment;
computer for access to other systems, eg email, supply, etc;
better grade control outcomes (fewer misallocated loads);
cordless phone; and
and
handheld radio.
high quality data for analysis and business improvement,
With regard to systems and the supporting technology which in turn can lead to further improvements in
infrastructure there is only one criterion, they must work! operating efficiency.
The fleet management system (FMS) The mutual respect derived from success further strengthens
the whole mining process. Key stakeholders take the trouble
To digress slightly from the main theme of this paper to communicate as team players should. This leads to fewer
and briefly discuss the functionality required of a fleet
surprisesfor management and a harmonious workplace. The
management system used in Mine Control. The mandatory latter harmony encourages skills and knowledge retention in
basic functionalitycan be described as:
the workforce, again further aiding sustainable performance
Recording, against vehicle and operator, of time stamped gains.
changes to: location, critical status and measures of work
Whilst we have discussed the human resources considered
done, logged as it happens feeding a real-time decision
necessary for success quite extensively, one could still be
support capability. forgiven for asking the question: What does this Mine Control
Other functionality that can then be considered depending workgroup look like? Benchmarking around the world tends to
upon the requirements of the operation including:
indicate that an adequately resourced workgroup comprises:
optimised dispatch algorithm, A unifymg leader whose roster overlaps those of all others
high precision vehicle guidance, and in the workgroup. This person facilitatestraining, oversees
interface to onboard condition monitoring. internal standards, and liaises on technical/operational
Computer infrastructure issues with key stakeholders. Lastly, this person covers
sickness and annual leave.
To support the functional components of successful Mine
On larger sites (say >30 - 40 trucks), it is relevant
Control provided by technology some considerations for
to consider two Mine Controllers on all shifts. This
control room design include:
allows one to look after the basics, whilst the other does
console should provide easy access to computers and administrative tasks, resolves problem and helps facilitate
cabling without disrupting dispatcher, efficient shift changes. One person simply cannot cope and
backup power supply for communications and computer standards will start to slip.
equipment, On larger sites, it is not uncommon to have two or
computers located in adjoining room or enclosed console more technical support personnel. These personnel
to minimise ambient fan noise, are commonly rostered on day shift only. Typical skill-
account for heat load from computer and communications sets for such people include very high level computer
equipment, and literacy, and good working knowledge of communications
0 backup computer for Mine Control system. infrastructure. A deep and fundamental understanding
In summarythe Mine Control facilityshould be appropriately of the system and its logic algorithms and variable
located, well equipped, well appointed, maintained and the parameters constitutes the cherry on the cake. When
supporting technology infrastructure must work. not dealing with the above, these technical personnel
undertake analysis and are actively involved in business
improvement projects.
CONCLUSIONS
Do not be surprised to see five Mine Control personnel at
It is concluded that: a benchmark mine site.
Dispatch/fleet management systems are not a silver bullet
for production problems. They should be considered as HOW DO YOU RATE?
part of the tool set that supports the business process of Revisit our success factors. Score your site against each of
Mine Control. these.
There are six key factors for sustainable success in Mine Do you have high calibre controllers? Could you take your
Control. The five most important are the human factors, controllers and put them out in the pit as supervisors?
namely: high calibre controllers, relationship with Are your supervisors prepared to hand decision making to
supervisors, systems administration, process ownership the controllers?
R BALLANTYNE
Are your systems diligently administered? Every day? Are ACKNOWLEDGEMENTS

