We are Metallurgists, Not Magicians
Landmark Papers by Practising Metallurgists
Edited by D Pollard, G Dunlop and J Herzig
Spectrum 23
We are Metallurgists,
Not Magicians
Landmark papers by practising metallurgists
Spectrum 23
Edited by D Pollard, G Dunlop and J Herzig
Published by:
The Australasian Institute of Mining and Metallurgy,
Ground Floor, 204 Lygon Street, Carlton Victoria 3053, Australia
© The Australasian Institute of Mining and Metallurgy 2017
No part of this publication may be reproduced, stored in a retrieval system or transmitted
in any form by any means without permission in writing from the publisher.
All papers published in this volume were peer reviewed before publication.
The Institute is not responsible as a body for the facts and opinions advanced in any of its publications.
ISBN 978 1 925100 62 4
Desktop published by Belle Doley, Ross Palmer, Kelly Steele, Alexandra Talbot and Mia Wotherspoon
Printed by:
Focus Print Group
2 McIntyre Street, Burwood Victoria 3125, Australia
Front cover image (top):
Construction at Newmont Boddington Gold, photo courtesy of Newmont Australia.
Back cover image:
Just before sunrise at the Phu Kham Copper-Gold Operation process plant, photo courtesy of PanAust Limited.
introduction
The majority of papers included in this volume have been selected from AusIMM MetPlant and MillOps conferences from 2008 to 2016. The
compilation was inspired by the realisation that the Institute’s metallurgy conferences deliver papers of enormous value to practising metallurgists
and, unless they happen to be conference delegates, the information contained in them is otherwise lost to the wider metallurgical community.
They provide a unique resource for professionals at all career stages, whether working in design, construction or the operation of metallurgical
processing plants.
The editors reviewed over 40 AusIMM conferences identifying metallurgy papers relevant to the objectives of the present publication, viz:
•
•
•
they are relevant to good design, construction and operation of metallurgical processing plants
they do not have an ‘expiry date’
they pass on relevant experience and inspiration to the next generation of plant designers and operators.
The papers have been updated, a few are new for the volume, and some have been re-written to combine the core narrative of two or more
conference papers. This compilation presents them in themed chapters, grouped with relevance to the challenges of the future as well as the
present.
Most authors expressed how privileged they were to have their papers included and how the initiative to have this wisdom distilled in one
place was a noble one. Many responded with additional comments, some of which are incorporated overleaf as testimonials.
Through this publication the editors hope that fewer wheels will be re-invented, and the accounts of successful ideas and innovative thinking
will inspire pathways to improved plant designs, operating efficiencies and profitability. Metallurgists in planning, plant design, construction,
operation, improvement and management will find the volume essential reading.
ACKNOWLEDGEMENTS
The editors are grateful for the suggestions and feedback from members and ex-members of the Metallurgical Society Committee, including Dean
David, Aidan Giblett, Andrew Newell, Philip Stewart and Peter Tilyard.
Janine Herzig joined the original editors as the paper selections were being finalised. The three editors worked together, contacting and
liaising with authors and securing the generous sponsorship from industry necessary to ensuring a hardcopy as well as electronic version. Janine
edited all of the revised papers, making an invaluable contribution.
David Pollard, Geoff Dunlop and Janine Herzig, Editors
testimonials
“A splendid outcome and a worthwhile contribution to the industry.”
E McLean
“My fellow authors and I are honoured that our papers are
considered for inclusion. The list of papers is very impressive.”
D Bennett (PanAust)
“I’m honoured to have some of my papers considered as ‘landmark papers’ and to be included as
a ‘Practical Metallurgist’! A very worthy project.”
T Napier-Munn
“I would be honoured … I see it as a great idea.”
P Thwaites (XPS)
“It is a great honour to be included in a such august company
and I must congratulate you on this wonderful outcome
and legacy to your vision, drive and management skills.”
A Newell
“Honoured to be included and love the title.”
P Bartsch
about the editors
David Pollard is a consultant metallurgist with an interest in education and professional development. His employment after graduation was
with the Port Kembla steelworks, and later as a metallurgy lecturer at University of Melbourne and the South Australian Institute of Technology
(now UniSA). For over a decade he managed the professional development activities of the Australian Mineral Foundation (AMF), providing short
courses and conferences across Australia and internationally, across a range of high level technical and management topics. Many of these courses
were offered in association with AusIMM conferences.
David brought the Metallurgical Plant Design (MetPlant) conference series to the AusIMM when AMF closed, and has convened the series
with Geoff Dunlop since 2002, and from 2017 with Janine Herzig. He has twice been Chair of the Adelaide Branch, and was Chair of the AusIMM
Metallurgical Society (MetSoc) from 2008 to 2013.During this time MetSoc initiated the groundwork for AusIMM to publish the third edition of
the Sir Maurice Mawby volumes, Australasian Mining and Metallurgical Operating Practices.
We are Metallurgists, Not Magicians was initially undertaken by David and Geoff, with the support of MetSoc, and will add to the professional
development resources for practising metallurgists and processing plant designers.
Geoff Dunlop is a metallurgical engineer who began his career at Mt Isa, rose to Concentrator Superintendent, and moved to a site research
role where he carried out early investigations on the geometallurgy of the Mt Isa orebodies. He characterised the ‘3000’ and Lead/Zinc orebodies.
He moved to AMDEL in 1970 as a project metallurgist, and developed process flow sheets and systems to provide financial information about
performance. In the early 1990s he worked with AMDEL collaborating with MIM Process Control and JKTech to develop instrumentation for
Indian, Burmese and Iranian processing plants. He continued to consult for some years following his retirement, including for BOC gases regarding
automated equipment for plants, as well as for international clients.
In 1998 Geoff edited, with Norton Jackson and Peter Cameron, the AMF Conference proceedings for ’Mineral Processing and Hydrometallurgy
Plant Design‘, a conference convened by David Pollard. This collaboration led into the partnership between David and Geoff on the AusIMM
MetPlant conference series from 2002.
Janine Herzig began her graduate career in Mount Isa as a metallurgical engineer with MIM in the Lead/Zinc Concentrator. She then moved into
the mineral sands sector with Iluka in various locations across Queensland and Western Australia, before being appointed Principal Metallurgist
for the Murray Basin Operations feasibility work and execution. After ten years with Iluka, Janine accepted the role of Mineral Processing Manager
with AMDEL and ultimately as the General Manager of the Minerals and Industrial Division, where she oversaw programs spanning multiple
commodities and mineralisation styles. Key projects included the first major ODX geometallurgy program under WMC then BHP, which involved
establishment of a major new metallurgical and geochemical testing facility, the Prominent Hill feasibility study, and management of a major
expansion of the on-site laboratories for the Whyalla steelworks and processing plants. Janine established a consulting business after leaving
AMDEL in 2009 conducting due diligence, desktop reviews and project management across a range of commodities including industrial minerals,
graphite, precious and base metals.
She has served on the AusIMM Adelaide Branch Committee since 2005, holding the positions of Chair, Secretary, newsletter editor,
and member on the South Australian scholarship interview panel. She continues to serve on a range of AusIMM conference organising
committees and other specialist taskforces. She was honoured to be invited to assist David Pollard and Geoff Dunlop with this publication,
and believes that it will become the essential reference for all metallurgists, plant designers, consultants, students and new professionals
for many years to come.
Sponsors
Platinum Sponsors
Gold Sponsor
Silver Sponsor
Bronze Sponsors
Our purpose is to create value and
improve lives through sustainable
and responsible mining
Follow us on:
www.newmont.com
PanAust is an Australian-headquartered copper
and gold producer with producing assets in Laos
and a portfolio of organic growth projects in
Laos, Papua New Guinea, Myanmar and Chile.
PanAust’s producing assets are the Phu Kham
Copper-Gold Operation and the Ban Houayxai
Gold-Silver Operation, both located in the
Company’s 2,600 square-kilometre Phu Bia
Contract Area in Laos.
In Papua New Guinea, PanAust holds an 80 per
cent interest in the Frieda River Copper-Gold
Project, with the remaining 20 per cent held by
joint venture partner, Highlands Pacific.
In Myanmar, PanAust holds an 80 per cent interest
in Wuntho Resources Company Limited (WRCL)
and has established a joint venture with Myanmar
Energy Resources Group International Company
Limited (MERG), a Myanmar-based company which
holds the remaining 20 per cent of WRCL. WRCL
owns 100 per cent of the rights to areas covered
by three Exploration Licences in the Sagaing
Division of Myanmar.
At PanAust, operational excellence and
sustainable business outcomes are pursued in a
complementary manner. This strategy, which sees
excellence in one area underpin, encourage and
facilitate excellence in the other, has been the
driving force behind PanAust’s success. PanAust is
determined to remain Way Ahead.
For more information, visit www.panaust.com.au.
gold SPONSOR – Ausenco
Ausenco is a world leading engineering company providing consulting, project delivery and asset management and optimisation solutions to the
resources, energy and infrastructure sectors. We add value to our clients’ projects and deliver results even in the most challenging environments.
Ausenco has a global footprint with offices and current projects in Australia, the Pacific, Asia, Africa and North and South America. Our experts
thrive on their ability to create ingenious solutions that improve our clients’ return on investment. We’re proud to be a Gold Sponsor of this
publication and support the continued development of excellence in Minerals Processing.
CONTENTS
Overview
‘We’re metallurgists, not magicians!’
E McLean
3
Back to the future – why change doesn’t necessarily mean progress
P D Munro and P A Tilyard
11
Back to the future – still on the dark side
P D Munro
19
Undermining productivity – when good key performance indicators go bad
J Pease
27
Presented at MetPlant 2011, this paper has been updated and edited for this compilation.
Presented at MillOps 2009, this paper has been updated and edited for this compilation.
Presented at MillOps 2016 as a follow-on to the paper: ‘Back to the future - why change doesn’t necessarily mean progress’, this paper has been updated and edited
for this compilation.
Presented at MillOps 2014 as a PowerPoint presentation, this paper has been written specifically for this publication.
Geometallurgy
Geometallurgy – what do you really need to know from exploration through to production?
K Ehrig
33
Integrating geometallurgy with copper concentrator design and operation
G Harbort, K Jones, D Morgan and C Sola
37
Integrated mining and metallurgical planning and operation
P L McCarthy
55
Presented at MetPlant 2013, this paper has undergone minor edits for this compilation.
This paper was presented in various forms at MetPlant 2011, MillOps 2014 and MillOps 2016. It has been updated with a new geometallurgy model, undergone
significant edits and is the integration of the papers presented at the three events outlined above.
This is a compilation of three papers presented and published for MetPlant 2011, MetPlant 2013 and MetPlant 2015.
Project economics
Guidelines for economic evaluation of projects
P Card
65
Sensible cost cutting for resource projects
D Connelly
69
When does further processing at the mine site make sense?
C Fountain, S La Brooy and G Lane
75
The ABC of Mine-to-Mill and metal price cycles
P Cameron, D Drinkwater and J D Pease
85
Base metals concentrate sales contracts – change Pavlov and the dog
P D Munro and S E Munro
91
Presented and published at MetPlant 2011, this paper has been significantly reviewed and updated for this compilation.
Presented at MetPlant 2011, this paper has undergone significant edits and updates for this compilation.
Presented at MetPlant 2008, this paper has had minor edits for this compilation.
Presented at MillOps 2016, this paper has been reviewed and abridged for this compilation.
Presented at MetPlant 2015, this paper has undergone minor edits for this compilation.
Project design
Karouni Gold Project from drill core to commissioning
K Nilsson and D Connelly
101
Upgrades, modernisations, automation and expansions … where will the expertise, capability and skills come from in the future?
R Coleman, J King and T Hunter
107
Presented at MetPlant 2015, this paper has undergone minor edits for this compilation.
Presented at MillOps 2016, this paper has undergone minor edits for this compilation.
x
115
Is an 80th percentile design point logical?
D David
Presented at MetPlant 2013, this paper has undergone significant edits for this compilation.
Measuring and taking notice of orebody variability – an essential ingredient for reliable plant design
D David
121
Presented at MetPlant 2015, this paper has undergone significant edits for this compilation.
Getting optimum value from ore characterisation programs in design and geometallurgical projects associated with comminution circuits
S Morrell
131
Presented at MillOps 2009, this paper has been updated and edited for this compilation.
137
Cost-effective concentrator design
G Lane, P Dakin and D Elwin
Presented at MetPlant 2011, this paper has undergone minor edits for this compilation.
Project management and delivery
145
Fatal flaws in technical due diligences
A J H Newell
Presented at MetPlant 2015, this paper has undergone significant updates and edits for this compilation.
155
Guidelines for mineral process plant development studies
P R Whincup
Presented at MetPlant 2008, this paper has undergone minor edits for this compilation.
163
Project delivery
G Lane and E Skinner
Presented at MetPlant 2013, this paper has undergone minor edits for this compilation.
173
Mineral project management – a perspective from four decades in the industry
J S Dunlop
Presented at MetPlant 2013, this paper has been updated and edited for this compilation.
183
Keeping projects on the rails
J Canterford
Presented at MetPlant 2011, this paper has undergone significant edits for this compilation.
187
Operations versus projects – how do people think and what are the implications?
G Lane and B Clements
Presented at MillOps 2012, this paper has undergone minor edits for this compilation.
193
Performance testing – when, what and how?
G Lane, M Davis, E McLean and J Fleay
Presented at the Project Evaluation 2007 conference, this paper has undergone minor edits for this compilation.
Unit design and development
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
D Bennett, I Crnkovic, P Walker, A Hoyle, A Tordoir, D La Rosa, W Valery and K Duffy
201
The subject matter for this paper was written, published and presented at three AusIMM conferences: MillOps 2012, MillOps 2014 and MetPlant 2015.
Significant effort has gone into combining, updating and editing this combination of the three papers for this compilation.
235
What can go wrong in comminution circuit design?
C Bailey, G Lane, S Morrell and P Staples
Presented at MillOps 2009, this paper has undergone significant edits for this compilation.
243
Do we need a gravity circuit or not? A case study in applying best practice
A Giblett and K Afewu
Presented at MetPlant 2015, this paper has undergone minor edits for this compilation.
Solvent extraction of uranium – towards good practice in design, operation and management
P Bartsch and S Hall
249
Presented at MetPlant 2011, this paper has been updated and edited for this compilation.
257
Design of copper-cobalt hydrometallurgical circuits
G Miller
Presented at MetPlant 2008, this paper has undergone minor edits for this compilation.
263
A review of best practice in gravity circuit design and operation
A Giblett, A Bax, G Wardell-Johnson and W Staunton
Presented at MetPlant 2013, this paper has undergone minor edits for this compilation.
xi
Leach residue and pregnant liquor separation – process and capital comparison of counter-current decantation and counter-current
washing with vacuum filtration
R Klepper and P McCurdie
273
Presented at MetPlant 2011, this paper has been updated and edited for this compilation.
285
Filtration test work – extracting the whole story for studies and design
G Bickert and B Länger
Presented at MetPlant 2011, this paper has been updated and edited for this compilation.
297
Considerations for effective gold process development
A Giblett, D Appelhans and R Dunne
Presented at MetPlant 2013, this paper has undergone minor edits for this compilation.
307
Energy efficient ball mill circuit – equipment sizing considerations
A Jankovic and W Valery
Presented at MetPlant 2013, this paper has been updated and edited for this compilation.
315
Advances in dense medium cyclone plant design
T J Napier-Munn, G Gibson and B Bessen
Presented at MillOps 2009, this paper has undergone minor edits for this compilation.
Sampling, metallurgical accounting and control
327
Using metallurgical data to drive continuous improvement
R Dunne
Presented at MillOps 2016, this paper has been updated and edited for this compilation.
333
The importance of sampling in the mineral industry
R J Holmes
Presented at MetPlant 2013, this paper has been updated and edited for this compilation.
A progressive, iterative approach leading to reliable metallurgical accounting sampling systems
F F Pitard
343
This paper was written for MetPlant 2015 but not presented at the conference. It has undergone significant edits for this compilation.
349
Sampling, corporate governance and risk analysis
G J Lyman and F S Bourgeois
Presented at MetPlant 2015, this paper has been significantly edited and updated for this compilation.
355
Metallurgical management tools for continuous improvement
W McCallum and R Dunne
Presented at MillOps 2012, this paper has undergone minor edits for this compilation.
Measuring the influence of sample size on the precision and accuracy of gravity gold estimation
A Giblett and T J Napier-Munn
361
Presented at MetPlant 2015, this paper has been updated and edited for this compilation.
365
Defining practical metallurgical accounting discrepancy limits for gold operations
A Giblett, R Dunne and K McCaffery
Presented at MillOps 2012, this paper has undergone minor edits for this compilation.
Manual control, process automation or operational performance excellence – what is the difference?
P Thwaites
373
Presented at MillOps 2014, this paper has been significantly updated with new information for this compilation.
401
Unlocking processing potential by empowering our operators
X Li, M S Powell and W McKeague
Presented at MillOps 2012, this paper has been updated and edited for this compilation.
Current developments in the operation and control of autogenous and semi-autogenous grinding mills in Australia
J Karageorgos, Y Atasoy and D Baas
411
Presented at MetPlant 2008, this paper has undergone minor edits for this compilation.
Environmental management and sustainability
Mine waste risk minimisation by integrated waste management and process optimisation
D Brett
419
Presented at MetPlant 2013, this paper has been updated and edited for this compilation.
427
Sustainability, made easy – a business improvement case study
G Corder, N Currey and G Becker
Presented at MetPlant 2013, this paper has been updated and edited for this compilation.
xii
Continuous improvement – how changes in metallurgy can have a substantive impact on environmental impact and closure: lessons
learnt from environmental auditing
P Mulvey and G McMillan
437
Presented at MetPlant 2013, this paper has undergone minor edits for this compilation.
Process improvements and case studies
Improving fines recovery by grinding finer
J D Pease, M F Young and N W Johnson
445
Cunning solutions to process improvement
T J Napier-Munn
455
How to prioritise process improvements
D J Hill
461
Some practical problems in running statistically valid plant trials and their solution
T J Napier-Munn
469
Presented at MetPlant 2004, this paper has had significant updates and edits for this compilation.
Presented at MillOps 2014, this paper has undergone minor edits for this compilation.
Presented at MillOps 2014, this paper has had a significant update for this compilation.
Presented at MetPlant 2008, this paper has undergone minor edits for this compilation.
The optimisation of semi-autogenous grinding and ball mill based circuits for mineral processing by means of versatile and efficient
high pressure grinding roll technology
S W Kirsch and M J Daniel
475
Presented at MillOps 2009, this paper has been significantly updated for this compilation.
Optimisation opportunities for high pressure grinding rolls circuits
M S Powell, M M Hilden, C M Evertsson, G Asbjörnsson, A H Benzer, A N Mainza, L M Tavares, B Davis, N Plint and C Rule
483
An update on applications of high frequency screens in closed grinding circuits
S B Valine, J E Wheeler, B Packer and A N Cavendor
499
The influence of liner wear on milling efficiency
P Toor, T Perkins, M S Powell and J Franke
509
Development of the Fosterville gold mine heated leach process
M Binks and P Wemyss
523
Collector – addition point and consumption
C J Greet, W J Bruckard and D MacKay
529
The Ausenco carbon reactivation kiln initiative
J K Claflin and S R La Brooy
537
Lessons learnt and performance – installing and commissioning an Ausenco carbon reactivation kiln in Africa
J K Claflin, S R La Brooy and D Preedy
541
Carbon management in a high gold price environment
S R La Brooy and J K Claflin
549
The impact of gravity gold recovery at Kalgoorlie Consolidated Gold Mines
A Giblett, D Hillier, K Parker and V Ramsell
559
Agglomeration – the key to success for the Murrin Murrin heap leach
D Readett and J Fox
569
Author index
575
Presented at MillOps 2012, this paper has been significantly updated for this compilation.
This paper is a compilation of two papers from MetPlant 2011 and MillOps 2012. It has been significantly updated with additional case studies and information
obtained for the Alban Lynch publication Comminution Handbook.
Presented at MetPlant 2011, this paper has been significantly updated for this compilation.
Presented at MetPlant 2011, this paper has been significantly updated for this compilation.
Presented at MetPlant 2008, this paper was revised several years later for Transactions of the Institutions of Mining and Metallurgy – Mineral Processing and
Extractive Metallurgy, and has been further updated and edited for this compilation.
This paper was specifically written for this compilation as an author foreword to the following two papers.
This paper was presented at MetPlant 2015. Context for this paper and the paper immediately below has been provided via an author foreword (above), which also
provides significant new insights and updates.
This paper was presented at MetPlant 2013. Context for this paper and the paper immediately above has been provided via an author foreword (above), which also
provides significant new insights and updates.
Presented at MillOps 2012, this paper has undergone very minor edits for this compilation at KCGM’s request.
Presented at MetPlant 2011, this paper has been updated and edited for this compilation.
xiii
Overview
Contents
‘We’re metallurgists, not magicians!’
E McLean1
ABSTRACT
The decision-making checklists and the hierarchy of information for process and
plant design are generally wide-ranging and complex.
Grades, mineralisation and ore types in an orebody, the mine development
sequence and results from metallurgical and physical test work present complex
and competing influences for flow sheet development, values for design, equipment
selection, extraction/recovery performance and operating strategy.
Challenges are presented on several fronts: how to manage these variables;
understand their impacts on performance; and best provide a cost-effective plant
design and robust process that meets daily, weekly and monthly operating objectives.
Throughput, production and grade or product specifications need to be met. The
orebody is not homogeneous; its range of metallurgical and physical characteristics
may be moderately to highly variable. Average values derived from test work results
are unlikely to be suitable for design purposes.
This paper assesses some of the competing influences and conflicting issues in design
and equipment selection, and their effect on production and performance expectations.
THE DESIGN SEQUENCE
Each item of process equipment, the flow sheet, process logic and the operating
strategy for the treatment plant are there by design and purpose. This is because a
sequence of process engineering activities has led to their implementation, installation
and construction in the project.
The development activities for the project follow the staged sequences shown in
Table 1. Although not exhaustive, this table simplifies a complex procedure and
identifies the main process and engineering decision-making activities. In practice,
iterations following reviews, revisions at each stage and the linked sequence of stages
increase the complexity in project delivery.
Reviewers of the performance of project delivery, plant start-up and operation
(Agarwal, Brown and Katrak, 1983; Performance Associates International Inc, 2017)
identified fundamental issues that contributed to financial shock, whether it was due
to attaining nameplate capacity at a slower rate than anticipated, or not achieving
this at all. They noted that, notwithstanding experienced and competent mining,
design and planning personnel on the project, the chief technical problems, start-up
difficulties and poor performance for new mineral processing operations, occurred
across all stages of the project from the following items:
•• representativity of samples – high heterogeneity of orebodies, difficulties with
adequate sampling, extent of testing was inadequate (stage I in Table 1)
•• definition of the process – complexities of the process route, influence of process
water (stage II)
•• engineering – design deficiencies with mechanical equipment, equipment
design and/or installation is inappropriate (stage III)
•• operation – operators lack knowledge and skills to properly operate the plant,
management is unable to cope with the problems experienced during initial
operation, maintenance is not executed properly, particularly during the initial
operation (stage IV).
The authors stated that although problems with mechanical equipment were
expected, these risks can be reduced through efficient design and scale-up. They
attributed a number of these problems to basic design deficiencies such as incorrect
specifications, unit capacity and duty.
1. FAusIMM, Manager Minerals Consulting,
Ausenco Minerals & Metals, South Brisbane
Qld 4101. Email: eddie.mclean@ausenco.com
The activities in Table 1 match elements in the schematic project flow chart (Lane
et al, 2008) in Figure 1, which shows the relationship between design inputs and
project outcomes:
•• stage I input = design inputs (green boxes)
3
E McLean
Table 1
Development sequence for principal process design and engineering activities.
Stage
Development activities
I – Input
Geology, mineralisation, resource model
Description of key activities in development stage
Discuss deposit(s) with exploration geologists
Review long- and cross-sections showing lithology and structures
Geological interpretation, reconcile metallurgical classification for zones or domains
Sample selection and sampling protocols Representative samples and the basis for selection
Characterisation – metallurgical and physical, of the major ore types, ore zones or domains
Variability – metallurgical and physical, of parameters that influence design and production
Production composites, as required, for selected periods during the initial two to three years of operation
Test work, by laboratory and vendor
II – Definition
III – Engineering
IV – Deliver
Project
Preparation of samples, ensuring integrity of sample, preventing oxidation
Batch or semi-continuous small or large scale (pilot) for flow sheet development and to obtain data for process design
Reproducibility and robustness of tests; simulation of operating conditions; check/repeat and diagnostic tests
Analysis of data
Collate and compare, interrogate for trends and relationships, investigate outlier data
Ore sequence in likely mine plan and weighting by ore characterisation
Processing strategy and process alternatives, flow sheet development
Design value
Recommend value/specification for unit process and for process stages
Define duty point in range of operating conditions, considers variability
Calculation – size and select equipment
Text book or standard methodology, with efficiency factors
Include equipment factors – industry applicable, vendor recommended
Include operating factors – experience, reconciliation of operation with previous design and performance
Include agreed or nominated engineering margin, owner margin, allowance for engineering risk
Tender and select for procurement
Vendor receives engineering specification, criteria, data sheets, drawings
Vendor uses proprietary modelling, calculations and database to size and recommend
Vendor submits equipment that complies with throughput and duty required
Vendor selection also considers commercial aspects: performance warranties, penalties for non-performance,
reputation, competition (cost)
New or second-hand equipment may be offered
Technical and commercial evaluation by engineer – recommend to purchase
Install, start-up and ramp-up operations Implementation plan to coordinate and schedule activities
Install and commission in accordance with standards and procedures
Maintenance – planned and in accordance with recommendations
Operator experience and training
FIG 1 – Relationship between design inputs and project outcomes (from Lane et al, 2008).
4
we are metallurgists, not magicians
‘We’re metallurgists, not magicians!’
•• stage II definition = project design concept (orange oval)
•• stage III engineering = project outcomes: engineering
design equipment selection and quantities and take-offs
(blue boxes)
•• stage IV delivery = project execution: implementation,
procurement, construction and commissioning (yellow
boxes).
The project design concept and definition stage are pivotal:
all information channels through here for interpretation and
analysis, and then forms the basis for subsequent engineering
design and planning for the execution phases of the project.
Activities and decisions in this stage are able to impact on the
project cost, schedule and plant performance.
Although all items are integral to a successful project outcome
and some are critically important such as ore representativity
and sampling, this paper reviews the following activities
described in Table 1 that have a high ability to influence the
final project outcome:
•• analysis of the data to develop a viable processing strategy
•• values for design to specify the duties for the unit
processes
•• calculations to size and select the process plant and
equipment.
THE DATA SEQUENCE
Practical and robust interrogation, and interpretation and
analysis of the data from the body of metallurgical information
available are fundamental to successful design. The quality
of data at various stages of project development is also of
consequence with respect to its validity and applicability for
design.
Samples and data
Samples for metallurgical and physical testing are usually
domain-orientated and can be categorised as characterisation
samples and variability samples.
Characterisation samples normally represent a lithology,
oxidation, alteration, mineralogy, grade, or spatial property
of the ore, as appropriate, for which metallurgical or physical
unit parameters are determined by the test work for that
specific ore classification. Variability samples are usually
from each domain or ore classification in which one property
varies with, for example, metal grade, sulfur grade, depth,
location in the geological structure.
Data generated from tests on both these types of samples
can be associated with blocks of ore in a resource model; some
of the more common applications are described as follows:
•• hardness and competency values populate blocks of ore
in which the dominant lithology or alteration type is
identified (characterisation)
•• leach recovery values populate blocks of free-milling
gold ore in which the extent of oxidation is identified,
ie for wholly oxidised, partially oxidised or transition,
or fresh ore (characterisation)
•• a relationship of recovery with a variable (or variables)
can be described by an algorithm to populate blocks
of ore; for example, in gold ores, relationships may
be established between leach extraction and feed
grade, leach extraction and sulfur grade, and extent of
refractory gold and arsenic grade (variability)
•• mill feed from a mine plan and ore schedule can be
interrogated to provide information over a period
we are metallurgists, not magicians
(month, quarter, half-year) to support production and
cash flow forecasting.
Resource composites which are blended by ore types and
grade to represent samples such as whole-of-pit, life-of-mine
and annual averages do not necessarily provide the basic
information and ‘building blocks’ for flow sheet development
and design that characterisation and variability samples do.
Plant operation is unlikely, if ever, to treat such samples
on a sustained basis. Specific ore properties that may cause
metallurgical and operational difficulties are diluted in a
composite sample and any impact these have on performance
is dampened, or perhaps not identified in the test results.
For a design based on average values plus allowances for
mechanical and vendor margins, there is a high likelihood that
the expected performance and corresponding production will
not meet planned targets when treating ores over sustained
operating periods with characteristics significantly above
the average value. Furthermore, the opportunity to catch
up production when ore with characteristic values below
the average are treated may be constrained by physical and
volumetric limits in the plant and equipment; for example,
pump capacity, launder size, screen area, thickener area,
filter rate.
Data analysis
In the early stages of project development when drilling and
geological interpretations are ongoing and resource definition
is in its formative stages, the body of test work data is limited.
Data is generally in small sets of information and the data
selected is typically the maximum or the average of the range.
Incremental increases in input data values may be made to
assess trends and to assess sizes of plant and equipment.
As well, benchmarking of the values may be made with
comparable industry data or in-house data from similar ore
types and similar applications to improve confidence at this
early stage.
As the project develops, not only does the volume of data
available from several stages and iterations of metallurgical
test work programs increase, but so too does the complexity
of information from the amount of detail available with an
improved understanding of ore characteristics and variability.
Superimposed on these are a number of ore feed schedules
to the plant with variable tonnage, grade, mineralisation and
lithology, and different ore exploitation sequences depending
on the mining strategy, mine development plan, pit financial
modelling and resource constraints.
Analysis of the characterisation and variability data by ore
type to determine values for design includes:
•• a statistical analysis which identifies the average, the
standard deviation, minimum and maximum values,
values at nominated percentiles (default is typically 75th
or 80th) of the data population
•• an assessment of outlier low/high data, analysis and
inclusion with basis and comments noted, or discard
with justification
•• a plot of values in an ordered sequence which graphically
shows the distribution, the location of the bulk of the
values, and where the outliers or sparsely populated
values lie
•• identification of values, which characterise the major ore
types, unit process performance, process streams and
water quality
•• determination of a weighted value appropriate for
principal blends of ore in the feed to the plant.
5
E McLean
Sources of data
Input data and information for studies from concept to final
feasibility, for basic engineer packages (BEP) or front-end
engineering design (FEED) are from several sources. These
can be grouped as follows:
•• Project, operating and production parameters, which
are set by the client. These are based on financial and
organisation parameters. These are usually agreed
and fixed, and form part of the design basis, such as
operating schedule and availabilities.
•• Test work (by laboratories, vendors, and specialists),
resource models and consultants’ reports. These are
the main sources of data for the project as they contain
detailed test work and analyses for the resource, its
geology, mine plan, all unit processes, the water quality
and environmental controls. This body of work also
contains reports from client’s specialists, processing
specialists, other consultants and third-party audits.
•• Operating practice and industry standards. Sources
of data are in-house experience, operating history
on similar treatment plants, published or confirmed
operating information.
•• Public domain information such as vendor data,
handbooks, regulatory and environmental standards,
industry codes, service and performance catalogues
on like-for-like applications, published engineering or
discipline text books, and published regulations.
•• Engineers’ database, operating and commissioning
experience. This would apply in instances, typically as
follows: if data from previous sources are not available;
if the number of test samples and testing is considered
limited and/or incomplete; if the test work does not
replicate the unit process performance well enough; if
test data is from batch operation and needs adjustment
for a continuous basis; if scale-up factors need to be
allowed for based on previous experience.
As a guideline, the hierarchy of sources of information and
data for process and plant design is usually in the order listed
(above). The client parameters (first group) normally form
the financial, operational and production basis for the project.
Test work data (second group) are ore and deposit specific
and consequently form the prime source of information for
flow sheet development and unit process design. Data from
the remaining three groups of information would be used in
preference where these could be demonstrated to be more
applicable and more representative than those in the first
two groups.
Design value
The design value is the input value used in calculations to
size and select the processing equipment item (eg crusher,
grinding mill, hydrocyclone, thickener, float cell, filter) or a
unit process (eg carbon adsorption circuit, counter current
wash circuit). The value can be a grade or an ore/mineral
characteristic, an attainable unit parameter, a rate or capacity
for the nominated duty and operating conditions for a unit of
process equipment or a specific circuit.
The design value does not necessarily relate to production
schedules or integrate to a mass balance. The mass balance for
the plant represents an operating condition for a continuous
circuit in steady state in which inputs equal outputs for all
items (ie solids, water, metals and elements). Plant mass
balances are prepared for an average ore condition and for
a series of cases which represent specific feed parameters or
operating conditions.
6
The design value does not include factors such as catchup capacity, any additional design margins, factors or other
allowances. Design factors and margins should be assessed
and identified in the calculation where necessary to achieve
the design objective, specific equipment duty or operating
condition. Factors include allowances for items such as wear,
screen pegging, minor surges, emergency relief, certain startup or shutdown conditions or short-circuiting. Operating
margins are allowed for materials handling and in selected
process equipment such as conveyors, cyclone feed pumps,
other slurry pumps, reagent pumps, water pumps and air
services.
Equipment sizing and selection
Calculations for equipment sizing show the design value, the
input flow(s) and additional factors where these materially
affect the equipment size and basis of selection. The input
flow(s) for the calculation are normally from a nominated
(usually at, or near maximum) operating case or from a shortterm event which may include a step-change due to a batch
flow stream addition. The calculation using the average mass
balance condition is also normally carried out to provide a
measure of the range of operating performance for the unit.
Selection and recommendation of the unit process
equipment or plant should meet two main criteria:
1. function efficiently in accordance with the parameters
and specifications nominated
2. be cost-effective from both initial capital and life cycle
assessments.
Selecting the unit process equipment or plant achieves the
desired balance between the competing demands of:
•• undersizing equipment to minimise costs, which may
increase the risk of underperformance, and
•• oversizing equipment which, whilst providing a
measure of confidence to meet the operating duty, not
only over-commits initial capital for the equipment item
(or unit process) but also can increase costs for concrete
works and steel structure to install this item.
CASE STUDIES
Case studies for selection of design values for key process
criteria and for the design basis of process areas in gold plants
are described in the following sections.
Selection of design values
The selection of a design value from their population of data
is described for comminution, gravity and feed grade cases.
An example of variability data analysis for gold recovery
with sulfur is given. Data sets are from different projects and
provide examples of the basis used to select a value for design.
Comminution – competency
The drop weight test parameters ‘A’ and ‘b’ are measures of
the competency of an ore used in grinding calculations for
semi-autogenous and autogenous grinding mills. Figure 2
shows the distributions of Axb values with samples tested for
three ore types in a supergene/hypogene type orebody with
skarn mineralisation.
The Axb value selected for design from the distribution
plots for each ore type was based on the 25th percentile in
each ore type (lower Axb values represent more competent
ore). The design value for each ore type and the characteristic
blend of ore types in the mine plan are shown in Table 2.
we are metallurgists, not magicians
‘We’re metallurgists, not magicians!’
FIG 2 – Correlation Axb values with samples tested for three ore types.
Table 2
Distribution of ore types by the principal mine plan blends.
Lithology
Axb
(ore
type)
Work index
ball
kWh/t
Mine plan
(wt %)
years 1–3
Mine plan
(wt %)
years 4–8
Type 1
77
12.8
50
0
Type 2
76
11.5
20
20
Type 3
38
15.9
30
80
Total
--
--
100
100
The (weighted) Axb design value for each mine plan blend
is 65 for the 1–3 year period and 46 for the 4–8 year period. The
corresponding Bond ball mill work indices are 13.5 kWh/t
and 15.0 kWh/t, respectively.
The ore blend is less competent (higher Axb value) and
less hard (lower work index) in the initial three years, and
consequently the unit power requirements for the semiautogenous grinding (SAG) mill and ball mill are lower
during that period compared to the latter five years. Operating
options allow either:
1. ability to treat higher tonnages during the initial years
based on the installed power and nominated throughput
for years 4–8
2. expansion of facilities and increase grinding capacity to
maintain the nominated throughput for years 4–8.
In this project, option (1) was determined to provide more
value to the project.
FIG 3 – Bond ball mill and rod mill work indices (WI)
for samples from the dominant type.
Gravity recovery
Gravity recovery tests on ground samples (5 to 10 kg
per sample) using laboratory scale centrifugal bowls
together with shaking tables or vanners measure the gravity
recoverable gold from steams in a ball mill circuit closed
with cyclones. The grind size for these types of gravity tests
is usually at the design P80 grind size, or it may be up to
two to three times this value which recognises that the bleed
stream from the grinding circuit is coarser than the cyclone
overflow final product size. The weight recovered to the final
gravity concentrate in this type of test is normally in the range
0.03–0.09 per cent by weight from feed. Figure 4 shows the
results from standard gravity recovery tests on three ore types
with auriferous quartz vein stockwork mineralisation.
Gravity recovery was moderately high and variable; no
trend was observed by ore type or by feed grade. Recoveries
at the 90th percentile, 15th percentile and average were
assessed. The laboratory recovery at the 90th percentile was
49 per cent; this was rounded up to 50 per cent for design
of metal recovery in downstream processing of the gravity
concentrate.
Gravity recoveries for the average and 15th percentile
from the laboratory tests were 39 per cent and 29 per cent,
respectively. These results from laboratory tests, which
Comminution – hardness
Standard Bond work index rod mill and ball mill tests are
measures of the hardness of an ore and are used in grinding
calculations for rod and ball mills; they also form an integral
component of SAG mill grinding calculations. Figure 3 shows
the distribution of work indices with samples tested for a
volcanic, volcaniclastic, co-magmatic intrusive type orebody.
The work index values selected for design from the
distribution plots, based on the 75th percentile, were rod
mill work index 28.9 kWh/t and ball mill work index
21.9 kWh/t. These values provide sufficient flexibility for
the mill to accommodate the majority of ore types and ore
feed conditions (with blending). Work index values at higher
percentiles would add capital cost and an additional margin
over the required feed rate case.
we are metallurgists, not magicians
FIG 4 – Gravity recoverable gold from three ore
types in quartz-veined gold deposit.
7
E McLean
were carried out in ideal, batch conditions, were discounted
to account for the lower mass pull by weight per cent to
concentrate in the plant, the type of gravity circuit installed,
the recovery effort, continuous operation and ore variability
in the plant. The discount factor for low-sulfide and low
specific gravity gangue is typically in the range 0.60 to 0.80.
The adjusted average gravity recovery for the mass balance
was 30 per cent. For downstream leaching, carbon-in-leach
(CIL) and carbon in pulp, the carbon circuit and metal
recovery were designed for periods of lower gravity recovery
and a gravity recovery of 20 per cent was adopted.
The silver in feed and to gravity was tracked to assess its
deportment and accountability. As in many gold and gold/
silver ores, silver recovery to the gravity concentrate was very
low, and significantly less than that for gold.
Front-end engineering design grades
The cumulative distribution of copper feed grades from
the mine plan and ore schedule for a long-life copper-gold
porphyry mine is shown in Figure 5.
The copper feed grade distribution in years 4–12 was
selected as the basis for design for the flotation and concentrate
handling areas. The 75th percentile was used for the design
copper value. Although gold grades in this copper-gold ore
were higher in the initial three years than in the year 4–12
operating period, gold has an economic contribution but no
material influence on the sizing of the copper concentrator
equipment.
Although the copper feed grade was lower in the initial
three years, it was preferred that higher unit capacity was
installed at the outset to avoid future disruptions to ongoing
operations from construction of additional flotation and
concentrate handling facilities. Additional capacity was
therefore available in the initial years of operation for
catch-up and ability to cope with any short-term grade or
mineralisation fluctuations.
Gold recovery variability
An apparent scatter of gold recovery results from a number of
routine, standard grind and cyanide leach tests on the gravity
tail for a large open pit gold resource, was resolved (Smith,
2005) by assessing the response by sulfur grade increments.
The leach residue versus leach feed grade trends for four
sulfur increments were plotted, shown in Figure 6. Each
sulfur grade represented approximately equal proportions of
the sulfur distribution in the mine model for the ore reserves.
FIG 6 – Gold residue and leach feed grade relationship by sulfur increments.
A relationship of residue gold grade with feed grade for
sulfur grade levels was obtained. Further statistical analysis
of this database was required to develop a model which
estimated the leach residue gold grade for the range of sulfur
grades expected in the ore schedule, and thus to calculate
gold recovery from the corresponding feed gold grade.
Design basis
A case study for the design basis for areas in a conventional
gold plant is described. The circuit comprised the following
key process stages: crushing, semi-autogenous and ball mill
(grinding) with pebble crusher (SABC), gravity, leach/CIL,
thickening, cyanide detoxification, desorption, regeneration,
gold room cathode, gravity preparation and smelting.
Plant feed characteristics
FEED distributions by tonnage and by feed grades for ore to
the mill, sourced from three open pits are shown in Table 3.
The yearly intervals correspond to optimised pit development
to maximise gold and silver grades, as well as revenue. The
majority of the ore comes from Pit 1, with ore from Pit 2
boosting gold and silver grades in the first two years. Pit 3
is the lowest grade and is developed towards the end of the
project life.
Characterisation – comminution and recovery
Ores from all pits were characterised by lithology and
alteration types and tested for their comminution competency,
hardness and abrasion properties. Three categories were
identified and grouped as ‘hard’, ‘medium’, ‘soft’ according to
unconfined compressive strength (UCS), Axb, rod mill work
index and ball mill work index values. Ore type distribution
in the mill feed tonnages by period is shown in Table 4. The
corresponding recoveries of gold and silver, based on the
characteristic gravity and leach/CIL recovery performance
Table 3
Mill feed from open pits – distribution by weight and average grades.
FIG 5 – Copper cumulative distribution from mine
and ore schedule (LOM: life-of-mine).
8
Year
Pit 1
(wt %)
Pit 2
(wt %)
Pit 3
(wt %)
Total
(wt %)
Grade Grade
Au (g/t) Ag (g/t)
1–2
80
20
0
100
2.2
20
3–7
95
5
0
100
1.6
9
8–9
60
15
25
100
1.6
6
10–12
0
20
80
100
1.1
2
we are metallurgists, not magicians
‘We’re metallurgists, not magicians!’
Table 4
Distribution ore competency/hardness and recoveries by year.
Year
Hard (%)
Medium (%)
Soft (%)
Total (%)
Recovery Ag (%)
Au-equivalent index
1–2
20
60
20
100
Throughput index Recovery Au (%)
0.90
88
60
1.35
3–7
10
60
30
100
1.00
84
50
1.00
8–9
10
40
50
100
1.05
84
50
1.03
10–12
0
5
95
100
1.25
80
40
0.78
by ore type for the blend of ores in each mill feed category, are
shown in this table. Comparative throughput and production
performances are expressed as a throughput index and a
gold-equivalent index (converts silver value to equivalent
gold) in the table.
Grinding
The ore hardness variability in the mine resource block model
showed that a significant number of blocks in the first two
years of operation had a predicted work index above those
expected for the following five to seven years. The design
values for comminution were based on treatment of ores
for years 3–7 in the schedule which comprised nearly threequarters of all ore scheduled for the initial seven years. The
comminution circuit power and mill sizes were selected based
on comminution parameters for this period and a throughput
index of 1.0 was assigned. Although the average throughput
during this initial two-year period was 90 per cent of the design
throughput based on the installed power of the selected mills,
production for the first two years on an equivalent recovered
gold basis (Table 4) was 35 per cent higher than at the design
throughput due to the higher grades and higher recoveries
with the harder ore blend.
This strategy avoided oversizing and over-capitalising the
grinding circuit and underutilisation of grinding power in
subsequent years at the design throughput rate. This also
prevented a flow-on of higher capital costs if the downstream
process plant and equipment were increased in size to match
the grinding circuit throughput.
Gravity
Gravity recovery tests showed that gold recovery was
moderately low and highly variable for all ore types. Gold
recovery was about 20 per cent for the lower grade main pits
and approximately 10 per cent for ores from the (smaller) highgrade pit. The impact of gold deportment to gravity ahead of
the leach/CIL circuit was relatively minor, particularly in the
initial years treating the higher grade ore.
For design of the leach/CIL circuit, it was assumed that the
leach circuit would treat all gold in feed, that is, no gravity
gold recovery. This ensured that all gold was recovered in
leaching and adsorption when the gravity circuit was not
operating and also provided a small operating margin for
variability in gold grades to leaching.
Silver gravity recovery was consistently very low,
typically less than five per cent to a final gravity concentrate
in all tests on various ore types. For design purposes, it was
assumed that all silver in the feed was available for leaching
and carbon adsorption.
Precious metal recovery
A carbon adsorption and desorption circuit which met
the duty required for the initial two years was selected for
this operation. Although the throughput rate was lower,
ore grades and recoveries were highest in this period and
we are metallurgists, not magicians
thus directly affected solution and carbon inventories. The
maximum gold and silver inventories, and the resulting
carbon transfer, desorption and regeneration systems to
manage these metal inventories occurred during the initial
two years of operation (Table 4). Consequently, this period
was the basis for the design of the carbon adsorption, elution
and regeneration circuits.
As the soluble ratio of silver to gold during the initial two
years was about 6:1, silver management was the main design
consideration and determinant for the carbon circuit design
and operating strategy.
Gold room
Design of the gold room and smelting activities was based on
treating peak weekly inventories of gravity recoverable gold
and electrowinning cathode clean-up.
Volumetric capacity
The amount of ore classified as ‘soft’ increased with mine
life and is the major to dominant proportion of ore feed to
the mill towards the end of operations, in years 8–12. During
this period the grinding circuit was able to treat above-design
throughputs and most likely up to treating a 25 per cent
higher feed rate. This helped to offset the loss in production
due to falling grades and decrease in recovery.
Slurry viscosity and rheology tests on all ore types,
particularly the characteristic moderate- to high-alteration
ores in the soft category, provided comparative information on
pulp density, viscosity, slurry transportation properties and
hydraulic gradients in the circuit. The slurry characteristics
and flow properties directly affected duty and specifications
for equipment such as slurry pumps, trash and safety
screens, launders, intertank screens and the thickener. Hence,
provision was made in the initial design for equipment, pipes
and launders to handle larger flows for periods when poor or
adverse slurry properties dictated, or allowances were made
to subsequently add to or upgrade existing equipment.
CONCLUSIONS
Data obtained from metallurgical and physical ore testing
programs is appropriate and applicable to the extent that the
sample selection is representative and correct preparation
protocols are practiced.
The recommended methodology for sample selection as a
basis for flow sheet development and process design is by
characterisation and variability categories. Characterisation
accounts for metallurgical and physical properties of the ore
by domain, zone, ore type or specific geological classification.
Variability encompasses subsets of domains such as
mineralisation, range of grades of economic metal, range of
grades of deleterious or penalty elements, ratio of grades of
metals, spatial location along strike and at depth.
Practical and robust interrogation, interpretation and
analysis of the complex database of information and detailed
9
E McLean
data available are fundamental to successful design. This
is not possible without an understanding and an iterative
assessment in conjunction with the mine planning group
of ore feed schedules to the plant by tonnage, grade,
mineralisation and lithology; of ore exploitation sequences,
mining strategy, mine development plan, pit financial
modelling and resource constraints.
The selection of the appropriate design value for each unit
process or process stage is critical. Coupled with relevant
design margins, calculation factors, scale-up, benchmarking,
operating experience and input flow conditions, the unit
process equipment or plant selected should meet two
main criteria: to function efficiently in accordance with the
parameters and specifications nominated; and to be costeffective from both initial capital and life cycle assessments.
The engineer, scientist, or metallurgist who can make
apparently impractical things happen and is able to
produce outcomes that achieve objectives shows initiative,
is innovative, creative and resourceful. The inexplicable or
impossible events outside these parameters are magic.
10
ACKNOWLEDGEMENTS
The inspiration for this paper comes from beleaguered
colleagues, many of whom have been in the firing line from
project managers and production superintendents to explain
‘why’, ‘why not’ and ‘when’.
REFERENCES
Agarwal, J C, Brown, S R and Katrak, S E, 1983. Taking the sting out of
out of project start-up problems, adapted from a presentation at
The American Mining Congress, September 1983, San Francisco.
Lane, G, Staples, P, Dickie, M and Fleay, J, 2008. Engineering design
of concentrators in Australia, Asia and Africa – what drives the
capital cost?, in Proceedings Procemin 2008 V International Mineral
Processing Seminar (ed: R Kuyvenhoven, C Gómez and A Casali),
pp 30–38 (GECAMIN: Santiago).
Performance Associates International Inc, 2017. So you are investing
in a mining project – what usually goes wrong [online]. Available
from: <http://www.perfnet.com/wp/white-papers/investingmining-project-usually-goes-wrong/> [Accessed: 20 July 2017].
we are metallurgists, not magicians
Contents
Back to the future – why change
doesn’t necessarily mean progress
P D Munro1 and P A Tilyard2
ABSTRACT
There have been enormous changes in mineral processing in the past four decades.
For example, grinding mill power has increased by an order of magnitude, regrinding
is done to -10 µm and flotation machines are 100 times bigger. Operating staff have
unprecedented opportunities for online monitoring and performance control of
mineral processing plants. Sophisticated instruments can provide a plethora of data
characterising the mineralogy and surfaces of particles. Digital computers allow
complex calculations on huge amounts of data including modelling and simulation of
machine and plant performance.
However, all these changes have not necessarily led to better metallurgical results. An
analogy can be drawn with the thoroughbred racing industry in Australia. Significant
advances in scientific knowledge in animal genetics, physiology, biomechanics and
nutrition applied to the business have resulted in only a two per cent reduction in
winning times for the Melbourne Cup and Caulfield Cup since the 1920s.
A critical look at some mineral processing metrics suggests similar failures to
improve performance despite putting in more resources. In fact, certain parameters
such as operating times and plant start-up performance are considered to have
remained static or even deteriorated.
There has been an emphasis on ‘process’ at the expense of ‘outcomes’. The industry’s
strength has been in finding technical (or ‘hardware’) solutions while its weakness
has been at the people end of the business in maximising and consolidating the gains
from the technologies. Some trends in plant design over these years have exacerbated
the apparent deskilling of operating and technical staff. Despite unparalleled options
for communications, some staff are embarrassingly uninformed about technical
developments in their fields.
The ‘boom and bust’ cycles of the industry, together with trends in tertiary education
and the effects of fly-in, fly-out (FIFO) operations, raise serious questions about the
sustainability of human capital in the mineral processing sector.
This paper by two experienced mineral processing engineers, with contributions
from other senior practitioners, reviews these trends. While there may be an element
of ‘the older we are, the better we were’, it is an attempt to identify the issues and
propose solutions.
INTRODUCTION
The terms ‘mineral processing engineer’, ‘metallurgist’ and ‘graduate’ are used
interchangeably in this paper.
Most observations and examples have been drawn from base metals sulfide
concentrators using flotation as the separation method. However, the authors
have enough experience with other operations, such as gold leaching and iron ore
processing, to expect that these have similar issues.
The authors have attempted to contrast the mineral processing sector that they entered
as new graduates at the beginning of the 1970s with the current situation in 2009.
While there may be an element of ‘the older we are, the better we were’ in this
paper, it is an attempt to identify the issues and propose some solutions.
1. FAusIMM, Senior Principal Consulting
Engineer, Mineralurgy Pty Ltd, Taringa Qld
4068. Email: pdmunro@bigpond.com.au
2. FAusIMM(CP), Former Group
Metallurgist, MMG (now retired).
Email: tilyards@bigpond.net.au
AROUND 40 YEARS AGO
An industry snapshot in 1970:
•• The typical starting salary for a graduate mineral processing engineer was
~$3500/a; accounting for overhead makes the cost to the employer ~$5000/a,
which after six weeks’ annual leave and 8 h/d = $2.72/h cost to the employer.
Using the Australian Consumer Price Index as a multiplier gives a cost in
11
P D Munro and P A Tilyard
2009 of $26/h. Another way of looking at this is that
at 1970 metal prices of US$1300/t for copper, US$35/t
oz for gold, US$350/t for lead, US$2750/t for nickel,
US$1.88/t oz for silver and US$300/t for zinc with
the exchange rate of A$1 = US$1.12, the graduate’s
annual salary including overhead had the following
approximate metal equivalents:
•• 4.3 t of copper
•• 161 t oz of gold
•• 16 t of lead
•• 2 t of nickel
•• 2988 t oz of silver
•• 18.7 t of zinc.
•• Employment conditions relative to Australian norms
were generous and encouraged young people to seek a
professional career in the industry; McCarthy (2006) has
commented on this issue.
•• Fly-in fly-out (FIFO) did not exist. People lived in mining
towns where socialising and talking ‘shop’ gave you a
good appreciation of other disciplines such as mining
engineering and geology.
•• Joining the Australasian Institute of Mining and
Metallurgy (the AusIMM) was almost a condition of
employment and contributing to local branch activities
was expected.
•• There were no personal computers.
•• The internet did not exist, with no time consumed
reading emails!
•• Telephone calls were expensive and making an STD
(subscriber trunk dialling) call required permission.
•• Fax machines did not exist.
•• Copying was changing over from wet process
duplication to ‘Xeroxing’.
•• Process control computers had just arrived with 4–8 KB
of memory, programmed in Assembler.
•• 4 KB of memory cost $4000 (these are 1970 dollars).
•• There were a few electronic calculators around. One of the
paper’s authors was hugely impressed as a new graduate
to find out that the mill clerk at Mount Isa Mines Limited
(MIM) had one with a square root function!
•• Companies had on-site technical capabilities with the
larger ones such as CRA and MIM doing world-class
research.
•• Information was provided through a company technical
library. This housed all significant reports and circulated
journals. Graduates were expected to read the technical
literature in their professions.
•• The assay function had not been ‘outsourced’ and the
company’s chief chemist was a source of wise counsel on
all analytical and chemical matters.
•• Large complex sulfide flotation concentrators could
be competently run by three metallurgists: control
metallurgist, metallurgist control plus one graduate. The
mill clerk adequately coped with most data collecting
and reporting functions.
•• Flotation was controlled using the vanning plaque (or
‘pan’) confirmed by wet chemical assays every two
hours. This was a skill that the metallurgist had to
master to acquire any credibility with the operators. The
‘pan’ had the advantage of providing real-time semiquantitative mineralogy as well as an estimated ‘assay’.
12
•• The design of flotation plants brought the operators and
metallurgists close to the froth allowing observation
and giving an excellent ‘feel’ for the process. As the
American baseball commentator Yogi Berra once said
‘you can see a lot by looking’.
•• Comminution was done by multistage crushing with
rod milling plus ball milling giving stable grinding
throughput so the metallurgical ‘narrative’ focused on
separation performance.
•• ‘Running time’ was the only thing that mattered, with
both maintenance and operational groups zealously
guarding their reputation on minimising downtime.
There was significant focus at a high level on areas for
improvement.
•• The flotation section was operated under a strict
‘theory x’ (command and control) set of guidelines with
operators allowed to adjust parameters (such as air, froth
depth, reagent additions etc) only within limits set by the
metallurgist. This reflected the high level of accountability
the metallurgist had for separation performance.
•• Prototype on-stream analysers (OSA) started appearing
in flotation plants.
•• Quantitative mineralogical data were collected by
manual point counting. It was not uncommon for a site
to have a mineralogical laboratory.
•• Computer models of mineral processes were empirical
and run on main frame machines, programmed on
punched cards. Conclusions from the models were
tentative because of their novelty.
•• Metallurgists routinely:
•• were held accountable for metallurgical performance
daily and subjected to relentless queries; this may not
be in accord with current ‘warm and fuzzy’ human
relations approaches but it certainly provided focus
and was character-building – the authors observed
that those metallurgists who survived and thrived
in this trial by ordeal had the potential to become
competent plant managers
•• had to be very sceptical about the veracity of instrument
readings as much of the technology was in its infancy.
Pneumatic control with 3–15 psi air was the norm
•• checked crusher gaps with the ‘leads’
•• checked mill power draw
•• checked hydrocyclone spigots with callipers putting in
a maintenance work request when the wear exceeded
6 mm (¼ inch)
•• routinely inspected mill and flotation banks on plant
shutdowns, putting in maintenance work requests.
Decisions on equipment repair issues possibly
affecting throughput and metallurgical performance
were not abdicated to the maintenance department or
even the purchasing department
•• produced a detailed monthly metallurgical report,
commenting on performance on a sized basis
•• produced detailed cost comments – costing systems
were logically divided up into ‘cost centres’ for unit
processes such as ‘crushing’ and ‘grinding’ comprising
expense accounts (such as ‘electric power’, ‘50 mm
diameter grinding balls’ etc).
•• The monthly metal balance was done manually and
expeditiously, often by the mill clerk.
we are metallurgists, not magicians
Back to the future – why change doesn’t necessarily mean progress
•• Meetings were relatively infrequent and brief. Senior
staff generally exercised strict control to ensure they did
not become ‘talk fests’.
•• Despite the lack of the internet and without cheap
telephone calls, staff at remote sites were well aware of
technical developments in other locations. Publications
such as Jim Woodcock’s annual review Mineral
Processing in Australia (Woodcock, 1978) were eagerly
anticipated to provide details on operating practices.
Similar enthusiasm greeted other review articles such
those in the Mining Annual Review and World Mining.
•• Graduates were expected to be literate and numerate.
Memoranda and reports were carefully vetted by
superiors for both their technical and English content.
Senior staff were then competent enough in their
own literary skills to identify errors in grammar and
deficiencies in expression. Rewrites of offending sections
were obligatory and frequent.
•• Reports were written in long hand and subsequently
typed by a secretary. Graphs were hand drawn in India
Ink with the aid of ‘french curves’.
•• Future ore testing looked at the ore sources the plant
would process in future years.
•• Graduate training at larger operations was structured
with around six months spent doing ‘hands on’ jobs in
the plant including supervisory positions. The next step
was project work leading to time being responsible for
metallurgical performance. There seemed to be more
emphasis on achieving competency in core skills though
such jargon terms were not then used.
•• There were more opportunities for mentoring. It was
not uncommon to find metallurgists over the age of 50
at a mine site.
IN 2009
Observations on the situation in 2009:
•• Typical salary for a second-year graduate mineral
processing engineer on a FIFO basis was ~$102 000/a;
assume 8/6 roster with 0.5 d travelling each way gives
seven effective working days on-site. Subtracting four
weeks’ holiday per annum gives 24 effective working
weeks at 12 h/d = ~$50/h cost to the employer. Applying
the same overhead as in the 1970 case gives ~$70/h. At
2009 metal prices of US$4300/t for copper, US$900/t oz
for gold, US$1350/t for lead, US$13 000/t for nickel,
US$12.50/t oz for silver and US$1500/t for zinc with the
exchange rate of A$1 = US$0.75, the graduate’s salary
including overhead had the following approximate
metal equivalents:
•• 25.3 t of copper (~6 × the 1970 value)
•• 120 t oz of gold (~0.8 × the 1970 value)
•• 80.6 t of lead (~5 × the 1970 value)
•• 8.4 t of nickel (~4 × the 1970 value)
•• 8700 t oz of silver (~3 × the 1970 value)
•• 72.5 t of zinc (~4 × the 1970 value).
•• Employment conditions relative to Australian norms
are less generous, discouraging young people to seek a
professional career in the industry (McCarthy, 2006).
•• FIFO operations are increasingly the norm.
•• People do not tend to live in mining towns and do not
socialise much after work. Even on-site the 12 hour day
leaves little time for interdisciplinary interaction. An
unkind observation is that the solution to any problem
we are metallurgists, not magicians
is never further than the next rotation out. Support staff
are seen as a cost with harangues from accountants and
managers about the ‘head count’ and cost of supporting
people in the camp but neglecting to consider the value
they add.
•• There seems to be a lower level of participation in the
activities of the AusIMM. FIFO does not encourage
this, as attending technical meetings when a person is
rostered off is often viewed as ‘work’.
•• Everyone has access to a personal computer and the
internet with most of a person’s time spent in front of it
dealing with emails is a major preoccupation.
•• Process control computers are ubiquitous with the cost
of memory and data storage still falling in accordance
with Moore’s Law.
•• Very few companies have significant technical capabilities,
and most rely on research institutes, consultants and less
frequently inhouse technical groups.
•• The technical library was generally a victim of cost
cutting in the late 1980s and 1990s. Anecdotes tell of the
few remaining professional information technologists
(ie ‘technical librarians’) falling victim to the downsizing
that occurred following the global financial crisis of
2008. Historical technical work is usually imperfectly
filed and collated, often with important omissions. The
unspoken belief appears to be that everything you need
to know can be sourced from the internet. An unkind
comment from one senior engineer was that ‘the level of
inquiry was so basic that the answer is often found on
the internet!’ Conversations with metallurgists quickly
reveal significant lacunae in their knowledge. Nothing
more technical than the introductory textbook ‘Mineral
Processing Technology’ (Wills and Napier-Munn, 2006) is
found on most bookshelves. Those metallurgists with
an interest in reading the technical literature struggle
to convince senior management of the benefit of a
subscription to a data search/retrieval facility.
•• The assay function has been ‘outsourced’, sometimes to
the ludicrous point where it is no longer possible to have
any ‘spot samples’ processed! Some sites don’t even
have a rudimentary metallurgical laboratory.
•• A large complex sulfide flotation concentrator requires
six to eight metallurgists with no seeming improvement
in the quality of operation. The extra numbers are
needed because of FIFO, as support staff were culled
during the hard times and with metallurgists now often
performing clerical functions.
•• Flotation is controlled using OSA and (surprisingly)
still by two-hour assays in some plants. The current
design vogue of open air flotation plants, having the
cells packed together with walkways over the top, has
distanced the operators and metallurgists from the
process. A dissenting opinion on the ergonomics of
flotation plant design is contained in a paper on the
design of the Prominent Hill concentrator (Colbert,
Munro and Yeowart, 2009).
•• Comminution by autogenous grinding/semi-autogenous
grinding (AG/SAG) mills gives varying grinding
throughput with the ‘narrative’ focused more on the
grinding section to the neglect of evaluating separation
performance.
•• Previous simple classifications of plant downtime such
as ‘planned maintenance’, ‘unplanned maintenance/
breakdown’ and ‘operational’, ‘lack of ore, water or
13
P D Munro and P A Tilyard
power’ have become more complicated including
definitions such as ‘readiness’.
the mineral liberation and beneficiation process. It seems
to be for ornament rather than use.
•• The end result is that often no one is accountable for
actual running time, which used to be the mill manager’s
responsibility. Running times seem to have deteriorated
from values achieved in the 1970s and early 1980s.
•• It is disappointing to observe that quantitative
mineralogical data are usually one of the items discarded
as a response to low metals prices; however, this is
understandable if the metallurgists receiving such data
can’t use it effectively.
•• One nostalgic observation is that as soon as the
industry went away from fine crushing and rod and/
or ball mills the ‘rot’ set in. Availability of the ball
mills at Bougainville Copper Limited (BCL) inched up
to 99.1 per cent in 1983, utilisation of available time
was 99.7 per cent and total run time was 98.8 per cent,
equivalent to 8655 h/a. Compare this to the current
common SAG mill design run time of 8000 h/a
(91.3 per cent) and actual operating run time not much
better in many cases.
•• In 2006, one large mining company was designing
its iron ore plants for a run time of 74 per cent and
probably still is. It seems to have learned nothing from
the BCL crushing plant experience where line run time
was 96 per cent and secondary crusher run time was
91 per cent. Even experienced iron ore metallurgists
found these numbers difficult to believe. The secondary
and tertiary crushers and the tertiary screens were fed
from large surge bins which significantly contributed
to the operating efficiency of the BCL plant.
•• Flotation is now run under a laissez-faire regime where
seemingly everyone’s opinions are valued, regardless of
skill and knowledge, to avoid upsetting individuals.
•• Operators are allowed wide scope to adjust parameters
(such as air, froth depth, reagent additions etc) and
even to make circuit changes. One hears statements like
‘someone changed the collector addition last night’ at
the daily production meeting which doesn’t encourage
a culture of responsibility.
•• Democracy has been substituted for the scientific method,
eg the fact that in the 14th century most people believed
that the world was flat did not mean it was correct!
•• There doesn’t seem to be much accountability for
performance.
•• OSA and particle size monitors are supposed to be
standard equipment though there are some curious
exceptions. At some sites the OSA and/or particle size
measurement system is no longer operational through
systematic neglect or produces questionable data from
poor calibration procedures.
•• It is depressing to hear announcements trumpeting
the purchase of an OSA system or online particle
measurement as a technical breakthrough when such
items are assumed to be standard equipment in the 21st
century. The old adage ‘if you can’t measure it, you can’t
control it’ surely applies.
•• Quantitative mineralogical data can be rapidly produced
by automated X-ray methods such as QEMSCAN
(Quantitative Evaluation of Minerals by Scanning
Electron Microscopy) or mineral liberation analyser
(MLA; both, incidentally, are Australian innovations);
however, manual and automated point counting still
survive, with the practitioners of these supposedly
outdated techniques able to offer interpretation and
information as opposed to the data produced by the
automated X-ray systems.
•• Despite having the luxury of quantitative mineralogical
data few metallurgists are able to use the data to manage
14
•• Computer models of mineral processes are increasingly
phenomenological and can be run on powerful laptop
computers
•• Metallurgists less commonly:
•• are held accountable for metallurgical performance
daily and subjected to relentless queries
•• master the fundamentals of their profession, which
detracts from their later performance as plant managers
•• disbelieve instrument readings
•• check crusher gaps, mill power draw or hydrocyclone
spigot dimensions
•• routinely inspect grinding mills, hydrocyclones and
flotation banks on plant shutdowns, then submit
maintenance work requests when required
•• make decisions on equipment repair issues possibly
affecting throughput and metallurgical performance
and often abdicate these decisions to the maintenance
department
•• produce a detailed monthly metallurgical report
commenting on performance on a sized basis
•• exploit the power of the spreadsheet which is ideal
for manipulating mineral processing data to produce
a thorough understanding of the performance of the
plant according to the axiom of ‘size-by-size mineral
particle behaviour by liberation class’
•• are held to account for operating costs
•• read the technical literature of the profession.
•• The monthly metal balance is done in many cases
on expensive computer packages with no seeming
improvement in accuracy or speed. The computer can’t
tell you that the head sampler is not in its correct rest
position and is getting constant splash into it.
•• Meetings are much more frequent and last longer.
•• Despite the so-called communications revolution from
the availability of the internet, staff seem to be totally
unaware of past industry paradigms.
•• The authors have been astounded to find plants using
rubber or polyurethane spigots in hydrocyclones
taking SAG mill discharge. Hydrocyclone spigots were
supposedly ‘standardised’ to long-wearing ceramic units
during the late 1960s (Munro, Eaton and Burton, 1982).
•• Even when company intranets are set-up with a
metallurgical site and discussion page, there is a
reluctance to seek advice and/or information from the
wider pool of metallurgical expertise in the group.
•• Graduates often lack literacy and numeracy, and when
memoranda and reports are written (which isn’t often
enough in FIFO operations), they are often not well
constructed and painful to read.
•• With the exception of engineering design companies,
version control of reports and spreadsheets is almost
unheard of.
•• A recurrent theme is the disregard of the statistics
of variation as evidenced by claims of observed
we are metallurgists, not magicians
Back to the future – why change doesn’t necessarily mean progress
improvement well within the normal ‘noise band’ of
plant and laboratory performance.
•• Verbal communication is resolutely qualitative rather
than quantitative with numbers replaced by ‘larger’
versus ‘smaller’, ‘faster’ versus ‘slower’ etc.
•• It seems to be easier to get to the geologists’ office at
the mine site from head office in a capital city or from
the consultant’s office than it does from the nearby
concentrator office. ‘Future ore testing’ may have been
repackaged as ‘geometallurgy’ but you still have to talk
to the geologists.
•• Graduates spend around 25 per cent of the time doing
‘hands on’ training compared with 40 years ago.
•• Many operations supposedly have detailed programs for
the development of graduates; however, there is often a
gap between rhetoric and reality when the authors have
examined the knowledge and competencies of people
coming out of these schemes.
•• The age spread of site metallurgists is narrow, rarely
finding one over the age of 45.
•• The head office support function also appears to be
reducing with major mining companies cutting back on
experienced metallurgists in technical services functions.
•• Mentoring does not seem to be valued as much,
and while there may be ‘motherhood’ statements
about it, along with a supposed commitment to staff
development, the lack of any demonstrated commitment
to actually providing the resources for mentoring and
staff development belies this point.
IF THIS IS PROGRESS …?
Below are examples supporting the contention that mineral
processing outcomes are not uniformly improving:
•• Throughput variations as ascribed to the ore being
‘harder’ or ‘softer’. It is a long time since we heard a
metallurgist say something like ‘our current feed is
from bench ‘AB’ where the predominant rock type is
‘andesite/granite etc’; grindability data for this shows
a bond ball mill work index of ‘A’ kWh/t, drop weight
index a × b of ‘B’ etc’. Despite all the excellent tools for
data collection, manipulation and display, people seem
resolved to be qualitative rather than quantitative. It was
hoped that the Australian Minerals Industries Research
Association (AMIRA) P843 geometallurgy project
would change this, but it seemed to be driven more by
geologists than metallurgists when it ran its course.
•• In 2004, a large copper producer in a foreign country
asked a respected Australian research institute to
investigate why there was a difference in metallurgical
performance between its new plant with one type of
flotation machine for rougher duty and the older plant
which had another machine. After a detailed campaign
including measuring parameters such as superficial
gas velocity and bubble surface area flux it was found
that the rougher flotation cells in the newer plant had
significantly lower effective volumes because of the
accumulation of tramp oversized ore particles from
‘upsets’ in the grinding section. After the cells were
cleaned out copper recovery increased from 83 per cent
to the target of 90 per cent. It is a telling commentary on
how removed people now are from the flotation process
that someone has to come from another country to tell
you that your cells are full of rocks!
we are metallurgists, not magicians
•• In early 2006, a large copper concentrator did a thorough
maintenance overhaul of its flotation cells including
replacing worn impellers and stators, refurbishing level
control equipment and cleaning cell lips. On restarting
the plant the sulfide grade of the concentrate increased
from 78 per cent to 82 per cent with 2.5 per cent abs
higher copper recovery. Such a significant deterioration
in the condition of the flotation cells would and would
have been noticed much earlier if the metallurgists had
looked at the froth surface on the cells. One of the authors
had a similar experience in a copper zinc concentrator
where he suggested that the flotation cells be drained
and inspected on the next shutdown. The appalling state
of the impellers and stators (some had ceased to exist)
showed that no interest had been taken in the flotation
cells by metallurgical staff for years.
•• Daily data for mill products of ‘silver nitrate soluble
copper’, ‘acetate soluble copper’, ‘cyanide soluble
copper’ and ‘total copper’ were not used to explain
the performance of a large copper-gold concentrator.
Management’s concern and displeasure at high rougher
tailings losses could have been mollified by pointing out
that the proportion of acid soluble copper in the feed
had increased. The copper minerals in this category are
not recoverable by conventional sulfide copper flotation.
•• At a review meeting examining design options for a
major upgrade of a venerable concentrator, none of the
senior operating staff present knew the current plant
operating cost in $/t of ore treated or its components
according to activity.
•• At an overseas copper concentrator, a cost saving
initiative was suggested by a group of non-technical
employees assigned the task of improving plant
performance and was accepted by the metallurgical staff.
The instrument air compressor was decommissioned.
In due course the wet air from the plant compressors
destroyed the level controllers in the flotation section.
•• Two very large overseas copper concentrators do not
own a cyclosiser and the metallurgical staff stoutly
defend the fact that they have no data in the -37 µ size
region.
•• At an overseas copper concentrator, a graduate
metallurgist disputed the visiting consultant’s opinion
that, in the absence of OSA (the equipment had fallen
into disuse), two hourly spot assays would be useful.
•• Very few operations are able to present in a single
document, the performance of the plant over the past
say five to ten years.
MINERAL PROCESSING BASICS
The fundamental data requirements for managing a mineral
processing plant are the following:
•• The target metallurgical performance (eg concentrates
grade, metals recoveries, what you want to reject and
the priority for these targets).
•• An equipment list.
•• A mass balances for solids, elements, minerals and water
•• A simulation model of the comminution circuit with
grindability data for current ores.
•• Quantification of ore types (both grinding and
metallurgical performance) in the plant feed for future
years and relation to past performance.
•• Knowledge of the processing characteristics of the
orebody in a spatial sense. This should include
15
P D Munro and P A Tilyard
grindability, quantitative mineralogy, metallurgical
performance, distribution of precious and impurity
elements and minerals (eg Au, Ag, As, F, Hg, organic
carbon, talc etc) where applicable.
•• Grinding and regrinding mill power consumptions.
•• Element and mineral particle behaviour on a sized basis.
•• Mineral particle behaviour on a sized basis by liberation
class on a monthly basis.
•• Trade-off between concentrate grade versus recovery
and plant throughput versus separation section
(eg flotation) feed sizing.
•• Effect of concentrate grade and recovery on net smelter
returns (NSR).
•• Detailed chemical analysis of concentrates and concentrate
physical data such as transportable moisture limit.
•• A history of the metal balance/metal accounting with
accompanying narrative on all adjustments made
to the ‘first pass’ numbers. This should also include
reconciliation of data from dispatches and receipts of
products to customers.
•• Regular analyses of process and effluent waters.
•• Consumables usage, eg grinding media (also reported
as weight loss per kWh of mill power), reagents and
including water.
•• Operating cost data:
•• on Pareto graphs
•• by activity (eg crushing, grinding etc)
•• by expense account.
•• A simple financial model of the plant which incorporates
the technical drivers and their effect on NSR.
•• A metallurgical development plan which incorporates
the above trade-offs and outcome drivers etc.
•• An enthusiasm for capturing plant data on an ongoing
basis. Unfortunately, a compendium of the above data is
rarely seen when visiting a plant.
It is interesting to note that concentrator staff are likely
to have compiled job safety analysis documents for many
physical tasks in the plant but have minimal cogent
documentation on how to do the metal balance, production
forecasts or metallurgical planning.
GEOLOGISTS AS EXEMPLARS FOR DATA AND INFORMATION
We should take a lesson in data collection and information
management from our geological compatriots.
Mineral deposits can be discovered and mines can operate
for decades going through multiple ownerships with the
geological database remaining intact and up to date. You
hardly ever encounter situations where there isn’t any data for
a drill hole because the assay sheet was shredded five years
ago when the project geologists were ‘outplaced’ during the
last cut-backs or the samples weren’t analysed to save money.
Chief geologists are not in the habit of throwing away drill
core to eliminate the cost of storing it.
Contrast this to the metallurgical test reports on those same
geological samples. You would be an optimist to expect that
they would still be found in the concentrator department.
Essential fittings for many metallurgists’ offices are a plethora
of seemingly important test work and survey reports covered
in grime and the circular marks of beverage containers
haphazardly scattered over flat surfaces. Electronic copies of
reports when archived on computer servers have cryptic and
sometimes unintelligible titles and are not filed systematically
16
making retrieval a laborious and daunting undertaking.
Electronic files of plant operating parameters from process
measurement and control systems frequently have long periods
of missing or erroneous data. Sometimes complete archives of
metallurgical reports and plant data have gone missing when
a computer system was upgraded. Are mineral processing
engineers condemned to be ‘children of lesser data’?
THOROUGHBRED RACING Versus
DENTISTRY – A CAUTIONARY TALE
Quirk (2006) made some trenchant observations on the
performance of the thoroughbred racing industry. Winning
times for the Melbourne Cup, which is an open handicap
event and not a race of equals, have improved only
three per cent since the 1920s. Similarly, the W S Cox Plate,
which is weighted for age and supposedly a classic race for
the best horses, has shown only a two per cent improvement
over the same time period. These data are shown in Figure 1.
Similarly, analyses in the 1980s of the classic English races
the St Leger, the Derby and the Oaks showed that little or no
improvement in winning times had occurred in the previous
70 years.
By comparison, human athletes in the Olympic Games over
the same time period had an eight per cent improvement in
winning times for the 100 m, ten per cent for the 1500 m and
12 per cent for the 5000 m.
So, despite having an array of modern technologies in genetics,
nutrition, biomechanics, equine physiology etc available, the
thoroughbred racing industry, as measured by winning times,
has not improved its performance. The reason given for this is
the limited equine gene pool available for development.
The authors are not currently suggesting that the gene pool
of metallurgists is the cause of our perceived dissatisfaction
with the performance of mineral processing professionals;
however, the cautionary message from the thoroughbred
racing industry is that merely applying modern scientific
techniques does not ensure the desired outcome.
This is sharply contrasted to advances in modern dentistry
over the same time period where technology and innovation
have clearly improved the patient’s physical (if not his/her
fiscal) well-being (O’Rourke, 1995).
THE CONTEMPORARY MINERAL PROCESSING SITUATION
There have been significant improvements in mineral
processing technologies over the past four decades as per the
following examples:
•• Machines have become much larger, eg 20 MW SAG mill
compared with a 3 MW (max) ball mill. This has allowed
FIG 1 – Melbourne Cup and Cox Plate winning times (ten year averages).
we are metallurgists, not magicians
Back to the future – why change doesn’t necessarily mean progress
a single unit to replace a multiplicity of very small ones
for both comminution and separation duties.
•• Capital efficient AG and SAG mills instead of crushing
plants plus rod mills plus ball mills for comminution
(though some would argue this has been at the expense
of rapid plant start-up, fluctuating metallurgical
performance and increased downtime).
•• High-pressure grinding rolls (HPGR).
•• Comminution technologies that can economically grind
down to -10 µ.
•• Sorting technologies using sensing of multiple mineral
characteristics.
•• Very high field strength magnetic separators.
•• ‘High g’ gravity separators.
•• Carbon-in-pulp (CIP) and carbon-in-leach (CIL)
replacing solid liquid separation plus Merrill-Crowe
process for gold extraction.
•• Pressure oxidation and bioleaching for refractory gold
ores.
•• Solvent extraction plus electrowinning for copper leach
pregnant leach solution.
•• Selective flocculation.
•• OSA; online measurement of both coarse and fine
particle size distributions; online measurement of
chemical parameters such as redox potential, cyanide
concentration, dissolved oxygen level etc.
•• Expert control systems.
•• Ability to collect detailed quantitative mineralogical
data including QEMSCAN plus MLA, infra-red (IR) for
alteration etc.
•• Proven models for the simulation and control of mineral
comminution and to a lesser extent separation processes.
•• Computer aided design (CAD) for 3D visualisation
of plant layouts with accompanying efficiencies in
fabrication and plant construction.
•• The capacity to collect, analyse, manipulate, store and
retrieve data plus information on a scale unimaginable
in 1969–1970.
We should be wary about confusing:
•• movement with action
•• activity with achievement
•• change with progress.
The contention is that all this extra hardware and software is
not necessarily giving uniformly better performances across
the mineral processing sector.
There is some truth in the retorts to the above complaints
that:
•• some of the orebodies currently being treated are more
refractory than those processed 40 years ago
•• metallurgists are relatively more expensive to employ
than they were in 1969–1970 for most commodity
producers.
The most significant contributor to this is the dominance
of electronic information technology in the working life
of a metallurgist. This area, which without a doubt has
experienced the greatest advances over the last 40 years, is the
prime cause of process overcoming outcome.
MANAGEMENT RESPONSE
If management fails to comprehend and master the technical
basics of plant operation it will not understand the outcome
‘drivers’ and be less capable of directing the available
metallurgical talent pool.
The authors rarely encounter a coherent plan of metallurgical
development for a mineral processing plant with the most
precious resource on-site of ‘Competent Person hours’
appropriately focused on the most productive opportunities.
Management must recognise
metallurgist’s time from:
this
‘black
•• the ‘bureaucratisation’ of
overwhelming ‘outcome’
work
with
hole’
of
‘process’
•• producing abundant data but a paucity of information.
What value does a metallurgist add to the understanding of
a mineral processing operation by spending three hours per
day producing the daily report with a plethora of egregious
irrelevant details? Surely this is not an intellectually satisfying
outcome after spending four years at university?
Why such atrocious wastes of professional time occur in this
supposed ‘information age’ and are tolerated, considering the
higher ‘metal equivalents’ of metallurgist’s time, are questions
for serious reflection.
Corrective actions must include:
•• Putting the ‘mineral processing basics’ in place.
•• Freeing up metallurgists’ time; eg automatic collection
and processing of online measurements and hiring
sufficient clerical and support staff. With metallurgists
being relatively more expensive to employ than 40 years
ago, there is even less excuse to waste their time doing
jobs that can be farmed out to others.
•• Focusing the available professional time on a strictly
prioritised list of metallurgical opportunities with
quantifiable outcomes.
Further development of the ‘Mineral Processing Toolbox’
section of the AusIMM website to include examples of good
practice for some of the mineral processing basics cited, may
be an appropriate method to focus on the real issues driving
mineral processing outcomes.
Such a web-based facilitation mechanism could lift industry
performance, especially if companies would be prepared to
contribute suitable examples in a collaborative spirit, similar
to an AMIRA project.
CONCLUSIONS
By some criteria, the standard of operation of mineral
processing plants has not significantly improved in the last
40 years despite the ‘information technology revolution’.
Yet instances of poor metallurgical outcomes seem to be
too frequent and too serious to be classed as aberrations. The
suspicion is that the industry has some systemic problems
that have to be addressed.
Professional time for process monitoring and improvement
has been eroded by the ‘bureaucratisation’ of the metallurgist’s
work, despite the fact that, for most commodity producers,
process engineers are now relatively more expensive to
employ than in 1969–1970.
The authors contend that the decline in the productive
output of metallurgists is the erosion of the limited time
available to do professional work.
Metallurgists need to get back to the basics of their profession
to produce outcomes rather than being overwhelmed
by processes. There is a risk that information technology
we are metallurgists, not magicians
17
P D Munro and P A Tilyard
can ‘enslave’ rather than ‘liberate’ with data displacing
fundamental mineral processing information.
The concentrator manager has to take the leading role in
turning around the current unsatisfactory state of affairs.
ACKNOWLEDGEMENTS
The authors thank Mineralurgy Pty Ltd and MMG Ltd for
permission to publish this paper.
The following individuals are thanked for their observations
on these matters over the years: Gary Chilman, John Glen,
Greg Lane, Rolly Nice, Joe Pease, Geoff Richmond, Peter
Rohner, Stuart Smith and Michael Young.
The authors stress that the ideas, opinions and biases in this
paper are their own.
REFERENCES
Colbert, P J, Munro, P D and Yeowart, G, 2009. Prominent Hill
Concentrator – designed for operators and maintainers, in
Proceedings AusIMM Tenth Mill Operators’ Conference, pp 23–32 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
18
McCarthy, P, 2006. Message from the Managing Director [online],
Digging Deeper, AMC Consultants. Available from: <http://
www. amcconsultants.com.au> [Accessed: 5 June 2009].
Munro, P D, Eaton, R and Burton, E, 1982. Wear materials experience
in Mount Isa concentrators, in Proceedings Second AusIMM Mill
Operators’ Conference, pp 327–335 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
O’Rourke, P J, 1995. All the Trouble in the World: The Lighter Side of
Overpopulation, Famine, Ecological Disaster, Ethnic Hatred, Plague,
and Poverty, 340 p (Atlantic Monthly Press: New York).
Quirk, T, 2006. Correct weight [online]. Available from: <http://
www.onlineopinion.com.au/view.asp?article=5282&page=0>
[Accessed: 26 July 2017].
Wills, B A and Napier-Munn, T J, 2006. Wills’ Mineral Processing
Technology, Seventh Edition: An Introduction to the Practical Aspects
of Ore Treatment and Mineral Recovery, 456 p (ButterworthHeinemann).
Woodcock, J T, 1978. Mineral processing in Australasia 1978,
Australian Mining, pp 16–90.
we are metallurgists, not magicians
Contents
Back to the future – still on the dark side
P D Munro1
ABSTRACT
This is a reprise of the paper ‘Back to the future – why change doesn’t necessarily
mean progress’ presented to the Tenth AusIMM Mill Operators’ Conference in 2009.
It was observed that the ‘boom and bust’ cycles of the industry, together with trends
in tertiary education and the effects of fly-in fly-out (FIFO) operations raised serious
questions about the sustainability of human capital in the mineral processing sector.
In 2016, the industry found itself in a ‘bust’ after enjoying a decade of the longest
commodity price ‘boom’ in a century. In the boom did the industry make any
technological breakthroughs, and/or put measures in place to improve its human
capital, especially in the area of ‘professional formation’?
This paper makes observations on the current performance of mineral processing
engineers and the organisational milieu in which they operate. It questions whether
mineral processing engineers are actually ‘adding value’ at some sites given their
organisational and operational practices.
Some technical trends are examined and what they might mean for mineral
processing engineers.
Given the above, and demographic trends for mineral industry professionals,
changes to the seemingly current laissez-faire model of professional development
are proposed.
INTRODUCTION
At the Tenth Mill Operators Conference held in Adelaide in 2009 Peter Tilyard and
this author looked at ‘human capital’ in the mineral processing sector contrasting the
situation with the industry that it had entered at the beginning of the 1970s (Munro
and Tilyard, 2009). After seven years, it felt timely to reconsider the position, and to
see if the industry had improved the situation, especially given the perception that
the ‘minerals boom’ driven by China’s massive demand for raw materials was now
over and the mining industry seemed be coupled with the word ‘crisis’ in the media.
This is a personal view derived from over 45 years of experience in the business.
In the paper, the author uses the term ‘metallurgist’ to generically cover the job
titles/descriptions of primary metallurgist, mineral processor and process engineer,
while ‘concentrator’ and ‘mineral processing plant’ are used interchangeably.
Concentrator references are to a sulfide flotation plant but are just as applicable to
any mineral processing operation.
HUMAN CAPITAL – LESS ‘KNOW-HOW’ IN 2016
The lost decade
From 2000 the minerals industry enjoyed the longest commodity boom in a century.
In Australia for the 12 years to 2012, the world price of its mining exports more than
tripled and investment spending by the mining industry increased from two per cent
of gross domestic product (GDP) to eight per cent. Mackenzie and Cusworth (2016)
stated that from 2005 to 2012 private new capital expenditure in the Australian mining
industry grew from A$14.2 B in 2005 to A$94.5 B in 2012 for a mean annual growth
rate of 31 per cent.
Did the industry:
•• make any real technological breakthroughs?
•• improve the quality of its human capital by strengthening educational
institutions and skill development schemes for graduates?
1. FAusIMM, Senior Principal Consulting
Engineer, Mineralurgy Pty Ltd, Taringa Qld
4068. Email: pdmunro@bigpond.com.au
In the context of an industry that was intent on increasing production at any cost,
the answer to both of the above questions is ‘no’. While other economies with strong
mining industries in Africa and the Americas also enjoyed felicitous times, this author
believes that the industry’s processing knowledge base continued to erode.
19
P D Munro
Consider the following:
•• Decline in technical competency – we complained about
this in 2009 and it has continued to be noticed by others
(McCaffery, Giblett and Dunne, 2014).
•• Decreased research and development capability:
•• Closure of corporate research laboratories – Australian
examples are BHP Billiton, MIM Holdings, Pasminco
and WMC with diminished capability at what remains.
The conversion of former corporate research and
development facilities to commercial units offering
services to external customers must emphasise
investigating current problems at the expense of longer
term ones. Major companies, who traditionally were
major technical innovators are now very risk averse
and obsessive about intellectual property (IP) rights.
They have created bureaucracies in research and
development (R&D) making it unlikely that they will
ever achieve much by comparison with past efforts.
•• Reduced government activity for minerals research –
the United States Bureau of Mines closed in 1995 with
the loss of US$100 million/a in funding. Environmental
and sustainability issues are more likely to attract
governmental R&D dollars than minerals industry
ones. This is confirmed by looking at the respective
numbers of Cooperative Research Centres (CRCs) in
each sector in Australia.
•• One unsettling snapshot on the current status of mineral
processing is to compare the technical content on
processing of the third edition Australasian Mining and
Metallurgical Operating Practices, third edition (Rankin,
2013) with that in the first two editions (Woodcock, 1980;
Woodcock and Hamilton, 1993). While the third edition
should still be in every metallurgist’s library, it doesn’t
have as much detail as the first two. The author wonders
if the lack of content from operations of some major
companies indicates waning interest in, and commitment
to technical excellence from senior management.
•• Drivel spouted by mining companies about ‘people
being our greatest asset’ is just nonsense considering
examples such as the lack of meaningful industry
funding put into education and training, the usual knee
jerk responses in the economic downturn of not hiring
new graduates, not offering permanent jobs to those
coming off graduate training schemes and stopping
technical skills programs for professional development.
Peter McCarthy of AMC Consultants was right when he
said that real business improvement and value creation
comes from experts who have a deep understanding of
how a mining operation makes money (Anon, 2015).
That is, the industry needs more technical people doing
meaningful things, not less. The human resources sections
of companies don’t seem to make much noise about the
long-term effects on the organisation’s human capital
stock when these ‘downsizings’ happen. In 1998, the
Minerals Council of Australia published Back from the
Brink: Reshaping Minerals Tertiary Education, the National
Tertiary Education Taskforce discussion paper. With
regard to the issues raised in that document short-term
measures got us through the early 2000s but here we are
in 2016 with reports of mineral industry related courses at
universities still under continuing threat of closure. Why
would a university run a mineral processing course for a
small number of students requiring expensive laboratory
facilities when it could fill a lecture hall with full-fee
paying international students on a subject like ‘business
studies’? The author is also concerned about anecdotes
20
that computer simulation studies are being substituted
for laboratory time on subjects such as ‘unit operations’ to
save money in process engineering courses.
The situation is now worse than in 1998, considering that
the demographics of the population of metallurgists will
result in imminent loss of technical expertise as older people
leave the industry.
Professional formation and development
The usual suspects of authors (Drinkwater et al, 2011;
Drinkwater and Bianco, 2013; Drinkwater and Napier-Munn,
2014; McCaffery, Giblett and Dunne, 2014) have discussed the
need for professional formation and development.
This author will examine, in more detail, the conversion of
chemical engineers into metallurgists. Quoting directly from
Drinkwater (2015):
Data collected in Australia in 2011 (Lind, 2013) showed
that most metallurgists in Australia took their first degrees
in chemical engineering, chemistry, materials science
or engineering and a variety of other disciplines; only
25 per cent studied in programs with the words ‘mineral
process’ or ‘metallurgy’ in the name. It may be surprising
to some of our younger colleagues to learn that this is
not a new phenomenon. Many of our most prominent
senior metallurgists started out 30 or more years ago as
analytical chemists, materials scientists, steel-works cadets,
instrument technicians and (of course) chemical engineers.
Chemical engineers are the largest and most reliable
source of supply of new metallurgists. Chemical engineering
educators and practitioners point out that dealing with solids
is more difficult than gases and liquids (Merow, Phillips
and Myers, 1981; Nelson, Davies and Jacob, 1995; Merow,
2000). In the minerals business, the solid (ie ore) tends to be
heterogeneous and to change over time. Metallurgists spend
more time than chemical engineers in their tertiary studies
dealing with solids. The industry successfully converted
chemical engineers into metallurgists through a combination
of on-site training and formal courses. FIFO operations do not
meet the industry’s requirements for the development of a
metallurgist’s technical skills.
Formal continuing education programs such as Metskill
(Drinkwater and Bianco, 2013) are struggling in an economic
environment that has put cost reduction ahead of developing
its people.
In 2016, there are diminished employment and professional
development opportunities for metallurgists.
STILL GOING BAD
The author includes a list of some examples of egregious
operating practices, malfeasances and misunderstandings
to move the reader out of his/her comfort zone about how
the industry is running plants and offer a quantum of solace
through Schadenfreude (ie ‘well at least we aren’t as bad as that’):
•• The bizarre concept of ‘value engineering’ in plant design,
where the norm is to leave 25 per cent ‘fat’ in the design
for later removal – another way to look at it is the art of
rendering an economically marginal project technically
unviable by taking ‘bold’ decisions on equipment sizing,
selection and layout. For an alternative view on plant
design refer to a paper on the design of the Prominent
Hill concentrator (Colbert, Munro and Yeowart, 2009).
•• An acceptance of mediocrity in equipment running
times – Bougainville Copper showed what could be
done by understanding the basics (Tilyard, 2009; Tilyard
we are metallurgists, not magicians
Back to the future – still on the dark side
and Clarke, 2010). Alternatively there are managements
that expect Bougainville-type outcomes after reducing
the maintenance workforce.
•• The existence of three large concentrators without
ore:water ratio control on the primary grinding mill,
something that had supposedly become standard in the
1960s.
•• The fact that at a recent international conference
a colleague underwent the torture of listening to
a presentation extolling the virtues of adding a
hydrocyclone to a grinding mill and operating in closed
circuit. What will they think of next…gasp!
•• The installation of hydrocyclones without vortex finders,
enduring a long run of bad metallurgical results, only to
find out that poor classification in the grinding section
was the cause.
•• A new flotation circuit that experienced 99 flotation
cell ‘sanding’ events in 120 days and even more ‘near
sanding’ ones, causing poor metallurgy and resulting in
some blocks of cells decreed to be inoperable and thus
bypassed – the concentrator operating staff whinged
about the cells ‘having problems’ and being ‘no good’
but couldn’t trouble themselves to quantify the problem.
They did not share these facts with the engineers who
were designing an expansion of the flotation section
which was proceeding based on using the same cells.
This author expected that the operating staff would have
shown a modicum of interest and initiative in resolving
a problem that they would end up having to live with.
•• Incorrect maintenance decisions and practices affecting
metallurgical performance. At times, it seems as if all
the good words spoken and written about the need
to stop people operating in ‘silos’, and the dangers of
inappropriate key performance indicators, have been
ignored. When this author sees lowering maintenance
costs by increasing equipment service intervals causing
poor metallurgical performance, the author wonders if
concentrators operate in some parallel universe. Peter
Tilyard and the author commented on this in their 2009
paper and as soon as metal prices dropped, on too many
occasions, it has been ‘back to the future again’!
•• Metallurgists willingly abdicating important decisions
affecting metallurgical performance to the maintenance/
reliability engineering section – a probable contributor
to such aberrant outcomes is that while operating costs
can be measured to A$0.01/t, metallurgical benefits are
sometimes expressed more nebulously. This results from
a lack of data prescribed in the section ‘Mineral processing
basics’ below and the reluctance of too many metallurgists
to tear themselves away from their computer screens and
see what is happening in the concentrator. Two examples
suffice to demonstrate the problem:
1. A very large copper concentrator increased recovery
by three per cent after improved maintenance
restored the rougher flotation cells to the original
installed condition.
2. A maintenance section decided to cut the baffles out
of a flash flotation cell to reduce wear, resulting in no
feed going to the gravity concentration section.
•• Achieving 50 per cent running time for regrind
mills where the copper sulfide minerals in the final
concentrate were only 70 per cent liberated. The
predominant copper minerals were secondary sulfides,
so the operating staff comforted themselves that they
were making a concentrate assaying 22–25 per cent Cu
we are metallurgists, not magicians
when it was actually a low quality one with less than
50 per cent w/w copper sulfide minerals.
•• The testing of a new supposedly stronger flotation
collector, which actually had a shorter length carbon
chain than the existing collector – Sutherland and
Wark’s classic text Principles of Flotation (1955) states that
contact angle is dependent on the length of the carbon
chain. Knowing the chemical composition and structure
of reagents would prevent time being wasted on
meaningless testing of reagents. Two other irritants are
metallurgists trying to fix a mineral liberation problem
with reagents and not understanding the statistics of
testing. With a supposed sound grounding in statistics
during tertiary education and the availability of Tim
Napier-Munn’s seminal book Statistical Methods for
Mineral Engineers: How to Design Experiments and Analyse
Data (Napier-Munn, 2014) there is no excuse for not
doing tests rigorously. If the metallurgists can’t lever
off their education and training to add value, then as
mentioned below, why have them on-site?
•• The act of doing locked cycle flotation tests to ‘save time’
before establishing conditions using batch tests and
compositing drill core over unmineable intervals.
•• The installation of a new plate and frame filter press
at a cost of A$7.5 million to replace existing disc filters
for concentrate dewatering, only to find out that the
latter could do the duty when the concentrate thickener
underflow pump was upgraded, to allow delivery of
higher density slurry to the disc filters.
BIG DATA – A DIRGE ON PLANT INSTRUMENTATION:
LET’S DO GEOMETALLURGY INSTEAD
Instrumentation reality check
Mining company executives and the technical press extol
the virtues of ‘big data’ and the wonders this concept will
supposedly achieve towards improving operating outcomes.
There is a credibility chasm between the lofty concept of
‘big data’ and the reality in the current general unsatisfactory
state of plant measurement instrumentation and process
control systems. These are too often characterised by low
availabilities, inaccurate calibrations, discarded process
control strategies etc. The author distinguishes between
‘embedded instrumentation’ in a piece of equipment
eg ammeter, voltage, electric motor rev/min versus
instrumentation specific to a concentrator. Problems with
the measuring and sensing elements such as density gauges,
flowmeters, level sensors, on-stream analyser streams etc
are the inevitable result of reducing maintenance input; the
author can’t see how you can expect high running times from
part-time instrumentation support on a FIFO basis.
Another concern in larger companies is the tyranny
of enterprise data management systems typified by the
persistent encroachment of corporate information technology
(IT) interests into the process control area. IT people seem to
think that if it’s a computer then they want to own and manage
it. Too often the result is the imposition of inappropriate
software standards by people who may know something
about computers but very little about process engineering.
This ‘top down’ approach stifles innovation in process
control. This author has yet to hear of a mineral processing
plant controlled in real-time by SAP.
21
P D Munro
Geometallurgy
Consider one area where the industry actually does have data
but lag woefully in turning it into information resulting in
actions for the plant operators. As mentioned in Munro and
Tilyard (2009), geologists are better custodians of data than
metallurgists. They collect a plethora of data from drill core,
such as the following:
•• assays
•• rock type and alteration from
•• geological logging
•• petrological examination
•• photographs
•• mineralogy and textural analysis from
•• hyperspectral logging
•• mineragraphy
•• XRD (X-ray diffraction)
•• automated methods such as QEMSCAN and MLA
•• geotechnical data
•• rock strength data eg UCS, PLI
•• competence RQD.
To this should be added the location in 3D mineralised
space of every sample taken for mineralogical examination,
comminution and flotation tests etc. Unfortunately, current
operations often make little use of the extensive metallurgical
test work done for a feasibility study and fail to properly
integrate subsequent test work into this body of knowledge.
They seem to be two separate ‘silos’ rather than being
regarded as a continuum of knowledge. The author wonders
how interested people are in their profession when one
encounters operating metallurgists who either have not read
the feasibility study or, if done so, failed to understand it.
People talk about geometallurgy, but when all is said
and done there is a great deal more said than done. Some
operations do use geometallurgy in managing plant
operations (Butler et al, 2016) but remarkably few have
fully integrated it into their way of doing business. In 2016,
the fact that too many geologists and metallurgists are still
operating in ‘silos’ is totally the fault of the latter. The author
hasn’t observed any significant improvement in this over
the last seven years with people continuing to be resolutely
qualitative rather than quantitative. For example, throughput
variations continue to be ascribed to the ore being ‘harder’
or ‘softer’ rather than hearing a metallurgist at the morning/
daily meeting say something like ‘our current feed is now
from bench/stope ‘AB’ where the predominant rock type is
andesite; grindability data for this rock type shows a Bond
Ball Mill Work Index of ‘A’ kWh/t, JK Drop Weight/SMC
a × b of ‘B’ etc so the predicted throughput from the grinding
simulation model should be ‘X’ t/h’.
This author dreams of people at the daily meeting in every
concentrator reviewing a 3D representation of the orebody on
a screen that incorporates the data described, explaining both
yesterday’s performance and predicting future results.
We have the technology to do this, but don’t use it.
Bureaucratisation of the metal balance
Another gripe is the bureaucratisation of the metal balance;
it has become integrated into an enterprise data management
system promoted by accountants and the IT department.
There are very high ownership costs of these metallurgical
accounting add-ons to the enterprise data management
22
system. Claims that they are better supported than Excel are
just tosh; hasn’t anyone heard of Microsoft?
This author fears that they further encourage metallurgists
to spend even more time in the office rather than in the plant,
as they won’t tell you that:
•• the head sampler is stuck at the edge of the stream and
is receiving splash
•• the flotation operator is decanting the slurry sample and
losing the higher grade fines before filtration.
We are missing the point if the industry is looking for an IT
solution to what has always been an equipment and people
problem. A realistic treatment of measurement errors is often
lacking in these systems, a concept that many accountants who
drive them can’t comprehend as they deal in money, which can
be counted; rather than tonnes or assays which, are measured.
CONCENTRATOR OPERATIONS
Mineral processing basics
Nothing has changed from Munro and Tilyard (2009) in
the fundamental data requirements for managing a mineral
processing plant.
A particular peeve is the absence of a credible metallurgical
development plan. Some of the documents that claim to be
one are laughably unfit for purpose. Below are common
examples of bad practice:
•• A bewilderingly large number of ‘metallurgical
initiatives’ – some operations have hundreds of them.
The manager and technical superintendent should
allocate the professional time required to complete
each initiative then relate the total to the ‘discretionary
hours’ available for the metallurgists from both internal
and external resources. Going through such a list with
‘competent person hours’ as the limiting resource leads
to the realisation that the deposit will have been mined
out before even a small fraction of the initiatives have
been completed.
•• Competing priorities – this author has seen a set of
metallurgical initiatives with 30 ‘priority 1’ items. The
compilers of such a list must have lacked the basic mineral
processing data listed above to appropriately allocate
priorities according to ‘pay-off’, ‘time to complete’ etc.
•• Vague targets – an example would be ‘increase copper
recovery by two per cent’ or slightly better, ‘increase
copper recovery in rougher flotation by two per cent’.
What this author wants to see is something like ‘increase
recovery by four per cent of composite particles that
are 50 per cent chalcopyrite or more with non-sulfide
gangue in the +75 μm -106 μm size fraction in rougher
flotation’. The first two examples above are ‘motherhood’
statements. They risk that if a metallurgical initiative
is implemented to ‘increase copper recovery by
two per cent’, its benefit could be diminished by factors
such as a change in ore type or operating practices. It is
much easier to demonstrate to senior management that
you actually did ‘increase recovery by four per cent of
composite particles that are 50 per cent chalcopyrite or
more with non-sulfide gangue in the +75 μm -106 μm
size fraction in rougher flotation’ as opposed to not
achieving the more general target. You retain credibility
with senior management who are rightly sceptical of
previous promises about metallurgical improvements
which, if aggregated, often would give a metal recovery
over 100 per cent. This leads you to understand the
process according to the axiom of size-by-size mineral
we are metallurgists, not magicians
Back to the future – still on the dark side
particle behaviour by liberation class but this a fundamental
of the profession. This author wants a focused target
instead of a general one where the result can be obscured
by the ‘noise’ of plant operations.
If the concentrator manager doesn’t have a credible
metallurgical plan based on process fundamentals then
one will be imposed by senior management, generally with
unsatisfactory results.
An excellent management book is Moneyball: The Art of
Winning an Unfair Game by Michael Lewis (2003), which
describes how the Oakland Athletics baseball team managed
by Billy Beane successfully competed with a player roster that
lacked the stars fielded by other teams and had a budget only
one-third of that of the New York Yankees (read the book –
the Hollywood movie left out the mathematics, which was the
really interesting part).
Quoting directly from the book:
One cliché is the ‘five-tool player’, a player who can hit for
average and power, and run, field and throw like a demon.
You still hear it all the time.
Right. And everyone wants five-tool players. There’s just
very few of them on the planet! But virtually every player has
one tool. So, we started saying, ‘Well, let’s investigate each
player’s strength. Is there a way to combine all these strengths
and cover up some of their deficiencies? We’re not going to be
able to do away with their deficiencies. But as a team, can we
do it all?’ It was about piecing together players we had on our
roster and building a team that could do everything. You had
to find value where it wasn’t readily apparent.
One lesson from Moneyball was that you can’t assume
that a collection of apparently talented individuals will
serendipitously give the required result. Hiring someone who
looks like the ‘complete metallurgist’, analogous to the ‘fivetool player’, and just leaving him/her to it under a ‘laissezfaire’ approach isn’t getting us the required outcomes listed
above under ‘Mineral processing basics’.
A second learning from Moneyball is that you can compete
by focusing on the separate components that contribute to the
desired outcomes. If you accept that you need the items in
the ‘Mineral processing basics’ list then you should prescribe
that they have to be produced, it’s not some kind of an option;
then work out the processes and actions required from both
internal and external resources to make it happen.
Running under the premise that a ‘champion will somehow
spontaneously emerge’ isn’t delivering the goods.
Meetings – still the practical alternative to work
The limiting resource to getting anything done is ‘competent
person hours’ so it’s a wonder that too many concentrator
managers have a long and unfocused morning/daily meeting.
While such gatherings can be social successes, I deplore the
waste of time of having up to 20 people milling around for an
hour. The morning/daily meeting should not be an exercise
in participatory democracy if you have to endure a junior
metallurgist drone on for ten minutes explaining 24 hours of
‘screen grabs’ of primary crusher operating variables in 15
pastel colours. All the people at the meeting need to know
about the crusher is the tonnes in the run-of-mine stockpile,
crusher operating hours and the tonnes in the crushed ore
stockpile.
In this author’s dream world, the geometallurgy section
of the meeting should not be a long-winded discourse on
geology, but emphasise the nature of the future ore feed and
what actions the operators should take in the next 24 hours
to meet the metallurgical performance targets. As a young
we are metallurgists, not magicians
metallurgist, the author witnessed a senior manager fully
comprehending the operations of two large concentrators and
two smelters in a little over an hour including travelling time
between each plant. The emphasis was on information rather
than data, the future rather than past, with metallurgists
held accountable for performance and subjected to ruthless
questioning about any deviations. We can learn from the
‘orders group’ concept used in military forces on how to run a
successful meeting but if you want ‘participatory democracy
then …’. This author has not sighted any evidence showing
that the workforce has superior skills in chemistry, physics
and mechanical aptitude compared to 45 years ago, excluding
the ability to send text messages. Indeed, with respect to issues
like vaccination, ‘organic’ food and homeopathy etc, one
could believe that society has become more technologically
illiterate!
SOLID LIQUID SEPARATIONS – NEGLECTED UNIT OPERATIONS
Solid liquid separations – uninteresting and unloved?
Mineral processing is simplistically described as ‘LSD’, that is:
•• liberation
•• separation
•• disposal.
The ‘disposal’ part of the business for both product and
waste streams has been ancillary to the other two sectors of
‘liberation’ viz comminution + classification and ‘separation’
as done by gravity concentration, flotation and other methods.
Taking the case of a sulfide flotation concentrator, previously
mineral processing engineers have tended to focus their
intellectual efforts on the ‘interesting’ activities of grinding
and flotation; the former is usually the highest cost unit
operation and the latter the bit where it is easily discerned if
money has been made or lost.
An example of neglect in solid-liquid separations is the
indifference about the operation of clarifiers and thickeners
(the two duties hereafter combined in the single unit operation
of ‘thickening’) which are ubiquitous in concentrators and
extractive metallurgical plants. Now into my fifth decade in
the business, this author has seen a considerable number of
them, and the operation of many could only be described as
‘poor’ to ‘appalling’. Depressingly, it doesn’t appear to be
getting any better with the passage of time. Malfunctioning
solid-liquid separation unit operations threaten the viability
of hydrometallurgical process plants and recycled thickener
overflows containing fine solids can materially affect flotation
performance.
Thickening is a good example of this neglect. Consider the
following thickener fundamentals:
•• In a thickener, there are two flows: simplistically
solid going downwards and liquid going upwards.
Calculation of the rise velocity of the liquid going
upwards gives a number that is usually higher than the
settling velocity of the finest particles in the feed. Note
that the settling rate of a ten micron quartz sphere in
water is 0.6 cm/min at 25°C (Dorr and Bosqui, 1950).
Hence the importance of adequate flocculation (and
coagulation in water treatment) to create agglomerates
with a higher settling velocity than the rise velocity.
•• Hydrometallurgical plants often have multiple thickener
operations eg counter-current decantation. Taking a
desired 92.5 per cent operating time with a total of
five thickeners in series in the flow sheet means that each
thickener has to be performing as specified 98.5 per cent
23
P D Munro
of the time ignoring all other causes of downtime and
poor operation!
qualities. Water treatment could become common in
concentrators.
In a mineral processing world of finer particle sizings,
due to the nature of ores treated, leading to more difficult
solid-liquid separation duties, the following are some of the