blocks entered, road networks up to date, new starters
Tim Riley, formerlySuperintendent ContinuousImprovement,
created, etc?
Mt Keith Operations who provided tremendous help in
Do you have someone who is the visible owner of the
converting a workshop presentation into a conference paper.
process, who can ensure that standards are maintained
and changes implemented? All the many clients and hosts at sites visited over nearly 30
Do you have management support or are they always years, helped formulate the ideas expressed in this paper. In
criticising the system and looking for ways to eliminate particular, those who were participants in the many vigorous
the cost of manning and supporting Mine Control? discussions as to the best way to go about Mine Control.
Do you have quality infrastructure, is the control room
well set up, is it in the right location, do your systems work
(all the time) and is the technology maintained?
Long-Term Mine Planning at the
Largest Phosphate Mine in Brazil
R de Lemos Peroni, L N Capponi2,J F Costa3, E Gontijo4and
D C Veiga5
ABSTRACT
Tapira Mine (Tapira, MG, Brazil) is the largest phosphate ore and concentrate producer in Brazil.
The mine is an open pit, which until recently operated with a large equipment fleet and now uses
a smaller fleet operated by contractors. The annual production reaches 17 Mt of ore plus 14 Mt of
waste at a stripping ratio of 0.82.
Defining the geological boundaries and forecasting the annual production can be a difficult task
if not performed on a reliable model. The geological model has been totally reviewed as well as the
estimation process. Ordinary Kriging has been used to build the new block model with P205and
contaminant grades. After estimation, block model validation and classification criteria definition
to assess resources and reserves, the result is given by the life-of-mine plan is defined after pit
optimisation and design.
The Lerchs-Grossmann algorithm implemented in the NPV Scheduler was used to define the
long-term production targets based on the main parameters Pz05and the contaminants that are
just important to the beneficiation process.
INTRODUCTION
The TAPIRA complex has been known and characterised 50 x 50 m) the current pit and approximately 200 x 200 m
for more than 50 years, particularly through exploration at the external area. The pink boundary line represents the
carried out between 1953 and 1954 by the Brazilian Agency limit of the alkaline dome which delimits the mineralised area
of Mining Production (DNPM). Currently the mine produces comprising about 7 km north-south by 5 km east-west and the
17 million Mt/a of ROM and moves about 14 Mt/a of waste. It orange boundary represents the current pit limit representing
is important for the mine to keep the stripping ratio constant approximately4.3 km.
and to feed the process plant with an average grade of The samples were composited in 5 m intervals to avoid
6.0 per cent P205,so those are the main production goals to support problems after linear combination between samples
keep stationary. Stripping ratio is contracted and must not be by the estimation technique, and to match the operational
so variable to not impact on budget, as head grade must be bench height. This procedure generated the final data set that
kept constant due to the concentration grade after processing. has been used to re-estimate the resources of the deposit.
The quality of the ore associated with the world demand The samples were collected over many years and by different
for fertilisers has motivated the company to proceed with a drilling campaigns. Hence, some elements or minerals
wide review of its grade model. Based on this review, a work were not always analysed, so the final database sometimes
program was carried out to update the resources/reserves has missing samples for some of the oxides. The two main
classification plus the life-of-mine plan (LoMP) and designed oxides are the total Pz05and the CaO, which is an important
the intermediate pits for medium term strategic mine mineral to determine the process recovery for each ore type
planning. Figure 1 shows an aerial view of the current pit at and to obtain the amount of apatitic P205 from the total
Tapiras mine. P205content. Thus, the final database was then made up of
18 239 P205tsamples and 16 813 analyses of CaO. Note that
SAMPLING AND DATA SET there is a difference between the P205t(total phosphate) and
The deposit has been sampled since the 1960s and after a P,05ap (apatitic phosphate). To account for samples that do
thorough detailed review of the data set, in terms of quality and not have Pz05apanalysis the company uses a relationship that
confidenceof the samples, the final database is now composed allows one to transform P20$ in Pz05ap,this relationship is
of 1084 drill holes spread across the deposit (Figure 2). called the calcium/phosphate ratio for the deposit and follows
The samples are distributed in a dense grid (approximately the equation below:
1. Adjunct Professor,Federal university of Rio Grande do Sul, Av Bento Goncalves, 9500 Bloc0 IV, Predio 75, Sala 116 CEP 91509-900, Brazil. Email: peroni@ufrgs.br
2. Senior Mining Engineer, Fosfertil S/A, Rodovia MG 341, km 25, Fazenda Boa Vista, Tapira-MG CEP 38185-000, Brazil. Email: lucianocapponi@fosfertil.com.br
3. MAuslMM, Associated Professor, Federal Universityof Rio Grande do Sul, Av Bento Gonplves, 9500 Bloco IV, Predio 75, Sala 104 CEP91509-900, Brazil. Email: jfelipe@ufrgs.br
4. MiningTechnician, FosfertilS/A, Rodovia MG 341, km 25, Fazenda Boa Vista, Tapira-MG CEP 38185-000, Brazil. Email: edilsongontijo@fosfertil.com.br
5. Geologist FosfertilS/A, Rodovia MG 341, km 25, Fazenda Boa Vista, Tapira-MG CEP 38185-000, Brazil. Email: diegoveiga@fosfertil.com.br
R DE LEMOS PERONI eta/
if -
RCP 2 1.35 E 0 , a p = E0,t GEOLOGICAL MODEL AND GRADE
where:
RCP =
E0,t
~
1
i f R C P < i . 3 5 - E O a --CaO
- 1.35
EST1MATI0N
The current geological model is composed of a mixture of
geological domains combined with weathering zones. The
geological interpretation is comprised of nine different
RCP is calcium/phosphate ratio lithologies as per the descriptions in Table 1. The two main
CAO is calcium oxide percent domains are the FCR and BEB, which represent 90 per cent of
the mineralised samples.
PP5f is total phosphate
The approach given in this work has grouped the geological
PzO5aP is apatitic phosphate domains into geostatistical domains - (GDs) based on the
FIG 1 -Aerial view of the current pit.
FIG 2 -Geological limit and the exploration drill holes available at the deposit.
LONG-TERM MINE PLANNING ATTHE LARGEST PHOSPHATE MINE IN BRAZIL
TABLE 1
Main rock types found in theTapira Mine.
Geological description
FCR I Phoscorite- Group of rocks consistinq of modal variations in apatite, maqnetite and maqnesian silicates (olivine and/or phloqopite).
SLX I Silexite - Rocks composed essentially of cryptocrystallinequartz, with high hardness.

CAP I Overburden- Includessoil, clay and laterite, not mineralizedin phosphate or titanium.
ENC I Country Rocks - Includes quartzites, Dhvllitesand schists of Canastra Grow. This litholoqvwas disreqardedfor the separation of GDs.
found in David (1977), Annels (1991), Sinclair and Blackwell
(2002),JORC (i999), CIM (2000) among others. At Tapira,
a geometrical constraint was used to classify resources in
MEASURED/INDICATED and INFERRED that took into
account both variables (P205t and CaO) measuring the
distance of the samples from the block. If one considers that
the main parameter of interest to determine the resources
P,O, CaO classification was obtained from the P205t and the PzO5ap
GD1 FCR CBN was derived from the RCP, it was decided to categorise the
resources considering the two parameters together. A field
I GD2 I DNT,BEB,eFEN I FCR, BEB, e FEN I was created to identify block distance considering the P20$
I GD3 I CBN,SIEeSU( I DNT, SIE, CAP e SH I samples according to the following rule:
The grade estimate was carried out using ordinary kriging

(Matheron, 1963) which was performed in each GDfor each
oxide, which means that, for the example shown here, six
estimates were generated and the individual estimates were IOO< = DIST P205<200 - CAT P205= 2
combined within a single block model using percent block
model of each domain to superimpose the different block 200< = DIST P205<400 - CAT P205= 3
models estimated using samples from each domain. The
basic statistics for the variable P20$ for each GD is shown
in a sequence of histograms presented in Figure 4. The same DIST P205>= 400 - CAT P205= 4
results were obtained for the CaO variable and the histograms
and the statistical summary are presented in Figure 5. And a field to identify block distances from CaO samples
that follows the same idea. The logic follows below:
RESOURCE C LA5 5 IFICAT10N CRITERIA
The block model estimation was performed using the two o< = DISTCaO <loo - CATCaO = 1
main parameters (P20$ and CaO), but resources are defined
based upon the apatitic phosphate (P205ap)content. Detailed
discussion about resources reserves classification can be loo< = DISTCaO < 2 0 0 - CATCaO = 2
LIhology Analysis W05T Lithology Analysis CaO
0.00 t
0.W
Maan 17.1
Mean [XI
FIG 3 -Average grade x standard deviation of variables P,O,t and CaO to GDs individualisation.
I""
FIG 4 - Histogramsfor each GD - P,O,t.