depressing realities of thickener operations:
2. Tailings – this is the ‘big one’. The mining industry has
a poor and increasingly indefensible record managing
tailings dams; this author prefers the word ‘dams’ to the
euphemism ‘tailings storage facilities’. Doubters about
this should spend time on the internet using the search
term ‘tailings dam failures’. To dismal incidents such as
Certej (Romania) 1971, Buffalo Creek (USA) 1972, Val
di Stava (Italy) 1985, Omai (Guyana) 1994, Merriespruit
(South Africa) 1994, Los Frailes (Spain) 1998 and Baia
Mare (Romania) 2000 we can now add the recent events
of Mount Polley (Canada) 2014 and Samarco (Brazil)
2015. Over the last century Azam and Li (2010) estimated
the failure rate of tailing dams at 1.2 per cent compared
with 0.01 per cent for conventional water retention dams.
What an indictment on our industry! This author sees
a future where dry stacking of tailings will be enforced
by regulation. An example is the proposed 72 000 t/d
Rosemont copper-molybdenum-silver operation of
Hudbay Minerals near Tucson, Arizona, USA, which
will only be permitted if tailings are dry stacked. If dry
stacking is mandated then the unit operation of filtering
tailings will become as important as grinding in terms
of plant throughput and operating cost. Metallurgists
are the best qualified people to design and operate
such dewatering facilities. If this industry continues
its current lack of interest in the ‘back end’, the civil/
geotechnical engineers will ‘eat our lunch’ just as the
mining/geotechnical engineers have got into the paste
fill business.
•• Flocculent systems malfunctioning with incorrect
addition and subsequent dilution; flocculents once
formed are ‘maltreated’ in their passage to, and through,
the thickener feedwell.
•• Feedwells in poor condition leading to excessive
turbulence damaging flocs, side feeding etc.
•• Thickeners acting as a further stage of cleaner flotation
with voluminous froths on the surface exacerbated by
upwelling air bubbles; technologies to de-aerate flotation
concentrate froths have been available for over 15 years
(Garraway and Kaboth, 2001) and should be standard
for such duties.
•• Solid-laden concentrate thickener overflows reminding
one of the Dutch farm adage ‘too wet to plough, too dry
to drink’.
•• Thickener overflow launders choked with deposited
solids.
•• Thickener underflow density reduced to cope with
inadequate pumping systems: this author has to
suppress a tirade of vituperation when told by a process
engineer that the downstream filtration stage ‘works
better at lower solids density’ when what he meant was
that the underflow pumping system was undersized and
couldn’t deliver a high density slurry to the filters. How
could someone claim to be a professional metallurgist
after making such an asinine statement?
•• Too often the suggested remedy is to install additional
instrumentation when the problems are caused by
flawed process fundamentals.
Process engineers are encouraged to get ‘back to basics’
(has anyone ever made a credible case for leaving them?)
on thickening. This includes learning from the excellent
work done by CSIRO in the AMIRA P266 Project ‘Improving
Thickener Operations’.
Paradigm revolution in concentrator operations
– driven from the ‘back end’
The author discerns a trend that processing aspects of mining
operations and projects will be increasingly driven from the
‘back end’ influenced by two main issues:
1. Water – generally, the mining industry will not be
allowed access to water until competing demands for
human, agricultural and pastoral uses have been met.
The result is that the industry has to maximise both
water recovery from tailings and recycling of water. This
is not a novelty for Australian mineral processing plants,
but it is for some of the jurisdictions where we work. The
move to dry stacking of tailings discussed below means
operating the plant with a higher proportion of recycled
water than the industry has considered the ‘norm’.
This reduces the ability to maintain water quality by
bleeding out deleterious chemical species via the tailings
stream. A build-up of certain chemical species will affect
the metallurgical performance of flotation which is a
separation process dependent on the quality of mineral
surfaces. Johnson (2013) has discussed treatment of
recycled process water to achieve specific chemical
24
THE FUTURE FOR METALLURGISTS
Metallurgists – are they needed?
McCaffery, Giblett and Dunne (2004) asked whether we still
needed metallurgists. This author accepts that we still do, but
would like to see more evidence that they are making a positive
economic contribution as this author is currently unconvinced
about the value of much of the ‘on-site professional hours’,
particularly in FIFO operations.
Consider a typical FIFO site with the following professional
staff:
•• concentrator manager
•• metallurgical superintendent
•• two plant metallurgists
•• two project metallurgists.
Annual cost to the employer would be at least A$1 million/a.
If the reality of the metallurgists on-site:
•• have gaping deficiencies in the list of ‘Mineral processing
basics’
•• are being used for clerical rather than technical tasks.
Managers and senior professionals are to blame for the
allocation of metallurgists to routine and mundane tasks
instead of process improvement ie ‘discretionary time’
(Munro and Tilyard, 2009). Consider things such as not
being able to manage a good plant survey and not doing
statistically valid test work. (This author has difficulty
remembering when he last went to a concentrator and
the metallurgists could show him actual corrected and
reduced efficiency curves for the hydrocyclone classifiers)
•• are unable to sustain gains
•• lack process or corporate memory
we are metallurgists, not magicians
Back to the future – still on the dark side
•• have no awareness of current and past operating
practices for the type of ore being processed and
seemingly little interest in finding out
Metallurgists are too narrowly focused on liberation and
separation; solid-liquid separations are likely to be the future
‘drivers’ of concentrator design and performance.
•• are guilty of other technical malfeasances as mentioned
has been mentioned.
A new model for the deployment of metallurgists could
have less technical people on-site using experienced
external technical people in a scenario that provides genuine
professional development of junior metallurgists.
If metallurgists on-site aren’t using the technical skills they
supposedly acquired during four years of tertiary education,
then how could a company justify having them? Is their
presence some kind of ‘Potemkin village’ to reassure outsiders
that metallurgy matters?
A possible new model
A person more interested in outcomes than appearances
might consider how he/she might alternatively spend the
A$1 million/a outlaid for the metallurgical team. Here are
some heresies to current practices:
•• the concentrator manager could be either a metallurgist
or a skilled production superintendent
•• the metallurgical development plan is produced and
monitored by head office metallurgists and/or external
consultants
•• clerks perform many of the tasks of the plant metallurgists
•• one can buy a lot of external ‘Competent Person hours’
for the yearly salary of a metallurgist on-site who is
effectively a clerk and should be quite demotivated
acting as one.
This new model has less technical people on-site and
uses experienced external technical people in a scenario
that provides genuine professional development of junior
metallurgists. There are analogies to the South African
‘mining house’ model where site managers had a reporting
line on technical issues to consulting engineers and practices
in the oil industry where much of the expertise comes from
specialist consultants.
CONCLUSIONS
The ‘mineral boom’ since 2000 has not led to any technological
breakthroughs nor improved the quality of our human
capital by strengthening educational institutions and skill
development schemes for graduates.
It is the author’s opinion that the situation is now worse than
in 1998, considering that the demographics of the population
of metallurgists will result in imminent loss of technical
expertise as older people leave the industry.
The industry is not behaving appropriately to both improve
the skills of its metallurgists and ‘convert’ chemical engineers
to metallurgists. This is particularly so given the large number
of FIFO operations.
A meaningful use of ‘big data’ would be to fully integrate
geological and metallurgical databases to make geometallurgy
a practical reality.
Concentrators are still running with the usual glaring
examples of egregious operating practices, malfeasances
and misunderstandings. Poor metallurgical results due to illinformed maintenance cost savings make you wonder if the
industry will ever learn from previous mistakes.
Having the basics of mineral processing on-site to effectively
operate a concentrator is still the exception rather than the rule.
ACKNOWLEDGEMENTS
The author thanks Mineralurgy Pty Ltd for permission to
publish this paper.
The following individuals are thanked for their continuing
observations on these matters over the years: Peter Colbert,
John Glen, Rolly Nice, Joe Pease, Peter Rohner, Tom Shouldice,
Stuart Smith, Jorma Tuppurainen and Michael Young.
The author stresses that the ideas, opinions and biases in
this paper are his own.
REFERENCES
Anon, 2015. Mining consultant rejects rivals’ advance, Mining
Magazine, November.
Azam, S and Li, Q, 2010. Tailings dam failures: a review of the last one
hundred years, Geotechnical News, December, pp 50–53.
Butler, C, Dale, R, Robinson, S and Turner, A, 2016. Geometallurgy
– bridging the gap between mine and mill: a case study of the
Degrussa geometallurgy program, in Proceedings GeoMet 2016,
pp 77–88 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Colbert, P J, Munro, P D and Yeowart, G, 2009. Prominent Hill
concentrator – designed for operators and maintainers, in
Proceedings Tenth AusIMM Mill Operators’ Conference, pp 23–32 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Dorr, J V N and Bosqui, F L, 1950. Cyanidation and Concentration of Gold
and Silver Ores, second edition, p 23 (McGraw-Hill: New York).
Drinkwater, D, 2015. What should be in the Mineral Processing
curriculum?, The AusIMM Bulletin, 2:44–47.
Drinkwater, D and Bianco, N, 2013. Developing technical excellence
in young Australian metallurgical professionals – a new graduate
development program, in Minerals Industry: Education and
Training, IMPC Monograph (eds: J Cilliers, D Drinkwater and K
Heiskanen), pp 117–129 (Indian Institute of Minerals Engineers:
Jamshedpur).
Drinkwater, D, Bradshaw, D, Tilyard, P and Munro, P, 2011.
Professional development for metallurgists – improving technical
skills, in Proceedings Metallurgical Plant Design and Operating
Strategies (MetPlant 2011), pp 39–48 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Drinkwater, D and Napier-Munn, T, 2014. Why good professional
development is key to profitability in the mining industry,
Chapter 20: Mineral economics and human capital, in Proceedings
XXVII International Mineral Processing Congress (IMPC 2014) (ed: J
Yianatos), pp 28–37, Chile.
Garraway, B and Kaboth, K, 2001. The Frothbuster – innovative
thickener froth reduction technology, in Proceedings 33rd Annual
Operator’s Conference Canadian Mineral Processors, pp 17–24,
(The Canadian Institute of Mining, Metallurgy and Petroleum:
Montreal).
Time management to ‘create Competent Person hours’ is
still a problem.
Johnson, N W, 2003. Issues in maximization of recycling of water in
a mineral processing plant, in Proceedings Water in Mining 2003,
pp 239–246 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Many metallurgical development plans need to be trimmed
down to reflect the ‘possible’ and to have meaningful priorities.
Lewis, M, 2003. Moneyball: The Art of Winning an Unfair Game (W W
Norton & Company: New York).
we are metallurgists, not magicians
25
P D Munro
Lind, G, 2013. Minerals industry engagement in metallurgical
education in Australia’, in Minerals Industry: Education and
Training (eds: J Cilliers, D Drinkwater and K Heiskanen)
pp 87– 104 (Institute of Minerals Engineers: Jamshedpur).
Napier-Munn, T J, 2014. Statistical Methods for Mineral Engineers, How
to Design Experiments and Analyse Data, JKMRC Monographs
No 5 (Julius Kruttschnitt Mineral Research Centre University of
Queensland: Indooroopilly).
Mackenzie, W R and Cusworth, N, 2016. The use and abuse of
feasibility studies – has anything changed, in Project Evaluation
2016, pp 133–148 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Nelson, R D, Davies, R and Jacob, K, 1995. Teach’em particle
technology, Chemical Engineering Education, 29:12–16.
McCaffery, K M, Giblett, A and Dunne, R, 2014. Sustaining
metallurgical competencies, Chapter 20: Mineral economics
and human capital, in Proceedings XXVII International Mineral
Processing Congress (ed: J Yianatos), pp 15–27, Chile.
Merow, E W, 2000. Problems and progress in particle processing,
Chemical Innovation, 30:35–41.
Merow, E W, Phillips, K E and Myers, C W, 1981. Understanding Cost
Growth and Performance Shortfalls in Pioneer Process Plants (The
Rand Corporation: Santa Monica).
Minerals Council of Australia, 1998. Back from the Brink: Reshaping
Minerals Tertiary Education, National Tertiary Education
Taskforce Discussion Paper, February.
Munro, P D and Tilyard, P J, 2009. Back to the future – why change
doesn’t necessarily mean progress, in Proceedings Tenth AusIMM
Mill Operators’ Conference, pp 5–11 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
26
Rankin, W J (ed), 2013. Australasian Mining and Metallurgical Operating
Practices, third edition (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Sutherland, K L and Wark, I W, 1955. Principles of Flotation,
pp 98– 100 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Tilyard, P J, 2009. The Bougainville Copper concentrator: a
retrospective – part 1, The AusIMM Bulletin, 5:73–81.
Tilyard, P J and Clark, N, 2010. The Bougainville Copper concentrator:
a retrospective – part 2, The AusIMM Bulletin, 6:69–74.
Woodcock, J T (ed), 1980. Mining and Metallurgical Practices in
Australasia, third edition (The Australasian Institute of Mining
and Metallurgy: Melbourne).
Woodcock, J T and Hamilton, J K (eds), 1993. Australasian Mining and
Metallurgy, second edition (The Australasian Institute of Mining
and Metallurgy: Melbourne).
we are metallurgists, not magicians
Contents
Undermining productivity – when good
key performance indicators go bad
J D Pease1
ABSTRACT
All organisations, from businesses to sports teams, have key performance indicators
(KPIs). These are always well-intentioned. A small number are crucial – indeed, they
are ‘key’. But without constant vigilance KPIs grow like weeds and do more harm
than good. It takes strict discipline and a deep business understanding to maintain
truly ‘key’ performance indicators that benefit the organisation.
ON ARCHAEOLOGY, TREES AND HOME INSULATION
An archaeological team had uncovered the remains of an ancient settlement, and
needed to recover as much as it could before the next wet season. Luckily there
was a local population of poor, hardworking villagers who enthusiastically joined
the dig. As the work progressed, the organisers were initially excited by the large
amount of ancient pottery recovered. But gradually they became disappointed that
they rarely recovered a complete piece – almost everything was broken. As ever more
fragments were returned, this moved from disappointing to perplexing to statistically
inexplicable. The archaeologists checked their numbers, reviewed data from other
sites, considered all the technical factors. Finally – and sadly too late – someone asked
them how they paid the villagers. ‘Well, we pay them for each piece they bring us’.
The villagers may not have had much formal education, but they certainly
understood management KPIs.
A small Australian town had become a desirable destination for ‘tree changers’. A
new council was elected with a mandate to protect the town’s leafy green image. They
passed a by-law that all trees over a certain girth would be listed and protected by the
council. Special approval would be needed to cut or even prune them. That weekend
the sound of chainsaws rang out around the town. Some owners with trees above the
girth cut them down before they were listed, fearing the development restrictions of
a ‘listed’ tree would reduce their property value. Worse, many young smaller trees
were cut down too. These should have been the next generation of trees, but their
owners opted to remove them while it was legal.
The most well-intended laws had precisely the opposite effect.
Recent governments have excelled at this art of unintended consequences. From roof
insulation to solar feed-in tariffs to training schemes, programs have been introduced
that are as well-intended as they are wasteful. They seem to say ‘here is a bucket of
money, you can help yourselves, but we can trust you all to do the right thing …can’t
we?’ It is alarming how often well-intended KPIs cause disastrous outcomes. It is not
that they are simply inefficient or ineffective. Their careless design encourages the
worst behaviour. It would have been better not to have any KPIs at all.
Sometimes this is called the law of unintended consequences. I just call it lazy and
dumb. Thoughtless. Honestly, what did they think was going to happen?
OF BANKERS, MINERS AND FOUNDING FATHERS
We would like to think big business is smarter than that. But consider the Global
Financial Crisis (GFC). How many bankers paid back the bonuses they ‘earned’ by
writing the bad mortgages that led to the GFC?
Or consider miners. Though, here the story takes longer to unfold.
1. FAusIMM, Senior Principal Consulting
Engineer, Mineralis Consultants Pty Ltd,
Brisbane Qld 4066.
Email: jpease@mineralis.com.au
In the 1950s, the underground mine at Mount Isa was one of the richest in the world.
Work was hard – mining relied on hand-held drills and considerable manual labour.
It was tough work in a cramped space at over 40°C and high humidity. It was quite
reasonable that miners were paid relative to effort – metres drilled or tonnes mucked.
The miners worked hard and both they and the company were rewarded. This was
capitalism at work, aided by an appropriate KPI and reward system. But technology
27
J Pease
changed, as it does. New drill rigs, new underground vehicles
were developed. Everyone knew that if the company bought
an expensive new drill rig to double productivity, then
the operator’s rate per metre should halve. But people are
people so that was hard to negotiate. There was only one
manager to negotiate with maybe 50 miners each with a deep
personal interest in capturing some of the productivity gain
for themselves. The manager was busy, had his own KPIs to
meet and couldn’t afford industrial unrest. Even if some of
the benefits leaked to bonuses the company was still much
better off and production wasn’t interrupted.
That seems okay – until it happens over several generations
of technology change. After about 40 years of this ‘creep’
underground miners in air-conditioned cabins were doing
less arduous work than road workers or truck drivers but
were paid three times as much. They also argued that they
shouldn’t lose money when the equipment broke down – it
wasn’t their fault the company couldn’t maintain it. So, they
negotiated a formula to pay them the average rate of the last
six shifts if the equipment wasn’t available. This encouraged
them to get a high ‘run rate’ for six consecutive shifts – even if
it meant pushing the equipment so hard it broke down on the
seventh shift; but that wasn’t their problem.
Perhaps the later generation of miners were simply lucky
to be in the right place at the right time. But the system was
broken. The mine was older and ore was lower grade and
more complex. Some ore zones had to be excluded because
they disrupted the plant. We asked the miners to avoid it or
tip it to waste. But at the same time, we paid them a handsome
bonus to mine it. What did we think was going to happen
when no one was looking?
In this case the original KPIs – simple $/metre – worked
well for a while. But they didn’t stand the test of time. The
founding fathers of the bonus scheme were not visionary
enough in their design. Once a bonus scheme is in place it
is really hard to change. You are trying to change the rules
during the game, in a way that costs people money. That is
not going to happen.
Considered in this context, the founding fathers of
democratic Constitutions deserve enormous respect. They
didn’t get everything right; no-one could foresee 200 years of
technological advances. But they had a wisdom about people
and what drives behaviour. Constitutions have stood the test
of time much better than most management KPIs.
THE PRODUCTIVITY CHALLENGE
After every mining boom we focus again on productivity. We
cut costs where we can, we reduce staff and we look for new
innovations. Yet most operating people know we can do a lot
better with the equipment we already have. We don’t need a
big scientific breakthrough or lots of new equipment. Most
managers can list the improvements they could make. If only
they could focus on the things they know to be important
rather than all the other ‘bumph’ they have to do.
So why can’t smart people do what they know they need to?
THE ANT AND THE GRASSHOPPER
Once a grasshopper managed a mill. Rather than understand
minerals, it understood the company. Instead of surveying
the plant, it studied the KPIs. Rather than understand the
client, the smelter, the grasshopper read the marketing
contract. It increased tonnage, reduced unit cost and outnegotiated the smelter. The grasshopper was promoted and
lived a prosperous life.
28
In a nearby village, an industrious ant managed the mill. It
strived to understand how all minerals behaved. It sampled
the plant and studied real mineralogical data. This took up
most of the ant’s time, but it refused to trust the Big Bad
Data on the office monitor. One summer, after studying
mineralogy and smelting thermodynamics, the ant installed
more equipment and improved concentrate quality. It became
even busier with more equipment mouths to feed, but the
smelter danced for joy. Overall far less energy and cost was
needed to make metal. Yet the ant missed its KPIs because
unit costs increased and recovery didn’t. One day the ant
missed a process-improvement meeting because it was busy
doing a plant survey. This disappointed management and the
ant was quietly moved to the safety department.
KENTROPY
We all know the second law of thermodynamics: without the
constant input of energy a system will move towards disorder
– entropy.
Our organisations are the same with KPIs. Without
significant energy and discipline our KPIs multiply and
proliferate. They are weeds that get out of control. It isn’t
because our organisations lack intelligent and hard-working
people. It is because they are filled with intelligent, hardworking people. As a group, that is what we do. Everyone
wants to make a ‘process improvement’. We all know that
others would work better if only they did things our way. The
more people we have in an organisation the more layers we
have and the more KPIs we have. And the less relevant they
become to our essential tasks.
How many readers can say they have a succinct set of KPIs
that enable them to focus on the few critically important
aspects of their job?
Our organisations aren’t stupid, but they end up doing
stupid things because they don’t apply constant energy into
resisting the inexorable sprouting of well-intentioned KPI
weeds. Without that constant energy, Kentropy prevails.
HAS ANYONE SEEN MY KEY?
The K in KPI stands for key. It must be the most overused word
in business. See if you can find a business memo not peppered
with ‘keys’. People now even talk about ‘the key KPIs’.
Readers are encouraged to google ‘KPIs’. There are scores
of titles such as ’20 000 KPIs used in practice’. Twenty
thousand KPIs? Seriously?
Yes, business is more complex and we all need to manage
multiple things. But most of us have so many KPIs it feels like
carrying a huge bunch of keys. Most of them just get in the
way. Without the right one it doesn’t matter how many others
you have in your back pocket.
THE WISDEN OF CRICKET?
In the 2000s, the Australian cricket team dominated the sport.
They had rare skills. With their coach they took the game to
a new professional level and made a science out of analysing
opposition teams, field placements, nutrition and mental
disintegration of opponents. They didn’t need much coaching
on the basics of the game because they were excellent at that
already. The performance focus was on other areas.
But over a couple of years the champions left the team.
Slowly it lost its aura. In response, management set more
and more targets and goals, training and nutrition regimes.
This culminated in four of the best players being sent home
before a crucial game because they didn’t submit written
we are metallurgists, not magicians
Undermining productivity – when good key performance indicators go bad
homework. Gradually two factors became apparent. First, the
game wasn’t fun for them anymore. But more importantly,
they simply weren’t good enough at batting, bowling and
fielding. Team management previously could take that for
granted. Now, without the basics, no amount of sports science
and psychology and KPIs were going to help.
Does this remind anyone of their own workplace? You are
probably overwhelmed by KPIs and standardisation. It wears
you down. You respond because you are queried about them
and your performance review and bonus depends on them.
But when was the last time management asked you
about your grind size distribution, your cyclone efficiency
curve, your mineral liberation, the source of diluents in
your concentrate? These are the very basics of running a
concentrator but if management doesn’t ask about them
people won’t focus on them. New graduates won’t embrace
them. Gradually the skills fade away. We no longer know the
basics. We can’t win.
Of course, other indicators are important and need your
attention. But they are not ‘key’. They are not the handful of
things you absolutely must get right for the business. Just as
cricket in its Wisden has volumes of statistics. But a batsman
at the crease only needs to know a few things: how many runs,
how many wickets, how many overs? Those true KPIs need
their full attention. If they get them right, the other statistics
look after themselves.
This is the difference between big data and the right data.
SOMETIMES IT IS BETTER TO BE SUBJECTIVE
One of the reasons we have so many distracting KPIs is our
beloved performance reviews. As rational and numerate people,
we eschew subjective measures. If something is important we
must be able to measure it. So we can’t measure ore quality
online but we can measure tonnes. We can’t measure valueadded at each step but we can measure $/tonne.
What do we think will happen then?
The most important characteristics of employees can’t be
measured numerically – their integrity; how they cooperate
and mentor others; their tenacity and willingness to do
whatever it takes; to work across the silos abandoning their
own KPIs and bonus because a higher priority arose in another
department. We know these people when we see them, but
we can’t put a number on it. They are the employees we
want; not the grasshoppers who argue about their score for
‘cooperation’ and assiduously manage their KPI spreadsheet
to maximise their bonus. We don’t like the grasshoppers, but
we reward them nevertheless. We all complain about the silos
in our organisation. Management says it wants to break them
down. It says that, but it pays us to do otherwise.
we are metallurgists, not magicians
The silos we bemoan are constructed from our very own
KPIs.
THE KEYS TO COPING WITH KEYS
The devaluation of KPIs is caused by the good intentions of
intelligent people. There is hope because it is well-intended.
Your approach depends on your role.
For researchers and suppliers: you can’t change your client’s
KPIs. But you need to understand them. If your business
proposition requires the client to work across the silos, then
find out if their KPIs will encourage and reward people to do
this. If they don’t find another client. If an organisation can’t
work across silos for its own benefit then it certainly won’t do
it for yours.
For supervisors and middle management: maybe your own
KPIs are exhausting and you can’t change them. But you can
simplify them for your own people. Give them goals that
let them focus on the basics of the business – and then hold
them accountable to that. Allow for enough subjectivity to
encourage the types of important and positive conduct and
performance that are hard to measure. Like a sports coach
you know that if your team gets the basics right, your job will
be easier and more secure regardless of your own KPIs.
For senior management: you can’t be an expert at everything.
You don’t have to have played elite sports to manage them.
But you need to understand the basics of the game and how
people behave and respond to incentives. You need to know
who to listen to in order to pick and develop a team who are
good at the basics. Then let your people focus on the basics
and hold them accountable for delivering them.
You must have KPIs that encourage – or at least don’t
discourage – people to work across the silos for the benefit of
the whole business. This is so obvious that it is rarely done.
The simple concept conceals devilish design and is resisted by
grasshoppers who like the simplicity of being able to manage
their own KPIs and bonus independent of the organisation.
Finally, you need constant vigilance against Kentropy.
Like the second law of thermodynamics it is relentless
and immutable. Your intelligent people (and our beloved
management consultants) will constantly propose helpful
new KPIs and systems. They will ask you to replace subjective
measures with ‘hard numbers’ that they have developed.
Everyone has a compelling argument, but the cumulative
effect will smother your organisation. You need constant
energy and conviction not to adopt them.
Just because you pulled out the weeds last year doesn’t
mean you don’t have to do it again this year.
29
Geometallurgy
Contents
Geometallurgy – what do you really need to
know from exploration through to production?
K Ehrig1
ABSTRACT
You are a metallurgist, geologist, geometallurgist or processing engineer and are
fortunate enough to work for an organisation who asks you to design and execute
a geometallurgy program. What seems like a simple request is actually not a trivial
exercise. ‘Recipe’ style manuals are not readily available. However, many ‘experts’
are lurking around willing to deliver solutions they claim will solve all of your
problems. Even though similar ore deposit styles (eg porphyry Cu, Fe-oxide Cu-Au,
Ni-laterites, calcrete hosted U etc) have many similar characteristics, each ore deposit
is unique. Hence your geometallurgy program needs to be fit-for-purpose. This paper
offers some insights gained via designing and executing geometallurgy programs
from exploration through to production, and working at a mine where you are held
accountable for your metallurgical performance predictors.
INTRODUCTION
Metal grades, mineralogy, rock and mineral texture are variably distributed across all
‘world-class’ mineral deposits. Hence the metallurgical performance of ores mined
and processed throughout the production life of an ore deposit will also vary. Mining
industry experience over the past 100+ years has shown that:
•• ‘recovery’ = f (mineralogy, grade, ore texture, process conditions) (Bojcevski,
2004)
•• future performance ≠ f (historical performance).
‘Recovery’ in this paper refers to any metallurgical performance parameter (eg ore
hardness, mill throughput, flotation recovery, leach recovery, concentrate grade,
reagent consumption etc).
Mineralogy exerts the primary control on metallurgical performance. Recovery
assumptions, based on a lack of data, usually prove to be misleading and imprecise.
These points are obvious, but we (as an industry) seem to be constantly trying to
convince ourselves that the above do not apply to our deposits. Think about the
revenue impacts and opportunity losses when we get it wrong.
The failure (or under performance) rate for mining projects remains relatively high.
More importantly, the failure rates have remained relatively unchanged for decades
(McCarthy, 2003). Based on numerous surveys to identify the reasons behind new
mining projects and brownfield expansions not delivering to forecast, McCarthy (2003)
identified the following feasibility study problem areas which lead to commissioning
and operational under performance:
•• mine design and scheduling (32 per cent frequency)
•• geology, resource and reserve estimation (17 per cent)
•• metallurgical test work, sampling and scale-up (15 per cent)
•• process plant equipment design and selection (12 per cent)
•• geotechnical analysis (9 per cent)
•• cost estimation (7 per cent)
•• hydrology (4 per cent).
The driver behind geometallurgy at Olympic Dam is to reduce the technical risk
to the current operations and future expansions caused by unexpected mining and
process performance issues due to variable ore properties.
1. MAusIMM, Principal Geometallurgist,
BHP Billiton Olympic Dam, Adelaide SA 5000.
Email: kathy.ehrig@bhpbilliton.com
The primary objective of Olympic Dam geometallurgy is to develop metallurgical
performance predictors that reliably describe the process performance of different
ore types and spatially distribute these into the resource block model for use as a
fundamental input into mine planning. The geomet models are predictive mineralogy
and recovery models which are applied to blocks in the mineral resource model to
33
K Ehrig
enable the estimation of mineralogy, metallurgical recovery,
and real value on a block-per-block basis.
The secondary objectives are to identify any ores which may
be ‘problematic’ to the current plant, and provide variability
data of suitable quality for process plant design and future
plant optimisations. And finally, the chemical, mineralogical,
physical property, and ‘recovery models’, collectively called
the geometallurgical model, provide the data required to
support the JORC Code (2012) ‘modifying factors’.
GEOMETALLURGY FUNDAMENTALS
There has been a rapid increase in the volume of published
geometallurgy literature over the past ten years. Numerous
papers are available which describe the definition, purpose,
benefits, and methodologies used in geometallurgy from the
perspectives of geologists, metallurgists, and process design
engineers. A few of these papers are listed with the references.
This section provides highlights regarding the links between
minerals and populating the resource block model and the
mine plan with geometallurgical data.
Minerals
Metals and elements occur in ore deposits as minerals. Minerals,
not elements or metals, are mined. Extractive metallurgy,
in particular, mineral processing, hydrometallurgy, and
pyrometallurgy extracts metals and elements from minerals.
Grades or concentrations of metals and elements have
been extensively used as proxies for mineral abundances.
Historically, it has been easier and more cost-effective to
assay a sample for a limited suite of elements than it is to
measure mineral abundances. However, over the past decade
relatively rapid mineralogy detection and measurement
methods such as automated scanning electron microscopy
(eg MLA, QEMSCAN etc) and short wave infrared (eg PIMA,
HyLogger, CoreScan etc) methods have become more readily
available thus permitting the measurement of mineral
abundances at an unprecedented scale.
In addition to the recovery relationship described above,
another critical yet useful relationship is:
mineral (wt per cent) = f (sample composition)
The relatively low cost of obtaining multiple element assays
via inductively coupled plasma optical emission (ICP-OES)
and inductively coupled plasma mass spectroscopy (ICP-MS)
methods opens the path to estimate mineral abundances on a
sample-by-sample basis based on multielement assays using
the above relationship. Obviously, the above relationship
needs to be calibrated against measured mineralogy.
If the recovery and mineral (wt per cent) relationships for
your deposit can be established and the distribution of the
process critical minerals can be quantified and mapped, then
a logical conclusion is that metallurgical performance can be
predicted across a deposit.
The geological database, mineral resource
block model and mine plan
Our understanding of the mineral deposit comes from the
systematic sampling provided primarily via drilling. In
general, as the sampling density increases, the confidence
level also increases. The geological database (includes
drilling, assaying, rock mass properties etc) is very likely to be
the largest data set (~100 000 to >2 million assayed drill core
samples) available to a project or operation. Typical resource
block models contain geostatistically estimated concentrations
34
of economic metals and deleterious elements along with
some minerals. The next level of ore characterisation beyond
assaying and geological/geotechnical/geophysical data is
the acquisition of quantitative mineralogy and metallurgical
performance parameters collected at a sufficient spatial
frequency to be incorporated into various block models. If
you are incredibly fortunate, the number of quantitative
mineralogy and metallurgical samples is only one to two
and two to three orders of magnitude less than the geological
assay samples, respectively.
The geostatistical techniques to estimate mineral
abundances into the space between the drill holes (typically
20–200 m spacing) or samples are straightforward because
mineral abundances are additive. However, this may, or
may not, be the case for metallurgical parameters. One way
around the potential problem of non-additivity is to develop
mathematical relationships which express each metallurgical
parameter as a function of additive properties such as metal
grades and mineral abundances. This has been the approach
taken to populate every block in the Olympic Dam resource
model (~20 million blocks) with abundances of 15 minerals
and >50 metallurgical performance parameters, in addition to
26 elements and metals and bulk density. This allows for the
assessment of a block’s real value to recover metal. To achieve
maximum benefit from your valuable geometallurgical data,
it must be incorporated into the mineral resource block model
because the resource block model is the primary input into
the mine plan.
Mine planning transforms the block model (eg 3D
orebody information) into a time-based 1D ‘metallurgical or
process engineer friendly form’ suitable for process design
and production planning. A well designed mine plan is
produced via an iterative process involving inputs from
the geology, geometallurgy, geotechnical, hydrology, mine
design, metallurgy, infrastructure, environmental, business
valuation, and financial teams.
Once a mine plan (there will be a continual stream of mine
plans) is available, where each block in the resource model
is tagged with a mining period, cumulative distribution
curves can be generated for any variable in the block model.
The curves can be produced to represent any mining period
and can also be filtered based on any variable in the block
model. This is a very powerful tool and provides the process
design engineer with some of the data necessary to design the
processing plant.
THE FIT-FOR-PURPOSE GEOMETALLURGY PROGRAM
In describing the fundamental aspect of geometallurgy, Sola
and Harbort (2012) quoted the following which was originally
sourced from Harry and Schroeder (2000):
You don’t know what you don’t know
You don’t measure what you don’t value
You can’t value what you don’t measure
If you can’t measure it you can’t control it
If you can’t control it you can’t improve it.
Several questions need to be considered when scoping and
then addressed in a geometallurgy program:
•• What, if any, are the technical confidence levels required
by your project funding institution? What are the
measures of the confidence levels?
•• Is the project time-constrained or resource-constrained?
•• What learnings (ie free lessons) are available from
similar deposit styles (or other deposit styles with
similar mineralogy) in the district, region, or globally?
We are metallurgists, not magicians
Geometallurgy – what do you really need to know from exploration through to production?
•• What are the market specifications for the final
products? What are the penalty elements/minerals and
their limits?
•• What are the payable elements/minerals
subeconomic elements/minerals?
and
•• Which minerals and concentration ranges are deleterious
to the process?
change with time. Waste today may become ore tomorrow.
Metallurgical performance is almost always impacted by
variations in mineralogy and grade. The number of samples
and variety of ore types required for testing will likely
decrease as the project advances into production where the
focus shifts to better understanding and managing local
(production) scale variability.
•• How do the process and deleterious minerals/elements
occur in the mineral deposit?
Model validation
•• How does mineral and textural variability across
the known extent of the mineral resource impact on
metallurgical performance?
Geometallurgy models should be treated with scepticism
until they are shown to be useful. Early in the geometallurgy
program, initially developed metallurgical performance
predictors can be used to predict the outcomes of future
laboratory testing outcomes. Model validation and model
updates should evolve as testing continues during the various
studies. The initial models can be used to predict the outcomes
of future laboratory testing. Full scale production is the
real validation of the metallurgical performance predictors.
During production, there will be repeated opportunities and
business requirements to improve the precision of the geomet
models so that they are useful for short-term production
planning!
Resource understanding and project evolution
From exploration into various project development stages
through to the end of production, resource classification
terminology such as inferred, indicated, and measured is an
attempt to demonstrate the level of geological knowledge and
confidence in the resource. This knowledge and confidence
should be increasing as the drilling metres increase. An
analogy in the engineering studies space is the increasing level
of cost accuracy required when progressing from conceptual
to prefeasibility (PFS) and feasibility studies through to
construction, commissioning and finally into production.
In an ideal world, resource confidence should be sufficient
to support the next stage of project development. However,
in the real world, drilling to support geological (and
geometallurgical) knowledge and confidence in the resource
to a level suitable for a PFS usually starts at the same time
as the PFS for mine planning and process design. The same
is true for the feasibility study, and even into production.
More simply put, the entire project team commences their
individual studies at the same time. The impact on the project
is that the mine plan and cut-off grade strategy are likely to
continually change as the geologists continue to discover and
improve the definition of ore zones within the deposit and the
geomet team continues to identify more potential constraints
or new revenue streams. Hence new ore types are likely to
be discovered, and more importantly the perceived relative
proportions of the ore types will also continue to change. The
impact on process design is that the ‘goal-posts’ will continue
to move during the engineering studies and a process plant is
often built which cannot meet project expectations.
Geometallurgy sampling strategy
Geological and resource ore types and domains may not
necessarily equate to equivalent metallurgical ore types
and domains. The metallurgist or process engineer needs to
understand that geological ore types are based on qualitative
visual estimates of mineral abundances, or as in the case of
alteration zones, simply the presence or absence of a specific
mineral or groups of minerals. Most geological descriptions
of host lithology, alteration, and mineralisation zones are for
the purposes of understanding the genesis of the ore deposit,
not defining metallurgical ore types. Resource domains are
usually linked to geological boundaries, but may be further
refined based on metal distribution patterns within an ore or
alteration type. However, when commencing sampling for
geometallurgical testing, the geology and resource ore types
and domains are all that is available.
Initially geometallurgical sampling must be designed to gain
a deposit scale understanding of the resource, so all mineral
combinations across the grade spectra require sampling and
testing. Remember that the classification of waste, marginal,
low, and high-grade ore is based on economic criteria which
We are metallurgists, not magicians
‘All models are wrong, but some are useful…’. (Box, 1979)
Data integrity and security
Spreadsheets should not be allowed for data storage solutions.
Spreadsheets can be used to analyse the data. Properly
designed and managed databases protect the integrity of the
data and also allow for full integration and interrogation of all
the technical data.
CONCLUSIONS
Back to the title of this paper, ‘Geometallurgy – what do you
really need to know from exploration through to production?’
•• mineralogy, mineralogy, mineralogy
•• how the mineralogy responds to likely extraction
processes across all grade ranges
•• the concentration of process and final product deleterious
elements and minerals
•• how the mineralogy changes, hence how the
metallurgical performance changes over the life of the
asset.
Remember that the primary purpose of geometallurgy
is to develop metallurgical performance predictors and to
incorporate expected performance into the block model
and mine plan. Geometallurgy is not process flow sheet
development, this still requires traditional metallurgical
testing programs.
REFERENCES
Bojcevski, D, 2004. Metallurgical characterisation of George Fisher
mesotextures and microtextures, MSc thesis (unpublished),
University of Queensland, Brisbane.
Box, G, 1979. Robustness in the strategy of scientific model building,
in Robustness in Statistics (eds: R L Launer and G N Wilkison),
312 p (Academic Press).
Harry, M and Schroeder, R, 2000. Six Sigma: The Breakthrough Strategy
Revolutionizing the World’s Top Corporation (Doubleday: New
York).
JORC Code, 2012. Australasian Code for Reporting of Exploration
Results, Mineral Resources and Ore Reserves (The JORC Code)
[online]. Available from: <http://www.jorc.org> (The Joint Ore
Reserves Committee of The Australasian Institute of Mining and
Metallurgy, Australian Institute of Geoscientists and Minerals
Council of Australia).
35
K Ehrig
McCarthy, P, 2003. Managing technical risk for mine feasibility
studies, in Proceedings Mining Risk Management Conference,
pp 21–27 (The Australasian Institute and Mining and Metallurgy:
Melbourne).
36
Sola, C and Harbort, G, 2012. Geometallurgy – tricks, traps and
treasures, in Proceedings 11th AusIMM Mill Operators’ Conference
2012, pp 187–196 (The Australasian Institute and Mining and
Metallurgy: Melbourne).
We are metallurgists, not magicians
Contents
Integrating geometallurgy with copper
concentrator design and operation
G Harbort1, K Jones2, D Morgan3 and C Sola4
ABSTRACT
The use of geometallurgical modelling with flotability component simulation provides
a design methodology with significantly less associated risk. The use of geological
data for optimisation of operating plants has become a significant part of the modern
process mindset. The underlying principle is to use spatial metallurgical information
to drive production planning, mine planning, blast design, blending strategies
and plant set-up. At a design stage the process designer can use geometallurgical
information to evaluate bottlenecks and potential design flaws and propose the best
investment strategy for the benefit of the project.
This paper provides a brief review of various copper sites that have implemented
geometallurgical studies. Two case studies are also presented. The first details the
geometallurgical characterisation of the Andash deposit and methodology used to
review the project’s detail design and projected production. The second discusses the
geometallurgical approach undertaken at the operating Northparkes mine, both for
circuit optimisation and life-of-mine planning.
INTRODUCTION
Early evolution of geometallurgy
Geometallurgy is considered a relatively recent development. It has been suggested
(Williams, 2013) that it evolved in the late 1980s or early 1990s. In reality, geometallurgy,
both as a practice and a science is far older. The practice of copper geometallurgy can
be dated to the first decade of the twentieth century. Dwindling stocks of high-grade
ore for direct smelting or gravity concentration, coupled with the discovery of massive,
low-grade Chilean porphyries were key drivers for its development (Yeatman, 1932).
This led to increased cooperation between geologists and metallurgists to achieve
economic extraction of copper from low-grade ores. Although practised by numerous
porphyry copper mines in an ad hoc manner, geometallurgy was not to gain favour
as a science for several decades.
Its initial academic development was to take the form of a collaborative research
project between the University of Melbourne and the University of Queensland
(O’Malley and McGhie, 1939). This investigation centred on the mineralogy of the
Black Star orebody at Mount Isa and its implications for mill practice. At the time it
was considered outstanding ‘…in focussing upon the mind of the reader, in a lucid
and emphatic manner, the metallurgical problems that result from complex sulphide
intergrowths …’ (Blanchard and Hall, 1939). Not only did the work correctly predict
metallurgical operation decades into the future, it also produced one of the first theses
on geometallurgy.
1. FAusIMM(CP), Technical Director, Amec Foster
Wheeler Australia East, Brisbane Qld 4000.
Email: greg.harbort@amecfosterwheeler.com
2. Senior Analytical and Evaluation Metallurgist,
Northparkes Mines, Parkes NSW 2870.
Email: kellie.jones@northparkes.com
3. MAusIMM, Technical Superintendent,
Northparkes Mines, Parkes NSW 2870.
Email: dylan.morgan@northparkes.com
4. Process Engineer – Geometallurgy,
Amec Foster Wheeler Australia East,
Brisbane Qld 4000.
By the 1950s geometallurgy was becoming an established concept. The first
symposium on geometallurgy was held in Salt Lake City in 1955 (Anon, 1956).
Discussions were very similar to those held at modern geometallurgy conferences.
Key issues were considered to be the need for greater interdisciplinary communication
and research (McQuiston, 1956; Kirkland, 1956; Reno Sales, 1956).
In the same era geometallurgical concepts played a crucial role in the formation of
resource reporting in the Soviet Union. Raw commodities were considered strategic
assets, with their treatment to provide maximum national benefit (Arden and Tverdov,
2014). Soviet geometallurgy, or technico economicheskiye obosnovaniye (technical
economic characterisation), included factors such as mineral complexity, treatment
options, product quality and economic performance (Tverdov and Mikishichev, 2014).
During this period detailed geometallurgical analysis at operating mine sites was
often constrained by a lack of resources. To quote Mount Isa Mines project metallurgist
A E O’Meara (1953):
37
G Harbort et al
The mill research metallurgical staff has endeavoured to
incorporate mineragraphy in solving mill problems and
determining the effectiveness of laboratory tests, as well as
examining minerals sent in by the geology department. For
many years the amount of equipment available was limited
and much of it locally made or improvised.
This would change in 1950 when a fully equipped
microscopy laboratory commenced operation at Mount Isa
Mines. Geologists and metallurgists shared offices, evaluating
the effects of comminution on mineral liberation and flotation
performance. Another example of geometallurgical success
was the Brunswick concentrator program detailed by Petruk
and Schnarr (1981).
In 1968 Quiston and Beachaud published their
comprehensive review of sampling and testing a virgin
deposit from a metallurgical perspective. They proposed
‘geometallurgy’ as a term to describe this perspective:
… since geology is inextricably interwoven with metallurgy
in gaining an understanding of the complexities of a deposit,
eventually leading to a definition of mineable reserves, with
the development of a flowsheet and engineering criteria for
the planning of a successful and profitable operation.
Further publications of note included the use of
geometallurgy for comminution (MacPherson, 1976) and
plant design (Young, 1983).
The appearance of automated mineralogy and the vast
amount of information it supplied was to have a major
effect on geometallurgical development. In the 1970s various
groups in Canada (Petruk, 1984), the United Kingdom (Jones,
1984), South Africa (Oosthuyzen, 1983) and Australia (Reid
and Zuiderwyk, 1983; Reid et al, 1984; Reid and Wittenberg,
1984) began to develop optical and electron beam methods
of automating the collection of point counting data. This
led to the development of the QEM*SEM, later renamed
QEMSCAN, and the Mineral Liberation Analyser (MLA),
automated instruments that allowed the rapid identification
and measurement of a desired number of mineral grains, drill
core or ore sample sections.
The modern era of geometallurgy
The modern era of geometallurgy can be considered to have
commenced circa 1985, where automated mineralogy and
increasing computer availability laid the groundwork for
it to become a common practice. QEMSCAN and later MLA
units were installed in commercial laboratories around the
world in addition to research departments in major mining
companies. It has become commonplace not only to interpret
the performance of operating plants by use of QEMSCAN or
MLA measurements but also to assess orebody composition,
including block models, and to predict plant performance from
drill core and ore samples. In this process, not only the mineral
abundances, but the nature of the gangue minerals, the mineral
associations and the natural grain sizes are all important
in predicting grind size, the design of flotation circuits and
the likely recoveries and losses. Relationships between ore
characteristics and concentrate grade and recovery were
obtained for ‘best practice’ of the day (Jackson, Gottlieb and
Sutherland, 1988). A further outcome of the QEMSCAN and
MLA interpretive capabilities was the reliable prediction from
studies of feedstock for grind size and liberation (Gottlieb,
Adair and Wilkie, 1994). This type of study was used by the
Hellyer mine in Tasmania to build up a mine model related
to plant performance and the creation of stockpiles based on
ore type (Lane and Richmond, 1993). Used at an early stage
in the mining life cycle the resultant data allows prediction of
38
processing behaviour and processing properties for stockpiling,
blending and block models (Fennel et al, 2005).
In the last ten years, there has been a flurry of activity
relating to geometallurgy. Excellent reviews of geometallurgy
methodologies have been published by Dunham and Vann
(2007), Walters (2008), Coward et al (2009), Williams (2013),
Dominy and O’Connor (2016) and McKay et al (2016). The
assigning of geometallurgical domains and their importance
has also been detailed by a number of authors, including David
(2007), Johnson and Munro (2008), Hunt, Berry and Bradshaw
(2011) and Sola and Harbort (2012). A useful benchmarking
of geometallurgy programs by type and depth of usage was
discussed by Lund and Lamberg (2014) and Lishchuk et al
(2015). Other areas of interest have included comminution
geometallurgy (Alruiz et al, 2009; Suazo, Kracht and Alruiz,
2010; Mwanga, Rosenkranz and Lamberg, 2015) and process
mineralogy (Evans et al, 2011; Kuhar et al, 2013). Although
most of the recent publications focus on geometallurgical
successes, a number also discuss potential problems (Kittler
et al, 2011; Sola and Harbort, 2012).
Copper geometallurgy
Australia
Within Australia, copper geometallurgy was initially
centred at Mount Isa Mines. Hoffmann (1964) reviews the
development of flotation treatment of Mount Isa’s chalcopyrite
ore between 1941 and 1963. Part of the development process
included microscopic examination of ore to reveal features
of metallurgical significance. Subsequent test programs
considered these features in terms of reagent addition and
flow sheet design. During operation, liberation studies were
also conducted to allow optimisation of concentrate grade
and copper recovery.
Geometallurgical evaluations have been conducted on
Australia’s iron oxide copper-gold (IOCG) deposits. Expansive
geometallurgy programs were conducted both for the
design of the Ernest Henry copper concentrator (Strohmayr
et al, 1998) and subsequent optimisation (Tew et al, 2003).
The former studies involved core mineralogy and flotation
tests on 42 drill holes. Although the drill holes represented
a good cross-section of the orebody, it was reported that
their compositing resulted in performance prediction issues
due to underestimating the microvariability of the orebody.
Latter investigations during operation used monthly modal
composites to calibrate and predict performance. Each mineral
was given a predicted flotation response in terms of grade
and recovery. In addition, samples were composited from
each mining block, with X-ray diffraction (XRD) conducted
and correlated with plant performance. Investigation showed
that feldspar and albite had a negative impact on throughput,
while quartz and magnetite had a positive influence.
For the Prominent Hill IOCG deposit, geometallurgy was
used for optimising mining block models, plant design
and operations trouble shooting. Khosrowshahi, Shaw and
McKevitt (2009) discuss the use of transfer functions to
assist in mine plan optimisation including copper speciation,
gangue mineralogy, contaminants, processing parameters
and concentrate quality. Barns, Colbert and Munro (2009)
discuss the mineralogy approach used for geometallurgical
domaining, test work and flow sheet development. Bradshaw
et al (2012) consider geometallurgy in their investigation of
variable performance when treating different ore blends.
Perhaps the most comprehensive IOCG geometallurgical
study has been that for Olympic Dam. The deposit has over
70 ore and gangue minerals, each with associated metallurgical
WE are metallurgists, not magicians
Integrating geometallurgy with copper concentrator design and operation
properties. In the initial comprehensive geometallurgy
program 500 to 1000 samples were used for full metallurgical
characterisation, with approximately 10 000 samples for
mineralogical characterisation (Liebezeit et al, 2011; Boisvert,
Rossi and Ehrig, 2013). Olympic Dam Operations maintain a
geometallurgy program to provide technical support to both
the mining and processing departments. Metallurgical and
mineralogical assessment is conducted on drill core within
five-year production plans. Predictive mineralogy algorithms
are used to estimate throughput, copper recovery and
concentrate grade (Nzama and Kapo, 2014).
Although the geometallurgy programs for the larger
producing Australian copper mines have received significant
recognition, much work has also been done on the smaller
polymetallic copper operations. A number of these relate to
mines within the Cobar Basin, New South Wales. The basin
is characterised by metal assemblages dominated by Pb-ZnAg ± Cu at the Endeavour Mine, Cu-Zn-Pb-Ag at the CSA Mine
and Cu-Ag at the Peak Mine. Loidl (2012) published a thesis
on the geometallurgy of Endeavour. The comprehensive study
included analysis of lithology and alteration, mineralogy and
petrography, grain size distributions and liberation and their
implication for processing. The CSA mine has transitioned
from being a lead and zinc producer to a copper producer. This
required extensive changes to the design of the grinding and
flotation circuits. Ongoing optimisation has utilised scanning
electron microscope analysis (James and Scamardella, 2000).
The Peak mine treats ore blends from five different deposits,
with significant challenges due to the inherent variability
between each of the orebodies (Hartog et al, 2014; Taylor, 2011).
Other polymetallic copper geometallurgy programs of note
included those of Hellyer mine (Richmond and Campbell,
1992) and the Thalanga operations (Gregory, Hartley and
Wills, 1987). At Hellyer core logging and mapping recorded
geological and geotechnical data which was considered of
importance to mining, metallurgy and exploration from the
onset. By selectively extracting various textures from the
mineralisation data set it was possible to construct plans
and sections showing the distribution of textural zones and
thus predict the ore type for each stope. These aided mine
planning and production scheduling. Ore types which were
difficult to treat were stockpiled separately and then blended
with other ore types to optimise metallurgical performance
(Downs, 1990). Diamond drill core was subjected to extensive
mineralogy to determine implications for possible processing
routes. The samples were further examined for evidence of
ore zones or mineralogical differences between different areas
of the orebody. This allowed the orebody to be classified into
several categories on the basis of texture with various flow
sheets tested for design purposes (Richmond and Lai, 1988).
The Thalanga operations provide an example where even
detailed geometallurgical analysis may not achieve targeted
performance. The primary ore zone consisted of massive
and semi-massive sulfides containing copper, lead, zinc and
silver, as well as gold mineralisation. Three supergene ore
types were designated originally according to the proportions
of the valuable sulfide minerals. Liberation data and test
work allowed a flow sheet to be developed for the recovery
of separate copper, lead and zinc concentrates in a threestage sequential flotation process. The major difficulty in
treating the Thalanga supergene ores was the separation of
copper sulfides and sphalerite. The poor flotation selectivity
existed in all three types of supergene. It appeared that in situ
activation of sphalerite has taken place during the process
of supergene alteration of the Central Thalanga deposit.
This was confirmed by the presence of copper sulfate in the
supergene ores and evidence of copper sulfate precipitation
WE are metallurgists, not magicians
occurring during stockpiling of the supergene ores after
mining (Wong et al, 1991). The same situation reoccurred
with treatment of the Reward deposit. Laboratory test work
found that there was deterioration in flotation performance
after only five days of ore aging and that this continued to
adversely affect performance the longer the aging. The
strategy to reduce this oxidation and subsequent deleterious
effects to plant performance was to mine at a rate consistent
with plant throughput, minimising stockpiles at both the mine
and plant. This strategy to minimise stockpiles impacted on
plant performance, as blending of ore for mixing of ore types
and to minimise head grade fluctuations was not possible.
Recovery loss due to the higher Fe:Cu ratios was investigated
by conducting mineralogy on flotation tails samples by
X-ray diffraction and electron microscope techniques. The
investigation found that there was no obvious mineralogical
reason for the difference in flotation response between the
high Fe and low Fe samples (Kilgariff, 2003).
South America
Yeatman in 1932 discussed two early examples of copper
geometallurgy from South America. The first was at
Chuquicamata in 1911–1912. Although drilling had identified
a large deposit of copper ore the mineralisation in the form of
brochantite was something entirely new in large commercial
copper deposits and presented new problems in ore treatment.
It was found that the oxidised ore could be leached by
sulfuric acid, a treatment method held with suspicion at the
time. Laboratory tests were conducted on 100 t of ore taken
from different sections of the deposit to determine solubility,
amount of sulfuric acid available and percentages of chlorine,
nitrate, iron etc. Construction of a 10 000 t/d copper leach
/ precipitation plant was started in 1913, with operation
commencing in May 1915. The second case was that of the
El Teniente deposit. The original 400 t/d gravity plant only
achieved 65 per cent recovery with a feed grade of 2.5 per cent
Cu, principally as chalcopyrite and, to a limited extent, bornite
and chalcocite. Samples were dispatched to London for testing
with the new minerals separation flotation technology. The
initial tests were not successful with investigation showing
that a considerable portion of the copper in the samples had
oxidised during transit. This led to one of the first protocols
for preservation of samples between collection and testing.
Additional fresh and protected samples of ore were sent to
London, with tests achieving a copper recovery of 85 per cent
to 90 per cent. This evaluation would lead to the installation
of flotation and form a basis for all large porphyry properties
that would follow.
South American copper geometallurgy has typically focused
on porphyry deposits, with CODELCO having a significant
role. At the Chuquicamata processing plant recovery was
modelled as a function of geomining and metallurgical data
and ore characteristics obtained from a historical database
(Compan, Pizarro and Videla, 2015). Operational data on
mill feed grades, ore hardness, particle size, mineralogy, pH
and reagents, representing several months of operation, was
collected and using multivariate regression techniques was
used to predict recovery. The recovery model was validated
using monthly plant data between January and July 2014.
The model prediction shows a correlation coefficient of
89.7 per cent and a mean absolute error of 2.75 per cent. Other
major geometallurgical studies by CODELCO were reviewed
by Beniscelli (2011) and included El Teniente, Radomiro Tomic
and Ministro Hales. The Radomiro Tomic study commenced in
1996. Mineral samples were analysed for total Cu (CuT), acidsoluble Cu (CuS), total Fe, Cl, S, Mn, Mg, Al, P, Ca, Na, F and
SiO2. Also, pH leach and acid dosage tests carried out prior to
39
G Harbort et al
a bench scale column leaching characterised the metallurgical
behaviour of two geological units. At El Teniente, a study was
conducted based on geological mapping information, optical
mineralogy and metallurgical data from 731 samples. This
allowed the generation of a model relating rock texture and
grindability for the main igneous lithologies. For the Ministro
Hales project geometallurgical characterisation allowed the
identification of key process issues in terms of mineralogical
species and major and minor chemical elements. These issues
included low copper recovery in the mixed minerals zone and
high final concentrate quality variability. In addition, Carrasco
et al (2005) presented the results of several heterogeneity
and Ingamells’ tests done in CODELCO deposits including
Chuquicamata, Mina Sur, Radomiro Tomic, El Salvador,
Andina and El Teniente. Tests considering the natural
variability were used to develop accurate sampling protocols
for further geometallurgical analysis.
At the Minera Escondida concentrator metallurgical issues
became apparent soon after start-up of the first phase of
expansion:
The orebody was proving to be more variable than expected
and the limitations of predicting behavior from the widely
spaced original diamond drill holes were apparent.
Additional drilling programs were initiated, and a joint
mining and metallurgy program of oreblock testing was
commenced, to determine mineralogical characteristics
and grinding and flotation response. This data was used
to schedule the mining sequence and allow blending, so as
to avoid peaks in hardness and clay content, however ore
exposure was often limited which meant high variations in
mill feed characteristics. (Kilgour, 1995)
The Sierra Gorda deposit was identified as having three
main zones of sulfide mineralisation. The hypogene zone
consists of chalcopyrite, pyrite, molybdenite and bornite.
Gold typically accompanies the copper sulfides. Atacamite,
brochantite, chrysocolla, vermicullite, lindgrenite, powellite,
ferrimolybdenite are minerals of note in the leached/oxide
zone. The supergene zone contains chalcocite, covelite,
digenite, bornite and chalcopyrite. The original flotation
circuit product recoveries and grades were developed by
Aminpro based on their proprietary simulation computer
model. The final flotation circuit design criteria used in
detailed engineering included modifications based on
benchmarking and additional ore kinetic and variability data
(Comi et al, 2013).
The Candelaria mine implemented a geometallurgical
project to increase the plant throughput by optimising rock
breakage from blasting, crushing and grinding. The ore was
characterised, and drilling, blasting, crushing and grinding
processes audited to develop site-specific models for each
process. Simulations were conducted to identify integrated
operating strategies in the mine and plant to increase mill
throughput. Recommended changes were implemented,
achieving throughput increases of ten to 20 per cent,
depending on the ore hardness (Muñoz et al, 2008).
A number of geometallurgical studies have been published
for Peruvian operations. At the Collahuasi operations studies
have been conducted on both comminution (Alruiz et al, 2009)
and flotation (Suazo, Kracht and Alruiz, 2010). The flotation
geometallurgy model evaluated flotation as a function of air
dispersion properties, the feed particle size distribution and
introducing a parameter (Ф), which represented the inherent
geometallurgical floatability of the ore. At industrial scale
the model was able to predict metallurgical results in a time
frame of several weeks with an average relative error of less
than two per cent.
40
Baumgartner, Gomez and Escobar (2016) discuss the
mineralogical characterisation and implications at the Cerro
Corona Cu-Au porphyry mine. The operation uses a suite of
geometallurgical tools including quantitative mineralogical
analysis (modal mineralogy, grain size liberation and
association), semi-quantitative XRD and rougher laboratory
flotation tests to determine clay types and their effect on
performance.
At Toquepala the orebody was characterised according
to lithology, mineralisation and alteration. The use of size
by mineral mineralogy allows ongoing optimisation of the
operating plant (Quiñones and Mattos, 2001).
During the design phase for the La Constancia project
(Klohn, Stephenson and Granados, 2016) a geometallurgical
study was undertaken to evaluate comminution and flotation
options. The study identified dendritic growths of sphalerite
and chalcopyrite into pyrite, the presence of copper oxides
in the supergene zone with associated recovery implications
and the copper activation of sphalerite both in situ and during
blending of supergene and skarn ores (Greig et al, 2009). The
initial geometallurgical model estimated copper recoveries
and concentrate grade and quality over two week periods
for the mine life to assist in mine planning and metallurgical
strategies.
Antamina is a complex, highly variable polymetallic
deposit. Ore is treated by campaign, where one of eight ore
types is treated over periods lasting from two to 30 days. Each
ore type has distinct mineralogical characteristics resulting
in variable metallurgical performance and treatment. In
one geometallurgical study, an assessment of mineralogical
deportment and texture led to effective penalty element
strategies being developed by metallurgical staff (Kormos
et al, 2010).
In Argentina, at the Minera Alumbrera operation,
geometallurgical studies commenced approximately ten
years before project development (Matar et al, 1986). They
continued through plant design (Keran et al, 1998) and into
operating philosophy (Harbort et al, 2000). The Instituto de
Investigaciones Mineras conducted an integrated study on
the deposit which included geological, geochemical and
topographic modelling, geostatistical reserve estimation,
mineralogical behaviour and prefeasibility studies. During the
design phase mineralogy was conducted on both bench and
pilot plant samples to determine optimum design. Operation
has used both monthly and survey specific modal mineralogy
and liberation data to optimise performance.
The Technological Characterisation Laboratory in the
Department of Mining and Petroleum Engineering –
University of Sao Paulo conducted a significant geometallurgy
program for the Sossego copper mine in Brazil (Kahn et al,
2014). The study included classifying samples by X-ray cluster
analysis and detailed automated mineralogy. The study by
Bergerman et al (2008) evaluated SAG mill variability via
ore characterisation, JKSimMet comminution modelling and
blast design methodology. Samples from both core and plant
feed were submitted for Drop Weight tests. Core representing
two years of the mine plan, along with plant samples were
taken as part of comminution surveys. The study allowed
modifications to be progressively implemented into both
blasting patterns and the comminution circuit, with good
correspondence to simulated results.
Fonseca and Sá (2005) published results from a
geometallurgical study on the Alemão IOCG deposit, Carajás,
Brazil. The basis of the program was a liberation study on
coarse and fine particles correlated with the performance of a
mini pilot plant. The results were used to indicate the primary
WE are metallurgists, not magicians
Integrating geometallurgy with copper concentrator design and operation
target size for grinding and to predict copper recovery from
the flotation rougher stage. The liberation by grade class was
used to indicate the particle size for the regrinding stage, prior
to flotation cleaning stages.
North America
The Colorado geometallurgy symposium gives an indication
of the significant focus on copper porphyry geometallurgy
in North America during the 1950s. A number of presenters
stressed the importance of a geometallurgical approach from
initial deposit discovery through exploration and into mine
planning and product (Clemmer, 1956; Richard, 1956). In
addition, Kildale (1956) discussed the more complex Cu-PbZn deposits, providing examples of the effect of mineral grain
size on both flotation and comminution performance.
The Cyprus Pima orebody in Nevada contained
chalcopyrite, bornite, native copper, chrysocolla and cuprite.
In the first year of operation, a geometallurgical program was
run in an effort to determine the best operating conditions on
Cyprus Pima ore. Seven operating strategies were developed
for the seven ore conditions that commonly occurred. Two
of these ore conditions had variations depending on the
tarnishing of sulfide particles. The ore types corresponded
to specific areas of the pit, allowing selective mining and
treatment (Ramsey, 1976).
Geometallurgy at the Lavender Hill concentrator was driven
by the complex copper mineralogy of the orebody. Chalcocite
occurred as rims around pyrite grains and as minute veinlets
along fractures within the pyrite grains, many contained a
network of chalcocite veinlets. Microscopic examination of
polished sections showed the chalcocite veinlets to be 10 µm
thick, or less. Oxide copper formed very thin coatings on the
chalcocite, barely perceptible under the microscope. Oxidation
took place rapidly after the ore was broken, particularly when it
was hot and humid. There was little oxidation during crushing,
but significant oxidation during grinding (Martin, 1957).
Geometallurgy at the Anaconda C E Weed concentrator was
instigated due to variability in the Berkeley Pit. The pit was
a very non-homogeneous mine with respect to ore type and
mineralisation. An understanding of the geology and mine
led to separate slimes and sands circuits and strategies for the
treatment of ‘sweet’, ‘sour’ and ‘hard’ ores (Palagi and Stillar,
1976; Wraith and Fulmor, 1964).
A more recent geometallurgy study is that of Bingham
Canyon which was designed as a proof-of-concept test
to develop a method of quantifying key geometallurgical
properties. The deposit was divided into 33 ore domains
by mine geologists and metallurgists on the basis of
lithology, alteration assemblages, fracture density, copper
mineralisation and metallurgical properties. This allowed a
combination of geological and mineralogical characteristics
to be determined that, when combined, impacted processing
(Ross et al, 2009).
operations was undertaken. An extensive laboratory program
was developed to assess the metallurgical response of the
various ores produced from the underground mine. In some
cases of extremely poor performance some ore types were
removed from ore reserves. In other cases, the process had to
be modified to optimise circuit response to the particular ore
feed. A considerable amount of mineralogical and microscopic
work was completed, indicating the need for better liberation.
In addition to the metallurgical laboratory work, pilot plant
campaigns were undertaken to confirm the conclusions
reached from bench scale studies (Urbanoski, 1993).
Kidd mine conducted detailed geometallurgical analysis
in 2010–2011 to determine optimum plant performance. The
analysis showed that the net revenue impact of copper recovery
improvement was neutral while zinc revenue decreased due
to the increased losses to the copper concentrate. The key
revenue driver for the project was found to be silver due to
higher than anticipated recoveries and prices (Leggett and
Morin, 2013).
The complex Laronde ores require a series of grinding,
copper/lead flotation and separation, zinc flotation, zinc
tails precious metals leaching, followed by a countercurrent decantation circuit and Merrill Crowe precipitation.
The plant treats a range of ore types requiring multiple
metallurgical processes and optimisations, as well as
operational and metallurgical philosophies that have
changed to adapt to custom centralised milling (Blatter,
Cayouette and Cousin, 2011).
Metallurgical results indicated slight differences when
ore from the Matagami Lake Cu-Zn mine was ground with
different grinding media. The grinds are compared on
the basis of size analyses and of the free pyrite, sphalerite,
chalcopyrite and galena grains in the flotation feeds. The
sizes of the free grains in all flotation feeds were different for
each mineral, decreasing in the approximate order pyritesphalerite-chalcopyrite-galena. This order corresponds to the
order of decreasing hardness for the minerals. In addition the
grain sizes for specific minerals were different when different
grinding media was used (Petruk and Hughson, 1977).
The Highland Valley Copper geometallurgical study
(Mitchell and Holowachuk, 1996) provides a classic example
of throughput optimisation via geometallurgy. The study
involved analysis of historical data and the comparison of
the different intensities of alterations and rock types with
actual milling rates. Milling rates derived from these studies
were further defined by processing many of the ore types
individually and in combination with one another. This
data was supplemented by both historic Drop Weight tests
and future ore tests. The study allowed an effective blending
strategy to be developed with mine operations, engineering
and geology personnel meeting daily to determine specific
blending strategies for each 24 hour period.
The approach used to produce geometallurgical domains at
the Pebble porphyry Cu-Au-Mo deposit in Alaska was to first
characterise representative samples from each geologically
distinct area of the deposit. The samples from each area
were studied using a combination of optical petrography
and automated mineral mapping to define the silicate and
sulfide mineralogy and the copper and gold deportment.
This allowed assessment of changes in gold deportment as a
function of alteration across the deposit (Gregory et al, 2013).
At the Kemess Cu-Au porphyry operation, concentrate
regrinding requirements were evaluated using a
geometallurgical approach. Treatment towards the end of
mine life resulted in a 22 per cent increase in grinding energy,
a seven per cent decrease in copper recovery, a 20 per cent
decrease in gold recovery and a three per cent Cu decrease
in concentrate grade. Mineralogical studies indicated poor
chalcopyrite/pyrite liberation, with finer grinding required
in concentrate regrind (Brissette and Roman, 2012).
Copper geometallurgy in Canada has a significantly higher
focus on polymetallic deposits than the rest of the Americas. In
order to achieve metallurgical improvements at the Brunswick
Cu-Pb-Zn mine a thorough review of the concentrator
The Mount Polley porphyry copper mine makes use of onsite automated mineralogy, both for optimisation and strategic
planning. An initial study compiled daily mineralogy data for
several weeks to generate mineral balances around the circuit.
WE are metallurgists, not magicians
41
G Harbort et al
This allowed strategies for mine planning, forecasting and
operational tuning to be developed (Dobbe et al, 2014).
The Bismark, Sabinas 2 and Tizapas concentrators
in Mexico have produced copper concentrates from
polymetallic deposits. In the last 15 years, emphasis has been
placed on the use of modal mineralogy, mineral association
and liberation characteristics to optimise operations and
metallurgical performance (Magallanes-Hernández and
Espinosa-Gómez, 2005).
Asia Pacific
The complexity of the Ok Tedi orebody resulted in numerous
geometallurgical studies being conducted over its history.
Sulfide porphyry ore consisted of chalcocite, digenite and
chalcopyrite with typical copper recoveries of 90 per cent.
Oxidation led to localised formation of malachite, azurite
and chrysocolla with lower and more variable recovery.
Skarn mineralogy is varied and includes magnetite skarns,
massive sulfides, gossans and calc-silicate skarns. In massive
sulfides and magnetite skarns, pyrite is abundant with copper
present as chalcopyrite, chalcocite, digenite and bornite. The
close association of copper and pyrite reduces both recovery
and concentrate grade. Recovery of copper and gold from
gossanous material was poorest of all the ore types. Specific
reagent regimes are required to achieve optimum performance
for the various ore types. In addition, ore hardness varies with
ore type and degree of weathering. A high level of cooperation
between geology, mining and metallurgy departments
was necessary in order to deliver blends of ore which gave
satisfactory metallurgical performance (England, Kilgour and
Kanau, 1991). Initial electron microprobe analysis of porphyry
samples confirmed the presence of fresh chalcopyrite, bornite
and pyrite and also showed that weathering products
of chalcopyrite vary in composition between chalcocite,
covellite, digenite, geerite and blaubleiblender ‘synthetic’
covellite. Fresh pyrite was found to exhibit natural floatability
(Afenya and Mwaba, 1991). Later work focused on the
mineralogical and metallurgical examination of fluorosilicate
to define and implement strategy for the reduction of
fluorine in concentrates (Pangum et al, 1998). One landmark
geometallurgy study was that of predicting ore aging within
potential new block caves and its effect on metallurgical
performance (Morey and Cantrell, 2011).
At the time of development Batu Hijau was the largest
mining project that has ever been undertaken. Detailed
production planning for the concentrator commenced in the
middle of 1997, two years before production commenced.
The planned production ramp-up was reviewed alongside
actual production experience from more than 30 new mining
projects dating back to the 1970s. The geometallurgical model
for copper recovery was based on lithology, mineralogy,
and depth with reference to 500 bench scale flotation tests
(DeMull, Spenceley and Hickey, 2001). Later work focused on
throughput prediction using a geometallurgy approach. The
deposit’s lithologies, their distributions and association with
mineralisation and geotechnical measurements served as the
basis for ongoing metallurgical studies. This allowed models
of Bond Crushing Work, Ball Work, Rod Work and Abrasion
indices, JKMRC impact breakage resistance and the JKMRC
abrasion resistance to be developed.
The Canatuan Cu-Zn deposit in the Philippines is comprised
of an iron oxide gossan underlain by massive sulfide. The
sulfide occurs as either massive or banded sulfide. The massive
sulfide is composed of >50 per cent coarse-grained pyrite,
chalcopyrite sphalerite, galena and tennantite. The banded
sulfide has lower pyrite content with bornite and covellite.
42
Chalcocite occurs as sooty coatings throughout the massive
sulfides, mainly in the interval immediately underlying the
current water table. This has resulted in complex association
of sulfides in the supergene-enriched zones as well as in areas
that are highly fractured and structurally exposed to water
penetration. A geometallurgical study involving mineralogy,
hydrology, structural geology and extensive metallurgical
tests was undertaken to evaluate the reasons for elevated zinc
content in copper concentrate. This indicated the presence of
in situ zinc and pyrite activation by copper, within zones of the
orebody. This resulted in a reagent scheme which exploited
the varying preferences of copper and zinc in relation to
pulp potential, allowing significantly greater selectivity to be
achieved (Umipig et al, 2012).
The King-king porphyry deposit copper mineralogy consists
of chalcopyrite, bornite, chalcocite, digenite and covellite.
Copper silicates are the most abundant copper minerals in the
oxide zone of the deposit. Gold is relatively abundant in the
oxide zone, with lesser abundance in the sulfides. Native gold
is occasionally observed on fractures and in quartz veinlets.
The King-king metallurgical study was based on an extended
sequential assay procedure to determine copper speciation,
correlated with the results of extensive flotation tests. The
model allowed prediction of copper and gold recoveries
throughout the deposit, based on sequential assay results
(Snider et al, 2013).
At the Sepon operations in Laos, geological logging
uses a set of predetermined tables for lithology, structure,
mineralisation, geotech, oxidation and alteration, in addition
to site developed metallurgical codes (metcode). These
metcodes are based on the metallurgical behaviour in the
Sepon processing plant. In addition the resource model is
coded by geometallurgical type for scheduling the process
plant and by lithology for determining mining costs and
geotechnical domains (Quigley, Hackhan and Broome, 2008).
At the Phu Kham concentrator, mineralogical examination
of monthly composites has been conducted since operations
commenced in 2008. This information, together with targeted
mineralogy, is routinely used for process understanding and
improvement. Evidence was provided of chalcocite-covellite
intergrowth with pyrite and rimming of pyrite. There was also
indication of copper activation of pyrite from soluble copper
species in weathered and transition ores. The mineralogy
data demonstrated that since primary chalcopyrite ores had
become the dominant source of plant feed, the major cause of
loss of copper in-plant tailings had changed from slow floating
finely liberated copper minerals to chalcopyrite locked in poor
quality coarse binary particles with non-sulfide gangue. This
knowledge, coupled with flotation modelling and simulation,
led to an increase in total recovery of both copper and gold
by six per cent into final concentrate by increasing mass
recovery into rougher concentrate, and debottlenecking of
rougher concentrate regrind, cleaning, and final concentrate
dewatering plants (Bennett, Crnkovic and Walker, 2011).
Variability in concentrate grades (24.7–37.6 per cent Cu) and
copper recoveries (60–90 per cent) during treatment of ore from
the Malanjkhand porphyry deposit, India, reflect its geology.
To obtain higher copper recoveries, a geometallurgical study
was undertaken to define the ore characteristics that affected
metallurgy and to assess applicable processes for recovering
the metals. It was found that chalcopyrite tarnished readily
and oxidised to chalcocite within several months. The
presence of both the tarnished chalcopyrite and chalcocite
affected chalcopyrite flotation and reduced copper recoveries
(Petruk and Sikka, 1987).
WE are metallurgists, not magicians
Integrating geometallurgy with copper concentrator design and operation
Europe – Middle East
In 2006 the Chelopech mine in Bulgaria commenced a
geometallurgical project to study the metallurgical complexity
of the ore and the key aspects of its treatment. A major aim of
the project was justification of possible changes in the process
flow sheet to achieve higher recoveries. The basis of the
program was samples collected during two full plant surveys
of the grinding, classification and flotation circuits. Sample
analysis included elemental and mineral composition, size
and nature of intergrowths, liberation, grindability and
floatability (Baltov et al, 2008).
Mikheevskoye in Russia undertook a study to improve
metallurgical performance through better understanding of
the variation in the porphyry orebody and mine planning.
Thirteen geometallurgical domains were established for the
ore zones which had similar characteristics during processing.
The study predicted that a geometallurgical approach could
potentially decrease the payback period for the project by
1.5 years and significantly increase the net present value
(Lishchuk et al, 2015).
The Sarcheshmeh concentrator in Iran implemented a
geometallurgical study to diagnose the reasons for fluctuating
copper recoveries. The study was instigated following reagent
and ball charge changes which were implemented as part of
a recovery improvement project. Five months of mineralogy
data was reviewed, with specific focus on the liberation of
chalcopyrite, covellite and chalcocite (Banisi et al, 2003).
Africa
The Mufulira copper mine in Zambia represented an earlier
age geometallurgy endeavour. Its copper minerals included
chalcocite, bornite, malachite, native copper, azurite, covellite
and chrysocolla. Gold was also present throughout the
orebody, as was graphitic carbon, which caused significant
difficulties in the flotation circuit. Geometallurgical planning
led to extensive efforts at blending, both underground and
in the crushing plant, to minimise the effect of ore variability
(White and Adair, 1948).
At the Tsumeb polymetallic mine in Namibia, the
presence of 40 economic minerals, some refractory, led to
treatment difficulties, even with high grades present. They
included chalcocite, bornite, tennantite, malachite, native
copper, cuprite, galena, cerunnite, sphalerite, native silver,
germanite, renierite, conichalcite, duftite, mottramite,
miretite, anglenite, smithsonite and willemite. To quote
Boyce, Venter and Adam (1970):
The distinct flotation characteristics of the different minerals
in the ore, and the rapid changes in ore grade and mineral
type, made the development of the current flow sheet a long
and at times, exasperating undertaking.
Over the concentrator’s history, multiple geometallurgical
studies were undertaken including in situ sphalerite activation,
overgrinding of chalcocite, tennantite selectivity, galena
flotation variability, intimate mineral locking and associations.
The implementation of an on-site daily mineralogical
analysis at Nchanga in Zambia allowed sharing of information
within the processing department, and between departments.
The geometallurgy program assisted in the diagnosis and
solution of processing problems. In one instance the program
was credited with achieving a one per cent increase in sulfide
recovery (Siddiqui, 2000).
The Ruashi mine in the Democratic Republic of Congo (DRC)
has significant variability in both copper and cobalt grades
across the resource. The constraint on acid availability and cost,
coupled with the need to stabilise plant performance, were the
WE are metallurgists, not magicians
financial drivers to focus on geometallurgical issues within
the orebody. This led to a strategy of ensuring a consistent
metal production profile, as a result of managed plant feed,
in terms of grade and geometallurgical characteristics, with
a mine to stockpile to plant philosophy built into the life-ofmine plans (Macfarlane and Williams, 2014).
Interestingly, there are similarities between Ruashi and
the Kolwezi concentrator 50 years earlier. Kolwezi, a Cu-Co
mine in the DRC also used early geometallurgy to stabilise
ore variability in its comminution and flotation plant. The
variable crushing requirements led to both in-pit and stockpile
strategies. Understanding of gangue variability assisted in
selecting and optimising flotation reagents (Saquet et al, 1962).
Another geometallurgy program in the DRC was conducted
on the Kamoa Copper deposit. The expansive study initially
demonstrated that representative, true samples of drill core
had been extracted from the drill core inventory. Copper
mineralogical studies identified bornite, chalcopyrite,
chalcocite, with lesser covellite and azurite. The variation
in copper speciation led to a mixed collector flotation
test program to optimise flotation of all of these species.
Liberation studies were used to define both primary grinding
and concentrate regrind requirements (Lotter et al, 2013).
Methods of integrating of geometallurgy with plant design
Traditional engineering design for flotation circuits uses the
‘Rule of Thumb’ approach. Typically this uses a nominated
maximum head grade for design. A scale-up factor is applied
to laboratory flotation tests for residence time. This scaleup factor is usually based on a designer’s experience at
other flotation sites and may or may not be relevant to the
circuit being designed, or the flotation equipment being
used. A small number of locked cycle tests are assumed to
represent the orebody and the results of these are often used
for financial analysis, independent of changes in mine plan
throughput and mineralogy. The engineering ‘rule of thumb’
approach is only strictly accurate where there is very little
variability in throughput, head grade and mineralogy. It may
provide accurate design for mature established operations
where a brownfield expansion is being considered but is
unlikely to provide accurate design for a new, greenfield
site. To overcome this Rule of Thumb approach to design,
geometallurgical test programs and data review have been
developed to provide realistic information into simulation
packages to provide data for process evaluation and plant
design. This approach was applied to the Andash project to
aid final design and risk minimisation.
A number of geometallurgical modelling techniques exist
and can be divided into three approaches as discussed by
Lishchuk et al (2015):
1. In the traditional approach the metallurgical response
of an ore in the mineral processing plant is calculated
from the chemical assays using mathematical functions,
which are often called recovery functions. The
functions are developed using variability testing and
statistical analysis to define the correlation between the
metallurgical response and feed properties.
2. The proxies approach uses geometallurgical tests for large
numbers of samples. The geometallurgical test is a smallscale test which indirectly measures the metallurgical
response. Normally the geometallurgical test results
must be converted with certain correction factors to give
estimates on the metallurgical results of plant.
3. The mineralogical approach’s main characteristic is the
continuous and systematic collection of quantitative
mineralogical information.
43
G Harbort et al
The geometallurgical development discussed in this paper
was conducted in a number of stages:
•• incorporation of mineral data, either from sequential
copper assays or mineralogy, into the resource model
and mine plan
•• development of floatability parameters based on the
rougher performance in either locked cycle tests or
rougher batch tests
•• generation of a floatability component model to simulate
the locked cycle tests or existing plant performance
•• calibration of the model so that model results equate to
either actual locked cycle test results or historical plant
performance
•• development of a JKSimFloat simulator incorporating
floatability data, proposed circuit design, flotation feed
rates and flotation machine characteristics
•• determination of mineral recovery with ores of various
copper mineralogies, based on variability tests and
mine plans.
CASE STUDIES
Two previously published case studies are reviewed in this
paper. The first is the Andash study (Harbort, Cordingley and
Phillips, 2011), a greenfields geometallurgical study.
The second is Northparkes Mines (Clarke et al, 2014; Jones
and Morgan, 2016), an operating site with both brownfield
and greenfield components. The brownfield component
relates to optimising the plant operation, while the greenfield
component relates to the prediction of future performance
with various circuit and mine plan options.
Andash
Location
The Andash Gold-Copper Project is located in the Kyrgyz
Republic, approximately 260 km from the capital Bishkek in
the Talas region of Kyrgyzstan.
Geological description
The Andash deposit is formed by the stratified sediments of the
Tohtonisai and Jangiturmush formations. These formations
are characterised by bodies and dykes of the Taldysui
monzo-diorite, granite and Permian dyke complexes. The
Tohtonisai formations, which forms the south-west part of
the deposit with a thickness of 300 m, comprises of moderate
alkaline basalts, trachy-andesites and associated tuffs, beds
of agglomerate basalts, tuff conglomerates and sandstone
layers. The Jangiturmush Formation forms the majority
of the deposit’s area with a thickness of 450–500 m. The
formation is formed by tuff conglomerates, tuff-gravelites,
tuff-sandstone with interlayers of tuffite, tuff-andesites and
agglomerate tuff of basalts.
The Andash copper-gold mineralisation is structurally
confined to a hydrothermally silicified tectonic breccia
and quartz alteration, extending from and terminating into
granodiorite. The orebody is zoned as follows:
•• The upper section of the ore zone spreads from surface
to a depth of 40 to 75 m and is composed mainly of
cataclasised granodiorite porphyry. Hydrothermally
imposed ore minerals include chalcopyrite, native gold
and with hypergene minerals including cuprite, covelline,
bornite, native copper, chalcosine and hydrotite.
•• The middle section of the ore zone stretches from
40–105 m with sharply increased gold and copper
44
content. The ore is formed by carbonate-chloritesericite-quartz alterations. Hypergene minerals are
significant and are represented by malachite, azurite,
covelline, cuprite, native copper, drogoethite, bornite
and chalcosine.
•• The lower ore zone is at depths of 60–175 m. This zone is
mainly formed by chloritised and sericitised cataclasised
granodiorite porphyry with increasing chalcopyrite and
pyrite content.
Numerous barren dykes of various compositions are present
within the orebody.
Treatment
The proposed Andash ore processing flow sheet, as shown
in Figures 1 and 2, was designed to recover gold bearing
copper concentrate through the following principal unit
processes:
•• three-stage crushing, ore storage and reclaim
•• single stage ball milling (expanded to primary and
secondary ball milling in Phase 2)
•• flash flotation (Phase 2)
•• sequential rougher flotation (utilising split sulfide and
oxide circuits)
•• combined cleaner, cleaner scavenger and recleaner
flotation
•• sulfide concentrate regrind (Phase 2)
•• concentrate dewatering, filtration, storage and load-out
and tailings thickening (Phase 2)
•• disposal and decant water return.
Geometallurgical evaluation
During the development and design of the Andash
concentrator, numerous flotation tests were undertaken
including variability tests. The variability tests confirmed that
the orebody exhibits significant variability in total copper
rougher recovery and minor variability in rougher gold
recovery. The wide copper variability was directly related to
copper mineral variability. The recovery of copper sulfides
was relatively robust with 96 per cent rougher recovery when
no copper oxides were present, declining to 90 per cent at a
copper oxide to copper sulfide ratio of 0.8. At very high oxide
copper content the sulfide recovery decreased significantly.
The rougher copper oxide recovery was heavily dependent
on the ratio of acetic acid-soluble copper, eg malachite to
weak acid-soluble oxides such as tenorite. When copper
carbonates were the only oxide copper minerals present, their
recovery approached 75 per cent and declined with higher
weak acid-soluble copper content. Due to the wide variability,
it was determined that the locked cycle tests could not
effectively be used as a measure of life-of-mine performance.
At this stage the test sample was approaching exhaustion,
with limited sample available to conduct further flotation
tests. A geometallurgical approach was adopted to evaluate
performance over the mine life.
Resource and mine plan development
Sequential assay results were input into the resource model.
Development of a mine plan based on this geometallurgical
resource model allowed a mine plan and concentrator
schedule to be developed based on copper mineral species
(sequential copper extractions). The scheduled feed grade to
the concentrator on a copper species basis is shown in Figure 3.
WE are metallurgists, not magicians
Integrating geometallurgy with copper concentrator design and operation
FIG 1 – Andash Phase 1 circuit flow sheet.
FIG 2 – Three-dimensional view of the Andash circuit.
WE are metallurgists, not magicians
45
G Harbort et al
Key points in the schedule are:
•• At the commencement of mining, heavily oxidised
material was stockpiled for end of mine life treatment.
This resulted in an initial concentrator feed where the
majority of copper was acetic acid-soluble, eg malachite.
•• During the second year of operation, the starter pit
penetrated a high-grade sulfide pod, with a significant
increase in both feed grade and aqua regia-soluble copper,
ie chalcopyrite. Year three represents a significant mine
expansion with lower grade copper carbonate material
feeding the concentrator. This period coincided with the
planned expansion of the concentrator facilities.
•• During the fourth year, increased amounts of copper
oxides and copper carbonates are mined from the
expanded pit and presented to the concentrator. The
deepening pit also results in significant amounts of
chalcopyrite being mined and treated.
•• In the ensuing three years, the head grade to the
concentrator remains relatively constant, although
the copper mineralogy changes significantly. Acetic
acid-soluble copper is progressively replaced with
chalcopyrite as the pit deepens. The amount of copper
oxide material also decreases.
•• Year seven represents treatment of stockpiled material
with substantial amounts of copper oxides, although
acetic acid-soluble copper still represents 50 per cent of
the concentrator feed.
Floatability component modelling
Ore floatability components were calculated from the sulfide
dominant composite rougher kinetic locked cycle test results.
As no bubble size measurements were conducted, a bubble
size of 1.1 mm was assumed. It was also assumed that the
batch cell froth recovery was 100 per cent and entrainment
was zero. As the model utilised a mineral based system it was
assumed that acetic acid-soluble copper was malachite, weak
sulfuric acid copper was tenorite, cyanide soluble copper was
chalcocite and the copper extracted in the final aqua regia was
chalcopyrite.
The floatability parameters indicated that chalcopyrite was
highly floatable in the sulfide rougher feed and was depressed
in the oxide flotation stage. Both malachite and tenorite had
low floatability in the sulfide rougher. Both became floatable in
the oxide rougher with the weighted floatability of malachite
being approximately double that tenorite.
A simulation was the created to match the configuration
of the locked cycle test, including cleaner stages. The model
results were then compared to the measured locked cycle
FIG 3 – Life-of-mine mineral grade in the concentrator feed schedule.
46
results to see if floatability was constant across the locked
cycle test. The comparison indicated that mineral flotabilities
in the respective rougher concentrates were not constant
across their respective cleaner circuit feeds. Flotabilities in the
cleaner concentrate were, however, conserved into recleaner
feeds. The major variations in the cleaner feeds were:
•• a decrease in the floatability of tenorite and malachite in
the sulfide cleaner feed
•• an increase in the floatability of chalcopyrite in the
sulfide cleaner feed
•• a decrease in the floatability of tenorite in the oxide
cleaner feed.
A series of floatability transfer matrices were incorporated
into the model to calibrate it against the locked cycle test. This
provided acceptable recoveries of copper and sulfur, with the
modelled final concentrate grade matching the locked cycle
test in terms of copper and sulfur. Modelled iron grades were
lower in the final concentrate however and this resulted in
over estimation of non-sulfide gangue. To correct the iron
imbalance non-sulfide gangue was adjusted in the model to
contain iron. (Later QEMSCAN – Quantitative Evaluation
of Minerals by Scanning Electron Microscopy – analysis of
the concentrate indicated the higher iron levels were caused
by the flotation of copper-goethite). The calibrated model
was then compared against the second locked cycle test and
showed good agreement. A further two locked cycle tests
were conducted with different configurations and simulated.
The simulated results agreed with actual results and it was
decided that the model could effectively be used to replace
further locked cycle tests and simulate full circuit operation.
Integration with plant design
Seventy-eight simulations were conducted on the Andash
flotation circuit developed during the DFS. This circuit consisted
of sequential sulfide and oxide circuits and split sulfide and
oxide cleaner circuits. Each simulation was conducted on the
basis of one month of the concentrator feed schedule.
These simulations were initially conducted to determine the
variability of copper and gold recovery and concentrate grade
throughout the planned operating life. They also provided
detailed mass balances across each month of the mine plan
and were used as a cross-check of the DFS design. A number
of significant problems soon became apparent from using a
rule of thumb approach. Although the simulations showed
only minor variation in recoveries, they indicated major
variation in stream pulp flow on a month-to-month basis. It
was expected that these variations would become worse over
shorter time frames. The sulfide cleaner and recleaner circuits
typically operated within acceptable limits. This was not the
case with the oxide cleaner circuit where significant variations
were evident and circuit stability was expected to cause
operating difficulties and loss of metallurgical performance.
The oxide recleaner circuit would never have operated to
design without major modifications.
An initial solution evaluated was to combine sulfide rougher
concentrate and oxide rougher concentrate and treat them in a
combined cleaner circuit utilising the same amount of flotation
equipment specified in the DFS design. This provided a major
improvement in the ratio of anticipated residence time to the
nominated design residence time and hence circuit stability.
Simulations, however, failed to match those of the test work,
with any recovery benefit undermined by large decreases
in concentrate grade caused by the copper oxide flotation
chemicals generating high levels of gangue entrainment. A
modified combined cleaner circuit was simulated with staged
addition of rougher sulfide concentrate and rougher oxide
WE are metallurgists, not magicians
Integrating geometallurgy with copper concentrator design and operation
concentrate to the cleaner circuit. This provided an increase in
copper recovery of approximately 2.5 per cent when treating
sulfide ore and little tangible benefit when treating oxide
ore. Considering the ore mineral split would represent an
overall improvement in copper recovery over the mine life of
0.5 per cent. In addition, the revised circuit required two less
flotation cells and was expected to be overwhelmingly more
stable than the DFS split-cleaner circuit design. The modified
circuit was incorporated into the Andash detail design. In
addition, the simulation mass balances were used to calculate
cleaner circuit make-up water requirements. This resulted in
all water addition piping to the cleaner circuit being changed
in the detailed design to further maximise circuit stability.
•• The cleaner Jameson cell tailing is treated by the cleaner
scavenger mechanical cells, where the concentrate
discharges to recleaner Jameson cell flotation. The tailing
from the cleaner scavenger is pumped back to the feed of
the rougher scavengers.
•• Concentrate from the recleaner Jameson cell discharges
to the final concentrate hopper. Tailing from the
recleaner Jameson cell is returned to the feed of the
cleaner Jameson cell.
Geometallurgical evaluation
The copper-gold porphyry mineralisation at Northparkes
is hosted by the Late Ordovician Goonumbla Volcanics
and occurs in stockwork quartz veins and disseminations
associated with potassic alteration. Sulfides are zoned
laterally from the centres of mineralisation. The central
portions are bornite-rich with minor chalcopyrite, zoning
outward through equal portions of bornite + chalcopyrite,
to a chalcopyrite-rich zone. Pyrite increases outward at the
expense of bornite. Beyond the chalcopyrite zone, pyrite is the
main sulfide and copper sulfides are minor to absent.
From June 2011 to November 2013 Northparkes conducted
mineralogical evaluation of monthly composites to evaluate
plant performance. Conducted by the Rio Tinto Technology
and Innovation group, these production investigations were
complemented by analysis of comminution samples, selected
flotation survey samples and drill core samples from potential
future concentrator feed sources. Historical plant surveys
resulted in the ongoing development of model simulations.
In 2013, the 2012 un-sized, copper only JKSimFloat simulator
was expanded to include mineralogy and a floatability
component and used to check the accuracy of performance
estimates for the future. Simulations were conducted over
the period of 2011 to 2013 to confirm validity of the model.
In March 2015, Northparkes installed a process mineralogy
facility on-site. The laboratory included a TESCAN TIMA
scanning electron microscope (SEM) and was the first SEM
facility to be commissioned on a mine site in Australia. The
facility has not only continued with the analysis of monthly
plant composites, but has also supported: metallurgical test
work programs; the evaluation of alternative ores and feed
material; plant and laboratory reagent trials; grinding media
trials; plant optimisation and improvement; changes to plant
configuration and provided laboratory support to external
consultants working on NPM projects.
Treatment
Optimisation
Northparkes
Location
Northparkes Mines (Northparkes), an unincorporated
joint venture between the China Molybdenum Company
(80 per cent), Sumitomo Metal Mining Oceania (13.3 per cent)
and Sumitomo Corporation (6.75 per cent), operates block
cave and open cut mines and an ore processing plant located
27 km north of Parkes in central New South Wales.
Geology
The Northparkes ore processing flow sheet is shown in
Figure 4 and Figure 5. The concentrator consists of two
modules. Each module has its own grinding circuit, flotation
circuit, concentrate thickening circuit and filtration circuit.
The tailings from the two modules are combined in a single
tailing thickener before being deposited in tailing facilities.
Flotation takes place in two distinct but similarly configured
modules, each fed by its own grinding circuit. The tertiary
cyclone overflows of each module feed the main flotation
circuits with the following configuration:
•• Preflotation is performed in tank cells, with frother,
sodium hydrosulfide and thiocarbamate promoter
added.
•• The preflotation tailing is further treated with xanthate
collector and frother. The conditioned pulp flows
through a series of banks of mechanical rougher and
scavenger cells, wherein down-the-bank xanthate is also
employed.
•• Further scavenging in a post-flotation stage is conducted
in large tank cells. The resulting post-flotation
concentrates are recirculated back to the tertiary
grinding circuit, while the post-flotation tailings are
combined in a single thickener before being deposited in
the tailing facilities.
•• Concentrates from preflotation, roughers and rougher
scavengers are combined to feed the cleaner Jameson cell.
Concentrate from the Jameson cleaner cell discharges to
the final concentrate hopper.
WE are metallurgists, not magicians
Routine analysis of monthly plant composites has allowed the
measurement of plant performance over time and provided
a baseline of feed characteristics versus performance for
comparison. Northparkes have been conducting quantitative
mineralogical analysis of plant composites for more than
11 years. Important information obtained from this data
has included a measure of the variation in the mineral
make-up, ratios and copper source minerals of the plant
feed; individual mineral recoveries by size; liberation and
locking characteristics; grain size versus particle size; mineral
association and mineral maps. Monthly mineralogy is
routinely input into the circuit flotation simulations, together
with other operating data to calibrate and validate the model
accuracy.
The data has shown the reality of actual recoverable copper
versus the actual non-recoverable copper. This is an important
distinction when evaluating performance and changes to the
circuit. Analysis of the concentrate has illustrated the nature
of impurities and levels of penalty elements and facilitated
the development of a test work program or strategy, to reduce
these. The microchemical analysis of gold particles by energy
dispersive X-ray (EDX) analysis has detected the presence of
gold and silver tellurides in the ore. This data has driven a
phase of laboratory method development to improve the fireassay process and quality of NPM gold analysis, particularly
in regard to shipment assays.
Both equipment changes and circuit improvement projects
have incorporated geometallurgical data. This has driven a
reduction in the technical risk of specific projects with the
47
G Harbort et al
FIG 4 – Northparkes Module 1 circuit flow sheet (shaded areas represent potential future changes).
FIG 5 – Three-dimensional view of the Northparkes circuit.
48
WE are metallurgists, not magicians
Integrating geometallurgy with copper concentrator design and operation
data used to evaluate, justify, test and validate major plant
upgrades. This has proven to be a major drawcard when
justifying funding for an expensive project. As an example,
installation of additional scavenger flotation cell capacity
demonstrated recovery improvements of around one per cent.
Mineralogy information plus simulations using mineralogical
inputs were involved in all aspects of the project, with data
used to successfully evaluate the risk, drive decision-making,
optimise and validate changes. Mineralogy was also used
to critically review the post-commissioning performance of
the new cells. Additional implementations have included
installation of froth washing and subsequent optimisation
to decrease entrainment. The bypass of a primary cyclone to
reduce overgrinding of already liberated copper minerals and
optimisation of reagents once the bypass was implemented.
The Northparkes on-site laboratory has frequently been
utilised to support work programs. Mineralogical analysis
is used to evaluate change in reagents and grinding media
from bench tests and in-plant. Performing the mineralogy
and some of the assay components of the test work in-house
has given the site control over quality assurance and project
integrity.
In addition to optimisation in the concentrator,
geometallurgy has also been used for mine optimisation.
Northparkes are currently mining the Endeavour 48 (E48)
orebody via a modified block cave technique. This orebody
still has 15 years of mining life, yet up until recently there was
very little mineralogical information available. To remedy
this, mineralogical analysis has commenced on material from
the current drawpoints to gain information about the samples
that make up the daily feed blend. The mineralogy and assay
combination is used to provide mineral abundance, elemental
deportment and textural data to map cave drawpoints. The
data shows the mineralogical variability of the orebody
which is currently mined according to fragmentation, caving
characteristics and copper grade inputs. The long-term
objective would be to add mineralogy data to that criteria.
Some drawpoints contain very fine clay-bearing material
that has the ability to move like a liquid. This is extremely
dangerous in the cave environment as the material has
the potential to cause a fines rush. This fine material was
investigated using SEM imaging and EDX microchemical
analysis techniques to infer the size, illustrate the morphology
and measure chemical composition. Based on this data,
additional mining controls including exclusion zones have
been implemented where similar mineralogy indicators exist.
Life-of-mine planning
The incorporation of mineralogical data and floatability
components into simulations to predict future performance
for analysis has provided a more robust method of
prediction than the historical method of prediction using
only a head grade and throughput relationship to predict
recovery, based purely on laboratory flotation results.
This improved understanding allows for the development
of circuit modifications and circuit options earlier in the
planning process as well as supporting the development of
future concentrate specifications. The simulation of future
performance provides an improved ability to manage future
expectations and ore reserves, becoming a critical tool in
optimising the mining sequence over the life-of-mine. Further
expansion of the model will introduce the ability to predict
gold and impurity behaviour as well as the copper sulfides.
WE are metallurgists, not magicians
CONCLUSIONS
The use of geometallurgy with copper concentrator design
and operation has a long history. Recent years have seen it
move from an ad hoc methodology at a few sites to a wellestablished procedure.
The Andash case study details the use of geometallurgy in
greenfield plant design with limited information available.
Geometallurgical modelling using sequential copper assays
and floatability component modelling provided a design
methodology with significantly less associated risk than
traditional design.
The Northparkes case study illustrates the use of
geometallurgy, via process mineralogy and mineral
component modelling. The use of geometallurgy has
facilitated a common goal of mill optimisation, stability and
more realistic forecasting through orebody knowledge and
the sharing of information between disciplines.
ACKNOWLEDGEMENTS
The authors would like to thank Northparkes Mines and
Amec Foster Wheeler for allowing this paper to be published.
The case studies utilise three papers previously published by
the AusIMM and the authors are grateful for the permission
to extract the content from these. Previous authors of those
papers are also acknowledged, namely Guy Cordingley,
Marius Phillips and Danica Clarke. Sections of the introduction
were extracted from an unpublished paper – ‘Technological
advances in flotation’, written with the guidance of Professor
Alban Lynch in 2011.
REFERENCES
Afenya, P M and Mwaba, C C, 1991. Flotation of Ok Tedi sulphide
minerals, in Proceedings PNG Geology, Exploration and Mining
Conference, pp 191–196 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Alruiz, O M, Morrell, S, Suazo, C J and Naranjo, A, 2009. A novel
approach to the geometallurgical modelling of the Collahuasi
grinding circuit, Minerals Engineering, 22:1060–1067.
Anon, 1956. Success depends on: geologists, mining engineers,
metallurgists, presented to Symposium on Greater Cooperation
for Metallurgical Planning; Second Annual Rocky Mountain
Minerals Conference, Salt Lake City, October, Mining Engineering,
July, p 698.
Arden, H and Tverdov, A, 2014. Resource and reserve valuation
practices in CIS countries, in Mineral Resource and Ore Reserve
Estimation: The AusIMM Guide to Good Practice, second edition
pp 825–834 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Baltov, I, Hristov, N, Evtiminova, K, Dontchev, I and Kuzev, L, 2008.
Metallurgical challenges in the recovery of copper and gold from
the Chelopech mine, Bulgaria, in Proceedings XXIVth International
Minerals Processing Congress (ed: W D Duo and S C Yao), pp 1785–
1792 (Science Press: Beijing).
Banisi, S, Iran-Manesh, H R, Shayestefar, M R and Shekarchyan,
H, 2013. Mineralogical tracing of metallurgical results – the
Sarcheshmeh copper mine case, in Proceedings 35th Annual
Meeting of the Canadian Mineral Processors, Ottawa, 21–23 January,
pp 417–433 (Canadian Institute of Mining, Metallurgy and
Petroleum: Montreal).
Barns, K E, Colbert, P J and Munro, P D, 2009. Designing the optimal
flotation circuit – the Prominent Hill case, in Proceedings Tenth
AusIMM Mill Operators’ Conference, pp 173–182 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
49
G Harbort et al
Baumgartner, R, Gomez, P and Escobar, G, 2016. Comprehensive
mineralogical characterisation at the Cerro Corona Cu-Au
porphyry mine – the fundamental key for geometallurgical
applications, in Proceedings The Third AusIMM International
Geometallurgy Conference, pp 221–230 (The Australasian Institute
of Mining and Metallurgy: Melbourne).
Beniscelli, J, 2011. Geometallurgy – fifteen years of developments
in Codelco: Pedro Carrasco contributions, in Proceedings The
First AusIMM International Geometallurgy Conference, pp 3–7 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Bennett, D, Crnkovic, I and Walker, P, 2011. Recent process
developments at the Phu Kham copper-gold concentrator, Laos,
in Proceedings 11th Mill Operators’ Conference 2012, pp 257–272 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Bergerman, M G, Morais, J, Castro Rosa, M A N and Delboni Jr, H,
2008. Variability studies for the Sossego 41 ktpd grinding circuit,
in Proceedings International Mineral Processing Conference (ed: W D
Duo and S C Yao) volume 1, pp 407–416 (Science Press: Beijing).
Blanchard, R and Hall, G, 1939. Discussion – the mineralogy of the
Black Star ore body and its relation to milling practice at Mount
Isa, The AusIMM Proceedings, 116:487–490.
Blatter, P, Cayouette, J and Cousin, P, 2011. Agnico-Eagle Laronde
plant: metallurgical challenges present and future, in Proceedings
43rd Annual Canadian Mineral Processors Conference, pp 3–15
(Canadian Institute of Mining, Metallurgy and Petroleum:
Montreal).
Boisvert, J, Rossi, M and Ehrig, K, 2013. Geometallurgical modelling
at Olympic Dam Mine, South Australia, Mathematical Geosciences,
45:901–925.
Boyce, J H, Venter, W J and Adam, J, 1970. Beneficiation practice at
Tsumeb concentrator, in Lead and Zinc (eds: D O Rausch and B
C Mariacher), vol 1, pp 542–570 (American Institute of Mining,
Metallurgical, and Petroleum Engineers: New York).
Bradshaw, D J, Wightman, E M, Evans, C L and Triffett, B, 2012.
Characterising the reasons for the difference in performance of
two ore blends in early processing at OZ Minerals Prominent Hill
operation, in Proceedings 11th AusIMM Mill Operators’ Conference
2012, pp 373–278 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Brissette, M and Roman, E, 2012. Concentrate grade and regrind size
improvement on the east pit hypogene ore at Kemess Mine, in
Proceedings 44th Annual Canadian Mineral Processors Conference,
pp 27–38 (Canadian Institute of Mining, Metallurgy and
Petroleum: Montreal).
Carrasco, P, Campos, M, Tapia, J and Menichetti, E, 2005.
Heterogeneity and Ingamells’s tests of some Chilean porphyry
ores, in Proceedings Second World Conference on Sampling and
Blending, pp 139–150 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Clarke, D, Harbort, G, Morgan, D and Sola, C, 2014. Northparkes
Mines – flotation and planning for the future, in Proceedings 12th
AusIMM Mill Operators’ Conference, pp 203–208 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Clemmer, J B, 1956. Metallurgical planning for porphyry copper ores,
presented to Symposium on Greater Cooperation for Metallurgical
Planning; Second Annual Rocky Mountain Minerals Conference,
Salt Lake City, October 1955, Mining Engineering, July, pp 701–702.
Comi, T, Garcia, C, Kaleta, R and Kurzydło, P, 2013. Sierra Gorda
– the most modern Cu-Mo mine in Chile, in Proceedings Mineral
Engineering Conference MEC2013 (ed: P B Kowalczuk), SwieradowZdroj, 16–19 September, pp 109–121 (Wroclaw University of
Technology: Wroclaw).
Compan, G, Pizarro, E and Videla, A, 2015. Geometallurgical model
of a copper sulphide mine for long-term planning, Journal of The
Southern African Institute of Mining and Metallurgy, 115:549–556.
Coward, S, Vann, J, Dunham, S and Stewart, M, 2009. The primaryresponse framework for geometallurgical variables, in Proceedings
Seventh International Mining Geology Conference, pp 109–113 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
50
David, D, 2007. The importance of geometallurgical analysis in
plant study, design and operational phases, in Proceedings Ninth
AusIMM Mill Operators’ Conference, pp 241–248 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
DeMull, T J, Spenceley, J and Hickey, P, 2001. Planning and teamwork
lead to successful start-up at Batu Hijau, presented to SME
Annual Meeting, Denver, 26–28 February, SME Preprint 01–159.
Dobbe, R, Bajic, S, Kalstrom, M, Palko, A, Martin, C, Thorpe, R, Lamson,
G and Baron, J, 2014. On site automated mineralogical analysis at
the Mount Polley porphyry copper mine, Canada, presented to
XXVII International Mineral Processing Congress, Santiago.
Dominy, S C and O’Connor, L, 2016. Geometallurgy – beyond
conception, in Proceedings Third AusIMM International
Geometallurgy Conference, pp 3–10 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Downs, R C, 1990. Geological mapping and logging at Hellyer, in
Proceedings Mine Geologists’ Conference, pp 1–7 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Dunham, D and Vann, J, 2007. Geometallurgy, geostatistics and
project value – does your block model tell you what you need
to know?, in Proceedings Project Evaluation 2007, pp 189–196 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
England, J K, Kilgour, I and Kanau, J L, 1991. Processing copper-gold
ore at Ok Tedi, in Proceedings PNG Geology, Exploration and Mining
Conference, pp 183–190 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Evans, C L, Wightman, E M, Manlapig, E V and Coulter, B L, 2011.
Application of process mineralogy as a tool in sustainable
processing, Minerals Engineering, 24(12):1242–1248.
Fennel, M, Guevara, J, Canchaya, S, Velarde, G, Baum, W and Gottlieb,
P, 2005. QEMSCAN Mineral Analysis for Ore Characterisation
and Plant Support at Cerro Verde, in XXVII Convencion Minera,
Arequipa, Peru, pp 1-11 (Instituto de Ingenieros de Minas del
Perú: Lima)
Fonseca, E and Sá, L R B M, 2005. Methodological evaluation of a
liberation study in sulfide copper ore, in Proceedings Centenary
of Flotation Symposium, pp 953–958 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Gottlieb, P, Adair, B J and Wilkie, G J, 1994. QEM*SEM liberation
indices for grinding classification and flotation, in Proceedings Fifth
Mill Operators’ Conference, pp 5–14 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Gregory, M J, Lang, J R, Gilbert, S and Hoal, K O, 2013. Geometallurgy
of the Pebble porphyry copper-gold-molybdenum deposit,
Alaska: implications for gold distribution and paragenesis,
Economic Geology, 108:463–482.
Gregory, P W, Hartley, J S and Wills, J A, 1987. The Thalanga
massive sulphide deposit, in Proceedings Pacific Rim Congress 87,
pp 573–577 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Greig, D, Widenbar, L, Oliver, R, Harbort, G, Cuttriss, C, Kerr, T F
and Cummings, R, 2009. Mineral processing and metallurgical
testwork (Section 16), in Constancia Copper Project, Definitive
Feasibility Study technical report NI 43–101, Rev 0, 28 September.
Harbort, G, Cordingley, G and Phillips, M, 2011. The integration
of geometallurgy with plant design, in Proceedings MetPlant
2011, pp 339–363 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Harbort, G, Lauder, D, Miranda, J and Murphy, A, 2000. Size by size
analysis of operating characteristics of Jameson Cell cleaners at
the Bajo de Alumbrera copper/gold concentrator, in Proceedings
Seventh Mill Operators’ Conference, pp 207–220 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Hartog, J, Beehan, V, Karageorgos, J and Beeson, D, 2014. Optimisation
of the Peak Gold Mine’s processing plant through advanced
process control, in Proceedings 12th AusIMM Mill Operators’
Conference, pp 335–344 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
WE are metallurgists, not magicians
Integrating geometallurgy with copper concentrator design and operation
Hoffman, C W, 1964. A review of developments in flotation treatment
of Mount Isa chalcopyrite ores, AusIMM Proceedings, 103.
Hunt, J, Berry, R and Bradshaw, D, 2011. Characterising chalcopyrite
liberation and flotation potential: examples from an IOCG
deposit, Minerals Engineering, 24(12):1271–1276.
Jackson, B R, Gottlieb, P and Sutherland, D N, 1988. A method for
measuring and comparing the mineral grain sizes of ores from
different origins, in Proceedings Third Mill Operators’ Conference,
p 61 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
James, B and Scamardella, M, 2000. The Re-opening of CSA Mine,
Cobar NSW, in Proceedings Seventh Mill Operators’ Conference,
pp 137-147 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Johnson, N W and Munro, P D, 2008. Methods for assigning domains
in the primary sulphide zone of a sulphide orebody, in Proceedings
Ninth International Congress for Applied Mineralogy, pp 597–603 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Jones, K and Morgan, D, 2016. On-site process mineralogy – linking
departments and disciplines with the common goal of increasing
plant performance, in Proceedings 13th AusIMM Mill Operators’
Conference, pp 149–156 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Kuhar, L L, McFarlane, A J, Chapman, N M, Meakin, R L, Martin,
R, Turner, N L and Robinson, D J, 2013. Calibration and testing
of a geometallurgical leaching protocol for determining copper
mineralogical deportment, in Proceedings Second AusIMM
International Geometallurgy Conference, pp 177–186 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Lane, G S and Richmond, G D, 1993. Improving fine particle flotation
selectivity at Hellyer, in Proceedings XVIII International Mineral
Processing Congress, pp 897–904 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Leggett, A R and Morin, G R, 2013. Maximising concentrate revenues