200< = DISTCaO <400 - CATCaO = 3 And finally the block is going to be classified as INFERRED
resource (INF) if at least one of the fields considered has the
DISTCaO> = 400 - CATCaO = 4 value 3, according equation below:
l1or
CATP20,= 1and CATCaO = 3
Having the two fields, the classification was performed
according to the syntax below, considering the outcome of each
comparison. The block is going to be classified as MEASURED CLASS1 = 3 if CATR 0, = 3 and CATCaO = 3
resource (MEAS), if and only if, the block is considered
MEASURED by both fields, ie, the number 1 is attributed to K T P O , = 2 and CATCaO = 2
both distance fields, according the rule shown above.
CLASS1 = 1if {CATP205= 1 and CATCao = 1

I"'
CATP20,= 3 and CATCaO = 1
where:
The block is going to be classified as INDICATED resource
(IND) if at least one of the fields considered has the value two, CLASS1 = 1means MEASURED resources (MEAS)
according equation below: CLASS1 = 2 means INDICATED resources (IND)
CLASS1 = 3 means INFERRED resources (INF)
CATP,O, = 1and CATCaO = 2
CLASS1 = 4 means NOT CLASSIFIED resources
or
CATP,O, = 2 and CATCaO = 1 (NOT CLASS)
or After the classification,the gradeltonnage curve was built for
CATP,O, = 2 and CATCaO = 2 the deposit (Figure 6). The solid lines represent the tonnage
(el
FIG 5 - Histograms for each GD - CaO.
for each resources category and the dashed line represents the the GREEN line represents the area where remobilisation
average grade, which can be read on the right hand vertical of infrastructure will allow for the mining in the next
axis. The amount of ore of the deposit is very large, but it is 20 years; and
directly linked with the cut-off grade (COG) applied, and it the BLUE boundary is the final pit limit determined by the
can be seen that above five per cent P20$ the curve is very conjunction between the extent of the alkaline dome and
steep and consequently very sensitive to changes. This means the waste dumps that are already built at the northern end
that small variations in grade have a big impact on resource of the pit.
quantities. The incremental pits and the ultimate pit were generated
using the NPV Scheduler (Earthworks Corporation pty Ltd,
LONG-TERM MINE PLANNING 2009) and using the optimisation algorithm implemented in
The strategic mine plan was developed considering the software (Underwood and Tolwinski, 1998). Pit design
some parameters that limited the areas available for the was performed using DATAMINE design tools. Considering
the physical constraints above, the optimisation was carried
development of mining as per the list below, referred in
out in three stages until the exhaustion of resources. The input
Figure 7. The orange line represents the current pit outline to of each optimisation is the output of the previous stage.
give an idea of mine advance:
In Figure 8, the final pit is shown after smoothing by defining
the RED boundary represents the limit of the existing and locating access roads and accounting for pit geometry, and
environmental permits and proximity to the current geotechnical design criteria (it was assumed a global angle to
process plant limit; the whole deposit of 29O).
FIG 6 -Grade tonnage curve for the deposit.
FIG 7 - Pit limitsfor each stage of mine planning advance.

The definition of ore reserve will be based on the combination RESULTS
of P205COG plus the limits of RCP (0.9 5 RCP s 3). Only the After the three runs of the NPV Scheduler and some
blocks classified as MEASURED and INDICATED (M and I) adjustments and analysis necessary in any long-term mine
that felt within the ultimate pit will be considered as reserves planning exercise, the graph presented in Figure 10, was
PROVEN and PROBABLE (P and P) respectively (JORC, produced. It was obtained with the standardisation of the
1999;CIM, 2000). The graph presented in Figure 9 shows the head grade over the life-of-mine. The dashed blue line shows
tonnage of the deposit for this ultimate pit shell. Considering that the head grade fluctuates around the goal of six per cent
a production of 17 Mt/a, the expected mine life for the deposit and in average for the ultimate pit it stays around seven per
is 61 years. cent. It was possible to smooth out the stripping ratio (dashed
132 SEVENTH LARGE OPEN PIT MINING CONFERENCE I PERTH, WA, 27 - 28 JULY 2010
FIG 8 - Ultimate pit configuration after optimisation and pit design.
FIG 9 - Parameterisationcurve for the ultimate pit shell.
FIG 10 - Results from the long-term mine planninggoals.
R DE LEMOS PERONI e t d
TABLE 3
Resourcesand reserves board forTapira's deposit at a five per cent cut-off grade.
Measured 1 985523121 I MIand16.49% I Proven / 514649533 1 PandP7.00 1