at the Xstrata Copper Kidd concentrator, in Proceedings 45th
Annual Canadian Mineral Processors Operators’ Conference, pp 465–
478 (Canadian Institute of Mining, Metallurgy and Petroleum:
Montreal).
Liebezeit, V, Smith, M, Kittler, P, Macmillan, E and Lower, C, 2011.
Geometallurgy data management – a significant consideration, in
Proceedings First AusIMM International Geometallurgy Conference,
pp 237–246 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Lishchuk, V, Koch, P-H, Lund, C and Lamberg, P, 2015. The
geometallurgical framework: Malmberget and Mikheevskoye
case studies, Mining Science, 22:57−66.
Jones, M P, 1984. Recent developments in the rapid collection of
quantitative mineral data, in Proceedings Second International
Congress on Applied Mineralogy, pp 449–464 (Metallurgical Society
of AIME: Warrendale, Pa).
Loidl, G, 2012. The Elura orebody: a multidisciplinary study
investigating geochemical mineralogical and geometallurgical
ore characteristics, PhD thesis, University of Adelaide, School of
Earth and Environmental Sciences.
Kahn, H, Antoniassi, J, Shimizu, V and Uliana, D, 2014. X-ray
diffraction cluster analysis and automated mineralogy of Sossego
copper ore, Brazil, presented to XXVII International Mineral
Processing Congress, Santiago.
Lotter, N O, Oliveira, J F, Hannaford, A L, Amos, S R and Broughton,
D W, 2013. Flowsheet development for the hypogene geomet
unit in the Ivanplats Kamoa copper project, in Proceedings
45th Annual Canadian Mineral Processors Operators’ Conference,
pp 71–85 (Canadian Institute of Mining, Metallurgy and
Petroleum: Montreal).
Keran, V P, Zumwalt, F and Palmers, J, 1998. Design aspects of the
Minera Alumbrera concentrator circuit, presented to SME Annual
Meeting, Orlando, 9–11 March.
Khosrowshahi, S, Shaw, W J and McKevitt, M B, 2009. Ore block
optimisation for open pit grade control, in Proceedings Seventh
International Mining Geology Conference, pp 141–146 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Kildale, M B, 1956. Metallurgical planning in connection with leadzinc or lead-zinc-copper mines, presented to Symposium on
Greater Cooperation for Metallurgical Planning; Second Annual
Rocky Mountain Minerals Conference, Salt Lake City, October
1955, in Mining Engineering, July, pp 706–707.
Kilgariff, B, 2003. Thalanga processing plant – as many lives as a cat
and just as curious, in Proceedings Eighth AusIMM Mill Operators’
Conference, pp 129–141 (The Australasian Institute of Mining
and Metallurgy: Melbourne).
Kilgour, I, 1995. Development of the Escondida Concentrator 1990–
1995, in Proceedings XIX IMPC, pp 145–148 (Society for Mining,
Metallurgy and Exploration: Littleton).
Lund, C and Lamberg, P, 2014. Geometallurgy – a tool for better
resource efficiency, European Geologist, 37:39–43.
Macfarlane, A S and Williams, T P, 2014. Optimizing value on a copper
mine by adopting a geometallurgical solution, The Journal of The
Southern African Institute of Mining and Metallurgy, 114:929–936.
MacPherson, A R, 1976. A simple method to predict the autogenous
grinding mill requirements for processing ore from a new deposit,
Presented to SME-AIME Fall Meeting, Denver, September.
Magallanes-Hernández, L and Espinosa-Gómez, R, 2005.
Mineralogical and metallurgical results at Peñoles’s mining
operations, Mexico, in Proceedings Centenary of Flotation
Symposium, pp 887–894 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Martin, H K, 1957. Milling practice at the Lavender pit concentrator,
Mining Engineering, November, pp 1229–1235.
Kirkland, A G, 1956. Mining engineer’s view of cooperation, presented
to Symposium on Greater Cooperation for Metallurgical Planning;
Second Annual Rocky Mountain Minerals Conference, Salt Lake
City, October 1955, Mining Engineering, July, pp 707–708.
Matar, J A, Perucca, J C, de Salmuni, N S, Berenguer, T and Carrizo, M,
1986. Geological research as a basis to define metallurgical
behaviour of the disseminated Cu-Au orebody of the Bajo la
Alumbrera, presented to SME Annual Meeting, New Orleans,
2–6 March.
Kittler, P, Liebezeit, V, Ehrig, K, Macmillan, E and Lower, C, 2011.
‘It Seemed like a good idea at the time …’ common mistakes
in geometallurgy, in Proceedings First AusIMM International
Geometallurgy Conference, pp 133–138 (The Australasian Institute
of Mining and Metallurgy: Melbourne).
McCaffery, K and Wirfiyata, F, 2011. Applied Geo-metallurgical
characterisation for life of mine throughput prediction at Batu
Hijau, presented to Fifth International Conference on Autogenous
and Semiautogenous Grinding Technology, Vancouver,
25–29 September.
Klohn, S, Stephenson, D and Granados, H, 2016. Constancia project
process plant design and start up, in Proceedings 48th Canadian
Mineral Processors Operators’ Conference, Ottawa, January,
pp 100–112 (Canadian Institute of Mining, Metallurgy and
Petroleum: Montreal).
McKay, N, Vann, J, Ware, W, Morley, C and Podkiewicz, P, 2016.
Strategic and tactical geometallurgy – a systematic process to
add and sustain resource value, in Proceedings Third AusIMM
International Geometallurgy Conference, pp 29–36 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Kormos, L, Oliveira, J, Fragomeni, D, Whiteman, E and De la Cruz,
J C, 2010. Process diagnosis using quantitative mineralogy, in
Proceedings 42nd Meeting of Canadian Mineral Processors, pp 97–
119 (Canadian Institute of Mining, Metallurgy and Petroleum:
Montreal)
McQuiston Jr, F W, 1968. Metallurgist’s view on milling base
metal ores, presented to Symposium on Greater Cooperation
for Metallurgical Planning; Second Annual Rocky Mountain
Minerals Conference, Salt Lake City, October 1955, Mining
Engineering, July, pp 698–700.
WE are metallurgists, not magicians
51
G Harbort et al
McQuiston Jr, F W and Bechaud Jr, J, 1956. Metallurgical sampling
and testing, in Chapter 3.2 – Mine Development, in Surface Mining
(ed: E P Pfeider), pp 103–121 (The American Institute of Mining,
Metallurgical, and Petroleum Engineers: New York).
Mitchell, J and Holowachuk, N V, 1996. Ore blending at Highland
Valley Copper, its necessity, effects, and benefits, in Proceedings
SAG 1996, vol 1, pp 1–24 (Canadian Institute of Mining,
Metallurgy and Petroleum: Montreal).
Morey, M and Cantrell, R, 2011. Ore ageing test work for the Ok Tedi
skarns, in Proceedings MetPlant 2011, pp 615–629 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Muñoz, A, Álvarez, L, Colacioppo, J and Valery, W, 2008. Process
integration and optimisation at Freeport-McMoran Candelaria
mine, Copiapo, Chile, in Proceedings PROCEMIN 2008, Santiago,
pp 303–315 (Gecamin Digital Publications: Santiago).
Mwanga, A, Rosenkranz, J and Lamberg, P, 2015. Testing of ore
comminution behavior in the geometallurgical context – a review,
Minerals, 5:276–297.
Nzama, S C and Kapo, M K, 2014. Managing the giant – a mine
geologist’s perspective to the mining of the Olympic Dam
deposit, in Proceedings Ninth International Mining Geology
Conference, pp 451–459 (The Australasian Institute of Mining
and Metallurgy: Melbourne).
O’Malley, G B and McGhie, R R, 1939. The mineralogy of the Black
Star ore body and its relation to milling practice at Mount Isa, The
AusIMM Proceedings, 116:459–486.
O’Meara, A E, 1953. New microscope laboratory at Mount Isa, The
AusIMM Proceedings, 171:219–229.
Oosthuyzen, E J, 1983. The Application of Automatic Image Analysis
to Mineralogy and Extractive Metallurgy, special publication 7,
pp 449–464 (Geological Society of South Africa).
Palagi, C G and Stillar, S S, 1976. The Anaconda C. E. Weed concentrator,
in Flotation A M Gaudin Memorial Volume (ed: M C Fuerstenau),
second edition, pp 1029–1042 (Society for Mining, Metallurgy &
Exploration – American Institute of Mining Engineers: New York).
Pangum, L S, Lauder, W, Manlapig, E V, Mackinnon, I R and
Glatthaar, J W, 1998. Mineralogical and metallurgical examination
of fluorosilicate mineral flotation in the Ok Tedi Circuit, in
Proceedings Mine to Mill 1998, pp 19–28 (The Australasian Institute
of Mining and Metallurgy: Melbourne).
Petruk, W, 1984. Image analysis measurements and data presentation
for mineral dressing applications, in Proceedings Second
International Congress on Applied Mineralogy, Los Angeles,
22–25 February, pp 127–140 (Metallurgical Society of AIME:
Warrendale, Pa).
Petruk, W and Hughson, M R, 1977. Image analysis evaluation of the
effect of grinding media on selective flotation of two zinc-leadcopper ores, ClM Bulletin, June, pp 128–135.
Petruk, W and Schnarr, J R, 1981. An evaluation of the recovery of
free and unliberated mineral grains, metals and trace elements
in the concentrator at Brunswick Mining and Smelting Corp Ltd,
CIM Bulletin, 74(833):132–159.
Petruk, W and Sikka, D B, 1987. The formation of oxidized copper
minerals at the Malanjkhand porphyry copper deposit in India,
and implications on metallurgy, in Process Mineralogy VII (eds:
A H Vassilou, D M Hausen and D J T Carson), pp 403–420 (The
Minerals, Metals and Materials Society: Warrendale).
Quigley, P, Hackhan, D and Broome, J, 2008. Sepon Copper Mineral
Resource statement, in OZ Minerals Limited Mineral Resources
explanatory notes, 30 June, pp 8–26.
Quiñones, S and Mattos, R, 2001. Metallurgical control through
mineralogy in the Toquepala concentrating plant, XXV Convencion
de Ingenieros de Minas – Arequipa, Septembre.
Ramsey, T, 1976. The Cyprus Pima concentrator, in Proceedings A M
Gaudin Memorial Symposium, pp 1079–1099 (American Institute of
Mining, Metallurgical, and Petroleum Engineers: New York).
52
Reid, A F, Gottlieb, P, MacDonald, K J and Miller, P R, 1984. QEM*SEM
image analysis of ore minerals: volume fraction, liberation and
observational variances, in Proceedings Second International
Congress on Applied Mineralogy in the Minerals Industry, pp 191–204
(Metallurgical Society of AIME: Warrendale, Pa).
Reid, A F and Wittenberg, J C, 1984. Technical note – rapid production
of high quality polished sections for automated image analysis of
minerals, The AusIMM Proceedings, 289:93–97.
Reid, A F and Zuiderwyk, M A, 1983. QEM*SEM: automated image
analysis and stereological applications to mineral processing and
ore characterization, Acta Sterol 1983; 2/Suppl I, in Proceedings
Sixth International Congress, Gainesville, 9–14 October, pp 205–208.
Reno Sales, R H, 1956. Full cooperation between departments is a
must policy for mining enterprises, presented to Symposium on
Greater Cooperation for Metallurgical Planning; Second Annual
Rocky Mountain Minerals Conference, Salt Lake City, October
1955, Mining Engineering, July, pp 709–711.
Richard, K, 1956. Geologist-metallurgist cooperation in porphyry
copper exploration, presented to Symposium on Greater
Cooperation for Metallurgical Planning; Second Annual Rocky
Mountain Minerals Conference, Salt Lake City, October 1955,
Mining Engineering, July, pp 703–705.
Richmond, G D and Campbell, J J, 1992. Innovation at the Hellyer
Concentrator, in Proceedings AuslMM Annual Conference,
pp 191–196 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Richmond, G D and Lai, K F, 1988. Metallurgical development of
the Hellyer Ore, in Proceedings Third AusIMM Mill Operators’
Conference, pp 9–14 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Ross, J, Appleby, S K, Hoal, K and Botha, P, 2009. Quantitative
mineralogical study of ore domains at Bingham Canyon, Utah,
USA, presented to SME Annual Meeting, Denver, 22–25 February.
Saquet, J J, Apraxine, V, Lakaye, J and Troch, P, 1962. Kolwezi mining
and metallurgical operations of Union Miniere du Haut-Katanga,
Mining Engineering, December, pp 70–81.
Siddiqui, F, 2000. On-site mineralogy – SIDMIN program, in
Proceedings 32nd Annual Meeting of Canadian Mineral Processors,
Ottawa, 18–20 January, pp 459–475 (Canadian Institute of Mining,
Metallurgy and Petroleum: Montreal).
Snider, J W, Ibrado, A S, Hester, M G, Earnest, D, Aronson, J G,
Roman, R J, Rehn, C C and Harbort, G J, 2013. Mineral processing
and metallurgical testing, in King-king copper-gold project
preliminary feasibility study, NI 43-101 technical report, Rev 0,
25 February, pp 90–104.
Sola, C and Harbort, G, 2012. Geometallurgy – tricks, traps and
treasures, in Proceedings 11th AusIMM Mill Operators’ Conference,
pp 187–196 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Strohmayr, S J, Barns, K E, Brindley, S K and Munro, P D, 1998.
Mineralogy controlling metallurgy at Ernest Henry Mining, in
Proceedings Mine to Mill 1988, pp 13–18 (The Australasian Institute
of Mining and Metallurgy: Melbourne).
Suazo, C J, Kracht, W and Alruiz, O M, 2010. Geometallurgical
modelling of the Collahuasi flotation circuit, Minerals Engineering,
23:137–142.
Taylor, B, 2011. The Perseverance gold copper deposit – a new Peak
Mines challenge, in Proceedings Eighth International Mining Geology
Conference, pp 367–378 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Tew, A, Strohmayr, S, See, A and Harvey, R, 2003. Mineralogy
controlling metallurgy at Ernest Henry Mining revisited, in
Proceedings Eighth AusIMM Mill Operators’ Conference, pp 211–
217 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Tverdov, A and Nikishichev, S, 2014. Mine design stages in Russia, in
Mineral Resource and Ore Reserve Estimation – The AusIMM Guide
to Good Practice, second edition, pp 815–824 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
WE are metallurgists, not magicians
Integrating geometallurgy with copper concentrator design and operation
Umipig, C V, Israel, E E, Hutalle, G G and Williams, S R, 2012.
Canatuan Cu/Zn flotation metallurgy – dealing with zinc preactivation, in Proceedings 44th Annual Canadian Mineral Processors
Operators’ Conference, pp 155–168 (Canadian Institute of Mining,
Metallurgy and Petroleum: Montreal).
Urbanoski, L S, 1993. Road to recovery – metallurgical improvements
at Brunswick Mining and Smelting Corporation Limited, in
Proceedings International Symposium World Zinc’93, pp 205–212 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Walters, S G, 2008. An overview of new integrated geometallurgical
research, in Proceedings Ninth International Congress for Applied
Mineralogy, pp 79–82 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
White, J and Adair, R B, 1948. Mufulira Copper Mines Limited,
Concentrator, Northern Rhodesia, presented to AIME New York
Meeting, February 1948, technical publication no 2250, Class B,
Mining Technology, January (American Institute of Mining and
Metallurgical Engineers: New York).
WE are metallurgists, not magicians
Williams, S R, 2013. A historical perspective of the application and
success of geometallurgical methodologies, in Proceedings The
Second AusIMM International Geometallurgy Conference, pp 37–47
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Wong, K Y, Breem, A M, Doherty, R J and Phelan, J M, 1991. Flotation
treatment of supergene complex sulphide ores at Thalanga, in
Proceedings Fourth AusIMM Mill Operators’ Conference, pp 205–
212 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Wraith Jr, W and Fulmor, T G, 1964. Anaconda’s Butte Concentrator,
Mining Engineering, May, pp 55–73.
Yeatman, P, 1932. Problems at the Chuquicamata and Braden copper
properties, in Choice of Methods in Mining and Metallurgy: a Record
of Experience in Making Engineering Decisions, pp 1–34 (McGrawHill: New York; London: published for American Institute of
Mining and Metallurgical Engineers).
Young, K A, 1983, Metallurgical aspects of mining – project
development, in Proceedings AusIMM Sydney Branch, Project
Development Symposium, pp 221–228 (The Australasian Institute
of Mining and Metallurgy: Sydney Branch).
53
Contents
Integrated mining and metallurgical
planning and operation
P L McCarthy1
ABSTRACT
To design and operate a successful mine the technical specialists must agree on what
is ore, what is not ore, and how quickly the orebody is to be mined and processed.
These seemingly simple questions defy easy analysis and can push mining engineers,
geologists, metallurgists and others to the limit of their capabilities. The questions
are rarely addressed adequately in feasibility studies or in operational planning. To
optimise the mine and processing plant requires proper determination of cut-off
grades or values, prediction of feed variability on an hourly, daily or longer basis,
right-sizing the mining rate for optimum product quality, understanding the cost of
capital and the expectations of investors, and many other considerations that involve
geology, geotechnical engineering, mining, metallurgy, environmental effects and
economics. This paper addresses the key matters that should be considered as inputs
to the practical issues of metallurgical plant design and operation.
INTRODUCTION
The mine planning process includes the mining, processing, infrastructure,
environmental and social aspects of an operation. It begins with a sequence of
increasingly detailed feasibility studies and continues throughout the life of the
mine through several long-term and short-term planning processes. These processes
all require the selection of a value descriptor that best reflects the value that can be
obtained from the material, which may be expressed as a metal grade, dollar value or
other measure. At various points in the sequence of mining and processing activities
there is a need to choose a particular value of the value descriptor, called a cut-off,
to enable decisions to be made about what to do next with raw or semi-processed
material. Table 1 of The Australasian Code for Reporting of Exploration Results,
Mineral Resources and Ore Reserves (‘the JORC Code’, 2012) provides a useful
checklist of the many considerations that determine whether in situ mineralisation
that is above the mining cut-off is ore (and should be processed) or is not.
Selection of appropriate cut-offs is an outcome of the mine planning and
optimisation process. While approximate cut-off grades may be used early in the
resource estimation process, these are superseded by better, value maximising cut-off
grades as planning progresses. The process is iterative and cut-offs are influenced,
among other things, by the chosen mining and processing rate. The two variables,
cut‑off and mining processing rate are the primary levers that can be used by planners
to optimise the mining and processing plan.
The selected mining rate determines how much of the above cut-off material can
practically be recovered, how much below cut-off material is included in the feed sent
to the processing plant, and the quality and variability of that feed on various time
scales. Slow mining allows careful selectivity; fast mining generally reduces recovery
and increases dilution.
MINING CUT-OFF
The optimum rate of mining is closely tied to the cut-off grade chosen. Hall and Hall
(2006) observe that:
1. HonFAusIMM(CP), Chairman, Principal
Mining Consultant, AMC Consultants Pty Ltd,
Melbourne Vic 3000.
Email: pmccarthy@amcconsultants.com
…the major parameters that a mining company can make independent decisions about
are typically the mining method(s), mining sequencing, production rate, and cut-off
grade (or ‘cut-off’). Since the size and shape of the orebody and hence possible mining
methods and the range of feasible production rates may vary significantly with cut-off,
it is often the cut-off that is the key driver of value of the operation.
This is illustrated in Figure 1, which shows that a smaller operation optimised
using a higher cut-off grade is more robust than an operation optimised using high
price assumptions.
55
P L McCarthy
mine more carefully and selectively. While quick progress
can be made in a single year, longer term averages should be
considered when sizing the processing plant.
The floor of an open pit can be advanced quickly but, as
a rule of thumb, the open pit production limit is around
eight benches per annum, or 80 m vertical advance using
ten metre benches. Though some very large pits and small,
short-life gold pits can achieve over 100 m per annum within
the operating area. However, an average sink rate of 40–60 m
per annum is more likely for longer term planning in most
pits.
The vertical rate of advance in a pit is influenced by the:
FIG 1 – Risks and rewards of optimum cut-offs (Hall and Hall, 2006).
CONSIDERATIONS ON MINING AND PROCESSING RATE
A design rate of mining and processing is selected in every
mine feasibility study, although any attempt to optimise that
rate is rarely documented. To maximise return on investment,
it has long been recognised that both the capital investment
per unit of output and the operating cost per unit of output
should be minimised. In general, both cost measures decrease
as the scale of the project increases, so the initial temptation is
to ‘push the orebody to the limit’.
However, the technical and commercial risk both increase as
the scale of the project increases. Hoover (1909) states that the
lower the production rate, the lower the required investment,
the longer the income stream and the lower the risk to the
investor. While this was well before the advent of discounted
cash flow (DCF) analysis, the point made by Hoover remains
a good one.
Until the last third of the twentieth century, most mine
developers did not have ready access to project capital and
so they had to develop projects using mainly a combination
of new shareholders’ funds and retained earnings. Available
capital was a key consideration in sizing a new project. More
recently, and particularly during the ‘mining boom’, there
has been an assumption that any scale of project will attract
project finance if it satisfies the hurdles set by bankers. Capital
constraints were not commonly included in the project
optimisation process. Bigger projects were generally thought
to be better. In reality, the interests of existing shareholder
owners of a mineral deposit may be best served by a modest
scale of development, with restricted use of external capital.
The value of a smaller project as measured by net present
value (NPV) may be lower, but the risk-adjusted value to
existing shareholders may be greater.
The uncertainties in mining investment are many. Mineral
prices are cyclical and to a large extent unpredictable. Over
the life of a mine, these prices usually fall in real terms. The
mineral resources being mined are finite and can be highly
variable in both size and quality, while ground conditions can
vary significantly with depth and location. Costs are difficult
to predict and subject to periods of rapid escalation. These
uncertainties tend to favour a more modest capital investment
and hence exposure to risk, provided the expected product
cost falls within an acceptable range.
MINING RATE LIMITATIONS
There are physical limits to the rate that any orebody can
be mined. High rates of mining are associated with greater
day-to-day or month to month production volatility, with a
tendency for dilution of the ore to become excessive at high
rates as the pressure of production reduces the ability to
56
•• number of activities that are included in the mining
cycle, eg grade control, presplitting, drill and blast,
loading
•• rate at which each activity can be carried out, as
determined by the size and number of equipment
•• total available floor area
•• bench height, as this determines the vertical advance per
mining cycle
•• need to develop a drop cut to establish the bench, or a
sump for dewatering, as is the case in mining in the base
of pits compared to a mining a cut-back.
Once the vertical advance rate is established, the average
production rate can be determined from the available tonnes
of ore per vertical metre (t/vm) within the pit design.
Similar considerations apply to underground mines. The
production rate from an underground mine is not usually
limited by the rate at which ramp or decline development can
be advanced, as a decline face can typically be advanced at
40 m per week on a 1-in-7 (14.3 per cent) grade, which is a
vertical advance rate of 297 m/a. Rather, the production rate
is limited by the number of available working faces which
in turn depends on the amount of predevelopment, possible
rates of ongoing lateral development, infill drilling, stope
turnaround times, backfilling and so on, with interference
between these activities.
In steeply dipping deposits underground mining can occur
on several levels simultaneously, but the mine production
rate can still be related to the ore t/vm that will be mined. This
relationship can be expressed as the ‘effective vertical advance
rate’, or the relationship between actual mining rate and the
t/vm available in the deposit. For example, a 1 000 000 t/a mine
with an average 20 000 t/vm would have an effective vertical
advance rate of 1 000 000/20 000 or 50 vertical metres per
annum (vm/a).
Work by the author and others has shown that the risk of
failure increases as the vm/a increases. A century ago, a rate
equivalent to one level per annum or approximately 30 vm/a
was considered a reasonable basis for planning. Tatman
(2001) was able to derive an empirical formula relating the
risk of failure to the geometry of the deposit and the rate of
mining and to conclude that for modern mines in steeply
dipping tabular deposits thicker than 10 m, risks were:
<30 vm/a
low risk
30–70 vm/a
moderate risk
>70 vm/a
high risk.
In a 2014 confidential study of 12 current Australian mines
using sublevel open stoping methods, the author found that
average vertical advance rates varied from 23–71 vm/a with
an average of 43 vm/a, while the maximum single-year rates
varied from 31–79 vm/a with an average of 56 vm/a. Only one
mine sustained a rate higher than 61 vm/a. As a generalisation,
We are metallurgists, not magicians
Integrated mining and metallurgical planning and operation
special circumstances are required for any underground mine
to sustain a rate above 60 vm/a.
These observations describe the limits that might apply
to mining rates for open pit and underground mines.
The optimum rate can only be determined after detailed
scheduling of alternative mining plans and the completion of
an optimisation study that balances revenue against capital
and operating costs for the entire mining and processing
operation.
MINING RATE AND HEAD GRADE
The assumption that ‘economies of scale’ will result from
increasing throughput rates needs to be balanced by an
awareness of the adverse effects of increasing the rate beyond
a level that is supportable by the resource. For each scale of
operation considered, it is a reality that for any intended head
grade, at the associated intended cut-off grade, the actual
head grade achieved will fall as the mining rate increases. This
effect is known to people at operations but is not generally
recognised in current ore reserve estimation methodology.
Once recognised, this dependence of head grade on mining
rate can be quantified and used to establish the economically
optimum mining and processing rate (McCarthy, 2010).
MINING FLEXIBILITY
It is unfortunately true that a mining operation presents
challenges on every time scale from daily to annually. The
failure of many ‘mine to mill’ studies to deliver promised
improvements is largely because the mining engineers
are fully absorbed in meeting existing challenges and have
limited capacity to vary the method and sequence of mining
to deliver a better or more predicable product to the process
plant. Monthly, quarterly or annual mining schedules are
based on the ore reserve model, which is frequently found to
be deficient at those scales. Changes to the mining sequence
are made on the run, and the challenge for the mining engineer
is to deliver the scheduled tonnes, of any quality, above the
mining cut-off.
Some mining methods, such as sublevel and block caving,
allow of no short-term control of product quality, although
differential draw may be possible on a timescale of years.
Provided it is above the mining cut-off, in the short-term the
ore is delivered as it presents at the drawpoints. While other
methods in both open pit and underground mines can be more
selective, the pressure to deliver the tonnes to schedule often
precludes any management of product quality by scheduling
a blend of material from several mining areas. Of course, the
higher the rate of mining, the less selective the mining can be.
ORE STOCKPILES
Nineteenth-century mines used hand-pushed rail trucks,
typically of around 400 kg capacity, to move ore from chutes to
the shaft and up to surface. Each truck could be marked with
chalk as to its source. Stamp batteries were fed separately in
groups of five stamp heads. It was possible to blend ore from
different sources to achieve a steady quality in each stamper
box. Alternatively and more commonly, parts of the mill were
tuned to run on ore from different stopes or orebodies. In a
gold mill, high-arsenic ores might be run over strakes and
gravity tables into barrels, while clean ores were first mercury
amalgamated on copper plates. Thus the mine and mill were
closely aligned, with feedback on ore grade being given to the
miners for individual stopes or areas. The many ore trucks
formed a stockpile which decoupled the stopes from the
hoisting and processing system. Old illustrations show trucks
lined up at the plat, waiting to be hoisted, with empties ready
to be refilled (Figure 2).
By the mid-twentieth century orepasses acted as stockpiles
at each stope to the haulage level, and from there, in parallel
with the hoisting shaft, down to a skip-loading station. This
allowed for a great deal of blending of ore from different
sources, while decoupling the mining activity from hoisting
and processing. The ability to reconcile stope grades and to
run parallel lines through the mill had been lost, but cost
savings and greater throughput were achieved.
Today many mines are designed without significant
stockpiles. Block caving mines in particular may have no
effective stockpile capacity between the drawpoint and
the surface stockpile. Even the surface stockpile may be
eliminated in normal operation, with the inclined conveyor
from the underground mine delivering crushed ore directly
to the secondary crusher. In these circumstances the miners
have no ability to manage product quality. In other operations
such as sublevel open stoping or longhole retreat stoping
without shaft hoisting, the old approach of using orepasses as
stockpiles has been eliminated and ore is hauled to the surface
in trucks. This provides an opportunity to manage ore quality
by blending using two or more dump points at the run-ofmine (ROM) stockpile (or dump at a low-grade stockpile for
future processing). Management of ore quality then becomes
the responsibility of the reclaim operator, but may require
resampling to establish the variability and location of material
within the stockpiles.
The more challenging the mining situation, the greater the
stock levels need to be including developed (exposed) ore
stocks, drilled stocks, broken stocks and ROM pad stocks. If
these stock levels are adequate then variability and volatility
There is evidence that more selective mining is possible if it
is designed into the process from the outset and the planned
mining rate is adjusted downward to allow it to take place.
A good example is the selective mining of acid-generating
and benign waste from a pit, with the benign waste used for
capping. This is a legislative requirement which cannot be
compromised, so it is given appropriate attention.
The best way to achieve predictable feed for the process
plant is to develop an accurate orebody model in which all
the variables of consequence are modelled faithfully, and to
set the mining rate low enough so that selective mining can
be practiced on every time scale. This requires competence in
the emerging specialty of geometallurgy, a healthy geological
budget for drilling and modelling, and an uncommon
appreciation of the benefits of mining at a rate lower than the
maximum possible.
We are metallurgists, not magicians
FIG 2 – Ore trucks at the plat (from Dicker’s Mining Record, 1867).
57
P L McCarthy
can be reduced to a minimum. The mine should be designed
so that all stockpiles, including orepasses in an underground
mine, have adequate capacity to smooth the short-term
surges to a level acceptable for the entire system, including
ore processing. This is a commonly overlooked requirement.
variability in ore quality or of the distribution in space of
valuable material, process contaminants, ore hardness and so
on. At best, the mining schedules are presented as monthly
averages for the initial few years and as quarterly or annual
averages thereafter.
Some orebodies are amenable to visual grade control; others
require assays on a short turnaround to allow ore selection
decisions to be made. The latter can suffer from dilution and
high-grade volatility if the grade control program is not well
designed and given priority at the laboratory.
As initially constructed, the processing circuit must be
designed to cope with or be adapted to the expected range
of ore qualities, with an ability to respond quickly to any
changes. Alternatively, with a less flexible circuit, the cost
and revenue impacts of possible variations in ore quality
must be examined to ensure they fall within acceptable limits.
The impact may be greatest in the first year, when orebody
knowledge is weakest and cash flow is critical. As experience
is gained during mining, and infill or grade-control drilling
advances, the orebody model can be greatly improved and
the processing circuit adapted to reality.
The reliability of mining equipment has an effect on ore
quality. Delays in mine development (accessing ore in an
underground mine or prestripping in a pit) can lead to
periods when low-grade or high-impurity ore is all that is
available. Breakdowns in ore-production equipment can
lead to increased dilution because it is human nature to be
less concerned about dilution when there is insufficient ore
available to feed the mill. Hence the old saying ‘waste plus
ore equals more ore!’
Proper mine design, planning, scheduling and maintenance
require good management. Ultimately, the capabilities of
the mine management team will determine the quality and
regularity of mill feed.
PROCESSING RATE
Unlike the mining rate, the processing rate is not physically
limited by geology. A plant of any capacity can be built,
at a cost, although there are step changes in the capacity
of available components that make particular rates more
attractive. Sometimes environmental constraints limit the
hours for activities such as stockpile reclaim and crushing
but this can be addressed by using larger equipment. The
availability of services such as power or water may place an
absolute limit on the size of plant, or impose a large capital
cost burden for going beyond that point.
Once a plant is operating, the processing rate may be limited
by feed characteristics. For example, harder ore than expected
may limit the milling rate while wet, clayey ore may limit the
crushing rate. The copper to sulfur ratio may limit throughput
in a smelter. For these reasons plant components and the
overall plant capacity may be oversized to some extent as
compared to the selected mining rate. If this is recognised by
management as an allowance for variable ore quality then no
problem arises, but invariably when the ore quality is good
the higher capacity is pushed back to the mine as a demand
for a higher mining rate, with adverse consequences.
For some products the production rate may be limited
by the market. The output of large individual producers of
products such as iron ore may influence the market price, so
that the optimum production rate is less than it would be in
an unlimited market. In some circumstances there may be a
limit to the road or rail haulage capacity for concentrate, or
the availability of ship loading facilities.
OREBODY MODELLING
A resource model is initially based on exploration drilling only.
Its accuracy is limited by the exploration budget, by difficulty
of access to the mineralisation, and by a lack of experience
with the deposit. This resource model is used by mining
engineers to design a mine at the feasibility study level, with
their predictions of mining dilution and mining recovery used
to estimate an ore reserve. The mining schedules produced
in the feasibility study are used by the plant engineers and
metallurgists to design the plant. At this stage there may be
a very poor understanding of the hourly, daily or weekly
58
The planning process begins with a good 3D geological
(resource) model developed by an experienced geologist
familiar with the deposit type. Geological domains are
identified, such that a common set of rules can be applied to
determine local variations in metallurgical responses within
each domain. The domain boundaries may be structural,
mineralogical, alteration or lithological. Poor geological
modelling and domaining are the leading causes of failure
in geostatistical modelling for grade estimation and for
modelling metallurgical parameters.
Domains should be defined beyond the ‘orebody’ to include
all material that could find its way into the ore stream.
Metallurgically, adjacent domains may have little or nothing in
common. For example, the waste rock adjacent to the orebody
across a sharp contact may be much harder and more abrasive
than the ore, and if the mining method will cause 25 per cent
mining dilution of ore with waste rock then the crushing,
grinding and wet plant performance of the waste rock must
be thoroughly understood. Similarly, dilution from nearby
carbonaceous shales (such as a hanging wall zone) may be
preg-robbing or a talc-rich fault zone may impact on filtration,
although neither is considered part of the orebody proper.
Once the domains (both within and near the orebody)
have been described (as ‘wire frames’ or solid objects in a 3D
computer model), representative samples from each domain
can be subjected to laboratory-scale test work to determine
the rock’s response to each mineral processing operation.
Conventional geostatistics can then be used to model the
distribution of metallurgical responses throughout each
geological domain. This results in a model comprising a large
number of blocks in 3D space, each block being assigned
all the geological, geotechnical, geometallurgical and other
characteristics needed for mine planning and scheduling.
The mine scheduling process can then produce from the
geological block model not only a schedule of tonnes and
grade but also a schedule of metallurgical performance and
other characteristics such as ground support requirements or
even water inflows to the mine.
The geostatistical approach used to model metallurgical
performance need not be complex. Even the simplest
approaches using the ‘rule of nearest neighbours’ (which says
each block in the model is likely to perform in a similar way
to the nearest sample) or the ‘rule of gradual change’ (which
calculates a distance-weighted average of characteristics
based on nearby samples) is likely to provide a significant
improvement in predictability of plant performance when
compared with having no geometallurgical model. However,
the more advanced geostatistical methods are not difficult to
apply and will further refine the result. Selection of the best
techniques is the subject of ongoing research.
We are metallurgists, not magicians
Integrated mining and metallurgical planning and operation
Samples for metallurgical testing are usually composited
from diamond drill core. Hardness testing (for crushing and
grinding) typically requires 10–20 kg of sample, with some
tests requiring 100 kg, while flotation testing usually requires
at least several kilograms of sample (Barratt and Doll,
2008). Large metallurgical samples excavated from near the
surface of a deposit are unlikely to be representative of the
orebody at depth. Shafts sunk for the purpose of obtaining
large metallurgical samples may also yield unrepresentative
samples, or samples that represent performance in only
one geological domain. Such exercises may be compared to
searching for a lost wallet under a streetlight, because it is too
dark to search elsewhere.
In order to be useful in developing a geometallurgical
model, test results must satisfy the following (ibid):
•• results must reflect the properties of a ‘small’, identifiable
interval of drill core
•• the location of the interval must be identifiable in 3D
space (to connect it to the block model)
•• the values being distributed through the orebody must
be reasonably additive, allowing unknown blocks in
the model to be estimated by interpolating two or more
known samples.
Sufficient sample material to achieve these aims may be
available from drill core for a large porphyry copper open pit
which has a large selective mining unit (SMU or minimum
mining block size). For smaller, more complex deposits and
many underground mines where assays are obtained for
each one-metre sample interval, the production of composite
samples of sufficient size for metallurgical testing may defeat
these aims and blur the modelling results. For example Barratt
and Doll (2008) propose sample intervals of 15 m of HQ
(63 mm) core for a JK SMC® (drop weight) test and 45–105 m
of HQ core for a Bond test. It may be necessary to develop local
correlations between the large-scale tests and other properties
such as point load strength, rock quality designation (RQD),
fracture frequency and mineralogy in order to obtain sufficient
data to create a meaningful 3D model.
The new techniques of geometallurgical modelling
are useful in improving process plant performance and
predictability in large orebodies such as porphyries, mined
with big equipment. In these situations zoning may allow
prediction of changes over a period of years. However, many
medium to small-scale mines may not be amenable to the
techniques being developed. The problem is that variability
of the factors affecting metallurgical performance occurs
at a scale smaller than can be sampled for metallurgical
testing, and at that scale the measured properties are not well
correlated with performance. The same problem exists for
geotechnical modelling, where very limited success has been
achieved in predicting ground conditions and stability using
geostatistical methods.
mined, stockpiled and processed separately. The premining
block grades were estimated from diamond drill assays by
geostatistical methods using a 150 g/t Au top cut. Overall, Gill
Reef was estimated from drilling to have a grade of 7.5 g/t Au
which reconciled (for comparable blocks) with a mine head
grade of 7.6 g/t Au. Individual block reconciliations, with a
linear best-fit line, are shown in Figure 3.
It was possible to reconcile mill recovered grade against
mine head grade for 45 blocks ranging from 2064–9799 t, with
a mean size of 4426 t. Head grades ranged from 2.2–12.6 g/t
with a weighted mean of 7.6 g/t. Reconciled metallurgical
recoveries ranged from 71.5–96.7 per cent with a weighted
mean of 89.1 per cent. Individual block reconciliations, with a
logarithmic best-fit line, are shown in Figure 4. A constant tail
grade model was not a good fit to the data.
With sufficient experience in similar orebodies it should
have been possible to predict from the drill results that mining
about 240 000 t of ore would yield a head grade of about
7.5 g/t Au and a metallurgical recovery of about 90 per cent.
However, these long-term averages would be of little use
in predicting or optimising process plant performance on
a daily or weekly basis at the actual mining rate from Gill
Reef of 2000–3000 t per week. Figure 5 shows the predicted
metallurgical recovery for each mining block (based on the
predicted head grade and the grade-recovery relationship)
and the actual metallurgical recovery, with a weak linear bestfit relationship. It is clear that metallurgical performance was
not predictable at a useful level.
While attempts could be made to improve the metallurgical
model by modelling other geological features such as
carbonaceous shales or associated sulfides, the innate
geological variability makes success unlikely at the actual
scale and rate of mining. At much higher mining rates, or
FIG 3 – Predicted and actual block grades, Gill Reef.
CASE STUDY – PREDICTING THE UNPREDICTABLE IN GILL REEF
The Bendigo orebodies have been shown to suffer from an
extreme nugget effect, making prediction of grades from drill
data difficult. For this reason Gill Reef was mined, processed
and reconciled in small batches and provides a more detailed
picture of ore variability and plant performance than is
usually available. The author examined the data to see
whether metallurgical recovery could have been predicted
from the predicted head grade of each batch.
Gill Reef is a distinct quartz reef at the Kangaroo Flat
mine that was mined between 2009 and 2011. A total of
55 discrete ore blocks totalling 243 497 t from Gill Reef were
We are metallurgists, not magicians
FIG 4 – Head grade and metallurgical recovery, Gill Reef.
59
P L McCarthy
•• Surface mines suffer significantly higher variability in
ore production from budget than underground mines.
This is surprising, but reflects the ability to switch the
mining fleet from ore to low-grade to waste in reaction to
changed circumstances, including geological variations,
weather and equipment downtime. Another reason may
be that open pit schedules are less conservative than
underground schedules.
•• Despite the mining variability, processing rates in
surface mines run closer to budget, perhaps due to
having larger stockpiles.
•• Variability in mined and processed grades is similar,
suggesting that little use is made of stockpiles for
blending and that grade control outcomes are similar in
the surface and underground mines studied.
FIG 5 – Predicted and actual metallurgical recovery.
when considered over longer periods, the volume-variance
relation would allow greatly improved predictability.
VOLATILITY AND VARIABILITY
Production volatility refers to the relative variation in a
parameter from one time period to the next, while variability
refers to the variation from budget or plan from one time
period to the next. The volatility of parameters such as feed
tonnage, head grade, metallurgical recovery, throughput or
product output can be measured hourly, daily, monthly etc.
The more volatile the measure, the less use is being made of
the installed capacity and hence of the capital invested and
of the fixed component of operating cost. One of the key
symptoms of a system that has been pushed beyond its stable
capacity is an increase in production volatility.
A mining project designed for a 1.0 Mt/a rate with
five per cent volatility needs an installed capacity of
1.05 Mt/a. If a decision to increase the rate by ten per cent
to 1.1 Mt/a leads to an increase in volatility to 15 per cent,
then the installed capacity must be increased to 1.27 Mt/a,
an increase of 20 per cent. If the capacity is only increased
by ten per cent, the increased volatility could lead to a slight
reduction in output when the mine and mill are considered
together as a system.
Another useful measure is volatility, the average percentage
change in a measure from one time period to the next. The
level of planned and actual volatility drives the stock
requirements and levels. When volatility is low, mining and
processing are efficient and capacity is being used effectively
with costs minimised.
Table 2 shows the month to month volatility from the
benchmark study mentioned above. It can be concluded that:
•• ore tonnage mined is nearly twice as volatile in surface
mines compared with underground mines
•• on other measures, underground mines are more volatile
than surface mines.
Carter (2010) observes that volatility and variation in any
business process creates uncertainty, whether in determining
mining volumes and plant feed or in optimising maintenance
schedules and supply chain management. The same holds for
the management of working relationships. It is this variation
and volatility which skews outcomes, and which can be
reduced by implementing rigorous planning, scheduling,
resourcing and execution processes, and most importantly,
by clarifying roles and accountabilities at each level.
The specifications for the processing plant should reflect
real hourly or daily mining outcomes, not a smoothed and
idealised schedule. Table 1 summarises the unpublished
results of benchmark studies of 44 underground mines
and 21 open pit mines conducted by AMC Consultants.
The monthly variability shown is the average absolute
difference (as per cent of budget) between the mine budget
and production over a 12 month period. High variability
indicates that a mine is not operating as intended, and that
daily variability may be much greater than the steady hourly
throughput for which the plant was designed. It can be
concluded that:
When volatility and variability are assessed on an hourly or
daily basis they are seen to be much greater than the monthly
measures in Tables 1 and 2. For example, Figure 6 shows
daily plant tonnes at Anglogold Ashanti’s Mponeng mine.
The figure also shows the results of a business improvement
initiative which contributed to a 15 per cent increase in
throughput over the historical average. This is an especially
significant result at the Mponeng plant, long regarded as the
flagship operation within the group. Before the initiative,
ore from the Mponeng mine would regularly be trucked to
neighbouring plants for processing, as the mill struggled to
cope. After the initiative the plant had improved productivity
to the point that it now had spare capacity. Emphasis on
stabilised processes also resulted in a 20 per cent reduction in
sodium cyanide consumption.
Table 1
Monthly variability from budget over a 12 month period.
Table 2
Month-to-month volatility over a 12 month period.
Underground mines
Underground mines
Surface mines
Min
(%)
Max
(%)
Ave
(%)
Min
(%)
Max
(%)
Ave
(%)
Ore mined (t)
5
43
14
7
68
29
Head grade
7
33
13
3
38
11
Ore processed (t)
5
54
12
4
39
Processed grade
3
33
13
2
31
60
Surface mines
Min
(%)
Max
(%)
Ave
(%)
Min
(%)
Max
(%)
Ave
(%)
Ore mined (t)
6
34
14
11
62
25
Head grade
5
28
11
3
22
9
9
Ore processed (t)
3
32
13
4
22
10
10
Processed grade
5
23
12
3
16
8
We are metallurgists, not magicians
Integrated mining and metallurgical planning and operation
FIG 6 – Mponeng mine daily plant tonnes (over 12 months) (after Carter, 2010).
OPTIMISING THE MINE AND PROCESSING PLANT
MANAGING THE DOWNSIDE RISK
The key parameters over which a mining company has
control are the size of mine (large or small, low or high-grade,
as determined by the cut-off grade), the mining method, the
production rate, the mining sequence, the processing method
and the amount of money that will be spent on getting these
things right, which includes exploration drilling, geological
modelling and bench and pilot plant testing. Other aspects
such as power supply, water supply, concentrate transport
and logistics generally have a more obvious engineering
solution and are ancillary to the optimisation process.
The need to manage risk was well understood in the past. A
small project was built, often with second-hand plant, and
then cash flow from the operation, or equity funding from the
now-reassured investors, was used for a series of expansions
and optimisations. If there was a problem with the initial ore
reserve or cost estimates, the exposure of shareholders to this
problem was minimised and managed.
The objectives of optimisation must be aligned with the
corporate objectives of the owner. Some stated corporate
objectives, such as maximising annual ounces of gold
production or maximising mine life, cannot be optimised.
Clearly, increasingly large subeconomic projects will satisfy
the former objective while decreasingly large subeconomic
projects will satisfy the latter. Short-life projects carry the risk
that most of the production will be delivered into a trough
in the product price. Sensitivity analysis based on a range
of price scenarios will identify the parameters that yield an
acceptable risk.
There is also the problem of capital allocation between
competing projects. If there is no restriction on the available
capital then corporate value is maximised by maximising the
NPV of every available viable project and carrying all of them
through to production. In the real world, where available
capital is restricted, the corporation must select projects
for investment using some ranking technique. Economic
theory says that projects should be ranked using the present
value ratio (PVR), which is the ratio of NPV to initial capital
investment. For simplicity, the capital investment is usually
taken to be the total of negative cash flows prior to achieving
positive cash flows. If the perceived risks are similar, projects
with higher PVRs are selected before those with lower PVRs.
A project with a high NPV but a low PVR may require more
capital than the corporation (or the investment community)
is able or willing to risk, or if developed it may displace
alternatives which would have provided a better aggregate
return on investment.
From the above, the project should be designed to maximise
the project NPV at the corporation’s agreed discount rate,
provided this leaves it with a PVR that will make it an
attractive investment. Arguably, the plan should be changed
to improve the PVR, even at the expense of NPV, if this will
allow the project to proceed in competition with others.
We are metallurgists, not magicians
The author was involved in a large gold project that had
high geological risk. It could be initially developed using a
nearby idle plant, either by purchasing the plant for a modest
sum or by toll milling, before building a full-scale plant.
This plan did not give a satisfactory project NPV because it
delayed full-scale production by several years. A compromise
was found whereby a small plant would be built on-site
and later expanded. However, as the time for construction
approached, deteriorating estimates again made the NPV
unattractive. This was addressed by deleting the smaller stage
and building the larger plant immediately. This enhanced
the NPV substantially in spreadsheet models. When the
geological problem proved intractable the project failed, with
capital losses around three times what they might have been
with the original toll-milling proposal.
This outcome might have been avoided using the value at
risk (VaR) approach which is widely used in the financial
sector. VaR is the maximum loss not exceeded with a given
probability defined as the confidence level, over a given
period of time. It is commonly used by security houses or
investment banks to measure the market risk of their asset
portfolios (market VaR), over time periods of one day to a few
days. However VaR is a very general concept that has broad
applications.
For example, a Monte Carlo approach to modelling net cash
flow outcomes for a particular project development option
might show that 95 per cent of outcomes have a net cash
result better than minus $50 M. In other words, the cash loss is
expected to be greater than $50 M only five per cent of the time.
This approach must have a constrained time period applied,
such as the time to project payback or a fixed number of years.
The various project development options can be modelled
and a decision made based on both the expected NPV and the
VaR for each development option. Ultimately, the corporation
must be able to absorb and manage the ‘worst-case’ outcome.
CONCLUSIONS
The design of metallurgical plants should be undertaken
in the context of the broader optimisation of the mining
61
P L McCarthy
project. Often the appropriate metallurgical process can be
selected early in the optimisation process and thereafter the
plant design is simply a matter of good engineering, without
strategic options. By contrast the size of mine (large or small,
low or high-grade, as determined by the cut-off grade), the
mining method, the production rate and the mining sequence
are strategic decisions which form critical inputs to the
engineering design of the metallurgical plant.
These strategic decisions should be made by a team
which includes members who have an appreciation of the
implications of those decisions to every part of the proposed
operation. Subsequent engineering design for the mine,
processing plant and infrastructure can then proceed within
appropriate constraints.
In recent years, many project investment decisions have
been made on the assumption that unlimited project finance
is available. Due to global economic circumstances this is no
longer the case, and a more traditional approach to project
optimisation is called for. After consideration of risk, a
modest-sized, staged development may provide better
shareholder returns than the largest project that an orebody
can theoretically support. Staged development may require
multiple parallel processing circuits and smaller, more
selective mining machines operating at higher cut-off grades.
The VaR approach is one way of quantifying the downside
when considering alternative project scales and development
paths. Its use, in combination with tradition NPV analysis,
provides a more complete picture of the options available to a
company when it sets out to develop an orebody.
ACKNOWLEDGEMENTS
The author acknowledges AMC Consultants Pty Ltd for
permission to publish benchmarking results and Unity
Mining Limited for permission to publish the Gill Reef data.
REFERENCES
Barratt, D J and Doll, A G, 2008. Testwork programs that deliver
multiple data sets of comminution parameters for use in mine
planning and project engineering, in Proceedings Procemin 2008,
Santiago, Chile.
Carter, C, 2010. Project one holistic transformation plan to achieve
strategic goals, Anglogold Ashanti.
Dicker’s Mining Record, 1867. Dicker’s Mining Record and Guide to
Gold Mines of Australia (newspaper), 28 November.
Hall, B and Hall, A, 2006. Doing the right things right: identifying and
implementing the mine plan that delivers the corporate goals, in
Proceedings International Mine Management Conference 2006, CDROM (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Hoover, H C, 1909. Principles of Mining, pp 153–160 (McGraw-Hill:
New York).
JORC Code, 2012. Australasian Code for Reporting of Exploration
Results, Mineral Resources and Ore Reserves (The JORC Code)
[online]. Available from: <http://www.jorc.org> (The Joint Ore
Reserves Committee of The Australasian Institute of Mining and
Metallurgy, Australian Institute of Geoscientists and Minerals
Council of Australia).
McCarthy, P L, 2010. Setting plant capacity, Transactions of the
Institutions of Mining and Metallurgy, Mineral Processing and
Extractive Metallurgy, 119(4):C184–C190.
Tatman, C R, 2001. Production rate selection for steeply dipping
tabular deposits, Mining Engineering, October, pp 62–64.
62
We are metallurgists, not magicians
Project economics
Contents
Guidelines for economic evaluation of projects
P Card1
ABSTRACT
‘Let’s get discipline and quality into easy to understand evaluations!’
Project managers and metallurgical plant operators accept poor quality evaluations
because they are generally unaware of what they should expect.
To provide guidance, the Australasian Institute of Mining and Metallurgy (the
AusIMM) established a subcommittee of practitioners that collated evaluation
systems and procedures regarded as best practice. This has been developed into a
free website for the greater mining industry: <www.economicevaluation.com.au>,
which is maintained by the author. It offers a range of modules and worked examples
that describe in plain language, the step by step sequence of conducting an economic
evaluation.
These practices apply directly to projects and metallurgical plant evaluations,
including those with a heavy technical basis. They can be applied to technical problem
solving where there are no monetary computations.
Project managers and metallurgical plant operators should demand that economic
models and evaluations follow the six principles. They should become working tools
which are easy to understand, fit for purpose, consistent, rigorous, record sources of
input data, have key graphs and are rapid to audit.
INTRODUCTION
Poor workmanship is common but usually accepted
If industry made an award for the worst performance in metallurgical plant project
design and plant operation then the odds-on favourite would be the economic
evaluation!
If an economic evaluation is sophisticated, complex and terribly clever with Excel so
that only one or two experts can use it, then it probably is worst practice. But if anyone
with only basic knowledge of evaluation can readily follow it, sees the correct data
being employed and feels it is easy to understand then it is on the way to best practice.
Project managers and plant operators need to take control and demand economic
evaluations they can quickly follow during their busy working day.
The world’s best practices are readily available on the internet and are in active use
in the mining industry. But for historical reasons most project managers and plant
operators will accept poor quality work in this arena. What is accepted for economic
evaluation would not be tolerated in its sister disciplines of geology, mining,
metallurgy, engineering and accounting. Fortunately this is rarely due to sloppy
management by project managers and plant operators, but due to lack of awareness
of what they should expect and demand.
As a horrible start, most professionals do not even call it by its correct name of
‘economic evaluation’ but talk of ‘financial modelling’ or even more incorrectly
‘financial analysis’ (more later). Project managers and plant operators would insist on
metallurgists doing the processing study work, insist on engineers doing the design
and Engineering, Procurement, Construction Management (EPCM), would want
accountants to do the books, but probably would accept almost anyone willing and
able to perform the economic evaluation. This person probably would be allowed to
do the evaluation, however they believed was best. This is because the discipline of
economic evaluation has evolved over recent decades to bridge between operations/
engineering and accounting without an academic or professional foundation.
1. MAusIMM, Consultant – Economic
Evaluations, Aspendale Vic 3195.
Email: mpcard@tpg.com.au
Older style project managers and plant operators see economic evaluation as a back
room activity to be hurriedly completed, typically when the last of the cost estimates
are finished the night before the submission is due to management. They see the
activity as a mathematical computation to fill in the paperwork with net present value
(NPV), internal rate of return (IRR), payback etc. Fortunately, these are a dying breed.
65
P Card
These older style managers do not really understand
they are designing or operating a business, but live in a
closed world of professional engineering or hands-on plant
operating. They are very confident that they are ‘working on
the important stuff’ and ‘getting things done!’ They do not
realise that contemporary managers demand an economic
evaluation up and running from Day One as a tool to steer the
project or operating plant through the study process into the
optimum state of business. The economic evaluation specialist
should be the second-best role in the team; after the leader.
Mantras of best practice
A few mantras of best practice in economic evaluation are:
•• ‘If you do not readily understand and comprehend my
evaluation then you do not have a problem, I do!’
•• ‘Every worksheet should be as easy to read as a school
text book!’
•• ‘Do not try to impress with sophisticated Excel functions,
but use your intelligence to convert complex interactions
into simple steps on the worksheet.’
•• And definitely not ‘Trust me! I am the expert in
evaluation modelling!’
Financial analysis versus economic evaluation
Completion of a project requires two separate money-focused
activities:
1. economic evaluation
2. financial modelling.
The first, economic evaluation, is all about understanding
the business health of the project. What cash will be required
to establish, operate and pay taxes versus the cash generated
or saved by the project. This is the simple economics of cashin and cash-out over time. Economic evaluation does not
worry where the cash comes from (that is, financing) but
rather it wants to understand the cash generating power
of the underlying project including what sort of prices and
operating environment is required to cover costs and generate
an economic return (the discount rate is usually before
financing).
Everything is computed in cash in the year it is actually spent
or received. There is no accounting depreciation, no accounting
charges for future closure and no other non-cash items.
There is no equity raisings nor company borrowings, and
so no financing charges on borrowings during construction.
Taxes are included and computed before the apparent benefit
of debt.
The second, financial modelling, is all about sourcing the
cash to establish and run the project until it becomes selfsustaining. What mix of company internal cash flow, debt,
new equity, convertible notes, derivatives and hedging will
the company use to progressively pay for owner’s costs,
EPCM, first fills, commissioning, ramp-up to commercial
operations and interest on these borrowings?
How will each type of capital raising impact company
profits, its balance sheet and share price?
Economic evaluation might be best lead by a person with
a technical-operating background whereas financing might
be best lead by an accountant. Both perform spreadsheet
modelling of the future business but they are very different
in purpose and process. Both should be presented in simple
language and easy-to-understand concepts.
The two activities should not be woven together because:
•• each is a stand-alone decision
66
•• combining their mathematics is very tricky, especially
adjusting the discount rate as debt is introduced (do
people still get fooled by false claims of improving
project returns by using debt?)
•• the spreadsheets will become unnecessarily complex
to use and audit and so alienate all in the team except
specialists in finance.
The financing spreadsheets could be appended to the end
of the economic evaluation workbook providing the flow is
one-way and nothing feeds back to the economic evaluation
worksheets.
Evaluation, valuation, modelling
‘Modelling’ is a component of, but not all of ‘valuation’ which
is a component of, but not all of ‘evaluation’. They form
a hierarchy with ‘modelling’ at the bottom providing the
hands-on computations that feed results for various cases and
scenarios up into ‘valuation’. This ‘valuation’ quantifies and
characterises the value of the project or metallurgical plant. In
turn this ‘valuation’ feeds up as one element in the intellectual
activity of ‘evaluation’ by fully understanding the project or
metallurgical plant as a business.
Anyone thinking that economic evaluation is all about
pouring numbers into a spreadsheet model to get NPV and
IRR is living in the past. Today it is all about having a working
knowledge of the whole project or metallurgical plant from
ore in the ground through all activities and influences to the
market. It encompasses nearly everything from engineering
to paying taxes. It is about understanding the key drivers
and key interactions of the business. It is getting a helicopter
view of the total entity, deciding how it fits the existing
business and helping to test ideas and create better projects
and metallurgical plants. It is all about putting the ‘E’ into
Evaluation. Yes, the economic evaluation specialist should
have the second-best job in the team: after the leader.
Best practice in economic evaluation
There are six principles at the top of the hierarchy.
KEY PRINCIPLES
The key principles to which one should adhere when
performing spreadsheet modelling are simple but extremely
effective. They are:
•• easy to follow
•• tailored-to-purpose
•• transparent
•• disciplined, rigorous and consistent
•• recording sources of all data
•• rapid to audit.
KEY PRACTICES
Flowing down from these six principles are key practices to be
incorporated in an economic model. They are:
•• A non-expert should readily understand the function
of each worksheet, how it is arranged into component
parts, how the data is entered, the computations and the
relative importance of the parts.
•• The visual flow down and across each worksheet should
be intuitive and logical.
•• Each worksheet should have a bold heading followed
by a brief outline of its purpose, and where helpful, its
important links with other worksheets.
We are metallurgists, not magicians
Guidelines for economic evaluation of projects
•• Sections within each worksheet should be in discrete
work blocks, with obvious subsections and subheadings
using a cascading layout for subtitles.
•• starts with a brief overview of results including multiple
graphs of the four cash flows (see below) and all
important inputs and outputs
•• Visually each work block should be self-contained with
an obvious step-by-step development toward a bold
subtotal for that work block.
•• is audited by the specialist, by experts in their areas
(example: metallurgist audited the processing section)
and usually an external specialist
•• The separate work blocks should be in a logical
sequence down and across the worksheet and their
aggregation is obvious.
•• is intuitive
•• Complex and extended computations should be shown
in a series of small steps so that the logic is visible and
the input parameters are obvious. There is no need to
interrogate the algorithm. Explain the logic of complex
algorithms in a note.
•• Usually if a row of data that has already been presented
above is needed again in a work block then the entire
row should be repeated so that there is visual flow of
the logic. If referenced from another worksheet then it
should be coloured coded.
•• Key inputs and results should be shown in graphs as a
self-check and for rapid understanding.
•• The ‘Data Group and Outline’ facility can be used to define
worksheet structure, collapse related groups of rows or
columns to reduce visual clutter, and aid navigation.
WORKED EXAMPLE OF MODELLING
Worked examples of economic models in Excel for
concept studies, prefeasibility studies and final feasibility
studies can be downloaded (free) from the website:
<www.economicevaluation.com.au>.
READY-MADE EVALUATION MODELS
There was unanimity amongst the AusIMM practitioners that
ready-made economic evaluation models, where users fill in
the blanks, were too dangerous to use. Experience is that these
black-box models transgress the six principles of best practice,
but more importantly have a bad history. Their computations
cannot be audited, the models must be exceedingly complex
(or deficient) to accept a wide variety of scenarios and they
simply cannot be trusted.
WHAT TO EXPECT FROM SPREADSHEET MODELLING
A project manager or metallurgical plant operator should
demand that every economic evaluation model under his/
her management:
•• is absolutely rigorous in its construction
•• is in discrete blocks of simple steps with clear headings
and obvious end results
•• has input data coloured (example: blue) so that it is
immediately visible and the project manager can skim
across it to quickly check validity of all inputs
•• has input data with a source (when, who and what)
typed in the row above so the project manager can
immediately see if it is the correct version
•• has data referenced across from another sheet coloured
(example: green) so it is immediately recognised
•• is absolutely consistent across rows, so that algorithms
do not have hidden changes
•• has every item of input data exposed in a row before
being used and absolutely no fresh data entered as data
hidden in algorithms
We are metallurgists, not magicians
•• provides trust in what you see without drilling down
into algorithms.
USING EVALUATIONS IN FEASIBILITY STUDIES
Project managers and metallurgical plant operators appear to
use three phases of study for projects:
1. Concept or scoping studies to assess if the project fits
company strategy and at least one alternative has a
reasonable likelihood of being economically viable:
•• For these brief studies the modelling and evaluation
usually can be relatively coarse and simple. It needs
to explore the range of outcomes and key drivers of
success and failure. It needs to weed out pet projects
and support only quality concepts.
•• The evaluation person needs to work cooperatively
between the project manager and the many experts
inside and outside the company who have the
knowledge. The specialist may be a ‘back room’ type
from any background, but better if an active specialist
with an operating/engineering background.
•• The project manager is likely to work closely with the
evaluation specialist interacting every day or two.
2. Prefeasibility studies that assess the complete range of
project alternatives. The divergent thinking in this phase
generates the most value:
•• For these studies the modelling and valuation needs to
be detailed enough to differentiate the economics and
character of each alternative. This phase excites those
who are creative but objective. It needs to confirm the
attractiveness of the selected alternative and define its
business character.
•• The evaluation person needs to frequently interact
with specialists inside the project team and outside,
drawing out the complete information, going back and
confirming it has been correctly modelled (audits) and
being absolutely objective. The evaluation specialist
must resist the temptation to be too clever with Excel
functions but keep the model simple so everyone can
readily understand, audit and feel it represents the
alternative truly. The specialist should not be a ‘back
room’ type but probably from an operating/technical/
engineering background with a broad understanding
of the business and a work ethic that is energetic,
collaborative, creative and accurate.
•• The project manager is likely to see the evaluation
model as an engine to assess alternatives and generate
ideas. He/she is likely to rely heavily on the evaluation
specialist to understand where each of the alternatives
is heading and to help steer the study activities to best
effect. Daily contact, with creative thinking, would be
common.
3. Final feasibility studies that thoroughly assess and define
the alternative to be taken into execution or construction.
This is convergent thinking with lots of detail:
•• For these studies the modelling and evaluation will
become quite detailed. The whole business needs to
67
P Card
be thoroughly defined. The modelling will evaluate
different methods and different equipment within the
selected alternative. It may need to work in nominal
terms in some areas and integrate with company’s
accounting and financing activities.
•• The evaluation person needs to work in detail as part
of the enlarged study team.
•• By this time the project manager should know if the
existing evaluation person is the right fit for this phase.
Contact may become periodic but the evaluation
model would continue as a vehicle to help steer the
Feasibility Study.
FOUR CASH FLOWS
Anyone opening a worst practice evaluation model is faced
with a myriad of parameters and a convoluted array of
computations that are understood only by the evaluation
specialist. Key outputs such as NPV are buried amongst the
worksheets. Auditing would be tedious and take days.
Best practice has simplified valuations into four streams of
cash flow:
1. revenue cash stream (production, stocks, sales, prices,
debtors)
2. capital cash stream (capital expenditure (CAPEX),
creditors, tax deductions)
3. operating cost cash stream (operating expenditure
(OPEX), creditors)
4. taxes cash stream (royalties and income tax).
Their sum represents the net cash flow each year and this can
be simply discounted to give NPV, or used to compute IRR.
Second and third level computations such as working
stocks, debtors, creditors and tax deductions for CAPEX (‘tax
depreciation’) are computed as high-level calculations within
these four cash streams. Their impact on project managers’
and metallurgical plant operators’ decision-making will be
minimal so should be reduced to a few simplified rows.
THE TWO BOOKENDS OF ECONOMIC VALUE
In most mining industry businesses there are two ‘bookends’
which dominate the economics: the ore resources in the
ground and the market. One end determines how big and
good the business can be and the other end determines how
profitable it will be. In between are all the very important and
exciting projects and operations to make it happen and make
it improve. It has been joked that a whole team of engineers
and metallurgists slave for ages on production, CAPEX and
OPEX while somewhere in a back office a few people generate
price forecasts that swamp the valuation. None-the-less
the production, CAPEX and OPEX are critical and must be
forecast with appropriate quality.
68
DO BIGGER PROJECTS NEED BIGGER ECONOMIC MODELS?
One of the world’s greatest iron ore mines was acquired in
the 1970s with the economic evaluation model provided to
the company’s Board being just one page of very easy-tofollow, manual computations. Contrast it with the volumes
of modelling and synthesis that would be required today. In
a strange way that one page from the 1970s was as potent as
all the evaluation study work that we generate today. There
in amazing simplicity for the Board members was the heart of
the acquisition decision, namely the risks in the forecasts of
price, mineral resource, production and costs over the years.
Senior executives in the world’s biggest mining companies
have made it clear they would love simple one page models of
major investments, major acquisitions and life-of-mine plans.
This is not to suggest that all evaluation models should
be one page. Quite the reverse: a detailed evaluation model
should be a centre-piece of a major project or metallurgical
plant. It should draw together all the component parts as a
business so the team understands the relative importance of
each part, where to focus and how to optimise the overall
design. It should be a tool used every day by members of the
project team/metallurgical plant to challenge and test their
ideas. Detailed evaluation models are needed to steer the
project toward the best configuration and to assess the risks.
Ironically, Management/Board need a simple, easy-tounderstand evaluation ‘one page’ model to help it understand
the big decision of whether to invest in the project whereas
the project team might need a big, detailed, working model
to optimise the project’s configuration and the design of its
component parts.
HALL MARKS OF BEST PRACTICE ECONOMIC EVALUATION
From day one, project managers and metallurgical plant
operators must set their expectations and demand that
economic evaluation is:
•• easy to follow – it may not be simple but anyone in the
team should be able to follow it like a school text book
•• fit-for-purpose in that detail matches importance – begin
as simply as possible and add complexity only when
warranted
•• rigorous, transparent and intuitive – trust what you see
•• fully documented – readily see where every piece of
data was sourced
•• graphs – to find errors and to give quick visual
understanding
•• audited – both the mechanical computations and the
results.
We are metallurgists, not magicians
Contents
Sensible cost cutting for resource projects
D Connelly1
ABSTRACT
As the resources industry becomes tougher with rising costs and cyclical metal prices,
mining companies need to continue to reduce their costs and do more with less.
Many new resource projects fail to come in under or on budget based on feasibility
studies. In particular, capital expenditure (CAPEX) proves difficult to achieve
for resource projects. Many resource projects have failed because of aggressive
plant CAPEX cost cutting, which results in projects with no surge capacity and
an inability to achieve design throughput. In addition, the plants are not operable
because of the omissions. In recent times, many businesses have also experienced
the ‘costs’ of simply implementing aggressive operating cost-cutting measures as a
strategy for solving business performance problems. This paper looks at technologydriven, employment-related luxury, department and restructuring cuts. In the past,
companies used an incremental approach based on the use of past budget information
as an integral part of the budget construction process. The use of performance reports
and management information systems (MIS) is examined, along with the role of
continuous improvement in achieving sensible cost cutting.
INTRODUCTION
All companies look for ways to reduce costs and increase profits, and resource
companies are no different. With the scale of mining projects, a cost saving of a
few per cent could mean millions of dollars for the company. Budgets are prepared to
determine where money is spent and where it could possibly be saved. This applies
to all resource companies, whether an exploration company looking to build a plant,
an engineering design company or a mining company in production.
Historically, the resources industry has always gone through peaks and troughs.
Understandably, mining companies want to take advantage of the good times by
maximising output of existing mines or constructing new ones. Designing a plant
to these favourable market conditions during a mining boom can be a dangerous
exercise. If the metal price suddenly drops, a low ore grade or high processing cost
prospect that was previously viable, could result in a mine closure. To avoid this,
companies must look at cost-cutting measures to increase the viability of their project
for a range of commodity prices.
A plant in the design stage has a much larger scope for cutting costs from the
CAPEX and operating expenditure (OPEX) than an existing operation through circuit
optimisation or process changes; however, overly aggressive process cuts to reduce
the CAPEX may cause the process to run inefficiently during operation and could
reduce the overall availability of the plant, causing extra costs and less revenue.
Process changes and their operability implications must be thoroughly considered
by the design company. Mining companies should not simply cut costs to the design
process as this may cause greater costs to the project in the future (for example, Murrin
Murrin, Browns project, Ravensthorpe nickel project, Port Hedland hot briquetted
iron project, Rapu Rapu copper zinc project).
1. MAusIMM(CP), Director/Principal Consulting
Engineer, Mineral Engineering Technical
Services Pty Ltd (METS), Midas Engineering
Group, Perth WA 6000. Email:
damian.connelly@metsengineering.com
An existing operation that is struggling to be profitable may need to make some
significant operating cuts. Rather than one large cut, which could have serious negative
implications to the personnel or process, several small cuts should be implemented
to achieve an overall significant saving. These cuts could be related to employment,
luxury, department or restructuring, and are applicable to both engineering design
companies and production companies. Production companies also have scope to
refine the operating costs through power conservation, process automation and
consumables optimisation. It is important that prior to implementation, all cuts be
assessed to ensure that the negative impacts are minimised. Continuous improvement
(CI) plays an important role in the cost-cutting procedure. Streamlining information
communication using MIS is just one aspect that could be included for CI.
69
D Connelly
BUDGETING AND BUDGET CUTS
Creating a budget should be the first stage of any project.
It is a process of predicting and controlling the expenditure
over the life of any given project. Budgets are the foundation
of an organisation’s financial success. The importance
of creating a budget is that it forces an organisation to
consider the expectation for its products and services with
the required resources to meet that expectation. In addition,
budgets can transform an organisation’s higher priorities
into the appropriate resources required to achieve those
aforementioned priorities. The potential problems could
be highlighted in a sufficient time to acquire the corrective
actions to be performed. A baseline can be created against
which the actual results can be compared.
Due to the variability of the resources industry and the
difficulty in predicting future costs, the budget is a document
which should be continually monitored. When a planned
budget is overshot, the consequences can vary from mere
frustration to anger or even litigation, if it involves a new
engineering design and construction project with other
companies. Overspending an allocated budget will result
in the reputation of the company being marred and good
business relationships being severed. (See Figure 1, which
highlights impact of changing budget).
Strategic budget cuts
Strategic budget cutting is a common organisational policy.
Applying the right combination of budget cutting and
strategic growth is a fundamental input for the long-term
success of any resource organisation. It is vital for managers
to approach budget cuts practically so as not to affect the
organisation’s capability.
There are many ways to perform a budget cut, yet extreme
cut-backs can eventually affect the growth of an organisation.
Many organisations today have come to understand how an
overly enthusiastic implementation of budget cuts can have
the unintended effect of inhibiting their revenues. Profits
were increased in the short-term due to the severe budget
cut, but the overall revenues were then decreased due to the
lack of growth. In some cases, budget cuts have caused the
subsequent declines in customer service and product quality.
There are several types of budget cuts and these include
luxury cuts, employment status related cuts, technology
driven cuts, department cuts and restructuring cuts.
Luxury cuts
An approach to reduce the budget is to focus on the costs of
supplies and services along with employee related expenses.
Travel and other related benefits spending can be minimised;
for instance, by employees only being offered economy class
when travelling instead of business. Additionally, many
organisations often spend a good deal of money on office
spaces or other items to impress their clients and competitors
(for example, the use of hire cars, video conferencing instead
of travel, mobile phone calls, alcohol, dining expenses).
Employment status related cuts
By employing consultants and independent contractors,
organisations will be able to minimise the overhead salaries
spent on permanent employees. Although independent
contractor arrangements can make a significant saving to an
organisation, there are some downsides that need to be taken
into consideration. An unintentional effect of converting
several former employees into independent status may create
unfortunate tax complications in an organisation. Some
organisations offer year-end bonuses instead of pay rises to
avoid a fixed commitment. There is a growing trend to utilise
temporary workers to support some parts of an organisation’s
operations such as during the shutdown period of an operating
plant. Certainly, temporary workers will never be as fully
invested in the company as would permanent employees and
this could result in less efficiency and productivity.
Technology driven cuts
With the advancement of technology, organisations have
started to find ways to reduce the current workload. The
application of technology will help to reduce the number
of employees needed and other associated costs to perform
the tasks. It may even allow employees time to be in part
redirected to improving the efficiency of other areas of the
operations. Although technology can be a good way to reduce
the operating budget, it is costly to establish and requires
time to implement new technologies. The initial investment
may discourage some companies from implementing
the technology. Technology is a capital investment and
consideration of the depreciation, ongoing maintenance
and replacement should be assessed adjacent to the realistic
savings. New technology introduces risks and is not applicable
for new projects unless piloted first.
Department cuts
Department cuts usually involve a participatory process where
managers of each operation unit will identify the prospects
to reduce their budgets without affecting the operation of
an organisation. Managers can start by checking the surplus
funding in the past that may have been given but is not
entirely essential to maintain the level of services. When there
is a vacant position available, managers will decide whether
it can be held without affecting the everyday operation of the
organisation. Managers should make sure that the budget cut
does not affect the organisation’s core strategies or the key
clients, customers and constituencies’ interest.
Restructuring cuts
FIG 1 – The cost of not getting the budget right.
70
When an organisation undergoes a major restructure, this
generally entails a shift from budgeting to strategic planning.
This change occurs as the demand from the market has
changed. Restructuring an organisation is not an easy process
and it involves the re-examination of the services it offers,
re-evaluating the departments and managers it should keep,
and the determination of the employees it needs depending
on their role in delivering new services and the application
of technology as an alternative to additional employees in
its operations. The result of the restructuring cut is usually
a significant cost reduction. The organisational changes in
the 1990s where whole layers of middle management were
we are metallurgists, not magicians
Sensible cost cutting for resource projects
retrenched including specialists with long-term business
knowledge is a specific example.
Continuous improvement
Continuous improvement can be described in several ways.
In the simplest terms, CI is the development of ongoing
improvement in quality and efficiency within a company. The
objectives of CI are to:
•• provide a more disciplined approach to CI projects and
initiatives
•• allow for greater consistency across global sites with a
standard methodology and tools
•• allow for a quicker response to CI opportunities
by facilitating the organisation of teams and the
development of both teams and individuals
•• improve safety, cost, production and productivity
performance
•• help all employees understand the importance of, and
the ways to, improve processes and organisational
relationships
•• help capitalise on global best-practices and share
knowledge across sites.
The steps of CI are:
1. identify the opportunity that will lead to continuous
improvement
2. study the opportunity with respect to the key business
needs for the organisation
3. define the current state by researching and understanding
the current process, system or organisation
4. develop the future state then develop and test an
implementation plan
5. implement the solution and maintain high levels of
communication to monitor the status of scheduled
activities
6. follow up and document the new procedures that have
been implements and assess the effectiveness.
Together these points seek to improve existing operating
procedures, quality and efficiency.
Management Information System
Management Information System is the practice of
managing data so that information can be delivered with
insight, understanding and value for the employees in an
organisation. It is the process of organising an information
database which is easily accessible, well defined and flexible.
Although techniques and technologies will change, these
principles will remain in the core information management
model. Through the MIS, information is primarily delivered
to the right person in the right structure at the right time with
a cost that adds net value to the organisation. For those that
apply their knowledge gained through experience it will
produce a result with more effective outcomes.
MIS is an important tool to streamline communication
throughout an organisation and serves to reduce time
generally used to search for information. This time saving
ultimately makes an organisation more efficient and hence
more profitable.
Design Stage Cost Cutting
Optimisation during the design stage of a project is an
important measure in reducing the CAPEX of a new project.
However, reducing the CAPEX of the project does not mean
we are metallurgists, not magicians
reducing the cost spent on design. Mining companies may
be reluctant to spend large amounts of money on the design
stage of a project because until production commences they
will see no revenue for all their expenditure. Cutting corners
at this early stage, although saving in the short-term could
prove costly in the future. A high quality plant design can
ultimately save on plant CAPEX and reduce unexpected costs
in the future.
Typically, in the early stages of design for a mineral
processing operation a number of options will be investigated
to determine the best possible design for the plant. These
options will be compared on the basis of process performance,
CAPEX and OPEX.
Unit design
It can be tempting for mining companies to only design a plant
to the nominal operating conditions in order to save on CAPEX.
Ultimately this could be detrimental to plant performance and
result in less availability or throughput. Increased flow rates
due to surges need to be accounted for around the plant and it
is important that the equipment is able to handle this. Pumps
are a typical example of this and the design should incorporate
a 10–15 per cent allowance for surges.
Another trap is designing equipment to handle the average
rather than the maximum ore characteristics. Ore hardness
is one characteristic where the maximum value should be
taken when designing equipment. Although sizing to these
specifications may result in larger, more expensive pieces
of equipment being required, designing the plant to handle
average ore hardness can result in lower crusher throughputs,
higher recirculating loads and less overall plant utilisation.
Similarly, individual equipment availability needs to be
considered. In a processing system with little surge capacity,
the overall availability is equal to the lowest availability,
not the average. If a critical piece of equipment is offline the
entire process stops. A high maintenance piece of equipment
like belt filters may have an availability of 65–70 per cent
due to cloth replacement and cleaning. If an overall plant
availability of 90 per cent is desired, standby filters would be
needed to achieve this. Removing these standby units to save
on CAPEX will ultimately result in lower plant throughputs
and smaller revenue.
Alternatively if there is equipment which is known to have
a low availability, but the capital required makes the use of
multiple units unviable, instead of expecting unreasonable
availabilities these units could be decoupled from the process.
That is, placing adequate surge capacity before or after the
units to allow for the increased downtime. These units will
require a higher throughput than the rest of the processing
plant to maintain the overall plant throughput with less uptime. A common example of this in a minerals processing
operation will be a crushing circuit which has been decoupled
from the processing plant through the use of a crushed ore
stockpile. A crushing circuit will generally have an availability
of around 80 per cent.
Plant layout
Plant layout can affect the CAPEX for a project. Optimising
the plant location and plant layout can produce significant
saving when considering access and plant roads, material
handling distances, electrical and communications
infrastructure, and plumbing. The design must be optimised
with these factors in mind without having the equipment
so close that access for maintenance is compromised.
Computer-aided design packages (for example, SolidWorks
by Dassault Systèmes) can be used to virtually construct
71
D Connelly
the plant and be used to produce quantities. Optimising the
design in the virtual world can enable direct savings in the
real world. A more compact plant but with room for later
expansion flows through to savings on civil, piping, electrical
and construction costs in both the short and long-terms.
Contingency
Mineral processing projects in development will always
have a contingency added to the price. Although this may
seem like an ambiguous addition to the price it is in fact an
important item in the project cost. This cost represents the
project unknowns which experience tells us will eventuate as
the project progresses. Contingency is directly related to the
risk of the project whether that is political, environmental or
otherwise. The more risk associated with a project the higher
the contingency will need to be. For example, if a project is
to be constructed in an area with political instability or a
tropical area that is prone to cyclones, an allowance will need
to be made in the event that the project becomes delayed or
requires extra capital to complete. As the project develops and
gets closer to production, the chances of one of these events
occurring suddenly decreases, as does the contingency.
The contingency for a project in the prefeasibility stage may
be between 20–30 per cent of the direct costs. However, as the
project gets to the detailed design stage, it would be expected
that this number will drop to 10–20 per cent. This contingency
must be determined on a case-by-case basis and these
numbers can vary. Estimating contingencies can be a black
art. To reduce the CAPEX of the project it may be tempting
for some companies to reduce or remove the contingency.
Although there is a possibility that the contingency will
not be needed, it is there for a reason and cannot simply be
dismissed. Reducing or removing this could prove costly if
the contingency is required (see Figure 2).
Engineering Shortcuts
Engineers have a duty to provide their services in a manner
consistent with the standard of care of their profession.
A good working definition of the standard of care of a
professional is that level or quality of service ordinarily
provided by other normally competent practitioners of
good standing in that field, contemporaneously providing
similar services in the same locality and under the same
circumstances. An engineer’s service need not be perfect. As
the engineer is using judgement gained from experience and
learning when providing professional services, and is usually
doing so in situations where a certain amount of unknown
or uncontrollable factors are common, some level of error is
allowed.
When you hire an engineer you purchase service, not
insurance, so you are not justified in expecting perfection
or infallibility, only reasonable care and competence. An
engineer who makes a mistake causing injury or damage is
not sufficient reason to lead to professional liability on the
part of the engineer. For there to be professional liability,
it must be proven that the services were professionally
negligent, that is, they fell beneath the standard of care of the
profession. When hiring an engineer, there is an expectation
of risk acceptance that this professional engineer potentially
may make a mistake whilst using reasonable diligence and
best judgement.
The standard of care is not what an engineer should have
done in a particular instance; it is not what others believe
an engineer should do, or how others say they would have
done. It is what competent engineers have actually done in
similar circumstances.
Operability
Operability is the ability to keep equipment, a system
or a whole industrial installation in a safe and reliable
functioning condition, according to predefined operational
requirements such as:
•• consideration of the operator
•• distributed control
•• slopes on sump floors – too flat is a problem
•• access to power, air and water points
•• walkways and access
•• flat launders, which lead to sanding
•• skirting – spillage
•• dust control
•• clean up in crushing plants
•• surge between unit operations
•• low head height
•• clear signage, pipes labelled
•• attention to particular areas such as reagent mixing and
lime mixing
•• floor space around mills
•• absence of or insufficient bunding height
•• insufficient process water storage
•• ball charging
•• telemetry etc.
Maintainability
In engineering, maintainability is the ease with which a
product can be maintained to:
•• correct defects
•• meet new requirements
•• make future maintenance or expansion easier
•• cope with a change to maintain the plant
•• access to pumps for repairs or change out
•• gain access for mobile equipment and personnel for
maintenance, clean up etc
•• prevent lubrication of equipment under spillage or
placement well outside the area of operation
•• provide spare cyclones
FIG 2 – Value-adding through project life.
72
•• have common equipment
we are metallurgists, not magicians
Sensible cost cutting for resource projects
•• install cranes in grinding area, flotation area
•• install bypass facilities (thickeners, trash screens)
•• separate the acid wash column or hopper
•• avoid the obvious (for example, vent exhaust next to a
high voltage switchyard etc).
CONSTRUCTION COST CUTTING
hand sales, sourcing of appropriate equipment is generally
done during the definitive feasibility study (DFS), the last of
the project before construction.
There are four factors that should be considered when
determining the viability of using second-hand equipment,
and these are:
1. equipment costs
2. the condition of equipment
Labour costs
3. suitability for process requirements
Labour represents a significant percentage of the construction
costs. When looking to reducing these costs the ‘per hour’
labour rate should not be the determining factor; rather the
efficiency of the workforce. It may be tempting for some
companies to hire cheaper, unskilled or foreign workers for
plant construction to reduce the labour rate. However, if one
of these workers is half the price but takes three times as long
to do the work, it is not a cost-effective decision.
4. spares.
Examples exist of projects where cheap, foreign electricians
have been used for African projects and poor supervision
resulted in extended commissioning problems, the need for
rewiring and damage to equipment.
Plant modularisation
Plant accommodation and site buildings have used modular
design for some time now to save on construction and
installation costs. The same benefits are now being seen
with plant and equipment modularisation. Companies
that produce modular equipment have a standard design
and modify it as per the process requirements, reducing
engineering costs. The equipment is commissioned off-site
before being dismantled and packed into shipping containers.
Once at site the equipment can simply be ‘bolted together’
and it is ready for use. Using equipment modularisation can
significantly reduce construction times and hence capital cost.
Innovative construction techniques
With construction costs representing a significant portion of
the CAPEX, there are companies that are marketing innovative
techniques and materials to reduce the cost of the project.
There are several products and construction methods that
look to reduce the amount of concrete used in construction.
One method uses earth retaining walls constructed from steel
mesh. Another is the use of steel arches overlaid with earth
to form bridges, underpasses and stockpile tunnels. Both
save time and money which is associated with the concrete
structures that would otherwise be required.
Reducing construction capital through unit hire
It is possible to reduce the CAPEX of a processing operation
through the hire of equipment and infrastructure. However,
the reduced initial cost comes at the price of an increased
OPEX. This could be an option investigated by junior miners
looking to enter production but are finding it difficult to raise
the initial capital. The mining fleet, mobile plant and the
accommodation camp are typical examples of units that can
be hired for this purpose.
Second-hand plants
The option of using second-hand equipment in the construction
of a processing plant can mean considerable savings to the
overall CAPEX of the project. However, this needs to be
considered on a case-by-case basis as it is possible for the costs
to outweigh the savings. The use of second-hand equipment
can be considered at any stage of the project. However, due to
the project timeline compared to the urgency of most secondwe are metallurgists, not magicians
Equipment costs
The equipment cost does not only include the purchase of
the equipment but also the dismantling, re-conditioning,
transportation and re-assembling at site. This should be
considered carefully if the equipment is located in a remote
mine site or is difficult to dismantle and transport, it may be
cheaper to purchase new equipment.
The condition of equipment
Although second-hand equipment may be cheaper than new,
its condition needs to be considered. If substantial refurbishing
is going to be required, the initial saving on CAPEX will be
outweighed by the cost required for the plant to become
operational. In addition, used equipment may require extra
maintenance during operation and this possibility must be
taken into account.
Suitability for process requirements
Second-hand equipment is unlikely to meet the exact
specifications as detailed by the proposed equipment list. Some
concessions may need to be made and it must be determined
whether the equipment can handle the process requirements.
The concept is fine; however there are numerous examples
where the execution has been poorly handled resulting in
the savings being less than originally considered. Other
issues include a lack of drawings, lack of vendor support or
unknown historical problems with the equipment.
Spare parts
This is necessary to avoid the cost of having to specifically
manufacture one-off parts.
OPERATIONAL COST CUTTING
It can be difficult to achieve any one cost cut on an existing
operation that is going to make a great deal of impact on the
overall operating cost. Also, a large cut will inevitably impact
negatively on the process or personnel of the organisation.
Therefore, the most sensible way of achieving any significant
saving is to implement several small cuts. With regards
to resources, the sector can be broken up into two distinct
industries; these are mineral consulting companies and the
production companies. Consulting companies would use the
strategic budget cuts as detailed previously to achieve costs
cuts in the organisation. Production companies can also use
these strategic budget cuts; however they also have scope to
reduce the operating costs of the processing plant.
Reagents and consumables
Optimising the operating costs with regards to plant reagents
and consumables needs to begin at the first contract tendering
phase. The tenders need to be reviewed to obtain the best
product supplier and this may not always be the cheapest
option. The lead time for the products needs to be considered
73
D Connelly
and also the supplier’s production rate, particularly if they
are supplying to an ever increasing market. It is necessary
that supply can be maintained to the plant, especially if that
product is critical to the process.
This has been the case with cyanide when a global shortage
impacted on the smaller supply companies. The operations
holding contracts with these smaller companies were paying
enormous margins at the time because of the difficulty to get
the product. The larger companies, on the other hand, were
able to maintain supply to their clients but would not provide
product to the companies without existing contracts in place.
It is also important to watch the usage of reagents and
consumables, particularly with the more expensive ones such
as cyanide in a gold plant. High usages should be investigated
and lowered if possible. In some cases it may be appropriate
to use automatic control to moderate reagent and consumable
use. It should be investigated whether the capital outlay will
be repaid through cost savings.
operation; however the capital to implement these changes
could be significant. These improvements could be related
to automation of sampling, online analysis or automatic
reagent addition. These methods would be categorised under
‘technology driven cuts’ as mentioned previously. Although
the improved efficiency could reduce the operating costs this
needs to be weighed against the, often substantial, capital that
must be outlaid for their installation. Flexibility to cope with
changing ore types and flow sheets is a must.
CONCLUSIONS
Power saving
Sensible cost cutting is an important measure in ensuring
the viability of an organisation through a range of market
conditions. This is particularly important in the resources
industry where the cyclical metal prices mean that boom and
busts are inevitable. Cutting costs to remain profitable during
the busy times is often necessary to avoid business or mine
closures. Careful budgeting, implementing an MIS and CI
are just a few of the measures that should be implemented to
keep costs down at all times. If budget cuts are required, large
cuts should be avoided due to the detrimental effect upon
personnel or the process. Several small and well considered
cuts should be implemented, as discussed in this paper. These
small savings can cumulatively produce an overall significant
saving to the CAPEX and/or OPEX costs.
Using power efficiently and saving energy where possible
can result in decent operating cost savings, especially if this
power supply is diesel generated. In some cases an external
contractor can be brought in to determine the savings
achievable through power saving. However, it is necessary
to assess whether the cost of engaging an expert is worth the
savings.
The author would like to thank Robert Hanna from METS,
various companies, colleagues, engineers at various sites,
METS staff and other consultants for their contribution. Also
the management of METS for their permission to publish this
paper and the constructive criticism of various drafts.
There are numerous examples of reducing cyanide or
flocculant consumption where considerable savings have
been made because technical people were prepared to
challenge what had been used historically and undertake test
work to support the case for reducing reagent usage.
One way of saving power is to ensure that the operating
curves of process pumps are within the most efficient range.
Depending on the size of the pumps in the system this could
equate to a significant cost saving. There are numerous
examples of cost savings for semi-autogenous grinding (SAG)
mills and crushing circuits by using circuit surveys and
simulation studies.
External contractors
It is important that costs for external contractors are monitored.
If contractors are not managed closely, they may charge more
than budgeted for the job. One provision would be to never
hire a contractor for a large job on an hourly rate. Over the life
of the project, the cost blowout could be substantial.
Improved control
Improving the control systems of an existing processing
plant does have the ability to increase the efficiency of the
74
ACKNOWLEDGEMENTS
REFERENCES
Australian and New Zealand College of Anaesthetists (ANZCA),
n/d. How to carry out a continuous improvement project,
Guidelines on continuous quality improvement [online].
Available
from:
<http://fpm.anzca.edu.au/resources/
educational-documents/guidelines-on-continuous-qualityimprovement> [Accessed: 2 March, 2011].
Mackenzie, W and Cusworth, N, 2007. The use and abuse of feasibility
studies, in Proceedings Project Evaluation Conference, pp 65–76 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Maddox, D, 1999. Strategic budget cutting [online], The
Grantsmanship Center. Available from: <https://www.tgci.
com/sites/default/files/pdf/Strategic%20Budget%20Cutting_1.
pdf> [Accessed: 17 July 2017] (John Wiley & Sons).
Petty, J, n/d. Budgeting and one day reporting: developing and
managing a budget, towards one day monthly management
reporting, Course notes.
we are metallurgists, not magicians
Contents
When does further processing at
the mine site make sense?
C Fountain1, S La Brooy2 and G Lane3
ABSTRACT
A century ago, new mines were often accompanied by smelters. In Australia, smelters
were built at the Daydream mine near Broken Hill; at Kuridala and Mount Elliott near
Mount Isa; Broken Hill; Mount Lyell; Mount Morgan; Mount Isa; and more recently at
Tennant Creek; Kalgoorlie; and Olympic Dam. Now the big miners eschew site-based
processing in favour of shipping concentrate to smelters in Europe and Asia.
Yet there are times when on-site processing clearly makes sense. Few companies
ship gold ore or concentrate to China for gold extraction. New smelters are being
built in Zambia. Solvent extraction and electrowinning plants frequently produce
copper from oxide ores and concentrates while pressure leaching is increasingly
being considered for copper sulfide ores. Some argue that on-site processing is best
because the waste products can be returned to the ground from whence they came,
avoiding potentially large disposal costs in more-populated areas. Treatment and
refining charges could once again turn in favour of the smelter operators.
This paper examines when it makes sense to use on-site pyro- or hydrometallurgical
processes in today’s environment and comments on how current trends might alter
the balance in future.
INTRODUCTION
In a keynote address to the Ecological Society of Australia’s annual conference in
November 2007, Paul Ehrlich said:
… Australia is still, in the 40-some years I’ve been coming here, striving to become
a third-world country, to be a place that just exports its natural capital as fast as
possible, unworked upon, until it becomes truly poverty stricken. (Ehrlich, 2007)
In the same year, Chip Goodyear was in the process of vacating his leadership role
at BHP Billiton Limited (BHP). He was quoted as saying that BHP should concentrate
on mining and ore, and leave ‘to others the skill set of processing that material’
(Roberts, 2007).
It was not always the case.
When the Broken Hill Proprietary Company Limited was floated in June 1885 to
develop the deposit that gave BHP its name, a smelter was not far behind, opening
in May 1886 (Blainey, 1968a). Other companies built their own, so there were at least
five smelters operating along Broken Hill’s line of lode by 1891. BHP’s first smelter
was preceded in the district by a smelter at the Daydream mine near Silverton, which
treated the small ore arisings from the district.
Other early smelters included copper smelters at Burra in South Australia; Mount
Morgan, Mount Elliott, and Mount Cuthbert in Queensland; and Mount Lyell in
Tasmania. Lead and copper smelters were later developed at Mount Isa; a nickel
smelter at Kalgoorlie; and copper smelters built at Tennant Creek and Olympic Dam.
A short-lived nickel-copper smelter was built at the Radio Hill mine near Karratha in
Western Australia.
1. MAusIMM, Operational Readiness Manager,
Nyrstar Port Pirie Smelter, Port Pirie SA 5540.
Email: chrisjan@iinet.net.au
2. FAusIMM, Principal Process Consultant,
Ausenco, Perth WA 6000.
Email: stephen.labrooy@ausenco.com
3. FAusIMM, Chief Technical Officer, Ausenco
Minerals & Metals, South Brisbane Qld 4101.
Email: greg.lane@ausenco.com
Similar stories can be told in North America and other parts of the world. In 1978, a
survey of converter practices elicited responses from 16 copper smelters in the United
States (Johnson, Themelis and Eltringham, 1979), most of them near the deposits that
provided their feed. Today only three remain.
Most mining companies today follow Goodyear’s injunction and sell concentrate
to others. However, there are some who are still building processing facilities at or
near mine sites: new smelters have recently been constructed in Zambia; pressure
leaching, and roast, leaching and electrowinning plants are being built at some sites;
and hundreds of millions of dollars have been spent on nickel-laterite leaching plants.
75
C Fountain, S La Brooy and G Lane
This paper considers the drivers for site-based processing
and outlines guidelines for project developers to assist in their
choice between on- and off-site processing options.
The developers of new mining projects need to decide
how much processing there will be at the mine site. The
options range from direct shipment of ore with virtually no
processing (for example, hematite lumps and nickel-laterite
ore) through to the production of a finished commercial
product (for example, magnesium alloy wheels at Solikamsk
in Russia). Table 1 lists examples of different degrees of mine
site processing.
Hardly any mine sites take processing further than refined
metal to fabrication of finished products. The operation at
Solikamsk in Russia, where magnesium alloy automotive
wheels were produced for export at a magnesium chloride
mine, is an exception to the rule.
In most cases, the selection of the degree of further
processing will be made on economic grounds. In some cases,
it might be dictated by political considerations; the former
Soviet Union is not the only place where economics might be
ignored for ‘social’ reasons. However, it is the economic case
that is the focus of this paper.
DISCUSSION
The economics of on-site processing changed dramatically
during the twentieth century. Some of the changes favour
reduced on-site processing, while others have worked to
increase it.
Factors limiting on-site processing
The factors limiting on-site processing include:
•• technological change
Table 1
Examples of materials shipped from mine sites.
Type of
material
Examples of
materials
Examples of
locations
1
Ore
Iron ore
Nickel laterite
Bauxite
Pilbara,
New Caledonia, Indonesia,
Philippines,
Weipa
2
Concentrates
(flotation, gravity,
dense medium
separation,
magnetic,
electrostatic etc)
Cu
Ni
Zn
Sn
Mineral sands
Cadia,
Cosmos,
Century,
Renison (when operating),
Bemax,
Iluka Eucla Basin
3
Intermediate
products
Alumina
Nickel sulfide matte
Nickel hydroxide
Nickel oxide
Wagerup, Pinjarra,
Kalgoorlie nickel smelter,
Ravensthorpe,
Goro
4
Crude metal
Lead
Gold bullion
Blister copper
Anode copper
Mount Isa,
Kalgoorlie Cons gold mine,
Mount Isa (before 1979),
Mount Isa
5
Refined metal
Ni
Cu
Murrin Murrin,
Olympic Dam
6
Manufactured
product
Mg alloy wheels
Solikamsk,
Russia
76
•• lower transport costs
•• falling product prices
•• higher labour costs
FURTHER PROCESSING
Category
•• higher capital costs
•• higher employee turnover
•• lack of infrastructure
•• changing orebodies
•• spare capacity in existing processing facilities.
Technological change
Concentration techniques significantly improved over the
past 120 years. Remember that the froth flotation process was
developed to treat sulfide ore and tailings in Broken Hill.
Before then, practically all the copper mined in the United
States came from underground mines chasing high-grade
veins, averaging 2.5 per cent copper (Wills and Atkinson,
1991). The high cost of transport could cripple lower-grade
deposits. Smelting was often a means of concentrating the
valuable metals to reduce transport costs.
The past century has seen the invention and improvement
of froth flotation, improved grinding technologies allowing
better liberation of minerals (the development of the McArthur
River deposit depended on the development of ultra-fine
grinding techniques such as the IsaMillTM), the development
of solvent extraction – electrowinning (SX–EW), carbon-inpulp, pressure oxidation, and heap leaching technologies.
Improved concentration techniques have lessened the need for
local smelters. The fine grain of the Mount Isa lead–zinc deposit
made it impossible to produce high-grade lead concentrate,
so a lead smelter was built at the mine. The McArthur River
mine, developed following the invention of the IsaMillTM finegrinding technology, ships concentrate to smelters.
On the other hand, SX–EW has allowed the economic
production of copper metal at site from low-grade oxide
deposits.
Higher capital costs
Smelters were once cheap to build. The Daydream smelter
was little more than a blast furnace and a chimney. Smallscale smelters built to last only a short time could be viable.
However, they had their costs: Blainey (1986b) wrote of the
problems of lead poisoning in Broken Hill – once strong men
throwing lead fits, the difficulty of raising kittens or puppies
… or even children. Such conditions are unacceptable today
in nearly every corner of the world.
When the Mount Isa lead smelter was commissioned in
1931, it was fitted with a baghouse to limit lead emissions and
so the citizens of Mount Isa were spared the lead problems
encountered by the early residents of Broken Hill.
Emission controls increase the capital costs of smelting
and have contributed to a trend to build larger smelters.
High production rates reduce the relative cost of emission
controls by spreading their unproductive imposts over larger
quantities of product. Ramachandran et al (2003) showed
that the size distribution of smelters has increased in most
countries over the years.
These higher capital costs have made it impractical to build
smelters for small mines with short lives and encouraged the
sharing of capital expenditure through toll smelting by large,
often significantly depreciated smelters. It would be difficult
for an MIM to justify a copper smelter with the initial capacity
of 15 000 t/a of the Mount Isa smelter (Pritchard, 1980). AGIP
Australia attempted such a feat with its Radio Hill smelter near
we are metallurgists, not magicians
When does further processing at the mine site make sense?
Karratha in 1990, with the IsasmeltTM plant there designed to
produce 1.5 t/h of nickel–copper matte (Bartsch et al, 1990),
but this plant was quickly closed by low nickel prices (Player,
Fountain and Tuppurainen, 1992).
Lower transport costs
Lower effective energy costs and larger, more efficient trains,
trucks and ships have reduced the cost of transport. They
make possible the shipment of lower-grade (and hence value)
products from mines, particularly those near coasts and ports.
Thus, it is possible for nickel mines in New Caledonia and the
Philippines to ship ore containing about 1.6 per cent nickel
to Townsville for processing (Fittock, 2006). Lower transport
costs have reduced the benefit of mine site processing.
Falling product prices
Until recently, there has been long-term downward pressure
on metal prices (Fountain, 2002). In the mid-1990s, WMC
Limited allowed an average 1 per cent real fall in prices each
year (Morgan, 1995).
This downward pressure, combined with rising costs,
compressed margins, made it harder to justify new capital
and increased the need to use existing facilities that had
already depreciated capital. It encouraged brownfield rather
than greenfield smelter capacity expansion.
Higher labour costs
Metal prices have fallen relative to labour costs. In 1900, it took
an American on the average 84 minutes to earn the equivalent,
before tax, of the price of a pound of copper; in 2002, it was just
over three minutes (Fountain, 2002). The resulting increased
standard of living in the developed world has reduced the
incentive for people to work in remote, undesirable regions.
The consequent workforce shortages have been compounded
by the actions struggling mining companies took to survive:
reducing their support of universities and apprentice-training
schemes, and reducing their hiring and training of graduates
and other young people.
To attract and retain a skilled workforce in residential mine
site processing plants a premium must be paid to them. This
leads to a further incentive to ship the material to a location
where such a workforce is more prepared to live or is cheaper
to employ.
Higher employee turnover
An increase in career opportunities has resulted in higher
employee turnover. The workforce was very stable at the
Broken Hill mines throughout most of the twentieth century.
The mines paid well and the Barrier Industrial Council’s
policies favoured employing local men and single women
over married women and those people ‘from away’.
and action at the expense of understanding and planning.
Processing plants with a stable workforce will be at an
advantage over those with high-turnover and chaos.
Lack of infrastructure
Mine sites in remote locations usually lack the infrastructure
that is commonly available in industrial areas of the world.
The need to build power stations or long power lines, water
storage and treatment facilities, and perhaps oxygen and other
industrial gas production facilities, raises the capital investment
required for further processing. Many of these facilities are
available almost on tap in developed industrial areas.
Government assistance to single companies is often
constrained by international trade agreements, but
governments are less constrained when it comes to providing
infrastructure that will service multiple customers. It is
difficult for a state or national government to provide a
connection for a remote project to an electricity grid, but it can
do so more easily if it is setting up an industrial estate that will
attract many companies.
Changing orebodies
Mining companies normally mine the easy deposits and the
easy parts of deposits first. The shallow Broken Hill ores
mined in the 1880s were oxidised and readily smelted in cheap
blast furnaces. As the surface mineralisation was depleted,
the miners reached the deeper sulfide deposits from which
the zinc minerals had not been leached. The lead–zinc ores
could not be smelted economically with existing technology
and the ‘sulfide problem’ provided the incentive to develop
the flotation process.
Many of the orebodies discovered today are complex.
Complex orebodies can require complex treatment processes
and complexity increases capital cost.
The McArthur River deposit did not become an orebody
until the IsaMillTM fine-grinding technology was developed
(Fountain, 2002). Even with grinding to an 80 per cent
passing size of 7 µm, McArthur River produced a bulk leadzinc concentrate that needed to be treated by the Imperial
Smelting Process (Nihill, Stewart and Bowen, 1998). MIM
Holdings (now Glencore) bought the existing Avonmouth
and Duisburg lead-zinc smelters to treat McArthur River
bulk concentrate rather than building a new smelter for the
task. It subsequently sold the Duisburg smelter and closed
the Avonmouth smelter, and sold the bulk concentrate to
smelters operated outside the original MIM Holdings group
of companies.
Spare capacity in existing facilities
The situation is further complicated by a management
philosophy that assumes that plant managers do not need
to understand how the plant works – they just need to be
able to manage the people who do. Such managers, who
rely on others for the technical knowledge, often cannot tell
when they are getting poor advice. To make matters worse,
managers often change jobs frequently.
There are many established processing facilities around
the world. Construction of a new processing facility must
be weighed against using an existing facility with largely
depreciated capital. Some of these facilities were constructed
to process ores or concentrates from local sources that have
been depleted or are too high cost to compete with cheaper
imported feedstocks. Others sought to reduce their unit
costs by importing concentrate to supplement local sources.
Examples included the Queensland Nickel nickel production
plant at Yabulu; north of Townsville in Queensland; and
the New Boliden smelters at Skelleftehamn in Sweden and
Harjavalta in Finland.
High employee turnover makes it difficult to optimise
complicated processes and creates a culture of fire-fighting
The existing facilities most likely to be able to accept
imported materials are those near a port.
These days, however, mining companies that do recruit
skilled people are struggling to retain them. It takes time for
people to learn how to run a complicated process well. Many
employees leave before providing employers with a return on
the time and money invested in their training and development.
we are metallurgists, not magicians
77
C Fountain, S La Brooy and G Lane
Other considerations
Smelting sulfide ores and concentrates these days usually
results in sulfuric acid by-product. Smelters with local markets
for sulfuric acid, such as an established chemical industry or
an acid-consuming leaching process, have an advantage over
those that must transport acid over long distances, particularly
over land. Local acid demand helped the profitability of
Japanese smelters in the 1980s and the production of acid for
heap leaching helped the Miami smelter survive when many
other American copper smelters were closing.
Smelters and other complex processing technologies can
have long ramp-up times, disadvantaging them against
existing plants. Ramp-up times can be decreased by using
well-established technologies with good training programs
(McNulty, 1998; Arthur and Hunt, 2005).
Factors promoting further on-site processing
Despite these disadvantages, there are circumstances that
promote on-site processing. These include:
•• difficulty producing a saleable product
•• problems concentrating valuable minerals
•• high transport costs
•• additional transport problems
•• government subsidies
•• high treatment and refining charges
•• new technologies
•• simple, low-cost process flow sheets
•• environmental considerations
•• government decree.
Product saleability
It is sometimes not possible to produce a salable product
from a mineral deposit without further processing. This is
particularly the case at times of surplus supply when smelters
and refiners can be selective in their feedstock purchasing and
can afford to eschew concentrates containing minor element
assemblages that are costly to treat or handle. A zinc smelter
might reject a high-manganese zinc concentrate when times
are good but purchase and blend it with other low-manganese
concentrates when zinc concentrate is scarce.
Western Australian mineral sands miners adopted the
Becher synthetic rutile process to convert almost unsalable
grades of ilmenite into a premium synthetic rutile product
(George, 1993).
St. Joe Minerals Corporation installed a multiple-hearth
roaster at its El Indio mine in the Chilean Andes to remove
arsenic from its copper concentrate to make it marketable
(Smith et al, 1985).
Ability to produce a concentrate
Some ores are difficult to concentrate with economic
recoveries. Fine gold does not respond well to conventional
concentration techniques. Consequently, many gold
operations use some form of leaching followed by fire-refining
to produce gold bullion. Gold can be produced economically
from gold grades as low as 0.5 parts per million (Wills and
Atkinson, 1991). This would not be viable without on-site
processing. Even at $1500 per ounce, it would be uneconomic
to transport ore containing such a low concentration of gold
any significant distance.
The Teck CESL hydrometallurgy process might be suitable
for treating low-grade bulk copper-nickel concentrates (Teck78
Cominco, 2008a) but Teck-Cominco (now Teck Resources)
could not justify its use at its Highland Valley mine, which
produces readily marketable, high-grade concentrate (TeckCominco, 2007). Vale built a 10 000 t/a CESL demonstration
plant to process copper-gold concentrate at its Sossego mine in
Brazil’s Carajás region (Teck-Cominco, 2008b) with a view to
applying it to the high-fluoride Salobo concentrate (Defreyne
and Cabral, 2009). Vale’s website makes no mention of using
the process in its current Salobo flow sheet.
High transport costs
There are still parts of the world where transport costs are
high. Zambia, for example, is a landlocked African country
with transport routes through other African countries. Mopani
Copper Mines, Konkola Copper Mines, First Quantum Minerals,
and a joint venture between China Nonferrous Metal Mining
(Group) Company Limited and Yunnan Copper Industry
(Group) Company Limited have either refurbished existing
smelters or built new ones this century. The Mopani smelter at
Mufulira (Ross and de Vries, 2005), the First Quantum Minerals
smelter (Glencore Technology, 2016) and the Chinese smelter
at Chambishi use Glencore’s Copper IsasmeltTM technology.
The Konkola smelter at Nchanga is using an Outokumpu flash
furnace (Konkola Copper Mines, 2006).
Another example might be the Kennecott smelter near the
Bingham Canyon mine in Utah. Rio Tinto spent US$1.1 billion
on the new smelter and refinery that was commissioned near
Salt Lake City in 1996 (West, 1999).
Improvements in recovery
Additional on-site processing might improve recoveries of
valuable minerals. Concentrator operators are sometimes
forced to sacrifice recovery to achieve a concentrate grade
that is economic to transport. Additional on-site processing
steps could allow higher recoveries by providing the ability
to treat lower-grade concentrates. The Mount Isa lead smelter,
for example, was designed for lead concentrate grades lower
than typically transported due to the fine nature of the Mount
Isa ore and the difficulties experienced producing high-grade
concentrate. As a result, the blast furnace is fluxed using
higher levels of lime than commonly used in lead smelting
(McLoughlin, Riley and McKean, 1980).
Additional transport problems
Materials with high uranium levels can be difficult to permit
for transport. WMC (now BHP) reduced the transport issues
for its copper product by building a processing plant at
Olympic Dam to remove the uranium from the concentrate
and to produce cathode copper at the site.
Government subsidies
Governments sometimes offer subsidies to persuade
companies to maximise local processing. These might include
assistance with infrastructure development, royalty discounts,
tax-free holidays, or outright cash grants. The Queensland
government, for example, offered a royalty discount for
downstream processing of nickel, cobalt, copper lead and
zinc ores in the past (Queensland Government, 2003). The
government of Trinidad and Tobago has offered tax-holidays
to companies building smelters and petrochemical plants
using local natural gas.
High treatment and refining charges
Treatment and refining charges vary, depending on the
balance between concentrate supply and smelter demand. The
construction and expansion of copper smelters in China has
we are metallurgists, not magicians
When does further processing at the mine site make sense?
placed downward pressure on these charges, to the benefit of
concentrate sellers. In the June 2005 – June 2006 period, BHP
was paying US$115 per t for treatment of its copper concentrate
and US$0.115 per pound for refining. They were able to reduce
these charges to US$60 and US$0.06 for the period June
2006 – June 2007 (Carlisle, 2006). Spot sales have achieved
even lower prices and The Sydney Morning Herald reported a
settlement at US$45 and US$0.045 between ‘an Indian smelter
and a major Western miner’ (Reuters, 2007). Treatment charges
subsequently rose progressively to be over US$100 and refining
charges over US$0.10 by 2015 (Outotec, 2015).
Low treatment and refining charges make justification of
further on-site processing difficult, but high treatment charges
tilt the balance in favour of building additional capacity. It is
easier to justify building a smelter at a mine site when the
industry as a whole needs additional capacity, although
greenfield capacity still needs to compete with cheaper
brownfield expansions.
New technologies
New technologies can favour additional mine site processing.
The development of SX–EW technology led to increased
production of cathode copper at mine sites, particularly from
low-grade oxidised copper orebodies. The development of
IsasmeltTM technology has lowered the cost of copper smelting.
Carbon-in-pulp, BIOX® and other new technologies have
resulted in more gold mines being developed, mines at which
on-site processing is a necessity. Relatively new technologies
such as BioheapTM and variations on pressure oxidation are
emerging as potential competitors to the established secondary
processing options, principally smelting, for copper sulfides.
One of the driving forces for the recent implementation
of pressure oxidation technology for copper concentrate
treatment at Freeport McMoran’s Cyprus Bagdad operation
was the production of acid on-site for use in its oxide heap
leach operation. Similarly, the Sepon operation in Laos uses
the pressure oxidation of pyrite to provide ferric ion and acid
to the copper leaching process.
Simple, low-cost process flow sheets
Given the difficulty in attracting and retaining skilled
operators and professionals to mine sites, it is an advantage
if the process is simple and robust. Simplicity tends to mean
low-capital flow sheets. Low-capital reduces the economic
risks of on-site processing.
Environmental considerations
Processing concentrate to extract valuable metals produces
waste products, whether the process is smelting or
hydrometallurgical. Waste disposal can be difficult in heavily
populated areas. A mine site is often the best place to dispose
of those wastes. The slags from the Mount Isa smelters can be
used as underground fill (their pozzolanic properties make
them a substitute for cement).
Slags and other metallurgical wastes can also be deposited in
mine site tailings storage facilities. Titanium dioxide pigment
producers in developed countries prefer to use the chloride
process with high-grade feed (rutile, synthetic rutile or highgrade titania slag) because the waste disposal costs of the older
sulfate route are significantly higher. This has encouraged
the production of synthetic rutile in Western Australia and
upgraded slag at Rio Tinto Iron and Titanium Inc in Quebec.
Transporting ores or concentrates uses more energy than
transporting final product. If human emissions of carbon
dioxide are causing the world to warm, it would often be more
sensible to produce metals at the mine site than transporting
we are metallurgists, not magicians
gangue minerals half-way around the world, perhaps to a
less-efficient plant. However, imposing strict carbon emission
caps on resource industries developed-world nations is likely
to have the perverse effect of driving metals processing
activities to countries without caps, potentially resulting in an
increase in total emissions.
Government decree
In some instances, governments require further processing to
be undertaken as a condition of mining rights. Development
of Voisey’s Bay nickel deposit was delayed over the insistence
by the provincial government of Newfoundland and Labrador
that Inco build a processing plant in the province.
In 2008, the Zambian government introduced a 15 per cent
levy on the export of copper concentrates to encourage further
processing in Zambia (Montia, 2008).
In such circumstances, the developer must determine
whether the deposit can withstand the additional capital
impost, and either comply or walk away.
Comments on specific metals
Aluminium
Aluminium is produced from bauxite ores. Some mines have
associated refineries that produce alumina as an intermediate
product (for example, Alcan’s Gove operations), while
others ship bauxite to refineries (for example, Comalco’s
Weipa operations). Aluminium smelting consumes large
amounts of electrical power (Australian smelters consumed
about 15 MWh for every tonne of aluminium produced in
2006 (Australian Aluminium Council, 2006)). Consequently,
smelters tend to be built where electricity prices were cheap at
the time of construction – near hydroelectricity in Washington
state, Quebec, Tasmania and Norway, and increasingly
through natural gas production in Gulf states.
Copper
In the past few decades, most owners of new sulfide copper
mines have elected to install concentrators at mine sites and
ship concentrates to smelters located elsewhere. As discussed
earlier, smelters have become larger and more capitalintensive with time, making them less likely to be built
on-site.
China’s late start to heavy industrialisation, its extreme
demand for materials, and a focus on economic growth ahead
of environmental concerns has allowed smaller Chinese
smelters to remain economically viable for longer. However,
economic, governmental and environmental pressures are
driving the move to larger production plants with associated
oxygen and acid plants. Sulfur capture in Chinese smelters
was around 85 per cent in 2006, the same as the Chilean rate,
but below the world average of 90 per cent, and 95–97 per cent
level in developed countries (Diaz and Mackey, 2007).
Except for Zambia and possibly China, most recent increases
in capacity have come through upgrades of existing smelters,
including the expansion of Southern Copper’s Ilo smelter
and Vedanta’s Tuticorin smelter using new, large-capacity
IsasmeltTM furnaces.
Hydrometallurgical processing options for copper minerals
increased in importance following the development of solvent
extraction, which largely eliminated the cementing process
with its transport of scrap iron. Processing is still limited by the
acid consumption of the ore and availability of acid, but new
technological developments, such as pressure leaching, might
lead to additional on-site processing. One way of coping with
79
C Fountain, S La Brooy and G Lane
acid demand is integration of different processes so that acid
generated in one is available for leaching in another, reducing
net acid requirements and avoiding transport issues. BHP’s
Olympic Dam complex uses smelter acid to leach uranium.
Another approach, recently implemented, is to use pressure
oxidation of sulfide concentrates to generate acid for leaching,
thus avoiding the cost of an acid plant; examples include Sepon
in Laos (Sherrit, Pavlides and Weekes, 2005), Kansanshi in the
Congo, and Phelps Dodges’ Baghdad and Morenci mines in
Arizona (Dreisinger, 2006).
Current hydrometallurgical technologies offer a range
of processes – biological, fine grinding, ferric leaching and
pressure oxidation (Dreisinger, 2006) – that between them
cater for different feeds (ore or concentrate) and different
levels of acid demand from other on-site processes. No one
process suits all applications. For sites with existing SX–EW
facilities that are running out of oxide ore, hydrometallurgical
processing offers a way of continuing on-site processing using
existing equipment.
Gold
Gold metal is conventionally produced at the mine site;
payable ore grades are at the g/t level and the capital cost
of cyanide leaching and associated gold recovery is low
unless the ore is refractory. Even when refractory ores
require pretreatment to make them amenable to conventional
cyanidation, on-site processing is still better than transporting
ores long distances. This is normally the case even when the
contained value in the concentrate exceeds that of base metal
concentrates. An exception occurs when such an ore can be
concentrated and transported to an existing facility close
enough that the capital saved offsets the transport costs.
Oceana Gold’s Reefton (New Zealand) mine is an example:
Reefton produced a pyritic flotation concentrate containing
highly-refractory, sulfide-encapsulated gold at about 30 g/t;
the concentrate was transported 600 km by rail to Oceana’s
plant at Macraes Flat for treatment by pressure oxidation in
an existing autoclave. The Reefton operation was placed on
care and maintenance at the end of 2015 (OceanaGold, 2016).
Payment for precious metal values in concentrates can be a
factor. Smelters normally pay for gold and silver recovered
during smelting or refining. However, treatment charges
and the effective costs of late payment can be avoided if the
gold can be recovered by gravity treatment (Gray, Katsikaros
and Fallon, 1999) and as practiced at Newcrest’s Cadia and
Telfer operations.
There is no payment of gold value in antimony concentrates,
making alternative on-site processing more favourable. New
England Antimony Mines Limited used a thiourea leach to
extract gold from its antimony concentrate (Hisshion and
Waller, 1984).
Iron
Extensive economies of scale enable long-distance rail and
maritime hematite ore transport. Between 1997 and 2000
Australian free-on-board (FOB) costs fell from US$9.71–
7.08/t of iron ore (AME, 2002). With the failure of BHP’s hotbriquetted iron process, there is currently little appetite for
further processing of hematite ore near the mine site. On the
other hand, upgrading magnetite ore to produce concentrate
pellets can reduce transport costs sufficiently to justify the
investment (for example, the Savage River and Port Latta
complex in Tasmania).
80
Lead
Most lead concentrates are exported to distant smelters.
However, mine site smelters might be warranted if the
concentrate grade is low, incurring high transport costs, or if
the ore is composed of oxidised lead minerals. In the latter
case, lead metal poses less risk to communities than moving
more-soluble forms of lead, such as lead carbonate.
Oxidised lead minerals can be smelted with a relatively
low capital cost using technologies such as rotary furnaces,
which are commonly used for recycling lead sulfate battery
materials (Rao, 2006). The product is crude lead bullion that
can be exported for refining.
Nickel
Limonite nickel-laterite (low-grade iron ore with around
1–1.5 per cent Ni) suitable for pressure acid-leaching (PAL) or
Caron processing have typically traded at 10 per cent of the
nickel value in the ore, while saprolite (magnesium silicate
ores with around 2–2.5 per cent Ni) suitable for smelting to
ferronickel or matte have typically traded at 25 per cent of the
value of the nickel content (Australian Mining Journal, 2001).
With development of PAL, the economics of laterite shipping
are becoming marginal.
Guidelines for decision-making
The authors offer the following guidelines to help organisations
with their decisions about the extent of on-site processing.
Is there an existing facility that will buy an unprocessed product?
This is the first question that should always be asked. Using
existing facilities limits the capital expenditure for a new
mining project. However, treatment charges and transport
costs might make new processing facilities attractive relative
to existing ones.
Shortages in existing processing capacity will result in
higher treatment charges and a better case for building new
capacity. However, new capacity needs to come in at the
lower end of the operating cost-curve so that the operator is
not left with a high-cost asset when the shortage abates and
treatment charges fall.
Special issues, such as the presence of deleterious minor
elements (for example, fluorine, arsenic or radioactive
elements), might make a product difficult to market in an
untreated form.
Undertake a value-chain analysis
It is important that mining companies understand where the
value is added to their products. This means examining the
potential net profit at each step of the processing chain and the
total return on capital invested. It would include considering
the cost of transport and relative energy costs at the mine site
and at potential alternative sites.
Such an analysis will usually identify a logical point to hand
processing to another facility.
Is the project capital-constrained?
If a project is capital-constrained, it is often better to sell
lower-grade products than spend additional capital to
upgrade them. The situation can change over time as a
positive cash flow develops and provides financial resources
for further investment.
we are metallurgists, not magicians
When does further processing at the mine site make sense?
Are there government incentives or mandates for further processing?
Determine whether any government incentives or mandates tilt
the balance in favour of further processing. Are the incentives
sufficient to provide a return on investment that justifies the
additional processing? If there are further processing mandates,
is the project robust enough to withstand any opportunity costs
created by this mandate? If not, the opportunity to lose money
is probably better left to someone else.
Care needs to be taken to ensure that government incentives
are incentives. Tax-holidays to encourage companies to a
region might not be as valuable as first thought once the tax
effects of plant depreciation in the early years of operation
are considered.
Consider the valuable metal content of the product
Table 2 shows typical metal contents in various materials.
Most materials transported have a valuable metal content
greater than 25 per cent. Materials containing less valuable
metal than this would normally require further processing.
Mount Isa over a period of six months and, as a result, set a
new benchmark for smelter commissioning (Arthur and Hunt,
2005). A new processing plant is more likely to be successful if
the plant’s owner can find a similar plant on which to train its
workforce before commissioning.
Consider the complexity of the process
Simple processes are more likely to be successful at the mine
site. The greater the complexity, the more difficult it will be to
find a workforce capable of sustaining the process and plant
performance used in the initial economic justification.
It is difficult to achieve high plant availabilities with highlycomplex processing systems with minimal surge capacity,
and long ramp-up times reduce or destroy a project’s net
present value.
Avoid the economies-of-scale trap
Long inland transport routes are more likely to favour onsite processing. Economies of scale and dedicated railways
might mitigate this effect, as in the case of the Pilbara railways
owned and operated by Rio Tinto and BHP.
When a project is uneconomic at the initial scale, the
temptation is to increase output to reduce unit costs through
economies of scale. This can produce significant benefits, but
increasing size to achieve a target rate of return carries risks.
These include an increased financial loss if the project fails
to achieve its cost, quality and throughput targets, and the
possibility that increasing the scale of production will alter the
supply–demand balance, depressing the price of the product.
This latter effect increases the risk to the project.
Are there environmental reasons for further processing at site?
What is the projected mine life?
Transport distances
As discussed, there might be environmental reasons to
upgrade a product at site. If there is an environmental driver,
site-based processing might make the difference between
receiving regulatory approval to proceed and a refusal.
As with government mandates for further processing for
social reasons, the project proponent will need to determine
whether the project is sufficiently robust to proceed under
these circumstances.
The mine life needs to be sufficient to allow a reasonable
return on the investment in additional processing. The larger
the capital expenditure and the more complex the process, the
greater the mine life needs to be.
If the project is near a port, it is possible that the processing
plant could operate well beyond the life of the orebody by
treating imported ore or concentrate.
Resist justifying on-site processing by selling potential by-products
Consider the ability to attract a workforce
If it is relatively easy to attract a high-quality workforce,
additional processing options are more likely to be successful.
It is better to have an experienced core group to build the rest
of the workforce around than to try to train an entire staff with
no experience. Inexperienced employees will make mistakes,
extend ramp-up times, and perhaps develop a culture that
prevents the project ever realising its potential.
Opportunities for training
Good training, particularly hands-on training, is important
when commissioning new processes. When the Yunnan
Copper Corporation chose to build an IsasmeltTM plant at its
smelter in Kunming, it sent metallurgists and operators to
The authors are aware of projects where the principals
were keen to undertake further processing. The proposed
processing plants were not economic in their own right, so
their ‘project champions’ sought to bolster the economics
by selling by-products. In one case, this led to a scenario
of a billion dollar (1990s dollars) industrial complex in the
Australian outback to develop an 8 Mt tonne mineral deposit.
In another case, there was a short period of wild enthusiasm
when it was thought that by-products previously regarded as
wastes (with associated disposal costs) had such a huge return
that they converted a marginal project into a potential bonanza
… until it was realised that overly optimistic assumptions had
been made about the value of the by-products and the capital
and operating costs of producing them.
Table 2
Relative metal contents.
Class of material
Ore
Iron
Aluminium
Ore (55% Fe)
Bauxite (26–30% Al)
Concentrate
Magnetite pellets
Intermediate
DRI or HBI (93% Fe),
HIron™ (96% Fe)
Crude metal
Pig iron (93–95% Fe)
Refined metal
Steel
Copper
Nickel
Lead
Limonite (1.5% Ni), Saprolite (2.5% Ni)
Concentrate (25–37% Cu)
Concentrate (12–27% Ni)
Alumina (52% Al)
Matte (40–80% Cu)
Precipitated mixed sulfides (23% Ni), Oxide
or hydroxide (40% Ni), Matte (45–72% Ni)
Refined Al (>99.7% Al)
Refined Cu (>99.85% Cu)
Blister Cu (95–99.5% Cu)
Sulfide (50–70% Pb),
Carbonate (60–70% Pb)
Lead bullion (97% Pb)
Refined Ni
Refined Pb (>99.9% Pb)
Note: DRI – direct-reduced iron; HBI – hot-briquetted iron.
we are metallurgists, not magicians
81
C Fountain, S La Brooy and G Lane
Miners can feel offended by lack of payment of ‘full value’
for minor elements (for example, gold, silver, indium,
and cobalt). However, extracting these elements requires
additional processing steps that add to complexity (Fountain,
2013). Careful consideration of the capital and operating costs
of these additional steps is needed to justify their construction.
One of the key actions that a company developing a
new deposit need to undertake is a value-chain analysis
to understand the logical end point of processing. The end
point will depend on the type of mineral mined and the
location of the mine. It might be affected by government
incentives or subsidies.
By-products might provide a supplementary income stream,
but projects that depend on this stream are often not robust.
Exceptions include the use of sulfuric acid for leaching.
On-site processing facilities will be easier to justify in times
when demand for processing exceeds global capacity (as
occurred during the decades of rapid growth after World War
II) than times with surplus processing capacity.
Consider the quality of the existing infrastructure at the mine site
Additional mine site processing will increase the demand for
water and electricity in particular. The availability of plentiful
water and cheap electricity can make the difference between
shipping a relatively low-grade material and upgrading
it at the mine site. The need to install additional electricity
generating capacity or power lines can often kill upgrading
projects in remote areas.
Will further processing improve metal recoveries?
If further processing results in higher recoveries, are they
sufficient to justify the additional capital and operating costs?
What effect will complexity have on development schedule?
Increasing the complexity of the project will increase the
complexity of the permitting and ultimate financing of the
project, and increase the development duration. Does this fit
with the owner’s capability and expectation?
When should novel processes be considered?
Novel processes are sometimes considered as a saviour for
processing difficulties with complex orebodies. When used in
greenfield locations, the application of novel technologies can
suffer from over enthusiastic expectations of performance,
operating cost and capital cost. A recent example was the
development of the PAL technology in Western Australia,
which resulted in underachievement of expectations for onsite nickel production from laterite ores on three simultaneous
projects. Extensive pilot and demonstration plants are
required to establish the technical and cost drivers for novel
processes and applications, particularly where the project is
the only or major source of cash flow for the company.
CONCLUSIONS
While many lament the export of relatively unprocessed
products from mines, the decline in mine site processing in
recent years has been driven by a variety of economic and
social factors. These include increasing capital costs associated
with emissions controls, the existence of large facilities that
toll-smelt or buy concentrates or ores, the rise of Chinese
smelters over the past decade, poor mine site infrastructure,
and an increasing reluctance by skilled people to work in
remote areas.
There are still circumstances that favour mine site
processing. These include:
•• difficulty of producing a marketable product without it
•• problems concentrating valuable minerals (such as fine,
disseminated gold)
•• high transport costs
•• simple, low-cost process flow sheets
•• environmental considerations
•• government sanction and approval.
82
Advocates of additional mine site processing need to keep
a clear focus on the economics of the process. Increasing
economies of scale or earning additional income through selling
by-products might not be as attractive as at first glance. Lowcost and simple flow sheets are easier to operate and expose the
company to less risk than complex, high-cost flow sheets.
REFERENCES
AME Mineral Economics, 2002. Australia boasts lowest iron ore
production costs in 2001 [online], Reed Business Information.
Available from: <http://www.ferret.com.au/> [Accessed: 11
March 2008].
Arthur, P S and Hunt, S P, 2005. IsasmeltTM – 25 years of continuous
evolution, in Sustainable Developments in Metals Processing
(eds: M Nilmani and W J Rankin), pp 73–94 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Australian Aluminium Council, 2006. Sustainability report 2006
[online]. Available from: <http://www.aluminium.org.au/>
[Accessed: 28 March 2008].
Australian Mining Journal, 2001. Trade in nickel-cobalt laterite ores
[online]. Available from: <http://www.theajmonline.com/>
[Accessed: 9 March 2008].
Bartsch, P, Fountain, C and Anselmi, B, 1990. Radio Hill project, in
Pyrosem WA (eds: E J Grimsey and N D Stockton), pp 232–250
(Murdoch University Press: Perth).
Blainey, G, 1968a. The Rise of Broken Hill, pp 20, 24 and 63 (Macmillan
of Australia: Melbourne).
Blainey, G, 1968b. The Rise of Broken Hill, pp 86–95 (Macmillan of
Australia: Melbourne).
Carlisle, C, 2006. BHP Billiton secures big drop in Japan copper
smelter charges, Mineweb, 5 October 2006.
Defreyne, J and Cabral, T, 2009. Early copper production results from
Vale’s hydrometallurgical CESL refinery [online]. Available from:
<http://www.teck.com/media/CESL-Publication-Copper-ValeMetallurigcal-Results-ALTA-2009.pdf> [Accessed: 18 July 2017].
Diaz, C and Mackey P, 2007. Discussion of copper smelting
technology practice and innovation [online], in Proceedings
The Carlos Diaz Symposium on Pyrometallurgy – Cu 2007,
vol 3. Available from: <http://cim.v1.myvirtualpaper.com/
report/2007061901/?page=10> [Accessed: 9 March 2008].
Dreisinger, D, 2006. Copper leaching from primary sulfides: Options
for biological and chemical extraction of copper, Hydrometallurgy,
83:10–20.
Ehrlich, P R, 2007. Keynote address to the Ecological Society
of Australia’s annual conference, Perth, 25–30 November
2007 [online]. Available from: <http://www.abc.net.au/
rn/inconversation/stories/2007/2109181.htm#transcript>
[Accessed: 24 February 2008].
Fittock, J, 2006. Nickel and cobalt refining by QNI Pty Ltd, Yabulu,
Qld, in AusIMM Monograph 19 Volume. Available from: <http://
www.ausimm.com.au/Content/docs/mono19model_paper.
pdf> [Accessed: 24 February 2008].
Fountain, C, 2002. IsasmeltTM and IsaMillsTM—models of successful
R&D, in Proceedings AusIMM 2002 Young Leaders’ Conference (The
Australasian Institute of Mining and Metallurgy: Melbourne).
we are metallurgists, not magicians
When does further processing at the mine site make sense?
Fountain, C, 2013. The whys and wherefores of penalty elements
in copper concentrates, in Proceedings MetPlant 2013 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
George, I W, 1993. Synthetic rutile operations of RGC Mineral Sands
Limited at Capel and Narngulu, WA, in Australasian Mining and
Metallurgy – The Sir Maurice Mawby Memorial Volume, Second
Edition (eds: J T Woodcock and J K Hamilton), pp 1301–1304 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Glencore Technology, 2016. Installations – IsasmeltTM [online].
Available from: <http://www.isasmelt.com/EN/Installations/
Pages/Installations.aspx> [Accessed: 20 November 2016].
Gray, S, Katsikaros, N and Fallon, P, 1999. Gold recovery from copper
gold gravity concentrates using the Inline Leach reactor and weak
base resin [online], in Proceedings Oretest Copper Gold Symposium.
Available from: <http://www.gekkos.com/documents/17-gold
recoveryfromcoppergoldgravityconcentratesusingtheinlineleach
reactorandweakbaseresin.pdf> [Accessed: 9 March 2008].
Hisshion, R G and Waller, C G, 1984. Recovering gold with thiourea,
Mining Magazine, September: 237–243.
Johnson, R E, Themelis, N J and Eltringham, G A, 1979. A survey
of worldwide copper converter practices, in Copper and Nickel
Converters (ed: R E Johnson), pp 1–32 (The Metallurgical Society
of the American Institute of Mining, Metallurgical, and Petroleum
Engineers: New York).
Konkola Copper Mines Plc, 2006. Expansion projects [online].
Available from: <http://www.kcm.co.zm/expansion_projects.
php> [Accessed: 28 March 2008].
McLoughlin, B C, Riley, J F and McKean, G R, 1980. Development
of optimum fluxing procedures for Isamine lead concentrates, in
Lead-Zinc-Tin’80 (eds: J M Cigan, T S Mackey and T J O’Keefe)
pp 309–320 (The Minerals, Metals & Materials Society:
Warrendale, Pennsylvania).
McNulty, T, 1998. Developing innovative technology, Mining
Engineering, 50(10):50–55.
Montia, G, 2008. Zambia introduces 15% copper concentrates levy
[online]. Available from: <http://www.metalmarkets.org.
uk/2008/03/10/zambia-introduces-15-copper-concentrateslevy/> [Accessed: 29 March 2008].
Morgan, H, 1995. Interview on The Bottom Line, Australian
Broadcasting Corporation Television, 18 February 1995.
Nihill, D, Stewart, C M and Bowen, P, 1998. The McArthur River
mine – the first years of operation, in Proceedings 1998 Annual
Conference, pp 73–82 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
OceanaGold, 2016. OceanaGold operates mines in the three largest
goldfields in New Zealand [online]. Available from: <https://
www.oceanagold.com/our-business/new-zealand/> [Accessed:
20 November 2016].
Outotec, 2015. Copper markets and TC/RCs [online]. Available from:
<http://www.outotec.com/en/Search-material/Customernewsletter-Smelting-News/Issue-1-February-2015/Coppermarkets-and-TCRCs/> [Accessed: 20 November 2016].
Player, R L, Fountain, C R and Tuppurainen, J M I T, 1992. IsasmeltTM
developments, in Mervyn Willis Symposium and Smelting & Refining
Course, July, pp 20:1–20:13 (The University of Melbourne).
Pritchard, J, 1980. Copper smelting at Mount Isa Mines Ltd, Mount
Isa, Qld, in Mining and Metallurgy Practices in Australasia
(ed: J T Woodcock), pp 340–344 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Queensland Government, 2003. Explanatory notes for SL 2003
No. 174, Mineral Resources Regulation 2003 [online]. Available
from:
<http://www.legislation.qld.gov.au/legisltn/sls/ris_
en/2003/03sl174e.pdf> [Accessed: 9 March 2008].
Ramachandran, V R, Diaz, C, Eltringham, T, Jiang, C Y, Lehner,
T, Mackey, P J, Newman, C J and Tarasov, A, 2003. Primary
copper production – a survey of world copper smelters [online],
in: Volume IV: Proceedings Pyrometallurgy of Copper, Herman
Schwarze Symposium, Book 1, pp 3–106 (Canadian Institute:
Montreal Canada). Available from: <http://www.hatch.
ca/non_ferrous/articles/primary_copper_production.pdf>
[Accessed: 9 March 2008].
Smith, E H, Foster, J W, Minet, P H and Cauwe, P H, 1985.
Selective roasting to de-arsenify enargite/pyrite concentrate
from St. Joe’s El Indio Mine – from pilot plant to commercial
operation, in Complex Sulfides: Processing of Ores, Concentrates
and By-products (eds: A D Zunkel, R S Boorman, A E Morris and
R J Wesely), pp 421–440 (The Metallurgical Society: Warrendale,
Pennsylvania).
Rao, S R, 2006. Resource Recovery and Recycling from Metallurgical
Wastes, pp 199–200 (Elsevier: Oxford).
Reuters, 2007. Japanese smelters, BHP discuss copper fees [online],
The Sydney Morning Herald, 6 December. Available from:
<http://www.smh.com.au/news/Business/Japan-smeltersBHP-discuss-copper-fees/2007/12/06/1196812912011.html>
[Accessed: 28 March 2008].
Roberts, J, 2007. BHP denies delays at Olympic mine [online],
The Australian, 8 September. Available from: <http://www.
theaustralian.news.com.au/story/0,25197,22382128–5005200,00.
html> [Accessed: 24 February 2008].
Ross, J and de Vries, D, 2005. Mufulira smelter upgrade project
– ‘Industry’ smelting on the Zambian copper belt [online].
Available
from:
<http://www.isasmelt.com/downloads/
XTpaper_Mopani_MSUPyromet05.pdf> [Accessed: 9 March
2008].
Sherrit, R, Pavlides, A G and Weekes, B L, 2005. Design and
commissioning of the Sepon copper pressure oxidation circuit,
in Proceedings First Extractive Metallurgy Operators Conference,
pp 21–27 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Teck-Cominco, 2007. Annual Report 2006 [online]. Available from:
<http://www.teckcominco.com/DocumentViewer.aspx?eleme
ntId=115704&portalName=tc>, p 21 [Accessed: 31 March 2008].
Teck-Cominco, 2008a. Annual Report 2007 [online]. Available from:
<http://www.teckcominco.com/DocumentViewer.aspx?eleme
ntId=124227&portalName=tc>, p 26 [Accessed: 31 March 2008].
Teck-Cominco, 2008b. Applied technology [online]. Available from:
<http://www.teckcominco.com/Generic.aspx?PAGE=Technolo
gy+%26+Innovation+Pages%2FResearch+Centre+Pages%2FCES
L&portalName=tc> [Accessed: 31 March 2008].
West, G, 1999. Breakdown leaves giant copper furnace out of action,
The Australian Financial Review, 12 January, p 22.
Wills, B A and Atkinson, K, 1991. The development of minerals
engineering in the 20th Century, Minerals Engineering, 4:643–652.
we are metallurgists, not magicians
83
Contents
The ABC of Mine-to-Mill and metal price cycles
P Cameron1, D Drinkwater2 and J D Pease3
ABSTRACT
In the 1990s metal prices were trending in a long-term decline and the usual cost-cutting
exercises were adopted. Cuts to research and development made innovation difficult
and the outlook was grim. But necessity is a strong motivator, and these conditions
spurred the Mine-to-Mill movement, which optimised across organisational silos and
utilised new technology tools, innovative software and increased computing power.
Many applications of Mine-to-Mill exploited the fact that comminution is usually
the site processing bottleneck, and that blasting is more efficient at breaking rock than
grinding. This approach sought to:
•• understand and characterise rock breakage from mining to the mill
•• develop models and simulators for blast design, fragmentation, crusher and
mill circuits
•• develop tools to measure in real-time the particle size distribution of rocks on
conveyors and run-of-mine (ROM) muck piles
•• ensure effective communication across the silos between geologists, blast design
engineers, mining engineers and metallurgists.
These methods and the technology tools available in the 1990s were widely adopted
at numerous sites around the world with significant benefits. Semi-autogenous
grinding (SAG) mill throughput increases of 10–20 per cent were common. This
was the advance the industry desperately needed. It was low capital cost; it was
obvious. It was here to stay.
Except, in too many cases, it didn’t. In the minerals price boom of the 2000s Mineto-Mill was no longer necessary to ‘survive’, even though it was still good business.
Some operations, including some of the early success stories, slipped back to old
habits of optimising within organisation silos instead of across them. Fortunately,
other operators still embraced Mine-to-Mill and developed new techniques. Mine-toMill was still successful, but not as widely adopted as expected.
Now the boom has ended and operations are again under cost pressure. Since
these are the same circumstances that created Mine-to-Mill, it seems time to achieve
the wider adoption.
Since the 1990s many new or advanced technology tools are available for Mine-toMill projects: blasthole sensors, new explosives formulations, new blasting techniques
and modelling, ore tracking devices, improved image analysis to determine size,
texture and colour of coarse ore, grade sensors, whole-stream simulation tools, even
more powerful computing hardware and data analysis software.
If we could achieve so much in the 1990s, how much more can we achieve today
when we have the same imperative, the same potential and a larger number of high
technology tools?
INTRODUCTION
The Mine-to-Mill methodology was developed during the commodity price downturn
of the 1990s and was widely adopted at numerous sites around the world. Mine-toMill relies on the fact that comminution is usually the site processing bottleneck, and
that blasting is more efficient at breaking rock than grinding.
1. MAusIMM, General Manager Australia,
Split Engineering, Mission Beach Qld 4852.
Email: pcameron@spliteng.com
2. MAusIMM, Principal Consulting Engineer,
Mineralis Consultants Pty Ltd, Brisbane Qld
4066. Email: ddrinkwater@mineralis.com.au
3. FAusIMM, Senior Principal Consulting
Engineer, Mineralis Consultants Pty Ltd,
Brisbane Qld 4066.
Email: jpease@mineralis.com.au
This approach and the technology tools available in the 1990s were widely adopted
around the world with significant benefits. Semi-autogenous grinding mill throughput
increases of 10–20 per cent were common. We thought this approach to managing
operations was here to stay, since it appeared to be free money for operators; however
during the subsequent boom and focus on ‘production at any cost’, while some
operators advanced the approach, many reverted to old habits of optimising within
organisation silos. Mine-to-Mill wasn’t crucial to their survival.
Now that the commodity cycle is back in downturn, the simple productivity gains
offered by Mine-to-Mill look attractive again. Since the 1990s many new or advanced
85
P Cameron, D Drinkwater and J Pease
technology tools have been developed to enhance the approach.
Now is the time to adopt the new technologies, and to embed
them into a more universal and lasting industry change.
TABLE 1
Mine-to-Mill case studies 1996 to 2002.
Mine site
Metals produced
Country
Product increase
PAST PRACTICE AND ESSENTIAL REQUIREMENTS
Highland Valley
Cu
Canada
10%
In its broadest sense, Mine-to-Mill integrates all aspects of
geometallurgy and production steps with processing and
marketing. In this paper, the authors confine their comments
to the subset of Mine-to-Mill that focuses on the integration
of blasting with comminution and separation. They refer to
this as ‘Advanced Blasting for Comminution’ – the ABC of
Mine-to-Mill.
Alumbrera
Cu Au
Argentina
13%
Porgera
Au
Papua New Guinea
15%
KCGM Fimiston
Au
Australia
18%
Cadia
Au Cu
Australia
14%
Red Dog
Zn
USA
12%
An excellent summary of the steps involved in any Mine-toMill project groups them into three main areas (McKee, 2013):
BHP Iron Ore
Fe
Australia
3% increase in lump
1. First and foremost, good data. Data needs to be collected
about the ore before it is mined and as it passes through
the production chain, about the equipment and processes
used in production, and the process performance
and cost.
2. Good, robust analytical tools and models to evaluate
options and identify optimum operating points for a
range of feed types and conditions. The best models also
account for economic factors such as metal price and
operating cost.
3. Finally, any Mine-to-Mill project requires tools for
ongoing monitoring, assessment and evaluation.
Findings need to be validated and optimisation kept ontrack to fully realise the benefits.
Further, these non-technical factors are critical:
•• sustained management support
•• availability of staff with specialist skills
•• an enabling organisation structure.
Early Mine-to-Mill projects included scheduling to smooth
ore variations, building stockpiles to ‘campaign’ different
ore types, or redesigning underground activities to eliminate
‘cash negative’ ore while rescheduling surface operations to
eliminate the supposedly ‘fixed’ costs associated with them.
These examples show that Mine-to-Mill projects did
whatever was required to improve overall mine site
performance. They were orebody and situation specific. Many
case studies reported significant gains, typically in the range
10–20 per cent productivity improvement across the mine
site, with little or no capital expenditure. These were dramatic
improvements by any measure compared with working in
isolated silos (McKee, 2013).
MINE-TO-MILL IN 1996
By 1990 higher speed computational power enabled
innovative software for mathematical modelling and
simulation of industrial processes including mining and
mineral processing. The AMIRA Project ‘Optimisation of
Fragmentation for Downstream Processing’ (1996 to 2002;
Table 1) was a collaborative research project to exploit this
capability and develop operating strategies to enhance mining
and downstream processing activities. The history, concepts
and case study projects are presented in McKee (2013).
In 1996 the technology tools available to Mine-to-Mill were
limited to elementary versions of tools such as (or similar to):
•
JKSimBlast blast design software
•
Kuz–Ram fragmentation model to predict particle size
distribution of the blast in the stockpile
•
Split-Online, or other image analysis of the particle
size distribution at the truck tip to the primary crusher,
86
KCGM – Kalgoorlie Consolidated gold mine.
and on the conveyors from the primary crusher to the
SAG mill feed
•
JKSimMet, simulation software for comminution circuits.
There was minimal information about mineralogy, and
work-arounds had to be devised to deal with complex,
multicomponent ore types and non-standard mining and
processing scenarios. Early Mine-to-Mill projects relied
heavily on the knowledge, experience and desire for
cooperation of the project team. Refer to Figure 1 for a
schematic representation of the work process.
Importantly, the projects and the survival imperative of the
times engendered dialogue, collaboration and cooperation
between geologists, blast and mining engineers, metallurgists
and General Managers (Kanchibotla et al, 1998; Lam et al,
2001; Valery et al, 2001). In 2012, Karen McCaffery suggested:
Mine-to-Mill (and geometallurgy) is just code for making
the effort and putting the processes in place to record, in
an accessible format, an understanding of the orebody,
how changes in the orebody and operating practice drive
productivity and production, and understanding the
operating parameters in the mine and mill which can be
manipulated to improve productivity and operating cost.
It is what people in mining and processing at sites should
be doing as a normal part of their day-to-day business.
(personal communication, 2012)
NEW TOOLS AND NEXT GENERATION OF MINE-TO-MILL PROJECTS
Since 2002, there have been marked developments in image
analysis, GPS, simulation software, radio-frequency tracking
devices for ore, in-plant instrumentation to measure flows,
online particle size monitors, mineral liberation analysis,
geometallurgy and equipment monitoring instrumentation.
Though some early adopters lost their way, new adopters
achieved outstanding improvements, such as Antamina’s
45 per cent increase in SAG mill throughput followed by a
further 10 per cent later (Rybinsky et al, 2011; Valery et al, 2012).
Other case studies have been discussed by Bennett et al (2014);
Hart et al (2011); Dance et al (2007); Diaz et al (2015); Renner et al
(2006); McCaffery et al (2006) and Gomes et al (2010).
Now the productivity imperative has returned it is time for
more operators to learn from the old and the recent successes
and adopt the ABC approach using the new technology tools,
which include:
•• more complex software for blast design, analysis and
management
•• advances in fragmentation modelling
•• image analysis with comprehensive rock fragmentation
software for automated particle sizing at truck dump
and conveyor belt locations
we are metallurgists, not magicians
The ABC of Mine-to-Mill and metal price cycles
FIG 1 – Extract from the Julius Kruttschnitt Mineral Research Centre (JKMRC) Mine-to-Mill brochure 1999.
•• end-to-end modelling packages such as the Integrated
Extraction Simulator (IES) developed by CRC ORE
(Cooperative Research Centre Optimising Resource
Extraction) and provided by JKTech
•• high energy explosives to improve post
fragmentation (Hawke and Dominguez, 2015)
•• tools such as Split FX® that processes 3D point clouds
from LIDAR (light detection and ranging) scans and
photogrammetry to automatically characterise fracture
attributes
•• rock tracking devices such as Metso SmartTagTM
(La Rosa et al, 2007) to track ore from blasthole through
mining, handling, stockpiling and mill feed. Combined
with mill performance and online size data, the effects
of ore and blast changes can be correlated with their
impact on processing, refer to Figure 2
•• GPS digital drilling systems to guide, monitor and save
drill hole patterns
•• measured-while-drilling (MWD) data that captures rock
hardness measurements from blasthole drills to reliably
categorise rock types
•• blast movement monitors accurately measure threedimensional blast movement to minimise ore loss and
dilution and significantly increase ore yield (LaRosa and
Thornton, 2011)
we are metallurgists, not magicians
blast
•• electronic detonators programmable in one millisecond
steps to provide blast flexibility and precision
•• image analysis of particle size distribution in the
stockpile
•• MineWare’s Argus shovel monitor to improve shovel
and operator performance, optimise truckloads and
reduce costs
•• prompt gamma neutron activation analysis (PGNAA)
like GeoScan or CB Omni for online measurement of
87
P Cameron, D Drinkwater and J Pease
FIG 2 – Using a SmartTagTM to track ore batches from blasting (Metso image, from La Rosa et al, 2007).
elements in rock streams on belt conveyors for grade
engineering or stockpile blending
•• safe, accurate and quick measurements of the Crusher
Closed Side Setting (CSS) from 6 to 220 mm through the
‘C-Gap’ digital measurement tool
•• accurate and reliable online particle size monitors
(PSM) and sampling stations to sample and measure the
particle size distribution and per cent solids of the total
grinding circuit classifier product
•• acoustic monitoring software to monitor operating
conditions in SAG mills and allow optimisation
•• automated quantitative mineralogy provides mineral
liberation and association data (using area scan of
particles) and elemental distributions of ore and waste
minerals
•• geometallurgy – the integration of geological, mining,
metallurgical, environmental and economic information
to maximise the net present value (NPV) of an orebody
while minimising technical and operational risk.
20 YEARS LATER – WHAT HAVE WE LEARNED?
For many of the early adopters, the initial success with Mineto-Mill fell victim to organisation systems and personal
incentives that ‘sprang back’ to default when survival was
no longer in doubt. Any reinvention of Mine-to-Mill needs to
recognise why we failed to hold the gains first time, then set
about to fix those flaws.
The fatal flaw was that earlier designs didn’t lock the
new operating methodology into organisational and
management systems. When profits rose in the boom,
attention was focused on expanding output and resources.
Blasting engineers were rewarded for reducing their cost
per tonne and miners were rewarded for increasing tonnage.
Metallurgists were incentivised to increase tons and recovery
and to reduce costs, but were rarely encouraged to increase
product quality beyond ‘good enough to sell’. Smelters
remove impurities at much higher cost than the concentrator
– but that was the smelter’s problem.
If we are to truly succeed with Mine-to-Mill in future,
implementation must be supported, not undermined, by our
organisation systems. We need to design key performance
indicators (KPIs) that encourage integration and work across
the silos to provide mine site targets.
Mining operations are complex. We need as many good
measures as we can find; they are getting better but are still
imperfect. We need to distil the complexity to the simple
fundamental basics of what makes a good integrated
organisational team, and then set the minimum few KPIs
that let clever people get on with their own jobs while using
their initiative for the group benefit. This will keep the
shareholders onside.
TABLE 2
Mine-to-Mill case studies 2003 to 2016.
Antamina and Cerro Corona demonstrated what can be
achieved by applying some of the new tools in a Mine-to-Mill
project. Models of drilling and blasting, crushing and grinding
were combined with a knowledge of different ore domains,
plant surveys and SmartTagTM tracking of blasting changes.
The result was a significant increase in mill throughput
from 2750 to 4400 t/h and a 25 per cent reduction in specific
energy (kWh/t) (Valery et al, 2012). While at Cerro Corona,
throughput was increased by 6 per cent overall (as much as
15 per cent for harder ores) and the SAG specific energy was
reduced by over 9 per cent (Diaz et al, 2015).
Mine site
Metals produced
Country
Product increase
Batu Hiju
Cu
Indonesia
10 to 15%
Antamina
Cu
Peru
45 to 60%
Los Bronces
Cu
Chile
15 to 20%
Cerro Corona
Au
Peru
15% hard ore, 6% overall
Phu Kham
Cu
Laos
8%
Ahafo
Au
Ghana
8%
Morila
Au
Mali
10%
Case studies from 2003 to 2016 are listed in Table 2, and
there are even further operations that have benefited from
Mine-to-Mill but chosen not to publish the results.
Oyo Tolgoi
Cu
Mongolia
25%
Iduapriem
Au
Ghana
21 to 32%
88
we are metallurgists, not magicians
The ABC of Mine-to-Mill and metal price cycles
CONCLUSION
We know Mine-to-Mill works. We made it work well with old
technology. We have more tools and more appropriate data
now. We have a business imperative, so we can make it work
much better again.
We have to recognise the complexity of sites and the
differences between sites. The principles of integration will work
everywhere, but the crucial components will be site-specific.
Off-the-shelf solutions provided by external management
groups won’t work; the components of the solutions need to
be specifically assembled by multidisciplinary teams of people
who know the details and constraints of each site, and who
understand the difference between minimising costs in silos
and maximising mine site profits.
This time we need to support Mine-to-Mill with an
understanding of human and organisational behaviour,
and lock that into our incentive systems. We need to greatly
simplify the overwhelming list of performance targets and
distil them to the critical few that encourage the behaviour
needed to integrate the operation. Then we can allow our
clever people to get on with the job.
ACKNOWLEDGEMENTS
This paper was inspired by observation and comments from
many people that Mine-to-Mill is often overlooked, yet the
same problems which drove the original AMIRA Project are
now repeating themselves.
The authors thank the contributions made by the many
suppliers of the technology tools that are now available for
Mine-to-Mill projects. Valuable contributions were made by
Walter Valery and Sarma Kanchibotla, both of whom have
been associated with Mine-to-Mill projects almost since their
inception.
REFERENCES
Bennett, D, Tordoir, A, Walker, P, La Rosa, D, Valery, W and Duffy,
K, 2014. Throughput forecasting and optimisation at the Phu
Kham Copper-Gold operation, in Proceedings 12th AusIMM Mill
Operators’ Conference 2014, (The Australasian Institute of Mining
and Metallurgy: Melbourne).
Dance, A, Valery, W, Jankovic, A, La Rosa, D and Esen, S, 2007.
Maintaining the benefit – how to ensure mine-to-mill continues
to work for you, in Proceedings Ninth AusIMM Mill Operators’
Conference, (The Australasian Institute of Mining and Metallurgy:
Melbourne).
we are metallurgists, not magicians
Diaz, R, Mamani, H, Valery, W, Jankovic, A, Valle, R, and Duffy, K,
2015. Diagnosis of process health, its treatment and improvement
to maximise plant throughput at Goldfields Cerro Corona, in
SAG 2015 Conference Proceedings, Vancouver, Canada.
Gomes, M P, Tavarez Jnr, L, Nunes, E, Colacioppo, J, Jankovic, A and
Valery, W, 2010. Optimization of the SAG Mill Circuit at Kinross
Paracatu, Brazil, Comminution 10.
Hart, S, Rees, T, Tavani, S, Valery, W and Jankovic, A, 2011. Process
integration and optimisation of Boddington, in SAG 2011
Conference Proceedings, Vancouver, Canada, 25–28 September.
Hawke, S J and Dominguez, L A, 2015. A simple technique for using
high energy, in Proceedings 11th International Symposium on Rock
Fragmentation by Blasting, Sydney, Australia, pp 321–326.
Kanchibotla, S S, Morrell, S, Valery Jnr, W and O’Laughlin, P, 1998.
Exploring the effect of blast design on SAG Mill throughput
at KCGM, presented at the Mine to Mill Conference, Brisbane,
Australia.
La Rosa, D, Valery, W, Wortley, M and Ozkocak, T, 2007. The use
of radio frequency id tags to track ore in mining operations, in
Proceedings 31st International Symposium on Computer Applications
in the Mineral Industries (APCOM), Santiago, Chile, 23–25 April
2007.
Lam, M, Jankovic, A, Valery, W and Kanchibotla, S, 2001. Increasing
SAG mill circuit throughput at Porgera gold mine by optimising
blast fragmentation, in Proceedings SAG 2001 Conference,
Vancouver, Canada.
McCaffery, K, Mahon, J, Arif, J and Burger, B, 2006. Batu Hijau
– controlled mine blasting and blending to optimise process
production at Batu Hijau, in Proceedings SAG 2006 Conference,
Vancouver, Canada.
McKee, D J, 2013. Understanding Mine-to-Mill, 96 p (Cooperative
Research Centre for Optimising Resource Extraction: Brisbane).
Renner, D, La Rosa, D, DeKlerk, W, Valery, W, Sampson, P, Bonney
Noi, S and Jankovic, A, 2006. Anglogold Ashanti Iduapriem
mining and milling process integration and optimisation, in
Proceedings SAG 2006 Conference, pp 249–264, Vancouver.
Valery, W and Rybinski, E, 2012. Optimisation process at Antamina
boosts production and energy efficiency: helping a large copper/
zinc mine meet the economic challenge of processing harder ore
types, Engineering and Mining Journal, September 2012.
Valery, W, Morrell, S, Kovovic, T, Kanchibotla, S and Thornton,
D, 2001. Modelling and simulation techniques applied for
optimisation of Mine to Mill operations and case studies, in
Proceedings of VI Southern Hemisphere Conference on Minerals
Technology, Rio de Janeiro, Brazil, 27–30 May.
89
Contents
Base metals concentrate sales contracts
– change Pavlov and the dog
P D Munro1 and S E Munro2
ABSTRACT
The commonly traded base metals concentrates of copper, lead, tin and zinc are sold by
miners to custom smelters using contracts whose terms were generally established at
the beginning of the 20th century or even earlier. Technical clauses in these concentrates
for items such as payable metals and penalty elements do not reflect current realities of
extractive metallurgical technologies and hygiene/environmental issues.
It is an axiom that the miners and smelters have a symbiotic relationship. While the
existing antique and arcane commercial arrangements serve the purpose of transferring
money from the buyer to the seller, they are not likely to result in the industry achieving
better overall technical and economic outcomes. By examining the cases for copper
concentrates and zinc concentrates in particular, it will be shown that the practical
application of these antiquated technical clauses does not drive the miners to produce
the better quality concentrates that should be desired by the smelters. This is despite
development of mineral beneficiation and hydrometallurgical technologies that would
allow the miners to make better products.
The adage that any extractive metallurgical problem in the smelter is cheaper to solve
in the concentrator still generally holds.
While nickel concentrates are only touched on lightly in this discussion because sales
terms are more opaque than those for the other base metals, it is expected that the same
conclusions will apply.
INTRODUCTION
The commonly traded base metals concentrates of copper, lead, tin and zinc are sold by
miners to custom smelters using contract terms generally established at the beginning
of the 20th century or even earlier. Technical clauses in these concentrates for items
such as payable metals and penalty elements do not reflect current realities of extractive
metallurgical technologies and hygiene/environmental issues. The discussion is
confined to the consequences of these technical clauses rather than commercial ones
such as treatment and refining charge, price participation, quotation period for metal
pricing and so on.
This paper focuses on the situation for the traditional base metals because of the
greater transparency of sales terms and the common use of flotation as the beneficiation
method (tin being the exception by using mostly gravity concentration). Nickel
concentrates are only touched on lightly in this discussion. The cases of other metals
such as molybdenum, tantalum and tungsten are not discussed except where it relates
to flotation performance.
Lewis and Streets (1979) provided a comprehensive and masterly review of base metal
smelter terms from the miner’s point of view which is a good primer for the novice in
the area. Details on the contracts for the individual metals are found in the Australasian
Institute of Mining and Metallurgy publications Cost Estimation Handbook for the
Australasian Mining Industry, Monograph No 20 (1993) and Cost Estimation Handbook, 2nd
Edition, Monograph No 27 (2012).
The reader is referred to the relevant sections for individual concentrates from these
publications cited in the references list of this paper as follows:
•• copper – Wilson and Chanroux (1993a); de Sousa (2012)
•• lead – Wilson and Chanroux (1993b); Watters (2012)
1. FAusIMM, Senior Principal Consulting
Engineer, Mineralurgy Pty Ltd, Taringa Qld
4068. Email: pdmunro@bigpond.com.au
2. MAusIMM, Senior Process Engineer,
Mineralurgy Pty Ltd, Taringa Qld 4068.
Email: semunro@bigpond.com
•• tin – Lewis (1993); Kettle (2012)
•• nickel – Cunningham (1993); Selby and White (2012)
•• zinc – Wilson and Chanroux (1993c); Wise (2012).
On reviewing the sales terms cited in the above papers and those for current contracts
sighted in the course of our work, the striking impression is the continuity ie the lack
91
P D Munro and S E Munro
of change over the last 20 years. The reader does not get the
impression that the base metals industry is now operating in
a situation where it has to deal with vastly heightened public
concerns about toxic elements such as arsenic, lead, and
mercury plus the issue of safely disposing of smelting and
refining residues and wastes etc.
Taking a holistic approach, it seems wrong to make
concentrates containing excessive waste and toxic elements
thereby moving the associated disposal and containment issues
from a remote mine site to a smelter located near a population
centre.
A significant proportion of base metals concentrates are low
quality as measured by ‘weight per cent valuable mineral’.
Quantitative mineralogy often shows both high amounts of
misplaced free diluent minerals and unliberated valuable
minerals in composite particles with both non-sulfide and
sulfide gangue minerals.
The base metals industry is inherently flawed if current
commercial arrangements encourage the miners to make low
quality concentrates without appropriate incentives to use the
best available technologies to make better ones.
THE QUALITY OF BASE METALS CONCENTRATES
Measuring up to an industry benchmark?
The industry usually measures the quality of base metals
concentrates by chemical content of the valuable constituent
for example per cent Cu, per cent Zn etc. A more appropriate
measure of the efficiency of the mineral beneficiation process
(flotation for sulfide minerals or mostly gravity in the case of
tin) is by per cent valuable mineral in the concentrate.
Examining the particular case of sulfide mineral flotation
we use the benchmark of Johnson (February 2015, personal
communication) that an industrial flotation process should
be able to make a concentrate with 85 per cent w/w valuable
mineral from a feed with >80 per cent mineral liberation
(liberation is defined as particles of composition ≥98 per cent
w/w of the target mineral).
Table 1 shows the grade of base metal concentrates using a
composition of 85 per cent w/w valuable mineral.
Contrast the above base metal sulfide concentrate with a
magnetite concentrate (magnetite 72.4 per cent Fe) assaying
68 per cent Fe which contains 94 per cent w/w magnetite.
Some comments on Table 1 are warranted:
•• Molybdenum concentrate – quality is emphasized
probably from the element’s major use as an alloying
element in steel and the situation of a relatively few
‘converters’ roasting the sulfide concentrate to molybdic
oxide (MoO3). By-product molybdenum concentrates
from porphyry copper mines are often subjected to
a chloride leach at the mine site (The Brenda Process –
Jennings, Stanley and Ames, 1973) to reduce copper
and lead levels. Molybdenum concentrates are often
≥55 per cent Mo.
•• Zinc concentrate – while sphalerites in volcanogenic
massive sulfide and Mississippi Valley-type deposits are
relatively ‘pure’, others contain significant amounts of iron
with the mineral being more appropriately described as
marmatite ([Zn, Fe]S). This is the case for Broken Hill ore.
Manganese can also be present in the sphalerite as is the
case for Broken Hill, Dugald River and Gamsberg. Many
zinc concentrates contain >90 per cent w/w sphalerite or
marmatite. A preceding flotation step for copper and/or
lead usually prevents other sulfide minerals reporting to
the zinc concentrate.
•• Copper concentrate – copper concentrates can be
surprisingly poor quality with some of the world’s largest
producers having <60 per cent w/w copper sulfides in
their products despite hubristically quoting grades over
30 per cent Cu. This is because of the presence of the
high copper value supergene copper minerals chalcocitedigenite (79.9 per cent Cu), covellite (66.4 per cent Cu)
and bornite (63.3 per cent Cu).
•• Lead concentrate – those from Mississippi Valley-type
and carbonate-hosted ores can be high quality but those
produced from texturally complex ores such the sedextype deposits in the Carpentaria – Mount Isa Inlier of
Queensland and the Northern Territory, Australia plus
certain volcanogenic massive sulfide deposits such as
Hellyer and Woodlawn struggle to reach 50 per cent Pb.
•• Nickel concentrate – a distinction is made between
concentrates produced from ores with a significant
amount of pyrrhotite gangue such at Sudbury in Canada
and deposits common in Western Australia. The latter
mostly komatiite-hosted nickel deposits are sulfur-poor
and magnesium-rich with altered phyllosilicate minerals
of concern being:
•• the end product ta
•• the end products of the serpentine subgroup such as
chrysotile, antigorite, lizardite etc. Ores with a high
pyrrhotite content can have a significant amount of nickel
associated with the pyrrhotite in pentlandite ‘flames’
and in solid solution so rejecting pyrrhotite decreases
nickel recovery. Concentrate with high per cent Fe is
desired to maintain the required Fe:MgO ratio in the
feed of flash smelting to matte which treats ~75 per cent
of nickel sulfide concentrate (Crundwell et al, 2011). The
high MgO concentrates come from those deposits with
altered phyllosilicate minerals. The mineral separation
problem is difficult under circumstances where the
non-ssulfide gangue can be hydrophobic in the case of
talc and with serpentine minerals positively charged
at alkaline pH while sulfide minerals are negatively
charged causing agglomeration of the mineral species.
As a consequence of the mineralogical factors for
these two types of deposits concentrate grades above
TABLE 1
Base metal concentrate quality as per cent valuable mineral.
Concentrate
Mineral
% w/w Valuable mineral
Grade
Molybdenum
Molybdenite 59.9% Mo
85
50.9% Mo
Zinc
Sphalerite 67% Zn
85
56.9% Zn
Copper
Chalcopyrite 34.6% Cu
85
29.4% Cu
Lead
Galena 86.6% Pb
85
73.3% Pb
Nickel
Pentlandite 34.2% Ni
85
29.1% Ni
92
We are metallurgists, not magicians
Base metals concentrate sales contracts – change Pavlov and the dog
20 per cent Ni are rare. The sulfide nickel segment of the
industry is more integrated than copper, lead or zinc
with major smelters such as Falconbridge, Kalgoorlie,
Norilsk and Copper Cliff receiving most of their feed
from the companies’ own mines.
NEED FOR DIALOGUE
Miners versus smelters
Many participants in the base metals sector do not understand
the technical situations facing both the miners and smelters;
‘technical’ in this context includes environmental factors. This
is compounded by the lack of meaningful dialogue between the
participants. The tremendous growth in the custom base metals
smelting and refining sector in China over the last 20 years has
been accompanied by a retreat from the smelting and refining
business by ‘western’ or ‘first world’ mining companies.
Companies with fully integrated mining + concentrating +
smelting + refining assets seem to becoming rarer along with
possession of the associated ‘in-house’ technical knowledge
across the whole sector.
Contact between miners and smelters is increasingly through
the marketing department of the former meeting the concentrate
buying department of the latter where the ‘deliverable’ is the
concentrate sales contract. The participants in this dialogue/
negotiation are naturally focused on commercial issues rather
than technical ones.
Myths, misconceptions and misunderstandings
Miners
Some common misconceptions that miners have about smelters
that we have encountered are as follows:
•• Smelting and refining is a profitable business that
makes money by ‘gouging’ the miner. This myth is
demolished for copper concentrate by looking at the
history of treatment charge (TC) + refining charge (RC)
from say 1987 to 2014. Up to 2006 the smelters received
15–25 per cent of the value of the contained copper but
since 2007 coinciding with the disappearance of price
participation, their share has fallen to only ~5–8 per cent
due to a concentrate supply deficit caused by the huge
expansion of smelting + refining capacity in China. Thus
the miners have appropriated 50–80 per cent of the value
of the copper in the concentrate that formerly went to
the smelters. The economics of smelting lead, nickel, tin
and zinc concentrates can be expected to be similarly less
attractive than mining the ores containing these elements.
For all base metals the miner receives the largest share of
the value of the contained metal in the concentrate and
operators of high-grade deposits get superior returns
because of their higher comparative advantage. Compare
this to the position of the smelters who receive a minor
and diminishing proportion of the value of the metal in
the concentrate and have limited comparative advantage
as they use the same technologies for smelting and
refining.
•• Smelters ‘want sulfur in the concentrate because it is free
fuel’. It is reasonable to assume that if sulfur is a desirable
constituent in a concentrate then it would receive a
payment. We have never seen a payment made for sulfur
after reviewing hundreds of base metals concentrates
sales contracts over the last five decades.
•• Smelters ‘want silica because it slags out the iron thus
removing the need to buy flux’. See comment for sulfur
We are metallurgists, not magicians
above. Using copper concentrate as an example it is
a fruitless exercise to attempt to optimise the content
of copper, sulfur and silica for a given grade-recovery
curve in a particular beneficiation system. Selecting the
target copper grade of the concentrates fixes the copper
recovery and the sulfur and silica grades.
Miners may not appreciate the increasing environmental
and industrial hygiene pressures applying to the smelters.
These are genuine problems exacerbated by the proximity of
many smelters to population centres. Some smelters operate
in jurisdictions that impose a yearly site limit for particular
elements/compounds, ie limiting the total amount of the
element or compound allowed on the site for the year.
Fountain (2013) has reviewed the situation for elements of
concern in copper smelting and refining.
This is an evolving situation with a seemingly endless
procession of new issues. Some years ago we were aware that
a smelter was concerned about the chromium level in a zinc
concentrate. This seemed strange because chromium is not
known to affect the performance of the conventional roastleach-electrowinning process. However, the leach residue
of this zinc smelter was treated pyrometallurgically to fume
out remaining zinc and lead leaving a discard slag possibly
used as construction fill material. We surmised that because
hexavalent chromium (Cr6+) is a known carcinogen, it would
an easier for the smelter to state that there was negligible
chromium in the slag than to expect that the relevant
regulatory authorities could draw a distinction between Cr3+,
which is not a carcinogen, and Cr6+.
Smelters
The misunderstanding that smelters have about miners is
ignorance that many of them could actually make better quality
concentrates given appropriate incentives.
The miner has the choice of moving along a given graderecovery curve in a particular beneficiation system to make a
higher grade concentrate at a lower recovery or conversely a
lower grade concentrate at a higher recovery. Taking copper
as an example, for at least the past 50 years the economics have
always favoured recovery at the expenses of concentrate grade.
The miner can move to a new grade-recovery curve
(conventionally drawn with concentrate grade on the y-axis
and recovery on the x-axis) pushing the curve up and to the
right by changing the flow sheet and operating conditions. In
many cases the higher concentrate grade from operating on
this new grade-recovery curve will not be accompanied by a
significant recovery loss. This is because the flow sheet and
operating conditions changes can be confined to the cleaner
flotation block. The major loss of the valuable mineral is usually
from the rougher flotation block with the cleaner flotation block
usually contributing less than a third of the total.
CONCENTRATES IN PARTICULAR
Copper
Payment disincentive for superior quality
Normal payment is for 96.5 per cent of the copper content
subject to a minimum deduction of one unit (Lewis and Streets,
1979; Wilson and Chanroux, 1993a; de Sousa, 2012).
Table 2 shows how payment terms vary with the copper
content of the concentrate.
Miners making a low-grade concentrate at 15–20 per cent
Cu appear to be treated relatively generously compared with
producers of higher grade materials. Lower grade concentrates
93
P D Munro and S E Munro
A 1.7 per cent absolute increase in metal payment seems
inadequate recompense for the miner who makes a 75 per cent
Pb concentrate compared with one at 45 per cent Pb that has
only 60 per cent of the lead per tonne of concentrate. The lower
grade material probably leaves the smelter with more sulfur to
be eliminated as sulfuric acid. So-called ‘black acid’ from lead
and zinc smelters is disadvantaged in the market because of
concerns about toxic elements such as mercury, cadmium etc
entering the food chain via phosphate fertiliser. Other usual
diluent elements such as zinc and iron must be removed as
slag, increasing lead loss and subsequent disposal cost of what
must be considered a ‘sensitive’ material as it contains elements
such as lead, zinc, and cadmium.
will produce more sulfuric acid and/or make more slag
increasing copper loss. The higher slag volume increases
disposal cost and requires more silica flux if the concentrate
isn’t self-fluxing.
Arsenic problem
Current ‘world scale’ copper smelters produce ~300 000–
500 000 t/a of copper (Schlesinger et al, 2011). Assuming a mean
feed concentrate grade of 25 per cent Cu then the 500 000 t/a
unit requires >2 000 000 t/a of concentrate feed. Using the
limit of <0.5 per cent As content for concentrates imported
into China (Anon, 2006) means such a smelter has to deal with
up to 10 000 t/a of arsenic or 13 200 t/a as arsenic trioxide.
Halving the arsenic content of the feed to 0.25 per cent As still
produces an alarmingly high amount of arsenic that has to be
stored because of insufficient demand for it.
Technology changes
The conventional lead smelting and refining route is as follows
(Sinclair, 2009):
Arsenic does not assist in the smelting of concentrate to
anode copper with its only technical function needing to be of
sufficient concentration in the electrolyte when electrorefining
anodes to eliminate the formation of floating slimes and assist
the precipitation of antimony and bismuth in the slimes layer.
•• sulfur elimination in a sinter plant
•• bullion production in a blast furnace
•• thermal refining:
•• drossing in two stages to remove copper and arsenic
Some very large copper deposits such as La Granja and
Tampakan remain undeveloped because metallurgical test
work has shown unacceptably high levels of arsenic in the
copper concentrate.
•• oxygen softening or Harris process to remove arsenic,
antimony and tin
•• desilvering by the Parkes process
•• vacuum dezincing
Arsenic penalties (Wilson and Chanroux, 1993a; de Sousa,
2012) are generally in the range US$2.50–5.00 per 0.1 per cent
As over 0.2 per cent As. Concentrate >0.5 per cent As cannot
be imported into China with the Japanese and South Korean
smelters unwilling to take material above 0.15–0.2 per cent
As. High arsenic concentrates enter the market by incurring
a negotiated higher treatment charge reflecting the cost of
shipping to a destination and the cost of supplying ‘clean’
material to blend down the arsenic content. This is usually
done through a trading house.
•• bismuth removal by the Kroll-Betterton process, and
•• caustic refining.
New lead smelting technologies include bath smelting
processes such as the popular Shui Kou Shan (SKS) process
from China (15 plants operating and a further 15 under
construction as of 2013), Ausmelt, Isasmelt and less commonly
the Queneau-Schuhmann-Lurgi (QSL) process with minor
contributions from other technologies such the Kivcet and
Boliden Kaldo processes.
Thus the industry’s tactics of dealing with the arsenic issue
are the somewhat passive responses of rejecting ‘high arsenic’
material and limited attempts at blending down the arsenic
level.
Electrolytic refining by the Betts process has never been as
well accepted by ‘western’ countries as in Japan and China.
The enormous growth of the Chinese lead smelting industry
means that the Betts process with associated limited thermal
treatment is now probably the dominant refining process.
Lead
Normal payment is for 95 per cent of the lead content subject to
a minimum deduction of three units (Lewis and Streets, 1979;
Wilson and Chanroux, 1993b; Watters, 2012).
For the conventional lead smelting process a certain amount
of zinc in the concentrate improves sinter quality (Sinclair, 2009)
as the formation of an appropriate proportion of zinc melilite
(hardystonite) gives the desired microstructure for optimum
blast furnace operation with high softening temperature and
open porosity (McLoughlin, Riley and McKean, 1980).
Table 3 taken from Sinclair (2009) shows how payment terms
vary with the lead content of the concentrate.
New lead smelting technologies do not make sinter thus
removing the need for zinc in the concentrate. The now
Payment disincentive for superior quality
TABLE 2
Copper concentrate metal payments.
Concentrate copper content (%)
Copper in concentrate paid for (%)
Copper content paid for (%)
50
48.25
96.5
45
43.425
96.5
40
38.6
96.5
35
33.775
96.5
30
28.95
96.5
25
24
96
20
19
95
15
14
93.3
94
We are metallurgists, not magicians
Base metals concentrate sales contracts – change Pavlov and the dog
TABLE 3
Lead concentrate metal payments (after Sinclair, 2009).
Lead in concentrate paid for (%)
Lead content paid for (%)
75
Concentrate lead content (%)
71.25
95
65
61.75
95
60
57
95
55
52
94.5
50
47
94
45
42
93.3
dominant SKS process treats the lead-rich slag from the firststage bath smelting process in a conventional blast furnace.
J Tuppurainen (February 2015, personal communication)
advises that blast furnace practice is to maintain a zinc:iron ratio
of feed material less than 0.7:1 to effectively remove the zinc in
the discard slag. Zinc contained in the blast furnace slag can be
recovered by fuming but maintaining reducing conditions at
high temperatures is an expensive process.
Zinc in lead concentrate has moved from being a ‘necessary
evil’ in the conventional sinter plant + blast furnace process
route to being a nuisance in processing by the current dominant
lead smelting technologies. This leads to the question whether
zinc contained in lead concentrate is being penalised too lightly
by the smelters.
Conversely, the predominance of electrolytic refining over
thermal refining calls into question the validity of a severe
penalty charge for an element such as bismuth which is readily
removed by the Betts process. Bismuth and antimony are
usually intimately associated with the lead mineralisation and
not rejected by physical beneficiation.
Tin
Tin concentrate contract terms affected by quality
The deductions and charges for tin concentrates shown in
Lewis and Streets (1979) and later papers (Lewis, 1993; Kettle,
2012) are generally more variable than for other base metals
concentrates. This reflects tin contents varying from 20 per cent
Sn to 70 per cent Sn; a range rarely encountered in other
concentrates. Charges for penalty elements such as US$120–150
per 1 per cent As and $300–500 per 1 per cent Cu are very high
compared with other concentrates.
Tin concentrate sales contracts increase the base deduction
with increasing iron content which is logical considering
the difficulties of recovering tin from the tin-iron alloy
called ‘hardhead’.
Quality change from hard rock source
Glen (February 2015, personal communication) notes that
depletion of alluvial tin resources means that relatively more
tin concentrates are being produced from ‘hard rock’ deposits
with custom tin smelters concerned about elevated levels of
arsenic levels (Glen, 2015). Arsenic is a real issue in tin smelting
as its removal from molten tin is by the addition of aluminium
which forms a dross containing the two impurities. Wright
(1966) describes this dross as ‘… probably the most dangerous
substance met in classical metallurgy’ because of its propensity
to evolve arsine on contact with water or moisture in the air.
Arsine is 50 times more toxic than hydrogen cyanide.
Custom tin smelters are examining treatment of incoming
tin concentrates at their plants by flotation to reduce arsenic
(presumably contained in arsenopyrite) to acceptable levels.
We are metallurgists, not magicians
While it would be preferable that this would be done at the
mines, the small-scale of many operations may make it
impractical.
Zinc
Payment disincentive for superior quality
Zinc concentrate sales terms pay for 85 per cent of the contained
zinc subject to a minimum deduction or free metal allowance of
eight units. This deduction is an antique term, possibly related
to the relatively low first pass recovery of zinc in the now
obsolete horizontal retort process dating from the 18th and
19th centuries. Modern electrolytic plants which produce over
90 per cent of the world’s zinc (Sinclair 2005), achieve well over
95 per cent zinc recovery.
Table 4 taken from Sinclair (2005) shows how payment terms
vary with the zinc content of the concentrate.
Thus the smelter provides little incentive for the miner to
make a concentrate much over 52 per cent Zn except to save
transport costs to the smelter. Lewis and Streets (1979) observed
a similar ‘flattening’ out of payable zinc for concentrate grades
above 54–55 per cent Zn.
The elimination of by-product revenue has changed the
economics of the zinc smelting business. In the 1980s zinc
smelters could expect to receive payment for products such as
cadmium, cobalt, copper residue, and mercury, lead residue
containing silver and sulfuric acid extracted during processing
of the concentrate (Hamdorf and Woodward, 1980). Few zinc
plants in the ‘western’ world are currently paid for these byproducts and it is reasonable to assume that the Chinese
smelters will reach this position in future. Toxic elements
associated with the extractive metallurgy of zinc such as
cadmium, lead, mercury and thallium are other ‘hot-button’
issues for the public.
Iron problem
Table 5 taken from Sinclair (2005) for a conventional roast-leachelectrowinning plant shows the dependence of zinc extraction
in primary leaching on the iron content of the zinc concentrate.
High iron levels in the concentrate increase both the amount
of primary leach residue that has to be processed by secondary
leaching and the amount of leach residue for ultimate disposal.
It is anomalous that payment terms shown in Table 4 level
out for concentrates assaying 52 per cent Zn and above
considering that disposal of iron-rich residue from acid
leaching of calcine and/or concentrate is a major problem
for zinc smelters (the term ‘smelters’ includes all zinc
extractive metallurgical plants). This is particularly true
for zinc smelters located near population centres. Storage
of jarosite or goethite-type leach residue in ponds is under
regulatory pressure. Conversion of the leach residue to
hematite or alternatively making a relatively benign slag by
95
P D Munro and S E Munro
TABLE 4
Zinc concentrate metal payments (after Sinclair, 2005).
Concentrate zinc content (%)
Zinc content paid for (%)
Zinc content as free metal to smelter (%)
Zinc content paid for (%)
46
38
8
82.6
48
40
8
83.3
50
42
8
84.0
52
44
8
84.6
54
45.9
8.1
85.0
56
47.6
8.4
85.0
TABLE 5
Primary leach zinc recovery and residue composition (after Sinclair, 2005).
Zinc in concentrate %
Iron in concentrate %
Zinc recovery %
Zinc in primary leach residue %
Primary residue amount % of concentrate
48
12
82.8
22.8
36
50
10
85.6
22.8
32
52
8
88.3
22.5
28
54
6
90.7
21.7
24
56
4
92.9
20.9
19
pyrometallurgical treatment such as the Waelz process or
fuming by smelting is expensive.
The case of the Budel zinc smelter in The Netherlands is a
portent for the future of leach residue disposal. Since 1999
on-site storage of residue has been banned with the former
leach residue ‘waste’ now mandated to be a ‘lead product’
sent for smelting by another party. Budel has achieved this
by using low iron zinc concentrate from the Century Mine in
Queensland as its major feed source to minimise the amount
of this ‘lead product’ (Sinclair, 2007).
bearing their full share of downstream processing costs. In a
situation where the industry has a finite capacity to deal with
input iron in concentrate, removing some iron from highgrade concentrates allows more room for iron in low-grade
concentrates albeit at lower returns for miners producing
such materials.
Ameliorating the iron issue does not resolve the problem of
elements such as cadmium, mercury and thallium which are
usually too closely associated with the sphalerite/marmatite to
be removed by physical beneficiation.
Given this concern about iron it seems surprising that the
penalty for iron in a zinc concentrate hasn’t changed in nearly
20 years being US$1.50 per 1 per cent Fe above 8 per cent Fe
(Wilson and Chanroux, 1993c; Wise, 2012).
Greater use of existing technologies and practices
A new zinc concentrate contract?
Miners can often make higher grade base metals concentrates
from flotation by:
We suggest that the zinc industry move to a new zinc
concentrate sales contract that more appropriately aligns the
capabilities of the miners with the needs of the smelters. This
new contract would comprise the following:
•• Treatment charge reflecting the actual cost of producing
zinc metal from concentrate. Sinclair (2005) estimated
operating costs for a 200 000 t/a electrolytic plant as of June
2000 to be US$383/t (or US$191.50/t of zinc concentrate
with 50 per cent recoverable metal content). Power in
his calculation was US$0.04/kWh. Cost of jarosite leach
residue disposal into rubber-lined ponds holding up to
five years of production was US$6.42/t of zinc.
•• Realistic payment for contained zinc eliminating the
‘hocus pocus’ of ‘eight units of free metal’ and adjustments
to the treatment charge.
•• Iron penalty commensurate with the cost of dealing
with the leach residue. A concentrate with 50 per cent
recoverable zinc and 8 per cent Fe generates at least 0.3 t
of residue per tonne of zinc metal produced.
Obviously there are going to be ‘winners’ and ‘losers’ among
the miners from such a contract. Miners with deposits of high
iron sphalerites will receive a lower return than miners able
to make high-grade concentrates. However, this is required
to achieve a better overall industry outcome. We could argue
that currently producers of low-grade concentrates are not
96
MINERAL PROCESSING SOLUTIONS
•• increasing mineral liberation of the target mineral feed
going to the cleaner flotation block; too often the
benchmark condition mentioned above of >80 per cent
mineral liberation for the valuable mineral in the feed is
not achieved
•• improving mineral separation; this is evidenced by the
presence of excessive amounts of free diluents such as
iron sulfide minerals and non-sulfide gangue.
Johnson and Munro (2002) have reviewed the technologies
and plant practices developed to treat complex sulfide ores
such as the zinc-lead-silver ores of the Mount Isa Inlier (Young
et al, 1997) and difficult volcanogenic massive sulfide ores such
as Hellyer (Lane and Richmond, 1993).
Better understanding of pulp chemistry together with
using ultra-fine grinding and washed froth cleaning have
been successfully applied in concentrators such as Century
Zinc (Barham and Kirby, 2001; Obeng et al, 2013), McArthur
River (Rossberg and Pafumi, 2013), Mount Isa – George Fisher
(Pafumi et al, 2013), Phu Kham (Bennett, Crnkovic and Walker,
2012) and Prominent Hill (Barns, Colbert and Munro, 2009).
A seemingly forgotten powerful flotation practice is reverse
flotation as exemplified by the so-called Brunswick process
(McTavish, 1980) which was used at Brunswick Mining and
Smelting and later at Kidd Creek. Sphalerite was depressed
by heat and sulfur dioxide addition with pyrite removed
We are metallurgists, not magicians
Base metals concentrate sales contracts – change Pavlov and the dog
as the concentrate. The standard by-product molybdenum
separation from copper-molybdenum ores can be considered
to be a type of reverse flotation process upgrading the copper
concentrate.
If these existing technologies and plant practices have to be
applied to ‘difficult’ ores to achieve ‘ordinary’ metallurgical
outcomes then using them on ‘easy’ ores such as a porphyry
copper deposit would achieve ‘extraordinary’ metallurgical
outcomes compared with current ones.
One consequence of better quality concentrates will
probably mean miners and smelters having to cope with
finer particle size distributions. For the miners this means
froth pumping and concentrate dewatering. This shouldn’t
be alarming as the plants mentioned above are successfully
dealing with these issues. A far bigger potential problem is
the solid-liquid separation issues with the movement to dry
stacking of tailings if regulators ban the conventional ‘wet’
tailings dam. For the smelters finer concentrate particle size
distribution means more dust.
The reason that these technologies and practices aren’t used
more widely by the miners to make better quality concentrates
is because technical terms in the sales contracts do not provide
any monetary incentive to do so.
As mentioned above a magnetite concentrate at 68 per cent
Fe contains 94 per cent w/w magnetite which is much higher
quality than the base metal concentrates in Table 1. Higher
capital and operating costs are needed to produce magnetite
concentrate compared with direct shipping ore (DSO) so
a high quality product has to be made to get the premium
offered over the DSO material. Blast furnace iron producers
are much more conscious of feed quality than base metals
smelters. An argument against using iron ore sector practice
as a parable for base metals is that iron ore is a low value
commodity mined from very large deposits where recovery
isn’t important. One tonne of iron ore at 30 per cent Fe with
iron ore at US$100/t has the same in situ value as one tonne
of copper ore at 0.5 per cent Cu with copper at US$6000/t.
Large porphyry copper deposits can be the same size as iron
ore deposits.
Magnetite concentrates are high quality because they have
to be, customers won’t accept poor quality. It exemplifies
Sinclair’s reiteration (2007) of the business truism ‘that if you
don’t have customers you really don’t have a viable project’.
A viable technical solution to the arsenic problem?
High arsenic levels in copper concentrate are usually caused
by the presence of the copper arsenic sulfide minerals enargite
(Cu3AsS4) and tennantite ([Cu, Fe]12As4S13 or Cu13As4S13).
These minerals behave in flotation like other copper sulfide
minerals and report to the concentrate. Arsenic in the form
of arsenopyrite should be readily dealt with as it behaves
like pyrite.
The proven mine site solution to high arsenic concentrates
is roasting to remove the element as arsenic trioxide. This
was done at Lepanto and El Indio and is currently practised
at Codelco’s Ministro Alejandro Hales operation (formerly
known as Mansa Mina). A less appealing solution from both
the point of view of the miner and custom smelter is for the
former to produce copper metal on-site by smelting or the
pressure oxidation + solvent extraction + electrowinning route.
•• Separation of the copper concentrate into a high tonnage
lower arsenic portion and a low tonnage higher arsenic
portion by exploiting Eh-pH flotation separation
‘windows’ for the different copper mineral present.
Ametov has demonstrated the possibility of a reasonable
separation of chalcopyrite + pyrite versus chalcocite
+ bornite (Ametov et al, 2014). Ma and Bruckard (2009)
show that similar ‘windows’ exist for the arsenic carriers.
It is unreasonable to expect that there will be clean
separation with all the copper arsenic minerals reporting
to one product.
•• Hydrometallurgical treatment of the high arsenic
concentrate by an emerging technology such as the
Toowong process (Nakon and Way, 2014; Turner, 2015)
which uses an alkaline leach. This has the advantage
over roasting of leaving the main copper sulfide minerals
substantially unchanged. Arsenic is fixed using a range
of conventional methods and the stabilised product
can be disposed into a suitable engineered facility.
Alternatively all the copper concentrate is treated
through the Toowong process reducing the arsenic
content to 0.05 per cent As.
SUMMARY
The technical terms in contracts for the sales of for base
metals concentrates show little change over the last 20 years
or even the last 50 years. This is despite all sectors of the
primary base metals industry now operating in a situation
where they have to deal with vastly heightened public
concerns about toxic elements such as arsenic, lead, and
mercury plus the issue of safely disposing of smelting and
refining residues and wastes etc.
For mineral beneficiation by sulfide mineral flotation, many
miners including some of the world’s largest producers do
not achieve our benchmark that an industrial flotation process
should be able to make a concentrate with 85 per cent w/w
valuable mineral from a feed with >80 per cent mineral
liberation.
Mineral processing developments over the past 20 years, such
as better understanding of pulp chemistry, ultra-fine grinding
and washed froth cleaning have been successfully applied to
the treatment of complex or ‘difficult’ ores.
Using them on ‘easy’ ores such as porphyry copper deposits
would achieve ‘extraordinary’ metallurgical outcomes
compared with some current mediocre ones leading to a general
improvement in the quality of base metals concentrates.
The smelters, after appropriate dialogue with the miners,
should alter the technical clauses in base metals concentrate
sales contracts appropriately affecting economic returns. This
should elicit the appropriate Pavlovian response in terms of
concentrate quality.
If the industry doesn’t hang together on this then its
constituent components will hang separately!
ACKNOWLEDGEMENTS
The authors would like to acknowledge Mineralurgy Pty Ltd
for the permission to publish this paper.
Examination of the technical literature suggests the interesting
possibility of a mine site treatment route as follows:
We thank Mr Craig Walter formerly of MIM Holdings
Limited and later of Trafigura Beheer BV for his acute
observations, wise guidance and patient mentoring in the areas
of concentrate sales and metals marketing over 30 years.
•• Product of a copper concentrate by conventional or
enhanced means.
Mr Jorma Tuppurainen Senior Principal Consulting Engineer
with Mineralurgy Pty Ltd kindly provided technical advice
We are metallurgists, not magicians
97
P D Munro and S E Munro
from his considerable expertise in the extractive metallurgy of
copper, lead, nickel and zinc.
Mr Roderick Sinclair is thanked for his very useful
publications The Extractive Metallurgy of Lead (2009) and The
Extractive Metallurgy of Zinc (2005) from which we have drawn
heavily in writing this paper.
REFERENCES
Ametov, I, Harmer, S, Grano, S and Alford, R, 2014. Selective flotation
of chalcopyrite and pyrite from a mixed copper mineral concentrate
under controlled redox potential condition, in Proceedings 12th
AusIMM Mill Operators’ Conference, pp 131–140 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Anon, 2006. The specification for limit on harmful element content of
the heavy metal concentrates products (English trans) National
Standard of the People’s Republic of China GB 2044–2006.
Barns, K E, Colbert, P J and Munro, P D, 2009. Designing the optimal
flotation circuit – the Prominent Hill case, in Proceedings Tenth
AusIMM Mill Operators’ Conference, pp 173–182 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Bennett, D, Crnkovic, I and Walker, P, 2012. Recent process
improvements at the Phu Kham Copper-Gold concentrator,
Laos, in 11th AusIMM Mill Operators’ Conference, pp 257–272 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Crundwell, F K, Moats, M S, Ramachndaran, V, Robinson, T G and
Davenport, W G, 2011. Extractive Metallurgy of Nickel, Cobalt and
Platinum-Group Metals (Elsevier: Amsterdam).
Cunningham, W H, 1993. Nickel, in Cost Estimation Handbook for the
Australian Mining Industry (eds: M Noakes and T Lanz) (Chapter 16)
pp 356–358 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
de Sousa, A, 2012. Copper, in Cost Estimation Handbook, Second
Edition (Chapter 20) pp 483–486 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Fountain, C, 2013. The whys and wherefores of penalty elements in
copper concentrates, in Proceedings MetPlant 2013, pp 502–518 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Hamdorf, C J and Woodward, O H, 1980. Zinc production by
electrolytic zinc company of Australasia Ltd, Risdon, Tas, in
Mining and Metallurgical Practices in Australia (ed: J T Woodcock),
pp 254–261 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Jennings, P H, Stanley, R W and Ames, H L, 1973. Development of
a process for purifying molybdenite concentrates, in Proceedings
International Symposium on Hydrometallurgy (eds: D J I Evans and
R S Shoemaker), pp 868–883 (Society of Mining Engineers of AIME:
New York).
Johnson, N W and Munro, P D, 2002. Overview of flotation technology
and plant practice for complex sulfide ores, in Mineral Processing
Plant Design, Practice and Control Volume 1 (eds: A L Mular,
D N Halbe and D J Barratt), pp 1097–1123 (Society for Mining,
Metallurgy and Exploration: Littleton).
Ma, X and Bruckard, W J, 2009. Rejection of arsenic minerals in
sulphide flotation – A literature review, International Journal of
Mineral Processing, 93:9–94.
McLoughlin, B C, Riley, J F and McKean, G R, 1980. Development
of optimum fluxing procedures for isamine lead concentrates, in
Lead-Zinc-Tin ‘80 (eds: J M Cigan, T S Mackey and T J O’Keefe),
pp 309–320 (Metallurgical Society of AIME: Warrendale,
Pennsylvania).
McTavish, S, 1980. Flotation practice at Brunswick Mining, CIM
Bulletin, 115:115–120.
Nakon, D and Way, D M, 2014. Method for Treating Arsenic Containing
Materials, US Patent 8771619 B2 (Xstrata Queensland Ltd).
Obeng, D, Twining, D, Milsom, J, Brown, M, Smith, M, Horton,
R, Ackroyd, T and Smith, W, 2013. Century Mine MMG, in
Australasian Mining and Metallurgical Operating Practices (ed: W J
Rankin), pp 1661–1682 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Pafumi, J, Rossberg, F, Golling, J, Armstrong, C, Walker, R, Grobler, J,
Boyer, J, Meyer, R and Beer, A, 2013. George Fisher Mine Xstrata
Zinc, in Australasian Mining and Metallurgical Operating Practices
(ed: W J Rankin), pp 1639–1660 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Rossberg, F and Pafumi, J, 2013. McArthur River mine xstrata zinc, in
Australasian Mining and Metallurgical Operating Practices (ed: W J
Rankin), pp 1543–1668 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Schlesinger, M E, King, M J, Sole, K C and Davenport, W G, 2011.
Extractive Metallurgy of Copper, 5th edition (Elsevier: Amsterdam).
Selby, M and White, D T, 2012. Nickel, in Cost Estimation Handbook,
Second Edition (Chapter 20), pp 502–507 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Sinclair, D F, 2007. Project evaluation – the customer’s perspective, in
Proceedings Project Evaluation 2007, pp 203–208 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Sinclair, R J, 2005. The Extractive Metallurgy of Zinc (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Sinclair, R J, 2009. The Extractive Metallurgy of Lead (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Turner, D, 2015. The Toowong Process, Core Resources Pty Limited,
Albion, Queensland.
Watters, T, 2012. Lead, in Cost Estimation Handbook, Second Edition
(Chapter 20), pp 493–495 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Wilson, P C and Chanroux, C, 1993a. Copper, in Cost Estimation
Handbook for the Australian Mining Industry (eds: M Noakes and
T Lanz) (Chapter 16), pp 346–348 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Wilson, P C and Chanroux, C, 1993b. Lead in Revenue Calculations and
Marketing of Cost Estimation Handbook for the Australian Mining
Industry (eds: M Noakes and T Lanz), (Chapter 16) pp 337–339 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Kettle, P, 2012. Tin, in Cost Estimation Handbook, Second Edition
(Chapter 20), pp 511–514, (The Australasian Institute of Mining
and Metallurgy: Melbourne).
Wilson, P C and Chanroux, C, 1993c. Zinc, in Cost Estimation Handbook
for the Australian Mining Industry (eds: M Noakes and T Lanz)
(Chapter 16), pp 364–367 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Lane, G S and Richmond, G D, 1993. Improving fine particle flotation
selectivity at Hellyer, in Proceedings XVIII International Mineral
Processing Congress, vol 3, Flotation 1, pp 887–905 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Wise, W, 2012. Zinc, in Commodity Marketing of Cost Estimation Handbook
Second Edition, (Chapter 20) pp 516–521 (The Australasian Institute
of Mining and Metallurgy: Melbourne).
Lewis, P, 1993. Tin, in Cost Estimation Handbook for the Australian
Mining Industry (eds: M Noakes and T Lanz) (Chapter 16) pp 360–
362, (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Lewis, P K and Streets, G C, 1979. An analysis of base-metal
smelter terms, in Proceedings Eleventh Commonwealth Mining and
Metallurgical Congress (ed: M J Jones), pp 753–767 (The Institution
of Mining and Metallurgy: London).
98
Wright, P A, 1966. Extractive Metallurgy of Tin, pp 138–139 (Elsevier:
Amsterdam).
Young, M F, Pease, J D, Johnson, N W and Munro, P D, 1997.
Developments in milling practice at the lead/zinc concentrator of
Mount Isa mines limited from 1990, in Proceedings Sixth AusIMM
Mill Operators’ Conference, pp 3–12 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
We are metallurgists, not magicians
Project design
Contents
Karouni Gold Project from drill
core to commissioning
K Nilsson1 and D Connelly2
ABSTRACT
Troy Resources acquired the Karouni Gold project in Guyana, South America in 2014
and immediately proceeded to a Pre-Feasibility Study and project implementation.
The study considered two open pit mines feeding a conventional carbon-in-leach
(CIL) gold plant with a capacity of 1 Mt/a. The study assumed a total of 2.61 Mt of ore
with an average grade of 3.84 g/t gold with a recovered gold production of 303 526 oz
over a three year mine life.
The project development involved drilling metallurgical test holes, securing
second-hand major pieces of equipment where suitable and arranging shipping to
Guyana. The remote location meant logistical considerations were paramount. The
plant consists of conventional two-stage crushing with single stage ball milling and a
gravity circuit. The tailings circuit includes cyanide detoxification (DETOX).
The extremely high rainfall presented challenges with water management and
the necessity to dispose of clean water to the environment. The DETOX was also
mandatory due to the highly sensitive environment and past history in the area. The
criticality of this meant that in the design, any cyanide in the tailings would precipitate
in a plant shutdown.
Lessons were learned and the project has been both profitable and successful for
Troy Resources in a new operating location. In addition, the project has created jobs,
delivered royalties and brought new skills into an area where previously there was
little current mining activity.
PROJECT HISTORY
The Karouni project was previously owned by Azimuth Resources. The project
transferred to Troy Resources following their acquisition of Azimuth Resources in July
2013. DDrilling by Azimuth established inferred resources of 1.6 Moz for the combined
Smarts and Hicks deposits, with Smarts being the higher grade of the two (see Figure 1).
Guyana is host to several active junior mining companies, and also has activity from
the major minors as well, including BHP, Newmont and Lamgold. In 2012, mining in
Guyana produced over 400 000 oz of gold.
A majority of the population of Guyana lives along the coastal areas, with a growing
amount of exploration occurring in the interior of the country. Resources within the
interior include diamonds, semi-precious stones and gold (see Figure 1).
GEOLOGY
The Guyana Shield has been correlated with the Leo-Man Shield of West Africa and
it is generally accepted that prior to the opening of the Atlantic during the Mesozoic,
the two shields formed a contiguous craton. The Archaean Imataca Complex can be
correlated with the Archaean Liberian Province, the Central Guyana Granulite Belt
with the Dimbroko Zone in Ivory Coast, the Barama-Mazaruni greenstones with
the Birimian greenstones and the Trans-Amazonian tectono-thermal event with the
Eburnean Orogeny.
METALLURGICAL TEST WORK
1. Project Director, Troy Resources Ltd,
West Perth WA 6005.
2. MAusIMM(CP), Director/Principal Consulting
Engineer, Mineral Engineering Technical
Services Pty Ltd (METS), Midas Engineering
Group, Perth WA 6000. Email:
damian.connelly@metsengineering.com
The test work program investigated three ore zones; primary, shear and oxide, over
a range of processing parameters. This included comminution, gravity, cyanide
leaching, variability test work and ore process characterisation. Additionally, some
tests were conducted on material classified as sulfide. This material was subjected to
preliminary comminution testing and checked for gold grades.
Samples were selected based on geological data, orebody cross-sections and a review
of the core logs and assays as provided by Troy Resources to METS. Composites were
101
K Nilsson and D Connelly
FIG 1 – Karouni Gold project location.
formed based on this information including assay data for
further testing, both for variability and master composites.
one of the sulfide samples, with the high sulfide result being
classified as hard (see Table 2).
Comminution results
SAG mill comminution
Comminution testing was conducted on all areas of the
orebody.
SAG mill comminution testing is conducted to further evaluate
ore hardness and competency for size reduction, with a
smaller starting size than that of the crushing tests. The shear
material was seen to be the least competent, being classified as
soft–medium. The primary and sulfide materials were seen to
be more competent as hard to very hard. These rankings were
achieved on the basis of the A*b values (material hardness),
and the ta. values (resistance to abrasion) (see Table 3).
Unconfined compressive strength
The oxide material was classified as very weak, while the
primary and shear materials were classed in the range of
weak to medium-strong. The sulfide ore tested was seen to
be the most competent material, being classified as mediumstrong (see Table 1).
Simulations indicated SAG milling was an option for the
process plant.
Crushing work index
Bond ball mill work index
The crushing work index results showed that all material was
classified as ‘very soft’. There will be no identified issues as
a result of processing this through the crushing circuit. The
results ranged from 0.4 kWh/t through to 17.9 kWh/t for
The Bond ball mill work index results indicated that all
samples were medium to hard, across all zones tested. The
range of values was between 12.57 kWh/t and 15.28 kWh/t,
Table 1
Unconfined compressive strength (UCS) results summary.
102
Hole
Zone
UCS ranking
Average
Max
Min
Std dev
99
Primary
Medium-strong
25.65
45.62
12.93
14.29
100
Oxide
Very weak
2.75
3.30
2.20
0.55
100
Primary
Weak
8.85
14.35
3.50
4.43
101
Shear
Medium-strong
29.53
80.57
8.37
27.08
102a
Primary
Weak
21.76
39.56
10.59
7.52
102a
Shear
Weak
16.67
45.50
4.48
14.60
103
Primary
Medium-strong
28.14
32.52
20.75
5.26
103
Sulfide
Medium-strong
34.62
133.31
6.55
31.66
We are metallurgists, not magicians
Karouni Gold Project from drill core to commissioning
Table 2
Crushing work index (CWi) results summary.
Hole
Zone
#CWi
CWi(kWh/t, average)
Max
Min
Std dev
Ranking
99
Primary
21
3.13
4.79
2.39
0.73
Very soft
100
Oxide
20
1.46
3.52
0.37
1.05
Very soft
101
Primary
15
3.87
6.01
3.38
0.73
Very soft
101
Shear
20
4.64
9.62
3.77
1.80
Very soft
102a
Primary
30
4.24
7.25
3.44
1.12
Very soft
102a
Shear
15
5.14
11.21
2.41
2.18
Very soft
103
Primary
21
4.42
9.51
2.64
1.97
Very soft
103
Sulfide
34
5.87
17.93
2.37
3.03
Very soft
103
Shear
19
5.12
7.69
2.36
1.43
Very soft
Table 3
SAG mill comminution (SMC) results summary.
Sample ID
A
b
A*b
Rank
ta
Rank
SDD102a – Primary
58.7
0.63
37.0
Hard
0.35
Hard/mod hard
SDD103 – Primary
54.6
0.53
28.9
Very hard
0.28
Very hard
SDD103 – Shear
51.9
0.95
49.3
Medium
0.48
Medium
SDD103 – Sulfide
69.9
0.36
25.2
Very hard
0.24
Hard
SDD102a – Shear
46.3
1.81
83.8
Soft
0.83
Soft
with the highest value being recorded by the sulfide ore, and
the lowest by the shear ore.
Bond rod mill work index
Rod mill work index testing was conducted on the primary
and shear master composites. Both results showed the material
to be hard, with results of 19.9 kWh/t and 17.5 kWh/t for the
primary and shear zones respectively.
Bond abrasion index
The primary and shear master composites were tested. The
primary material returned a result of 0.13, which classifies this
ore as mildly abrasive. The shear material returned a value of
0.09, which is non-abrasive. These results indicate that there
will not be any undue wear experienced in a crushing circuit.
Master composite testing
The gravity recovery varied between the three composites,
with the oxide displaying a gravity recovery of around
30 per cent, the shear 40 per cent and the primary 60 per cent.
Given the high gravity performance, it was recommended
that a gravity stage be included in the processing circuit flow
sheet.
Quantitative Evaluation of Minerals by SCANning
Electron Microscopy (QEMSCAN) – mineralogy
A QEMSCAN analysis was carried out on heavy liquid
separation sinks and floats following desliming of a subsample
of the primary master composite material. The floats contained
a majority of quartz, feldspar and other silicates. Based on
work conducted in the test work program, the inclusion of the
feldspar and silicates shows negligible negative effect on the
rheological properties of the primary ore.
Master composites of primary, shear and oxide ores were
formed. While not all tests were conducted on all composites,
the primary, shear and oxide master composites were used
to assess optimal leaching conditions and determine process
performance.
The analysis of the sinks showed that the sulfides,
identified in the head assays, were present as pyrite. Other
major minerals present in the sinks included iron oxides/
hydroxides, and iron silicates. These results give basis to the
assumption that the unleached gold remaining in the leach
residue may be locked in pyrite.
Gravity – cyanidation leach optimisation
The QEMSCAN results showed no major cause for concern
from a processing point of view.
A series of tests were conducted to determine the optimal
grind size and cyanide dosage for the leaching process. It
was determined, after the first round of testing, that a 24 hour
leach time would be adopted due to negligible improvements
in leach recovery being achieved from longer leach times.
Cyanide was very low and lime consumption as expected.
The test work program, conducted on the oxide, primary and
shear master composites, determined that a P80 of 63 microns
would be used for subsequent testing and throughout the
process plant, with a cyanide dosage of 250 ppm. Under these
conditions all samples displayed fast leaching kinetics and
high recoveries.
We are metallurgists, not magicians
Oxygen uptake testing
The oxygen consumption ranged between maximum values
of -0.024 mg/L/min for the oxide and -0.035 mg/L/min as
a maximum value seen for both the shear and primary ore
types. As a result air sparging will suffice.
Carbon loading testing
Carbon loading tests were conducted on the primary and
shear master composites, with the slurry from a bulk leach
on gravity tailings being used to generate a representative
carbon loading feed.
103
K Nilsson and D Connelly
The primary material had a k value of 177 h-1, and an n value
of 0.74.
Table 5
Outotec thickener test results summary.
The k and n values for the shear test were 202 h-1 and 0.70
respectively, and therefore indicate no issues for processing.
Parameter
Rheology
Rheology testing was conducted on all three master
composites. Based on the expected design of the process
plant, the tests were conducted at a P80 of 63 micron, at 40, 50
and 60 per cent solids by weight. A range of shear rates was
applied to the slurries, with shear rates of 5 s-1 and 200 s-1 being
taken as low shear and high shear conditions respectively for
the analysis.
Based on the current process plant design, with correct
thickener control it is not expected that any viscosity issues
will be seen through the process plant.
Thickening
The tests were conducted on RC chip samples (SRC730 and
SRC711) provided by Troy Resources, with three flocculants
being tested. It was hence recommended to proceed with the
Magnafloc 351 as the preferred flocculant in the design work
(see Table 4).
Outotec thickening test work
Thickener design test work was conducted by Outotec in
two stages; initially flocculant screening was conducted on
five flocculants including the previously tested Magnafloc
351. From this testing it was determined that Magnafloc 155
would be used for the dynamic settling tests.
The test work results showed that the optimal feed density
for all ore types tested was 15 per cent solids. Other parameters
from the test work are summarised in Table 5.
Based on the results of this work, there should be no issues
in thickening the Karouni ore or cyclone overflow.
Bulk leach and detoxification
In order to conduct the detoxification test work, a 10 kg amount
of sample from the primary and shear master composites was
used as feed slurry.
The primary material showed strong performance at all
ratios, with the cyanide level, both free and WAD cyanide,
being below 1 ppm after two hours. The shear material was
unable to achieve free or WAD cyanide levels below 1 ppm
after eight hours at the lowest ratio, however at ratios of 3
and 4 both the free and WAD cyanide was seen to be below
1 ppm. Cyanide destruction testing was based on the INCO
sulfur dioxide system.
VARIABILITY TESTING
Variability testing was conducted on five primary samples
only, due to the amount of available mass for this ore zone,
and the limited mass of the other ore zones (see Table 6).
Ore Type
Oxide
Shear
Primary
Feed density (% w/w)
15
15
15
Flocculant dosage (g/t)
30
30
20
U/F density range (% w/w)
47.4–53.2
49.0–63.4
55.9–63.5
Overflow clarity (mg/l)
~100
~100
~350
The cyanide consumption was low for all samples tested,
while the lime consumption was at levels below concern.
PROCESS PLANT DESIGN
The process plant design is based on two-stage crushing to
a covered crushed ore stockpile with a nominal 3800 t live.
The primary jaw crusher is a Metso C120 and a Nordberg HP
500 secondary crusher. This is based on operating 6500 hr/a.
The product will be a P80 of 10.3 mm from a double deck
screen. Crushing was selected based on low wear rates,
energy efficiency and process certainty compared to a SAG
mill.
A second-hand ball mill 5 m diameter by 7.32 m long was
sourced out of Canada using a 40 per cent ball charge with
a 3.1 MW motor. The mill is close circuited by cyclones with
the cyclone overflow going to a 25 m preleach thickener.
The thickener underflow is pumped to the first CIL tank.
A preleach thickener was included to allow flexibility of
cyclone overflow and grind size, increased residence time in
the CIL and reduced cyanide and lime consumption. This also
assists in clean-up of spills inside the bunded areas of a high
rainfall area.
Two Knelson XC30’s gravity concentrators are used to
recover coarse gold before being processed using intense
leach reactors. The high gravity gold recovery made this
essential for higher overall recovery, faster cash flow and
improved gold security.
Cyanide leaching is using six CIL tanks. The nominal leach
residence time is 24 hours. The loaded carbon is stripped in a
pressure Zadra stripping column.
The gold is electrowon and smelted in the gold room. The
carbon is regenerated after stripping on an as-needed basis.
DETOX
The sensitive environmental area required a DETOX to
minimise the potential impact on the environment.
The leached slurry in CIL tank 6 will pass from the final
launder into the DETOX circuit, at DETOX tank 1 (300-TK008) for cyanide destruction.
The DETOX circuit contains three agitated tanks, (DETOX
tank 1, 2 and 3 – 300-TK-008/009/010) and will be bunded
Table 4
Reverse circulation chip settling results.
Property
104
SRC 730
SRC 711
Magnafloc 351
Magnafloc 345
Magnafloc 351
Magnafloc 345
Free settling velocity m/h
25.50
27.06
25.88
27.85
Hindered settling velocity m/h
1.14
0.78
0.72
0.60
Underflow % solids w/w
55.3%
51.6%
56.2%
44.9%
We are metallurgists, not magicians
Karouni Gold Project from drill core to commissioning
Table 6
Variability leach results summary.
Sample
Head grade
(Au g/t)
Gold recovery (%)
Leach time (h)
0
2
4
8
24
NaCN consumption
(kg/t)
Lime consumption
(kg/t)
P-V1
2.59
58.4
76.1
77.9
79.5
80.4
0.09
1.25
P-V2
11.12
42.6
78.8
87.3
92.4
92.8
0.12
1.15
P-V3
3.94
43.5
70.0
75.5
76.4
79.2
0.08
0.91
P-V4
7.80
79.3
88.4
92.2
95.6
97.4
0.10
0.83
P-V5
4.58
58.2
78.5
84.8
87.9
95.2
0.07
1.03
separately to the CIL tanks and operate in series. This will
provide three stages of DETOX to ensure that the cyanide is
completely removed through the 7.5 hours total residence
time. As with the CIL circuit the DETOX tanks will be able
to have a tank bypassed as required for maintenance. The
DETOX tanks will all be dosed with sodium metabisulfite
(SMBS), while DETOX tank 1 will also receive a feed of
copper sulfate (CuSO4) to maintain a minimum 50 ppm Cu2+
as part of the detoxification process. Spent hydrochloric acid
(HCl) and acid wash water from the elution circuit will also
be pumped into DETOX tank 3. The DETOX tanks will be
highly agitated and be air sparged to drive the reaction which
destroys the cyanide. The slurry will gravitate between the
tanks in this circuit. Controls are in place to alarm and even
shut the plant down if cyanide is detected in tanks 2 and 3
above a preset level.
Tailings
A conventional tailings dam will be constructed on-site. Due
to the high rainfall run-off water collected around the process
plant and TSF, or from mine dewatering activities, will be
handled in stilling ponds prior to discharge into the river
system. This will ensure the water released is treated and
hence suitable for release.
The tailings are not acid forming (benign sulfides) and this
is a positive for the project.
Water management
Water will be fed to both the raw water dam (700-DM-001) and
the water purification feed tank (700-TK-001) from the river or
bore field. The raw water dam will provide water to be used as
both feed throughout certain areas of the process plant (elution
circuit, gravity fluidisation and reagents), and for top up water
for the process water dam (700-DM-002). Mine dewatering
water will be used preferentially for processing and the water
balance indicates excess water will need to go to stilling ponds,
tested and then disposed of into the environment.
Reagents
The reagents and services area of the process plant will be used
for both reagent preparation and storage prior to distribution
throughout the process plant.
The reagents will be mixed in a bunded area in covered
tanks to prevent heavy rain damaging or causing losses to
the reagents. Bunding will be placed around key areas to
also provide additional levels of separation between reactive
reagents.
Air
Air is used throughout the process plant. This includes air
for the workshop, sparged into the leach circuit to improve
leaching kinetics, and instrument air. Air will also be used in
the Detox circuit. This will be provided by blowers installed
in this area of the process plant. Air will be supplied from
one of two air compressors (800-CP-001/800-CP-002). This air
will be stored in the air receiver (800-RA-001) until required
throughout the process plant.
Diesel
Diesel will be used on-site as the primary source of fuel
and heating. Diesel will be used for the power station, the
smelting furnace, elution heater and the carbon regeneration
kiln, as well as fuel for site vehicles. Power supply is based on
packaged sets using diesel fuel. This will supply power to the
camp, process plant and facilities. Troy Resources will own
and operate the power station.
CONCLUSIONS
Small resource gold projects can be profitable and the risks
can be managed using suitable second-hand equipment. The
concept of minimal engineering, that is, project management
of a group of subconsultants to execute the project did result
in substantial savings in time and money. The use of a secondhand ball mill (never used) saved time and money. The simple
flow sheet has low technical risk and can be readily expanded
should the need arise. The DETOX circuit was critical given
past problems in Guyana with cyanide spills. Troy Resources
has gained valuable experience in project development in a
new international area and has established a presence which
will allow the evaluation of other resource opportunities in
the immediate area of the plant. The project impact will be
very positive for Troy and the local community. Exploration
in the region is likely to find new resources within proximity
to the process plant and the operations future may continue
beyond the present resources.
ACKNOWLEDGEMENTS
The author would like to thank colleagues and persons
associated with the Karouni Gold project for the education
and life experience received during this exciting period of
the development, birth, nurturing and growth of a new
gold project in Guyana, South America. Thanks are also
due to the management of Troy Resources for permission to
publish this paper.
Cyanide, hydrochloric acid, sodium hydroxide, flocculant,
copper sulfate, sodium metabisulfite and antiscalant will be
mixed on-site for delivery and use within the plant.
We are metallurgists, not magicians
105
Contents
Upgrades, modernisations, automation and
expansions … where will the expertise,
capability and skills come from in the future?
R Coleman1, J King2 and T Hunter3
ABSTRACT
The industry that supports operations and new projects is suffering a rapid decline
in the current severely curtailed capital investment environment. This is evident by a
decline in the number of skilled teams, who are now centred in fewer locations. Other
features are an aging and retiring ‘baby boomer’ generation, coupled with a low
uptake of new graduates and the uncertainty in predicting future capital expenditure
programs and requirements. Combine this with a reluctance of cash-poor operations
to modernise and adapt in an environment of low metal prices, and there is a real
uncertainty as to how service providers will meet industry needs as they evolve.
Trends amongst service providers will be put forward as well as developing options
and alternatives. These embrace a whole gamut of agile alternatives in the areas of
partnering, alliances, commercial and shared risk models and mutual innovations.
These will be discussed with some case studies and suggested approaches.
Particular emphasis will be upon brownfield expansions and upgrades which have
emerged recently as a preferred lower capital and focused way to achieve the best
practical outcome at lowest cost. There are several specific features of such brownfields
projects including the requirement for very close cooperation between the execution
team, operators, maintenance and suppliers and a maximum usage of site resources
and ingenuity. Often, required investment hurdle rates are high, downtime allowed
is minimal and site resources usage is maximised. This requires mutual trust, good
planning and high health, safety, environmental and community (HSEC) standards
and practice. In many ways, the practices and skills needed are the exact opposite of
the historical ‘big Greenfields’ project with which the industry has been geared for
more recently.
Another developing area has been with innovative commercial models and
practices. This is a direct result of the more challenging investment criteria being
adopted and the relative hunger of service providers. It has become more common
to look at these commercial-costing alternatives very early in the project cycle as an
essential project tool and not just as a possible option. These commercial solutions
cover such alternatives as:
•• leasing and rental
•• deferred payments
•• build–own–operate–transfer and build–own–operate
•• whole-of-mine preferred spares and maintenance terms
•• aggressive performance targets with real assurance penalties.
Once again this requires very close mutual dealings and trust, and sufficient time
allocation compared to a conventional approach.
INTRODUCTION
1. MAusIMM, Account Director, South East Asia
Pacific, Outotec, West End Qld 4101.
Email: rob.coleman@outotec.com
2. Head of Sales Minerals Processing, Outotec,
Frenchs Forest NSW 2086.
Email: jason.king@outotec.com
3. FAusIMM(CP), Senior Consultant, Clayfield Qld
4011. Email: tomhunter134@gmail.com
The broader Mining Equipment Technology and Services (METS) industry has
grown since the early 2000s in response to the demand from the industry and is a
significant employer and economic contributor in its own right. This has recently been
documented and recognised particularly by governments (Hunter, 2014; Austmine,
2013). The METS subsector dealing with metallurgical technology development,
process plant upgrades and new projects, had grown at a real pace and likely peaked
in 2012. It is now in serious decline and losing capacity, capability and confidence.
There are several reasons for the decline, which include:
•• the well-documented and recognised step change in metal prices since 2013
from previous buoyant levels (Fiscor, 2015)
107
R Coleman, J King and T Hunter
•• the ‘hangover’ psychology from the project excesses, in
terms of bloated capital cost compared to expectations,
delayed deliveries and fierce project competition for
scarce engineering and construction resources
•• the generational change occurring as the ‘Baby Boomers’
move on from active project involvement
•• the severe contraction in the skill level and number of
teams now available for projects (Federation University
Australia, 2014).
So the challenge is how to provide the level of service
demanded by our industry in a sustainable and competent
manner. This paper looks at the challenges, some potential
solutions, illustrated by some recent project examples,
particularly in the ‘new’ environment since 2012.
BACKGROUND
There is a wide range of options available to operations and
projects wanting to spend capital (Luxford, 2006; Hunter and
Broome, 2008; Revy, 2008; Lane and Clements, 2012). Looking
through them, one can identify the following six points:
1. The classic ‘in-house’ owner’s team with several ongoing
projects, who know the owner requirements well and
have developed systems and techniques attuned to the
site.
2. The ‘hybrid owner’ model where a trusted outside
small group coordinates the project. This has many of
the advantages of the classic ‘in-house’ but can be more
flexible in handling a variable project portfolio.
3. The ‘owner project team’ where a specific team is
recruited for the project. Typically, this works well
for a substantial complete project cycle of some years,
for example prefeasibility study (PFS) to definitive
feasibility study (DFS) (Hunter et al, 2009), to execution
and commissioning.
4. The traditional ‘engineering, procurement, construction
and management’ (EPCM) approach where the owner’s
team engages a substantial study-execution team
on a reimbursable basis (Loots and Henchie, 2007).
The EPCM acts as the project’s ‘arms and legs’ under
the direction of the small owner’s team. The owner is
ultimately responsible for the project. The ‘long-term
reimbursable alliance model’ is a variation of this, where
the EPCM imbeds a long-term core team into the owner
structure and uses that to generate project deliverables
and services.
5. The lump sum, engineering, procurement and
construction (EPC) or design and construct (D&C)
approach where the owner’s team carries out substantial
definition and invites lump sum prices for a well-defined
scope-of-work with the intention of having minimal
variations upon execution. This involves potential risk
for both parties because of the tight contract definition,
transfer of risk and the issue of variations (if and when
they occur).
6. The technology driven project where the owner’s desire
to involve a new, improved, or even revolutionary
technology provider or partner in the project (Wasmund
et al, 2011). The driver here is the technology provider
and the execution model is ‘adapted’ to meet the needs
of the researcher/technology owner or developer as the
owner deems the greater execution risk is outweighed
by the perceived advantages of the technology.
The chosen model will be governed by:
•• company policy
108
•• practice and history
•• options available in the market place
•• degree of technical innovation
•• competitive climate
•• owner preferences.
The above list is not exhaustive and often combinations and
hybrids are adopted to fit. In the current climate of minimal
capital expenditure, fierce competition and an industry
demand for predictability and economy there has been a strong
trend towards lump sum and well-defined scopes rather
than the more flexible reimbursable models. This approach,
however, does demand larger and more skilled owner teams
and may also involve greater preparation time and earlier
project expenditure on early contractor involvement (ECI).
A key (and often overlooked) consideration for a project
typically going through the phases of studies, early
involvement and approvals, engineering and execution, is that
of trust and predictability between the parties. This is much
treasured and encapsulated in the salutation ‘no surprises’.
Constant multilevel open communication, sharing openly the
inevitable project challenges, and accurate forecasting, are
intrinsic to achieving this aim.
In summary, there are many project models and a large
number and variety of providers who are hungry and keen
for work.
DRIVERS AND TRENDS
Drivers in the new environment
The industry itself and the associated METS area are acutely
aware of how the industry has changed since 2012. The
previous extraordinary emphasis on capital expenditure (and
poor discipline in operating costs) in the drive to maximise
production and capacity in the buoyant price environment has
been very rapidly replaced by a completely new discipline and
ethos. This is where capital expenditure is very constrained
and mainly focused on essential HSEC drivers and often
only where there can be a very rapid return on expenditure.
‘Brownfields’ has become the main area of focus with the aim
of minimising operating costs with particular emphasis on
‘debottlenecking’ and ‘incremental efficiencies’. Many areas
of previous outsourcing and use of contractors have now been
returned to owner involvement as this is perceived (often
correctly) as adding value and promoting increased owner
awareness of operations and options. Particular drivers that
can be identified include:
•• a complete change in focus from production and tonnage
maximisation to capital and operating cost discipline
(Clarke, 2015)
•• an acute bottom down driven capital expenditure
discipline and budget
•• a fundamental review of operating practices and
expenditure on operating costs
•• changed philosophies towards outsourcing and service
provider models
•• changed perceptions by financiers and investors on
the sector’s prospects, relative attractiveness, future
profitability and drivers; this has totally changed the
environment for the industry at all levels
•• the owner/operator view that the extraordinary
demands up to 2012 for services had resulted in
bloated costs (and charges) from the METS sector and
we are metallurgists, not magicians
Upgrades, modernisations, automation and expansions … where will the expertise, capability and skills come from in the future?
poor performance in terms of productivity, relative
international costs and project efficiency
•• competition for resources (particularly in engineering,
project management and construction resourcing) was
a particular factor with coincident buoyant commodity
pricing in energy and the ‘once in a generation’
establishment of the huge regional liquefied natural
gas (LNG) projects in Western Australia, Queensland,
Northern Territory and Papua New Guinea
•• commercial discipline was viewed by many owner/
operator organisations as having been compromised
during the boom and needed to be ‘normalised’
•• many services organisations had expanded laterally,
usually in response to client demand, into extra services
outside their traditional core business.
This was therefore a new paradigm for many in the
industry. There were memories of years like 1987, 1992–1993
and 2001–2002 where downturns had occurred, but the speed,
depth and completeness of the 2013–2016 downturn has been
viewed as almost unique probably due to rapid industry
growth in the decade to 2012.
Recent trends
This paper has not tried to identify specific trends, which
can vary by company size, history and specific practices,
geography and commodity, but have chosen those that in
our view are the most germane and relevant to the Australian
industry in 2016. Obviously in the current environment, we
are faced with issues, such as high industry unemployment
and underemployment particularly at graduate level, and
also in geosciences, engineering and construction, as well as