Indicated I 646079596 I 1 Probable I 364723845 1 I
Inferred I 260533463 I I Infresources I 145530778 I i
Not classified 1 10661 731 I I I i
Total 1 1902797913 1 879 373 378 I
purple line). The huge peak seen in 19 th year is attributed REFERENCES

to the proximity of the boundary constraint imposed by mine Annels, A E, 1991.Mineral Deposit Evaluation: A Practical
planning. This will be addressed in a detailed review and a Approach, 436 p (Chapman and Hall: London).
relaxation of the constraints when the operation gets closer to
David, M, 1977.Geostatistical ore reserve estimation, Developments
this point in time. Hence, Figure 10 shows the result in terms
in Geomathematics 2 , 364 p (Elsevier Scientific Publishing
of life-of-mine planning to the exhaustion. Company: Amsterdam).
In Table 3 the final reserves statement are presented for Earthworks Corporation Pty Ltd, 2009. NPV Scheduler 4 Help
a COG of five per cent. As mentioned before, the M and I Contents.
resources turned into P and P reserves based upon the blocks JORC, 1999. Australasian Code for Reporting of Mineral
that fell within the ultimate pit acknowledging the feasibility Resources and Ore Reserves ( f i e JORC Code), Joint Ore
of economically and technically mining each block and the Reserves Committee of The Australasian Institute of Mining and
geological knowledge of the blocks considering the resources Metallurgy, Australian Institute of Geoscientists and Minerals
classification criteria adopted for this deposit. Inferred Council of Australia [online]. Available from: <http://www.jorc.
resources that felt within the pit shell were not classified as erg>.
reserves, keeping the total as only the sum of P and P reserves. Matheron, G,1963.Principles of geostatistics, Economic Geology,
The goal of the next drilling campaigns is to better investigate No 58, pp 1246-1266.
the areas classified as IND and INF resources in those areas Sinclair, A J and Blackwell, G H, 2002.Applied Mineral Inventory
that are planned to be mined in the first years of production. Estimation, 381 p (Cambridge University Press: Cambridge).
ACKNOWLEDGEMENTS The Canadian Institute of Mining (CIM), 2000.Resource and reserve

definitions: CIM standards on mineral resources and reserves -
The authors would like to thank to Fosfertil S/A for its support Definitions and guidelines, CIM Standing Committee on Reserve
along the development of this study and authorisation to Definitions, The Canadian Mining and Metallurgical Bulletin,
publish this paper. This work is part of a complete review of 93(1044):53-61.
the estimation techniques used at Tapira's mine, which it is Underwood, R and B, Tolwinski, 1998. A mathematical
still ongoing, and will advance in further developments and programming viewpoint for solving the ultimate pit problem,
become the standard for other Fosfertil operations. European Journal of Operational Research, (107):96-107.
ORGANISING COMMITTEE
Richard Flanagan (Conference Chair)

Bob Adam
Mark Adams
Barry Grant
John Hearne
Andrew King
THE AUSIMM
Miriam Way (Director of Conferences and Events)