much reduced overall spend on services by the industry.
Relevant trends include:
•• a much reduced new project study and execution
capacity and capability
•• capability and project teams becoming concentrated into
fewer regional and global centres (for example, Perth,
Brisbane,
Toronto,
Vancouver,
Santiago
and
Johannesburg), and many former ‘powerhouses’ having
little real capability now (for example, Melbourne)
•• the increasing usage (particularly with demonstrated
and consistent performance) of ‘best value engineering
centres’, such as those in Mumbai, Delhi, Dubai,
Shanghai, Kolkata and Manila. It does take time and real
corporate commitment to incorporate this potential into
reality and to achieve consistent performance
•• in-sourcing of many aspects for projects and execution,
and the re-emergence of some internal engineering
teams within organisations
•• a much more rigorous contractual, commercial and cost
environment
•• a trend back to longer term alliancing often with
innovative and open commercial arrangements
•• a very open dialogue between owners and potential or
incumbent service providers
•• opportunities for small and single-person consultants
and contractors, particularly for studies, audits and
upgrades
•• realisation by owners that significant technical and
productivity gains can only be achieved by longterm alliances and joint developments with ‘original
equipment manufacturers’ (OEMs), research and
technology developers and leading edge service
we are metallurgists, not magicians
providers. These arrangements tend to be more specific
and early results driven compared to those of the early
2000s which were more industry wide and often driven
by the research or developer partner
•• a much higher degree of accountability in providing
services or solutions is being demanded
•• the highly competitive new environment has
concentrated the industry to fewer large providers and
several new smaller entrants. The more junior sector
for instance, looks for early study engagement with
commitment to their project and financing demanding
a lump sum or EPC approach. This requires a nimble,
client centric and innovative execution model well
removed from the reimbursable EPCM model of the
early 2000s.
WANTS, NEEDS AND ISSUES
We have previously discussed the industry trends and
drivers, and will now focus on the wants, needs and issues of
the models presented earlier.
Classic ‘in-house’ owners team
Generally, a larger organisation with multiple sites and a
history/confidence in client knowledge/alignment and
projects delivery coupled with well-developed systems. This
requires:
•• competence and alignment
•• commercial/contract agreement
•• a predictable costing model justifiable to management
•• ability to align with corporate HSEC imperatives
•• comprehensive systems aligned to the owners
•• adherence to project milestone expectations/
commitments.
Hybrid owner model
Here a specific outside team is recruited to act as the owner,
usually on a single project or program. What is required
includes:
•• a high degree of comfort and trust in the service
providers
•• a willingness to align and provide/source the areas that
the owner team are unable to provide
•• a staged and flexible commercial arrangement with a
high degree of openness/transparency
•• ability/willingness/appetite for varying execution
models and partners (note that systems alignment can
be an issue as can process coordination, so specialist
help in these areas may be needed. Often metallurgical
process expertise is provided by an outside expert who
may have historical owner links.)
Owner project team
The owner project team may be used for a longer term or
larger project of a cycle of some years. This is particularly so
for a larger EPCM project where the owner team needs the
authority and expertise and risk assumption inherent is such
arrangements. Special owner challenges exist for this case,
which is basically a combination or version of the two already
considered, and includes:
•• the use of systems and procedures that fit both the
parent owner and the contracted EPCM
•• retention and alignment of the team
109
R Coleman, J King and T Hunter
•• decisions as to the degree of intrusion/checking into the
EPCM and the degree of alignment and trust with the
EPCM (the ‘man matching’ syndrome)
•• project manager alignment, retention and drivers can be
an issue.
There are many contractors who are very comfortable to
work on the owner side in this team environment and it fits
well in career development. One challenge is that each brings
a unique experience, preferences and skill set and that team
alignment particularly on systems use can be a challenge.
Traditional engineering, procurement, construction and management
The EPCM approach is where the owner’s team essentially
uses an ‘agent’ to perform most of the tasks under the
direction of the owner team. This has worked well in the
past but perceived shortcomings were evident to many in
the period of 2002 to 2014 (Walker, 2015; Moore, 2015). These
issues involved:
•• a shortage of skilled project personnel for both
EPCM’s and owners led to issues with team retention,
incentivisation, alignment and performance
•• intense competition for people driving up rates and
affecting productivity and quality
•• excessive ‘project churn’
exacerbated these issues
of
people
(turnover)
•• the traditional clear delineation of owner and contractor
roles, functions and responsibilities became a real issue.
It should be remembered that EPCM had evolved in the
industry to provide a solution to the need to efficiently exercise
projects in a timely manner with maximum owner control
and accountability. When properly done by competent teams
in a ‘normal’ market it is still an efficient and well understood
project mechanism.
A recent trend has been to use the EPCM model and
expertise for longer term ‘alliancing’ where the client selects
one or more alliance partners (usually from traditional EPCM
practitioners) to work on a longer term embedded basis in the
owner team. The intention is to have a dedicated team that is
able to cover internal project generation, definition, approvals,
execution, commissioning and alignment. This has long been
the norm in the oil-gas-petrochemical area and combines
the advantages of a competitive contracting environment
with dedicated owner teams who build-up client and project
familiarity and expertise, and project execution efficiency.
Lump sum engineering, procurement and construction
The lump sum, EPC or D&C approach involves slight
variations on a simple execution method namely that the
project is defined by the owner who invites bids to do the
complete job. The issue for the owner is in defining the scope
of the project in sufficient detail to enable a contractor to cost
and price their bid. Likewise, such an effort by the bidders is
very expensive in terms of dollars, commitment of experts,
the level of detail and clarifications needed, and in time taken.
It would not be unusual for a mid-size concentrator (if defined
to a 30 per cent engineering level) to take six to nine months to
arrive at a bidding cost commitment of some millions.
The issues here are:
•• the level and certainty of pre-work by the owner’s team
to reach the stage where it can be bid
•• the discipline required by the owner’s team in resisting
change, documenting and affirming variations and
using the systems needed for project tracking
110
•• for the contractor, the commitment to bid will only
be made with a good certainty of project actually
happening, a high probability of bidding success and
a level of comfort as to the owner’s competence and
financial strength; often a contractor will strive to have
an ECI status where they are paid (partially or wholly)
to invest the resources to give a bid
•• a lump sum bid by a substantial competent contractor is
very important to project participants such as banks and
other lenders/financiers.
Some owners believe the risk is lowered once this approach
is adopted and that all risk is now with the contractor. This is
certainly not the case as a good owner’s team needs to check
quality and progress, ensure what is being built is consistent
with owner expectations and the contractors obligations, and
to track and document inevitable variations.
Technology-driven project
All the above approaches have assumed conventional,
well-proven technology and process and would be greatly
complicated by new, first-of-a-kind or largest-yet-built
technology. The main driver in such projects is the special
needs of the process or technology provider and the degree
to which the execution strategy and practice needs to be
adapted. These particular technology drivers may mean that
the OEM or technology provider (TP) may adopt any of the
above roles in addition to their speciality one and may have a
great influence on the project model adopted and the project
participants. Areas of particular importance are:
•• the relationship, confidence and trust between owners
and OEM/TP
•• the level of development and maturity of the technology
– there is a vast difference between a larger sized piece
of conventional equipment and a completely new
multiphased, integrated hybrid process only proven at
demonstration scale; for example, mineral processing
with
hydrometallurgical
and
pyrometallurgical
downstream steps (Mezei, Todd and Molnar, 2006).
•• clear definitions of roles and responsibilities need to
be agreed and formalised with minimal number of
interfaces
•• the program needing to recognise the complexities
by having good ‘float’ and a realistic commissioning/
ramp-up schedule
•• multiparty involvement
imperatives
and
complex
reporting
•• each participant needing a technology ‘champion’
who believes that the project/process complexities are
outweighed by the advantages and potential.
COMMERCIAL CONSIDERATIONS
The typical execution models and study/practitioner issues
have been examined but often of overriding importance will
be the broader and detailed commercial matters. For the asset
owner, these include:
•• Risk profile and factors – a formal risk register needs to be
completed early and agreed internally by stakeholders. It
is often useful to do two; one from the owner viewpoint
and the other from that of potential contractors. This
helps in alignment later and identifying/quantifying
areas of potential dispute. A fundamental element of
risk mitigation on a project is that the particular risk
should always be allocated to the party best able (or
willing) to manage that risk. If a party is uncomfortable
we are metallurgists, not magicians
Upgrades, modernisations, automation and expansions … where will the expertise, capability and skills come from in the future?
or unfamiliar with the particular risk it will be the subject
of excessive provisioning and uncertainty.
•• Is the potential contractor capable of completing the
task through to a satisfactory outcome? Factors such
as: history, capability, capacity, team competence,
contractor internal execution model, subcontractors,
reference projects, reliance on variations for commercial
means should all be considered. This is often referred to
as ‘the smell test’.
•• The importance of project staging, particularly with
a large venture or one involving greenfields or new
technology features. This could involve multiple stages
from initial study/PFS/test work/pilot plant through
DFS, engineering, execution and commissioning. Is the
contractor capable of these?
•• Is the owner potentially committed for some years of
working productively with that contractor?
•• Should an execution model be adopted which plans for
use of multiple contractors at different stages whilst
ensuring project knowledge is not lost?
•• The execution model commercial/contract environment
will be fundamentally determined by whether it is
reimbursable or lump sum. For a lump sum, very
careful and detailed definition of scope is required and
intrinsically the contractor will build in a risk premium
and the cost of detailing variations. Conventionally this
is estimated at about 15 per cent above that estimated
with a reimbursable contract where the owner has a
much higher risk commitment but also greater flexibility
in changing scope.
•• The contract set-up needs to reflect the fundamentals
of how the job is being done and be consistent with
the practicalities of the parties’ performance. If it is too
one-sided, it will engender undesirable outcomes as the
lesser party protects its interests eg excessive liability
penalties will mean a behaviour focused on avoiding
those outcomes and this may be inconsistent with the
best overall project result.
There are many studies into project performance and
the factors behind underperformance or failure in the oilgas-petrochemicals area and more latterly in the minerals
industry. The major conclusion for successful performance
was the degree of front-end loading (FEL) on a project; the
more resources, effort, project team involvement and early
engagement of the contractor(s), the better. This would be
backed up by rigorous benchmarking and stage gate approval
processes.
Such systems have been adopted more widely in the
minerals industry over the last ten years, with some notable
successes. The system does, however, add earlier time and
costs to a project, but this is ultimately vastly cheaper than
making the mistakes later in execution. It does require a high
degree of owner discipline and commitment, systems and
project team talent and input.
EXAMPLES
Some indicative examples of recent projects will now be
provided. We acknowledge that there are many competing
models, companies and service providers in our industry who
energetically compete and this is very healthy. The project
developer faces the task of deciding how best to do the project
and with whom.
As a global supplier of equipment, industry solutions and
services, and a developer of new technology and techniques,
we are metallurgists, not magicians
Outotec itself adapts with its customers and for specific
applications the preferred execution models, partners and
commercial model. There is no ideal single solution. We
acknowledge that there are very competent competitors and
alternatives, ensuring owners have good choice with their
projects.
The opinions expressed on these selected examples are
those of the authors and are intended to generate ideas on
project improvement and how the optimal execution strategy
can be selected.
Alliances
‘Alliances’ are very case- company- and site-specific and can
cover the whole gamut of studies, engineering, approvals
and review, construction and execution, and commissioning,
upgrades, modifications, debottlenecking and operations/
maintenance. There are many advantages for these
arrangements particularly for larger, multisite operations
such as Tronox, Alcoa, South 32, Rio Tinto and BHP. Owner
management of the Alliance is best done with a specific
dedicated internal resource and an appropriate sponsorship
and hierarchy. The ability to Alliance with global miners is seen
as a key business imperative with many service providers who
also emphasize the need to align with the customer’s systems,
procedures, standards, reporting systems, approval hierarchy
and build-up ‘corporate and plant knowledge’ amongst the
seconded employees. A key issue with the seconded team can
be that they, over time, identify more strongly with the owner
than with their parent service organisation.
Jacobs is a good example of the effective use of Alliances.
Jacobs have 30 years of experience in creating value within
the framework of alliance relationships which has led to
the development of their alliance approach (Jacobs, 2013).
Their approach is based on the understanding that there is
no ‘one size fits all’ formula to forming a successful alliance/
partnership. Each one has to be tailored to the objectives and
goals as well as the needs of the client. Jacobs have developed
a range of models of Alliances. They all have somewhat
different characteristics but all have the common goals of
continuous improvement and bringing best practices and
lessons learned to their client’s business. More than 60 per cent
of their revenue is currently generated through Alliance-type
relationships. Therefore, forming and maintaining healthy
alliances that continually deliver value to their clients is key
to their success.
Owner project team
Masan Resources (MR) in Vietnam has developed the Nui
Phao polymetallic project some 70 km NW of Hanoi and this
is a large modern project producing a large range of tungsten
products, chemical grade fluorspar, bismuth metal and
copper-gold as a concentrate (Masan Resources, 2016). The
process plant is specific and complex involving many process
steps and unique metallurgically (Morgan, 2016).
Establishment of the project emphasized local capability
with western (Australia-India) engineering, along with
procurement and construction by the owner. For the rampup/debottlenecking/process optimisation phase, MR
chose to use a model of a strong plant site metallurgical
team, supported by a specialist engineering owner’s team
in Australia. Emphasis was put on partnering with OEMs
/service providers and maximising the use of in-country
construction capability. Projects are fast-tracked, fit-forpurpose and aggressively managed to align operations,
plant improvement, OEMs and technology suppliers and
the in-house team. This has enabled stepwise and regular
111
R Coleman, J King and T Hunter
improvements in plant performance at appropriate cost and
tight capital control. A similar approach has worked very well
recently at the Didipio project in Luzon, Philippines, operated
by OceanaGold (Walker, 2013).
Hybrid model
Carmen Copper Corporation (CCC) is a copper concentrator
situated on Cebu Island in the Philippines. The concentrator
was originally commissioned in 1955, and operated until 1994,
when low copper prices forced the mine to cease production.
The mine was rehabilitated in 2007 and brought back into
production at a rate of 40 000 t/d, as outlined by Morgan et al
(2014). The decision was then made to increase the production
from 40 000 t/d to 60 000 t/d. The decision was also made
to utilise existing installed equipment as far as possible to
reduce costs.
The client wanted a world-class project, with modern
process controls, utilising in-house capacity for construction,
and low cost detailed engineering. This resulted in a hybrid
model being utilised for the project:
•• metallurgical test work program designed and managed
by CCC
•• Outotec contracted to carry out the detailed design,
including a 3D model of the plant
•• a local Philippine company contracted by CCC to do the
detailed engineering
•• Outotec supplied and installed all the proprietary
equipment
•• CCC procured all non-Outotec proprietary equipment
•• construction carried out and managed by CCC
•• Outotec provided expert advice including construction
where requested
•• CCC managed the commissioning supported by vendors
and expert advice from Outotec when required.
CCC selected the above contractual model to fast track
the project compared to the standard EPC/EPCM model.
This enabled early works to be carried out facilitated by
utilising existing labour structures, circumventing the usual
mobilisation phase of traditional projects. The project was
designed, constructed and commissioned in 20 months
from receipt of a letter of intent. This was an excellent result
considering that the base starting document was a PFS,
compared to the standard practice of first doing a DFS as a
minimum.
Lump sum early contractor involvement
Alkane Resources Dubbo Zirconia Project (DZP) has one of
the world’s largest in-ground resources of rare metals and
rare earths (Alkane Resources, 2016). The mine is expected to
process 1 Mt/a over a period of 70 years or more. The DZP
will produce zirconium, hafnium, niobium, yttrium and rare
earth elements. This complex metallurgical plant flow sheet
fits in very neatly with the suite of process technology that
Outotec has developed over many years.
Outotec was engaged on an early contractor involvement
(ECI) basis for the project. The set-up is as follows:
•• initial value engineering phase – reviewing the front-end
engineering design (FEED) study carried out by others,
identifying gaps and opportunities for cost savings, and
process enhancements
•• technical input into the pilot plant test work program
ensuring that the process is fully tested, and all market
and vendor sample needs addressed
112
•• sufficient basic engineering on an open book basis to
derive an EPC cost for the project
•• fixing a mutually agreed EPC price
•• EPC project execution.
This represents a very flexible contracting model aimed at
reducing the risks for both parties, and therefore reducing the
risk premium associated with a traditional EPC contractual
model. This approach facilitates the client operating with a
small owner’s team, in line with an EPC strategy.
The initial design phase to produce a fixed EPC price is
based on a collaborative open book basis ensuring that Alkane
has sufficient input into the plant design and the costs, before
signing the EPC contract. This staged approach is designed
to reduce the risks of both contractor and client, utilising
the prior client knowledge in the design, then agreeing on a
mutually acceptable EPC price.
Build–own–operate and build–own–operate–transfer models
Outotec has recently delivered a total system solution for a
client in Australia to meet their underground mine backfill
and operational needs (Suvio et al, 2016). The contract
included an extensive backfill test work program, backfill
plant and underground piping EPC, system commissioning,
complete mine backfill system operation and maintenance.
This represents a classic build–own–operate (BOO) contract.
This contract contains a higher level of backfill engineering
and expertise and provides the customer with unique
expertise in state-of-the-art mine backfill engineering and
plant operation. The major customer advantage is a lower
cost operation, operated entirely by Outotec and providing
the customer with lower risk to their mining operations
throughout the life-of-mine.
This type of contract represents the ‘Ultimate Process
Guarantee’ as the contractor carries the full financial,
performance, operating and maintenance risk providing an
over the fence service to the client.
THE FUTURE
Looking to the future, we need to make some assumptions
before venturing into this crystal-ball space. We have
assumed a period of three to four years, with the industry
capital activity relatively restrained as it is at present. Whilst
this is hopefully a pessimistic view, we believe it is prudent
and consistent with many of the industry commentators and
the predictions made below:
•• more contractual, commercial disputes than hitherto
•• the EPC model will continue to be more attractive than
the reimbursable EPCM method
•• alliancing in all its forms will be a growing trend
•• dedicated owner teams for projects will be more widely
recognised as vital in terms of project success, necessity
and continuity
•• there will be very few ‘new technology’ projects and
plants as a result of owner conservatism and the muchdiminished recent capacity of in-house and industry
collaborative research, development initiatives
•• process plants will become more complex and often
incorporate ‘hybrid processes’ eg mineral processing
with hydrometallurgy
•• innovation will be directed to ‘peripheral’ areas such
as water recycling/treatment, energy usage, advanced
process control, remote operation and control, and
we are metallurgists, not magicians
Upgrades, modernisations, automation and expansions … where will the expertise, capability and skills come from in the future?
tailings management, as areas to give the greatest
predictable advantages.
Alternatively, if there is an appreciable recovery in metals
prices and market demand, we would expect that many of the
recent excesses and problems that emerged in 2008 to 2013 will
be repeated or even exaggerated. The capacity of the industry
for projects, from a skills and coordination viewpoint, will be
much reduced due to recent events. One could foresee a need
for strong self-discipline and clear thinking to avoid past issues.
ACKNOWLEDGEMENTS
The authors thank several Outotec and industry practitioners
for their input and comments, with special thanks to
Alan Dennis. The support of Outotec’s management is
acknowledged. As the authors collectively come from
backgrounds covering operations, research and development,
project development, equipment and solutions supply,
services delivery and studies, we have tried to use that
experience to give a balanced and wide view. A keynote
opinion piece does involve judgement and personal opinions
and is intended to engender debate and discussion. As such,
it is acknowledged that there will be a necessary degree of
interpretation and personal opinion. The authors are happy
to discuss this aspect and are open to further elaboration or
modification of their stated positions in this paper.
REFERENCES
Alkane Resources, 2016. Website homepage [online]. Available from:
<http://www.alkane.com.au/index.php/projects/currentprojects/dubbo/project-overview> [Accessed: 14 April 2016].
Austmine, 2013. METS National Survey [online], in Australia’s
New Driver for Growth, funded by the Department of Industry,
Innovation, Climate Change, Science Research and Tertiary
Education, p 2. Available from: <http://www.austmine.com.
au/Portals/25/Content/Documents/Austmine%20Survey%20
Highlights.pdf>.
Clarke, N, 2015. Improving productivity in processing operations,
paper presented at The International Mining and Resource
Conference (IMARC), Melbourne, November.
Federation University, 2014. Skills and Innovation in the Resources
and Mining Sectors.
Fiscor, S, 2015. Copper at the crossroads [online], Engineering &
Mining Journal. Available from: <www.e-mj.com/features/5307copper-at-the-crossroads> [Accessed: 12 June 2015].
Hunter, T C, 2014. Concentrators – past, present and future trends
for operators and service providers, in Proceedings 12th AusIMM
Mill Operators Conference 2014, Townsville, p 3 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Hunter, T C and Broome, A J, 2008. Current trends in project delivery
to the minerals industry, paper presented at The AusIMM New
Zealand Branch Annual Conference.
we are metallurgists, not magicians
Hunter, T C, Borthwick, J, Douge, M and Yusi, M, 2009. Project
feasibility studies – a necessary step or the best opportunity to
add value and certainty?, in Proceedings Project Evaluation 2009
Conference, pp 35–48 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Jacobs, 2013. Partnering/Alliances white paper, Jacobs.
Lane, G and Clements, B, 2012. Operations versus projects – how
do people think and what are the implications?, in Proceedings
11th AusIMM Mill Operators’ Conference 2012, pp 11–15 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Loots, P and Henchie, N, 2007. Mayer Brown article, worlds apart
– EPC and EPCM contracts: risk issues and allocation [online].
Available from: <http://fidic.org/sites/default/files/epcm_
loots_2007.pdf>.
Luxford, J, 2006. Project development and construction management,
in Proceedings International Mine Management Conference, pp 97–105
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Masan Resources, 2016. Website homepage [online]. Available from:
<http://www.masangroup.com/masanresources/en/projects/
nui-phao/highlights> [Accessed: 10 April 2016].
Mezei, A, Todd, I and Molnar, R, 2006. Can complex
hydrometallurgical pilot plants effectively reduce project risks?
Part 1 [online], SGS technical paper 2006-2004. Available from:
<http://www.sgs.com/en/mining/metallurgy-and-processdesign/pilot-plants>.
Moore, P, 2015. EPCM and contracting – adapting to conditions,
International Mining, February, pp 78–86.
Morgan, S, Serdzeff, S, Malaloy-On, C and Pagdalian, L C, 2014.
The benefits of technology partnerships during brownfield
upgrades, in Proceedings 12th AusIMM Mill Operators’ Conference,
pp 471–476 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Morgan, S, 2016. Performance and safety optimisation in Vietnamese
operations, in Proceedings Regional Technical Conference Mining
Vietnam 2016 (Mining Media International: Melbourne).
Revy, T, 2008. Trends in project development: where are they
heading? [online]. Available from: <http://www.ibram.org.br/
sites/1400/1457/00000190.pdf>.
Suvio, P, Palmer, J, del Omo, A and Kauppi, J, 2016. Holistic tailings
management solutions [online]. Available from: <www.outotec.
com>.
Walker, S, 2013. Didipio: a Philippines success story [online],
Engineering & Mining Journal. Available from: <www.e-mj.com/
features/3502-didipio-a-phillipines-success-story> [Accessed: 10
December 2013].
Walker, S, 2015. The EPCM perspective [online], Engineering &
Mining Journal. Available from: <www.e-mj.com/features/5029the-epcm-perspective> [Accessed: 05 March 2015].
Wasmund, B, Voermann, N, Haneman, B, Sarvinis, J and Sheehan, G,
2011. Implementing new technologies in metallurgical processes:
building plants that work, in Proceedings 50th Conference of
Metallurgists, pp 1–19 (Canadian Institute of Mining, Metallurgy
and Petroleum: Montreal).
113
Contents
Is an 80th percentile design point logical?
D David1
ABSTRACT
Clearly, a plant designed only to treat average ore at the nameplate rate will fail to
achieve nameplate in any typical year. To insert the necessary capability to achieve
nameplate it is common process engineering design practice that a plant must be
able to treat an ore with 80th percentile value (hardness or competence values for
example) at the nameplate rate. In the author’s experience, apart from a couple of
notable cases, this has been applied reasonably well as a principle, with widespread
success. However, in a number of recent instances it became clear that the use of
the 80th percentile number would have resulted in significant under-design of the
plant. This paper makes the case that the 80th percentile, as a principle, can have
serious flaws and its use needs to be assessed on a case-by-case basis. The discussion
in this paper is the first step in developing a new design principle, and the associated
methodology for selection of a design value, that will ensure plants are designed to
achieve their nameplate capacities.
INTRODUCTION
Designing a process plant has many conventions that are often taken for granted. One
of these is that the 80th percentile value of a key measure, like the Bond ball mill work
index (BWI), will provide an unquestionable margin of design safety in the plant. The
normal procedure is to take a set of test results, for example 20 individual BWI values
from around the orebody, arrange them largest to smallest and take the 16th largest
value. This is placed in the design criteria as the 80th percentile, usually alongside the
average BWI value, and is subsequently used to design the ball mill. Many years later
that number can assume almost legendary status as the work index of the deposit or it
may have assumed infamous status as being the main cause of the failure of the plant
to achieve nameplate throughput or grind size.
The purpose of this paper is to explore circumstances where the 80th percentile
value will not provide a definitive design point. Guidelines for avoiding being
fooled by a false design value are provided, together with the basis for alternative
design methodologies.
PROBLEMATIC DATA SETS
The annual plan
If the 20 values of a particular measurement are provided from a 20-year annual mine
plan, then failure to design for variability for that measure is virtually assured. One
of the most common values that dangerously makes its way into design criteria from
mine plans is head grade.
1. FAusIMM(CP), Technical Director – Process,
Amec Foster Wheeler, Perth WA 6000.
Email: dean.david@amecfw.com
A recent analysis of copper (Cu) head grade for a project determined the variability
of a number of related data sets ranging from the most inherently variable to the least.
The most variable data sets are those that look at the orebody in individual parcels
that may represent minutes or hours of plant feed. Two examples of such data (present
in virtually all projects) are the drill database and the block model. At the other end
of the variability scale is the annual mine plan. To arrive at the annual mine plan it
is normal to use the individual ore blocks contained in the block model in an orderly
and controlled fashion. The use of blocks (in an open pit example) is orderly because
it commences at the surface and must progress downward in some logical mining
sequence. The use of blocks is controlled because it is typical to attempt to obtain a
target head grade to the process plant when making selections from the blocks that
are immediately available to be ‘mined’. In many cases there is also a stockpiling
system in place (usually incorporated into the block sequencing procedures) between
to the mine and process plant to allow grade control to be achieved when the mine
cannot directly provide it. Usually the measure being controlled is the head grade
of the most valuable component, and this also is the measure for which believable
variability information is required for design.
115
D David
Almost immediately upon arrival at a copper mine for
commissioning the author was greeted with the bunded
flotation area of the plant freshly filled with copper
concentrate, mostly from the cleaners and recleaners. The
launders and pumps were continuing to overflow. On asking
what had caused the problem, the response was that the plant
had been processing 8 per cent Cu in feed for the last day and
they couldn’t do much about the problem immediately as the
stockpile was full of 8 per cent Cu ore!
The design criteria for this project were, in theory, prepared
on an even more conservative basis than the 80th percentile
because the ‘maximum’ grade had been used (100th
percentile). This value was 1.6 per cent Cu and, obviously,
nowhere near what was being dealt with by the plant.
Although this value was in the design criteria, the maximum
case mass balance assumed a head grade of only 1.39 per cent
Cu, the 80th percentile value. Both the 100th percentile and
the 80th percentile values were selected based on the data in
the annual mine plan.
An actual example from a design study of the variability
levels for the various data sets described at the commencement
of this section is shown in Figure 1. The ‘upper and lower limit
95 per cent’ lines represent the range that contains 95 per cent
of the data within that particular set. Note that this is not
the same project that had the 8 per cent Cu in flotation feed
problem. This ore has less variability and an average grade of
only 0.51 per cent Cu.
The 80th percentile value for the shift data set is 0.91 per cent
Cu and this would be a reasonable value to use for design
purposes. The shift data set also predicts that the plant
can expect to treat 1 per cent Cu (or greater) in feed for
2.5 per cent of the time (the ‘Upper Limit 95’ is actually the
97.5th percentile value) which is 27 × 8 hour shifts per annum,
or more than two per month. Clearly choosing to use the 80th
percentile annual, monthly or weekly mine plan Cu grades for
design purposes would dangerously underestimate the real
Cu grade variability that the plant must be able to process.
Interestingly, the core database 80th percentile would also
underestimate the shift 80th percentile value, while both the
metallurgical test data set and the block data 80th percentile
values would be acceptable estimates. However, all three
small-sample data sets overestimate the Upper Limit 95 value
by at least 100 per cent compared to the shift data.
When put in the context of real operational requirements for
a process plant, it becomes clear that the even a weekly mine
plan is not a valid source of process design information, let
alone an annual plan.
The composited sample
To save money, a recent client had only conducted five sets
of metallurgical tests at laboratory scale for a definitive
feasibility study (DFS) level design. Minimal background
information, apart from the data itself, was provided for
a review. The five BWI values provided all lay between
11.1 kWh/t and 11.8 kWh/t. The standard deviation (SD)
between the six tests was 0.3 kWh/t, approximately what
is accepted to be the inherent repeatability of the BWI test
itself. The standard deviation of the drop weight index (DWI)
results was only 0.4 kWh/m3, which is again similar to the
level of repeatability expected of that test.
The 80th percentile of the five BWI values was only
2 per cent greater than the average value and the maximum
value was only 3.5 per cent above the average. For the DWI
result set, the 80th percentile value was only 10 per cent higher
than average and the maximum value was not much higher
at 13 per cent above average. There were two possibilities to
explain these results, either the orebody was the first one in
the author’s experience where the material was ‘all the same’
or, something very strange had happened in the constitution
of the samples.
In response to requesting more information, the core
intervals making up each of the samples were provided. The
source of the problem was immediately obvious and it was
composite preparation. Through compositing all semblance
of variability had been eliminated from the test samples. Each
FIG 1 – Variability bands for grade data sets from a single orebody.
116
we are metallurgists, not magicians
Is an 80th percentile design point logical?
test sample was a composite of a minimum of 60 core intervals
from a minimum of four different drill holes. The stated
aim of this particular piece of work had been to define the
properties of five different areas in the orebody. Obviously, it
was confirmed that the five areas can be considered virtually
identical, on average, with respect to ball milling. However,
the test data set contains compositing that is equivalent to,
at least, the level of compositing inherent in an annual mine
plan (and maybe even equivalent to a five-year mine plan
basis). As demonstrated in the last section, annual plan data is
unsuitable because it contains a level of variability far below
what is needed for safe design.
Applying the relationship between annual variability and
shift variability from Figure 1, it is possible to provide an
estimate of what the true variability of BWI values might
have been, had multiple contiguous samples been tested
individually. From Figure 1 the 80th percentile value from
annual mine plan data is only 10 per cent above the average
value. In contrast, the 80th percentile on a shift basis is
74 per cent above the average value. Therefore, with the
benefit of having all data sets available, a reasonable estimate
of the design point, the shift 80th percentile, can be made from
the annual BWI 80th percentile as follows:
80th BWI Shift = Average BWI +
(80th BWI Annual – Average BWI Annual) × 7.4
This equation has been applied to the five-sample case above
where the only data available has annual-plan-equivalent
variability and is unsuitable for design purposes. The outcome
of this calculation is a more believable 80th percentile BWI
value of 13.2 kWh/t, 16 per cent above the average value of
11.4 kWh/t. Given the uncertainty in what the original data
set actually represents, the real 80th percentile value on a
production shift basis could be even higher.
This example clearly shows the danger of relying on
composited samples for design purposes. The circumstances
behind these composites are extreme, but it must be recognised
that any compositing reduces the inherent variability that will
exist in the set of test results. Knowing how samples have
been selected and prepared is essential to understanding the
design implications of test data.
As a variation on the compositing theme, in many projects
tests have been carried out on annual composites (for
example Year 1 composite, Year 2 composite etc). Taking an
80th or 100th percentile value from the set of annual composite
results will, again, dangerously underestimate the variability
the process design needs to cope with.
Multi-modal ore properties
Few orebodies consist of a single lithology or a single
geological classification of material. In many instances (but
not all) the comminution and separation properties of the
different geological units can be distinctly different and need
to be understood separately. If the geological differences also
correspond to metallurgical differences, then the geological
ore types are also valid geometallurgical ore types. In
addition, each geometallurgical ore type will display a range
of properties, and the proportions of each ore type in plant
feed will vary from shift to shift and year to year.
A classic example would be where the orebody has an
oxidised cap, a transition zone and fresh rock at depth.
The approach often seen by the author is for the fresh rock
to be represented by 20 or more samples, the transition by
five samples and the oxide by two, neatly matching their
proportion in the orebody.
we are metallurgists, not magicians
Provided the samples have been selected correctly, it is
certainly possible to derive a reasonable 80th percentile value
for the fresh ore. Conversely, it is not even possible to estimate
the degree of variability in key properties (let alone the 80th
percentile values) that characterise the oxide cap using only
two samples.
If the oxide cap represents 100 per cent of plant feed for the
first six months, then understanding the properties of that
material is of reasonably high importance. The plant will be
commissioned on that ore and it may play a significant role
when banker’s tests and warranty tests are being conducted.
In this situation, a minimum of ten spatially distributed
samples of oxide ore need to be tested before the variability
can be estimated. There is usually one guarantee with oxide
ore, it will be more variable than the fresh rock from which
it is derived.
The transition ore presents similar problems, not least of
which is definition. An ore is called transition because it is
part way between the original rock (in this case fresh ore)
and the geological layer above it (in this case the oxide cap).
Invariably, the transition ore will contain examples of the
end members (fresh and oxide) together with everything in
between. The high degree of variability in typical transition
ore demands that a reasonable number of samples be tested,
provided of course that the transition ore type represents
a substantial plant feed component for a long enough time
period to be considered separately in the design process.
The more components in the ore the more important it is
to have detailed mine plans guiding the design process. The
most common misconception for the inexperienced is that
the proportions of each ore type in plant feed can be taken
directly from the monthly or annual mine plan. The only
way a realistic appreciation of the variability in ore type
proportions can be gained is to understand how such ore gets
from the mine to the plant.
As an example, an orebody has two ore types with distinctly
different comminution properties and these ore types are
planned to be, on an annual basis, delivered to the plant in
a 50:50 ratio. Simplistically the ore type properties can be
averaged and then used for design. However, discussions
with the geologists reveal that the first ore type is on one side
of the orebody and the other ore type is on the other side of
the open pit. Discussions with the miners reveal that they are
only intending to have one shovel in ore at any one time and
that there is no intention to set up blending stockpiles, as it is
too expensive to double-handle all the ore. The 50:50 ratio is
not a controlled 50:50 blend and will be sequential processing
of one ore type, followed by the other. The resulting design
requirements are totally different to the requirements for
processing a controlled 50:50 blend.
Where does the 80th percentile come into this discussion?
The 80th percentile of the blend is obviously of little use. Each
ore type now needs to be understood individually in terms
of variability, and the number of test samples required has
probably doubled. If the 80th percentile value is to be the
basis of design, then the properties of each ore type need to be
measured to the degree where a reliable 80th percentile value
can be extracted from the results for each ore type.
Before proceeding to design the plant to treat each ore type
separately, a constructive discussion is needed across all three
disciplines to explore the implications of alternative mining
strategies. For example, introducing blending before the
primary crusher or reducing the shovel size and having two
(or more) ore faces supplying plant feed at all times. Beware
of the argument that often comes up in such discussions that
blending is happening in the coarse ore stockpile after primary
117
D David
crushing. Unless the coarse ore stockpile is a bed blending
arrangement where ore is stacked in layers and reclaimed
across the layers, then it can be safely assumed that no blending
of any consequence occurs in the coarse ore stockpile.
VARIABILITY EXPECTATIONS
Steve Morrell (2011) published the distribution of variability
levels that exists within the sag mill competence (SMC) test
database. Typically, SMC test samples are from contiguous
core intervals and are done in numbers large enough (per
orebody) to derive reasonable statistics. The variabilities of
the drop weight index (DWI, in kWh/m3) results from the
650 orebodies represented in the database (at the time of his
writing) were distributed as shown in Figure 2.
For an orebody with an average DWI value of 5 kWh/m3 and
a coefficient of variation (COV) value of 25 per cent, the 80th
percentile value must lie between 25 per cent and 50 per cent
greater than the mean value (between one and two standard
deviations greater than the mean). An 80th percentile value
of 7 kWh/m3 would not be unreasonable for this example.
However, the COV could be any value in the range from
5 per cent to 60 per cent, so 80th percentile values of 10 kWh/
m3 and 5.4 kWh/m3 are also possibilities for a 5 kWh/m3 ore.
All critical measures used for design will have a similar
(but usually less broadly spread) range of possible variability
values. Without having enough actual measurements from
tests performed in the correct manner on appropriate samples,
it is not possible to achieve a reliable estimate of variability
for a particular property in an orebody. This statement also
excludes the method applied in the compositing exercise
discussed previously. Although the final 80th percentile
estimate was superior to the estimate from the original data,
it was also far from definitive.
AN ALTERNATIVE APPROACH
As has been demonstrated in the examples above, the 80th
percentile value can be totally misleading and dangerous
as a design point if the data set from which it is derived is
not suitable for design purposes. In any data set typically
available as a foundation for design, the average value will
be a much more reliable number than the 80th percentile or
the SD. In the example given for the effect of compositing,
an estimate was made of a believable 80th percentile value
based on an assumption about the variability differences
known to exist between data sets calculated on varying time
scales of production. Notably the adjustment was made
relative to this well-defined value, the average of the data
set. The method employing scaling of variability (or 80th
percentile) between data sets is one alternative approach to
selecting the design point.
The concept of the 80th percentile is one that provides some
comfort in design, but it is also one that is arbitrary. Why
not use the 90th percentile, the 75th percentile or some other
value? In a recent design project some criticism was provided
in a review that the 80th percentile had not been used and
reliance was placed on what was considered a riskier
value, the 75th percentile. Regardless of the fact that the
75th percentile was the agreed design point with the client,
the results database was revisited and the design outcome
recalculated using the 80th percentile value. In this instance
the difference was less than 0.5 per cent in the mill power,
insignificant in the accuracy of the design. For this particular
deposit the SD of the value in question was very low but the
variability in the data set was considered valid for design.
The data set consisted of about 50 results, all from individual
contiguous samples according to an Australian Mining
Exploration Companies (AMEC) sample selection plan. The
insensitivity, in this instance, of the design outcome to the
selected percentile value was of concern, mainly because the
design point was not all that much greater than the data set
average. This particular instance has led to a re-focusing of
the basis of design towards using the mean value, rather than
any high percentile derivative (80th or 75th) of the data set.
Once the focus is on the mean value then the concept of
‘confidence’ can be introduced to design process. It is possible
mathematically to derive the upper and lower confidence
limits for a given mean value of a data population by
knowing the number of samples in the population, the SD of
FIG 2 – Distribution of orebody COV* values for competence measurement (after Morrell, 2011). *The coefficient of
variation (COV) is the standard deviation divided by the average value, expressed as a percentage.
118
we are metallurgists, not magicians
Is an 80th percentile design point logical?
the population and the degree of confidence that is required
in the mean value. The higher is the required confidence level,
the wider the range of possible mean values for any data set.
A range of possible mean values exists because, if exactly the
same sample selection process was conducted on the same
orebody, but different core intervals were chosen, the mean
result from testing this second sample set is almost certain to
be different to the mean result from testing the first sample set.
Therefore, it is possible to know what the highest likely mean
value is (the upper confidence limit of the mean) and Microsoft
Excel provides a function to simplify this estimation. As the
SD is an integral part of the calculation it is still necessary to
re-estimate the SD to a higher value if compositing or annual
planning has smoothed the available data set.
Confidence can be thought of as the inverse of risk. You
have selected good test samples, generated a result data
set and you have averaged it. How do you know if your
measured average value is a good estimate of the real average
for the orebody? As an example, a set of ten results give a
mean BWI of 9.1 kWh/t and the individual test results range
from 7 to 12 kWh/t. As a designer, you are happy to work
with a confidence level in the mean value of 90 per cent, as
this means that the risk of your measured average being
materially wrong is 10 per cent. A one-in-ten risk of being
wrong is unacceptable to most designers but the catastrophic
design risk level is more acceptable at one in 20.
The 10 per cent risk associated with applying 90 per cent
confidence limits is actually made up of two components.
The first 5 per cent risk component is that the measured
average is outside of acceptable limits on the high side. As
a result, the real work index of the orebody is found to be
even lower than the lower confidence limit. The design risk
is that a mill that is unnecessarily large and powerful will be
selected by using the 9.1 kWh/t average value. The second
5 per cent risk component is that the real average work index
of the orebody is found to be much higher than the upper
confidence limit. In this case the risk is that the selected mill
will be too small and consequently nameplate throughput at
nameplate grind size will not be achieved. Mill under-sizing
can be a major risk to project viability because the revenue
stream is likely to be compromised.
For the set of ten test results in the example above, the
90 per cent confidence limits were calculated to be 10.1 and
8.1 kWh/t respectively. In accordance with the discussion
above, there is a 5 per cent chance the real mean BWI for the
orebody is greater than 10.1 kWh/t and the 9.1 kWh/t average
value is too low and will compromise the design. There is
also a 5 per cent chance the real mean is less than 8.1 kWh/t
and a design based on 9.1 kWh/t will be conservative. As
the catastrophic risk lies in under-sizing the mill the sensible
designer will take 10.1 kWh/t as the average work index,
rather than 9.1 kWh/t. The risk of overdesign is increased
significantly but the risk of under-design is acceptable.
Now consider if 40 samples were tested and the individual
results all lay within the same range, 7 to 12 kWh/t, and the
average was again 9.1 kWh/t. The 90 per cent confidence limits
around this average value are much tighter and calculate to
8.6 and 9.6 kWh/t. The increased number of samples tested
has increased our confidence in the measured mean and
reduced the associated design risks. The sensible designer can
now use 9.6 kWh/t as a safe average value with exactly the
same risk level that existed when choosing 10.1 kWh/t based
on ten samples.
In selecting a safe design point the ten sample set forces
the designer to assume that the average could be as high as
10.1 kWh/t. The 40 sample set allows the designer to lower
we are metallurgists, not magicians
the average by 0.5 kWh/t to 9.6 kWh/t. The 30 extra tests
have probably saved $1.5 M in capital costs by allowing a
smaller mill to be installed without compromising on design
confidence. It is also clear that using the measured average of
a data set always incurs higher risk than using the calculated
upper confidence level for the mean value.
Another good thing to check is if the upper confidence limit for
the mean value calculates to be greater than the 80th percentile
of the data set. Consequently, by using the 80th percentile
value the designer is inadvertently designing with a number
that may actually turn out to be less than the upper possible
orebody average. This simple check shows definitively that too
few samples have been tested for the inherent variability that
exists in the orebody.
Having demonstrated a method to find a safe average value
(the upper confidence limit) it is now possible to construct a
new design methodology based on estimation of data set SD
coupled with the statistically-derived upper confidence limit
of the mean value. It will be argued (in a follow-up paper)
that such an approach is more robust than the tried, and often
wanting, 80th percentile based method.
CONCLUSION
The 80th percentile value can be a useful design number,
provided all the correct prerequisites are present in the
population that it is derived from. These are:
1.
The samples that the individual test results represent
must consist of individual lengths of contiguous core,
or something close to this ideal.
2.
There must be some understanding of how the
variability of the available data set is related to the
variability that needs to be catered for in plant design.
If a pathway to estimating shift-by-shift production
variability (for any variable) is available, then this can
provide useful guidance and should be followed.
3.
There must be some relationship between the available
data and the time sequence of presentation of ore to
the plant. For example, if the ore that the plant is to
be commissioned on is not clearly understood, then
an adequate number of samples of commissioning ore
need to be tested before commencing definitive design.
4.
The relationship between the mining practices to be
employed and the distributions of the various ores in
the deposit need to be understood at an operational
level, and not a smoothed monthly or annual ideal.
It is essential that the process design engineers have access
to all raw test data, and that they understand the basis of
sample selection, they understand any compositing that
has been performed on the samples and why it was done
that way, and they have confidence in the testing that has
been conducted. It is also essential to link the test results
to the ore types and to the time sequence in which the ore
is likely to be mined. Finally, the process engineers must
understand the operational reality of ore access and delivery
to the plant, including any assumptions regarding blending.
A good start to gaining these necessary understandings
is to ask the geologists for the drill database and to ask the
mining engineers for a sequenced block list. The resulting
conversations are usually very enlightening.
If a valid 80th percentile design value is not obtainable from
the available raw data, then a relatively simple pathway for
estimating a useful 80th percentile value (based on the relative
variabilities within available data sets) has been demonstrated.
119
D David
To increase the confidence that can be placed in a design
outcome, the basis for a new and robust design methodology
has been proposed that is linked to a worst-case estimate
of the mean value (at a desired accuracy level) and with a
defined confidence level based on the client’s appetite for risk.
gooey ore and the crunchy blue rock, they are both the same
geological classification. Acknowledgement is also given to
Steve Morrell, who has followed (and sometimes built) the
pathway to understanding variability.
ACKNOWLEDGEMENTS
Morrell, S, 2011. Mapping orebody hardness variability for AG/
SAG/crushing and HPGR circuits, in Proceedings International
Autogenous Grinding, Semiautogenous Grinding and High Pressure
Grinding Roll Technology 2011 (eds: K Major, B C Flintoff, B
Klein and K McLeod) paper 154 (Canadian Institute of Mining,
Metallurgy and Petroleum: Montreal).
Anonymous acknowledgement is given to those projects
supplying examples, good and bad, used in this paper. Special
acknowledgement is given to the geologist who advised (in an
operating pit) that there was no difference between the brown
120
REFERENCE
we are metallurgists, not magicians
Contents
Measuring and taking notice of orebody variability
– an essential ingredient for reliable plant design
D David1
ABSTRACT
The majority of semi-autogenous grinding (SAG) and ball mill design methodologies
are applied using a single design value for mill selection, such as the 80th percentile
ball mill work index or SAG mill comminution (SMC) value. Inherent in the selection
of this design value is a desire for the mill to accommodate unfavourable, but not
absolute worst-case, feed conditions. It has been observed in numerous ‘postmortems’ of unsuccessful designs and also in many optimisation studies, that the
selection of the design value has been poor. Statistical analysis of test work data
sets shows that inherent variability in the data can, by itself, lead to design errors.
Metallurgical testing is an attempt to measure the characteristics of millions of
tonnes of ore using only a few kilograms of sample. In this paper, some realistic but
synthetic test work data sets are used to explore old and new methods of defining
the design envelope and subsequently selecting a reliable point. The aim of the
paper is to provide practitioners with tools that reveal the adequacy (or otherwise)
of metallurgical test programs intended to support various levels of project design. It
also provides guidance as to how it is possible to statistically justify a comparatively
high design point in the absence of adequate test data.
INTRODUCTION
The words ‘it’s all the same’ are regularly heard in relation to orebody properties and
this phrase should always be an alarm bell to a mineral processing plant designer.
Nature does not provide orebodies that have consistent properties and this means
there will always be a range of values for any measure, be it for hardness or separation
characteristics.
Some of that variation in measurement comes from the repeatability of the
measurement technique itself and this should not be confused with ore property
variation. Test repeatability must be known and understood before using any test
results for design purposes.
The variation that will be discussed in this paper is the spread of test results
that naturally occurs when a set of samples, each sourced from a different threedimensional location in an orebody, is tested.
The degree of variability that is measured for the material will determine both the
design envelope and the safe design point for that measure.
It is easy to incorrectly measure variability through poor sample selection and
through compositing of samples inappropriately.
Correctly measuring variability leads to appropriate selection of design property
values and this, in turn, leads to a process plant design that can achieve its nameplate
performance requirements.
QUANTIFYING VARIABILITY
Ore variability measurement has two components. The magnitude of variability that
is measured and the belief you can place in that answer.
The magnitude of variability in a set of numbers is typically described as its standard
deviation (SD). The SD is almost always stated together with the average or mean value.
A typical example would be 10±1.0 where the average or mean is 10 and the SD is 1.
1. FAusIMM (CP), Technical Director – Process,
Amec Foster Wheeler, Perth WA 6000.
Email: dean.david@amecfw.com
The belief you can place in that average value is called the Confidence Level. This is
mathematically determined using an equation that includes the number of test results
(n) and the SD value. As you conduct more tests you get more confident that you have
actually measured the mean value and also that the SD magnitude is correct.
A typical statement of confidence is that there is 90 per cent confidence in the mean
to an accuracy of ±5 per cent. The statement ‘90 per cent confidence’ can also be
121
D David
written as ‘one time out of ten’. So, if the mean was measured
ten times (If the original calculations were based on a set of 20
correctly selected test samples then it is necessary to repeat
this by measuring another nine sets of 20 correctly selected
samples, 200 samples in all) we would expect nine of those
mean values to be within 5 per cent of this first mean value.
One of those readings would be expected to be greater than
5 per cent different from this first mean value.
TABLE 1
Synthetic grindability data set.
Sample
Result (kWh/t)
1
10.5
2
9.5
3
8.8
Now if a 5 per cent error is enough to cause serious problems
in the design then you have a one in ten chance of getting
the design wrong. If a 5 per cent high result is acceptable but
a 5 per cent low result is not, then things improve because
confidence works both ways. The one in ten problem result
could be a high result or it could be a low one. There is a
one in 20 chance that the result will be high (greater than the
original mean +5 per cent) and a one in 20 chance it will be
low (less than the original mean – 5 per cent). Because only a
low result is a problem we now have a situation where there
is a one in 20 chance of the design failing.
4
11.0
5
12.2
6
10.2
7
10.5
8
11.6
9
8.4
10
8.8
11
7.9
Is this acceptable or do we need to be more accurate? Do we
need to reduce the risk further? Do we simply overdesign and
mitigate the risk?
12
10.6
13
10.2
14
9.8
15
9.5
The answers to all these questions depend upon the risk
profile of both the project and the client. For new technology, a
one in 20 chance of failure may be an acceptable risk level. For
the application of well-established technology, it is usually
expected that the design should have effectively no risk of
failure, say a confidence level of 99 per cent or one in 100.
16
9.9
17
10.0
18
10.0
For a major mining house, where the project is one of
many and access to expertise is relatively easy then perhaps
a confidence level of one in 100 is also required. Often such
a project will be competing with other projects in which the
probability of success is similarly high.
19
11.1
20
9.7
For a junior miner with a single potential project and a high
degree of capital sensitivity and time pressure, perhaps a
one in 20 chance of failure is acceptable. It is up to those that
would supply the money to determine if they should ‘take
the chance’.
Mean or average
10.0 kWh/t
Standard deviation
1.05 kWh/t
90% confidence in average
0.39 kWh/t
The unfortunate part about statistics is that uncertainty is not
removed by analysis, it is simply quantified. For example, a
one in 20 chance of failure means that the set of data you have
generated and analysed could represent that one in 20 problem,
but there is no way to identify this without more testing.
Variability example
To demonstrate the quantification and implications of
variability a constructed example has been prepared. A set
of 20 ‘results’ of Bond testing is listed in Table 1, but in fact,
these numbers have simply been fabricated by the author to
illustrate the points.
The data set is plotted using an S-plot in Figure 1 to show
its variability in a visual fashion. For comparison, a normal
distribution line having the same mean and SD is also plotted.
The comparison shows that the data set can be considered
as closely following a normal distribution pattern. In other
words, the data is regularly distributed above and below the
mean according to known mathematical principles and that
the SD is statistically valid.
The 90 per cent confidence level in the average or mean
value is 0.39 kWh/t. This means that if another 20 samples
were selected and tested there is a 9 in 10 chance the mean will
be between 9.61 kWh/t and 10.39 kWh/t. There is a one in 20
chance the new mean will be greater than 10.39 kWh/t and
a one in 20 chance it will be less than 9.61 kWh/t. Literally,
we have 90 per cent confidence after analysing these 20
122
FIG 1 – S-curve of Bond work index result set with
normal distribution for comparison.
samples that we know the mean of the orebody hardness to
within 4 per cent either side of the measured mean value of
10 kWh/t. Making the simplistic assumption that SD can be
applied across the full range of the data set, a first-pass design
envelope can be prepared as is shown graphically in Figure 2.
Because only 20 samples have been tested out of countless
millions of possibilities in an orebody, we do not know exactly
what the average Bond work index (BWI) is, we only know
that it probably lies within the two red lines on the 50 per cent
rank horizontal line in the graph. To draw Figure 2, it has
been assumed that the confidence band remains unchanged
at 0.78 kWh/t (2 × 0.39 kWh/t) across the entire range of the
measurements in the data set. This is convenient and looks
we are metallurgists, not magicians
Measuring and taking notice of orebody variability – an essential ingredient for reliable plant design
FIG 2 – 90 per cent confidence envelope around Bond work
index data set assuming consistent standard deviation.
FIG 4 – Design envelope after considering 90 per cent confidence in
mean value and 95 per cent confidence limit in the standard deviation.
sensible on the graph, but it is not correct as the confidence
calculation has only provided information about the mean
value, not the remainder of the distribution. There will also be
uncertainty in the value of the SD. As will be demonstrated, the
uncertainty in SD is more of a problem for traditional design
methodologies than is the uncertainty in the mean value.
Rank Confidence Limit (RankCL) at various ranking levels.
The RankCL value is half the span from the upper line to the
lower line. At the mean the 90 per cent confidence interval
and the RankCL values are almost identical (0.39 kWh/t
90 per cent CI versus 0.41 kWh/t RankCL).
Upper and lower limits to the value of the SD can be calculated
from the data set using the method described by Sheskin (2007).
From this BWI data set the calculated SD is 1.05 kWh/t but the
upper limit for SD (to 95 per cent confidence) is 1.55 kWh/t.
The lower 95 per cent confidence limit to the SD is 0.80 kWh/t.
This method provides us with a measure of uncertainty about
the level of uncertainty we think we have measured. When
the design envelope is now modified (selectively) using this
information Figure 3 is the result.
The term ‘selectively’ is used because the envelope could
have been drawn in a number of ways, but the calculation that
maximises the possible data spread in the upper half of the
graph has been used. This analysis clearly demonstrates that
the range of possible values for a typical design number, like
the 80th percentile value, are significantly less certain than the
range of possibilities for the mean value. An alternative drawing
of the envelope to emphasise the effect at the lower end of the
data range would demonstrate a similar level of uncertainty
in the 20th percentile ranking value as has been demonstrated
for the 80th percentile ranking. When the maximum level
of uncertainty is estimated across the full distribution the
envelope can now be realistically drawn as per Figure 4.
Figure 4 represents a much more realistic understanding of
the uncertainty associated with choosing any percentile value
in the distribution than is represented by Figure 2. This issue
is further quantified by plotting what has been termed the
In this data set the uncertainty in knowing the 80th
percentile value is 0.71 kWh/t which is 75 per cent higher than
the uncertainty at the mean. Uncertainty in knowing the 90th
percentile is more than double the uncertainty in the mean.
From Figure 3 and Figure 4, the lower limit (with 95 per cent
confidence) for the 80th percentile design point is only just
above the measured mean at 10.3 kWh/t. The upper limit
is 11.7 kWh/t. The difference between these two values
represents a 14 per cent difference in power requirement
ultimately selected for plant design.
The statistical analysis is indicating that these 20
measurements give a particular picture of the distribution
of hardness in the orebody. However, testing another 20
samples from this same orebody and simply choosing the P80
value for design could have just as easily resulted in design
points anywhere from 10.3 kWh/t to 11.7 kWh/t without
being statistically inconsistent with the data set that has been
analysed here.
In a traditional analysis, the actual 80th percentile ranked
data point is towards the lower limit of this statistical band
at 10.6 kWh/t. The mathematically calculated 80th percentile
value (using the Percentile function in Microsoft Excel for
example) is consistent with the normal distribution curve at
10.9 kWh/t. The major point of this paper is to argue that
these traditional estimates are inadequate for design using
this data set because this orebody, sampled and tested in the
traditional way could just as easily have provided a design
point as high as 11.7 kWh/t.
Alternative design point selection methods
One obvious alternative is to choose a percentile rank higher
than the 80th percentile that gets closer to the statistically
valid upper limit of 11.7 kWh/t. The 90th percentile point is
only 11.1 kWh/t and the 95th percentile point is 11.6 kWh/t.
Therefore, in this data set the second highest measured value
would become the design value. However, from Figure 5
it has also been shown that this 95th percentile point could
easily be 1 kWh/t higher.
FIG 3 – Effect on the design envelope of applying the upper
standard deviation (SD) limit to the upper mean value
and the lower SD limit to the lower mean value.
we are metallurgists, not magicians
If the normal distribution curve was used in mathematically
calculating a percentile (rather than simply choosing the
second highest measurement) it would be very safe to adopt
the 95th percentile point (11.74 kWh/t). If the Excel percentile
function was used, then the 90th percentile would also appear
to be suitable as it gives a value of 11.55 kWh/t.
123
D David
The ultra-conservative approach equation
Design point = UCL90 + UCL95SD
[2]
where:
UCL95SD
is the upper limit, at 95 per cent confidence,
of the SD of the data set
For the example data set this approach gives a design value
of 11.93 kWh/t.
This approach is recommended where the data does not
follow a normal distribution and is especially applicable when
insufficient tests (say less than 10) have been conducted to
establish the true mean and variability in the design property.
The outcomes of the various design point selection methods
are compared in Table 2.
FIG 5 – Calculated confidence in measured Bond work index
value either side of the mean (50th percentile) value.
The first option, choosing a measured point, is the least
favoured method. In this data set, choosing a particular
measured point is relatively simple and presents few problems
apart from the relatively low confidence level that has been
identified when choosing points far from the mean. But this
data set is regular in its shape while many other real sets are
not. One or two outlying points at the top of the distribution
would invalidate any method where an upper measured
point is simply chosen.
The second method, calculating percentiles based on a
fitted standard distribution, also assumes regularity in the
data set. If the set follows the normal distribution curve
closely, then using the basic normal distribution techniques
are appropriate, although a value ranked higher than the
traditional 80th percentile is needed to be prudent in design.
This comparison suggests that the traditional design
method is likely to underestimate the power requirements of
the circuit by at least 7 per cent. A plant failing by 7 per cent
to achieve its nameplate throughput rate may be enough to
reduce the profitability of the project to zero, or even to a lossmaking scenario.
Detailed equations
The equations behind the two methods will be explained in
a form suitable for use in Microsoft Excel. It is assumed the
20 measured data values are in a single range of 20 cells that
has been assigned the name BWI. This means that the SD term
in Equation 1 is calculated as STDEV.S(BWI). The capitalised
function name STDEV.S is a new function in Office 365 version
of Microsoft Excel and has equivalents in older Excel versions.
UCL90 can be calculated as:
UCL90 = AVERAGE(BWI) + Conf90[BWImean]
[3]
Conf90[BWImean] = CONFIDENCE.NORM(0.1,BWISD,20)
[4]
The third method (using the percentile function) seems
to lie somewhere between the first two and is subject to the
constraints of the normal distribution assumption.
where:
The recommended method for choosing a safe and justifiable
design point is to apply some of the statistical calculations
used to analyse the problem above. The difference between
this approach and the traditional methods is that it recognises
both the uncertainty in the mean value and the uncertainty in
the degree of variability of the data.
for 20 samples, and where:
Recommended design point selection procedures
Elsewhere, the author proposed alternative methods for
selection of the design point based on the mean value
and variability in the data set (David, 2017). Equation 1 is
the recommended method and is considered a prudent
conservative approach. Equation 2 represents an ultraconservative approach and is advisable when risk tolerance
is very low or the data set has high uncertainty.
This allows Equation 1 to be rewritten as:
Design Point = AVERAGE(BWI) + CONFIDENCE.
NORM(0.1, STDEV.S(BWI),20) + STDEV.S(BWI)
[6]
The value 0.1 is derived from the choice of confidence level
of 90 per cent. The value 20 is the number of test samples.
Equation 6 is the recommended equation for selecting the
design point.
The calculation of UCL95SD for Equation 2 is more
complicated, as described here.
UCL95SD = BWISD*((n-1)/CHIINV((0.05/2), n-1))0.5
[1]
[7]
UCL95SD = BWISD*((19)/CHIINV((0.025), 19))0.5
where:
UCL90 is the upper limit, at 90 per cent confidence, of the
likely values for the mean value of the data set.
Method
For the example data set this approach gives a design point
of 11.44 kWh/t.
This approach is recommended where the data follows
the normal distribution to a reasonable degree and also
where sufficient tests have been carried out to establish the
variability in the SD.
[8]
TABLE 2
Comparison of design point selections.
SD is the standard deviation of the data set
124
[5]
For 20 samples this equation simplifies to:
The conservative approach equation
Design point = UCL90 mean + SD
BWISD = STDEV.S(BWI)
Design value
(kWh/t)
Relative magnitude
(%)
80th ranked point
10.3
90
80th percentile() calculation
10.6
93
Conservative recommendation
11.44
100
Ultra-conservative approach
11.93
104
we are metallurgists, not magicians
Measuring and taking notice of orebody variability – an essential ingredient for reliable plant design
FIG 6 – Hard and soft ore random data sets.
And as a result, Equation 2 can be written as such:
Design Point = AVERAGE(BWI) + CONFIDENCE.
NORM(0.1, STDEV.S(BWI),20) + STDEV.S(BWI)*((19)/
CHIINV((0.025), 19))0.5
based on the mean and SD of each set individually, are
shown for comparison.
[9]
where:
0.025 is derived from the choice of SD confidence level (in
this example at 95 per cent) and 19 is the number of
samples minus one
Problems with the data set
For convenience and clarity, a fabricated data set that closely
complies with normal distribution parameters was used in the
preceding discussion. Many data sets generated in mineral
processing test work do not comply with this convenient
structure and can lead to problems with design, ranging from
minor to catastrophic.
Multimodal data
When two or more separate populations of properties are
embedded in the one set of data many simplistic analysis
assumptions are no longer valid. To illustrate multimodal
data the statistics (mean, SD and assumption of normal
distribution) of the example data set were used to generate
20 new random results representing what is a soft ore.
A second set of 20 hard-ore results were generated, also
somewhat randomly, by adding an amount varying between
4 and 6 kWh/t to the first set. Normal distribution curves,
To illustrate the resulting bimodality, all 40 values are
treated as a single set and shown as an S-curve in Figure 7.
An overall calculated normal distribution line has been
added that obviously does not represent the set well. A closer
fitting trendline has also been added using the polynomial
function in Excel.
The distribution cannot be considered to follow the normal
pattern and presents a few problems for design thinking.
To see the effect of the increase variability across the
40 samples, the upper and lower 90 per cent confidence values
for the mean and the upper and lower 95 per cent confidence
values for the SD were used to construct the potential upper
design envelope as per Figure 3. The resulting bimodal
envelope can be seen in Figure 8.
The 80th percentile could statistically be anywhere in the
range 13.9 to 16.3 kWh/t.
The four calculation methods used to generate Table 2
were applied to this new 40-point data set. The results are
compared in Table 3.
Again, the ‘conservative recommendation’ calculation
method is preferred to the traditional 80th percentile
methods as the result is closer to the upper end of the
statistically valid range. In this instance, the degree of
potential underestimation of design point using the
conventional methodologies is about 5 per cent.
FIG 7 – Bimodal combined data set.
we are metallurgists, not magicians
125
D David
FIG 8 – Multimodal upper design envelope.
Small data sets
To illustrate the effect of using smaller and smaller data
sets, subsets of the 40 bimodal data points used in the above
example were used as the basis for design calculations. A
random selection method was again used to avoid bias.
The base case for comparison in this analysis is the outcome
achieved using 40 data points. Example calculations have been
conducted selecting approximately 20 samples, 10 samples
and five samples out of the 40. Note that five samples are
generally considered a reasonable number of samples on
which to base a scoping study.
Note that the subset numbers are only approximate. This is
because the random selection process as it has been applied
cannot (by definition), result in a fixed subset size.
Each time the calculation is updated the selections change
as does the apparent shape of the S-curves. This is illustrated
by a selection of five graph updates shown in Figure 9.
On all five graphs the original blue data set is constant. All
other data sets are random plausible outcomes when less than
the full 40 samples are tested.
As statistically expected, if three to seven samples are
selected for testing (the ‘Approximately five samples’ cases
ranged from three to seven samples) the range of possibilities
for the design point outcome is highly variable. Even choosing
the maximum test result may not get close to being a safe
design value as Figure 9a illustrates.
measurement or mathematically calculating the 80th
percentile, always underestimate compared to the maximum
envelope value.
The conservative method either overestimates this value
or estimates a design point consistent with the value (it gave
16.0 kWh/t but only after 40 samples had been tested).
The ultra-conservative method provided an extreme result
(35.2 kWh/t) when only three samples were tested. This result
is a warning of the statistical inadequacy of testing only a few
samples and then trying to infer too much from the results.
The ultra-conservative method provided reasonable design
points when nine or more samples were tested.
It may seem unreasonable to adopt 21.6 kWh/t as a design
value after testing nine points and when the 80th percentile
value of the set is much lower at 15.1 kWh/t (Table 4).
However, the statistical analysis is saying that after nine
samples have been tested the variability at the top end of the
distribution is extremely poorly defined. It is feasible (without
knowing anything about the next samples to be tested) that
there is a significant amount of 20 kWh/t and harder material
in the orebody that could be detected when more tests are
carried out.
Design calculations were also conducted according to
the four methods previously described. The outcomes for
the full and reduced sample number cases represented in
Figure 9e are shown in Table 4.
In line with the naming of the methods, the conservative
and ultra-conservative approaches give higher design points
the fewer tests that have been conducted. This provides an
improved, some would say excessive, level of safety (compared
to traditional methods) if minimal testing is conducted. Most
importantly these high design points provide real incentive
for conducting more tests as the capital cost is most likely to
reduce as confidence improves.
From Figure 8, the upper boundary of the design envelope
is known to be 16.3 kWh/t. However, this value can only
be known after 40 samples have been tested. Each subset
result needs to be assessed as if the 40 samples have not been
tested. The first two methods, selecting the 80th percentile
A number of additional random calculations were captured
showing the potential for problems with small sample sets
In figure 10 and the associated Table 5 it can be seen that
only conducting six tests leads to a potentially problematic
hardness underestimation even using the conservative
TABLE 3
Comparison of design point selections – bimodal data.
TABLE 4
Comparison of design point selections for reduced
sample numbers – graphical case (E).
Method
Design value
(kWh/t)
Relative magnitude
(%)
Samples tested
40
21
9
3
80th ranked point
15.3
95.1
80th point
15.3
15.8
14.8
15.8
80th percentile() calculation
15.3
95.2
Percentile (80)
15.3
16.0
15.1
14.9
Conservative recommendation
16.0
100.0
Conservative
16.0
17.0
17.8
19.0
Ultra-conservative approach
16.6
103.6
Ultra-conservative
18.8
19.4
21.6
35.2
126
we are metallurgists, not magicians
Measuring and taking notice of orebody variability – an essential ingredient for reliable plant design
A
B
C
D
E
FIG 9 – (A–E) Random subset expectations if bimodal data set was only partially sampled.
TABLE 5
Comparison of design point selections for reduced
sample numbers – graphical case (F).
Samples tested
40
20
9
6
80th point
15.3
14.8
13.6
11.1
Percentile (80)
15.3
15.1
14.3
11.1
Conservative
16.0
16.8
16.4
14.8
Ultra-conservative
18.8
20.0
20.1
19.1
we are metallurgists, not magicians
method. The ultra-conservative method is the only one to give
a safe answer. The traditional methods provide dangerously
low design values. The results are a strong argument for
doing at least nine tests.
Another random cycle of the same calculation is shown in
Figure 11 and summarised in Table 6.
Again, the conservative method is preferred to the
traditional selection methods.
Based on conducting this calculation a number of times it is
recommended that if less than ten samples have been tested
127
D David
FIG 10 – Random subset expectations if bimodal data set was only partially sampled, case (F).
FIG 11 – Random subset expectations if bimodal data set was only partially sampled, case (G).
TABLE 6
Comparison of design point selections for reduced
sample numbers – graphical case (G).
TABLE 7
Guide for use of conservative design methods.
Stage
Scope
PFS
DFS
4 to 8 samples
C
UC
N/A
9 to 16 samples
C
C
UC
C
C
C
Samples tested
40
20
13
4
80th point
15.3
14.8
16.0
13.6
Percentile (80)
15.3
15.3
15.7
14.5
17+ samples
Conservative
16.0
16.1
17.0
17.7
Ultra-conservative
18.8
19.0
20.5
26.3
Notes: PFS = prefeasibility study; DFS = definitive feasibility study; UC = ultra-conservative;
C = conservative; N/A = not recommended.
the ultra-conservative method is required. However, if the
ultra-conservative method is arriving at design points twice
the magnitude suggested by traditional methods this should
be taken as a warning that inadequate sample numbers have
been selected for design purposes.
increase the sample numbers in the first column of Table 7 in
line with the increase in uncertainty in the data.
It is proposed that the 90 per cent confidence value, as a
percentage of the mean value, be used as a guide to sample
numbers.
However, at a scoping level it is often not possible to test
more samples, and the implications of slight underestimation
of power are not significant. This suggests a better approach
is to vary the design method according to the project stage
and available sample numbers. A guide for designers is
proposed in Table 7.
In the bimodal data example described above the
confidence level in the mean can be expressed as a fixed
value for the unchanging set of 40 samples but as ranges for
the subset cases (as these change randomly) when the sheet
is recalculated. The confidence levels (as a percentage of the
hardness measurement) are compared in Table 8.
This analysis appears to suggest that 16 samples is sufficient
for a definitive design. Using the conservative method and
with an orebody that appears to give test results of similar
variability to these examples the guidelines are useful.
Where the orebody is less variable the methods will simply
work better. Where the variability is much higher than these
randomly generated sets, then it is likely to be necessary to
Designers can use Table 8 as a reference to see if their data set
is similar to this example or if it is much worse than the example.
Simply calculate the 90 per cent confidence limit in the mean
for the data set and compare it with the table. If, for example,
15 samples have been tested for project XYZ and the 90 per cent
confidence limit is 15 per cent of the mean value, then the XYZ
data set (and more importantly, the XYZ orebody) is much
128
we are metallurgists, not magicians
Measuring and taking notice of orebody variability – an essential ingredient for reliable plant design
TABLE 8
The 90 per cent confidence limit in the mean measured value as a percentage
of the mean hardness measure – relationship to number of samples tested.
Samples
Average confidence as
percentage of mean
Range
4
21.5
18–26
5
15.0
11–21
6–8
15.2
12–20
9–12
11.4
9–13
13–16
9.0
7–10
17–25
8.0
40
7–10
5.6
more variable than the bimodal example. Fifteen samples of
XYZ ore have only given the statistical confidence level that six
to eight samples of the fictitious bimodal ore provided.
A new version of Table 7 would result for XYZ ore in which
the column one sample numbers would double. For orebody
XYZ the 15 samples do not provide enough confidence to even
use the ultra-conservative method for a definitive feasibility
study (DFS) design. More XYZ samples must be tested before
a DFS can be contemplated.
The analysis shows that the confidence in the mean is
strongly related to sample numbers. However, the calculated
SD values were relatively independent of the number of
samples tested. This is useful because it means that the SD is
a guide to overall variability, regardless of how much testing
has been performed. It follows that Sample XYZ would display
a higher SD than the bimodal sample under all circumstances.
CONCLUSIONS
The analysis in this paper shows that the inherent variability
in a measured data set is not considered fully in a traditional
design point selection method, such as calculating an 80th
percentile value.
It has also been shown that as the number of samples
tested decreases the traditional method is relatively poor at
providing a design point that can be considered safe.
Two statistical measures not typically invoked in
traditional design methods are the inherent uncertainty in
the calculated mean value and the uncertainty in the SD for
the measured sample set. Traditionally the mean and SD
values are taken as measured and the 80th percentile value
is considered a reliable indicator of where the plant needs to
be designed to operate.
In the synthetic examples analysed here the uncertainty in
knowing the mean of a measured data set is quantified and
the uncertainty in knowing a value like the 80th percentile
we are metallurgists, not magicians
is shown to approach twice the uncertainty of knowing the
mean. Any measured data sets’ inherent variability can be
evaluated using these same readily available statistical tools.
The mean value of any data set is its best-defined characteristic.
The level of confidence inherent in values distant from the
mean (such as the 80th or 90th percentile) is much lower than
the confidence that can be placed in the mean.
Design difficulties arising from variability issues are
exacerbated when the ore is bimodal and multimodal.
Design difficulties are also exacerbated when small
numbers of samples have been tested. The number of samples
needed to be tested to provide design confidence is directly
proportional to the inherent ore property variability. In a sort
of catch-22, the number of samples that need to be tested to
develop a reliable design for a particular orebody can only be
determined once enough samples (typically seven or more)
have been selected and tested.
Two methods of selecting design points have been proposed
and compared with traditional and simple methods. The
conservative method employs the confidence limit calculated
for the mean value coupled with the SD of the data set.
An ultra-conservative method has also been proposed
which considers both the confidence limit in the mean and
the confidence level the data set provides in the measured SD.
Traditional design methods are prone to design point
underestimation at all sample sizes and are particularly poor
with sample sets of six or less.
The conservative method is reasonably good for estimating
reliable design points with sample sets of eight or more in
the examples analysed. With samples sets of four to eight the
conservative method is prone to some underestimation, but
less severe than using traditional design methods.
The ultra-conservative method is prone to conservative
overdesign with sample sets of eight to fifteen. The method is
unreliable for sample sets less than eight and may randomly
return a conservative or optimistic result.
The traditional and proposed methods can be usefully
compared when selecting design points with real data sets.
They can also be used to analyse under or overdesign situations
in operating plants as this may provide a guide to avoiding
mistakes in future projects. Simply performing the statistical
confidence calculations and plotting the design envelope is a
useful guide to the adequacy of the number of samples tested
and can be used to justify conducting additional tests.
REFERENCES
David, D, 2017. Is an 80th percentile design point logical?, in We are
Metallurgists, Not Magicians (eds: D Pollard, G Dunlop and J Herzig),
pp 115–120 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Sheskin, D J, 2007. Handbook of Parametric and Nonparametric Statistical
Procedures, fourth edition, pp 197–198 (Chapman and Hall/CRC).
129
Contents
Getting optimum value from ore characterisation
programs in design and geometallurgical
projects associated with comminution circuits
S Morrell1
ABSTRACT
It is extremely difficult if not impossible to ensure optimum process efficiency without
having a thorough knowledge of the orebody. For the comminution circuit this means
having detailed data on the breakage characteristics of the ore, in particular how it
varies spatially. When planning campaigns to determine this information, questions
such as ‘What breakage tests should be used?’, ‘How many samples are required?’
and ‘What equations should be used to forecast comminution circuit performance?’
are often asked. The following paper presents a series of statistical analyses to help
answer these questions. Perhaps not surprisingly there are no simple answers, as
it depends upon how the characterisation data are planning to be used, as well as
the inherent variability of the orebody. Guidance is provided to help metallurgists
make informed decisions, which should result in a characterisation programme that
is cost-effective and statistically sound, providing a clear picture of the comminution
behaviour of the orebody in question.
INTRODUCTION
Inevitably the post-boom era has changed the emphasis of many mining companies
away from new project development and expansion to optimisation of existing
operations. However, regardless of the project, high quality laboratory ore
characterisation will remain of the utmost importance. In the case of new project
development, the need to understand the breakage characteristics will remain
as important as it has always been. In optimisation projects, however, where it
will be required to get the very best from existing comminution circuits, accurate
geometallurgical modelling will take on an increasingly important role. However,
these requirements, in both new development and optimisation projects, are likely to
be at odds with budgets as these will be tight. However, this does not necessarily mean
that accuracy should suffer, only that the ore characterisation test work program must
be more focused and streamlined. The following paper discusses several factors and
guidelines that should be considered when faced with developing such programs.
NECESSARY ATTRIBUTES OF A TEST WORK PROGRAM
From a comminution perspective, an effective model, whether used for design or for
geometallurgical purposes should be able to accurately predict the throughput of the
grinding circuit from information concerning the breakage characteristics of ores that
are planned to be delivered to the processing plant. To do so there are at least five
important requirements:
1. sufficient and relevant samples have been identified, extracted and tested
2. appropriate ore characterisation tests have been chosen to describe the
comminution properties of the orebody
3. the model(s)/equations chosen to describe the comminution equipment in
the circuit respond realistically to the values obtained from the chosen ore
characterisation tests
4. all of the above are integrated into an overall description of the operational
response of the grinding circuit that also takes into account non-ore related
influences eg equipment size, speed, ball load etc
1. Managing Director, SMCC Pty Ltd, Chapel Hill
Qld 4068. Email: steve@smccx.com
5. the final model/equations can convincingly demonstrate their accuracy through
validation using real plant data.
131
S Morrell
HOW MANY SAMPLES?
There is no easy answer to this question though it is true to
say ‘the more – the better’. If the deposit is highly variable
the required number of samples will be higher. Also, the
end use will also drive the number of samples required.
Hence if samples are required for a prefeasibility study the
number will be relatively low, whilst if samples are required
for the development of a geometallurgical model that can
accurately forecast daily grinding circuit throughput, the
number required will be at least an order of magnitude
higher. In all cases a staged approach to sample selection and
laboratory test work is recommended to ensure that costs are
kept to a minimum. Each stage should be designed to build
on the knowledge gained from preceding ones, particularly
concerning variability, both spatially within the pit as well as
in terms of absolute hardness values.
The prefeasibility study level is often the best opportunity
to start accumulating useful information of the comminution
properties of the orebody. At this stage, little or no
information is likely to exist on the comminution properties
of the orebody and hence the metallurgist is faced with the
decision of how many samples should be treated for this
very first investigation. A good starting point is to use the
distribution shown in Figure 1. This comes from the semiautogenous grinding (SAG) mill comminution (SMC) Test®
database which currently numbers over 40 000 separate test
results covering over 1500 different ore deposits. The figure
shows the coefficient of variation (standard deviation/mean
expressed as a percentage) of the measured drop weight
index (DWi) values from each deposit (Figure 2 shows the
distribution of mean DWi values from each deposit). The
DWi has been shown to be a very accurate measurement of
the hardness from an autogenous grinding (AG)/SAG and
high-pressure grinding rolls (HPGR) perspective and hence is
highly relevant in this context. The distribution of coefficients
of variation is bimodal, modal values being at 20 per cent and
30 per cent, the average being 30 per cent. Unfortunately the
database does not contain information other than SMC Test®
values and hence unfortunately the author is not in a position
to determine whether this bi-modality can be traced back to
broad geological descriptions of the nature and genesis of the
orebodies. However, it is tempting to hypothesise that there is
a very good physical reason why the distribution should have
such a bimodal shape.
The data in Figure 1 can provide the basis for some simple
calculations that can be used to guide the metallurgist’s choice
of how many samples should be taken and analysed in this
first step to investigate the orebody. Using classical statistics
and assuming the orebody has a variability of 20 per cent
(ie the lower of the modal values), then choosing a total of
10–15 samples should provide a mean hardness value for the
deposit with a precision of approximately 10 per cent at the
90 per cent confidence level. However, if the variability is in
the 30 per cent class, to obtain the same level of precision the
requisite number increases to about 30 and if the variability
is 40 per cent then sample requirement escalates to 45. The
minimum initial number of samples is therefore recommended
to be 10–15. Analysis of the data from these samples will
provide an indication of the true variability and can then be
used to estimate how many more samples (if any) are need to
satisfy the accuracy for the prefeasibility stage. If, for example,
the indicated variability from the initial 10–15 samples is
40 per cent then an additional 30–35 samples will be required.
As the design stages progress through to a final bankable
study, more definition is required to enable more accurate
forecasts of ore properties and hence throughput to be made
during at least the first few critical years of production. This
requires further ore characterisation. However, the results
from the initial stage of testing should provide a firm basis
on which to choose both the numbers and locations where the
samples should be taken.
It is pointed out that these guidelines assume that the
samples are representative of the whole orebody and should
be suitable to provide mean hardness values that relate to
the deposit in a global context only. For the development of
a geometallurgical model much more detailed information
FIG 1 – Distribution of drop weight index (DWi) variability values from the semi-autogenous grinding (SAG) mill
comminution (SMC) Test® database (based on 40 000 SMC Tests® covering over 1500 deposits).
132
we are metallurgists, not magicians
Getting optimum value from ore characterisation programs in design and geometallurgical projects associated with comminution circuits
FIG 2 – Distribution of drop weight index (DWi) mean values from the semi-autogenous grinding (SAG) mill
comminution (SMC) Test® database (based on 40 000 SMC Tests® covering over 1500 deposits).
is required, eg such a model may need to predict the
week-by-week or month-by-month throughput. In such
cases much more definition in terms of the hardness in
specific parts of the deposit is required. In relation to this,
knowing how hardness varies with depth is particularly
important as it indicates the extent to which throughput of
the comminution circuit will increase or decrease as time
progresses. Such knowledge is vital both for day-to-day and
long-term planning and is critical to the continued financial
success of the mine. As a result it is not uncommon for mines
to conduct ore hardness testing on hundreds of samples per
annum throughout the life-of-mine.
WHICH TESTS?
Clearly, the laboratory tests that are chosen need to be
compatible with the modelling approaches subsequently
used at the design stage or in the development of the
geometallurgical model; for example, if the AG/SAG mill
model in the JKSimMet software (by Julius Kruttschnitt
Mineral Research Centre (JKMRC), University of Queensland)
is to be used the laboratory test has to generate an A and b
value as well as the specific gravity. In addition to this, the
metallurgist should also be aware of what are the inaccuracies
inherent in the laboratory tests that have been chosen.
The study conducted by Dunne and Angove (1997) helps
to illustrate this point. Table 1 has been generated from the
data given in their study which looked at the variability in
results from sending the same samples to three different labs
and having Bond crushing, rod and ball mill work index
tests carried out. It is clear from their results that the inherent
variability in the crushing work index test is very large,
values from some laboratories being half those of others. Rod
mill and ball mill tests fared much better, the results from the
ball mill test showing a very good precision of 3.4 per cent
on average. The reasons for the very high variability in the
crushing work index test results is quite likely due to the
we are metallurgists, not magicians
use of non-standard equipment. This problem recently arose
when analysing the data shown in Figure 3. When considering
crushing work index (CWi) data from machines conforming
to Bond’s original design a reasonably good correlation had
been found between the parameter CWi*sg (specific gravity)
and the DWi (DWi units are in kWh/m3 so the use of CWi*sg
gives the same units). However, when data were added from
a modified machine design a very different correlation was
found. The differences between the CWi values from the two
types of machine are of the same order as those obtained from
the Dunne and Angove program. Metallurgists involved in
ore characterisation test work need to be aware of such biases
to ensure that the laboratory chosen to conduct the relevant
test are using standard equipment as well as understanding
TABLE 1
Accuracy of Bond tests (after Dunne and Angove, 1997).
Sample
P
G
R
Av
Min (kWh/t)
12.4
8.1
12.1
10.9
Max (kWh/t)
21.3
17.2
25.6
21.4
Coefficient variation (%)
27.3
43.5
40.9
37.2
Min (kWh/t)
15.9
16.9
15.8
16.2
Max (kWh/t)
20.9
18.2
17.9
19.0
Coefficient variation (%)
11.4
3.8
6.3
7.2
Min (kWh/t)
17.1
16.4
15.5
16.3
Max (kWh/t)
19.3
17.6
16.2
17.7
Coefficient variation (%)
4.6
3.6
2.2
3.4
Crushing work index (CWi)
Rod mill work index (RWi)
Ball mill work index (BWi)
133
S Morrell
FIG 3 – Correlation between drop weight index (DWi) and crushing work index (CWi*sg), where sg – specific gravity.
that even when using standard equipment, some tests are
inherently less accurate than others.
In a study where the same sample was sent to eight different
labs to ascertain the variability from SMC Tests®, the results
shown in Table 2 were obtained and reflect the variability of
the test itself plus variability associated with any differences
between labs in testing machine or operating standards. As
can be seen the coefficient of variation was only 3.9 per cent
for estimates of the DWi and 3.8 per cent for the estimated A*b
value. It should be noted that the SMC Test® estimates of A*b
were done without reference to associated drop-weight tests,
ie they were not ‘calibrated’ with actual drop-weight test
data but used the raw SMC Test® results and global factors
TABLE 2
Accuracy of semi-autogenous grinding (SAG) mill comminution (SMC) Tests®.
Lab 1
Drop weight index (DWi)
A*b
9.6
29.2
9.5
29.3
Lab 2
8.9
31.6
8.7
32.2
Lab 3
9.7
28.8
9.5
29.5
9.0
31.2
8.8
32.0
Lab 4
9.3
30.0
Lab 5
9.4
29.9
9.6
29.3
Lab 6
9.1
30.9
8.6
32.5
9.6
29.3
9.5
29.5
Lab 8
9.1
30.8
9.1
30.9
Mean
9.2
30.4
Standard deviation
0.36
1.2
Coefficient variation
3.9
3.8
Lab 7
134
derived from the SMC Test® database. The magnitude of the
coefficient of variation is very low, indicating a very precise
test. The result can be compared with the value of 5.7 per cent
quoted by Stark, Perkins and Napier Munn (2008) from doing
repeat full drop-weight tests using the same drop-weight
machine at JKTech, University of Queensland.
WHICH MODELS/EQUATIONS?
The choice of which models or equations to be used should
be driven by their demonstrable ability to predict the
performance of existing plants. Once again ‘the more – the
better’ rule applies. The more (and varied) data that exists
to prove the accuracy of the technique, the stronger is the
argument to adopt it.
In design studies the most relevant data with which to
evaluate the suitability of a technique should be those from
existing comminution circuits whose performance has
been predicted then checked against high quality operating
data from the same circuit. Developers of design and
geometallurgical models will claim their techniques and
models to be suitable and accurate, but if a large volume of
appropriately varied data cannot be presented to validate the
claimed accuracy, the metallurgist should be extremely wary
of utilising such models and techniques.
The development of the SMC Test® and the use of the
parameters that it generates in predicting comminution circuit
equipment and circuit performance has been well publicised
in respected international technical journals (Morrell, 2004a,
2004b, 2008, 2009, 2010). Figures 4–6 demonstrate its accuracy
in each of these roles using a large database of operating
plants. More recently the Global Mining Standards and
Guidelines (GMSG) Group has adopted the so-called ‘Morrell
method’ as one of its guidelines for predicting comminution
circuit specific energy (GMSG Group, 2016).
The SMC Test® is unique in that from the one test a number
of parameters are generated which can be used for a variety of
equipment, thus saving money which would otherwise have
to spent on separate tests for each different type of equipment.
Such is the case with Bond’s suite of tests which need separate
tests for crushers, rod mills and ball mills.
For validation of geometallurgical models, the most
appropriate data are from production records over relatively
long periods. Such an example is shown in Figure 7 and
demonstrates the accuracy of a model based on the use of
SMC Test® and Bond ball mill work index data.
we are metallurgists, not magicians
Getting optimum value from ore characterisation programs in design and geometallurgical projects associated with comminution circuits
FIG 4 – Observed versus predicted tumbling mill circuit specific energy.
(Note: ab – ag/ball; ag –autogenous grinding; abc – ag/ball/crusher; sab – sag/ball; sag – semi-autogenous grinding;
sabc – sag/ball/crusher; ss ag – single-stage ag; cr-ball – crushing ball, hpgr – high pressure grinding.)
FIG 5 – Observed versus predicted crusher specific energy.
FIG 6 – Observed versus predicted high pressure grinding rolls (HPGR) specific energy.
we are metallurgists, not magicians
135
S Morrell
FIG 7 – Example of the accuracy of a geometallurgical model (after Alruiz et al, 2009).
CONCLUSIONS
Given the tighter constraints on finances for ore
characterisation programs, metallurgists will need to be far
more selective in their choice of tests in future to ensure that
accuracy is not unnecessarily sacrificed.
The choice of appropriate test(s) should be made based
on measured precision as well as demonstrated accuracy
of the techniques that subsequently use the test results to
predict plant performance. This accuracy can only be truly
demonstrated from analysing large varied databases of
relevant plant performances. Regardless of claims by the
developers of various design and geometallurgical models
and techniques, about their suitability and accuracy, if
data cannot be presented to validate this claimed accuracy,
metallurgists should be extremely wary of utilising such
models and techniques.
In terms of the required number of tests, a staged approach
is recommended in which, as projects develop from the
prefeasibility stage, knowledge of the variability of the
deposit is progressively built and used to drive the number of
tests required in subsequent stages.
REFERENCES
Alruiz, O M, Morrell, S, Suarzo, C J and Naranjo, A, 2009. A novel
approach to the geometallurgical modelling of the Collahuasi
grinding circuit, Minerals Engineering, 22(12):1060–1067.
Dunne, R C and Angove, J E, 1997. A review of standard physical ore
property determinations, in Proceedings World Gold ’97 Conference,
Singapore, September, p 139.
Global Mining Standards and Guidelines (GMSG) Group,
2016. Morrell method for determining comminution
circuit specific energy and assessing energy utilization
efficiency of existing circuits [online]. Available from:
<http://www.globalminingstandards.org/wp-content/
uploads/2016/08/20150821_Morrell_Method-GMSG-ICEv01-r01-.pdf> [Accessed: 29 May 2017].
Morrell, S, 2004a. Predicting the specific energy of autogenous and
semi-autogenous mills from small diameter drill core samples
[online], Minerals Engineering, 17(3):447–451. Available from:
<https://doi.org/10.1016/j.mineng.2003.10.019> [Accessed:
29 May 2017].
Morrell, S, 2004b. An alternative energy–size relationship to that
proposed by Bond for the design and optimisation of grinding
circuits [online], International Journal of Mineral Processing,
74(1–4):133–141. Available from: <https://doi.org/10.1016/j.
minpro.2003.10.002> [Accessed: 29 May 2017].
Morrell, S, 2008. A method for predicting the specific energy
requirement of comminution circuits and assessing their energy
utilisation efficiency [online], Minerals Engineering, 21(3):224–233.
Available from: <https://doi.org/10.1016/j.mineng.2007.10.001>
[Accessed: 29 May 2017].
Morrell, S, 2009. Predicting the overall specific energy requirement
of crushing, high pressure grinding roll and tumbling mill
circuits [online], Minerals Engineering, 22(6):544–549. Available
from: <https://doi.org/10.1016/j.mineng.2009.01.005>
[Accessed: 29 May 2017].
Morrell, S, 2010. Predicting the specific energy required for size reduction
of relatively coarse feeds in conventional crushers and high pressure
grinding rolls [online], Minerals Engineering, 23(2):151–153. Available
from: <https://doi.org/10.1016/j.mineng.2009.10.003> [Accessed:
29 May 2017].
Stark, S, Perkins, T and Napier-Munn, T J, 2008. JK drop weight
parameters – a statistical analysis of their accuracy and precision
and the effect on SAG mill comminution circuit design, in
Proceedings MetPlant 2008, pp 147–156 (The Australasian Institute
of Mining and Metallurgy: Melbourne).
136
we are metallurgists, not magicians
Contents
Cost-effective concentrator design
G Lane1, P Dakin2 and D Elwin3
ABSTRACT
This paper discusses the factors that contribute to the cost-effective design of a
concentrator.
Concentrator design and layout outcomes are functions of the team (engineer’s
and owner’s) participating in each particular project. The benchmark for relatively
modest projects was set in the 1980s and 1990s during the ‘gold boom’ when
numerous cost-effective plants were designed and constructed on a lump sum basis
in a very competitive market. A number of factors contribute to lack of transference
of the lessons learnt in gold plant design to concentrator design including established
paradigms in the design and layout of concentrators, lack of experience in costeffective design, operator’s preferences for flow sheet and layout and simple lack of
appreciation of the impact of plant footprint on materials quantities and resultant
capital costs.
Experiences with recent copper concentrator projects (both small and large) are
used as case studies.
INTRODUCTION
Cost-effective concentrator design is not an isolated paradigm. It needs to interface
with the project infrastructure constraints, owner’s needs, vendor’s capabilities,
constructor’s logic and operator’s and maintenance team’s preferences. However,
cost-effective design has some ‘rules of thumb’, namely:
•• keep the execution strategy and plan simple, sift the ‘baggage’ from the facts
early, have a plan (and agreed scope) and stick to it
•• minimise the number of interfaces across all parties as every interface requires
‘management’
•• invest in good equipment as it saves you money
•• reduce plant footprint as capital and operating costs increase with increasing
plant footprint.
Capital costs will escalate if:
•• scope is poorly defined and the execution strategy meanders (scope and design
are not frozen)
•• simplicity is replaced with opportunism (hope)
•• pipe rack locations are used as the basis of plant layout
•• ‘expandability’ is a necessity.
SETTING THE SCENE
1. FAusIMM, Chief Technical Officer, Ausenco
Minerals & Metals, South Brisbane Qld 4101.
Email: greg.lane@ausenco.com
2. Principal Designer, Ausenco Minerals &
Metals, South Brisbane Qld 4101.
Email: phil.dakin@ausenco.com
3. Principal Designer, Ausenco Minerals
& Metals, South Brisbane Qld 4101.
Email: derek.elwin@ausenco.com
Nirvana for the project developer and plant designer is a ‘cost-effective plant’ that
meets expectations. There are many approaches to plant design that range from the
grandiose to the shoddy and mean. ‘Cost-effectiveness’ is a different paradigm. It
relies on sound judgement and a balanced assessment of what is required for the
circumstance. The ‘art’ of designing cost-effective carbon-in-leach (CIL) plants was
mastered in the late 1980s by Australian engineering companies who competed on a
lump sum turn key basis for plants in the 100 kt/a to 5 Mt/a throughput range. The
need for these plants arose through the development of CIL technology, the plethora
of modest grade opportunities in Western Australia in particular, the high gold price
and the flexible nature of gold metal marketing (Close, 2002). These circumstances
resulted in the need for innovation, short project schedules, low project capital costs
and a ‘money making machine’ approach. On occasions the commercial and design
‘recipe’ came unstuck. Notable projects, such as Three Mile Hill, initially failed to
meet throughput targets due, in part, to a ‘one size fits all’ approach and insufficient
test work to define the design.
137
G Lane, P Dakin and D Elwin
The lessons learnt from the ‘cost-effective’ design examples
of the Australian gold industry in the 1980s and 1990s are still
the basis of sound CIL plant design although the frequency of
projects in Australia is now low and there is greater focus on
Africa. The ‘art’ is practiced by some of the same individuals
and mostly in small to medium sized engineering companies.
Interestingly, as the size of the engineering company increases
the ‘art’ is diluted by other commercial imperatives. Larger
engineering businesses target Tier 1 mining clients. These
clients may express interest in cost-effective engineering but
the demands of their business, the number of interfaces and
the physical size of the projects makes meeting the aspirations
much more difficult. Tier 1 companies generally have long life
projects where the desire for durability, flexibility, life cycle
optimisation from the outset override the urgency and design
practicalities often associated with smaller projects.
‘Value engineering’ is often used to trim scope and cost
but the fundamentals and paradigm of the project are often
immovable or have a large coefficient of restitution (or
resistance) due to complex standards, systems and approvals
processes.
Cost-effective design needs good engineers and designs
and motivation from the client. For their six large copper
concentrators, Xstata has opted for a ‘standard concentrator’
approach as a method of reducing design costs. Another
project in South America opted for a ‘low cost’ design.
However, the design failed to meet many of the ‘rules of
thumb’ and the plant is struggling to meet throughput.
Another copper concentrator currently in design has had an
interesting history where the initial ‘high value centre’ design
failed to deliver a cost-effective outcome. Subsequent external
review raised a number of pertinent issues and proposed an
alternative design that went to the other extreme and was too
‘low cost’ to meet process needs. The final outcome will be a
balance between capital cost aspirations and sound design.
KEY DRIVERS
Typical concentrator layouts tend to be designed very
conservatively because designers:
•• Either don’t know how ‘close to the wind you can sail’
(fit for design) or aren’t sufficiently experienced to
understand the cost/risk/benefit relationships. This
results in the additional elevation of equipment based
on false ‘standards’ or previous practices, eg concentrate
launder slopes, setting height of mills and flotation cells.
•• Lack of innovation in design (as this requires experience
and effective risk management). This can include using
the topography and gravity instead of pumps, not using
pipe racks to dominate layouts or building close-stacked
vertical processes. Good examples are placing the
mixing tank on top of a storage tank or the motor control
centre (MCC) under the mill feed platform.
•• Tend to ‘re-invent the wheel’ instead of re-using,
improving or adapting previous proven design. Existing
design libraries aren’t well known about or published.
•• Don’t have field installation, commissioning and site
as-built experience, particularly on projects they have
designed. Designers with this experience are better
equipped to understand the basics of layout and
operability and then implement these lessons in future
designs.
•• Come from other industries, eg petrochemical or oil
and gas industries. As a result of their prior history they
have little or no relevant commissioning or operational
experience relevant to flexible but compact plant design.
138
Key drivers for a cost-effective plant design are listed below:
•• Optimise the plant footprint with the aim of reducing
concrete, structural steel, piping and electrical/control
cable and raceways. Push all areas close together,
eg grinding close to CIL/flotation, desorption and
reagents close to CIL/flotation, air /water services close
to plant, big drives close to the switchroom.
•• Keep elevation of the equipment to a workable
minimum. The elevation of run-of-mine bins, mills,
cyclones, flotation cells and thickeners are key drivers
(Lane et al, 2005).
•• Do not use dominating piperacks or large platform areas.
•• Have common platforms, stairs, pipe and cable ladder
supports.
•• Platework and lining should be kept lean, eg only put
wear liners in the chute areas exposed to wear and not
all internals.
•• Design with a fit for purpose attitude as though it is your
money you are spending.
•• Do not accept second best.
To progress design in a cost-effective manner the following
guidelines need to be applied (Lane and Dickie, 2009):
•• The orebody and its mineralogy, geometallurgy and
process responses need to be sufficiently understood to
allow process and market risks to be managed effectively.
•• Process flow sheets need to be ‘signed off’ in the first few
weeks of design. Any changes result in design change
notices that cause rework and administrative churn
within the project. If the flow sheets cannot be ‘signed
off’, detailed design is not ready to commence.
•• Duty/standby equipment needs should be defined in
the process flow sheets.
•• Process instrumentation and preliminary piping
diagrams need to commence early and be frozen from
a scope perspective at the 40 per cent design complete
stage.
•• Survey and geotechnical data generally hold up progress
when finalising the location and earthworks detail. Site
survey and geotechnical studies need to be completed in
the study and front end engineering phases.
•• Client maintenance preferences and local crane
availability impact on the decisions to use overhead
cranes, tower cranes, monorails or davit cranes for
various duties. These decisions need to be made early in
the plant layout process to avoid rework in all disciplines.
•• Local weather or environmental issues may define the
need for a plant under roof, inside buildings or with
other protection. Clear definition of environmental
needs is required prior to commencement of detailed
layout design.
•• The cost of installing plant in buildings, particularly in
the typical South American style, is high. Clear definition
of client preferences is needed prior to commencement
of layout design.
•• In-country materials of construction costs need to be
understood in order to make cost-effective structural
decisions (eg concrete versus steel).
•• Project expansion requirements and timing need to
be clearly defined in the front-end engineering design
(FEED) phase.
•• Concentrate transportation methods need to be defined
in the FEED phase (eg truck, rail or donkey).
we are metallurgists, not magicians
Cost-effective concentrator design
•• Reagent delivery and on-site storage requirements need
to be defined in the FEED phase based on plant access
limitations (seasonal weather and/or other social and
environmental factors).
•• The water balance needs to be finalised by area as the
design is developed with particular focus on the storage
method (ponds versus tanks).
•• Environmental approvals need to be finalised and
permitting requirements (traffic, run-off, dust, noise,
fumes, and materials safety) need to be clearly defined
in the FEED phase.
From project management and execution perspectives the
following issues need to be considered:
•• The ‘options’ need to be considered and evaluated prior
to detailed engineering and project execution proper.
Value engineering assessments can occur during the
design process but these need to be limited to low level
issues and not matters of scope or issues material to the
schedule. Value engineering exercises to contain capital
cost after 30 per cent engineering completion mean
that the project was not set-up initially with the correct
capital and/or design expectations.
•• Critical vendor equipment certified data needs to be
expedited. Detailed design can continue without vendor
data if the team has sufficient experience to understand
the impact that vendor data can have on design. Vendor
specifications by the engineering company may need to
be prescriptive to accelerate schedule. This compliments
the use of good equipment as if this equipment is similar
to that used on other jobs the vendor information can be
more easily expedited.
•• Simplicity in approach is ‘king’. Packaging aspects of
the engineering for completion by ‘low cost engineering
centres’ can be a recipe for disaster unless the packages
are well defined and managed. Engineering needs to
be progressed to between 40 per cent and 70 per cent
complete prior to remote completion, and slightly less if
key lead engineers migrate with the packages.
•• The fabrication strategy needs to be developed cognisant
of local (to the project) capability and capacity, low
cost offshore alternatives and logistical issues such as
the consolidation of equipment and fabricated items.
Preassembled modules may provide opportunities for
some locations where on-site fabrication costs are high
or people with the requisite skills are in short supply. A
logistics study is required at an early stage.
•• Steel sections standards differ between countries and
need to be reconciled with fabricator’s norms.
•• Project manager capability ‘to support the team
to perform at maximum capability’ is a key driver
particularly in maintaining a high level of clarity from
the client interface to the drawing floor.
•• Engineering managers need to be able ‘to lock down
the scope’, understand ‘fit for purpose design’, assign
responsibility, support the leads and motivate the team.
CASE STUDIES
Introduction
Project names are not used when discussing most of the case
study examples herein. Photographs from other publications
and the public domain are used as examples to illustrate
particular design features.
Large and small concentrators present different challenges.
Small concentrators (less than say 10 Mt/a and single
concentrate) are simpler to arrange as there is typically
one crusher, stockpile, semi-autogenous grinding (SAG)
mill, ball mill, flotation train and tailings thickener. Larger
concentrators with multiple SAG and ball mills, multiple
flotation trains and large service runs demand an additional
level of complexity for maintenance access, service equipment
and service runs.
Small concentrators
In many respects, small concentrators of less than 10 Mt/a
and particularly less than 5 Mt/a are easier to layout in a costeffective manner provided that all contributors are of a like
mind.
The following examples indicate what to avoid in order to
optimise project value. Figure 1 illustrates a number of design
features that increase plant footprint.
The space between the unit process operations are for
mobile crane access and potential expansion (regrind mills).
Mobile crane access is most effective when there is no impact
on plant arrangement. If pipe racks need to be installed or
FIG 1 – An example of a possible small concentrator layout.
we are metallurgists, not magicians
139
G Lane, P Dakin and D Elwin
extended to allow adequate access, the use of mobile cranes
may not be cost-effective and alternatives such as tower or
portal cranes should be considered.
The arrangement of the unit processes in Figure 1 requires
extensive pumping of slurry between unit processes and the
installation of large pipe rack ways. These can be avoided by
thoughtful design in most circumstances.
The derivation of the Figure 1 design is interesting in that the
original definitive feasibility study (DFS) design and layout
was a typical open air design on relatively small footprint that
had small pipe racks between facilities and maintenance access
by local davits and monorails. A metallurgical review with
about 15 per cent of the engineering completed resulted in the
inclusion of flash flotation and this changed the cyclone tower
design substantially. In addition, it was decided to leave room
for a possible regrind mill and the milling facility and flotation
circuit were separated to affect this change. Flash flotation
cleaning was added when engineering was about 25 per cent
completed and installed in the location allowed for the future
regrind mill. A maintenance study was conducted at about
40 per cent engineering complete and access for mobile cranes
increased the separation between the unit processes and pipe
rack lengths increased. In addition, overhead gantry cranes
were allowed over cyclones and the primary crusher requiring
significant structural strengthening due to the local seismic
conditions. Hence, this is a good example of a project where a
design approach was not frozen and maintained throughout
the engineering design and where a series of relatively minor
modifications led to a less than optimum outcome due to their
incremental impact on the layout.
leads to a more complex decision-making process and a greater
tendency to conservatism. The approach to design can also
change to one that is driven by key engineers and designers
to one that is driven by a more over-arching approach based
on ‘proven track record’ or prior designs.
At 12 Mt/a capacity with a single train SAG and ball mills
in the grinding circuit it is relatively easy to design a costeffective concentrator (Figure 2). These projects generally
require modest size teams and can be lead effectively by a
competent engineering manager using simple engineering
systems. As projects become larger, the team grows and
the infrastructure and technical issues increase, particularly
when equipment selection considers large capacities and/or
novel design features. However, it is still possible to design
concentrators with twin train grinding circuits with up to
25 Mt/a capacity with relatively simple and compact layouts
(Figure 3 for example) if the concept is set early, agreed by
the owner and conveyed effectively to the engineering group.
There are numerous examples of different approaches to
concentrator design throughout the world. Plants with large
Large concentrators
The design of large concentrators requires a large team of
designers, often multiple parallel lines of equipment and
critical consideration of operating and maintenance activities
due to the size of the wear items, the weight of replacement
equipment and the volume of consumables. In general, this
FIG 2 – The 12 Mt/a concentrator (Lane et al, 2008).
FIG 3 – The simple layout of a 25 Mt/a twin train SABC (semi-autogenous grinding/ball/crushing) concentrator.
140
we are metallurgists, not magicians
Cost-effective concentrator design
capacities, such as at <http://www.citicpacificmining.com>,
with multiple trains of the largest grinding mills have large
footprints. The grinding area layout is at the other end of the
spectrum from the concepts promoted in this paper. The mills
are elevated and separated and as a consequence the bulk
materials quantities are high. This combined with locationrelated costs for materials leads to high capital cost outcomes.
There are reasons for the layout, particularly associated with
the mill erection process where the mills were assembled
overseas, transported to site and lifted into place using a
purpose designed system, but the proportional costs of civil
and structural works associated with this style of plant design
are significantly greater than those for smaller plants such as
in Figure 2.
The South American market benchmark for large
concentrators has been set by Bechtel, eg Los Pelambres and
La Candelaria. However, the style of these plants leads to
high capital cost due to relatively large footprints compared
to Australian counterparts such as Cadia (Staples et al, 2008).
One of the more recent concentrators constructed in South
America is illustrated at <http://www.amec.com/andacollo>
and is said to be an ‘innovative, low cost design due to the
low-grade of the copper deposit’. However, there are aspects
of the comminution circuit layout and maintenance strategy
that offer opportunities for further improvement in the context
of cost-effectiveness. For example, the use of a Tower Crane
may have allowed a significant reduction in structural steel in
the cyclone tower by allowing overhead cranes to be removed
while the grinding floor layout is ‘relatively spacious’.
CONCLUSION
There is no panacea solution but there are some key issues
to consider in the design and layout of any plant as pointed
we are metallurgists, not magicians
out in this paper. However, the critical consideration is to
give the owner(s) what they want in meeting targets, budgets
and project timing. To achieve this, it is the engineer’s role to
optimise the design within the owner’s constraints to achieve
maximum value from the project.
ACKNOWLEDGEMENTS
The authors would like to acknowledge Eddie McLean
for reviewing the paper and providing suggestions for
improvement. As well as all those engineers and designers
who have contributed to the authors experience over the
years, including those from Ausenco.
REFERENCES
Close, S E, 2002. The Great Gold Renaissance, 282 p (Surbiton and
Associates: Melbourne).
Lane, G and Dickie, M, 2009. What is required for a low cost project? in
Proceedings Project Evaluation 2009, pp 199–204 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Lane, G, Green, S, Brindley, S and McLeod, D, 2005. Design and
engineering of flotation circuits in Australia, in Proceedings
Centenary of Flotation Symposium, pp 127–140 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Lane, G, Staples, P, Dickie, M and Fleay, J, 2008. Engineering design
of concentrators in Australia, Asia and Africa – what drives
the capital cost, in Proceedings Procemin 2008, p 29 (Gecamin:
Santiago).
Staples, P, Lane, G and Messenger, P, 2008. Horses for courses –
tailoring front end design to project requirements, in Proceedings
40th Annual Meeting of the Canadian Mineral Processors, Ottawa,
(Canadian Mineral Processors: Toronto).
141
Project management
and delivery
Contents
Fatal flaws in technical due diligences
A J H Newell1
ABSTRACT
The background and the basis for conducting a technical due diligence are discussed
as well as the source information, typical studies and potential limitations. Typical
short comings including ‘show-stoppers’ are presented for the metallurgical area,
with examples. Issues found in other areas such as geology, mining, infrastructure
and financial evaluations are briefly reviewed, highlighting the input and interaction
required by metallurgists with these disciplines. A checklist is provided as a guide
to maximise the success of a due diligence as well as a check on the quality of a
Feasibility Study.
INTRODUCTION
A due diligence is primarily required when funds are sought to develop a new
project or expand an existing operation or when there is a proposed merger or
acquisition. An investor, typically a bank and occasionally a trading house and/or
the merging or acquiring entity (usually a mining company or group of companies),
will select and appoint consultants to assist in the assessment of the project or
evaluation of the target asset.
There are two types of due diligence, namely legal/financial and technical, both
equally as important and both need to be found satisfactory before an acquisition or
merger can be progressed.
Historically, most technical due diligences conducted by RPM Global (RPM) have
been for financial institutions. Today, there are more potential operators (ie mining
companies) in the mix.
LEGAL/FINANCIAL DUE DILIGENCE
In a legal/financial due diligence, the investor conducts an investigation into both the
financial and legal status of the company, in an attempt to identify any current and
legacy issues, such as financial commitments, legal and ownership arrangements of
the company and the directors. It should be noted that the desirability for a merger
or acquisition can quickly evaporate with poor findings. This type of due diligence
should be conducted prior to undertaking a technical due diligence.
As an example, a client in the industrial minerals space planned a merger with a
Chinese producer. The technical due diligence went fairly well as one would expect
with an existing operation, except that the resources were not compliant with JORC
standards. However, the legal/financial due diligence (in this case completed in
parallel with the technical due diligence) discovered that not all of the shareholders
had been disclosed and moreover that there was money owed. Needless to say, the
merger did not go ahead.
TECHNICAL DUE DILIGENCE
A technical due diligence assesses the reasonableness of the technical and economic
aspects of a project in terms of practicality, viability and risks (RPM Global, 2011).
It examines the methodologies, supporting data and underlying assumptions used
for evaluating or managing a project based on supplied information, technical and
operating data from similar projects, site visits and interviews with key project or
operating personnel.
1. MAusIMM(CP), Executive Consultant,
Processing, RPM Global, Brisbane Qld 4000.
Email: anewell@rpmglobal.com
The main purpose of a technical due diligence is to identify ‘show-stoppers’
– aspects of the project that make it impractical or unviable – and thus protect the
investor from potentially losing money by investing the project. Often the technical
due diligence uncovers a raft of assumptions or unusual methodologies – ‘devils that
you don’t know’ – that seriously undermine the project. This does not necessarily
mean that the investor would not become involved in the project, particularly when
the underlying asset has significant potential (eg good orebody); it usually encourages
a re-development of the project by the owners.
145
A J H Newell
A successful and timely technical due diligence relies
on experienced personnel with a background in project
development as well as operations and an overlap with
other disciplines. For example, a metallurgist, besides being
competent in the processing technology associated with
that commodity as well as typical metallurgical responses
and costs, would require some knowledge of geology (the
geometallurgical aspects, such as ore types, as applied in
the resource model etc) as well as mining (mainly mine
scheduling, potential for and nature of dilution etc) to assist
in a more accurate and timely assessment.
Depending upon an individuals’ experience, a metallurgist
is often involved in assessing site infrastructure, except that
power stations, tailings storage facilities and rail are the
domain of infrastructure specialists. In addition, a metallurgist
often reviews the marketing arrangements and the nature of
the smelter/buyer contracts.
Technical due diligences are typically high level and
conducted over the time span of three to four weeks, depending
upon the size of the project and the number of assets. For a
major acquisition, a high level technical due diligence may
run into eight weeks or so. Occasionally a detailed technical
due diligence may be requested, which would take at least six
months or longer.
Two types of assets are the target of a technical due
diligence. Most assets that are investigated in technical due
diligences are projects supported by a Feasibility Study.
Consequently the main technical due diligence activities are
based around examining the reasonableness of the study,
approach/methodologies, supporting data and assumptions
and thus the likelihood of the project being technically and
economically feasible. In addition, attention is paid to the
nature of the project’s potential/upside as well as whether
risks and potential mitigations have been adequately
considered.
The other asset types considered in a technical due diligence
are an existing operation or operations, often becoming
available for either merger or acquisition at either the top or
the bottom of the mining cycle. These types of technical due
diligences are in some ways easier to conduct, since the asset
has passed the Feasibility Study stage and a body of industrial
data and actual costs is available for review. Occasionally, an
operating asset under technical due diligence is planning an
expansion. The expansion, which would be generally based
on an internal study, would hopefully follow the rigour and
accuracy of a Feasibility Study.
As a final comment, a generally less intense variation on a
technical due diligence is an independent technical review
(ITR) or independent engineer’s report (IER), which are
mainly used for an initial public offering (IPO) or stock market
listing. ITR/IERs have a similar content and approach to that
of a due diligence, however, are conducted at a higher level
(ie less depth) and often take the form of a fatal flaw analysis.
The nature of a Feasibility Study is discussed in some detail
in Appendix 1 since in the assessment of a project, the quality
of a Feasibility Study has a significant bearing on the outcome
of a technical due diligence. Some typical characteristics of
non-Western feasibility studies are also presented.
When technical due diligences are conducted on nonWestern Feasibility Studies, projects need to be evaluated
within the context of the local commercial, technical and
political environment and not necessarily compared with the
standard Western approach which addresses the risks raised
by Western investors and developers.
146
A Feasibility Study, when properly developed, has
progressed through the Scoping Study and Pre-Feasibility
Study stages. When a study stage has been bypassed, it
is quickly revealed in a Technical Due Diligence; there is a
general lack of methodology in the study approach, only a
focus on selected areas, various options remain open and
unresolved, marketing studies are not as advanced as they
should be, insufficient test work to support the study level –
captured in a document that is incomplete, contains gaps and
does not flow.
It goes without saying that studies should be conducted by
competent and experienced personnel with a proven track
record, either within the company or outside (eg consulting
companies or engineering, procurement, construction
management – EPCMs). As an example, a recent due diligence
of an industrial operation wishing to expand revealed a very
poor quality Feasibility Study, conducted by a fabricator. It
was the show stopper of all show-stoppers, fatally flawed in
many ways, as highlighted later in this paper. Technically,
it lacked metallurgical detail, such as mass balances, design
criteria, flow sheets, process and instrumentation drawings
(P&IDs), design criteria etc and was based on limited test
work employing samples from another deposit. Financially,
there was no breakdown of the capital and operating costs,
while the methodology and supporting data was poor.
The main areas examined in a technical due diligence
parallel those presented in a Feasibility Study, namely:
•• Geology – is often the most important area and has
significant potential for a fatal flaw; there needs to be
a high level of certainty that an orebody exists and
supports economic extraction, allowing subsequent
recovery of any project expenditures. Some significant
flaws that can be encountered include:
•• geological interpretation and the subsequent
conversion of drilling core data into a model
•• basis for classification of resources (inferred and
indicated).
•• Mining – has potential for fatal flaws, particularly
underground mines; major shortcomings include:
•• mine design and schedule – output not in terms of
ore types and is ‘lumpy’; significant variations in ore
grades, ore type blends and volumes
•• the optimum mining rate has not been determined
•• inappropriate underground mining method
•• proposed levels of costing accuracy not achievable
•• limited supporting geotechnical and hydrogeological
studies (particularly pit slope assumptions).
•• Metallurgy – often has a multitude of minor flaws
and occasionally some fatal flaws; suitable flow sheet
and equipment with reasonable costs are the main
considerations; more details are presented in the
following section.
•• Environmental – is a potential show stopper, sufficient
studies need to have been conducted and permitting
must be in place.
•• Infrastructure – can become a significant show stopper,
particularly for larger projects in remote locations and
typically for industrial minerals; inadequate water and
power supply studies; inadequate tailings disposal
planning; trade-off studies have not been conducted;
local climatic, seasonal or seismic conditions have not
been fully considered. In addition, the lack of or limited
sterilisation drilling as well as geotechnical studies for
the locations of the processing plant and tailings storage
we are metallurgists, not magicians
Fatal flaws in technical due diligences
facility can introduce unexpected costs and significantly
delay a project or limit production (eg Escondida).
•• Social/indigenous – has potential to be a show stopper
(cf Garg Island); local support for the project needs to
demonstrated.
•• Marketing – occasionally a show stopper (eg product
quality and or volumes not accepted by market).
•• Financial evaluation – only as good as the quality and
reliability of the inputs. Note that any project valuation
(eg net present value (NPV)) must be based on at least
Indicated Reserves if a public document and for a
Feasibility Study financiers generally like to see at least
six years of reserves.
•• Project execution – potential for flaws, occasionally
fatal flaw, generally due to optimistic construction
schedules (impact of seasons and climate has not been
fully considered), equipment delivery schedules and
ramp-up times.
An issue that can initially arise when conducting a technical
due diligence is getting access to documentation and data in
a timely fashion. This information is typically provided via
an online data room through a series of access protocols. The
quality of data can vary greatly and be in unusual places; for
example, in a recent technical due diligence of a large nickel
operation, some metallurgical studies were in geological
reports – which is not a place that a metallurgist would
naturally search.
Another strategy employed by target companies is the
provision of copious quantities of data, of variable relevance
and often out of date. Often the site visit produces the critical
data sought, however due to bureaucratic protocols, there may
be delays, some of which can be a strategy employed by the
target company. Both of these approaches were experienced
in a previously mentioned technical due diligence.
While site personnel tend to be knowledgeable, helpful and
communicative, this is not always the case. A relatively recent
technical due diligence site visit to a moderately large eastern
Australian copper operation revealed that a complete turnover
of technical staff had occurred within the previous nine months
or so and that site personnel had limited detailed knowledge.
executed, then the flow sheet may be fatally compromised and,
as a consequence, the process design and associated costs.
•• Test work sample issues:
•• samples for separation test work not being representative
•• limited range of samples
comminution testing
used
to
conduct
•• suitable test work sample locations for bench scale,
locked cycle and pilot plant test work.
•• Limited mineralogical studies:
•• particularly for more complex ores and subsequent
test work.
•• Limited comminution test work:
•• none – common in Russian and Chinese work
•• only Bond ball mill work index (BBMWi) and Bond
abrasion index (Ai)
•• poor understanding of liberation requirements.
•• No preconcentration test work (potential not identified).
•• Poor understanding of metallurgical losses.
•• Limited amount of separation test work:
•• insufficient amount of test work to convincingly support
flow sheet interpretation and metallurgical response
•• bench scale flotation test work – no locked cycle test
work (LCT)
•• no reproducibility studies, particularly when results
are variable – test work and thus flow sheet decisions
being made based on one result
•• no establishment of feed grade-recovery relationships
(important for financial model).
•• Limited amount of dewatering test work:
•• insufficient body of data to convincingly size and
select dewatering equipment
•• where concentrates are produced, no transportable
moisture limit (TML) studies
•• where ‘dry stacking’ is selected for the tailings, a range
of tailings samples reflecting the likely range in the
quantity of fines needs to be tested.
•• Flow sheet:
Generic flaws, which often prove fatal, result from:
•• not finalised
•• insufficient or unavailability of supporting data
(eg geological data, metallurgical samples etc)
•• unresolved technical issues
•• insufficient test work to support flow sheet selection
•• optimistic, incorrect or poorly based interpretations
of data (eg geological data, future metal prices, metal
recoveries etc)
•• not suitable for all ore types/blends that would be
presented to the plant
•• proposed levels of accuracy not achieved (eg capital and
operating costs).
•• ‘novel’ flow sheets:
A processing technical due diligence checklist has been
provided in Appendix 2 highlighting the many elements that
need to be considered. For a project, it is also a guide to what
a successful Feasibility Study should be addressing.
PROCESSING FLAWS
The most commonly encountered processing flaws are related
to the flow sheet. The flow sheet attracts much attention
because it is the basis for the processing plant design through
the mass and water balance, design criteria, equipment sizing
and selection as well as the determination of consumable and
labour requirements and finally operating and capital costs.
A flow sheet is based on test work samples, mineralogy
and the test work program. If any of these tasks are poorly
we are metallurgists, not magicians
•• insufficient flexibility
•• not been proven on a demonstration scale (cf pilot
plant and bench scale tests).
Other areas that are often flawed include:
•• Metallurgical recoveries:
•• limited understanding of sample or test work error
(reproducibility)
•• no establishment of feed grade-recovery relationships
•• no allowance for scale-up.
•• Plant design:
•• design criteria does not allow for the full range of
processing requirements over the life-of-mine (LOM)
and may require a later expansion
•• equipment selection and sizing unsatisfactory and not
well supported
147
A J H Newell
•• limited or lack of modelling particularly for milling
circuits; eg transfer sizing (semi-autogenous grinding
(SAB) / ball mill circuits).
•• Operating costs:
•• consumable estimates not based on first principles
(eg test work or simulation results), however not
always an issue, benchmarking may be acceptable
•• no current consumable unit cost quotations including
power
•• proposed levels of costing accuracy not achieved
•• no personnel list (employees and on-site contractors).
‘SHOW STOPPER’ PROCESSING CASE STUDIES
Three case studies are presented based on process development
issues concerning flow sheet design and equipment selection.
Each technical due diligences found a ‘show stopper’ that
was caused by the equipment vendor being involved in test
work and the flow sheet selection. Notably, the test work was
designed around the equipment proposed to be supplied. As a
result of the due diligence findings, the interest of the investor
waned and in two cases, the projects were not developed.
The key issues found were:
•• some misrepresentation about the capability of the
equipment
•• some misinterpretation of the test work results
•• metallurgical ignorance by the project owners (no
technical competency).
Case study 1
The Chinese client (an investor without any technical
personnel) was proposing to invest in an iron ore project in
Australia (hematite property in Western Australia). Initial
test work had been conducted with known equipment
(Eriez wet high intensity magnetic separation; WHIMS) with
positive results; however, these units have high unit capital
costs, high power requirements and low unit capacities. The
equipment vendor proposed to change the equipment to
conventional drum wet magnetic separators with rare earth
magnets, which would have significantly lower unit capital
and operating costs. However, they would be unable to
generate the magnetic forces required to recover the hematite.
The proposed equipment was not tested and the metallurgy
generated with the WHIMS test work was adopted for the
proposed equipment which indicated a very robust project
value (NPV) for the project.
Case study 2
A major Western gold producer allowed geological staff
to conduct a metallurgical study without metallurgical
staff input. An equipment vendor conducted the test work
(centrifugal concentrator). While the gravity concentration
stage was investigated, the remainder of the proposed flow
sheet, namely crushing, grinding, gravity concentration, direct
smelting of gravity concentrate and cyanidation of gravity
tailings, was not tested. Furthermore, no mineralogical studies
were undertaken on feed samples or intermediate products
eg gravity concentrate. In a subsequent study, it was assumed
that it would be technically and economically possible to
direct smelt the gravity concentrate with 100 per cent gold
recovery. When the technical due diligence challenged this
assumption, further test work was undertaken and it was
discovered that the gravity concentrate:
148
•• was mainly pyrite, assaying only 30 to 80 g/t Au when
5000 to 8000 g/t (0.5–0.8 per cent Au) is the threshold for
direct smelting
•• could not be easily upgraded
•• had poor gold recovery downstream (low gold
recoveries found with intensive leaching).
Case study 3
The project assessed was a multibase metal deposit (Cu-PbZn-Ag-Au) in Peru. The gravity concentration vendor had
taken over the test work program and not surprisingly, the
proposed flow sheet had eight stages of gravity separation,
some differential flotation and no tailings stream. Needless
to say, this flow sheet was fatally flawed and completely
undermined the Feasibility Study.
CONCLUSIONS
Technical due diligences are often challenging experiences
that generally reveal flaws, sometimes fatal, that justify the
expense of conducting such investigations. ‘Forearmed is
forewarned’ and identifying fatal flaws or ‘show-stoppers’
provides the investor or potential owner with the opportunity
to either walk away from the project or become involved
by applying a new approach to the project development or
operation.
While it is difficult to predict general fatal flaws or ‘showstoppers’ for specific projects, they typically arise from:
•• insufficient or lack of supporting data
•• optimistic, incorrect or poorly based interpretations of
data
•• poor costing methodologies
•• proposed levels of study accuracy not being achieved.
For processing, when assessing a project, the primary source
of potentially fatal flaws is associated with the development
and interpretation of the flow sheet, which is in turn based on
the nature of samples and the adequacy of test work.
Process design including equipment selection and sizing
and the development of capital and operating costs represent
two other areas where potential flaws may be found.
It is important that technically competent people manage
the test work program and subsequent process design and
study.
In the case of existing operations, the availability of records
and the ability to inspect the operation makes conducting a
technical due diligence relatively straightforward. Potential
flaws may arise through current or potential bottlenecks, the
quality of operational staff, the condition of the plant and the
ability of the plant to handle future ore types.
ACKNOWLEDGEMENTS
The author wishes to acknowledge assistance and input
from RPM colleagues namely Dick Addison and Don Larsen
(Principal Metallurgical Engineers, Denver) as well as Bob
Denis (Executive Geological Consultant, Brisbane).
REFERENCES
Pincock (now RPM Global), 2011. Independent engineer/due
diligence reviews [online], Perspectives 111. Available from:
<http://www.rpmglobal.com/wp-content/uploads/2015/08/
Issue111-IndependentEngineer.pdf>.
we are metallurgists, not magicians
Fatal flaws in technical due diligences
Pincock (now RPM Global), 2015. Minimum study requirements update
[online]. Available from: <http://rpmglobal.com/wp-content/
uploads/2015/08/Issue128-UPDATE-Minimum-EngineeringStudy-Requirements.pdf>.
APPENDIX 1 – FEASIBILITY STUDY
A Feasibility Study, whether claimed to be Bankable,
Definitive or some other rarely met descriptor, is the
culmination of a series of studies, typically starting with
Conceptual and progressing through Scoping and PreFeasibility. The intent of the study process is to decrease the
project risk with increasingly thorough technical studies that
define an optimal project development with accurate costings
and financial evaluation.
Note that the major mining houses, which typically conduct
these studies in-house, use different terminologies for
these study phases however employ the same processes in
determining the optimum project solution. Some companies
(eg Vale) have a preference for the front end engineering
design, which typically involves three front end loading (FEL)
stages (FEL 1, FEL 2 and FEL 3). This approach has a primary
focus on establishing the engineering design and subsequent
capital and operating costs; it relies on a body of information
(eg metallurgical test work) of Feasibility Study standard
that needs to be completed before the preliminary study
(FEL 1) commences. FEL 1 typically consists of a mass, water
and energy balance while FEL 2 prepares the preliminary
equipment design, layout, schedule and cost estimates. FEL 3
is more akin to the final design phase (the EPCM stage in
Figure 1), where major equipment specifications are prepared,
definitive estimates are prepared as well as electrical
equipment, line and instrument lists, preliminary 3D model
and project execution plan. Unlike a Feasibility Study, where
all of the project information is located in one document, this
body of information is separate to the FEL study process and
generally not readily accessible.
For the assessment of projects, Feasibility Studies are the
main source of information while supplemented by site visits,
other documentation as well as meetings with managers
and subject experts, the outcome of a technical due diligence
depends heavily on the quality and accuracy of the associated
Feasibility Study.
Table 1 presents a summary of the study progress, showing
the decreasing level of risk and increasing level of detail
and accuracy by decreasing the number of options through
trade-off studies, particularly in mining, processing and
infrastructure.
As in a technical due diligence, a Feasibility Study address
each major study discipline, that is geology, mining,
processing, infrastructure, social/indigenous, environmental
and financial analysis, except that in a technical due diligence,
it tends to be undertaken by individual subject experts rather
than a team.
The relationship between studies, with a metallurgical
flavour, and the project development cycle is presented in
Figure 1.
A number of projects experience problems when a study
stage is bypassed, for example jumping from a Scoping Study
to Feasibility Study or from a Conceptual Study to a PreFeasibility Study. This occurrence lies almost exclusively in the
domain of junior mining companies and appears to occur due
to a combination of inexperience, attempts to save money and
decrease the project development period. In fact it rarely does
and often substantially increases both costs and the chances of
FIG 1 – Relationship between project studies and project development cycle.
we are metallurgists, not magicians
149
A J H Newell
TABLE 1
Project studies.
Project Stage
Conceptual Study
Scoping Study
Pre-Feasibility Study
Feasibility Study
EPCM
Activity
Examine concept
Examine options
Trade-off studies
One option
Construction
Resource
Internal
Internal/consultants
Internal/consultants/EPCM contractor
Internal/consultants/EPCM contractor
EPCM contractor
Basis
Comparable projects
Experience
Benchmarked costs
Database
Probable flow sheet
Estimate based on data from similar
projects
Some cost factoring
Fixed flow sheet
Layout drawings
Preliminary process and information
drawings
Some estimates based on similar
projects
Budget quotations (key equipment and
consumables and contract rates)
20% of the engineering
Engineering drawings
Process and information drawings
Material take-offs
Firm quotations (most equipment, all
consumables and contract rates)
100% of the
engineering
Engineering drawings
Material take-offs
Contracts
±25
±15
±5
Capex accuracy
±100
±50
Contingency
±50
±20
±15
±10
±10
Opex accuracy
±50
±40
±20 to ±25
±10 to ±15
±5
Study/activity costs
12.5–25 k
125–250 k
625–1500 k
2.5–5 million
25–50 million
Duration
4 weeks
3–6 months
6–12 months
12–18 months
18–36 months
Geology
Limited drilling
Geological understanding
More drilling
Develop resource model
Ore types
Inferred Resources
Geotechnical and hydrological studies
Infill drilling
Indicated Resources
Inferred Resources
Exploration potential
Geotechnical and hydrological studies
Infill drilling
Resource model
Condemnation drilling
Measured and Indicated Resources
Geotechnical and hydrological studies
Drilling
Continue to upgrade
resources
Mining
Desktop study
Probable production rate
Probable mining method
Equipment and labour
requirements
Typical capex and opex
Whittle Studies
Mining options
Mine scheduling studies
Capex and opex
Whittle Studies
Optimum production rate
Mine scheduling studies
Reserves
Blast fragmentation studies
Capex and opex
Mine design
Schedules
Mining fleet
Reserves
Detailed capex and opex
Prestripping
Developing audits
and drives/shafts
Purchase equipment/
hire contractor
Processing
Desktop study
Typical flow sheet and plant
Product volumes
Typical flow sheet
Mass and water balance
Typical capex and opex
Mineralogy
Limited bench-scale testing
Ore characterisation
Basic comminution data
Process options
Probable flow sheet
Mass and water balance
Equipment selection and sizing
Power and water requirements
Develop capex and opex: budget
quotations for key equipment and
consumables, labour
Bench-scale testing
Variability testing
Comminution testing
Dewatering testing
Flow sheet fixed
Mass and water balance
Equipment selection and sizing
Power and water requirements
Develop capex and opex: budget
quotations for key equipment and
consumables, labour
Pilot plant (if required)
Detailed comminution testing
Comminution and process modelling
Mass and water balance
Equipment selection and sizing
Power and water requirements
Detailed capex and opex firm
quotations: equipment, consumables
Site preparation
Source equipment/
installing
equipment/building
and commissioning
plant
Training operators
Infrastructure
Review
Outline
Infrastructure requirements: sources of
power and water
Roads, accommodation, logistics
Infrastructure options
Infrastructure design
Indicative contracts for power and
water (if appropriate)
Install infrastructure
(tailings dam)
Environmental
Regulatory requirements
Data collection and baseline studies
(usually takes two years)
Prepare and submit environmental
impact statement
Monitoring programs
Social and
Indigenous
Review
Discussions with locals
Establish needs
Develop strategy and initiate
negotiations
Continued negotiations
Signed agreements
Marketing
Desktop study
Product uses/potential for substitution
Identify market and buyers
Existing and future market size
Competitors: existing producers and
future projects
Future prices
Discussions with potential buyers
Refine study
Product samples for buyers
Negotiations
Financial analyses
Assumed values
Exchange rates
Exchange rates
Detailed NPV analysis: various scenarios
High level analysis
Discount rate
Discount rate
NPV: used to compare options (tradeoff studies) and sensitivity analysis
NPV: used to compare options (tradeoff studies) and sensitivity analysis
Preliminary
Detailed
Risk and mitigation
analysis
Identify
Off-take agreements
Review
Based on Pincock (now RPM) (2007, 2009). EPCM – engineering procurement construction management; NPV – net present value.
project failure. Moreover, this approach assumes that the project
is technically and economically viable, which undermines
the purpose of the study process. The Pre-Feasibility Study,
150
for example, serves a very important role and represents a
critical stage where the options are investigated, tested and
resolved (so-called trade-off studies), allowing a single project
we are metallurgists, not magicians
Fatal flaws in technical due diligences
development pathway to be interrogated in greater depth with
regards to costs during the Feasibility Study.
•• No comminution (also with Russian TEOs) or dewatering
test work conducted.
A problem often encountered with studies conducted by
junior mining companies is that they have been conducted
in a haphazard fashion, typically by different groups of
consultants assigned with a particular section obviously
working in isolation. This is typically driven by ‘the rush to
market’ syndrome, where attracting finance overshadows
the need for a thorough understanding and study of the
project. This approach is often characterised by the presence
of conflicting data indifferent parts of the study as well as
limited test work, general lack of supporting detail, optimistic
capital and operating cost estimates and the use of current
and unrealistic metal prices for revenue calculations.
•• Equipment sizes are generally much smaller in size and
more numerous:
Employing adjectives such as Definitive (should be by
definition) or Bankable (confirmed by a bank?) to a Feasibility
Study does sound warning bells, suggesting a marketing
rather than a study document.
•• the value of the mined material is included in the
mining operating cost, which is included as a cost in
the processing operating cost
Often significant differences in critical numbers are found in
various parts of the study and other supporting documents,
making it difficult to form an opinion. In these cases, there
would not appear to have been a study manager controlling
and integrating the overall study process, resulting in little
or no communication occurring between the disciplines.
Studies are an iterative process and require integration
through review and communication by all disciplines. For a
processing engineer, knowledge of the geological (or block)
model, the mining schedule as well as what constitutes a
marketable product are critical elements in developing and
designing a flow sheet that would successfully process LOM
feed and thus maximise the project value.
Note that there are so-called Western and non-Western type
Feasibility Studies; in most cases, non-Western Feasibility
Studies, with the general exception of the Russian TechnioEconomicheskiye Obosnovaniye (TEO), do not meet the
standards of a Western Scoping Study let alone a Western type
Pre-Feasibility Study. They tend to follow a formulaic approach
as prescribed by guidelines held by the Design Institutes.
The main differences encountered with non-Western
Feasibility Studies are:
•• Often a geological model may not exist (except Russian
TEOs):
•• does not include ore types
•• issues with quality assurance / quality control (QA/
QC) procedures as well as assaying methods.
•• Resources are not JORC compliant; while other
methods of estimating resource are reasonable (for
example, Russian), they cannot be simply ‘converted’
into a JORC classification without undertaking a full
review, which is very difficult to do when:
•• the drill core has been destroyed or lost (cf when the
metal is classified by a country as strategic eg uranium
in the USSR or tungsten and rare earths in China)
•• the actual coordinate system used to locate drill holes
is not known.
•• Limited mine planning and lack of detailed mining
schedules.
•• Source of samples not always revealed and often some
uncertainty about representativity:
•• where samples cannot be obtained for test work
(mainly occurs earlier study phases examining
underground deposits), metallurgical responses are
based either on local, provincial or country metallurgy
for that deposit type.
we are metallurgists, not magicians
•• spiral classifiers commonly used; unusual to see
hydrocyclones
•• thickeners are rarely used.
•• Marketable product grades are lower with higher
levels of penalty elements accepted (eg 18 per cent Cu,
40 per cent Zn and 55 per cent Pb).
•• Rarely issues with infrastructure – sources power and
water are typically provided by the state.
•• Different methodology for calculating operating costs:
•• depreciation is included as an operating cost which
is normally considered as a separate cost in typical
Western financial analyses
•• product transport costs for projects located in
development regions are typically subsidised by the
buyer.
A Feasibility Study considers the following disciplines,
which parallel the technical due diligence assessment process:
•• Geology – ownership of tenures; regional and deposit
descriptions; geological or block model; data acquisition
(drilling methods, core recovery, logging, sampling,
assaying, twinning etc); density determinations; quality
of the resource estimate (inferred and indicated – data
validation, modelling method, QA/QC, interpretation
(‘geo-imaginitus’), grade/tonnage relationships etc),
exploration history and upside.
•• Mining – open cut relatively straightforward,
underground more complicated and depends
upon the deposit style (cf thin veins and need for
backfilling), mining permits, mining design, waste
dumps and stockpiles, trade-off studies, mine schedule
and production plans, pit wall slope, rock stability
and strength (geotechnical studies), dewatering
requirements (hydrogeology), equipment, capital and
operating costs (methodology, budget quotations,
quantity take-offs, consumable and personnel estimates,
accuracy, contingencies, equipment/services sources
(currency), transport and insurance, dewatering and
reserves (proven and probable)).
•• Metallurgy – location and nature of samples, test work
program, mineralogy, preconcentration, comminution
parameters, separation and dewatering; water
quality; technical management of waste streams
(eg detoxification); flow sheet; mass and water balances;
design criteria; equipment selection, sizing and list;
comminution and preferably flow sheet modelling;
process description; drawings – general arrangement,
process flow diagram (PFD), P&IDs and one-line
electrical; consumable requirements (power, water,
labour, milling media etc); capital and operating costs
(methodology, budget quotations, quantity take-offs,
factoring, accuracy, contingencies, mobile equipment,
equipment/services sources (currency), transport and
insurance etc); product schedule.
•• Environmental – suitable and sufficient baseline studies
have been conducted; appropriate regulatory approvals,
permitting and licences eg details of environmental
impact assessment or statement (EIA/EIS); stakeholder
151
A J H Newell
engagement, environmental management systems
(EMS) and environmental/social management plans
(EMP) completed; reclamation and closure plan;
operating costs; closure costs.
•• Infrastructure – requirements/sources/distribution/
reticulation of power and water; if connected to grid
power, route map; logistics (roads, rail, airports,
ports etc – getting equipment, consumables and
personnel to site; getting product to market); buildings
(offices, warehouses, maintenance/repair shops,
accommodation); allowance for location, climate and
seismicity (buildings); construction schedule.
•• Social/indigenous – needs to demonstrate a good
relationship with local people; support for the project,
program of assistance and ongoing interaction.
•• Marketing – suitability of product and product
specifications, size of market, future of market
(competition, growth, substitution etc), product terms
(penalties, credits etc), smelting and refining terms,
future product prices, product transport route map,
arrangements and costs etc.
•• Financial evaluation – typically based on discounted cash
flow methodology, with NPV used as the measure;
only as good as the quality and reliability of the inputs:
capital and operating costs (eg estimate basis, nature of
the contingencies), future metal prices, exchange rates,
inflation rates, discount factor etc). Any NPV must be
based on at least Indicated Reserves if a public document
and for a Feasibility Study bankers like to generally see
at least six years of reserves.
•• Project execution – management plan and procedures
(owner/EPCM structure and organisation, key
personnel and qualifications, project responsibility list
(owner-managed and EPCM-managed projects), EPCM
contractor procedures (engineering, procurement,
construction and reporting etc)), implementation
schedule, list of probable contracts and facilities, likelyqualified contractor listings, equipment delivery (long
lead items, logistical plans etc).
In addition, bankers considering projects in non-Western
parts of the world also like to know that the project would
meet the Equator Principles (Equator Principles Association,
2011). These are series of guidelines adopted by financial
institutions for determining, assessing and managing
environmental and social risks.
Site inspection
Operation
•• how the plant is being operated – number of operators,
operator activity, housekeeping etc
•• condition of the plant – need for equipment replacement,
nature and magnitude of maintenance costs
•• process bottlenecks
•• review sample collection including final concentrates,
assay facility, laboratory (routine and future ore testing).
Future plans – ore sources, grades, ore types, mineralogy etc.
•• Will future ores have different characteristics or grades
that will affect the metallurgical performance in terms:
•• throughput
•• recoveries
•• product grades
•• What procedures are in place to measure and record any
changes in ore types? For instance:
•• weekly testing on site
•• off-site testing
•• frequency
•• How would the current flow sheet and equipment
handle any changes in ore characteristics?
•• Any potential bottlenecks?
•• increased hardness – comminution circuits
•• grade changes – separation and dewatering circuits
•• Flow sheet changes required?
•• potential costs
•• downtime for installation
•• Equipment changes or upgrades required?
•• potential costs
•• downtime for installation.
Future water and power supplies – quantities, availability,
supply agreements.
Reconciliation between mine and plant
Tailings Storage Facility (TSF) capacity – need for another TSF.
Future costs – basis for sustaining capital, any capital
upgrades/expansions, supporting studies, production and
revenue forecasts etc.
Project (largely based on a Feasibility Study)
Nature of Feasibility Study
•• internal or external?
REFERENCES
Equator Principles Association, 2011. Equator Principles [online].
Available from: <http://www.equator-principles.com/>.
Pincock (now RPM), 2007. Minimum report contents for engineering
studies, Perspectives, 70 (December).
Pincock (now RPM), 2009. Minimum engineering study requirements,
Perspectives, 95 (March).
APPENDIX 2 – PROCESSING DUE DILIGENCE CHECKLIST
Existing operation
Historical and current records – production, operating costs,
recoveries, product grades etc.
152
•• where they suitably competent to conduct or manage
the Feasibility Study?
Ore types
•• have they been identified?
•• reasonableness of classification (degree of oxidation,
degree of alteration, mineral gangue ratios
(eg chalcopyrite/pyrite ratio), mineral/mineral ratios,
impurity or gangue mineral (eg As level))
•• included the geological model (and thus the mine
schedule)?
Samples – nature of samples and composites: number, type
(drill core, air core etc), location, representative in nature,
mining year etc.
we are metallurgists, not magicians
Fatal flaws in technical due diligences
Mineralogy – conducted for all ore types and composites;
identified economic and impurity types, associations, grain
size and rock types; degree of liberation.
Test work program
•• undertaken by reputable and independent test
work provider (cf equipment vendor – under some
circumstances)
•• conventional or ‘novel’ or using ‘novel’ equipment
•• if ‘novel’, a pilot plant or pilot scale test and
demonstration plant required
•• has addressed the obvious flow sheet options
•• sufficient test work has been conducted on an
appropriate range and number of samples
•• has preconcentration been examined
•• has site water or synthetic water been used in test work
•• ore variability test work been conducted
•• material handling – flow, rill angles, abrasiveness, slurry
viscosity etc
•• other characteristics: oxygen demand etc
•• feed grade-recovery relationship been established for
each ore type and mining composite
•• has the product been satisfactorily characterised in
terms of:
•• source for values used in calculations (eg ore, water,
product and intermediate stream specific gravity
(SGs), mass and metal stage recoveries).
•• sufficient flexibility to handle potential
variations (‘average’ and ‘design’ cases).
process
Metal losses
•• metal losses for each ore type has been satisfactorily
identified:
•• addressed the potential for improved recoveries or
concentrate grades.
Metal recoveries and product grades
•• demonstrated as a function of ore types, mining
composite types and head grades
•• full analysis of final concentrate (credit and penalty
elements and levels identified).
Design criteria
•• basis and source for selection of Design Criteria; is it
reasonable and supported?
•• transfer sizing for SAG – ball mill circuits
•• circulating loads
•• product and intermediate product per cent solids
•• full assay – credit and penalty elements
•• media and reagent consumption rates
•• dewatering requirements and equipment requirements
•• scale-up factors used to estimate residence time
(eg conditioning, flotation, heap leaching etc).
•• if a concentrate, TML
•• other properties such as size range, friability etc.
•• have the tailings been satisfactorily characterised for
range of ore types and blends in terms of:
•• mineralogical characteristics:
•• potential for pollution and acid generation.
•• dewatering properties
•• backfill and ‘dry stacking’ (where the size range,
particularly the amount of fines, is important).
•• environmental studies such as detoxification of process
water for release into the environment.
Flow sheet
•• basis for and reasonableness of flow sheet selection
•• conventional or ‘novel’
•• unnecessarily complex
•• appropriate use of surge capacity within the flow sheet
(eg stockpiles, bins or holding tanks with sufficient
residence to decouple critical parts of the flow sheet)
•• potential bottlenecks.
Modelling
•• conducted in a satisfactory fashion for the comminution
circuit:
•• basis for circuit and equipment sizing and selection?
•• identified behaviour and impact upon design criteria
of different ore types and mine composites.
•• where appropriate, for the separation circuit.
Mass, water and energy balance
•• basis: optimum mining rate study, flow sheet, other data
or sources:
we are metallurgists, not magicians
•• addresses needs of climatic and operating culture
requirements (eg cold climate, regulated holiday
periods etc).
Equipment
•• basis for selection, number and sizing: reasonable
•• equipment source
•• equipment vendor: reputable, critical spares identified,
process and equipment warranties, availability for
commissioning
•• duty specifications: reasonable.
Mine schedule
•• an optimum mining rate has been determined:
•• basis: process capital and operating costs used
•• limited by market.
•• includes ore types and thus mining composites.
Water
•• water quality:
•• impact upon separation and final product:
•• need to wash final product
•• need to produce pure water.
•• fresh and process water circuits:
•• complexity
•• adequate storage.
•• recovered water: assumptions, recovery from TSF,
allowances for evaporation losses etc
•• water sources available to sustainability supply
processing water requirements (including mine water).
153
A J H Newell
Power requirements
•• basis (installed equipment, typical operating load,
simulations etc).
Trade-off studies
•• have been applied in a satisfactory manner to select
between processing, flow sheet or equipment options:
•• for example, ‘dry’ stacking, finer grind etc
•• comminution circuits.
Product schedule
•• basis:
•• appropriate mine schedule
•• reasonable metal recoveries
•• achievable product grades
•• varies appropriately with ore types or mine blends.
Process description
•• cost elements:
•• quantities reasonable: consumables, power, water,
labour schedule
•• unit prices reasonable: with quotations for key
equipment; based on data from similar, current projects
•• EPCM costs – reasonable for the location and
complexity of the project
•• what has been factored: piping, installation,
commissioning, first fill etc – reasonable and supported.
•• basis for determination of sustaining capital costs and is
it reasonable?
•• contingency: appropriate for commodity,
complexity and country location.
project
Marketing
•• has a marketing study been conducted?
•• suitable standard?
•• has it satisfactorily identified:
•• current market, uses and growth
•• satisfactorily describes the process flow sheet.
•• potential buyers and actual market size for
proposed product quality and volumes
Operating cost
•• methodology/basis: reasonableness (benchmark against
similar projects in same operating environment)
•• competition:
•• other operations (expansions) or projects
starting up
•• accuracy (typically ±10 per cent feasibility study (FS)):
has it been achieved or demonstrated?
•• potential substitution.
•• cost elements:
•• what has been included?
•• product specifications and terms (credits, penalties,
size ranges, moisture content, other properties)
•• quantities reasonable: consumable, labour schedule
•• price history and future prices.
•• unit prices reasonable with quotations for key
consumables including power
•• labour costs: salaries and on-cost reasonable
•• maintenance cost basis: factoring, reasonable.
Capital cost
•• methodology/basis: reasonableness (benchmark against
similar projects in same operating environment)
•• accuracy (typically ±10 per cent FS): has it been achieved
or demonstrated?
154
Risks
•• has a risk analysis been conducted?
•• how does it assess the degree of risk?
•• does it identify and rank all potential risks?
•• does it provide any recommended mitigating measures?
•• does it analyse residual risk (after application of the
mitigation)?
we are metallurgists, not magicians
Contents
Guidelines for mineral process
plant development studies
P R Whincup1
ABSTRACT
This paper presents guidelines for studies required for the development of mineral
processing facilities from initial feasibility studies through to commissioning. Mining
project schedule and cost overruns can often be attributed to inadequate metallurgical
test work, engineering and cost estimating leading up to commitment to the project.
In some cases this may result from lack of understanding of, and commitment by
the project proponent to, the requisite metallurgical and engineering studies during
the development stages. Guidelines for metallurgical test work, process development,
engineering and estimating requirements for each stage of precommitment studies
are described together with those for the engineering phase.
INTRODUCTION
The continuing rapid rise in metal prices has resulted in an unprecedented global
increase in the number of mineral project developments. There have however been
instances where companies in their rush to exploit resources, have overlooked or
cut short some of the necessary metallurgical and processing studies necessary to
ensure that a project is properly implemented and performs in line with expectations.
Consequences have included cost and schedule overruns and less than optimal plant
performance. This has led to disaffected shareholders, non-performing loans and
involuntary and disruptive changes at board and senior management level.
Requirements for the various levels of study leading to commitment of funding
for mineral project construction have been well documented (White, 2001; Noort
and Adams, 2006; Cusworth, 1993; Warren, 1991). This paper focuses on the mineral
processing aspects of these studies for which the outputs are:
•• throughput and recovery models as well as operating cost and capital cost
estimates for the project financial model
•• realisation cost information comprising transport cost, treatment and refining
charge (TC and RC) data including penalty element deductions and paid metal
recoveries
•• process plant operating costs to mine planners for pit shells/cut-off grade
determination and mining schedules, which are used in an iterative financial
modelling process to determine the project scale
•• flow sheet and design criteria for the process plant that provides for process
variability.
These outputs result from metallurgical test programs, engineering cost studies and
this paper provides guidelines for the study managers and project metallurgists at
each study level:
•• scoping
•• prefeasibility
•• feasibility
•• engineering.
Processing studies will usually interact closely with other studies contributing
to an assessment of project feasibility, which include mineral resource, mining,
infrastructure, environmental and marketing studies.
SCOPING STUDIES
1. FAusIMM, Whincup and Associates,
Lower Templestowe Vic 3107.
Email: whincup@optusnet.com.au
A scoping study would typically commence following an exploration success to:
•• define the range of process options
•• establish the project scale
155
P R Whincup
•• provide first pass metallurgical recoveries and ore
processing costs for resource cut-off grade estimates
•• provide first pass cost estimates for a preliminary
evaluation of the prospect.
Expenditure on extensive sampling and metallurgical test
work is usually not justified at this stage. It could be limited
to optical mineralogy followed by the minimum bench scale
test work necessary to establish indicative metallurgical
parameters and would be based on an assumed flow sheet.
Examples include an agitated cyanide leach or a roughing/
cleaning flotation test at one or two grind sizes with a typical
reagent regime. Limited comminution would be undertaken,
which may include determination of approximate work
indices using comparative methods.
Samples would typically be diamond drill hole quarter core
covering identified major mineralisation types and should
be selected in consultation with study geologists. For most
projects a sample weight of between 5 and 10 kg for each
mineralisation type should be sufficient. Scoping study test
work would typically cost US$30 000 to US$50 000 including
sample collection and freight.
Results of test work would provide the basis on which
to develop process options. For each option a strengths,
weaknesses, opportunities and threats (SWOT) analysis is
recommended as at a low level of cost certainty this may be
the only way to differentiate between options. From these
options a process route would be selected as the basis for the
scoping study.
It is important to focus on selecting a processing rate or
project scale at this time. It is surprising how often detailed
studies are attempted without serious efforts to establish the
project scale. It requires capital and operating estimates to
be conducted over a range of treatment rates for the entire
project including the mine, infrastructure and services. The
‘base case’ mining and treatment rate may be determined by:
•• Observing the best project net present value (NPV)
return over the range examined, although sometimes
the numbers are too approximate and unless there is an
obvious NPV difference or ‘step change’, this method
may be unreliable.
•• Using a rule of thumb by assuming a mine life (no less
than five years or greater than ten years) and applying
this to the expected mining inventory size. For example,
the base case treatment rate of a potential base metal
resource of 80 Mt could be 8 Mt/a based on a ten-year
project life. In this case a range of preliminary capital and
processing cost estimates at say 4, 6, 8, 10 and 12 Mt/a
would be conducted.
During the preliminary evaluation one also needs to consider
the likely price cycle of the commodity. In the above example
the life of the project would be expected to see production
through at least one base metals ‘low cycle’.
Unless capital data are available from a recent ore processing
plant of the type and capacity envisaged, preliminary capital
estimates will usually require some engineering and vendor
pricing, which would typically cost US$50 000 to US$100 000.
Capital estimates would be based on:
•• direct cost estimates for other commodities (for
example, steelwork, concrete and piping) factored
from the estimated mechanical equipment cost and/
or estimated installed electrical load; most mineral
process plant engineers will have in-house factors for
determining these, for example, percentages of the
mechanical equipment cost or $ kW-1 installed
•• indirect costs (for example, engineering, procurement
and project management) determined as percentages of
the directs total
•• from the ‘base case’ capital estimate, estimates covering
the range of treatment rates would be made by scaling
‘base case’ capital using, for example, the 6/10 rule:
Capital 2 = Capital 1 × (Rate 2/Rate 1)0.6
•• review of step changes in capital. These could be additional
costs arising from issues such as additional process lines,
change in water supply or electricity sources.
Figure 1 shows an example of the capital estimating
process for a large simple base metals mineral processing
plant covering the required range of processing rates for the
example described.
In this example a contingency of 20 per cent was allowed
and represents the lower limit of the range of contingency
allowances applicable to a properly conducted scoping study.
At this and subsequent stages of the project an enthusiastic but
inexperienced project proponent may be tempted to delete or
reduce the contingency. This is an early warning sign that a
project could be heading for cost overruns.
Benchmarking at this level of study is valuable as a check but
care needs to be taken that a comparative project and scope are
being examined. Capital estimates using this methodology are
considered to be accurate to no better than ±30 per cent.
Preliminary processing cost estimates for each treatment
rate would be produced from either current cost data from
similar operations or from first principles. The setting-up of
a processing cost model that reflects fixed and variable cost
components is recommended. Once established, the model
can then be used over a range of processing rates and refined
as the project develops.
Table 1 shows typical sources of scoping level processing
cost estimates.
For projects where a concentrate would be produced for
transport to a downstream processing facility realisation costs
must be taken into account at the scoping study stage as they
usually impact materially on project economics and resource
cut-off grade. Realisation costs include concentrate transport,
treatment and refining charges and can amount to 10–
15 per cent of the in situ ore value. Indicative transport costs
can be obtained from specialist road transport and rail freight
operators. Treatment and refining charges are available from
commodities research groups.
It is recommended that at the scoping stage a risk and
opportunity register be established and reviewed during
each subsequent study stage.
Time required for a scoping level processing study will
be dependent on availability of data; however, for planning
purposes a minimum of six months is recommended.
•• assumed flow sheet showing all major mechanical
equipment and ‘base case’ major process flows
PREFEASIBILITY STUDIES
•• preliminary layout sketches
The prefeasibility study (PFS) has three functions:
•• ‘base case’ mechanical equipment and electrical load lists
•• mechanical equipment pricing using recent pricing from
other projects or single vendor budget pricing
156
1. evaluate all process options by establishing preliminary
financials for each
2. select one or two options for more detailed cost analysis
we are metallurgists, not magicians
Guidelines for mineral process plant development studies
FIG 1 – Scoping level comparative capital estimates (US$M).
Table 1
Basis for preliminary ore processing cost estimates.
Expense element (notional, simple base metals plant)
Basis
Operating and maintenance labour
Conceptual manning schedule. Total employment costs from recent industry remuneration surveys or
similar operations. Employee related government charges can be sourced from government websites.
Grinding metal
Annual grinding mill relines cost from other similar projects or single vendor pricing.
Annual crusher relines cost from other similar projects or single vendor pricing.
Typical grinding media consumption and current pricing.
Consumables
Typical or preliminary test work consumptions and current pricing.
Maintenance materials and services
5 per cent of direct capital cost. Allowance for lubricants.
Technical services (eg assays, metallurgical consultants, audits)
Allowances based on similar projects.
Services (eg freight, engineering, other consultants)
Allowances based on similar projects.
Energy
Preliminary electrical load list, diversified load or an allowance of 35 kWh t-1 of plant throughput. For grid
power use available gazetted prices. For diesel generated power use current or recent comparable buildown-operate vendor pricing and the diesel price selected for the project.
Water
Allowances for other energy sources. Project unit cost based on a consumption of 1 kL t-1 treated.
3. refine capital and operating cost estimates, metallurgical
recoveries and concentrate quality ranges for project
financial modelling.
These objectives would be met by metallurgical test work,
and engineering and cost studies.
Test work would be aimed at providing sufficient data on
which to:
•• undertake comparative evaluations of process options
•• establish the key preliminary design criteria on which
to base the engineering work needed to upgrade capital
and processing cost estimates to prefeasibility level.
Test work will typically also produce samples for tailing
storage, environmental and marketing studies.
The source of sample material for prefeasibility test
work would be as for the scoping study test work, drill
core. However, a minimum total sample weight for each
mineralisation type of ~50 kg would be required for bench scale
we are metallurgists, not magicians
concentration or extraction testing. An additional 80–100 kg of
unbroken composite core sample material would be required
for comminution testing. Residual sample and selected test
products should be retained in storage until at least completion
of plant performance testing or abandonment of the project.
It is strongly recommended that the detailed test program
be developed well in advance of sample selection, in
consultation with the selected laboratory and with one or
more specialist metallurgical consultants to reduce the risk
of significant additional sample material and test work being
required at detailed feasibility or design stage to resolve
flow sheet uncertainties.
The scope of comminution test work on a composite sample
or, depending on variability of the lithology, a number of
individual samples of the major lithology types, should be
sufficient to establish the comminution circuit. Test work
would usually include:
157
P R Whincup
•• unconfined compressive strength (UCS)
•• Bond crushing work index (CWI)
•• Bond rod mill work index (RWI)
•• Bond ball mill work index (BWI)
•• abrasion index (AI).
SAG Mill Comminution Tests (SMC Tests®) may not be
required at this stage if there are other strong indicators that the
mineralisation would or would not be suitable for SAG milling:
•• mineralisation is not from a very competent uniform
zone or a fully oxidised clayey zone
•• UCS >180 MPa
•• BWI >20 kWh t-1
•• RWI is not significantly higher than the BWI and both
are not significantly >15 kWh t-1.
Morrell (2009) has provided guidelines for the number of
comminution samples required using classical statistical
analysis of comminution parameters starting with a minimum
of ten samples, respectively representative of each production
year, if possible, as well as guidance on selection of tests, test
equipment and modelling techniques. It is advisable to have
preliminary comminution parameters of potential ore types
benchmarked for SAG milling amenability using a specialist
comminution consultant.
Initial prefeasibility leaching, flotation, gravity and other
beneficiation test work would focus on elimination of process
options. For example, combinations of flotation, gravity and
cyanide leach tests on a copper–gold ore would be aimed
at resolving questions such as whether to include a gravity
circuit for gold removal, merits of intensive cyanide leach on a
gravity gold concentrate or cyanidation of an auriferous pyrite
flotation concentrate. Comparative capital and processing
costs may be required to identify preferred process options.
Bench-scale batch tests will usually suffice but the number
and complexity of tests required will be specific to the
mineralogy of the prospect.
Following determination of a preferred process route some
optimisation test work should be undertaken, particularly
to determine the grind/recovery relationship for major
mineralisation types, and in the case of a concentrate, the
grind/recovery/concentrate grade relationship.
A single locked cycle test and single test on reground
middling for each major mineralisation type would usually
be the limit of prefeasibility test work.
Reagent optimisation would not normally be done at this
stage unless reagent selection has potential to materially
impact project viability.
The study metallurgist should consider engaging an
independent third party to review the metallurgy and
processing aspects of PFS and subsequent studies leading up
to commitment to the project.
An allowance of at least US$70 000 to US$100 000 is
recommended for prefeasibility level test work for simple
metallurgical processes. Complex processes such as those for
refractory gold mineralisation treatment would need to be
estimated on a case by case basis but a cost within a range of
US$200 000 to US$500 000 would not be unexpected.
Engineering at prefeasibility level would usually be
undertaken by an engineering consultancy experienced in
the design type of the mineral processing facility anticipated.
For most ores any one of over 20 internationally recognised
engineers would be appropriate. Selection of the engineer
would be based on considerations of cost, relevant experience,
quality, availability of people and location.
158
Prefeasibility engineering would typically cover delivery of
an engineering and cost study covering:
•• key design criteria
•• preliminary flow sheets and piping and instrumentation
diagrams (PIDs)
•• preliminary mass balance, including a water balance
•• site selection and layout drawings
•• a limited number of preliminary general arrangement
(GA) drawings, plans and sections taking into account
safety, operability and maintainability; it is not
unusual to commence the development of 2D and,
in some instances, 3D computer-aided design (CAD)
models at this stage to provide GA and plan layouts
with sufficient detail, enabling preliminary materials
takeoffs (MTOs) for cost estimating
•• preliminary mechanical and electrical equipment lists
•• preliminary electrical load list
•• preliminary commodity pricing
•• capital cost estimate
•• processing cost estimate
•• preliminary schedule including a capital disbursement
schedule
•• study report.
Engineering design should take into account known
environmental and regulatory constraints.
Capital estimates would be typically based on:
•• mechanical and electrical equipment pricing using a
single vendor quotation
•• structural steelwork, plate work, concrete, major
piping and architectural MTOs and single vendor
written quotation
•• factored costs for other commodities shown in Figure 1,
including architectural – use of factors for prefeasibility
direct capital estimates assumes that the process plant
is typical of those from which the engineer has derived
the factors; however, if the process plant is known, for
example, to have an unusual amount of pipe work or
speciality pipe or plate work, MTOs and pricing should
be used for that commodity
•• owner’s preproduction capital from preliminary
quantities and current rates, and should contain an
allowance for spare parts based on a percentage of
mechanical equipment capital
•• estimated feasibility level metallurgical test work and
engineering costs
•• an assessment of working capital
•• other indirect costs as percentages of directs
•• a preliminary engineering and construction schedule.
It is recommended that during prefeasibility engineering a
work breakdown structure (WBS) be developed for the entire
project (including mining, infrastructure and indirect costs)
and the estimating package set-up.
Capital estimates produced for a PFS should have an overall
accuracy in the range from ±20 to ±25 per cent.
The processing cost model developed at the scoping study
stage would be updated and refined, and include:
•• a preliminary ore processing and production schedule
•• a revised manning schedule and current industry rates
applicable to the location
we are metallurgists, not magicians
Guidelines for mineral process plant development studies
•• estimates associated with on-site accommodation and
rotational travel
•• consumable costs determined using rates from test work
and current vendor pricing
•• maintenance materials as a percentage of the direct
capital cost
•• allowances for services
•• electrical energy costs based on electrical load list and
written vendor pricing
•• other energy from estimated consumption derived
from preliminary equipment vendor data, engineering
and current pricing, taking into account freight and
storage for items such as diesel fuel oil and liquefied
petroleum gas (LPG).
The processing cost model should be extended to cover
preproduction capitalised processing costs and set-up on a
quarter by quarter basis for at least four years from project
commitment and annually thereafter. The required time to
commission the process plant and ramp-up to capacity needs
to be considered at this stage to assist in assessing working
capital requirements.
Realisation costs should be updated based on:
•• product quality determined from metallurgical test work
•• preliminary transport studies including vendor budget
pricing
•• current or predicted industry treatment and refining
costs, penalties, deductions and price participation
arrangements; for smaller companies, use of a mineral
commodity marketing consultant is suggested.
At this point the financial and technical aspects of the
project are reviewed and further test work, options evaluation
and value engineering may be required before committing
to feasibility level studies. Corporate self-discipline may be
required so as not to rush into a detailed feasibility study with
significant technical issues unresolved.
A time of 8–12 months for the PFS could be assumed for
study planning purposes.
Detailed feasibility studies
At this stage of the project there is a reasonable expectation
that the project will proceed and metallurgical test work and
process plant engineering would be undertaken on this basis.
Study results may form the basis for a project funding request.
During the process plant feasibility study all design level
metallurgical test work should be completed together with
30 per cent of the engineering.
Dependent on equipment lead times, it may be prudent to
complete sufficient test work and engineering to allow ordering
of long delivery equipment (eg grinding mills) prior to project
approval. On more than one occasion urgent additional
comminution test work and grinding mill specification work
have been required after project commitment to allow the
mills to be ordered to meet the committed project schedule.
Design level metallurgical test work should be commenced
early in the study and should be scoped in consultation
with the proposed laboratories, a recognised comminution
consultant and, if applicable, a metallurgical consultant
specialising in the subject metallurgy and processing
techniques (for example, flotation).
Early in the feasibility study dedicated metallurgical
samples should be taken. Sample locations should be selected
in consultation with resource geologists and a consulting
mineralogist. Samples should include dilution waste rock.
we are metallurgists, not magicians
Morrell (personal communication, 2008) has advised that
while large diameter diamond core PQ (85 mm diameter)
size samples may be taken, use of smaller diameter core,
for example, NQ (50 mm diameter) as comminution sample
material is satisfactory. Generally few contemporary test
procedures, in particular the drop weight test that forms
the basis for the SMC Test®, make any practical use of the
information from larger rocks.
Sample weight for comminution testing would typically be
700–1000 kg for each lithological domain. Ideally the domain
should be defined in terms of comminution properties, which
may not necessarily coincide with the mineralogical domains.
The additional sample weight required for other design
level test work and variability testing is likely to be an
additional 200–500 kg per geological/mineralogical domain
if these domains are not the same as those identified by
comminution properties.
Mineralogical investigations should be conducted on
samples or specimens from each geological/mineralogical
domain before finalisation of the test program and include:
•• mineralogical examination including multiple optical
evaluations
•• mineral liberation analyser (MLA) or quantitative
evaluation of minerals by scanning electron microscopy
(QEMSCAN) bulk modal analysis.
The following two-stage approach to comminution testing
is also suggested by Morrell (2009).
In the first, limited stage, sufficient samples are tested
to carry out a statistical analysis, following which a more
extensive program is undertaken, based on results of the first.
The wider the spread of results from the first stage, the more
samples would be needed for the second. The first stage would
typically involve four to five samples from each domain.
Comminution test requirements for samples from each
domain are as for the comminution testing recommended for
PFS together with:
•• Bond AI and UCS if not included in previous test
programs.
•• JK Mineral Research Centre (JKMRC) drop weight
tests or the recently developed JK rotary breakage tests
(JKRBT).
•• SMC Test®.
•• If geotechnical core is being point load tested then
consideration could be given to having point load testing
done on comminution test samples to provide a link
between the two databases. Point load tests correlate
quite well with the SMC Test® results and hence the
geotechnical data can provide a good indication of SAG
mill competency variability.
The crushed products from the drop weight and SMC Test®
can be reused for the Bond mill work index work if sample
quantity is a problem.
Pilot scale comminution testing is not generally required
as comminution consultant databases are now sufficiently
large to preclude the need for pilot scale testing; unless
rarer circuits are being designed (eg single stage autogenous
grinding milling or high pressure grinding rolls).
Other feasibility bench scale test work is process specific
but, for example, for a large copper–gold orebody for which
treatment by flotation to produce a saleable concentrate is
proposed, the test work should include for each ore type as
a minimum:
159
P R Whincup
•• roughing and cleaning batch tests to establish baseline
flotation conditions
•• providing any necessary processing cost input data (for
example, manning schedule).
•• bench scale locked cycle tests to establish optimum grind
sizes, flotation conditions and reagent regime leading to
definition of a standard test flow sheet.
Engineering deliverables will include:
Variability testing for the recovery and throughput using
the standard test flow sheets should be undertaken:
•• by production year in which composites representing
production periods are evaluated
•• characterisation of the deposit by testing a variety of
samples representing the spatial distribution of each
mineralogical and lithological zone within the deposit.
Pilot scale beneficiation testing needs to be considered. For
simple mineralogy, and established unit processes, pilot scale
test work is usually not justified. Indicators of the need for
pilot scale testing include unusually complex mineralogy
and use of new or uncommon technology. Between a clear
case for not including pilot scale testing and clear necessity
lies a range of situations for which consideration would be
given to time, cost and risk to arrive at a decision. As a general
principle pilot scale testing hydrometallurgical processes
must be considered as issues such as penalty element buildup and side reactions leading to scaling may not be apparent
during bench scale testing.
Test work samples should be made available to equipment
vendors to enable equipment specification and pricing. These
will include, for example, settling test work for thickener
sizing and viscosity testing for pump selection. Samples of
test work residues should be retained for testing by the tailing
storage facility engineer.
The cost of feasibility level test work will vary but the
following may be taken as a general guide for a large orebody
with three domains:
Mineralogy
US$100 000
Comminution
US$250 000
Bench scale testing
US$300 000 to US$400 000
Vendor test work
US$50 000
Pilot scale test work
US$250 000 to >US$1 000 000
These allowances exclude the cost of sample collection and
freight.
Feasibility process plant engineering should be awarded
to an appropriately qualified and experienced process plant
engineer in a process where tenders are evaluated on:
•• ability to meet the scope and deliverables
•• detailed design criteria
•• detailed flow sheets
•• mass balances for both design and operating departures
•• life-of-mine ore treatment and production schedule by
ore type
•• PIDs
•• detailed site layout drawings showing site roads,
hardstand, plant service buildings and services
(consideration may need to be given at this stage of
the project to provision for future expansion of the ore
processing facilities)
•• GA and plan/section drawings taking into account
safety, constructability, operability and maintainability.
The applicable CAD 2D or 3D model would be
considerably refined and optimised from that
commenced at prefeasibility level
•• mechanical and electrical equipment lists
•• electrical load list
•• data sheets and specifications for any critical long delivery
equipment (eg grinding mills and large transformers)
•• MTOs and written quotation pricing for all commodities
•• estimated construction hours and construction labour
rates
•• detailed capital cost estimate in the WBS format to an
accuracy of no less than ±15 per cent
•• processing cost estimate to an overall accuracy of no less
than ±15 per cent
•• engineering, construction and commissioning plans
and schedules including a quarter by quarter capital
disbursement schedule
•• plant and unit process performance guarantees
•• study report.
The design criteria and mass balance should provide for a
certain amount of variability however to accommodate shortterm variability may be unjustified from a capital perspective
and this variability would be taken up in operations by
stockpile management.
The process plant production and processing cost models
should be refined and updated to include:
•• price
•• ore processing and production schedule by ore type on a
quarter by quarter basis for at least two years following
commissioning and semiannually thereafter
•• acceptability to proposed financiers (usually decided at
the prequalification stage)
•• metallurgical parameters determined from test work
(recoveries and product quality) for each ore type
•• quality of the proposed study team
•• realistic ramp-up factors (recovery, plant availability
and product quality)
•• experience in the type of facilities proposed
•• availability and timing
•• location.
The project metallurgist plays a role in the engineering and
cost study by:
•• timely provision of results of metallurgical test work
•• providing input to the plant design operating and
maintenance philosophies
•• making process related decisions
•• participating in hazard and operability (HAZOP) studies
•• initiating value engineering, if required
•• signing off key process documents
160
•• labour costs from feasibility study manning schedule
and rates agreed with operations management
•• consumables usage determined from metallurgical test
work results and process engineering (metal wear, power
and reagent consumptions) and written vendor pricing
•• maintenance materials as a percentage of the feasibility
capital
•• energy costs from feasibility engineering, electrical load
list and written vendor pricing
•• services costs (eg laboratory, freight and consultants)
from written vendor pricing.
we are metallurgists, not magicians
Guidelines for mineral process plant development studies
During the feasibility study revision of the resource model
may be required to take account of updated metallurgical
parameters and processing costs.
Realisation costs should be updated based on:
•• product quality determined from metallurgical test work
•• detailed transport studies and vendor written pricing
•• negotiated offtake agreements with product purchasers
(for example, smelters).
The updated capital, processing and realisation cost
estimates will be included in the project financial model. The
project metallurgist should be proactive in reviewing the
total model to ensure its completeness and accuracy from a
processing perspective and to reduce the risk of items being
omitted or double counted, particularly at the process plant/
mine and process plant/infrastructure interfaces.
The feasibility study should include a project risk analysis
of risks associated with the delivery and operation of the
process plant.
For planning purposes a minimum of 12 months should
be allowed for completion of processing facilities feasibility
studies.
ENGINEERING AND CONSTRUCTION
Engineering and construction is usually managed by an
engineering and construction company under, for example
an Engineering, Procurement and Construction Management
(EPCM) contract. This may include commissioning in
conjunction with the owner’s operations team.
The quality of the preceding feasibility study notwithstanding
the success of a project is very dependent on selection and
management of, in this example, the EPCM contractor. As
with the feasibility engineer, the EPCM contractor must be
selected on the ability to meet key performance criteria. A
dedicated owner’s team with experience in project delivery
(as opposed to operations experience) would be formed to
support and manage the EPCM contractor.
For the study metallurgist the engineering phase in the
development of a project commences the transition from
providing input to the final design and engineering of the
process plant to preparation for commissioning and operations.
During detailed engineering the study metallurgist will
assist with equipment selection, layout and process control
we are metallurgists, not magicians
and will review critical documents such as process design
criteria, mass balance and flow sheets and recommend these
for sign-off for construction. Some additional metallurgical
test work may be required for which sample material from
the preceding feasibility study test work would be used.
During the engineering phase facilities required for process
monitoring and control will be defined and there may be the
temptation to include all the control systems that might be
required. On the other hand, there will sometimes be pressure
by others in the owner’s team to remove control systems,
including sampling systems as the capital cost increases, often
without undertaking any value engineering. One approach to
resolve the potential conflict is to include all facilities that are
normally or typically included for the type of process plus any
for which there is demonstrated short-term economic value.
In cases where uncertainty exists, allowances should be made
in the estimate outside of the normal project contingencies
for inclusion of these items post-commissioning and where
economic benefit can be demonstrated.
At a point where ~60 per cent of engineering has been
completed, the estimate should be to an accuracy of no less than
±10 per cent, and becomes the control budget for construction.
REFERENCES
Cusworth, N, 1993. Predevelopment expenditure, in Cost Estimation
Handbook for the Australian Mining Industry (eds: M Noakes and
T Lanz) pp 252–259 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Morrell, S, 2009. Generating optimum value from ore characterisation
programs in design and geometallurgical projects associated
with comminution circuits, in Proceedings Tenth Mill Operators’
Conference, pp 167–170 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Noort, D J and Adams, C, 2006. Effective mining project management
systems, in Proceedings International Mine Management Conference,
pp 87–96 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Warren, M J, 1991. Pre-feasibility and feasibility studies: a case for
improvements, in Proceedings Mining Industry Optimisation
Conference, pp 1–11 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
White, M E, 2001. Feasibility studies-scope and accuracy, in Mineral
Resource and Ore Reserve Estimation – The AusIMM Guide to Good
Practice (ed: A C Edwards), pp 412–434 (The Australasian Institute
of Mining and Metallurgy: Melbourne).
161
Contents
Project delivery
G Lane1 and E Skinner2
ABSTRACT
Project delivery strategies are determined by project context, business case requirements,
scale of project, market conditions, owner’s capability and preference, contractor’s
experience, resource and skill availability, project location and type of project or facility.
Successful project management requires the setting of coherent, effective and
realistic business objectives and success criteria. The management team needs to lead,
prioritise, resource, align, plan, track and communicate the processes that enable the
realisation of these business objectives and success criteria.
For any given project, there is a unique balance between selection of a simple project
delivery strategy that minimises interfaces and a complex strategy that necessarily
engages the use of expertise and experience from numerous parties. The right balance
for a project between these ‘book-end cases’ is a function of how the project owner or
managing contractor defines its objectives and chooses to manage risk, cost, schedule
and project quality, in the context of the conditions affecting the project.
Selection of an effective project delivery strategy requires a detailed appreciation of
the range of external and internal environmental factors which shape and influence
the project. These factors include the project constraints imposed by the business,
market or operational environments and the selection of an effective mix of primary
(head contract), secondary (mix of vertical and horizontal packages) and tertiary
(type of individual contract) strategies. The selection of the primary, secondary and
tertiary strategies also shapes the project, the size and responsibilities of the owner’s
team and the framework for delivery of outcomes.
How does an owner or prime contractor decide? What is the optimum for a given
project? Those with operations management backgrounds will have different views
to those with project and construction management backgrounds.
Is the simplest practical approach the best approach? Each and every interface
needs to be managed.
The owner will want a very experienced team that is engaged directly with the
owner. The team will have a track record of successfully performing projects using
contractors skilled in their respective disciplines and selected using competitive
tendering on brilliantly designed scopes of work and good contract terms, with a fully
defined project scope and appropriate schedule, for which all significant risks have
been addressed and no scope changes are outstanding. Budget and schedule will be
expected to have been developed with a full understanding of the scope, situation,
risk management requirements and so on.
The problem is that these conditions rarely exist. The challenge in selection of a
project delivery strategy is to resolve how best to address the inadequacies between
the ideal case and the present reality. The mechanism for addressing the inadequacies
defines the likely outcome of the project.
INTRODUCTION
Project delivery strategies are often discussed in the context of the advantages and
disadvantages of the common primary strategies:
•• reimbursable engineering
•• procurement
•• engineering, procurement, construction and management (EPCM) versus lump
sum engineering
1. FAusIMM, Chief Technical Officer, Ausenco
Minerals & Metals, South Brisbane Qld 4101.
Email: greg.lane@ausenco.com
2. Vice President, Project Delivery, Ausenco.
Email: ed.skinner@ausenco.com
•• engineering, procurement, and construction (EPC) (for example Gabrielson,
2007; Hundertmake et al, 2008; Loots and Henchie, 2007).
These advantages and disadvantages usually revolve around the key criteria for
project delivery:
•• cost
•• schedule
163
G Lane and E Skinner
•• quality.
Notionally, EPC provides cost security, but a poorly defined
EPC scope can lead to additional claims. EPC contracts
generally impose a quantum of risk on the contractor which
may invite a significant cost premium. Poorly managed
EPC contracts may see either the owner bearing more costs,
or the contractor seeking to cut costs at risk of schedule or
quality outcomes for the owner. EPCM has the advantage of
flexibility, transparency and quality control with the potential
disadvantages of cost and schedule growth where project
definition and/or controls are inadequate.
Beneath this primary strategy level there are a multitude
of options that determine the shape and outcomes of the
project. Selection of a complex packaging and delivery
strategy in the interests of say, cost economy or community
engagement leads to an increased number of interfaces.
Increased interfaces lead to complexity for the owner and
contractors requiring higher levels of management and
control. This often provides fertile ground for interface
related claims from contractors.
A small project using proven technology and applying
a simple project delivery strategy can be managed with a
small team and simplified systems. Increased scale of project,
the introduction of one or more novel technologies or more
complex interfaces often demand corresponding modifications
to team size and competency, and the complexity of systems
required to deliver the project. Interface management often
becomes a key determinant of team size and project systems
for larger, more complex projects. Team capability needs to be
matched by trust and accountability by the team to deliver a
large project successfully.
This paper discusses the project delivery options that are
available to a project developer/owner. The various levels
of delivery strategy are summarised to provide context for
how factors that are internal and external to the project drive
strategy selection.
Finally, some case study examples of strategy selection
processes are provided to illustrate both success and where
improvement is possible.
PROJECT DELIVERY OPTIONS
Each project is unique, as are the factors affecting the
business case for the project, the market conditions predicted
to prevail during the production life cycle and the owner’s
circumstances. In forming a business case for a project,
owners need to take a long-term view with careful and
holistic consideration of all the factors affecting a financial
investment decision. These considerations may include the
following interrelated factors:
•• size and richness of the resource
•• costs of extraction of the resource
•• return on investment (ROI)
•• market for the product
•• production scale of the project
•• costs of establishment of production facilities
•• sources of labour skills for construction and operation
•• sources of construction materials and operating
consumables
•• external factors (government, community, competitors,
stakeholders)
•• time to market
•• long-term positioning of the owner’s operating cost
profile in context of the market.
164
Decisions around the factors affecting the business case
have a significant bearing on the investment decision, the
project delivery strategy, project timing requirements and
subsequent operations.
For any given project, there is a unique balance between
selection of a simple project delivery strategy that minimises
interfaces and a complex strategy that necessarily engages the
use of expertise and experience from numerous parties. The
right balance for a project between these ‘book-end cases’ is
a function of how the project owner or managing contractor
defines its objectives and chooses to manage risk, cost,
schedule and project quality, in the context of the conditions
affecting the project.
Post the business case planning phase, the commonly
applied options available to the project are summarised
below in the context of primary, secondary and tertiary
level strategies that address the head contract, management
method and delivery approach, respectively.
Primary delivery strategies
At the primary level the delivery options are:
•• Owner self-perform – the project owner’s team directly
manages and controls contractors or employees that
complete the works.
•• Augmentation strategies – the owner’s team is supported
by a specialist project delivery team to directly manage
and control contractors or employees that complete the
works. The supporting mechanism may be one of the
following:
•• Program management – adopts a holistic approach to
project delivery, with capabilities extending across
the complete project life cycle. Hence, program
management is used in the context where a contractor
takes on a broader role than just a project to deliver
multiple interrelated outcomes (Table 1). Program
management is sometimes also used to address a
series of concurrent projects which together form
a program of works (eg complex sustaining capital
works projects). By contrast, ‘project management’
usually refers to a single project which can be delivered
independently of any concurrent activities by an
owner and which for purposes of this paper, does not
extend across the project life cycle.
•• Alliance arrangements – can result from the owner
incorporating contractor’s personnel into the project
management team. True alliances result in ‘salt and
pepper’ teams where project outcomes are shared on
a pain/gain basis. Pseudo-alliances can result from
‘body-hire’ style models where the contractors risk
exposure is negligible.
•• Outsourcing strategies – the most common primary
strategies:
•• EPC (fixed price) or lump sum turnkey (LSTK)
approaches are used to transfer the bulk of the project
design and construction risk (and potential reward) to
the contractor. To be successful, the realised rewards
need to be significant for both parties and the scope
of delivery needs to be well-defined. LSTK projects
may involve extended performance warranties over
and above typical EPC delivery where the contractor
accepts responsibility for performance of the entire
project as a functional unit. EPC delivery may include
nearly all project scope of work, or subelements of the
project which can be adequately segregated and tested
for performance in isolation from the interfacing
We are metallurgists, not magicians
Project delivery
Table 1
Comparison of projects with programs.
Project management
Program management
Fixed duration
Undetermined duration, or determined across a suite of deliverables which may have imprecise delivery timeframes
Defined objectives
Defined/negotiated objectives accommodating a broader delivery framework
Task focused
Goal focused
Focused on specific task/project
Focused on multiple projects or across multiple phases in a single project life-cycle
Manager as overseer
Leader as creative thinker in addition to manager as overseer role
Single deliverable
Multiple interrelated outcomes
elements of the project. In such cases, contractors are
reluctant to accept risk of performance outside their
control. This classically arises where the feed products
from one subelement of the production chain are
required in sufficient quality, quantity and feed rate to
performance test another system.
•• EPCM (fee for service) is the most common head
contract model used in the minerals business. All
services expended to meet the scope of services
defined in the head contract are reimbursable and
performance warranties are typically diminished
when compared with EPC contracts due to increased
owner input to engineering design and construction.
EPCM delivery is generally for a scope that includes
the majority of design and construction works within
a project. Some elements may be retained by the
owner, typically those related to mining and external
authorities.
Secondary delivery strategies
Secondary level strategies relate to the method of managing
the scope of services. The packaging strategy is tied to the
primary delivery strategies and may be constrained to
respond to external factors (eg financing, equity partner
requirements, permitting, community requirements, time to
market) or internal factors (eg owner organisation, capacity,
level of project definition). Secondary level strategies can
relate to front end engineering design (FEED), organisation
of early site works, or permitting. Activities may be tied to
the packaging approach and selection between vertically
integrated packages, horizontal discipline based packages
and/or ‘EPCM self-perform’ delivery.
Vertical packages are typically for blocks of work that
are supplied as complete systems, where the contractor
is required to perform all works from the ground up to
point of handover of a functioning asset. Examples include
accommodation villages, primary crushing stations, power
stations, technology packages or other similar elements
of project that can be effectively designed, procured and
constructed as a packaged scope of work. Vertical packaging
is viable where the contractor can economically perform an
entire package of works and operate relatively independently
from other interfacing works.
Horizontal packages represent a practical solution where
quantities are large and/or scope overlaps between packages
provide a viable case for deploying contractors geared for
performance of works on a bulk scale. Examples include
site survey, bulk earthworks and drainage, roads, concrete,
steelwork, site erected tanks forming part of a broader
system, electrical bulk scopes or combination works such
as steelwork, mechanical equipment, platework and piping
(SMP) or bulk electrical and instrumentation packages (E&I).
We are metallurgists, not magicians
Project framing considerations will determine logical
selection of compensation methods. Common choices include
schedule of rates, cost plus and lump sum basis. The pros and
cons of these approaches are provided in Table 2.
As seen in Table 2, increasing time risk for lump sum
compensation arrangements relates to time investment by
the owner to achieve sufficient definition to support a lump
sum contract. Given that the lump sum contractor is accepting
cost and time risk, lump sum compensation arrangements
represent an incentive for contractors to complete works at
high levels of efficiency.
Owners and EPCM contractors can choose to ‘self-perform’
project works. This typically occurs:
•• when the owner or EPCM contractor has access to
equipment and can perform certain works economically
(eg use of a mine fleet for bulk earthworks)
•• for minor works which may not justify costs of
equipment mobilisation and demobilisation
•• where contractors have their own integrated or related
workforce
•• where contractors are unwilling or unable to accept the
risk associated with location, scope or supply.
Mixtures of the above and hybrid approaches are also
common. Table 3 illustrates the contracting approach used
by a major copper-gold project. This model included a head
contract with an EPCM contractor.
Tertiary strategies
Tertiary level strategies relate to the mix of methods of
delivery. These include:
•• design only
•• supply (and fabricate) only
•• design and supply
•• design, supply and construct (D&C)
•• construct only
•• hybrid types.
An example of the mix of delivery methods for a large
copper concentrator within an EPCM contract and using
various horizontal, vertical and EPCM self-perform delivery
models is provided in Table 4.
The selection of the primary, secondary and tertiary
strategies shape the project, the size and responsibilities of the
owner’s team and the framework for delivery of outcomes.
At the tertiary level, the owner may choose (for a range of
reasons), to combine supply and fabrication or fabrication
and installation scope, free-issue equipment and/or materials
by the owner or the head contractor, and a myriad of other
procurement, fabrication and contracting options.
165
G Lane and E Skinner
Table 2
Pros and cons of contract types.
Contract type
Cost plus
Schedule of rates
Lump sum
Positives
Can begin very quickly
Certainty on unit rates (for example, $/t steel,
$/m3 concrete etc)
Certainty of price
Can commence quickly while design is being
completed
Negatives
• no certainty of price
• no incentive by Contractor
to minimise cost or time
• owner takes the risk on quantities
• must define what each rate includes (for example,
for concrete)
• detail excavation
• formwork
• rebar
• concrete
• hold down bolts etc
Needs a higher level of engineering and
scope definition prior to contract bid
Increasing cost risk
Increasing schedule risk
Table 3
Example of contracting strategy – used for a major copper-gold project.
Area
Vertical, D&C
Fixed price
Concentrate and return water lines
Horizontal
Fixed price
Direct hire/sub-contract
Site preparation
Horizontal
Fixed price
132/33 kV substation
Vertical, D&C
Fixed price
HV distribution
Vertical, D&C
Fixed price
Administration building
Vertical, D&C
Fixed price
Store and workshop
Vertical, D&C
Fixed price
Field erected tanks
Horizontal
Fixed price
Primary crusher facility
Vertical, D&C
Fixed price
Conveyors and pebble crusher facility
Vertical, D&C
Fixed price
Concentrator area concrete
Horizontal
Fixed price
Concentrator structural steel
Horizontal
Fixed price
Mills installation
Horizontal
Fixed price
Mechanical – piping installation
Horizontal
Fixed price
Concentrator electrical and control
Horizontal
Fixed price
Tailings and water pipelines
Horizontal
Fixed price
Concentrate dewatering facility (Blayney)
Vertical, D&C
Fixed price
Process water ponds
Horizontal
Fixed price
Site access road
Self-perform
SHAPING FACTORS THAT INFLUENCE THE STRATEGY
The selection of the best delivery strategy depends on the
significant factors affecting the project and the strategic
elements related to the assembly of the selected solutions.
Early attention to shaping factors drives project stability
(Merrow, 2011). The early sharing of knowledge regarding
shaping factors and structure facilitates coherent alignment
among the project delivery team and stakeholders.
Significant factors include:
•• project business case
166
•• financing requirements
•• context and location
Type
Water storage
Tailings dam
•• market conditions (revenue side considerations)
•• external environmental variables with potential to affect
project delivery
•• market conditions (availability of owner/contractor
resources/competition for these resources, offshore versus
on-shore supply, stick build versus modular construction)
•• technical factors
•• operational needs
•• project scope
•• level of definition around basic data (process material
quality/morphology, services, geotechnical conditions,
permitting etc).
All the above can be quantified and valued as inputs into
the assembly of the delivery strategy. Additionally, several
known but unquantifiable factors may be significant in
strategy selection. ‘Unknown unknowns’ are generally
considered as part of contingency assessment for the project
and do not form part of a packaging strategy. The strategic
element in conceptualisation relates to the ability of the team
to identify the issues (and opportunities) that are impacting
and/or may impact on the project, to diagnose the causes, to
develop solution frameworks and to map these against the
available options.
On some occasions project owners/managers/directors
come to a project with experiences (‘baggage’) that
predetermine the strategy for the project. Some prefer the selfperform approach, some EPC, others favour alliances, and
still others choose hybrid approaches that have worked in the
past. Whilst experience from past projects is highly valuable, a
healthy project climate is one in which decisions are based on
facts and robust debate occurs around appropriate strategies
to address the facts.
Strategy complexity
The size of the combined owner’s and EPCM teams will be
driven by the complexity of the strategy and the accompanying
management systems and protocols.
Simple contracting strategies need a high level of clarity,
accountability and trust to be successful. The advantages
We are metallurgists, not magicians
Project delivery
Table 4
An example of a mix of contract types for a large copper concentrator.
Activity
Facility
Primary
crusher
O/L
conveyer
Stockpile
and reclaim
Grinding
Pebble
crusher
Site preparation
Flotation
Thickening
Moly plant
Concrete
handling
HV power
supply
Tailings
piping
C3
C5
Water
supply
C1
Engineering
C6
Concrete
C7
C8
Steelwork –
fabrication
C9
C10
Steelwork –
installation
C11
Site tanks
C12
Platework –
fabrication
C13
C14
C2
Platework –
installation
Mechanical
installation
C15
C4
C16
Piping – installation
C16A
Piping – spool fab
C17
Electrical and
control
C18
Major equipment
Free issue to contractors
and disadvantages of simple and complex strategies are
summarised in Table 5.
Larger (owner) companies attempt to manage risk through
complex systems and, on some occasions, complex management
structures. If not well managed, these risk management
strategies can hinder project success due to the inherent lack
of clarity and accountability, combined with complex decisionmaking and approval processes. The decision process becomes
defensive rather than balanced and accountable.
An example of simple contracting strategies working
successfully is in the West Australian goldfields where the
construction of small gold plants has predominantly been
based on simple LSTK or EPC contracts to experienced small
to mid-tier engineering companies. There were relatively
few failures in this model in a very competitive market place
(Close, 2002).
One example where an EPC contract failed to deliver to
expectations, in the same geographical area, was the Murrin
Murrin Nickel Project (Taylor, 2000) where the contractor bid
a project containing relatively novel technology on a lump
sum basis and completed engineering in several engineering
hubs. This mix of simple head contract and complex
engineering delivery, combined with the novel technology
caused the project substantial hardship in construction and
early operation.
Owner capacity and structure
Unless an owner’s business is structured for planning and
delivery of successive projects as occurs in the construction
industry and elements of the minerals and resources sectors,
several considerations arise for an owner embarking on a
development project:
•• What is the execution of the works program ‘core
business’?
•• What are the requirements of investors/partners?
•• What is the legacy requirement for the business after the
work is complete (eg completed asset or ongoing project
delivery capability)?
Table 5
Summary of advantages and disadvantages of simple and complex strategies.
Packaging Advantages
approach
Disadvantages
Simple
Owner’s team can be smaller where the level of confidence in
contractor delivery is high.
Low levels of complexity in interfaces lead to leaner control
structures.
Very successful with experienced contractors.
If the project is complex, higher levels of risk are assumed by the contractor and higher levels of
cost are implicit in contractor’s price.
A contractor who fails to perform their obligations represents a risk to the owner, which may
be difficult to address.
Not suitable for novel technologies.
Complex
Opportunity for owner to more significantly affect cost and risk
outcomes to potential advantages.
Opportunity for owner to share work where a contractor may be
under-performing.
Higher numbers of interfaces invite commercial complexity and higher levels of complexity in
control interfaces with consequent increase in project management systems and resources.
Owner needs to invest more in up front definition work to mitigate commercial exposure to
claims arising from poorly defined scope interfaces.
We are metallurgists, not magicians
167
G Lane and E Skinner
•• How well can the owner’s team perform the works?
•• What is time criticality?
•• What is the level of outsourced support required?
•• Management of delivery risk.
Secondary considerations include:
•• What organisation structure is required to manage the
project delivery and adequately deal with risks the
owner envisages could arise?
•• What technical requirements and controls should be
applied?
•• What are the key implementation risks and how should
they be managed?
•• adequacy of resources, utilities and services required for
the project
•• governance and security around achievement of the
project deliverables in accordance with the business
case, cost and time parameters
•• tranched investment structures which are tied to
achievement of certain milestones (approvals, permits,
progress), capacity of contractor/supplier market to
deliver equipment and services in context of the project
program
•• level of certainty around repayment (strength/depth
of market, probable fluctuations in revenue, defensive
measures against unforeseen events)
•• The impact of external factors on the project and how
best to manage them.
•• controls around surety of contractor performance, cash
flow and milestone payments
•• What processes, procedures, tools and systems are
required?
•• insurance conditions imposed on the project delivery
•• interface/external environmental risk management
•• Management of quality.
•• cost implications arising from key risks occurring
(sovereign risks, schedule slip, commodity movements,
labour/resource costs.
Gearing an organisation to deal with project delivery
consistently represents a substantial commitment on the
part of the owner. The sustainment of this organisational
capability between projects represents a carrying cost which
must be traded off against alternative strategies. Common
considerations in organisational planning are listed in Table 6.
Two examples of owners who have successfully delivered
self-managed projects are First Quantum and PanAust. Both
have outsourced engineering and self-managed construction.
The success of this methodology is linked to the ability of the
owner’s project team to maintain a project focus and resist
operational mindsets (Lane and Clements, 2012).
Financing structures
Capital requirements are significant for large projects. Where
an owner requires recourse to external funding for project
delivery, conditions may be imposed by equity partners or
lenders on how the project is shaped. Typically these will be
related to:
•• capacity of owner to succeed in its objectives
•• confidence in business case and factors critical to the
business case
•• off-take sales contracts secured
•• proof of concept where novel solutions or technologies
are proposed (eg prototyping, ore processing trials)
•• permitting and required approvals in place
•• community and governmental support for the project
Figures 1 and 2 illustrate the linkages between project
financing and the contracting strategy. Figure 1 compares
simple and complex financing arrangements and Figure 2
provides a simplified view of the payment interactions
between financing and contracted services and goods. As
the financing structure becomes more complex, the owner’s
ability to prescribe the contracting strategy reduces and the
financier’s influences increase.
Export credit finance may represent an attractive source of
funding for projects incorporating international contractor
content. Export credit finance arrangements are usually linked
to governmental incentives from the contractor’s country of
origin and commonly are tied to incorporation of goods and
services in the export contract from the contractor’s country
of origin. Terms are generally attractive for projects in that
the financing terms are generally low interest, long-term and
the opportunity to leverage imported goods and services may
represent an attractive alternative to locally sourced goods
and services. A disadvantage of this form of finance is the
time it takes to secure arrangements with the relevant foreign
governments via the selected contractor.
Insurance
Risk appetite varies between owners and is a function of the
owner’s capacity to absorb costs in respect of unplanned
outcomes affecting a project. The addition of equity partners
and lenders in a project financing strategy imposes additional
Table 6
Common considerations in organisational planning.
Considerations
Owner Self Perform
Integrated Team
PMC
Level of owner control
Full
Shared
Shared/devolved
High on owner controlled facilities/services,
lower on execution of works
Time to build/develop
team
Slow
Slow to moderate
Fast
Mgmt team required to overview contractor
(Owner/Integrated/PMC)
Develop in house
Build/adapt from contractor
standards
Adapt from PMC
Standards
Owner/
contractor
Owner/
contractor
Owner/
contractor
Low unless proven track
record
Depends on calibre of contractor
and team org structure
Higher
Higher
Higher
Depends on level of project
definition and package structure
Procedures
Credibility for external
investors
Cost of team
Lower
Higher
Higher
Holding cost of team
Higher
Medium
Lower (flexible pool)
168
EPC
EPCM
Construct Only
Blend of owner's mgmt team and contractors mgmt team indirects
Contractor risk
Contractor risk
Contractor risk
We are metallurgists, not magicians
Project delivery
FIG 1 – Simple and complex financing instruments.
FIG 2 – Sample payment release process incorporating banker’s engineer.
complexity around the conditions required to secure financing
for a project.
An owner will need to balance the amount of cost it bears
in context of an insurable event, against how much risk a
contractor and a third-party insurer would bear. On large
projects it is common to engage a broker for advice around a risk
management framework as package conceptualisation occurs.
This assures that contracts are structured with a considered
position around leveraging contractor standard insurance
programs as an integral part of the overall risk management
program the owner will assume. Early engagement of the
underwriter market is often cost beneficial as costs of the overall
risk management program and their effect on the project can be
tempered from the pretender period.
Owners typically need to think horizontally in terms of
phases of work (supplier works insurance, marine transit,
construction, commissioning and operations phase) and
vertically in terms of the classes of cover required to form an
adequate ‘wrap’ across the project.
We are metallurgists, not magicians
A typical project insurance scheme as used on the Goro (New
Caledonia) and Telfer (WA) projects is shown in Figure 3.
Stakeholders and local content
The owner is often developing and operating an asset which
will have a macro effect on local communities for a significant
period of time. Large projects can substantially distort factors
affecting employment opportunities, compete with or disrupt
local business and community, create demands on community
infrastructure which did not exist prior to the project, introduce
nuisance factors related to emissions/effluents, noise and dust,
competition for local resources, create disruption of cultures
and/or religious traditions, create fly-in fly-out (FIFO) cultures
where they may not have previously existed.
Conversely, the owner is also faced with needing to assure
itself of adequate access to utilities, services, transport
infrastructure and skilled/unskilled resources required to
sustainably operate the production assets.
A sensitivity analysis of the needs and wants of stakeholders
and affected communities is essential to sustainable project
169
G Lane and E Skinner
FIG 3 – Project insurance scheme.
construction and subsequent operations activities. Early
conditioning of the project to engage the affected groups in an
empathetic and constructive manner and an ongoing positive
engagement with these groups is essential to good project
governance and leads to reduced possibility of disruption and
related costs.
For the Telfer gold project in WA, care was taken to genuinely
engage local communities early through a combination of
Indigenous land user access agreements, positive engagement
strategies through a local community engagement plan
against which contractor performance was measured and
active engagement plans affording local communities early
access to information concerning opportunities to participate
in the project. For the Goro project in New Caledonia, a range
of Indigenous community groups and a small number of
monopolistic businesses heavily influenced project delivery/
contracting strategy. Additionally, regular engagement
sessions with the local community and government were
required in addition to periodic reporting on local contractor
engagement to the territorial government.
Permitting
Permitting is integral to the selection of the contracting
strategy. Permit applications in Australia typically involve
federal, state, local government and often land-owner
tiers, with various classes of construction phase permits
required. Large projects in Australia commonly require
several thousand permit applications and the processes can
be lengthy. Often traditional-owner groups will also impose
certain conditions on the project and the negotiations can be
lengthy and cost intensive. Whilst environmental and access
controls vary in jurisdictions outside Australia, the permitting
process is often of similar complexity.
It is important to consider the timing of activities for project
delivery and to periodically take stock of true project progress
(toll-gating reviews). A measured and gated approach forces
170
a process of assessment as to the true progression of key
activities and will assure that resourcing and expenditure is
matched to activities most necessary to progress the project
and keep within the cost and schedule parameters. On the
first stage of the Goro nickel project, permitting requirements
were substantially underestimated and engineering required
to support permit applications was intensive. Construction
activity at the site commenced early and the unexpected
delays in engineering delivery resulted in cost blowouts
associated with construction activity until the project was
suspended and the costly process of demobilising contractors
and re-assessment of the ‘go forward’ plan was completed.
DELIVERY STRATEGY SELECTION
How does an owner or prime contractor decide which project
delivery strategy is most appropriate? What is the optimum
for a given project?
Personal experience and personality plays a significant
role. For example, those with operations management
backgrounds will have different views from those with
project and construction management backgrounds (Lane
and Clements, 2012). Financiers want certainty of outcome
(cost and schedule).
Major mining companies tend to be protocol driven and risk
averse. Their businesses are often sufficiently large that they
have a macro effect on the regions or even countries in which
they operate. They are therefore burdened with complex
regulatory arrangements and protocols around their systems
of business, financial arrangements, community engagement,
project delivery and operations which may not apply for a
junior mining company.
Junior mining companies are often capital constrained and
more accepting of risk. By contrast with the major miners
who often have established finance facilities and large asset
bases against which to leverage credit, junior miners are more
likely to need to take a flexible approach to project financing
We are metallurgists, not magicians
Project delivery
and the concessions such arrangements may impose. Equity
partnerships are more common with the junior mining houses
and often a mix of credit and debt facilities are required.
•• the capacity of each contractor, based on shop visits and
discussion with existing clients (some local contractors
oversell their capacity)
The challenge in selection of a project delivery strategy is to
resolve how best to address the inadequacies between the ideal
case and the present reality. The mechanism for addressing the
inadequacies defines the likely outcome of the project.
•• track record for similar works
It is therefore beneficial to keep an open mind in the early
stages of project definition with regard to delivery strategy
and solutions. As project definition improves, options for
delivery become constrained to the point (usually when a
financial investment decision is made) where the project
execution plan can be finalised.
Good projects are characterised by solid basic data, high
levels of definition around the factors affecting project delivery,
a robust design with a well-defined production system and
aligned and appropriately motivated delivery resources among
the owner management and contractor teams. Projects which
are founded on a solid business case, a well-articulated delivery
plan and a skilled and experienced team which is encouraged
to work within the project plan and regularly challenge the
project to deliver on and improve the factors underlying the
business case are likely to be more successful.
As an example, consider a 5 to 8 Mt/a concentrator. An
established mining house may have sufficient income/
assets to substantially self-fund the project, perhaps with
minor assistance from banks or forward sales agreements.
As the banks would have ample surety, the owner is free to
determine the project delivery strategy which would typically
be EPCM where the owner can maintain substantial control
over the project with a modest project team, supported by the
corporate and operations team as needed.
A junior explorer would need to interest other parties
through equity arrangements, but maintaining control would
require substantial financiers’ support. This support would
be provided at increased risk and necessitate that the owner
passes as much risk as possible on to supporting contractors.
This scenario leads to LSTK or EPC contracting arrangements
involving performance warranties that mirror, in part, the
financiers’ needs (Lane et al, 2007). As well as derisking the
project for the financier, the owner’s team is smaller and the
contracting environment is simpler (if project definition is
sufficient). This approach can be extended to single sourcing
equipment and contractors where existing relationships
add value to the project (shorten schedule and increased
confidence in delivery).
•• capacity to innovate/add value for the owner
•• financial strength, including recent commercial history
•• other issues that may impact on cost and schedule such
as location, rates etc.
Contracting strategy
There are four key components to developing a viable
contracting strategy for a project:
1. understand contractor capability
2. determine the capacity and regulatory constraints to
mobilising suitable contractor resources (which may
include resources from other countries)
3. deeply understand the project
4. develop a work breakdown structure based on the above.
Local contracting capability
One of the deliverables from the feasibility study for a large
project should be a contractor capability report that clearly
defines the contracting capabilities of local fabrication and
construction contractors with respect to the following:
•• type of contractor by discipline, eg concrete, steel
fabrication, plate work, electrical etc
•• the capacity of each contractor, based on shop visits
and discussion with existing clients (local contractors
oversell their capacity)
•• the ability of the contractor to
requirements, based on shop visits
meet
QA/QC
•• financial strength, including recent commercial history.
Transport and logistics should be included in the contractor
framework.
The above data is required to assess whether/how local
contractors can support a large project and to define the
requirement for external contractors (higher cost). Note that
the use of external contractors as opposed to local contractors
needs to be justified and the basis of the justification explained.
External contractors
External contractors may be sourced from neighbouring
countries or overseas. For example, for projects in South-East
Asia, heavy steel fabrication may be sourced out of China,
Thailand, Vietnam etc.
EPCM and LSTK approaches considerations for a 5 Mt/a
project are summarised in Table 7.
Some external sourced contractors may have local offices, for
example electrical contractors, and may source their workforce
from a combination of national and international sources.
CONTRACTOR SELECTION
Once the local and external contracting opportunities have
been identified, the project manager will be able to make an
assessment of the most effective contracting strategy for the
project.
Definition of project parameters is typically undertaken
during a feasibility study. The project activities are defined on
an area and activity basis to allow a contracting breakdown
structure to be prepared. Thus, the key role of the project
director/manager is to ensure that all scope items are
identified and key issues, critical paths and required actions
are clearly defined.
CONCLUSIONS
The contracting capabilities of local and other pertinent
fabrication and construction contractors need to be defined
with respect to the following:
Successful project management requires the setting of
coherent, effective and realistic business objectives and success
criteria. The management team needs to lead, prioritise,
resource, align, plan, track and communicate the processes
that enable the realisation of these business objectives and
success criteria.
•• type of contractor by discipline, eg concrete, steel
fabrication, platework, electrical etc
Success requires a viable business case coupled with clearly
articulated project delivery strategies. The project delivery
We are metallurgists, not magicians
171
G Lane and E Skinner
Table 7
Engineering, Procurement, Construction and Management (EPCM) versus Lump Sum, Turn Key (LSTK) (self-perform) for a 5 Mt/a concentrator.
Parameter
EPCM
LSTK (self-perform)
Owner’s team size
Up to 10
Up to 5
Owner’s team capability
Sufficient to report to owner’s corporate management and
provide technical input/review
Sufficient to deal with variations in contract and monitor
adherence to scope
Contract complexity
Higher, primary, secondary and tertiary contracts and
associated interfaces
Low, only one contract if head contractor has construction
delivery capability
Suitability for simple technology
OK
OK
Suitability for novel processes
More suitable
Less suitable
Cost structure
Reimbursable
Fixed price with variations
Contractor’s fee structure
Up to 3% direct costs plus multiplier on cost of
employment
Hidden, but typically cost of employment plus 12% to 25%
of direct costs
Process/performance risk
Process risk borne by owner
Additional fees for process risk
strategies are the product of adequately funded project
definition processes with respect to basic data, external
environmental variables, process requirements, resourcing,
community requirements, capital and operating funds.
For any given project, there is a unique balance between
selection of a simple project delivery strategy that minimises
interfaces and a complex strategy that necessarily engages
the use of expertise and experience from numerous parties.
The balance between these ‘book-end cases’ is a function of
how the financier, project owner or managing contractor
defines its objectives and chooses to manage risk, cost,
schedule and project quality, in the context of the conditions
affecting the project.
An experienced project delivery team with a proven
track record on similar projects is essential for successful
delivery of large projects. Projects with the best levels of
definition and coherence of vision among the team will still
generate a wide range of unforeseeable challenges through
the execution process. Teams able to rapidly and effectively
identify and mitigate adverse conditions or risks are essential
to minimisation/avoidance of consequent delays and costs.
In general, simple contracting strategies are adequate
to deliver smaller projects of known technology and low
complexity. This approach should be extended to single
sourcing equipment and contractors where existing
relationships add value to the project (shorten schedule and
increased confidence in delivery).
Complex delivery strategies should be carefully articulated
and managed. They require higher levels of interface
definition and control and, hence, more management,
higher cost and result in greater ‘churn’. However, they do
offer the ability to include specialist (owner’s, contractor’s
and consultant’s) expertise and capabilities that can add
substantial value to the overall project.
REFERENCES
Close, S E, 2002. The Great Gold Renaissance (Surbiton and Assoc: Melb).
Gabrielson, A, 2007. Current trends in project delivery, CIM
Bulletin, 2(7).
Hundertmake, T, Valle Silva, A O and Shulman, J A, 2008. Managing
capital projects for competitive advantage, McKinsey Quarterly,
June.
Lane, G S and Clements, B, 2012. Operations versus projects – how do
people think and what are the implications? in Proceedings 11th
AusIMM Mill Operators’ Conference (The Australasian Institute of
Mining and Metallurgy: Melbourne).
172
Lane, G S, Davis, M, McLean, E J and Fleay, J, 2007. Performance testing
– when, what and how?, in Proceedings Project Evaluation Conference
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Loots, P and Henchie, N, 2007. Worlds Apart: EPC and EPCM Contracts:
Risk issues and Allocation (Mayer Brown: London).
Merrow, E W, 2011. Industrial Mega Projects, Concepts, Strategies and
Practices for Success (Wiley: USA).
Taylor, A, 2000. The outlook for the PAL process [online], in Proceedings
World Nickel Congress, Melbourne. Available from: <http://www.
altamet.com.au/wp-content/uploads/2012/12/The-Outlook-forthe-PAL-Process.pdf> [Accessed: 1 August 2017].
GLOSSARY OF TERMS
Alliance – where the owner and contractor work as a common
team to achieve performance outcomes agreed using open
and transparent decision-making processes and shared goals.
A good application of alliances arises where project definition
is poor, or where the capacity of the parties to segregate and
allocate risk is poor. In such cases, alliances seek to draw
upon the best resources of the owner and contractor teams
to achieve outcomes required for achievement of the project.
D&C – A variation on the EPC formula where
conceptualisation of the scope may be provided under a
separate contract and the contractor’s responsibilities are
limited to design finalisation and supply/construction and
usually, commissioning duties.
E&I – Electrical and Instrumentation – a typical aggregation
of construction services provided by contractors.
EPC – Engineering, Procurement and Construction – these
contracts place responsibility for delivery of the scope and
deliverables in the hands of the contractor with minimal
owner interface.
EPCM – Engineering, Procurement, Construction,
Management – delivery of scope and deliverables is placed
with the contractor however owner involvement can range
from minimal to significant levels. Where owner involvement
is significant, it can be challenging for owners to hold the EPCM
contractor to account for adverse outcomes with the project.
FEED – Front End Engineering Design – typically follows
a feasibility study as the lead in phase to project engineering.