Belinda Martin (Senior Coordinator, Conferences and Events)
Olivia Tet Fong (Coordinator, Publications)
iii
REVIEWERS
We would like to thank the following people for their contribution towards
enhancing the quality of papers included in this volume:
Bob Adam Andrew King
Mark Adams Paul Lewis
Allan Blair Peter Llewellyn
Ivy Chen Florian Morris
Jacqui Coombes Steve OGrady
Gerard Danckert Santi Pal
Peter Davies Peter Sage
Chris Davis John Sheppard
Phil Dight Martin Staples
Richard Flanagan CIive Workman-Davis
John Hearne
V
SPONSORS
Principal Sponsor
ix
EX HIBIT0R S
X
Davey Bickford is a resolutely high-tech initiation company that is committed to sewing its customers around the world.
-
Davey Bickford has been in business of Initiation Systems for the mining, quarrying and construction industries for 170 years and has over
20 years of digital blasting experience
HIGH-TECH INITIATION COMPANY

Driven by a culture of commitment Since 1839, Davey Bickford has been Davey Bickford is an independent well financed private company.
and precision, Davey Bickford is a a global specialist in pyrotechnic Its solid, long-term financial backing has enabled the company to
high-tech company where science initiation and is today a leader in forge enduring partnerships and is evidence of the sound strategic
and high-technology converge to many technological advances in the choices made over the years.
offer customers high value added Blasting Industry. Today, the A pure player in Initiation Systems, Davey Bickford is having USD
expertise through the synergies of company provides products and 80 million sales (2008) and over 400 employees.
know-how in areas like industrial services for many fields of application Davey Bickford can supply Initiation Systems on a global basis
processes, electronics, and automatic including Mines, Quarries, with offices and permanent staff in Europe, Canada, South
controls combined with knowledge in Construction, Seismic exploration, America, Australia. Davey Bickford has also built strong and
the sciences of chemistry and and other pyrotechnic security numerous distribution partnerships, particularly in the Electronic
geology. systems for various industries Initiation System domain.
.A Strong Innc.\citic I 7inch I<ecord
1631 In~entionand Piitent of the l3irkford Sate& t . t i s F ?MI2 C1 M \ u Igniter for Car Pa.rsi\e Safet)
I906 Inrention and Patent of the -1)ctonatinS CoIJ 2004 hew Generation >on-Electric Shock lube
1965 t Isctrir. I.)ctun,iior 2005 I aunch of the D \\O TRONICI 11
1971 Air Figliler .kitison S>stzini 2007 I aunch of the D\VEYTKOYICb 111
1 9 8 Launch of the I)AVI..Yl KOSICc Electronic Detonator 2008 Launch of Dateltronic Blasting Softwan (DID)
1999 PI MS*Igniter h i Car Pasine Saki! 2009 Launch of D.\Vk \ IXONIC.2 Remote Blaster
w
MISSION
Davey Bickford mission is to bring its customers highly innovative and value added solutions that help %em to improve their safety and profitability and to
lower the environmental impact.
As an example, blast induced vibration can be significantly controlled and reduced through the application of electronic blasting systems and vibration
control technology. This is not only beneficial in sensitive areas such as residential zones where there are concerns of potential damage and liability but
also in all mining operations where long term sustainable production processes are the rule.
Electronic blasting is precisely a technology area where Davey Bickford has put all of his efforts to become one of the world leaders thanks to continuous
research and developments investments and a constant touch with customers expectations.
The resources dedicated to basic and applied research help increase product performance and also the technical expertise at Davey Bickford. Most
importantly, they help the company anticipate market needs and take advantage of new technologies.
COMMANDING PERFORMANCE PASSIONFORPROOUCTO IN

With its vast knowledge of pyrotechnic initiation The teams at the Davey Bickford manufacturing facility have proven their ability to
technology, Davey Bickford has amassed an unsurpassed implement high-performanceproduction systems. The Hhry Plant is a place where the
pool of expertise. companys proven expertise and know-how converge within the chemical laboratories,
The synergy of know-how and skills is the motivating force the R&D, production and quality departments,
behind the companys performance in every market it
serves. The facility is fully qualified to mass produce initiation systems, from pyrotechnic
components to specific safe packaging.
Mines Seismic Exploration
Aeronautics Quarries And Construction Within this highly specialized pyrotechnic environment, the unrivalled experience of its
Security Systems staff makes this location a highly flexible and comprehensive production tool.
Hdry Plant in Numbers