LSTK – Lump Sum, Turn Key – such projects are generally
designed around a contractor delivering an entire production
system which must meet performance requirements across
multiple subelements.
SMP – Structural, Mechanical, Piping – a typical aggregation
of construction services provided by contractors.
We are metallurgists, not magicians
Contents
Mineral project management – a perspective
from four decades in the industry
J S Dunlop1
ABSTRACT
At its most basic level, minerals project management may be broken down into a series of
logical chronological phases, described in this paper as: exploration, discovery; scoping
and prefeasibility; feasibility and approval; project financing and commissioning;
operation; expansion and closure. Each phase has its own peculiarities and challenges,
any one of which can potentially upset the forward progress of a minerals project.
Reference is made to each of these project phases based on the author’s many years of
firsthand experience, to identify issues within the project management spectrum where
additional caution may be warranted. Left unaddressed, these issues can cause undue
delays in the project timeline, ultimately delaying approval at significant project cost.
Finally, reference is made to the Mine Manager’s Handbook (MMH), published by
the AusIMM. This is an excellent project management resource and is considered
to be very relevant for minerals treatment plant managers, whether the context be
constructing, operating, expanding or shutting down.
INTRODUCTION
We all recognise that minerals projects don’t just happen. They usually span years
from discovery to stable production, during which time many obstacles arise,
each having to be overcome whether they be technical, economic, environmental,
regulatory or various external factors. It is not surprising, therefore, that many
projects do not survive this process. Given that mineral deposits are becoming more
elusive and either deeper or harder to find, logic suggests that worthwhile projects
should not be shelved simply because they could not overcome the manageable
post-discovery hurdles.
Care therefore needs to be taken to identify the possible project ‘stoppers’ and manage
them effectively before they materialise. The concept of project phases is developed in
this paper with appropriate reference to previous publications on this subject.
The identified chronological phases of a mineral project are frequently referred
to in published literature. For example, Noort and Adams (2006) refer to project
phases in relation to the feasibility study process. Van der Merwe (2008) presented
a very similar view. In this paper, a somewhat broader view is taken, commencing
with a mineral deposit discovery and ending in mine closure. Each project phase is
summarised in the list below. They are expanded on by some personal observations
under each section which follows.
Where relevant, reference will also be made to previous MetPlant keynote addresses
on this subject, with the aim of providing continuity with some of the excellent
previous papers which touch on some of the themes discussed in this paper.
The project development phases discussed here are as follows:
•• exploration
•• discovery
•• scoping, prefeasibility and definitive feasibility
•• project approval
•• project financing and commissioning
•• operations
•• expansion
•• project closure.
1. FAusIMM(CP), Chairman, Alliance Resources
Limited, El Arish Qld 4855.
Email: jsdunlop@bigpond.com
Additional observations will be made in relation to:
•• capital markets
•• government understanding of the minerals industry.
173
J S Dunlop
EXPLORATION
Reference here to mineral exploration is made in the general
context of exploration activities carried out prior to the making
of an economic mineral discovery. More specifically, reference
here is to the exploration budget and how to manage it.
From a project management perspective, the questions often
asked are, ‘How much should we spend and over what period?’
This issue often arises at the annual budget approval process.
A theoretical approach to these questions was proposed by
Binon (1981) who described how to arrive at a budget figure
after establishing an agreed profit level and time frame.
In actual practice however, the theoretical or optimum
budget may not be available due to a shortage of funds,
especially for small capital explorers in a depressed share
market. In such a situation, the above questions remain
unanswered. This author has found it useful to start with the
same objective (deposit size, type etc) and then to assign an
acceptable maximum exploration cost, such as unit cost/oz
of precious metal or unit cost/t of economic mineral resource.
In this way, the project manager may determine the point
where acquisition may be a cheaper alternative to grass roots
exploration. For example, if the exploration budget implies a
gold discovery cost exceeding $50/oz (in the ground), it might
be cheaper to acquire already discovered mineral resources.
Unit discovery cost benchmarks are also recommended
as a management tool when testing the effectiveness of an
exploration campaign.
Exploration funding can also be managed by introducing a
joint venture (JV) partner to assume exploration costs by way
of a ‘farm-in’. This subject will be addressed later in this paper.
DISCOVERY
A mineral deposit discovery is the natural (though seldom
achieved) end point of a mineral exploration program. When
a discovery is made however, additional management issues
can arise. For example, is the deposit immediately economic or
are there remaining issues (such as metallurgical) unknowns?
Is the market fully informed (as the discovery is clearly
material)? What are the next steps?
Frequently, the market loses sight of the lead time to bring
a project to production, with the consequence that it loses
interest, further impeding the availability of capital required
for mineral resource delineation, let alone feasibility.
A useful perspective on this can be gained from a paper by
Duncan (1985), who described the development plans for the
Olympic Dam discovery at Roxby Downs, some ten years
after the initial discovery hole, (RD1) was drilled in 1975.
Another worthwhile example would be the history of the
Paddington gold mine in Western Australia, still working at the
time or writing. The early history is described by Nice (1986):
the discovery was made in 1981, the decision to develop in 1984,
based on Measured and Indicated Resources of 5.6 Mt which
later were drilled out to reveal an orebody of 8.4 Mt @ 3.2 g/t
(based on a gold price of $310). The plant was commissioned in
1985 at a throughput rate of 850 000 t/a. Being a less complex
and smaller project to Olympic Dam, the project lead time from
discovery was approximately five years, 18 months of which
was taken up by approval and construction.
This latter point is significant in the light of today’s far
slower project approval/construction timelines. It has been
this author’s experience that stakeholder expectations (not just
the public and the local community, but also the regulators,
permit issuers and engineers) need to be actively monitored
and addressed both widely and early. For this reason,
174
specialists in this area should be added to the management
team as soon as the feasibility process seems inevitable.
SCOPING, PREFEASIBILITY AND DEFINITIVE FEASIBILITY STUDIES
The feasibility stages of project development are well
understood and will be reiterated upon here. Noort and
Adams (2006) provide a very appropriate refresher which
includes orders of accuracy for each stage of the study
advancement towards final or ‘definitive’ status. Table 1
depicts the development study process, in terms of option
identification and estimation accuracy.
Aspects of the 2012 edition of the Joint Ore Reserves
Committee (JORC) Mineral Resource and Ore Reserve
reporting code (JORC, 2012) are relevant here, and readers are
directed to clauses 37–40 of the new code.
The use of discounted cash flow (DCF) methods to arrive
at a net present value (NPV) of a project during the scoping
and prefeasibility stages is common practice in the minerals
industry. If such estimates of project value are used in these
stages, it is likely that they will be based primarily on Indicated
and Inferred Resources (as was the case in the Paddington
example above). Such estimates should be for internal use
only and not for external or public reporting.
The new code requires a prefeasibility study to have been
completed before an Ore Reserve may be reported (clauses
38 and 39). It follows from this that a DCF analysis could be
reported along with such an Ore Reserve announcement. Still
being debated, however, are the conditions wherein Inferred
Resources should be included in the production schedule
underpinning the DCF model. In a previous MetPlant
keynote, Card (2011) provided extensive reference to the
AusIMM’s emerging economic modelling guidelines.
It is the author’s view that Inferred Resources only be used in
DCF-based project evaluations under the following conditions:
•• scoping studies – not at all
•• prefeasibility studies – in conjunction with (and following)
Measured and Indicated Resources, but only to assess
‘upside cases’ and not be included in the ‘base-case’
(Some stock exchanges, for example Hong Kong and
Singapore have strict rules concerning inclusion of
Inferred Resources in DCF models. Hong Kong forbids
their use entirely)
•• definitive feasibility studies (DFS) – as for prefeasibility
studies.
In the case of mineral project valuations under the VALMIN
Code (VALMIN, 2005), Inferred Resources should similarly
only be included in ‘upside cases’ and not be included in the
‘base-case’ evaluation (The code was updated in 2015, and
clarified this point, tending to support this interpretation.).
Another potential issue arising from the 2012 JORC Code
is centred on clause 51. This clause refers to the inclusion
of ‘in ground’ mineral value in public reports. Such in situ
or in ground financial valuations must not be reported
by companies in relation to Exploration Results, Mineral
Table 1
Project development phases (after Noort and Adams, 2006).
Study phase
Scoping study
Prefeasibility
Definitive FS
Iterations
Multiple
2–5
1 or 2
Accuracy
±30–50%
±20–25%
±10%
Contingency
~ 25%
~15%
~1%
Options
Full range
2–5
1–2
we are metallurgists, not magicians
Mineral project management – a perspective from four decades in the industry
Resources or deposit size. Whilst this is well overdue, a
potential problem arises in relation to compliant use of
the VALMIN Code for mineral deposit evaluation. In this
instance, ‘in ground’ mineral value is still used as one of the
rules-of-thumb, described by Lawrence (1994). This author
argues that in situ value still be retained as a valid rule-ofthumb in mineral asset valuation work, but not in the context
of mineral resource reporting under the JORC Code.
PROJECT APPROVAL
Whether a project advances to the ‘go-ahead’ stage is largely
dependent on three factors:
1. the inherent techno-economic strength of the DFS,
primarily through its financial argument
2. the completeness of the DFS, often referred to as its
‘bankability’
3. the speed of its approval by the project owners and
regulators.
These will now be discussed in the following subsections.
Feasibility studies
The content of the DFS is not usually the cause of a project
delay. The most likely causal factor is the quality of the
content. No attempt will be made here to discuss the optimal
or recommended DFS format. A detailed outline of this may
be found in the MMH (AusIMM, 2012).
Feasibility studies can be weakened by two critical factors:
excessive contingency and failure to recognise and satisfy
the critical ‘bankability’ issues. The first of these two factors
has two components:
1. excessive conservatism in ore grade, productivity and
costs
2. layer upon layer of uncontrolled contingency.
The second of these factors comes into play when the DFS
is completed without establishing achievement of critical
bankability issues (sometimes referred to as ‘fatal flaws’).
Usually these include:
Unfortunately, some benchmarks may not be known, such
as the required liquidity ratios, hedging levels and debt
service requirements of the lender. Ultimately, these gaps
may materially reduce the debt level that the project can
achieve. In a tight capital market, this factor alone may be a
potential project stopper.
Feasibility studies and joint ventures
Readers will be familiar with the very common joint venture
‘earn in’ condition, referred to in joint venture agreements
as the ‘completion of a bankable feasibility study’. Despite
what this author considers to be a professionally sound
understanding in the minerals industry as to what a BFS
requires, there has been much JV litigation resulting from
farm-in partners ‘passing off’ feasibility studies which were
claimed to be, but later shown not to be bankable.
It is therefore recommended that where JV agreements in
the future require this sort of clause, the agreement should
set out clearly the minimum standards for bankability status
to be achieved, and preferably with reference to an agreed
independent third party to arbitrate, should there be no
agreement.
Readers will be familiar with the term ‘fast-tracking’ where
feasibility study preparation, engineering and approval are
integrated to reduce project commencement delays. Often, this
situation results from a deadline of some sort to be met, such
as tenure of the project tenements or logistic or climatic issues.
An excellent account of fast-tracking was prepared by Hinz
and Aseervatham (1999), citing Rio Tinto’s experience with
its Hail Creek, Baranji and Gokwe projects. The compression
of tasks which results from fast-tracking increases the overall
risk of a successful project risk and should, in the author’s
view, be avoided if possible.
Project approval
Project approval in this context has four significant stakeholders:
1. project owners who are responsible for the equity funding
2. banks (or others) who are to provide the project debt
•• an inadequate ‘resource tail’ for the project
3. regulators (government or government agencies)
•• unproven metallurgy
4. the general public whose support for the project must
be obtained.
•• half-committed market arrangements with customers
•• a failure to recognise the minimum set of environmental
standards
•• an unclear position as to the lender’s minimum lending
terms.
For an account of project estimation conservatism leading to
a complete project re-think, readers are referred to Needham
(1985), who described in detail why the original Kidston gold
project was uneconomic until a leaner approach was taken to
every facet of the project. It was subsequently successfully built
and commissioned and operated successfully for many years.
It is suggested that detailed analysis of key project
development and operating assumptions aimed at discerning
the true base-case parameters, free of all contingency. Allocation
of the project contingency should then be added only when the
absolutely bare base-case has been attained and validated.
Metallurgists have a key role to play here, as described by
Whincup (2008), McCarthy (2002, 2011) and other previous
MetPlant keynote speakers.
Finally, the author recommends that, to avoid bankability
gaps, a list of all key bankability benchmarks (potential
fatal flaws) are established for each section of the feasibility
study, and systematically addressed as the study advances.
we are metallurgists, not magicians
It is worth remembering that the project owners cannot
approve a technically and financially sound project if their
company is not capable of raising the equity proportion of the
total project funding. Often the equity to debt ratio is only
known as negotiations for the project debt are advanced, and
in recent times, debt ratios as low as 20 to 25 per cent have
been offered as a maximum funding level.
Investment banks, which traditionally have offered project
finance, are becoming less popular since the Global Financial
Crisis (GFC) as they have become progressively more risk
averse and some would say, more aggressive towards
minerals ventures. Term sheets and loan agreements, in the
author’s view, have of late been characterised by unattractive
terms such as: high fees; intrusive security demands; tough
liquidity coverage; challenging project ratios and recourse
that places the borrowing entity’s directors at significant risk.
Compounding this funding dilemma is the emerging issue of
excessive project approval lead times. In one recent example,
approval for a NSW based gold project (Alkane’s Tomingley
gold project; despite this delay, the mine was constructed and
commissioned, achieving target head grade and production
guidance in its first two years of operation) took more than 18
months for its DFS and environmental plan to achieve final
175
J S Dunlop
approval for site earthworks. By this time, the capital cost of
the project had increased by about $25 M and the gold price
was in retreat. The subsequent ramifications were significant
as the reduced cash available for debt service (CADS) reduced
the available debt level, increased the required hedge cover
whilst reducing the overall project rate of return (IRR).
•• detailed loan agreement
•• drawdown of the loan funds
•• post-drawdown monitoring of repayments.
Project finance has lost much of its appeal in recent years for
a range of reasons, which the author considers include:
A notable description of the conventional regulatory process
appeared in 1997 (Chamberlain, Johnston and Joyce, 1997),
where the concept of stakeholder groups was identified in the
context of the approval of the Cadia mine project in NSW.
Readers interested in this subject in more detail are referred
also to chapter 4 of the MMH (AusIMM, 2012).
•• high fees
Approvals can also be overturned by regulators, thus
increasing uncertainty and project risk. This was illustrated
in 2010 when the Northern Territory government reversed
its approvals and support for the Angela Pamela uranium
project (Australian Uranium Association, 2010), citing concerns
expressed by local residents. At the time of writing, the project
remains stalled, further disadvantaged by a low uranium price.
•• operational needs.
Finally, the importance of the support of the general
public cannot be underestimated. The example above serves
to illustrate how pressure can be brought to bear on the
regulators by a range of project opponents. It is therefore
considered essential to monitor, establish and maintain
public support for new projects if potential delays are to be
avoided. The need for allocating high priority for this activity
is currently clearly reflected in the current opposition to coal
seam gas (CSG) exploration on pastoral land (no literature
reference is suggested here, as any online CSG search will
produce a plethora of argument both for and against).
PROJECT FINANCING, CONSTRUCTION AND COMMISSIONING
Project financing, construction and commissioning are given
mention here because they each play a more important role
than perhaps was the case in the past. The comments below
attempt to answer the question, ‘Why is this so?’
Project finance mix
The term project financing mix refers to the various component
parts of a total project financing package. Most typically, such
a package consists of equity and debt, though other forms of
finance are emerging as conventional debt has begun to lose
its appeal. Murray (1995) provides more background on this
theme and expands on frequently used banking terms.
Equity finance
The common forms of equity are:
•• free cash within the project company
•• cash raised by sale of existing assets
•• cash inflow from a farm-out
•• cash raised by rights issue or placement (dilutive)
•• cash raised by share purchase plan (SPP) (not dilutive).
Equity raising is vulnerable to the state of the share market
and the project company’s share price. In short, when the
capital markets dry up, it is very difficult to raise the level of
equity demanded by the lenders of debt. This is one reason
why funding is more complex than in the past, and why high
levels of project debt are becoming less attractive.
Bank finance
Conventional project debt arrangements consist of:
•• project funding term sheet
•• credit approval
176
•• high levels of security
•• onerous debt service arrangements
•• conservatism
•• inflexibility to the client’s time frame
Other forms of finance
Other forms of finance discussed here include:
•• convertible notes and bonds
•• export credit funding
•• customer terms.
These are all alternative funding options which have arisen
to offer alternatives to conventional project debt. Convertible
notes allow major shareholders (with available cash) to retain
an option to convert loaned funds to shares in the project
company or alternatively to redeem the loaned funds (with
interest, referred to as a coupon rate).
Export credit funding arises where the project’s product
customer is able to make use of funds provided by its host
government (usually at a lower interest rate compared to
conventional debt). For example, the Ichthys liquid natural
gas (LNG) project in Darwin has an estimated US$20 B
capital cost. Of this, US$11.2 B is to be provided by Japanese
Bank for International Cooperation (JBIC), Export Finance
for Australian Companies (EFIC), Nexi and five other export
credit authority (ECA) agencies in the form of direct loans or
guarantees. This actually resulted in the conventional debt
being scaled back (Schuler, 2013).
Finally, customer terms might include advance payment
for mineral output (sometimes referred to as a ‘prepay’),
a capital contribution to the project itself, tied to supply
arrangements, technology sharing or the purchase of an
equity share in the project itself.
Construction and commissioning
It is generally accepted that over the last decade, capital
and operating costs have escalated at rates which exceed
the traditionally enjoyed operating margins, based on the
prevailing metal prices. This, together with the project
funding challenges referred to above has caused a sharper
focus to be drawn on project construction and commissioning
costs. An excellent account of these activities was presented by
Luxford (2006), from which this author’s principal learnings
are summarised in the following sections.
Construction (project implementation)
Firstly, there is the engineering, procurement and construction
management (EPCM) decision. Do you outsource it, share it
or execute the roles in-house? Outsourced or shared EPCM is
the most frequent choice, usually where the owner assumes
EPCM responsibility for the mine plan and mining fleet
acquisition and the processing plant and infrastructure is
handled by a specialist engineering firm.
Next comes the selection of the owner’s management team.
Here the essential points, as suggested by Luxford, are:
we are metallurgists, not magicians
Mineral project management – a perspective from four decades in the industry
•• knowledge and experience in what is to be designed
and built
•• demonstrated track record in similar projects
•• honesty and integrity
•• intellectual and practical ability
•• ability to work well in a team.
It is considered essential to recognise that an owner’s
project team, separate to any ongoing operations or company
administration, be appointed. The project manager need not
be, and more often than not is not, the eventual operations
manager of the project. This arises from the fact that the project
construction manager has specialised skills in managing the
construction processes (such as contract management and
liaison with the EPCM engineers), which are themselves
different to the operations skill set.
The need for a project execution plan (PEP) is obvious
but often treated in an ad hoc manner. Luxford provides a
comprehensive outline of what such a plan entails. In the
author’s view a detailed PEP is essential, with clearly allocated
responsibilities for each set of tasks.
Engineering management responsibilities are also well
outlined by Luxford, who makes the point that, in addition,
the project management team must, in a timely manner:
•• provide user requirement specifications
•• examine and approve concepts and designs
•• review and approve equipment selections
•• approve standards.
Commissioning
(AusIMM, 2012). Of the various headings dealt with in that
text, the author sets out below his observations of what are
perhaps areas of potentially major project impact.
Regulatory considerations
For many years there has been confusion caused by the lack
of congruence between the various metalliferous and coal
mining regulations across the Australian states. This began
to change from the 2002 Conference of Chief Inspectors
of Mines and has now reached the point where in April
2009 the Workplace Relations Minister’s Council (WRMC)
endorsed the creation of Safe Work Australia (SWA), a new
independent body charged with progressing the concept of
harmonised work health and safety laws.
Following on from that, the National Mine Safety Framework
(NMSF) steering group is working with SWA to develop mine
specific regulations with national outreach and application. In
effect, each state was to have ‘mirror legislation’ in place by
1 January 2012, though only the NT and the ACT actually met
that deadline (as recently as 2016, complete harmonisation
has yet to be achieved).
The end result will be that all states will adopt ‘core
drafting instructions’ but retain some freedom in ‘non-core
areas’. The resulting legislative approach has been termed
harmonised legislation.
The three areas of key harmonisation may be summarised
as follows:
1. roles, functions and powers under the model legislation
(such as managers and mines inspectors) for mining
operations
Project commissioning is often misunderstood by operating
management. There are several stages involved in this
process. Luxford describes this process below:
2. management plans and records under the model (such
as hazard analyses)
Construction completion involves completing installation
and erection work. The phase is complete when all
components of the completed system have successfully
passed inspection and testing to verify they comply with the
contract. The only equipment operation at this point is to
check motor rotation.
These three points are expanded upon in the MMH
(AusIMM, 2012) at chapter 7, and familiarisation with them is
recommended. The harmonisation process will be completed
before too long and it is recommended that minerals
professionals keep apprised of developments.
No-load commissioning involves testing and verifying that
all equipment is ready to run, followed by operation of all
equipment under zero-load conditions. At the completion of
this phase, the plant has achieved mechanical completion and
is ready for commissioning. The plant is ready for handover
and is accepted by the owner. At this point, ownership is
usually transferred to the owner and the defects liability
period commences. Wet commissioning then follows with
the introduction of loads to the plant. In the case of a process
plant this would involve the loading of process media and
introduction of ore followed by:
•• adjustments and minor modifications
•• process guarantee performance testing
3. duties (such as duty of care) and other requirements.
Operations management
Perhaps the most controversial aspects of operations
management in the author’s experience has surrounded issues
to do with owner versus contract mining and joint ventures.
Some observations on these are offered below.
Owner mining versus contract operations
Detailed accounts of the main issues dictating a decision on
which operating methodology to adopt may be found in
Dunlop (2002) in the case of open pits, and Luxford (2005) in
the case of underground works. The arguments for and against
each operating strategy are clearly set out in these papers, along
with the operational backgrounds dictating each situation.
OPERATIONS
In recent times, commodity prices have been falling whilst
operating costs, capital costs and foreign exchange rates
have all been moving adversely, thus threatening operating
margins and making the economic viability of minerals
projects all the more challenging (In 2017, we are again seeing
effects of this in the so called ‘end of the mining boom’).
As a direct result, most operations in this country are now
back on a cost cutting footing. This has caused owners once
again to look at the contractor’s margin and demand that the
advantages of contracting out still justify the additional cost.
An excellent account of mine (and mineral processing
plant) operations management may be found in the MMH
Whilst this is a natural consequence of the economic times,
the author is still of the view that the reasons for and against
•• completion of as-built documentation
•• final punch list completion.
Finally, it is recommended that the construction handover
phase include a checking that all as aspects of the PEP
have been completed and that the EPCM engineer’s list of
deliverables has been completed and that those deliverables
are in place.
we are metallurgists, not magicians
177
J S Dunlop
contracting services out are still not sufficiently recognised
or understood. It is therefore recommended that decisions to
contract out or owner-operate are taken following a detailed
cost and benefit analysis following a careful review of the
literature on this subject.
Joint ventures
The next issue of potential major project impact centres on
JV. For an excellent general overview of JVs in the minerals
industry, readers are referred to Reynolds (1983). In Reynolds’
view, the principal reason for the creation of JVs is the need
to share project risk. He then describes the (presumably most
significant) risks as:
•• project size in relation to the market
•• financial magnitude of the project
•• whether new technology is involved
•• lack of certainty.
In this author’s experience, a large proportion of minerals
industry litigation has arisen out of JV arrangements, simply
because these unincorporated vehicles offer the greatest
potential for divergent participant interests. A worthwhile
example is that of the original Ok Tedi JV where the three
major participants, BHPB, Amoco Minerals and the PNG
Government had differing project development priorities.
Namely these were capital cost minimisation; maximised
NPV and early production, respectively.
It is probably a truism to suggest that farm-outs are only
born of necessity – that is, when all other options have been
investigated and shown to be impracticable. The party farmingout has to cede some control (often total management) in
order to see the project proceed, exacerbated by the farming-in
party assuming autonomous control and paying scant regard
to the new ‘minor participant’. Unless, therefore, there is good
will on both sides and strictly ethical behaviour, disputation
is highly likely. Should such disputation proceed to litigation,
the possibility is real that the project may be delayed for years.
The areas of disputation considered by this author to be
most common include:
•• failure to prepare and adequate JV agreement
•• failure to provide for equitable voting rights on the JV
operating committee (with regard to key decisions such
as the life-of-mine plan and annual budgets)
•• wilful abuse of JV terms as defined in the agreement
Substandard or otherwise deficient reporting has had
the result that many mines have no adequately detailed,
permanent record of their month-by-month operations, going
back in time. It is simply a matter of sound management that
adequate operation reports be prepared and appropriately
backed up and stored.
Readers are urged to adopt, as a minimum standard the
pro forma operations report format set out in Appendix 3 of
the MMH.
Offshore projects
Minerals professionals of all disciplines will readily recognise
the importance of identifying, monitoring, fostering,
maintaining and always improving stakeholder relationships
on their project. These issues are discussed in some detail in
chapter four of the MMH.
Offshore projects introduce another dimension to the picture,
where cultural or other norms may not be as we, the expatriate
skills providers, may well expect. It goes without saying that
great care and sensitivity is required in this situation and
careful choice of staffing for foreign work is advisable.
Another issue that often arises is that of the project
establishment coming ahead of basic regional infrastructure
such as power and water supplies, communications and
associated basic services. In the author’s experience, it is
essential that the project address this imbalance in some
meaningful and ongoing way, accepting that, in principle,
the prosperity of the regional stakeholder must advance in
line with the project itself, if stakeholder harmony and trust
is to be maintained.
‘Sustainable development’ was studied by the Mines and
Minerals Sustainability Development Project (MMSD), which
reported its findings (International Institute for Environment
and Development, 2002). There are useful observations in that
report which are still considered relevant more than ten years
after its publication.
Ethics
A simple definition of a professional (courtesy of the Oxford
dictionary) is:
...a person engaged or qualified in a profession: professionals
such as lawyers and surveyors; a person engaged in a specified
activity, especially a sport, as a main paid occupation rather
than as a pastime; a person competent or skilled in a particular
activity: she was a real professional on stage.
•• lack of good faith by the manager, possibly coupled with
a failure to exercise a fiduciary duty to the other party
(after writing this paper, the author became aware of a
landmark Supreme Court judgement in South Australia
where it was held that a manager had no fiduciary duty
to a minor JV participant – details upon request).
Perhaps a more relevant definition for the minerals industry
might read something along these lines: ‘a person of specialised
training or skills who, acting always with ethics, applies those
skills for the betterment of society and the industry, before
him or herself’. This definition is manifest in the first bullet
point of the AusIMM Code of Ethics (AusIMM, 2013).
At the outset, therefore, it is strongly recommended that a
detailed and painstaking review be applied to all JV agreement
drafts, and that expert advice be obtained before they are
executed. It is also suggested that the disputes clauses in the
agreement be broadly worded and provide for specialist,
binding arbitration as opposed to litigation.
In this latter definition, it is implied that a professional is a
person with a degree; one qualified to become a Member of
the AusIMM or an equivalent professional institution, though
there is no reason why tradesmen, para-professionals and
other service providers should not act ‘professionally’.
Operations reporting
Brief reference is made here to operations reporting, which,
in the view of the author, have lacked consistency over many
years. There are many reasons for this, but at the very least,
there has never been a widely adopted template or other
reporting convention to guide site operations personnel.
178
It is often assumed in general trade and commerce that
the rules of ethics do not apply – or at the very least have
low priority – whatever those rules may be. Consequently,
circumstances can arise where a professional is confronted
with disingenuous behaviour by others, and thus be tempted
to adopt the same tactics.
The Commonwealth Trade Practices Act (TPA) and its more
recent replacement, The Competition and Consumer Act 2010
we are metallurgists, not magicians
Mineral project management – a perspective from four decades in the industry
(CCA, 2010), are silent on ethics but make the following
requirement very clear: in trade and commerce it is expressly
prohibited to intentionally or unintentionally to engage in any
conduct which might mislead or deceive another party. This
requirement covers a key part of professional ethics (not acting
in a misleading or deceptive way), but does not cover the entire
spectrum of ethical practices a professional must follow.
As a general rule, the best defence is what is termed the
‘due diligence defence’ which, in plain language holds that
you cannot be guilty of misleading or deceptive conduct if
you can show that every effort was made not to do so. For
example, peer review of reports or opinions, risk audits and
other forms of due diligence will be taken into account in this
context. Mere disclaimers, absent the above actions, will not
achieve the same level of protection.
The obvious dilemma which emerges for professionals
leading public companies is embodied in the question: is
his or her primary obligation to the (profit of) company
shareholders or to the (betterment of) the public? As long as
company directors and executives continue to disregard this
dual obligation, corporate ethics will continue to lag behind
the required standards of profession ethics.
PROJECT EXPANSIONS
Project expansions are becoming more and more
commonplace, probably for two major reasons: the high
capital cost of start-ups making the ‘start small’ approach
attractive; and the simple economics of increased capacity
reducing unit production costs. Consequently, expansions of
operating plants are becoming a feature of a minerals project’s
normal operating life.
In the past, however, the obviously disruptive practice of
expanding an existing plant whilst continuing operation at
the old rate until ‘cut-ins’ were complete, was rare. This is no
longer the case and it is suggested that modern day mining
managers need to be equipped with the skills to recognise the
key requirements of an ‘internal’ expansion and to resource
the work appropriately.
In practice, all of the guidance set out under the construction
and commissioning headings (earlier in this paper) apply
equally in this context, though operations managers often
fail to recognise the fact sufficiently. There is a tendency for
operational and construction responsibilities to overlap unless
the operations manager takes organisational steps to separate
the two activities, as recommended by Luxford (2006).
These construction projects also require special care regarding
fitness-for-work and inductions for casual contractors
associated with the construction work. Contractors who come
onto the designated construction-site (as distinct from the mine
site) need to be inducted and cleared for work even though
their site residence time may be as short as one day. Time must
therefore be allowed for medical screening and drug clearance,
so as to ensure contractors do not commence work whilst their
prestart screening results are awaited.
Finally, it is essential that any additional site regulations
that apply pursuant to construction-related (non-mining)
regulations are clearly understood by all concerned.
PROJECT CLOSURE
Project closure (and, perhaps suspension of operations) is an
integral part of the overall mining cycle. Despite this, a search
of the published literature on ‘mine closure’ reveals most
papers which focus on the environmental compliance issue
post-operations. Whilst such compliance sits well as part of
the MMSD principles referred to earlier, the literature has
we are metallurgists, not magicians
little to say about sustainability in general in the locations
where mining has ceased. For example, it is difficult to find
accounts of communities which have not only benefitted
from the mining operations themselves, but have also
benefitted thereafter, by virtue of sustainable small business
ventures and the like which have endured post-closure
(The Bougainville Copper and Ok Tedi projects, (the author
worked at both) when looked at today, have achieved little for
the local communities in which the projects were established,
despite the best intentions).
Readers interested in this aspect of the mining cycle are
referred to McGuire (2003) who describes the mine closure
process at Rio Tinto’s Kelian gold mine in East Kalimantan.
In that paper, the concept of a ‘Mine Closure Committee’ is
introduced, along with the concept of involvement by the local
(and broader, regional) stakeholders in the committee’s work.
Today, a key part of annual audit requirements is a review
of mine closure plans and financial provision for those plans.
It is evident that much greater thought and detailed planning
and resourcing needs to be applied to this phase of the
mining cycle in the future. The sustainable mining principles
dictate that when mines close, the land and the communities
left where the mine once occupied ought to be better off for
the mine having been there, not just while the mine was in
operation.
THE CAPITAL MARKETS
The changing pattern of investment in world mining is
constant and the effects of those changes have a profound and
continuing effect here in our Australian minerals industry.
These observations were made by Aldous (1993) whose paper is
still recommended for those wishing to understand the concept
of mineral capital allocation in the world’s capital markets.
If we were to consider Aldous’ view in 1993 that capital
comes from where it is most available and goes to where it
is most welcome, and compare that view with the situation
today, it is clear that these fundamentals have not changed.
Suppose we were to rank Australia’s ‘capital attractiveness’
according to Aldous, we might well be disturbed by the list of
negatives, which currently include:
•• lengthy project approval timelines
•• relatively high project capital and operating costs
•• inflationary trends which outpace metal price growth
•• relatively high environmental and other compliance costs
•• relatively low rates of return compared to the risk profile
•• a small capital market compared to the rest of the world
•• an uncompetitive foreign currency exchange rate.
The combination of these factors, coupled with financial
instability in most of Europe and a sluggish US economy, has
caused the minerals equity and debt markets to dry up almost
completely for ‘small cap’ and mid-tier companies on the
Australian Stock Exchange (ASX) and the Toronto Venture
Exchange (TSXV). If this situation is allowed to continue,
significant damage will be done to what some would refer to
as our country’s cornerstone industry.
Whilst the capital markets will recover, following the endless
cycles we all recognise, the industry needs to find ways to
continue in the meantime. The solutions, it is suggested, lie in
the following strategies:
•• operating cost reduction
•• continuous improvement to lift productivity per
employee
•• leading in the application of advanced technology
179
J S Dunlop
•• faster project approval mechanisms
•• favourable government policies for projects of ‘national
importance’
•• favourable taxation incentives for small cap explorers
•• further development of Free Trade Agreements (FTAs)
with our closer trading nations, to further unlock
additional export credit.
GOVERNMENT UNDERSTANDING OF THE INDUSTRY
Following on from the aforementioned observations, there is
likely to be little disagreement with the proposition that the
current mining boom in Australia has stalled, to a greater or
lesser degree, due to the lack of capital attractiveness, to which
has been referred.
The Minerals Council of Australia (MCA) addressed this in
a paper by Ergas and Owen (2012), which concluded:
The easy gains from Australia’s early 21st century mining
boom are over, though large and enduring benefits are still
there to be secured from further resource-intensive growth
in emerging Asia.
Rebooting the mining boom calls for renewed policy focus on
securing the next generation of mining project investment
and delivering on potential export volume growth out
to 2025 (Author’s note: Very pertinent still in 2016.
Perhaps the ‘transition from the mining boom’ will not
be to something else but rather back to the next mining
boom in the endless minerals cycle).
That focus should be on tackling the ‘unfinished business’
across Australia’s export supply-chain, from exploration
and initial development through to final shipment, where
cost control, timeliness, flexibility and adaptability present
critical challenges. It is only by stripping out inefficiencies
across our export supply-chain that Australia can reap the
rewards on offer from the next phase of the boom.
As Marius Kloppers (then CEO of BHP Billiton) has
observed, companies themselves need to do the heavy
lifting to make our mining projects more cost competitive.
But they can only do so if the policy environment provides
the framework and tools for this to occur. At the moment,
it doesn’t.
To ‘re-boot the boom’, then, governments at all levels must
improve their understanding of the industry – particularly
now when it contributes to approximately 50 per cent of
national exports and is increasing. At present, its level
understanding, in the author’s view, is superficial at best and,
at worst, lacking in both breadth and depth when viewed in
the context of its national economic significance.
To illustrate the proposition, the industry can be viewed in
a range of ways:
•• commodity (base metals, precious
commodities, industrials and so on)
metals,
bulk
•• industry grouping (metalliferous, coal, petroleum,
industrial)
•• activity (exploration,
processing)
production
or
downstream
•• size (small caps, mid-tiers producers and multinationals).
Each of these groupings require a profound understanding
to formulate policy settings which will not only be of benefit
to those concerned, but will also address the structural
efficiency areas referred to by Ergas and Owen. It is the view
of this author that those responsible for governing and setting
minerals policies must improve their understanding of, and
180
liaison with our industry at all levels and commit to a much
higher priority level being accorded to the industry’s needs.
CONCLUSIONS
The personal observations of critical issues presented in this
paper span the many phases of mineral project development
from exploration, through development, operation, expansion
and finally closure. In summary form, the critical issue
identified are as follows:
•• exploration – the market continues to lose site of the real
lead time to develop a project following its discovery
•• discovery – stakeholders need to be identified and
engaged with as soon as possible
•• scoping, prefeasibility and definitive feasibility –
potential fatal flaws need to be identified and resolved
early in the process
•• project approval – current processes are far too slow and
completely uncompetitive when seen in an international
context
•• project financing and commissioning – conventional
debt funding is losing its attraction, in favour of
alternative financing means
•• operations – issues of concern relate to JV and corporate
ethics
•• expansion – operational and construction responsibilities
to overlap unless the operations manager takes
organisational steps to separate the two activities
•• project closure – mine closure plans need to be detailed
and involve a mine closure group of all relevant
shareholders.
Additional observations have been made in relation to:
•• capital markets – emphasizing the Australian minerals
industry’s current capital unattractiveness, and
corrective measures
•• government understanding of the minerals industry –
some thoughts on how the government might better
view and assist the industry.
ACKNOWLEDGEMENTS
The author wishes to thank the organising committee of
MetPlant 2013 for the opportunity to present this keynote
paper. He also gratefully acknowledges the many authors
cited in the paper, whose contributions to the topics discussed
in the paper provide an expanding pathway for those wishing
to concentrate further on any of the issues touched upon.
REFERENCES
As far as practicable, all references cited in this paper may be
found on <www.onemine.org> and downloaded for free by
members of the AusIMM, SME or SAIMM.
Aldous, R T H, 1993. The changing pattern of world mining investment
and the implications for Australia and her mining industry, in
Proceedings Centenary Conference, p 267 (The Australasian Institute
of Mining and Metallurgy: Melbourne).
AusIMM, 2013. Code of Ethics [online]. Available from: <http://
www.ausimm.com.au/content/docs/code_of_ethics_stand_
alone.pdf> [Accessed: July 20, 2017].
AusIMM, 2012. Mine Managers’ Handbook, Appendix 3 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Australian Uranium Association, 2010. NT decision on Angela
Pamela causes uncertainty [online], press release, 28 September.
Available from: <www.aua.org.au> [Accessed: 8 May 2013].
we are metallurgists, not magicians
Mineral project management – a perspective from four decades in the industry
Binon, L C, 1981. Determining the exploration budget, in Proceedings
SME-AIME Fall Meeting and Exhibit, Denver, November, Preprint
No 81–311 (Society for Mining, Metallurgy and Exploration;
American Institute of Mining, Metallurgical, and Petroleum
Engineers: Englewood).
Card, P, 2011. Guidelines for economic evaluation of mineral projects,
in Proceedings MetPlant 2011, August, 12 p.
CCA, 2010. The Commonwealth Consumer Act [online]. Available
from:
<http://www.comlaw.gov.au/Details/C2007C00619>
[Accessed: May 2013].
Chamberlain, C, Johnston, G and Joyce, C, 1997. Regulatory
framework for Mine Project Permitting, in Proceedings Mindev 97,
November, pp 193–198 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Duncan, I J, 1985. The competitive challenge in marketing Olympic
Dam copper, uranium and gold, in Proceedings The AusIMM
Annual Conference, p 6.
Dunlop, J S, 2002. Contract versus owner mining – an update on
Australasian open pit mining practice, in Proceedings Iron Ore
2002, Perth, p 205 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Ergas, H and Owen, J, 2012. Rebooting the boom: unfinished business
on the supply side. Available from: <http://www.minerals.
org.au/file_upload/files/publications/mca_rebooting_the_
boom_FINAL.pdf> (Minerals Council of Australia: Canberra)
[Accessed: May 2013].
Hinz, R and Aseervatham, R, 1999. Feasibility studies, EPCM and the
environment – case studies in integrated mine development, in
Proceedings PACRIM 99, Bali, October, p 123 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
JORC, 2012. Australasian Code for Reporting of Exploration
Results, Mineral Resources and Ore Reserves (The JORC Code)
[online]. Available from: <http://www.jorc.org> (The Joint Ore
Reserves Committee of The Australasian Institute of Mining and
Metallurgy, Australian Institute of Geoscientists and Minerals
Council of Australia).
McCarthy, P, 2002. Setting plant capacity, in Proceedings MetPlant
2002, Perth, pp 21–30 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
McCarthy, P, 2011. Predicting variations in mill feed, in Proceedings
MetPlant 2011, pp 19–25 (The Australasian Institute of Mining
and Metallurgy: Melbourne).
Lawrence, M J, 1994. An overview of valuation methods for
exploration properties, in Proceedings Mineral Valuation
Methodologies, p 205 (The Australasian Institute of Mining and
Metallurgy: Melbourne, in conjunction with The Mineral Industry
Consultants Association: Sydney).
Luxford, J, 2006. Project development and construction management,
in Proceedings International Mine Management Conference, p 97 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
International Institute for Environment and Development, 2002.
Breaking New Ground: Mining, Minerals and Sustainable Development
[online], 450 p. Available from: <http://pubs.iied.org/9084IIED.
html> [Accessed: May 2013].
McGuire, G M, 2003. Managing mine closure risks in developing
communities – a case study, Kelian Equatorial Mining, Indonesia,
in Proceedings Mining Risk Management Conference, p 385 (The
Australasian Institute of Mining and Metallurgy: Melbourne).
Murray, J, 1995. Project finance in the minerals industry, in Proceedings
Young Professionals Conference, p 59 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Needham, R, 1985. The Kidston Gold Mine: a development
challenge of a low grade deposit, in Proceedings AusIMM Annual
Conference, p 121 (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Nice, R W, 1986. The Paddington Gold Mine Project – a case study,
in Proceedings Second Project Development Symposium, Sydney, p 299
(The Australasian Institute of Mining and Metallurgy: Melbourne).
Noort, D J and Adams, C, 2006. Effective mining project management
systems, in Proceedings International Mine Management Conference,
p 87 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Reynolds, G C, 1983. Unincorporated joint ventures in project
developments in the mining industry, in Proceedings Project
Development Symposium, Sydney, p 349 (The Australasian Institute
of Mining and Metallurgy: Melbourne).
Schuler, S, 2013. Financing projects in hard times: export credit agencies
are the smart alternative [online]. Available from: <http://
www.corrs.com.au/thinking/insights/financing-projects-inhard-times-export-credit-agencies-are-the-smart-alternative>
[Accessed: May 2013].
VALMIN, 2015. Australasian Code for Public Reporting of Technical
Assessments and Valuations of Mineral Assets (The VALMIN
Code) [online]. Available from: <http://www.valmin.org/docs/
VALMIN_Code_2015_final.pdf> (The VALMIN Committee,
a joint committee of the Australasian Institute of Mining and
Metallurgy and the Australian Institute of Geoscientists).
Van Der Merwe, M J, Vegter, N M and Momberg, N 2008. Value
driven resource allocation and focus in the early stages of project
development, in Proceedings Lead and Zinc 2008, p 255 (The South
African Institute of Mining and Metallurgy: Johannesburg).
Whincup, P R, 2008. The role of the metallurgist in project
development, in Proceedings MetPlant 2008, p 25 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Luxford, J, 2005. Future directions in Australian contract mining, in
Proceedings Underground Operators, p 3 (The Australasian Institute
of Mining and Metallurgy: Melbourne).
we are metallurgists, not magicians
181
Contents
Keeping projects on the rails
J Canterford1
ABSTRACT
Although we have access to a wide range of sophisticated process design and
improvement tools in combination with a significant range of case studies, the mineral
resource industry still manages to be plagued by the negative publicity surrounding
operating plant failures because they were neither smart nor safe, together with
proposed projects that fail at the first or second hurdle. The public and the finance
industry are generally totally uninformed about projects that ramp-up to name-plate
capacity on-time and on-budget and even more so when ahead of schedule and
under-budget. As the saying goes, good news does not ‘cut the mustard’.
The following is a summary of the author’s sometimes jaundiced view about why
metallurgical plants do not live up to their technical and commercial expectations.
In no way should the summary be construed as a negative reflection on the skills of
the metallurgical profession. Rather, it is to be taken as a positive acknowledgement
of those skills and is intended to highlight some of the major considerations that
need to be assessed as a potential project transforms into a sustainable technical and
commercial reality.
While a considerable portion of the summary is directed at developing
hydrometallurgical flow sheets, the general principles are equally important to all
aspects of extractive metallurgy.
For convenience, the summary covers the following five basic discussion points –
project, people, process, patents and politicians.
INTRODUCTION
On any given day, every successful metallurgical plant can be characterised as being
a positive compromise between a series of potentially conflicting criteria, the more
significant including:
•• the mineralogical complexity of the feedstock
•• market requirements
•• owner/developer expectations
•• technical complexity
•• ongoing operating cost scenarios
•• sustainable environmental footprint.
Although we can talk about generic flow sheets, no two flow sheets are identical
when dissected in detail. While we certainly can and should learn from past and current
experience, what we must accept is that the balance between the conflicting criteria will
almost certainly alter during the life of any individual project. In turn, this means that
the ‘best’ flow sheet should have a reasonable level of flexibility to cater for a realistic
range of inputs and outputs, both metallurgical/environmental as well as corporate.
Some of the anticipated variations may be self-induced while others will be a result of
external forces over which the project owner and/or developer has no control.
If the conflicting criteria cannot be balanced and that balance maintained, then
the outcome will almost certainly be closure or at least the requirement for major
modifications to plant layout, operating procedures, changes to volume and
specifications of outputs etc. The capital and operating cost implications of such
modifications can be mind blowing. Without going into any detail, experiences at
Ravensthorpe and Goro are stark reminders of what can go wrong.
1. FAusIMM(CP), Process Technologies Australia
Pty Ltd, Deloraine Tas 7304.
Email: pr37666@bigpond.net.au
As a person with no practical metallurgical plant design experience but who is
sometimes asked to comment on what, how and why a flow sheet appears to have
failed or at least throw enough hand grenades at the evaluation stage of flow sheet
development, the simplistic answer provided is itself multifaceted covering:
•• the chemical/mineralogical integrity of all test work
•• realistic technical risk analysis
183
J Canterford
•• treatment of process and financial models with a fairly
high degree of scepticism
•• understanding what the market will buy and why and
how these requirements can be met.
It will come as no surprise that this author is very strongly of
the view that throwing hand grenades and proper evaluation
of the metallurgical plant design should be initiated at the
scoping stage and should be a continuing exercise well into
the operating phase and until name-plate capacity has been
achieved or exceeded.
SUCCESSFUL PLANT DESIGN DRIVERS
First and foremost, it is essential to match output volumes and
qualities with a realistic assessment of market requirements.
As noted below, some 15 years ago magnesium metal was
the metal of the future. Numerous proposals were being
bandied about. In Australia alone, the combined capacities
of the proposed magnesium metal projects in Tasmania,
Queensland, Victoria, South Australia and Western Australia
between 1995 and 2005 was in excess of the 300 per cent of
then total world consumption and more than 200 per cent
of the projected demand in 2015. Individually each of the
proponents considered their proposed projects to be superior
to all others and that only they would attain commercial status.
Of course reality was quite different and not one of those
proposed Australian projects has seen the light of day. As the
author of this paper participated in some of these proposed
projects he must confess to being guilty by association.
Consistent with his novice status as a metallurgical plant
‘expert’, he would like to put forward what he has termed
the ‘Five P’ set of criteria that should be seen as one of several
sets of criteria that need to be considered when developing
and maintaining a state-of-the-art metallurgical plant. This
set of criteria is made up of Project, People, Process, Patents
and Politicians. Clearly there is a strong interplay between
each of these components and each may individually or in
combination lead to initiation of an outstanding technical and
commercial success but conversely to an absolute failure.
From the author’s perspective it is important to expect the
unexpected and avoid any major deviation from the KISS
(‘keep it simple’) principle. The title of this presentation
indicates acquisition of information by osmosis and as such
the following comments cover some of the more relevant
observations made over the past 40 years.
Project
Geometallurgy is now properly recognised as a critical
evaluation step in establishing a technically and commercially
viable process flow sheet. Unless a detailed evaluation of the
chemical and mineralogical complexity of the resource under
consideration is executed then failure in some form is highly
likely to follow. One of the real challenges is to ensure that test
work samples should be properly characterised and that the
selected samples are representative of the expected resource
overall aspects of the entire project life. This means that there
must be close cooperation and coordination between the mine
planners (including the geologists), the client’s project team
and the process flow sheet developers.
While the distribution of the marketable metal(s) and
product(s) has a major influence on mining and processing
options, in reality it is how, where and in what form the
non-value components are present and how to reject them
efficiently that will be the key to success or failure.
Geometallurgy is also a major factor that assists in the
determination of realistic product outputs. Projects that can
184
be described as conventional in that they yield standard
products such as gold bullion via heap leaching of an oxide
gold ore or cathode copper via smelting/electro-refining
of high-grade chalcopyrite concentrates can be relatively
straightforward in that product specifications are well known
and there is an active, open market.
The situation with projects where each end-user sets their
own specific product specifications is quite different. For
example, virtually every end-user of dead burned magnesia
will have quite specific requirements for crystallite size and
shape in addition to other chemical and physical properties.
There are no universal product specifications. While magnesite
calcination may seem to be a relatively straightforward
metallurgical process, it is the physical structure of the
magnesium carbonate raw material as well as the calcination
conditions that determine the properties of the magnesium
oxide product. It follows that the owner/developer of a
magnesia production facility must be fully aware of detailed
market requirements and understand that it may not be
practical to service a broad spectrum of customers. Thus there
has to be an ongoing assessment of potential markets as it is
not a good option to tie production to a very limited number
of customers.
Practical metallurgical projects are generally based upon
consumables that are widely available although some, such
as sodium cyanide, will require a high level of process control
and mitigation strategies to be put in place to overcome any
plant failures however caused.
People
As practicing metallurgists, it is important to maintain a
high level of realistic technical input into senior management
considerations and, where appropriate, curb the misplaced
conviction by the technically illiterate that reliance on
simplistic in-ground evaluations and avoidance of locked
cycle continuous test work at the pilot and demonstration
level are just two of several ways of achieving a project that is
not technically and commercially sustainable.
It is not unusual for initial capital and operating costs to
be considerably higher than originally anticipated so that
cost-cutting procedures are instigated, particularly by the
corporate finance management team. This may involve,
for example, reduction in pipe and cable runs, reducing
overhead access, reduction of automation and monitoring
instrumentation etc. While there may be a saving on structural
steel, concrete, process control facilities etc, the result will
ultimately lead to increased maintenance costs, more complex
occupational health and safety issues, less reliable adherence
to operation within the designated bandwidth and potentially
extended shutdown periods. These technical issues must be
properly flagged. The project’s technical champion has a very
significant role to play.
‘Bean counters’ need to be controlled to avoid equipment
selection based on simple cost and availability terms, rather
than on a genuine ‘fit for purpose’ basis. There is no point
in gold plating the processing plant, but it should not be
deficient in practical operational and safety terms. Similarly,
there is no value in over-promoting the economic benefits that
might accrue.
Process
It is possible to develop a process flow sheet that is chemically
sustainable, at least at the theoretical and initial pilot/
demonstration scale, but which never reaches commercial
status because it is ‘impossible’ to engineer and operate
it and/or economic reality sets in. There are quite a few
We are metallurgists, not magicians
Keeping projects on the rails
proposed processing flow sheets that are best described as the
‘tail wagging the dog’. This is particularly the case where the
flow sheet incorporates recovery of every possible product in
the purest possible form. For example, many of the proposed
nickel laterite flow sheets incorporated recovery of metallic
magnesium, given that the magnesium content of the pregnant
leach liquor is many times greater than that of nickel. Only a
little bit of evaluation from the other side of the fence clearly
indicates that this approach is not sustainable. If the project
does not make commercial sense based on the nickel cash
flow with a nominal cobalt credit then the presumed financial
benefits from additional by-products will just be an illusion.
As a general but not totally universal comment, the ‘best’
process flow sheets are developed and commercialised by
project owners/developers. This is particularly the case with
mineralogically complex resources since exploitation of the
resource will not proceed without the successful development
of the required flow sheet.
This comment should not be taken as an unqualified
criticism of research and development (R&D) undertaken
by universities, research organisations and independent
companies. They certainly generate some great concepts.
However, such groups rarely have the financial and technical
expertise and abilities to undertake and achieve commercial
status for their concepts and are reliant on selling their knowhow. One must always question their ability to provide the
necessary support when their know-how is implemented
by others. In other words, the end-user should always very
seriously question the real value (technical and commercial)
of the licence to use purchased technology.
As noted above, magnesium metal has been vigourously
promoted as the metal of the future. Australia was at the
forefront of developing what was being claimed as novel
state-of-the-art processing regimes. Three major areas of
sustained R&D centred on:
1. purification and dehydration
magnesium chloride liquors
of
concentrated
2. development of alternative electrowinning electrolytes
3. improved cover gas technology.
Despite spending more than $50 M on their demonstration
facility, amongst other things the Australian Magnesium
Corporation (AMC) purification and dehydration technology
failed to deliver the anticipated outcomes. The chemical/
thermal decomposition of the organic dehydration agents
was soon found to be quite deleterious.
Electrowinning magnesium from fused salt electrolytes is
highly energy intensive as well as challenging in engineering
and operational terms. Molten magnesium metal is not the
simplest material to handle. Energy consumption during
electrowinning can be reduced by modification to the
composition of the fused salt electrolyte. One such option is
to add neodymium chloride. Unfortunately one side effect
is that rather than pure magnesium being produced at the
cathode, the end product is a magnesium-neodymium alloy.
Such alloys have several useful physical properties, but it was
soon worked out that the ‘loss’ of neodymium to the alloy
product was such that the volume of make-up neodymium
chloride would soon exceed that currently available. In other
words, the proposed magnesium metal production facility
would be dependent upon the establishment of an ongoing
rare earth production facility with a significant neodymium
output. The technical and commercial constraints so imposed
basically canned the concept of the use of neodymium chloride
additives to the magnesium electrowinning cell house.
We are metallurgists, not magicians
The cover gas technology developed by AMC and CSIRO
was and remains technically astute – it is unfortunate that it
has not been commercialised due to the total collapse of the
AMC project.
As noted previously, it is important to expect the
unexpected, especially when considering some of the more
complex hydrometallurgical flow sheets. For example, during
carbon dioxide leaching of several caustic calcined magnesia
feedstocks it was found that the resultant liquors had quite a
high soluble ferric concentration (several g/L) even through
the bulk pH was greater than 9.5. Under the operating
conditions it is possible to form soluble iron (III) carbonato
complexes even though all the available thermodynamic
data suggested that this was ‘impossible’. Development
of a suitable technique for selective iron removal from the
pregnant liquor proved to be quite challenging.
During the 1970s and 1980s the majority of hydromet flow
sheet development centred on the initial leach step, with
less emphasis on the downstream purification and recovery
steps. For example, chloride hydrometallurgy of base metal
sulfide ores was seen as the panacea of all environmental ills
associated with sulfur dioxide abatement with conventional
pyrometallurgical operations, although it is appropriate to
note that commercial reality is yet to be achieved. While
detailed knowledge of the leach step was generated, many
of the challenges with chloride hydrometallurgy relating
to separation and recovery of elemental sulfur, recovery of
precious metals from leach residues, deportment of nasties
such as arsenic, regeneration and recycle of the leachant, sulfate
control etc, were subjected to far less rigourous evaluation.
Fortunately this situation of unbalanced unit step development
for most hydrometallurgical flow sheets is now being corrected.
Two of the driving forces for a balanced assessment of all
hydromet unit steps are the need for and influence of a range
of internal recycle stages and the maintenance of a workable
process water balance. In some locations and for some flow
sheets there may be an excess of process water that needs to
be discharged into the local environment in an acceptable,
benign manner. In other situations, there will be a potential
deficiency of fresh suitable process water at reasonable cost,
so recovery and recycle of process water is a requirement that
adds to both capital and operating costs. Desalination plants
have their own problems including power supply, plant
duplication for continuation of supply and waste disposal.
For hydrometallurgical flow sheets in particular, all
definitive test work from feed preparation, which may
involve grinding and flotation, right through to final product
recovery, must be carried out using process water that will
be available at site. The physical and chemical properties of
the process water will have a significant effect on parameters
such density, viscosity, redox potential, oxygen solubility,
ionic strength, leach and precipitation kinetics etc, let alone
materials of construction considerations and actual equipment
design. It also follows that care must be taken in preparing
process models since many of the thermodynamic inputs for
most models are clearly deficient.
Infrastructure requirements and power generation
(particularly if ‘peak’ oil status is fact not fiction) will impact
on all remote sites as a cost burden as governments continue
to bail out of these areas.
Patents
While it is not unreasonable to protect intellectual property
(IP) by means of filing and executing patent applications,
from a practical perspective the situation is tending to get
‘out-of-hand’. For example, since 2000 more than 90 patent
185
J Canterford
applications on sulfuric acid leaching of nickel laterites have
been filed via the World Intellectual Property Organisation
(WIPO). The claimed flow sheets cover heap, vat, atmospheric
and high-pressure leach circuits, sometimes in combination
with downstream unit steps such as iron precipitation,
intermediate product recovery, solvent extraction and
electrowinning. Sometimes it is quite difficult to discern any
realistic technical differences between two or more claimed
flow sheets. Part of the problem relates to the definition of
patent laws and concepts such as novelty and inventiveness.
It must also be remembered that a patent is basically an idea
– it is not necessary to prove chemically and/or physically
that the concept actually works. In fact in some instances a
little technical nous indicates that the concept cannot work.
Another feature of patent law that is sometimes difficult
to fathom in technical terms is that known processes and
concepts can be incorporated into primary and subordinate
claims. For example, many of the nickel laterite sulfuric acid
leach applications incorporate current commercial practice
such as nickel-cobalt separation by solvent extraction using
Cyanex 272. From a practical perspective, it seems difficult to
understand how any claims relating to the use of this reagent
in recent applications are sustainable.
One apparent reason for the proliferation of patent
applications relates to the fact that patent challenges/
litigation is extremely costly in terms of time, money and user
licence fees, and conditions are often unacceptably onerous.
Thus ‘new’ flow sheets are devised to overcome claims of
existing novelty/inventiveness even though such flow sheets
may not be technically and/or commercially optimal.
It is now common practice to initiate a ‘freedom to operate’
review of potential patent infringement as part of the overall
risk analysis procedures during completion of feasibility
studies.
There are numerous examples of what are best described
as ‘nonsense’ patents. Probably the most technically
‘challenging’ – and that description is certainly offered with
186
a high level of derision – relates to the ‘neutralisation of an
acidic stream using sized limestone’. It was originally allowed
by the United States Patent and Trademark Office (USPTO)
on the basis of the word sized. Fortunately, industry finally
got the USPTO to accept a dose of technical reality and the
application was forced to lapse.
Politicians
The mining and metallurgical industries face several
challenges as they are forced to cope with ever increasing
volumes of more and more diverse bureaucracy. It is
unfortunate that the majority of our politicians and their
advisers are technically illiterate and are easily persuaded
by pseudo-scientific commentary. This can and does lead to
constrained operating parameters because of the imposition
of illogical rules and regulations as a means of placating vocal
opponents of our industries. Perhaps the most bizarre was the
push to ban the use of stainless steel because nickel has some
rather ill-defined carcinogenic properties.
Our industries must and do accept the imposition of logical
environmental constraints but it is a pity that the overall
record of compliance is not recognised by our politicians
and the public. Accidents do happen, but it would be safe to
assume that our record is somewhat more positive compared
with those of many other primary and secondary industries.
The challenge for us is to educate our politicians.
CONCLUSIONS
To continue to learn by osmosis it is constructive to
periodically assess why and how ‘wayward’ metallurgical
plants come into being. This paper indicates that this occurs
because of scientific-engineering incompetence, project
owner/developer avarice, insufficient attention to future
markets and product requirements, imposition of impossible
operating constraints, right through to obvious bad luck, but
it is more likely to be a combination of all such factors.
We are metallurgists, not magicians
Contents
Operations versus projects – how do people
think and what are the implications?
G Lane1 and B Clements2
ABSTRACT
The operations and project environments are very different, but overlap in many
areas. The development of mineral processing projects is goal oriented, schedule
focused and a contractual environment. The operating environment for a mineral
processing plant is oriented around longer-term commitments, focused on delivery
against budget expectations, and an environment heavily influenced by long-term
personal relationships, community relationships and people issues.
Further contrast is evident in the way expenditure is managed. Projects commit
large amounts of money quickly to meet schedule-based objectives. Operations
manage expenditure to minimise operating costs and are able to, or are forced to delay
decisions on large expenditures. However, these two environments need to come
together in many ways, ranging from greenfields plant handover to a new operations
and maintenance crew, to day-to-day interaction and management of brownfields
projects, often in close proximity.
This paper considers the different perspectives people bring to projects and early
operations, including the advantages and disadvantages of ‘operations’ input to
a project and models for managing value adding during project development,
implementation and debottlenecking.
INTRODUCTION
This paper is about the way people need to respond to the management of operations
and projects. The mining environment in 2012 was characterised by a shortage of
engineers and managers with project experience. This situation has existed for at least
three of the preceding five years with a short hiatus driven by the Global Financial
Crisis (GFC). The shortage caused a significant movement of personnel from mining
operations into projects, on either the owner’s or the contractor’s side of the fence.
The consequences of this are numerous but this paper focuses on one area key to
the delivery of a successful project, the behavioural characteristics of the project
management and leadership team.
Experience in mineral processing plant operations brings substantial value to
a project in terms of quality of outcome; however, the mindset associated with
operations is not the same as that required for a project. The development of mineral
processing projects is goal oriented, schedule focused and a contractual environment.
Because projects are not just driven by the quality of the outcome, schedule and
cost are the key measures of success up to, and including, commissioning. Quality
is heavily influenced by the design, procurement and contracting strategies. The
simplicity and effectiveness of the management systems, alignment of, and trust
between, the participants and speed of (correct) decision-making, become the key
behavioural requirements.
PROJECT MANAGEMENT 101
A successful project is one that meets or out-performs the project criteria agreed by
the contracting parties in terms of cost, schedule and quality (performance). The
project criteria are often based on the outcomes of a feasibility study, adapted to meet
financing and other external drivers.
Successful project outcomes are based on the following:
1. FAusIMM, Chief Technical Officer, Ausenco
Minerals & Metals, South Brisbane Qld 4101.
Email: greg.lane@ausenco.com
2. Former President/General Manager,
Sociedad Minera Cerro Verde, Freeport
McMoRan Copper and Gold, Arequipa, Peru.
•• a well-gated feasibility study (Biery, Hollands and Young, 2009)
•• an experienced project team
•• sound project planning and effective communication
•• simple interfaces (contracting and execution)
•• fast and effective macro-scale decisions.
187
G Lane and B Clements
The different types of execution strategy require very
different behavioural needs (Gabrielson, 2007; Hundertmark,
Siliva and Shulamn, 2008). Engineering, procurement,
construction, management (EPCM) style projects require
ongoing collaboration and interfacing with the project owner
and the associated operational team. This contract style
is characterised by ongoing scope change and continuous
change management. EPC or lump sum turnkey (LSTK)
contracts require much more rigid scope management and
change is managed on a contractual basis (rather than a ‘trend
register’ used for reimbursable EPCM projects). Hence, the
EPC approach gives cost surety, but is less flexible than the
EPCM approach, and quality outcomes can be sacrificed in a
poorly managed EPC environment.
Operations management is about managing continual
change on a relatively micro scale. Trends are compared
with annual budgets in a personalised environment that is
subject to continual change. Whilst the annual budget sets
a time frame for performance (and reward), the operating
environment requires management on a long-term horizon
to optimise project net present value (NPV) and other issues
such as community relations.
In a simplistic sense, there is a continuum of behavioural
environments that stretch from the microscale, continuous
improvement environment of operations, through various
forms for collaborative and reimbursable project management,
to the heavily contractual and short-term focus of EPC/LSTK
delivery. For the purposes of simplicity, we have characterised
this continuum using the three aforementioned examples:
the technical scope and simplifying and clarifying the interfaces
between the service providers.
KEY REQUIREMENTS OF A GOOD OPERATIONS MANAGER
An operations manager’s performance is measured against
a budget that is usually based on previous performance
(benchmarked or historical). Behaviour is heavily long-term
people oriented with a clear focus on incremental operating
and maintenance costs. This results in a thinking process,
which is heavily expenditure-driven and uses tools, such as
systems applications and products (SAP) in data processing,
to focus on cost management.
Incremental improvement is a critical part of operations
management. Reducing the time taken for effective
maintenance increases plant availability and revenue. The
balance between maintenance and availability is often the
key focus of the incremental improvement process and the
management process is more ‘microscale’.
In this process, the speed of decision-making is not critical.
Making a quick decision may result in an adverse outcome
that is easily compared with expected budget outcomes. In
this environment, it is more important to make the right or
optimised decision, than the quick decision.
Because of the long-term focus on availability and ease of
maintenance, operations personnel are typically much more
focused on installations designed for long-term performance
than on project deadlines and delivery to scope and budget.
1. operations management
KEY REQUIREMENTS OF A GOOD PROJECT MANAGER
2. EPCM reimbursable project management
Due to the nature of their roles, project managers are more
transient and less risk adverse. They are used to spending
large amounts of money quickly in order expedite the job or
maintain the schedule. Schedule often takes precedence over
cost, in part due to the high cost of delayed production.
3. EPC or LSTK project management.
The EPCM reimbursable approach also includes the
owner’s self-perform approach where an owner’s team takes
on the role of EPCM management rather than contracting to
a separate party.
PROJECT ENVIRONMENTS
In addition to the three management scenarios listed in the
previous section there are other factors that influence the
type of behaviour required for a successful project outcome.
These include:
•• greenfields and brownfields project environments –
brownfields projects typically require greater operations
input
•• small or large owner’s team environments – larger
owner’s teams typically include more operations and
maintenance experience/focus
•• conventional or novel technology environments –
novel technologies require an additional level of risk
management through test work, scale-up, engineering
design and start-up and are accompanied by increased
complexity in project relationship management.
A well-defined (and gated) greenfields project utilising
conventional technology with a small owner’s team is much
more likely to be a successful project than a brownfields project
with a large owner’s team using novel technology, simply due
to the complexity of the latter case and the number of interfaces
that require managing (personnel and technical). Clear
project definition becomes increasingly more important as the
complexity of the project increases. Clear project definition is
not a function of increased management, increased complexity
of procedures or layering of approvals. Clear project definition
is brought about by reducing the uncertainties associated with
188
A good project manager is milestone driven and requires
a management tool that is heavily focused on schedule and
work breakdown structure (WBS). Hence, SAP systems are
not appropriate and are replaced by project delivery systems
and other suitable workflow management tools that are cost
and schedule focused.
Time is a key differentiator between operations thinking
and project management thinking. During a design phase,
a typical medium sized copper concentrator project will
burn $2 M to $4 M per month. This excludes the cost of lost
production due to delayed delivery. Hence, delays in project
decision-making and poor decisions that result in rework
need to be avoided.
Micromanagement is to be avoided and it is important that
decision-making authorities are set at the lowest practical and
effective level. This requires an experienced team capable of
managing issues at the lowest level. Without a high quality
team, project managers find more and more of their time
taken up by low-level issues, to the extent that the focus on
the broader issues may be lost. This is a major challenge in the
current project environment as the large number of projects
and lack of experienced personnel mean there is little chance
of truly having an experienced team on your project.
If people from an ‘operations environment’ fill these project
management positions without adjusting their behaviour, the
result is extended schedule and increased cost.
Unlike operations managers, project managers are not
interested in process or maintenance optimisation once the
scope of work and services is agreed. Generally, a successful
project has a high-level of definition during its design phases
We are metallurgists, not magicians
Operations versus projects – how do people think and what are the implications?
and manages project execution to a defined scope. Variations
cost time (and money). One small design change, such as a
change in valve size may require modification of 50 engineering
documents. With associated approvals and sign-offs this will
cost much more than the new valve. Hence, whilst ‘fit for
purpose’ is a paramount criterion, changes that refine this
concept are not critical in a project manager’s paradigm.
The contractual and cost environment for a project manager
is demanding. Performance against contract timelines is
critical and if an existing strategy needs to be changed,
alternative suppliers are not easily sourced. Similar strategies
as those of an operations manager may be engaged, for
example using two suppliers for one commodity, but the
complexities of transferring structural steel fabrication or
erection scope from one supplier to another are far greater
than, for example, sourcing an alternative grinding media
supplier (in most cases).
The implications of the above for an EPC contract are
greatly magnified. Profit in a competitive bid environment
is driven by completion ahead of schedule, cutting corners
and maximising the margin in variations to the contract.
This environment magnifies the differences between the
operations paradigm and the project management paradigm.
OPERATIONS INVOLVEMENT – WHEN AND HOW MUCH?
Key questions from a project owner’s perspective are:
•• When is the best time to freeze design?
•• What defines a change outside the defined scope?
At some point during project execution the ambitions
of operations and the project collide. From the operations
perspective, operating knowledge needs to be embedded in
the project so that the plant can be effectively maintained
and operated. However, from the project perspective, early
involvement of the operations team will lead to increased
capital cost and extensions to schedule.
From both owner’s and EPCM contractor’s perspectives,
one of the biggest issues is when to get the operations
representatives involved in the project. If the operations
and maintenance personnel are brought on too late in the
execution process their input is not able to be included in
the design. If they are brought on too early and the input is
not critically assessed against project needs, the project cost
generally increases. The best approach is to have plenty of
operations and maintenance input during the feasibility stages
of the project. This is much more easily done on a brownfields
project or where the owner has existing operations, because the
operations and maintenance groups are well established and
are aware of the types of issues created by existing designs.
In general, any change creates rework and can delay the
project. The design usually needs to be frozen in parallel with
the receipt of vendor data, at about 20–25 per cent engineering
complete. After this point even minor changes result in
numerous changes to documents. This is often an issue at
the time when the client’s operating team joins a project as
their preferences and paradigms are often not met by an
existing design. The resultant tension needs to be carefully
managed. The first stage of managing this process is mutual
understanding of the needs of each party. This was one of the
drivers for this paper.
Operations team involvement is most effective:
•• during the project feasibility study when the operations
wish list can be assessed based on a cost/benefit analysis
without impacting on project schedule
We are metallurgists, not magicians
•• during project review to check for any fatal flaws that
may creep in during detailed design
•• as part of the operations readiness planning where
‘mod-squad’ projects can be identified early and plans
put in place to address serious defects.
WHAT CAN GO WRONG?
Input from operations is crucial to ensure rigorous project
design and without it projects can encounter trouble over issues
that could easily have been avoided. For example, poor chute
designs can create plugging segregation, or high-wear areas.
More critical are designs that have not considered maintenance
access to equipment. In some cases, poor access to maintenance
equipment can create safety hazards where maintenance teams
have to handle heavy materials in areas with little room to
manoeuvre. Finally, it is important to review the flow sheet
and layout for ‘fatal flaws’, ie design flaws that will prevent the
plant from performing as desired. The best time to catch these
is in the prefeasibility phase or at the latest during the feasibility
study. Safety issues and lack of access are issues where it can
be necessary to make design modifications post-20 per cent
engineering complete, even if it impacts the schedule. Hence,
it is important to include operations and maintenance thinking
prior to freezing the design.
For greenfields projects it is usually more difficult to
finalise operator input because the operations team is often
not assembled until engineering has begun, or sometimes
during construction. In these cases the success of the design
is dependent on the quality of the design team. For the best
results, it makes sense to have people on the design team that
have been involved in the post-commissioning, operations
and maintenance processes, but who understand the project
implications. Otherwise designs can be very logical, but not
consider common operations/maintenance issues.
Above all, clarity regarding cause and effect is a key
requirement to manage change during the reliability,
accessibility, maintainability, build-ability and operability
(RAMBO) process. ‘Wish list’ changes need to be differentiated
from fatal flaws in the design. Strong and consensus
leadership from the owner’s and Contractor’s teams is
necessary to ensure that design change is minimised once
critical commitments are made. Designers must understand
the impact of their design on plant operation and operations
personnel must understand the cost and schedule (cash flow)
impact of design change.
Another common problem, especially in greenfields projects,
is that operational maintenance management systems are
implemented too late in the process or are incomplete. This
can cause many problems during the commissioning phase
because critical spare parts or warehouse stocks may not be
available. This problem can be magnified if the feasibility
study does not consider or underestimates the cost and time
involved for operating spares. Cutting operational spares is
an easy way to reduce capital spending, but is a false economy
if operating time is lower during the commissioning/early
operating phase of the project.
There is nearly always a tension on the project between what
the operational team would like to have, what the project
can afford and what the designer would like to supply. The
degree of tension is a function of the contract type. Fixed
price contracts necessitate a highly contractual (and clear)
decision-making and change control process. This makes
owner input more difficult once contracts are let and requires
a more detailed and extended design period with associated
project schedule risk. Reimbursable contracts tend to be less
189
G Lane and B Clements
rigorously managed and even though change control is in
place, issues can arise due to poor decision-making processes
and rework and wasted time within the project. Projects with
large owner’s teams tend to have greater issues with decisionmaking than projects with small owner’s teams. Hence,
major mining houses have more issues controlling their
project budgets and schedules than junior mining houses
(irrespective of the size of the project).
CASE STUDIES
These case studies are based on the authors’ recent experiences
with major projects. They represent a contractors’ perspective
and an owners’ perspective of typical issues that arise during
project execution, mostly due to the way the various parties
react to project issues.
Contractor’s perspective
From an EPCM contractor’s perspective there are three broad
categories of project owner:
1. major mining house
2. mid-tier mining house
3. junior mining house (and explorers).
The extent of owner input, review and approval process
differs across this spectrum with major mining houses being
the most demanding and complex in terms of owner’s team
interface and impact on the control of project schedule and
cost. The following case study pertains to an experience with
a project with an owner in the range of mid-tier to major
mining house.
Cost-effective design is the aspiration of most project owners.
Several years ago when considering project development
options for a major new mine, discussions between the owner
and the EPCM contractor occurred around the nature of
the plant design for a new processing plant. The owner had
several operating plants and one of these was in the vicinity
of the new project. The agreed concept at executive level
was to design and build a ‘cost-effective’ plant and critically
assess operations and maintenance input to the design prior
to inclusion of ‘nice to haves’ in the design.
Several option studies were completed that focused on
the process flow sheet and whether to upgrade a nearby
concentrator to treat the ore from the new mine or build a
new stand-alone plant. The upgrade option (brownfield)
allowed the use of existing capital equipment and associated
infrastructure. The stand-alone plant option (greenfield)
allowed some of the existing paradigms to be tested, so both
had pros and cons.
there was substantial cost escalation associated with the plant
design in the transition from option study to execution that
was related to the growth in bulk quantities in the concentrator
design to meet the expectations of the contractor’s and
owner’s teams as they evolved.
Construction and commissioning of the brownfields
project adversely impacted on the operating environment
and production from the existing plant. Although this was
anticipated, the extent of the disruption caused by the addition
of new front end processes was underestimated. This was
particularly true of the amount of relearning required by the
operators as the plant changed over a period of 12 months.
Lessons learnt included:
•• keep the project execution plan simple and minimise the
contractor’s interfaces
•• clearly identify and continually communicate the vision
for the project
•• critically review and assess the project direction and
make courageous decisions (as a team) and thus avoid
the need to ‘cost-cut’ to maintain budget
•• unless tightly controlled by the project and owner’s
project directors, brownfields projects will maintain the
paradigm set by the plant operators
•• brownfields projects impact adversely and materially on
existing plant operations and are harder to commission
and ramp-up than a separate stand-alone (pseudogreenfields project).
Owner’s perspective
As previously stated, project schedule is arguably the most
critical factor in determining the success of a project. This
argument is supported by the economic drivers of almost
every project. A large-scale concentrator project has a high
capital cost and the best economic return of the project is
achieved by completing the project on time and ramping it
up smoothly to design productivity, including both design
performance of the equipment as well as design asset
efficiencies. So it is important, once the project scope is locked
down, for the operating and maintenance team to focus on
ensuring a smooth start-up.
A complex execution strategy further complicated the
execution phase and made project communications and
interfaces more complex.
One of the most valuable activities for the operations/
maintenance team to engage in is what is commonly called
‘operational readiness’. This involves ensuring that all of the
necessary resources are ready for the handover of the plant.
This involves hiring and training of the team, assembling the
required operational spares, and developing the maintenance
system such that surprises can be minimised. It is essential
during the commissioning phase that routine maintenance
activities are handled effectively. The best outcome is when
these activities can be planned into the scheduled outages
required by vendor run-ins. What needs to be avoided is
having a critical piece of equipment breakdown and not to
have the required materials on hand to complete the repairs
efficiently. Since cash flow is so critical during this period, lots
of time and money can be wasted in having to rush materials
in, including priority freight etc. In the worst-case, when
starting a project in a remote location, parts can take weeks
or longer to arrive on-site. The effort of finding work-arounds
for these issues is draining, and the lost opportunity can be
financially disastrous to the project.
The final outcome was a plant design that was robust and
flexible rather than ‘lean and mean’. Given the cost profile
of the project where the process plant costs are small when
compared with the mine infrastructure, this outcome is
reasonable in the context of the overall project. However,
In this case study, the operators’ team was involved early
on with the feasibility design. As a result, once the scope
was locked down, they were able to focus on the operational
readiness phase of the project. The start-up team was hired
and they used a local technical school to train a large number
As the work progressed from option studies to prefeasibility
to feasibility to implementation, the objective of establishing a
new paradigm became more and more diluted as new people
came onto the project. New people in the EPCM contractor’s
team brought their design experiences and preferences that
tended to be conservative rather than aggressive. As the
owner’s team grew, influence from the operations side of
the owner’s business increased and the pre-existing design
paradigm re-established itself within the project.
190
We are metallurgists, not magicians
Operations versus projects – how do people think and what are the implications?
of workers to be used in both operating and maintenance
activities. They also ensured a sufficient number of experienced
workers were on-board (hired from other operations) to operate
and maintain the equipment effectively. Numerous operating
and maintenance manuals were used to gain awareness of the
specific equipment used in the plant. One of the challenges in
this particular start-up was that the plant was designed with
equipment that was novel to almost everyone on the team,
so there were many unknowns creating distractions from the
establishment of routine maintenance activity.
Another challenge encountered in this case was that despite
it being a brownfields project, the projected plant was a
concentrator being constructed on a site with an existing
leaching plant. This minimised the amount of site knowledge
available to enhance the design. In addition, the projected
plant was much larger than the existing operation, so existing
resources were ill-prepared for the size and scope of the project.
Whatever the causes, the maintenance systems for the
new plant were not completely addressed during the
precommissioning phase. This made the routine maintenance
functions much more difficult than usual, because the
maintenance team were not always able to find the necessary
materials to repair the equipment. This resulted in many
instances of having to use suboptimum materials to repair
equipment or to rush materials to site, including air-freight in
some instances, in order to keep the equipment operating. Since
the correct materials were not always available, and because
of the lack of experience with the high-wear environment
in a concentrator, ‘breakdown maintenance’ (rather than
preventative maintenance) became more common, reducing
operating time of the plant.
Over time, the systems were implemented and the team
was able to install high-level maintenance systems including
reliability centred maintenance (RCM) techniques, predictive
maintenance etc. As a result the plant is currently achieving
excellent asset efficiency numbers. However, the same results
could have been achieved earlier if the maintenance systems
had been in place prior to the always-hectic commissioning
period.
In the end, a very successful debottlenecking project was
implemented. Many of the experiences and lessons from
commissioning and early operation were used in this project.
Some of those were probably thought of during the later
engineering phases and construction. But it is unlikely that
many of these plant modifications could have been imagined
prior to start-up. So, trying to implement these ideas during
the engineering phase could have improved some areas
marginally, but would have been costly, and a debottlenecking
phase would still have been necessary.
Lessons learnt:
•• experienced personnel who can focus on the operations
and maintenance needs and not be distracted by the
construction phase of the project
•• a dedicated budget.
•• Project debottlenecking is an important part of project
development and can’t be avoided by late design
changes during the initial project execution.
•• Plants containing new technology are more complex to
design, engineer, construct and commission due to the
learning curve.
CONCLUSIONS
In a recent communication, a senior project executive from
an owner’s team commented: ‘lessons are never learnt’ and
without an ‘A-team’, project teams will observe ‘that stuff
will never happen to me’ and ignore wisdom gained from
experience.
This paper states ‘the obvious’ in many ways, but ‘the
obvious’ only becomes obvious through experience. The
pressure on the current project development environment
means that experience can be in short supply. The behaviours
of, and interactions between, key project personnel drive the
outcomes of projects.
Project managers need clarity and a clear path to timely
completion.
Operations managers need a plant that is easily operated
and maintained, supported by a robust operational readiness
program.
The tension between the two sets of needs can add
tremendous value to a well-managed project when all those
contributing to a project understand and respect the diverse
requirements of a successful project.
ACKNOWLEDGEMENTS
The authors would like to acknowledge Ausenco and
Freeport McMoRan for permission to publish this paper and
the various people who contributed insights from within and
outside these and other organisations.
REFERENCES
Biery, F, Hollands, A and Young, R, 2009. Minerals and metals project
performance and improvement opportunities, in Proceedings
Project Evaluation 2009, pp 21–26 (The Australasian Institute of
Mining and Metallurgy: Melbourne).
Gabrielson, A, 2007. Eye on business – Current trends in project
delivery, CIM Magazine/Bulletin, 2(7).
Hundertmark, T, Siliva, A and Shulamn, J, 2008. Managing capital
projects for competitive advantage, McKinsey Quarterly, June.
•• Operational readiness requires:
We are metallurgists, not magicians
191
Contents
Performance testing – when, what and how?
G Lane1, M Davis2, E McLean3 and J Fleay4
ABSTRACT
Performance testing is often used by project owners and financiers to provide
protection from, or assessment of, technical aspects of project design, procurement,
construction and commissioning. The performance tests can range from comprehensive
performance warranties relating to process performance and production in the case
of lump sum turn key (LSTK) projects to less onerous demonstration tests for cost
reimbursable contracts.
This paper will discuss performance testing from both the contractor’s and owner’s
perspectives and focuses on:
•• when performance tests are required
•• what type of performance test/warranty is optimum for each type of project
•• how project performance should be evaluated.
The requirement for performance testing can be driven by the project financiers,
particularly in the case of smaller resource development companies. In such cases,
differentiation between owner’s risk and contractor’s risk can become contentious. At
the other end of the spectrum, larger mining houses may choose to heavily influence
the engineering and design process with the contractor supplying services on a
reimbursable basis. In that case, the performance testing process is often limited to
demonstration tests that validate performance for the owner’s corporate governance
requirements.
This paper was prepared due to the lack of published benchmarking of performance
testing and the resultant differences in expectations between a contractor and a project
owner that often surface late in the contract negotiation process. The paper provides
a checklist of requirements for both the owner and the contractor that may be used
to provide ‘common purpose’ early in the contract negotiation process.
INTRODUCTION
Process plant performance testing is often used by project owners and financiers
to provide protection from, or assessment of, technical aspects of project design,
procurement, construction and commissioning.
This paper was prepared in response to the lack of benchmarking of performance
testing and the resultant differences in expectations between technology vendors,
contractors, project owners and financiers that often surface late in the project
contract negotiation process.
The discussion pertains mainly to minerals processing projects, for example gold
plants, base metals concentrators and hydrometallurgical plants, and minerals sands
plants.
1. FAusIMM, Chief Technical Officer, Ausenco
Minerals & Metals, South Brisbane Qld 4101.
Email: greg.lane@ausenco.com
2. FAusIMM, Managing Director, Sivad Resources
Pty Ltd, Darlington WA 6070.
Email: michael.davis@sivadresources.com.au
3. FAusIMM, Manager Minerals Consulting,
Ausenco Minerals & Metals, South Brisbane Qld
4101. Email: eddie.mclean@ausenco.com
4. FAusIMM, Manager Metallurgy, Minnovo Pty
Ltd, Leederville WA 6007.
Email: john.fleay@minnovo.com.au
Performance testing obligations vary with the type of contact. For LSTK work,
performance warranties and resulting performance test requirements are used
to define the minimum performance required for a plant or project soon after
commissioning. For engineering, procurement, construction and management
(EPCM) work, performance testing is typically conducted within 12 months of the
completion of plant commissioning.
Achieving minimum performance is particularly important for LSTK contracts
where a contractor is generally given more freedom inflow sheet definition and
equipment selection, and uses this freedom to optimise between plant performance
and plant capital cost. For reimbursable contracts, where the owner’s input into plant
design can be extensive, the role of performance testing is more complex, and the
testing process can be linked to contractual bonuses and/or be used to protect the
contractor from ad hoc claims by the owner.
Performance tests are expensive to undertake and should only ever be conducted if
aspects of the plant or project, during the pretrial period post-ore commissioning, are
underperforming with respect to the warranted criteria. Poor performance can be due
193
G Lane et al
to obligations not being met by the project owner as well as
poor performance by the contractor. A clearer understanding
of the issues may result in additional and more cost-effective
rectification work being completed prior to a reassessment of
whether actual performance testing is required.
This paper discusses performance testing from the
technology vendor’s, contractor’s and owner’s perspectives.
The testing processes associated with plant preoperational
testing and commissioning are not specifically addressed,
other than where they interface with performance testing
post-commissioning.
HISTORICAL PERSPECTIVE IN AUSTRALIA
From a historical perspective, performance warranties played
a notable role in the early development of some major mining
companies and the, now ‘mid-tier’, Australian engineering
companies. Close (2002), in reporting the recent history of the
development of the Australian gold mining industry, noted
that the developing Australian engineering companies in the
1980s:
… provided a completion guarantee that the plant would
function to the specifications set out in the contract
that … this significantly reduced the risk for project owners
and made it easier for them to procure funding …
… significantly reduced the need for technical expertise in
the newly emerging gold mining companies ... and that ...
… the large firms were simply unable to compete with the
keen pricing and aggressive smaller local suppliers for the
design and construction of smaller plants. Neither were the
large traditional design and construction engineering firms
such as Fluor, Davy McKee, Bechtel and Dravo willing to
provide the completion and performance guarantees and
the financing options that the smaller companies such as
Minproc offered their clients.
The early 1980s saw the emergence of the LSTK contract and
accompanying performance warranty in preference to the
traditional EPCM contract. These were generally preferred
and commonly used by emerging, junior and new mid-tier
companies. Typically the project was their first, or among
the first of their ‘company makers’. Nearly all were for gold
projects.
In the 1990s, the proportion of LSTK projects decreased
and there was a return to EPCM type contracts favoured
by the newly established companies and major mining
houses. These companies had larger technical resource bases,
either in-house or at-call from their portfolio of operations.
Projects were typically carried out by personnel with project
development experience (from the LSTK days). Many of the
aspects of performance warranties were retained in the EPCM
work. This period saw extension of performance warranties to
base metal projects.
Toward the end of the 1990s, several nickel laterite projects
were developed with varying contract types and associated
warranties. Despite relatively onerous warranties and claims
on some of the engineers, the project owners and shareholders
were not protected from financial distress following poor initial
performance of some of these operations, whatever the cause.
The trend in the first part of the 2000s has seen a continued
dominance in EPCM type jobs with increased technical
participation and project direction by owners as well as
a tightening of expectations and expansion of scope for
performance warranties by owners and financiers. This
period has seen further extension of performance warranties
to the mineral sands industry.
194
Over a 25 year period from the early 1980s to the current
time, the use and application of performance warranties has
departed significantly with respect to the purpose and intent
from those initially used by the industry.
Of course, performance testing has been used for many
years to ensure unit process performance where technology
packages have been provided by vendors to projects, no
matter what the size and style of the project.
AIMS OF PERFORMANCE TESTING
A performance warranty, and the associated testing process,
is a method of risk management and quality control for
project development.
The requirement for performance testing can be driven by:
•• the project financiers, particularly in the case of smaller
resource development companies
•• larger mining houses where the testing process is often
limited to demonstration tests that validate performance
for the owner’s corporate governance requirements
•• engineering contractors where a vertical package LSTK
subcontracting approach is used
•• engineering contractors where equipment performance
criteria, such as mill power draw, are critical components
of a vendor supply package.
It is not uncommon for junior resource companies and their
financiers to attempt to manage process risk by using the
LSTK contracting approach and passing on as much of the
process risk to the contractor as possible. This approach is only
successful if the project and process design criteria have been
adequately defined in preceding feasibility study assessments
and the ore reserve and mine plan are well defined.
TYPES OF TESTS
The types of performance tests range from comprehensive
performance warranties relating to process performance
and production in the case of LSTK projects, to less onerous
demonstration tests of major equipment for reimbursable
contracts.
The consequences of failing a performance test vary and are
typically defined in the head contract or term sheet. Failure
by the contractor can require full rectification in the case of
LSTK works or re-performance of services in the case of a
reimbursable EPCM contract. The consequence of test failure
by equipment vendors is typically limited to the maximum
liability under the liquated damages clause of the contract.
The styles and types of tests are numerous. However, to
simplify matters for discussion, the performance tests can be
categorised into:
•• financier’s tests – typically between a financing body
and the project owner and linked to the conditions of
the loan (for example interest rate) to the project owner
•• LSTK performance tests – typically between the project
owner and the LSTK contractor
•• EPCM performance tests – typically between the project
owner and the EPCM contractor for reimbursable
contracts
•• vendor tests – required from major equipment vendors
to demonstrate that the supplied equipment performs in
accordance with the contract (for example power draw
for grinding mills).
Each of these is discussed in further detail in the following
sections.
We are metallurgists, not magicians
Performance testing – when, what and how?
FINANCIER’S TESTS
Financier’s or banker’s performance tests can be the most
extensive form of performance tests, as they can cover an
extensive range of project activities. These tests are typically
linked to the financing conditions for a project with a pass
resulting in reduction in project risk and a lower interest rate
for project financing to the project owner. The following list is
indicative of the type and extent of tests that may be required:
•• mine reserve test – validation of the mine reserve estimate
based on early mine development and production
•• ore to waste ratio test – based on early mine performance
comparing the basis of the financial model with mine
practice
•• ore delivery test – based on early mine performance
comparing the predicted delivery rates with mine practice
•• operating cost tests – based on early project performance
comparing the basis of the financial model with actual
project performance
•• environmental/social compliance – based on project
performance comparing the environmental/social
management plan with actual project performance
•• management and staffing – comparing actual management
and staffing levels with planned performance; both under
and over manning may require explanation
•• contract compliance and close-out – evaluation of
whether financial obligations have been met and
contracts closed in a satisfactory manner
•• financial management tests – assessment of project
financial management processes
•• product quantity and quality tests – assessment of
plant production and product quality compared with
forecasts, typically with a focus on measures of plant
throughput and metal recovery and quality
•• insurances in place – assessment of whether appropriate
insurances are in place
•• marketing tests – assessment of whether product
marketing systems are in place and capable of marketing
product as planned
•• power supply test – assessment of whether power
supply provides adequate starting and operating
capacity and reliability
•• water supply test – assessment of whether water supply
provides adequate operating capacity and reliability
•• tailings disposal test – assessment of tailings storage
performance against plan with a focus on settled
densities and impacts on any indicated sustaining
capital requirements above plan.
There are occasional attempts to ‘back to back’ as many
as possible of the financier’s tests requirements with LSTK
contractor performance. However, in practice, many of the
above tests are related to aspects of the project that are not
within the contractor’s scope, for example how well the
project was assessed in the feasibility stages, and how well
the plant is operated post commissioning. As a result, ‘back
to back’ or tripartite agreements offer little advantage other
than to bring the interested parties together (often at a stage
that is too late for beneficial effect).
The cost of financier’s performance testing is met by the
project owner.
LUMP SUM TURN KEY PERFORMANCE TESTS
LSTK or vertical contracts are used in the following cases:
We are metallurgists, not magicians
•• where the owner, typically a junior company, does not
have adequate technical expertise and experience to
manage a reimbursable contract
•• where a financing institution insists on a LSTK contract
for the above reasons
•• where a separable portion of a larger contract can be
identified and a specialist contractor can supply a
packaged plant at a lower cost to the project.
In all the above cases, the bidding process will force the
contractor to cut costs by simplifying the process flow sheet,
simplifying the engineering and construction services and by
procuring low-cost equipment, where possible.
The aim of the performance warranty is to ensure that the
outcome of the LSTK contract is a plant that delivers the
required performance. The performance may be measured by
items such as:
•• plant throughput
•• grind size (P )
80
•• for a gold plant:
•• leach efficiency
•• adsorption efficiency
•• elution circuit efficiency
•• WAD cyanide reduction circuit efficiency.
•• for a concentrator:
•• recovery of metal(s) to concentrate
•• concentrate grade
•• filtration circuit performance.
•• process or project specific performance criteria.
Thus, the LSTK performance warranty provides a measurable
quality control step at contract completion. In the 1980s, it was
relatively common for the performance tests to be completed
even though the plant was meeting performance targets.
More recently, best practice is to demonstrate performance
targets have been achieved using plant operating records and
waive the performance tests, thus saving on the testing costs
and upset to overall project operation.
The costs associated with LSTK performance tests in the first
instance are typically met by the project owner. Repeat test
costs are typically met by the party at fault in the preceding test.
With LSTK performance warranties the consequence of
failure by the contractor typically requires the contractor to
complete all necessary rectification works at no cost to the
owner.
ENGINEERING, PROCUREMENT, CONSTRUCTION
AND MANAGEMENT PERFORMANCE TESTS
EPCM performance warranties have a much reduced
consequence for the contractor as failure to achieve required
performance typically requires that the contractor provide only
the EPCM services for rectification at no cost to the owner.
Reimbursable EPCM contracts typically have substantially
more input from the owner into the flow sheet, process and
plant design and engineering. Plant equipment is typically
purchased by the owner and all drawings are reviewed and
signed-off by the owner.
As a consequence, poor plant performance that results
from a reimbursable EPCM contract is the outcome of shared
responsibility, with the owner and contractor contributing to
the outcome.
There is a recent trend toward EPCM project owners
asking for more detailed and onerous warranties, while still
195
G Lane et al
having substantial input to design and equipment selection.
This can lead to conflict during the design phase where the
engineer may raise issues with client design or equipment
preferences and seek written instructions from the client that
effectively waive the contractor’s warranty obligations. This
process needs to be thought through by both parties prior to
commencement of the project.
•• additional definitions
The bulk of the risk associated with poor performance of
this type of contract is with the project owner in the form
of lower than expected cash flow due to reduced throughput
or lower than expected metal recovery. Performance
warranties are typically aimed at ensuring that materials
handling systems (conveyors, pumps and pipelines) and
services (water and power systems) have adequate capacity.
Thus, the performance tests are restricted to throughput and
capacity tests rather than metallurgical performance tests
based on metal recovery or product quality.
•• test reporting and warranty claims
EPCM performance warranties are often associated with a
‘fee at risk’. In many cases the project owner may do better
to expend the cost of the fee on quality control and technical
review during the EPCM phase rather than entering into
performance warranties with a contractor.
VENDOR TESTS
Vendor warranty tests include all tests specified in contracts
between the owner or contractor and equipment vendors
for major equipment items, such as mills and filtration
equipment. The contractor typically supervises the conduct
and assessment of these tests and provides a written report to
the owner on the outcome of each test.
Typical vendor warranties are listed below:
•• primary crusher – throughput
•• semi-autogenous grinding (SAG) mill – power draw
•• ball mill – power draw
•• regrind mills – power draw
•• flotation cells – solid suspension and air dispersion
•• thickeners – overflow clarity
•• filters – throughput and moisture content of filter cake
•• blower performance
•• vendor package performance – for example lime slaking
and flocculant system.
Some of the listed vendor warranties require that test work
is conducted on ‘representative samples’ to enable the vendor
to size the equipment and measure commercial risk.
Vendor warranties enable the process risk to be shared
but do not necessarily obviate the owner’s risk. If the
representative nature of the sample tested is poor, lack of
technical precedents and poor risk management can result in
relatively minor cost to a vendor and large ongoing losses to
the project owner until the process is upgraded or rectified.
STRUCTURE OF PERFORMANCE WARRANTIES
AND CONTRACT ISSUES
The structure of the performance warranty can be made
reasonably consistent across all contract types and
circumstances. The performance warranty document is
typically a schedule to the contract documentation. As such,
the principal definitions and references are in the contract
rather than the performance warranty schedule. There a
number of key components to the performance warranty
schedule, typically consisting of:
•• introduction
196
•• warrantor’s obligations and warranties
•• warrantee’s obligations and warranties
•• procedures for testing (sampling, characterisation, data
capture and data analysis)
•• performance assessment and pass criteria
•• limits of liability
•• arbitration.
Aspects such as payment of costs, access to data and access
to site are best addressed under the obligations section of the
document.
A few ‘rules of thumb’ apply:
•• keep the document simple
•• don’t repeat information that can be referenced to the
main contract
•• don’t refer to other documents, such as design criteria
unless they are attached as schedules to the contract
•• clearly state the obligations of the warrantor and
warrantee with respect to services and work required
prior to and during the performance tests
•• clearly quantify the warranties required of both the
warrantor and warrantee for each relevant component
of the plant and whether these warranties are to be met
independently or collectively; each warranty should be
associated with a clear definition of the data required to
demonstrate compliance
•• clearly state the period over which each performance
test shall be run
•• detailed procedures are not necessarily as long as a
method of agreement and arbitration; agree immediately
on detailed procedures prior to the commencement of
the performance test
•• clearly define the pass/fail criteria for each performance
test, including the timeline and process for test
completion, reporting of warranty claims, arbitration
and rectification activities
•• refer to limits of liability as stated in the contract.
PERFORMANCE EVALUATION
The desired outcome for performance warranties is that all
tests are waived by the warrantee on the basis that project
operating records, during the pretrial period post-ore
commissioning, demonstrate warranted performance by the
warrantor (Lane and Messenger, 2005). This avoids the cost
and imposition of performance testing for all parties involved.
The failure of a performance test is of little benefit to either
party unless the relationship between the parties has broken
down to an extent that obvious areas of underperformance
cannot be negotiated (in the absence of legal teams) to a
mutually successful conclusion. The effort associated with
undertaking the performance test is better directed at the
engineering solution.
Performance testing places significant obligations and
warrantee requirements on the project owner as well as
the contractor. Churchill and Lane (1997) noted that for an
EPCM contract:
(Successful) Performance testing is reliant on the owner
supplying suitable material for testing. Failure to do so
within a reasonable period will result in default completion
of performance testing.
We are metallurgists, not magicians
Performance testing – when, what and how?
Of course, the perspective given to performance data and
project outcomes differs between parties and as a consequence,
performance tests are, on occasion, necessary. To minimise
the likelihood of disagreement over whether the plant has
reached performance targets during the pretrial period, the
following process is recommended:
•• prior to plant commissioning, jointly agree the data
that is to be used for assessing performance compliance
and agree to collate the data on a periodic basis
•• periodically, jointly review and evaluate plant
performance data to determine the nature of any cause
of poor performance
•• as warranted levels of performance are met, each party
should inform the other of the compliance
•• any outstanding warranties should be discussed and
plans to rectify performance developed.
The above process differs from the more common outcome
for LSTK, EPCM and vendor contracts where the warrantor
demobilises from site and lays low waiting for the liability
period to lapse if there is an indication of poor project
performance. This approach involves less cost (for both
parties) in the short-term but can lead to acrimony and a more
costly outcome in the longer term.
TIMING AND DURATION OF PERFORMANCE TESTS
The timing and duration of plant performance tests varies with
the test type and parameters. Vendor equipment testing may
be conducted over short durations, for example grinding mill
power draw tests, while financier’s or banker’s performance
tests are often over extended periods (months) to allow
assessment of operational and financial performance.
LSTK performance testing is often carried out soon after
commissioning and prior to final handover of the facility to
the owner. The test durations are sufficient to demonstrate a
suitable level of operability and availability (usually seven to
14 days, sometimes longer). LSTK contracts in the late 1980s
and early 1990s required completion of performance tests
We are metallurgists, not magicians
prior to the contractor demobilising from site. This model
is suited to owners who are new to the mining business and
benefit from the handover of a near-fully operational plant.
Alternatively, both parties may be best served by divorcing
initial handover of plant operations and completion of
performance tests. In this case, demonstration of compliance
is based on plant operating data. The best model is usually
project specific and dependent on location, the availability
of technical support, owner’s experience and the project
contracting risk/reward model.
Reimbursable EPCM contract performance tests may be
completed after the plant has been operating for an extended
period, and typically within the first 12 months of operation.
As the objective is to demonstrate compliance with the tests’
objectives based on operating data, it suits all parties to work
toward test compliance without the cost and potential upset
to normal operation of actually running performance tests.
The durations of reimbursable EPCM contract performance
tests are typically shorter (24 hours) than LSTK contract
performance tests with the focus on throughput testing rather
than operational/availability issues. As discussed earlier, this
change of focus is due to the much greater owner input into
plant design and equipment selection.
A CHECKLIST FOR FINANCIERS, OWNERS AND CONTRACTORS
A checklist of requirements for each of the various types of
performance tests is provided in Table 1 over the page.
REFERENCES
Churchill, S and Lane, G, 1997. Effective commissioning, in Proceedings
Mindev 97 Conference (ed: E Barnes), p 247 (The Australasian
Institute of Mining and Metallurgy: Melbourne).
Close, S E, 2002. The Great Gold Renaissance, The Untold Story of
the Modern Australian Gold Boom 1982–2002, p 141 (Surbiton
Associates Pty Ltd).
Lane, G and Messenger, P, 2005. Commissioning, in Advances in Gold
Ore Processing (ed: M Adams), Volume 15, p 176 (Elsevier).
197
G Lane et al
TABLE 1
A checklist of requirements for each of the various types of performance tests.
Performance test type
Banker
Lump sum turn key
Reimbursable
Vendor
Parties
Banks and owner
Owner and contractor
Owner and contractor
Owner or contractor and
vendor
Participant
B
O
O
O/C
O
C
C
V
Document structure
Interface with head contract or term sheet