*:* Located in Hery, Burgundy, 170 km south from Paris.
0:.Land: 100 acres
*: Manufactunng : 30 000 r n 2
*: 2 tesbng zones - 5 laboratories
9 Half a billion units produced 50 far.
I
M V E Y BICKFORD A W RAW PTY LTD
Davey BiCWord Australia Pty Ltd started trading in 2005 with Kim Greenock as Vice President, Sales and Business Development covenng
Australasia.
The first years was spent gaining all the required product registrations, insurances and product licences for all the states in Australia as well as
introducing Daveytronic systems to the Australian market.
Wkh a lot of hard w& the infrasbucturewas set in place for the import, stonng and selling of the Daveytronic products, the team gradually
increased to further develop our now working relationships with the end users.
OEWsr BlckfardAusballe prV LM. W Box 298, m t d E PARK WA 6917. Tel: +61(0)8 9207 1066 Fax: 4 1 (0)s9207 1833 Emall: admh@daveyMdrford.m.au www.dawy&kford.ew
LIEBHERR
Liebherr-Australia Pty Ltd is a member of the Liebherr
group, which consists of more than 100 autonomously
operated companies in over 50 countries around the
world.
All of the Liebherr companies come under the

ownership and management of Liebherr-International
AG of Bulle in Switzerland. Liebherr International
Liebherr-International (and hence the Liebherr group) is

wholly owned by the Liebherr family, making it one of
the largest family owned businesses in the world.
In 1983 the potential of the Australian market and

Liebherrs desire to further support customers in this
region saw the establishment of Liebherr-Australia in Liebherr Australia
Adelaide. -
Head Office Adelaide
In the early 199Os, Liebherr Australias involvement with

the mining industries via the supply of mining
excavators in Queensland and Western Australia
created a need for major branches in those states with
regional product support centres. This was followed by
the establishment of PT Liebherr Indonesia Perkasa, at
Balikpapan in Indonesia.
I
In 1995 Liebherr Australia took a major step into the Mining Trucks
Hunter Valley coal mining region of New South Wales
with the establishment of a branch at Mt Thorley near
Singleton. In that same year Liebherr-International
acquired the Wiseda mining truck business in the USA
and thus, Liebherr entered the mining truck market.
Liebherr Australia has continued to expand in Australia

and currently, in addition to the head office in Adelaide
has major mining focused branches in Mackay, Mt Mobile Cranes
Thorley and Perth, with service support centres in
Townsville, Mt Arthur Coal, (near Mt Thorley), Newman
and Telfer.
Liebherr Australia also includes a mobile crane and

crawler crane division in Australia and New Zealand,
further complementing Liebherr Australias commitment
to customer support.
I
Hydraulic Excavators
AUTHOR INDEX
Index Terms Links
B
Ballantyne, R 121
Bamford, W E 83
Beck, D 71
Beer, A J 45
C
Capponi, L N 127
Cerny, L 93
Costa, J F 127
Cruickshanks, S 39
D
Danckert, G 3
Davis, P 79
de Lemos Peroni, R 127
Dixon, J 93
F
Ferreira, N 103
Fillery, B 71
G
Gedeon, T 31
Gontijo, E 127
Guerrero, M 81
H
Heap, M E 45
Hendersonhall, B D 15
Hennessy, R 7
K
Kerr, N 15
Knights, P F 111
L
Lai, F J 83
Li,T 83
Lucas, D 15
This page has been reformatted by Knovel to provide easier navigation

Index Terms Links
M
Marton, A 57
McArthur, C 93
McElman, C 103
McHugh, C 7
Murray, C 65
N
Narendranathan, S 45
O
Oppolzer, F 7
P
Paton, S 111
Pennisi, C 15
R
Renstrom, A 115
Reusch, F 71
Reynaud, H 81
S
Schmid, M 81
Singh, U 93
Smith, S 7
Stokes, A 7
T
Tappeiner, S 81
Taylor, E 65
Taylor, K 31
V
Veiga, D C 127
Y
Yuen, S T S 83
Z
Zhu, D 31
This page has been reformatted by Knovel to provide easier navigation

Seventh Large Open Pit Mining Conference 2010 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Seventh Large Open Pit Mining Conference 2010 PDF

Uploaded by

Copyright:

Available Formats

SEVENTH LARGE OPEN PIT

MINING CONFERENCE 2010

The Australasian Institute of Mining and Metallurgy

All papers published in this volume were refereed prior to publication.

ISBN 978 1 921522 22 2

Desktop published by:

Innovation in an Uncertain World G Danckert 3

Enhanced Drilling in Rio Tinto CMcHugh,A Stokes, 7

Sydvaranger Magnetite Mine, Norway - Owner S Cruickshanks 39

Slope Monitoring in Open Pit Mining - A Case S Narendranathan, A J Beer 45

Minimising Crest Loss by Eliminating Surface A Marton 57

Ensuring Professional Engineering Capability - E Taylor and C Murray 65

Three-Dimensional Hydromechanical Simulation D Beck, B Fillery and 71

Linking Operator Competency with Performance P Davis 79

Capacity Increase of the Worlds Strongest Belt M Guerrero, H Reynaud, 81

Incorporating Economic Risk in Pit Slope Designs F J Lai, W E Bamford, 83

Mine Planning and Risk Management Associated L Cerny, J Dixon, 93

Payload Variance Effects on Truck Bunching P F Knights and S Paton 111

Productivity Enhancements to Increase Ore A Renstrom 115

Mine Control - Why do we do it, and how do we R Ballantyne 121

Long-Term Mine Planning at the Largest Phosphate R de Lemos Peroni, 127

FIG 1-location of RioTinto Iron Ore Pilbara operations.

Commissioning of remote control

Production performance of the automated drill

I 23/01/2006 I Drill enhancement proiect approved. 1

1 17/07/2008 1 Firstfully tele-remote percussion productionhole. 1

Automated productiondrilling of drill 21R03 24 hours seven days per week.

1 Description I Manual drilling Manual drilling with high precision

1 I Drill Drill Drill

Average rotary bit life (m) 21/05/2009 to 17/11/2009 750

Availability (%) 06/09/2009 to 07/03/2010 l___l_

Utilisation(%) 06/09/2009 to 07/03/2010 49 ~ 47 48 57

Description of the mine

Black Carbonaceous Shale (BCS) Hangingwall Sandstone (HWS)

Black Laminated Shale (BLS) Cambrian Limestone (CLS)

Hanging wall Siltstone-Shale (HWD) Clayey Gravels

I Unit I Abbreviation I UCS (MPa) I

Black Carbonaceous Shale (BCS) Hangingwall Sandstone (HWS)

Black Laminated Shale (BLS) Cambrian Limestone (CLS)

Hanging wall Siltstone-Shale (HWD) Clayey Gravels

I November 2007 1 - Batter scale failures 1104-1120 RL initiate redesian of SW wall. i

January 2008 4-. Batter

Batter scale wedge failures as mining progresses.

Month Southwest wall North wall

lonth Southwest wal North wall Northwest wall

TABLE 4 CONT ...

Month Southwest wall North wall Northwest wall

FIG 10 - Photographofthe northwestwall prior to failure, taken 12 June 2009.

FIG 14 - Photographof the north wall failure, taken 13 November 2009.

FIG 2 - Haul truck dumping a load into the run-of-mine bin.

FIG 3 - Multi-control situation in remote rock breaking process.

FIG 4 - Eye tracking for camera control.

CONCLUSIONS AND FUTURE DIRECTIONS

FIG 8 -Objective performancecomparison-goals and kicks.

SEVENTH LARGE OPEN PIT MINING CONFERENCE / PERTH, WA, 27 - 28 JULY 2 0 1 0 39

FIG 1 -Project location.

FIG 2 - Deposits and existing pit outlines.

1) Preliminary Geotechnical Investigation involving:

2) Formulation of a Preliminary Geotechnical Model:

0 As construction continued, safety bunds were created at the

Design The V W P sensors were installed along the centreline of

FIG 3 -Layout of Daydream instrumentation.

where: ac= 76.05 mm/P