Description of purpose




Definitions and terminology




Warrantor’s obligations


Warrantee’s obligations
Warrantor’s warranties




Warrantee’s warranties




Procedures
• sampling and test duration
• characterisation test work
• data capture
• data analysis






Can be agreed at a later date and subject to arbitration.
Arbitrating party to be named in Performance Testing document.
Performance assessment and pass criteria




Test reporting and warranty claims




Limits of liability




†
Plant performance warranty types
Throughput


Grind size


Plant availability












Process metallurgy
• metal recovery
• concentrate grade (concentrators)
• solution losses (gold plants)
• concentrate moisture (concentrators)
• equipment performance
• rheology
• head grade
• ore characteristics
• elution efficiency
• cyanide destruction




Power draw or equipment specific performance





? (mills)






Non-process engineering related tests
Mine reserve test

Ore delivery test

Mine performance tests

Operating cost tests

Environmental compliance

Management and staffing

Contract compliance and close-out

Product quantity and quality tests

Financial management tests

Insurances in place

Marketing tests

Power supply test

Water supply test

Tailings disposal test

B = banker, O = owner, C = contractor, V = vendor. † – Typically covered in head contract.
198
We are metallurgists, not magicians
Unit design and
development
Contents
Process development and throughput forecasting
at the Phu Kham copper-gold operation, Laos PDR
D Bennett1, I Crnkovic2, P Walker3, A Hoyle4, A Tordoir5,
D La Rosa6, W Valery7 and K Duffy8
The Phu Kham deposit in Laos People’s Democratic Republic (PDR) represents a
highly variable and complex copper-gold porphyry system, and the original 12 Mt/a
concentrator commissioned in April 2008 lagged a long way behind most others in
terms of recovery of copper and gold into concentrate.
A focus during the original Phu Kham concentrator design stage was to counter
the expected metallurgical challenges by including first class instrumentation,
on-stream analysis and sampling systems to be able to provide process measurement
and collection of data. Once in production, gathering key variability data from daily
laboratory test work and monthly mineralogical composites helped to unlock the
secrets of the ore, and laid the foundations for the subsequent capital investments
in extra capacity and increased recovery. Process control opportunities have been
advanced in parallel to manage the ore variability and to further enhance throughput
and copper-gold concentrate production. The involvement of people with the right
skills and experience throughout all stages of the Phu Kham operations development
has always been considered as critical to success.
The results of all the process development and enhancements from commissioning
until June 2016 is presented in Figure 1, with throughput increasing from 12 Mt/a to
19 Mt/a, copper recovery increasing from approximately 60 per cent to 80 per cent,
and gold recovery increasing from approximately 35 per cent to over 50 per cent.
ABSTRACT
1. MAusIMM, Principal Metallurgist, PanAust
Limited, Fortitude Valley Qld 4006.
Email: duncan.bennett@panaust.com.au
2. MAusIMM, Manager Process Technical
Services, Ok Tedi Mining Limited, Tabubil,
Western Province, Papua New Guinea.
Email: ivan.crnkovic@oktedi.com
3. MAusIMM, Director, Minmet Services
Pty Ltd, Wilson Beach Qld 4800.
Email: peter29walker@yahoo.com.au
4. MAusIMM(CP), Manager Minerals and
Metals, Ausenco, San Isidro, Lima 27, Perú.
Email: andrew.hoyle@ausenco.com
5. Lead Drill and Blast, Group Mining Technical
and Sustainability, AngloAmerican, London
SW1Y 5AN, UK.
Email: alan.tordoir@angloamerican.com
6. Principal Mining Engineer, CRC ORE,
Queensland Centre for Advanced Technologies
(QCAT), Pullenvale Qld 4069.
Email: david.la.rosa@crcore.org.au
7. FAusIMM, Global Director – Consulting and
Technology, Hatch – Mining and Minerals
Processing, Brisbane Qld 4000.
Email: walter.valery@hatch.com
8. MAusIMM, Process Consultant / Mining and
Mineral Processing, Hatch, Brisbane Qld 4000.
Email: kristy.duffy@hatch.com
The Phu Kham deposit represents a copper-gold porphyry system, with mineralisation
present in skarn, stockwork and disseminated styles. Significant folding and alteration
events have created a complex heterogeneous geotechnical and mineralogy horizon,
which affect plant throughput and metallurgical performance. Weathering and water
table contact have created a leached zone, overlying transition zones with supergene
chalcocite-dominant secondary copper mineralisation and clay-rich gangue.
Primary ore copper mineralisation is mainly chalcopyrite with minor bornite. The
major challenges to the copper-gold flotation process are a wide size distribution of
chalcopyrite mineralisation and poor primary grind liberation, a high pyrite content
in skarn ore requiring aggressive pyrite depression conditions, clay-rich gangue and
non-sulfide copper mineralisation in weathered zones, and a significant association
of gold with pyrite.
The Phu Kham concentrator has been developed as a conventional semi-autogenous
grinding (SAG) and ball milling circuit followed by rougher flotation, regrinding and
cleaner flotation to produce a copper concentrate containing payable gold and silver
values. The original 12 Mt/a copper-gold concentrator flow sheet design offered a
capital efficient compromise between high copper recovery bulk sulfide flotation
with large cleaning capacity, and lower recovery copper selective rougher flotation to
ensure concentrate specification of 24 per cent copper grade could be achieved.
Phu Kham commenced production of copper-gold concentrate in April 2008.
Flotation copper and gold recovery from commissioning was consistently poor, due
to the high levels of pyrite and problematic secondary and oxide copper species and
non-sulfide gangue. Incremental improvements in copper recovery were achieved by
2011 through conversion of conditioning cells to flotation cells in both roughing and
first cleaning and increased cleaner capacity through the installation of a Jameson
Cell. In 2012 the Phu Kham Upgrade Project (PKU) was commissioned to increase
mill throughput, along with additional rougher and cleaner capacity to increase
residence times and maintain copper production at the nominal 16 Mt/a throughput.
The marginal reduction in primary grind size due to the additional ball milling power
also provided a small recovery improvement.
201
D Bennett et al
FIG 1 – Phu Kham mill throughput and copper and gold recovery from commissioning until June 2016.
The Increased Recovery Project (IRP) commissioned in April
2013 targeted the major causes of copper and gold loss from
the circuit. Based on fundamental and detailed test work and
mineralogical analysis of concentrator streams, an optimised
design was developed which has increased copper and gold
recoveries by over five per cent and ten per cent respectively.
The IRP positioned the concentrator to process high pyrite
ores that were previously considered untreatable, and opened
further opportunities for maximising flotation recovery by
debottlenecking plant regrind and cleaning capacity.
To evaluate how to maintain design throughput rates with
increasing ore hardness over the life-of-mine (LOM), Phu
Bia Mining Limited conducted a throughput forecasting and
optimisation project for Phu Kham with the assistance of
Metso Process Technology and Innovation (PTI). The project
involved a review of the blasting, crushing and grinding
processes and development of a throughput prediction model
based on geometallurgical modelling for long-term planning.
The scope also included identifying opportunities for
increasing throughput and improving overall comminution
circuit performance when treating the most competent ore
types. A shorter term objective was to identify if and when
secondary crushing or other process changes will be required
to maintain the target throughput over the LOM.
This paper examines and discusses process and concentrator
flow sheet development, including projects implemented
since commissioning in 2008 to improve throughput and
copper and gold recovery with decreasing copper grade and
increasing pyrite content of ore feed, and increasing hardness
of primary ore.
Introduction
The Phu Kham operation consists of a copper-gold mine using
conventional shovel mining and truck haulage to a flotation
202
concentrator. The project is owned and operated by Phu Bia
Mining Limited. PanAust Ltd based in Brisbane, Australia
holds a 90 per cent interest in PBM through its wholly
owned subsidiary Pan Mekong Exploration Pty Ltd, with the
remaining ten per cent held by the Government of Laos PDR.
The Phu Kham copper-gold deposit is located in the
Xaisoumboun province as shown in Figure 2, approximately
120 km north of the Lao capital Vientiane. Access to the mine
is approximately four hours by road from Vientiane.
The Phu Kham 12 Mt/a concentrator was commissioned in
2008 for a capital cost of approximately US$150 M, placing it
in the lowest quartile for capital intensity for copper mineral
processing projects.
The installed plant was a compromise between a high
recovery but high capital intensity design, and a lower
recovery but technically lower risk and low capital intensity
design. The selective rougher flotation design was driven by
the complex and variable mineralogy and high pyrite content,
with over 90 per cent of pyrite required to be rejected in order
to produce a final concentrate of over 23 per cent copper.
With increasing depth of the pit since the commencement of
operations, the weathering profile of the feed has changed
such that the ore became primary dominant in 2010, with
chalcopyrite the main copper sulfide mineral. The complex
folding and alteration of the ore zones has meant continued
mining of supergene and oxidised areas within the pit, with
the copper mineralogy remaining diverse and varying from
native, oxide, secondary and primary copper species within
short time periods.
The development of the Phu Kham flow sheet was driven by
the poor recovery in comparison to other low-grade copper-gold
ores, and a need to counter decreasing ore grades from 2013.
Major projects implemented up until 2011 included increasing
rougher capacity by 25 per cent and increasing first cleaner
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
FIG 2 – Location of the Phu Kham copper-gold mine.
capacity by 16 per cent, and the installation of a Jameson Cell in
a cleaner scalper duty. In 2012 the operation was upgraded to
a nominal throughput of 16 Mt/a with installation of a second
13 MW ball mill, a further 33 per cent increase in rougher
capacity, 40 per cent increase in second cleaner capacity, and
33 per cent increase in third cleaner capacity.
In 2009 a project to achieve step-change in copper and
gold recovery from Phu Kham was initiated. A process
development study was completed in 2011, which showed
that it was technically feasible to generate a low-grade copper
and gold mineral concentrate by bulk sulfide flotation of
concentrator tailings suitable for leaching for recovery of
copper and gold into high-grade products. During the study,
opportunities for increasing copper and gold recovery in the
existing concentrator using standard processing methods
became apparent, and detailed mineralogical work and
metallurgical test work was undertaken to determine the
causes of copper and gold loss to tailings. The mineralogical
work revealed that up to 60 per cent of copper sulfide mineral
lost was in coarse non-sulfide gangue composites, and over
50 per cent of gold loss was in gold-pyrite composites.
The work presented an opportunity to recover these
composites by less selective rougher flotation, before upgrade
and additional recovery for both copper and gold into a
We are metallurgists, not magicians
23 per cent copper concentrate by regrinding of rougher
concentrate to 20 µm and additional cleaning flotation
capacity. Following extensive test work at bench, pilot, and
full scale which proved the concept and led to investment
approval, in 2013 the operation increased total recovery of
both copper and gold by approximately six per cent into final
concentrate through the Increased Recovery Project (IRP). A
second filter was installed during the project to dewater the
additional concentrate produced.
As mining extends deeper into the deposit, the operation
will experience an increased proportion of highly competent
ores which will have the potential to limit plant throughput,
through the semi-autogenous grinding (SAG) mill. Phu
Bia Mining commenced a throughput forecasting and
optimisation project in 2012 to evaluate how to maintain
design throughput over the LOM. Insufficient comminution
data in the mine block model created a lack of confidence in
the ability to predict mill throughput, particularly in the later
years of the mine life.
Geology and mineralogy
The Phu Kham geology is highly variable due to weathering,
alteration, faulting and folding. The deposit consists of
complex heterogeneous mineralogy horizons of copper-gold
203
D Bennett et al
stockwork and skarn mineralisation as shown in Figure 3.
Weathering and water table contact have created a soft
leached zone, overlying transition zones with supergene
chalcocite-dominant secondary copper mineralisation
and clay-rich gangue. The rock mass strength and degree
of weathering vary considerably across the deposit with
extremely competent (hard) rock found in the deeper levels.
Such variability causes a large range of plant throughput and
metallurgical performances.
Mineralisation is present in iron-rich skarns, silica-rich
stockwork and altered disseminated styles. Chalcopyrite
and bornite are the dominant primary copper minerals in
skarn, stockwork and disseminated mineralisation. Gangue
mineralogy is mainly quartz, mica and pyrite, with significant
kaolinite clay and talc-related magnesium silicate content
within the weathered zones.
A gold-enriched oxide zone on the Phu Kham orebody was
the resource for the heap leach gold mine which was built and
operated by Phu Bia Mining during the 2005 to 2010 period.
Below the oxide zone, there is a zone of supergene weathering,
with copper leached from the oxide zone re-precipitated
in contact with pyrite grains as particles and coatings of
chalcocite and covellite, with minor enargite and tennantite
copper arsenic sulfides. Significant copper enrichment in the
oxide and supergene zones is also present as oxide and native
copper species.
Skarns are present as replacement of carbonate minerals,
with disseminated grains of chalcopyrite and bornite in
banded to massive pyrite skarns and veinlets containing
pyrite, chalcopyrite and bornite in garnet, magnetite
and hematite-chlorite skarns. Pyrite skarns are common
throughout the mineralised system.
Stockwork mineralisation is present as fine fractures in
quartz veins. The fractures host pyrite, chalcopyrite and
bornite sulfide minerals. Minor chalcopyrite mineralisation
is also present in quartz-carbonate veins. Disseminated
mineralisation consists of scattered grains of bornite and
chalcopyrite in sericite altered host rock.
Gold occurs as small grains associated with pyrite and
copper sulfides throughout the mineralised system.
CIRCUIT DESCRIPTION
The original 12 Mt/a concentrator design and commissioning
in 2008 has been described in detail by Crnkovic et al (2009).
The description of the 2016 concentrator following the PKU
and IRP is provided in this section and the simplified flow
sheet is presented in Figure 4.
The crushing plant consists of a primary 55 in × 77 in
gyratory crusher, with single truck dump point above a pocket
designed to hold 200 t capacity equivalent to two 777D haul
trucks. Crusher discharge drops to a crushed ore bin of 200 t
capacity. The crushed ore bin is emptied by a variable speed
apron feeder onto a crushed ore transfer conveyor belt (CV001). The CV-001 conveyor transfers the ore to an 890 m long
overland conveyor CV-002, which moves ore to the coarse ore
FIG 3 – Phu Kham geological zones.
204
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
FIG 4 – 2016 Phu Kham 18 Mt/a concentrator simplified flow diagram.
stockpile with a live capacity of approximately 24 000 t. There
is additional dead capacity for storage of up to 300 000 t of ore.
Ore is reclaimed from the crushed ore stockpile by two
variable speed apron feeders onto a SAG mill feed conveyor.
SAG mill grinding media is added to the ore feed conveyor
via a spillage return hopper. Primary grinding is achieved in
a dual pinion 13 MW variable speed slip energy recovery/
hyper-synchronous drive 34 ft × 20 ft SAG mill in closed circuit
with scats return conveying including a high-lift conveyor to
overcome topography constraints.
SAG mill discharge is classified using an integral mill
trommel, with minus 12 mm product reporting to a 1.85 MW
cyclone feed pump. In the original circuit, cyclone feed was
classified in a single cluster of 18 × 650 mm diameter cyclones,
with cyclone underflow reporting to the No 1 dual pinion
13 MW drive 40 ft × 24 ft ball mill in closed circuit. In the
upgraded plant, a bleed stream of approximately 50 per cent
of the SAG mill feed mass is diverted via a transfer pump
to a 1.85 MW cyclone feed pump reporting to a cluster of
17 × 650 mm diameter cyclones, with cyclone underflow
reporting to the No 2 dual pinion 13 MW drive 40 ft × 24 ft ball
mill in closed circuit. Quicklime slurry is added to the SAG
mill and ball mills for flotation pH control to depress pyrite.
Cyclone overflow from the No 1 milling circuit reports
to a mixing box, where it joins a bleed of cyclone overflow
from the No 2 circuit to balance the volumetric feed split to
the No 1 rougher bank. The mixed cyclone overflow passes
through a multiple stage feed sampler before the No 1 bank of
ten 200 m3 tank cell roughers. The original plant had a single
200 m3 rougher feed conditioning tank before eight 200 m3
We are metallurgists, not magicians
rougher cells in the No 1 rougher bank, with the conditioning
tank converted to a ninth flotation cell in 2009. In early 2011, a
tenth 200 m3 rougher cell was installed and commissioned in
the No 1 rougher bank.
Cyclone overflow from the No 2 milling circuit reports to
a 70 m3 agitated tank with two discharge pumps, one which
reports to the mixing box before the No 1 rougher bank, and
another which reports to the No 2 rougher bank of five 200 m3
tank cells via a multiple stage feed sampler.
Dithiophosphate collector is used to recover copper sulfide
minerals in flotation while maintaining selectivity against
pyrite.
Rougher banks tailings pass through static dual fin pipe
samplers before reporting to final tailings mixing box where
it is combined with cleaner scavenger tailings. The mixed
tailings discharge to a metallurgical multiple stage sampler
and a final tailings sump. The final tailings sump discharges
slurry by gravity through two 750 mm diameter tailings lines,
which transport tailings approximately 1.5 km to a crossvalley subaqueous tailings storage facility (TSF).
Rougher concentrate from both rougher banks reports to
a common cyclone feed pump hopper before classification
in a cluster of 12 × 400 mm diameter cyclones, with cyclone
overflow reporting to a Jameson Cell feed hopper. Cyclone
underflow reports to two open circuit M10000 IsaMill™
regrind mills operating in parallel. IsaMill™ discharges
report to the Jameson Cell feed hopper. The 24 downcomer,
6500 mm diameter Jameson Cell operates in a cleaner feed
scalping duty, with concentrate passing through a pipe
sampler for process control, before reporting to the final
205
D Bennett et al
concentrate thickener feed sampler, and tailings reporting to
the conventional cleaning circuit.
The conventional cleaning circuit consists of three stages,
with the first stage in open circuit. The first stage of cleaning/
cleaner scavenging consists of two parallel banks of seven
70 m3 tank cells for a total of 14 cells. Cleaner scavenger tailings
pass through static dual fin samplers before reporting to the
final tailings mixing box. First cleaner and cleaner scavenger
concentrates report to the second cleaner, which consists of
seven 20 m3 cells. Second cleaner concentrate advances to the
third cleaner of four 20 m3 cells, while second cleaner tailings
return to the first cleaner feed pump hopper. Third cleaner
concentrate passes through a pipe sampler, before reporting
to the final concentrate thickener feed sampler. Third cleaner
tailings return to the second cleaners.
Final concentrate (combined Jameson Cell and third cleaner
concentrates) is sampled in a multiple stage metallurgical
sampler, before gravitating to a 15 m diameter high-rate
thickener. Thickener supernatant flows to a thickener overflow
process water tank. Thickener underflow at a nominal density
of 65–70 per cent solids is pumped to a mechanically agitated
filter feed tank of approximately 24 hours surge capacity, with
excess production bled to a second mechanically agitated filter
feed tank of approximately 8 hours capacity. The thickened
concentrate slurry in the 24 hour capacity tank is dewatered
using a 64-plate horizontal filter, with filter discharging into a
covered storage shed of nominal 8000 t capacity. The thickened
concentrate slurry in the eight hour capacity tank is dewatered
using a 40 plate horizontal filter, with the filter discharging into
a separate covered storage shed of nominal 5000 t capacity.
Concentrate is loaded into 25 t containers for transport by truck
to the Sriracha port in Thailand or ports in Vietnam.
Concentrator raw water is harvested from the Nam Mo River
before being pumped to a crusher process water tank and mill
header tank. The raw water is mainly used for cooling, pump
glands, flotation froth wash showers and fire water. Process
water is recovered from the TSF supernatant, and transferred
to a process water tank via two transfer stations.
process improvements and FLOW SHEET DEVELOPMENT
The Phu Kham 12 Mt/a concentrator was designed and built
to treat a high pyrite copper-gold skarn ore with significant
clay content, as described by Meka and Lane (2010). The plant
was commissioned in April 2008 and had ramped up to meet
nameplate capacity and operating time design by November
2008. A simplified flow sheet for the original 12 Mt/a
concentrator is shown in Figure 5.
A photograph of the 12 Mt/a concentrator in June 2008 is
shown in Figure 6.
2009–2010
Flotation cell conversions
In September 2009, the existing rougher conditioner and
cleaner conditioner tanks were retrofitted with flotation
mechanisms, thereby increasing the roughing capacity from
eight to nine 200 m3 cells, and increasing the cleaner capacity
from six to seven 70 m3 cells. The benefits arising from these
changes amounted to increased copper recovery in the rougher
flotation circuit by 3.5 per cent, and increased copper recovery
in the cleaner circuit by 2.5 per cent. The increased residence
time in each of the circuits resulted in higher recovery of slow
floating secondary copper minerals, and composite particles
FIG 5 – 2008 Phu Kham 12 Mt/a concentrator simplified flow diagram.
206
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
rougher residence time by 2.4 minutes at maximum design
throughput of 1750 t/h. In late January 2011, the tenth 200 m3
rougher tank cell was installed at the head of the rougher
circuit. At the time that the tenth rougher cell was installed,
the flotation feed rate was increased by about four per cent,
which effectively reduced the overall rougher residence time
increase from an expected 11 per cent to seven per cent. The
overall surveyed copper recovery improvement was between
0.4 per cent and 0.6 per cent, depending on throughput rate,
which met the project criteria.
FIG 6 – Southern view of the Phu Kham concentrator in June 2008.
particularly while transitional ore types were dominant at
this time. The improved recovery performance was validated
from using the database of daily rougher tail and cleaner
scavenger tail re-flotation tests.
FloatForceTM mechanisms
An investigation into using Outotec FloatForce™ rotor-stator
mechanisms in the rougher flotation circuit commenced
in 2009, starting with the first rougher cell 1B. The design
of the new mechanism was to deliver improved recovery
from increased mixing efficiency, by not allowing any air in
the central mixing area of the impellor thereby improving
mixing efficiency without affecting slurry pumping. The
installation of the FloatForce™ mechanism was simple, and
was commissioned without any issues. The conclusions,
drawn from extensive survey data around rougher cell 1B
(before and after installation), demonstrated an improvement
in rougher cell 1B copper recovery of 0.2 per cent. On the
basis of the survey data, the total predicted rougher copper
recovery increase for the nine rougher cells with FloatForce™
mechanisms was 0.3 per cent. The installation of the remaining
eight mechanisms was completed in June 2011.
2011
Tenth rougher cell installation
The ongoing rougher tail re-flotation tests continued to
highlight that there was a further copper recovery benefit
of approximately 0.6 per cent to be gained with the
addition of another 200 m3 tank cell, which would increase
Cleaner circuit debottlenecking
One of the major limitations in the original plant design had
been a lack of cleaning capacity, particularly with respect to
the second cleaner bank of four 20 m3 cells. The limitation
was that of carrying capacity, rather than residence time. The
cleaner circuit performance would generally deteriorate when
the cleaner feed copper metal units exceeded 9.1 t/h copper,
limiting copper metal production to a sustainable maximum
of approximately 8.5 t/h.
To further investigate the cleaning circuit capacity, a cleaner
circuit optimisation study was completed in February 2010.
From plant data, a mineral based floatability component
model was developed which allowed different cleaner circuit
configurations to be simulated. The option which gave the
optimum copper grade and recovery result was to install
additional cleaning capacity ahead of the existing cleaner
circuit, so effectively cleaner feed scalping. Different flotation
cell technologies were considered for this application, with the
Xstrata Jameson Cell meeting design criteria. The simulations
indicated that a 0.6 per cent improvement in cleaner recovery
could be achieved. The Jameson Cell was chosen because
of low installed cost, confidence in simulated performance
results, low performance risk, moderate installation risk and
low production continuity risk during installation.
The circuit simulations including the Jameson Cell
as a cleaner feed scalper indicated substantial recovery
improvement over the existing circuit at circuit feed rates
greater than 150 t/h, due to elimination of the carrying
capacity limitation as shown in Figure 7.
On this basis, it was decided to proceed with this design, the
Jameson Cell in a cleaner feed scalping simulation recovering
approximately 60 per cent of the copper present in the
FIG 7 – Effect of cleaner circuit feed tonnage on copper recovery for Jameson Cell cleaner feed scalping simulations.
We are metallurgists, not magicians
207
D Bennett et al
cleaner circuit feed across a cleaner circuit feed rate range of
100 t/h to 300 t/h. The Jameson Cell concentrate grade from
the simulation was 27 per cent copper, against a target of
25 per cent copper. The simulations showed that substantial
unloading of the remainder of the cleaner circuit would occur.
The result of this was that the third cleaner concentrate grade
was low at less than 22 per cent copper, however the net effect
was to produce an overall circuit final concentrate grade of
24 per cent copper.
Although there was a small cleaner recovery improvement
shown from the simulations performed using a cleaner
scalper, the real benefit was in maintaining cleaner recovery
when the cleaner feed rate is greater than 150 t/h.
The cleaner scalper cell was commissioned in March 2011.
Commissioning was carried out over a period of one week, and
no significant problems were encountered. The performance
evaluation of the Jameson Cell was remarkably consistent
with the expected performance from the equipment vendor,
the simulation data and from Phu Kham Metallurgical
Laboratory flotation tests simulating performance of the
Jameson Cell prior to commissioning. From surveys carried
out in February 2012, with the Jameson Cell online and
offline, the benefit of having the Jameson Cell in circuit was
determined to be 0.8 per cent increase in copper recovery.
In terms of overall cleaner circuit debottlenecking, the
objectives were achieved. The cleaner circuit with cleaner
feed scalping capacity is 10.1 t/h copper metal at 24 per cent
concentrate grade for a total 16 per cent copper metal
production increase.
2012
Phu Kham Upgrade project
The Phu Kham Upgrade (PKU) project commenced in March
2010 with a study to develop designs to ensure that copper
in concentrate production was maintained over 60 kt/a
after 2013 when plant copper feed grade was expected to
decrease. In order to maintain copper metal production, plant
nominal design throughput increased from 12 to 16 Mt/a
(1500 to 2000 t/h) and maximum instantaneous design
throughput increased to 2250 t/h. The plant upgrade concept
was not original, and had been studied in 2008 as part of a
copper production expansion project concept. Key aspects
of the upgrade designs for Phu Kham were the limitation
of available space for additional equipment, as the original
12 Mt/a plant design had not specifically made allowance for
any expandability.
The initial phase of the PKU included a plant debottlenecking
study, which consisted of analysis of the actual plant
performance and capacity data from 2008 to March 2010
against the original plant process design criteria. The purpose
of the bottleneck study was to determine aspects of the original
plant that either were, or would become bottlenecks with the
25 per cent increase in mill throughput. The key findings
from the plant bottleneck study are shown in Figure 8, which
shows that rougher copper recovery was 16 per cent below
design, cleaner copper recovery was seven per cent below
design, and mill throughput was three per cent below design
at 12 Mt/a. The mill throughput variance was a function of
rougher copper recovery and cleaning circuit capacity rather
than limitations in the grinding circuit.
The crushing and concentrate dewatering plant capacities
were also considered during the PKU bottleneck study. The
bottleneck study indicated that additional crushing capacity
would be required with capability for handling wet and
sticky ore which is a common feature of the transition zones
of the orebody. A mineral sizer in parallel to the existing
crushing plant, with product reporting directly to the coarse
ore stockpile was included in the PKU designs. Although
the concentrate thickener and filter performance had not
indicated that future concentrate production rate would
exceed capacity, limited data was available to confirm the
capacity against upgraded plant design criteria. Test work
was conducted to determine settling rates and filtration rates
for concentrate during the PKU to obtain the required data.
The basis of design for the grinding circuit upgrade has
been described by Hadaway and Bennett (2011). Two options
for increasing grinding circuit throughput after the SAG
mill to a nominal design of 16 Mt/a and primary grind at
80 per cent passing 106 µm or 75 µm were reviewed. The first
option was based on the original 2008 plant upgrade design
incorporating an additional 6.5 MW single pinion ball mill,
and the second for another 13 MW ball mill.
Data from JKTech grinding circuit modelling in 2009 was
extrapolated using the Phu Kham mine schedule to determine
throughput estimates at 106 µm and 75 µm for the two mill
FIG 8 – Phu Kham 12 Mt/a concentrator actual performance variances against design 2008–2010.
208
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
options. Pebble crushing was not considered in the evaluation.
The 6.5 MW mill was able to increase throughput at a 106 µm
primary grind to above the 16 Mt/a nominal design, however
was unable to meet the 16 Mt/a target at a significantly finer
primary grind. The 13 MW mill was able to achieve above
18 Mt/a for the 106 µm primary grind, and could achieve
above nominal design throughput at a 75 µm primary grind.
The effect of primary grind on flotation recovery was
reviewed based on feasibility study work from bench scale
batch tests in 2005. The study work indicated that the major
primary ore sources, in particular stockwork primary, were
relatively insensitive to primary grind size. Plant operations
mineralogy data from 2008 to 2011 monthly composites
indicated that minor sensitivity existed, with increases of over
five per cent in copper sulfide liberation with a primary grind
size decrease from 80 per cent passing 106 µm to 75 µm.
An economic analysis was conducted based on differences
in capital and operating costs for the two options at 16 Mt/a
throughput and 106 µm and 75 µm primary grind. The increase
in operating cost for finer grinding versus revenue benefits in
copper recovery showed that above $2.50/lb copper price, the
finer primary grind increased gross margin. Capital cost per
installed megawatt was 26 per cent less for the 13 MW mill
option, and the capital payback period for the 13 MW option
was significantly shorter.
A risk assessment was conducted for the 6.5 MW option
and the 13 MW option. The risk of the 13 MW option was
considerably lower than for the 6.5 MW option, mainly due
to the operating flexibility for periods of low-grade ore and
ore types with higher sensitivity of recovery to primary
grind. The 6.5 MW option was not able to take advantage
of economies of scale gained by increased throughput, or
the estimated one per cent increase in copper recovery at
the finer grinds, and would not be able to reach the nominal
16 Mt/a throughput at 106 µm primary grind after 2014. The
throughput at 106 µm and 75 µm primary grinds for the two
mill options is shown in Figure 9.
Based on the results of the risk assessment, the
recommendation for installation of an additional 13 MW ball
was accepted. Procurement of a second dual pinion 13 MW
drive 40 ft × 24 ft ball mill commenced in November 2010.
The dominant cause of the 16 per cent rougher copper
recovery shortfall shown in Figure 8 was a combination of
lower than design rougher residence time due to five per cent
lower rougher feed density, and a cleaner circuit capacity
constraint which limited the rougher mass recovery. The
dominance of transition ores with significant slow floating
secondary copper mineral content milled during the March
2009 to February 2010 period and the under-representation
of these ore types in the feasibility study test work provide
explanation for some of the copper recovery shortfall in
cleaning stages against design. The debottlenecking study
was developed for the flotation circuit to determine increased
capacity requirement at the PKU design 16 Mt/a throughput.
Rougher flotation feed density design for the 12 Mt/a plant
was 35 per cent solids. Actual operation rougher feed density
averaged 30 per cent solids due to the higher slurry viscosity
from kaolinite clay content not quantified during the feasibility
study. An extra 200 m3 rougher cell was required to achieve
the same residence time as at 35 per cent solids, which was
achieved by conversion of the rougher conditioning tank to a
cell in 2009. The reduced residence time from operating at the
lower rougher density at design tonnage throughput resulted
in a three per cent decrease in copper recovery, based upon
plant residence time–recovery data from July 2009 to February
2010. The PKU design therefore allowed for reduced rougher
feed pulp density, and a residence time calculation confirmed
that a 33 per cent increase in rougher capacity was required
for the 25 per cent increase in mill throughput at 16 Mt/a,
which would also provide an additional one per cent copper
recovery. A total of five 200 m3 rougher cells in addition to
the existing ten cells were included in the design, for a total
of 15 cells.
The cleaning circuit was not expected to require significant
expansion as a result of the 16 Mt/a upgrade, as the lower grade
mill feed would result in equivalent concentrate production
to the 12 Mt/a design throughput rate. The PKU design for
the cleaner flotation circuit also included the Jameson Cell
cleaner scalper although this had not been installed at this
time. However, cleaner circuit mass balance simulation data
including the Jameson Cell indicated that 40 per cent increase
in the existing second cleaner residence time and lip length
was required at 16 Mt/a. To gain this increase in second
cleaner capacity, the existing three 20 m3 third cleaner cells
FIG 9 – Throughput at 106 µm and 75 µm primary grind for additional 6.5 MW and 13 MW ball mill options.
We are metallurgists, not magicians
209
D Bennett et al
were combined with the four 20 m3 second cleaner cells to
create a second cleaner bank of seven 20 m3 cells, and four
new 20 m3 third cleaner cells were added for the PKU. The
simulation data for the cleaners also demonstrated that
upgrade of the first cleaner capacity was not warranted, as
the fine low-grade middlings recovered in the final cells were
not able to be upgraded to near final concentrate specification.
Prior to commencement of the detailed design phase of the
PKU study, maximum sustainable production rate (MSPR)
analyses were undertaken for the crushing and concentrate
dewatering plants to determine whether capacity expansion
was required for these areas of plant based on 16 Mt/a
production schedules. The MSPR was defined as the best
consecutive five days of performance, normalised using
plant-specific industry standards for annual availability to
allow for major scheduled maintenance. The MSPR for the
crushing plant also included seasonal variation due to the
tropical environment and the wet season impacts on crusher
productivity.
The primary conclusions from the performance review of
the Phu Kham crushing plant were that it had demonstrated
the target PKU production rate of 16 Mt/a over the June 2010
period, and approximately one-third of total crushing plant
downtime had been caused by events up and downstream
of the crushing plant while the plant was available to crush.
The low crusher utilisation of 64 per cent as a result of these
operating standby periods was equivalent to over 2.4 Mt/a
of additional crushing capacity at the target throughput of
2400 t/h and target utilisation of 75 per cent of total time. With
increases in haul fleet numbers for the PKU, improvements
in run-of-mine stockpile inventory, and a standby loader
available when there were delays in truck presentation to the
crusher, the MSPR demonstrated that increasing crushing
capacity was not required.
The design specifications for the 64 plate and frame filter
were for a filtration rate of 225 kg/m2/h, with an annual
design production rate of 311 000 t of concentrate. Actual filter
plant operating data was analysed to check filter performance
against the design capacity. MSPR for the filter was
determined to be 18 per cent above the life-of-mine maximum
concentrate production schedule, leading to deferral of capital
expenditure for the filtration plant. The main reasons for the
higher than design performance were; optimisation of filter
cycle settings following an improvement program including
operations, maintenance, and vendor support input and
change to filter cloth media type.
Following the review of the PKU design, engineering and
procurement services commenced for the 16 Mt/a PKU
project in January 2011, with commissioning commencing in
the third quarter of 2012. Ramp-up to nameplate capacity was
achieved within two months, and over 17 Mt/a annualised
throughput rate was achieved in the fourth quarter of 2012.
A simplified flow diagram for the PKU plant is shown in
Figure 10, with new equipment highlighted in mauve.
FIG 10 – 16 Mt/a PKU simplified flow diagram.
210
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
flotation option consistently achieved over 22 per cent copper
final concentrate grade, however ultimate copper recoveries
to final concentrate were lower than the bulk flotation option.
2013
Increased Recovery Project
The Phu Kham feasibility studies between 2004 and 2006
identified two options for flotation processing of Phu Kham
ore. The first option involved bulk flotation of the rougher
feed targeting a 25 per cent mass recovery into rougher
concentrate using non-selective amyl xanthate sulfide mineral
collector. The rougher concentrate was then reground to
80 per cent passing 38 µm and subjected to cleaner flotation
at pH 12 for pyrite depression. This process produced
high copper recovery results, however there was difficulty
achieving final concentrate grade of greater than 22 per cent
copper across all ore types, particularly transition chalcocitecovellite secondary copper mineral dominant ores. A rougher
feed photomicrograph is shown in Figure 11 with chalcocitecovellite intergrowth with pyrite and rimming of pyrite.
There was also indication of copper activation of pyrite from
soluble copper species in weathered and transition ores.
The second process option involved selective flotation in
roughing at pH 11–12 using a copper sulfide selective collector.
The rougher concentrates were again reground to 80 per cent
passing 38 µm and lime to pH 12 and sodium cyanide was
added to the cleaning stages to depress pyrite. The selective
The design of the original 12 Mt/a Phu Kham concentrator
was a compromise between the two process options, with
partially selective roughing being applied to minimise
pyrite gangue recovery into cleaner flotation feed. Sodium
cyanide addition to the cleaners was included in the original
design, however was never used with concentrate grade
over 22 per cent copper consistently achieved following
commissioning. This partially selective flotation process had
significant capital cost advantages over bulk flotation at a time
when the long-term copper price forecast was less than $2.00/
lb, due to the lower rougher concentrate regrind and cleaning
capacity required, and provided the best cost-benefit process
alternative while reducing risk of being unable to achieve
concentrate specifications using bulk rougher flotation.
Minimal work was performed during the feasibility
studies to test the sensitivity of final copper grade and
recovery on rougher concentrate regrind product particle
size. Grind size analysis was limited to two mineralogical
examinations which concluded that reasonable copper and
gold recoveries to rougher concentrate would result from
a primary grind of 80 per cent passing 106 µm, and that a
rougher concentrate regrind to less than 45 µm was required
to achieve an acceptable final concentrate grade. Mineral and
liberation analysis showed that associations between copper
sulfide minerals and pyrite did not indicate complex or fine
intergrowths that would adversely impact on the metallurgy.
The Phu Kham IRP commenced in 2009 as part of a
concept study to develop a process to increase copper and
gold recovery from Phu Kham ore. Since commencement
of operations in 2008, copper recovery had increased with
increasing proportion of chalcopyrite-dominant primary ores
replacing the chalcocite-covellite secondary copper mineral
dominant high clay and talc transition ores. The increasing
plant throughput and poor primary liberation with increasing
pyrite content caused copper and gold recovery to ‘flat-line’
as shown in Figure 12.
FIG 11 – Phu Kham transition ore rougher feed photomicrograph.
Cp – chalcopyrite, Ch/Cv – chalcocite/covellite,
Py – pyrite, Gn – non-sulfide gangue.
Bulk sulfide flotation bench tests in 2009 using isopropyl
xanthate collector on Phu Kham rougher tailings indicated
that up to ten per cent additional copper recovery and
70 per cent additional gold recovery could be achieved into a
scavenger concentrate of approximately 0.8 per cent copper.
FIG 12 – Phu Kham concentrator copper and gold recovery by year 2008 to 2011.
We are metallurgists, not magicians
211
D Bennett et al
The initial test program was designed to determine whether a
low-grade copper-gold concentrate suitable for downstream
hydrometallurgical processing to saleable products could be
recovered from the concentrator tailings.
The preliminary test program showed that a bulk sulfide
concentrate from plant tailings flotation could be upgraded to
over ten per cent copper concentrate grade, depending upon
copper sulfide mineral liberation, using a roughing, regrind
and two stage cleaning process similar to the Phu Kham
concentrator process. Figure 13 shows the copper graderecovery relationships with varying rougher concentrate
regrind power input of 10, 20 and 40 kWh/t.
The test results presented in Figure 13 clearly demonstrated
that finer regrinding of rougher concentrates from plant
tailings flotation would improve both copper grade and
recovery. Further flotation tests on plant final tailings samples
using roughing at pH 9 with amyl xanthate, followed by
regrinding of concentrate at 10 kWh/t power input, and two
stages of cleaning consistently produced a low-grade flotation
concentrate of approximately three per cent copper and 2 g/t
gold. Average copper recovery was 69.6 per cent and average
gold recovery was 54.1 per cent from 24 flotation tests as
shown in Table 1.
Since 2008, monthly plant composites had been submitted for
quantitative mineralogical analysis, and this data provided the
critical information used for recovery improvement process
development. The mineralogy data demonstrated that since
primary chalcopyrite ores had become the dominant source of
plant feed, the major cause of loss of copper in plant tailings
had changed from slow floating fine liberated copper minerals
(Crnkovic et al, 2009) to chalcopyrite locked in poor quality
coarse binary particles with non-sulfide gangue. The copper
loss in plant tailings from November 2011, representing a
typical month, is shown in Figure 14.
A digital photomicrograph of flotation tailings is shown in
Figure 15. The wide size range of the chalcopyrite particles in
non-sulfide gangue is evident.
Copper sulfide mineral grain size data for rougher tailings
is presented in Table 2. The data showed that under selective
rougher flotation conditions, recovery of coarse low quality
binary copper sulfide and gangue composite particles was
poor due to the fine copper sulfide grain size.
TABLE 1
Plant tailings flotation test results summary.
Copper
Gold
Stream
Grade
% Cu
Distribution
%
Grade
Au ppm
Distribution
%
Plant tailings (feed)
0.17
100.0
0.15
100.0
Concentrate
2.81
69.6
1.87
54.1
Final tailings
0.05
30.4
0.07
45.9
Gold recovery by flotation at Phu Kham had been poor
since commissioning, averaging approximately 40 per cent
to final copper concentrate product. Prediction of gold
recovery had also been demonstrated to be inaccurate during
plant operation, due to a lack of understanding of the key
mineralogical characteristics of gold occurrence and the
variability of gold occurrence across different Phu Kham
mineral assemblages. As part of the recovery improvement
study, the mineralogical reasons for gold loss into Phu
Kham tailings were examined to determine any potential
opportunities to increase gold recovery.
Diagnostic leach tests were conducted on Phu Kham plant
tailings samples. Results of the diagnostic leaching are
presented in Table 3.
The diagnostic leaching results in Table 3 demonstrated that
over 60 per cent of the gold in tailings was available for cyanide
leaching, either as free gold or partially liberated gold. This was
supported by the Albion gold leach test work on acid Albion
copper leach residues, which demonstrated that extraction of
gold to solution did not significantly increase with increasing
sulfide sulfur oxidation, as shown in Figure 16.
Laser ablation testing was conducted on the rougher tailings
to determine the proportion of gold locked with pyrite and
other sulfide mineral species. The laser ablation test results
were combined with the results from the diagnostic leaching
to provide a total gold association for the rougher tailings as
presented in Table 4.
Diagnostic leaching provided a measure of the unlocked
(cyanide soluble) and locked gold deportment, but the total
unlocked gold could not be split between fully liberated gold
FIG 13 – Phu Kham tailings flotation rougher concentrate regrind power grade – recovery relationships.
212
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
FIG 14 – Copper sulfide loss in Phu Kham flotation tailings by size and mineral association.
TABLE 3
Results of diagnostic gold leaching of Phu Kham flotation tailings.
Plant tailings
stream
% gold cyanide
soluble
% gold locked
in sulfide
minerals
% gold locked
in non-sulfide
gangue
Cleaner tailings
72
21
7
Rougher tailings
57
42
1
Final tailings
64
35
1
sample from the IRP laboratory test work was also conducted
to produce a gold-rich concentrate by gravity concentration
suitable for ADIS and photomicrograph analysis.
FIG 15 – Phu Kham flotation tailings photomicrograph.
TABLE 2
Rougher tailings copper sulfide grain size by size fraction.
Size fraction
Copper sulfide grain size (µm)
>106 µm
19.6
<106 >75 µm
13.0
<75 >53 µm
11.4
<53 >38 µm
9.1
<38 µm >C2
9.1
<C2 >C4
9.8
<C4 >C5
8.1
<C5
3.7
particles, and partially liberated (exposed in composites)
gold particles. Therefore, the diagnostic leach measure of
57 per cent cyanide soluble gold in Table 4 could not provide
definitive information for the cause of gold loss to tailings.
Final Phu Kham copper concentrate monthly plant
composites were analysed in 2012 by automated digital image
scanning (ADIS) at G&T Metallurgical Services to determine
the characteristics of gold and gold composite particles
recovered in flotation. Laboratory work on rougher tailings
We are metallurgists, not magicians
The ADIS work on the final concentrate showed an
average recovered gold particle size of 7.9 µm. The class and
mass distribution summary and area as a percentage of the
observed gold particles is presented in Table 5.
Binary particles of gold locked with pyrite in concentrate
were of particular interest. Only 8 per cent of the observed
particles were gold-pyrite binary composite particles,
however 71 per cent of the total mass of observed gold was
in these particles. The average surface area of the gold in the
gold-pyrite binary particles was 87 per cent of the total particle
surface area. An interpretation of this data indicated that for
a gold-pyrite particle to float into concentrate, the gold:pyrite
surface area ratio must be sufficiently large to overcome the
depression of the attached pyrite particle under high pH
flotation cleaning conditions. The liberated gold recovered
was typically fine, with an average particle size of 7 µm.
Photomicrographs of the gold showing some typical
particles in concentrate are presented in Figure 17.
The ADIS work found that the average gold particle size
was 19 µm in rougher tailings, approximately 13 times
larger than the average gold particle in final concentrate.
No binary particles were observed, only liberated gold and
gold-chalcopyrite-pyrite-gangue multiphase particles as
summarised by mass distribution in Table 6.
The liberated gold particles in tailings had a mean size of
35 µm, with the data indicating that liberated gold particles
above 20 µm in size are unlikely to be recovered to final
flotation concentrate, with a 13 µm particle the largest
213
D Bennett et al
FIG 16 – Copper and gold extraction in Albion leaching as a function of sulfide sulfur oxidation.
TABLE 4
Phu Kham rougher tailings gold deportment.
TABLE 5
Phu Kham final concentrate gold distribution data.
Gold mineral association % of total gold
Data
Gold locked in binary particles with:
Liberated
gold
Cp
Ch/Cv
Te
Py
Goe
Gn
38%
30%
8%
5%
8%
3%
3%
5%
Multiphase
Cyanide soluble
57
Pyrite locked
25
Association class
Other sulfide locked
17
Gold mass by association class
11%
16%
<1%
1%
71% <1%
1%
<1%
Non-sulfide gangue locked
1
Particle area % gold
100%
73%
52%
51%
87%
98%
9%
Total
100
Cp – chalcopyrite; Ch – chalcocite; Cv – covellite; Te – tennantite; Py – pyrite; Goe – goethite; Gn – gangue; MP – multiphase.
89%
FIG 17 – Phu Kham August 2011 final concentrate gold photomicrographs. NB: Au – gold, Py – pyrite.
TABLE 6
Flotation rougher tailing gold mode of occurrence, size and mass data.
Gold mode of occurrence
214
Average particle diameter (µm)
Average % of particle mass
Au
Gn
Au
Cp
Py
Cp
Py
Gn
Gold liberated
35
-
-
-
100
-
-
-
Gold adhesion multiphase
6
17
151
15
<1
<1
99
<1
Gold locked multiphase
8
28
104
27
3
2
80
15
Gold adhesion/locked multiphase
17
14
125
61
2
1
94
3
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
observed in the concentrate. Cyanide leaching would have
extracted 69 per cent of the gold particles observed.
The significant characteristic of the multiphase particles
observed in tailings was that greater than 95 per cent of the
total mass of the particles were gangue mass. The gold in the
multiphase particles had an average diameter of 9 µm, while
the pyrite gangue had an average diameter of 121 µm.
Photomicrographs of the gold showing some typical
particles in rougher tailings are presented in Figure 18.
Based on the concentrate and rougher tailings ADIS data, a
summary of the estimated recoveries of the main Phu Kham
ore gold association classes is presented in Table 7.
The downstream extraction of copper and gold from a
low-grade pyrite-rich concentrate produced by bulk tailings
flotation was developed to a prefeasibility level during
2010 and 2011 using Albion Process™ atmospheric leaching
technology. The Albion test work showed that copper
extraction of over 95 per cent using acid Albion leaching, and
gold extraction of over 85 per cent using a standard carbonin-leach process was achievable from a flotation concentrate
TABLE 7
Phu Kham indicative gold recovery by mineral association class.
Gold association class
Parameter
Liberated
Au
Au-copper
sulfide
Au-pyrite
Au-gangue
Gold recovery
70%
100%
10%
0%
of three per cent copper and 2 g/t gold, providing an overall
15 per cent copper recovery and 25 per cent gold recovery
increase for Phu Kham operations.
The process concept utilised bulk sulfide flotation with
potassium amyl xanthate collector (PAX), with the bulk
concentrate reground to 18 µm before cleaning flotation to
produce a three per cent copper concentrate. The concentrate
would then be leached in an Albion leach process with copper
cathode produced from the pregnant Albion leach solution
via a solvent extraction-electrowinning process. The Albion
leach residue would then be neutralised prior to carbon-inleach processing to produce gold dore.
The Albion test work program included a flotation pilot
plant at Phu Kham in order to produce a concentrate from the
plant tailings for pilot plant testing at the HRL/Core Resources
Albion pilot facility in Brisbane. The pilot facility consisted
of an M20 pilot IsaMill and a 1.5 t/h three stage flotation
plant. During the early stages of the pilot testing in October
2011, an updated capital estimate for the tailings retreatment
flotation and Albion hydrometallurgical process plants was
developed, with the updated estimate significantly higher
than previous estimates. The change in the project cost base,
coupled with the higher project risks for a tailings retreatment
plant incorporating hydrometallurgical processes, led to a
review of alternative recovery improvement strategies.
The prefeasibility tailings flotation study results had
indicated that there was potential for increasing copper
and gold recovery by mineral processing methods alone,
with reduced technical risk and capital intensity compared
FIG 18 – Phu Kham rougher tailings gold photomicrographs. NB: Au – gold; Cp – chalcopyrite; Py – pyrite; He – hematite; Ma – magnetite; Gn – gangue.
We are metallurgists, not magicians
215
D Bennett et al
to hydrometallurgical processing. The results from the
detailed mineralogical examination of Phu Kham tailings
during the Albion prefeasibility study were used to improve
understanding of the mechanisms of copper and gold loss
and develop a simplified process for improving recovery.
The first alternative was a ‘mainstream inert grinding’
process which classified the rougher tailings from the plant
to recover the plus 53 µm fraction containing the low quality
copper sulfide and gangue composites. This coarse fraction
of tailings would then be reground to 80 per cent passing
53 µm in order to further liberate the chalcopyrite prior to
an additional scavenging flotation stage, with scavenger
concentrate to report to regrinding with rougher concentrate.
Although the process was highly predictable (as it was based
on particle size rather than mineralogy), the capital and
power costs for the coarse fraction grinding were very high
due to the large tonnage of liberated and coarse silicate and
pyrite gangue in the rougher tailings required to be ground
for no value.
The second alternative was a ‘less selective’ sulfide rougher
flotation process followed by increased regrind capacity
in order to liberate the copper sulfide from coarse and low
quality composites, and increased cleaner circuit capacity
in order to treat the additional and lower grade rougher
concentrate mass. Absolute recovery improvement estimates
based on the test work for the ‘mainstream grinding’ and ‘less
selective flotation’ processes were similar at approximately
5–10 per cent for copper and gold. The advantage of the
‘less selective flotation’ process was that it was not a tailings
retreatment process and project value could not be eroded
with improvements in existing plant performance. Whilst
it was acknowledged that the mainstream inert grinding
process could also be performed on the plant feed, the ‘less
selective flotation’ process had significantly lower capital and
operating costs providing higher return on investment, so
was selected as the priority for further study.
The disadvantage of the ‘less selective flotation’ process
was that it was based on ore feed mineralogy, and therefore
generated a higher risk of mass balance errors during the
process design and uncertainty of achieving target copper
and gold recovery. The ‘less selective flotation’ process
development risks required that further test work be
conducted to confirm the recovery estimates and improve the
certainty of the economic case.
Project selection
The Phu Kham Albion flotation pilot plant work program
was changed to test the ‘less selective flotation’ process,
which successfully demonstrated that significantly increasing
the copper and gold recovery from roughing was achievable
utilising PAX with over 90 per cent copper recovery achieved
at 30 per cent mass recovery. The rougher concentrate grade
was significantly reduced due to the bulk sulfide flotation
regime employed during the test. The pilot scale tests proved
the concept that increasing mass recovery would lead to
increased copper recovery, whilst highlighting the pitfalls of
recovering too much liberated pyrite which would require
extremely aggressive pyrite depression conditions in cleaning
to produce a saleable grade concentrate.
In developing the design for a bulk sulfide rougher,
regrind and selective cleaning flotation process, three key
parameters were required to be identified; rougher mass
recovery to achieve maximum copper recovery across all ore
types, rougher concentrate regrind product size, and cleaner
capacity.
Due to the high variability in plant rougher copper and
gold recovery between the low and high pyrite content
ores, the rougher mass recovery likewise varied depending
on feed mineralogy. In order to develop the design rougher
mass recovery target, a review of the flotation recovery data
from daily rougher flotation testing of plant rougher feed
over a three-month period in 2011 was used to allow for the
variability in mineralogy of plant feed. Extensive rougher
flotation rate test data that had been used to develop early
Phu Kham recovery models were statistically analysed in
order to determine the mass recovery and copper recovery
curves for differing ore types, from weathered and altered
high pyrite (high S/Cu) material to chalcopyrite dominant
and low pyrite (low S/Cu) primary material.
The statistical analysis showed that rougher copper recovery
could be increased to 85 per cent by targeting a maximum
20–25 per cent mass recovery from rougher flotation across all
ore types, as presented in Figure 19.
FIG 19 – Rougher copper recovery and mass recovery for major ore types showing original mass recovery design and IRP mass recovery design.
216
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
The results show that the original design for rougher mass
recovery of 11 per cent of rougher feed was suitable for the
chalcopyrite-dominant primary ore types, but gave lower
copper recovery from the high pyrite and non-sulfide copper
species in chalcocite-rich transition ores. The results indicated
potential for an additional six per cent copper recovery
in roughing by increasing mass recovery to a maximum
25 per cent of rougher feed for all ore types.
observed as shown in Figure 13. Samples of plant feed were
tested at bench scale using roughing, rougher concentrate
regrind at increasing power input, and three stages of cleaning
to produce the copper and gold grade-recovery response
curves in Figure 21. The increase in rougher concentrate
regrind power input shows the benefit of the increased
liberation for both copper and gold grade and recovery into
flotation concentrates.
The results for rougher flotation gold recovery from the same
test work are shown in Figure 20. The gold recovery increase
in roughing with the higher mass recovery is more consistent
across primary, high pyrite and transition ore types than for
copper, which reflects the association of gold with copper
sulfide minerals and pyrite. The mineralogical analysis of
gold deportment and bench scale test work results concluded
that with the design rougher mass recovery increase and finer
regrind, gold recovery to final concentrate could be expected
to rise by at least six per cent.
With the variable mineralogy of Phu Kham ore, the
monthly composite mineralogy and laboratory scale test
work was used to determine the rougher concentrate regrind
size required to achieve maximum copper sulfide liberation
in cleaner feed, with a target of 80 per cent copper sulfide
liberation considered to be required to maximise recovery
and maintain final concentrate specification. Scan data on
two samples representing different Phu Kham ore types in
Figure 22 showed that maximum copper sulfide liberation was
typically achieved at 20 µm particle size, although 80 per cent
liberation was not necessarily achieved for all ore types.
Rougher concentrate regrind size optimisation work
commenced in 2011 to support the IRP process design. The
benefits for copper recovery of finer grinding of concentrates
from bulk flotation of Phu Kham tailings had previously been
Power input to achieve 20 µm regrind product size was
calculated from laboratory regrind signature plot data,
and daily surveys of the existing M10000 IsaMill to be 18–
FIG 20 – Rougher gold recovery and mass recovery for major ore types showing original mass recovery design and IRP mass recovery design.
FIG 21 – Copper and gold flotation grade-recovery responses as a function of rougher concentrate regrind power input.
We are metallurgists, not magicians
217
D Bennett et al
FIG 22 – Copper sulfide mineral liberation with particle size.
25 kWh/t, with an additional 3 MW M10000 IsaMill included
in the design to provide a total of 6 MW power input at the
maximum rougher concentrate mass recovery of 25 per cent.
First cleaner design
The increase in rougher concentrate mass recovery to
25 per cent of rougher feed, and the reduction in cleaner
feed particle size to 20 µm required a corresponding increase
in first cleaner flotation capacity. The 12 Mt/a plant had
25 minutes total first cleaner residence time for a 38 µm
cleaner feed size, and test work was designed to determine
whether this residence time needed to increase at the 20 µm
cleaner feed size due to potentially reduced kinetics of
the finer particles. Bench scale flotation cleaning tests and
cleaner circuit model simulations indicated that flotation
kinetics remained similar for copper sulfide minerals due
to improved liberation, and that no increase in residence
time was required for first cleaning until regrind product
size was below 10 µm. Figure 23 shows the plant cleaner
circuit chalcopyrite recovery by particle class and size.
The recovery of liberated chalcopyrite remained high with
decreasing particle size, which confirmed that the increased
chalcopyrite liberation from gangue at a 20 µm regrind size
would improve copper recovery.
The design included 100 per cent increase in first cleaner
capacity to allow for the increased rougher mass recovery, and
lower slurry density to provide improved dilution cleaning to
minimise fine gangue particle entrainment.
To validate the extensive test work and mineralogy data
results, four full-scale process plant ‘less selective flotation’
trials were conducted at Phu Kham between January 2012 and
April 2012. The trial method used was to decrease SAG mill
throughput by 50 per cent to avoid overloading the cleaning
circuit, reduce rougher cell residence time to the equivalent
post-PKU 16 Mt/a time of 30 minutes by reducing levels
and air in three of the ten cells, and using increased collector
and frother addition to increase rougher mass recovery to
25 per cent. The IsaMill power draw was increased to 2.8 MW
to target a 20 µm regrind product size. Immediately prior to
the trial periods, baseline plant surveys were undertaken to
obtain comparison data for the same ore type. The key results
of the four plant trials are summarised in Table 8.
The first plant trial period was over four hours to ensure
critical test criteria could be achieved with rougher mass
recovery target of 25 per cent. Although this was a preliminary
trial, survey results for the IRP process were positive compared
to baseline survey results, with eight per cent overall copper
recovery and 19 per cent overall gold recovery achieved
FIG 23 – Phu Kham cleaner circuit chalcopyrite recovery by particle class and size.
218
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
TABLE 8
Phu Kham ‘less selective flotation’ plant trial results.
Preliminary trials
Trial 1 – 15 January 2012
Trial 2 – 27 January 2012
Baseline
IRP
Difference
Baseline
IRP
Difference
Rougher recovery (%)
83
92
+9
76
88
+12
Overall copper recovery (%)
71
79
+8
68
79
+11
Overall gold recovery (%)
48
66
+19
45
56
+11
Concentrate grade (%)
22.1
24.1
+2.0
24.2
24.8
+0.6
Variability trials
Trial 3 – 21 March 2012
Baseline
IRP
Trial 4 – 24 April 2012
Difference
Baseline
IRP
Difference
Rougher recovery (%)
87
93
+6
81
87
+6
Overall copper recovery (%)
81
87
+6
75
80
+5
Overall gold recovery (%)
40
53
+13
36
56
+20
Concentrate grade (%)
25.7
23.8
-2.0
25.5
21.7
-3.8
into a two per cent higher copper grade final concentrate.
The second plant trial was conducted over a full 12-hour
shift period, with improvement in both overall copper and
gold recovery of 11 per cent at a 0.6 per cent copper grade
improvement in final concentrate.
March 2012, following PanAust Ltd board approval for the
IRP, the M10000 IsaMill was ordered as a turnkey package
including all the peripheral equipment to the mill along with
instrumentation. The package also included all steelwork but
specifically excluded mill foundations.
Two short four-hour variability trails were conducted in
March and April 2012 on primary ore with a good flotation
response and high pyrite ore with a poor flotation response.
Copper recovery improvement was consistent at over
five per cent for both ores, and gold recovery improvement
was 13 per cent and 20 per cent into final concentrate. The
short duration of the trials did not allow time for optimisation
of cleaner circuit performance.
The project schedule was developed to include the
completion of a front-end engineering phase (FEED) level
design and cost estimate by May 2012 using an external
engineering consultant. The FEED phase fixed the design
criteria, mass and water balances and tagged equipment lists
which were then be used in the detailed engineering and
procurement (EP) phase.
The economic evaluation at 6 per cent increase in copper
and gold recovery after the initial plant trials provided
a compelling investment case, resulting in PanAust Ltd
board approval in February 2012 for the development of the
re-named IRP at Phu Kham to allow early procurement of
long-lead equipment. The plant trial results also provided
significant confidence in the IRP basis of design.
Compression of the IRP schedule
During the early process design and project scoping, the project
case for the IRP was determined to be compelling enough for
the project development to be fast tracked with commitment
for long-lead capital items approved before completion of
the feasibility study. In addition to achieving the increased
plant recovery earlier, reducing the project schedule had the
added benefit that the PanAust construction team would be
able to commence work on the IRP immediately following
completion of the PKU in August 2012.
The longest lead item for the project was the M10000 IsaMill
with a delivery period of approximately 11 months. In early
We are metallurgists, not magicians
Chalcopyrite Recovery (%)
The trials were not used to confirm design mass recoveries
or regrind size. Higher final copper concentrate grades
were achieved despite the higher rougher mass recovery,
although the minor difference against baseline values was
not considered significant during the short trial. The ability
to conduct a plant trial and successfully demonstrate a
consistent metallurgical benefit at laboratory, pilot, and full
plant scale prior to implementation underpinned the high
degree of confidence in the ‘less selective flotation’ concept.
The detailed mineralogy-based understanding of the causes
of copper and gold loss to tailings developed over many years
was fundamental in gaining this confidence.
100.0
During the EP phase, the engineering consultant was
responsible
95.0 for engineering design and equipment selection up
to the recommendation for purchase decision, with approval
90.0
and procurement
performed by PanAust staff in Brisbane and
Laos. To avoid issues identified during the PKU with delays
85.0
in receipt of vendor data, the engineering consultant was
responsible
for liaison with vendors once the purchase order
80.0
was raised to ensure the engineering consultant maintained
75.0 control.
schedule
The basic project schedule for the IRP is presented in
70.0
Figure 24.
65.0
PROJECT DESIGN
60.0
Feasibility
55.0
In order to develop a business case for the project, initial
50.0 design was developed by PanAust during the
process
1 stage along with factored capital and operating
10
prefeasibility
Particle
cost estimates. The cost estimate was considered toSize
be (µm)
at a
higher level of accuracy than a normal prefeasibility stage due
Liberated Chalcopyrite
Chalcopy
to the recent completion of the engineering for the PKU and
Chalcopyrite‐Gangue Binary
Multipha
construction of the Ban Houayxai CIL plant by PanAust Ltd.
Fig 23 – Phu Kham cleaner circuit chalcopyrite recovery by particle class and size
2012
Project Stage
Q1
Q2
2013
Q3
Q4
Q1
Q2
Q3
Q4
Feasibility Study/FEED
Long Lead Procurement
Engineering and Procurement
Construction
Commissioning and Ramp-up
Fig 24 – Increased Recovery Project schedule
FIG 24 – Increased Recovery Project schedule.
219
23
D Bennett et al
In addition to process design criteria, the equipment lists
and brownfield plant layouts and integration plans were
developed to a detailed level by PanAust so that the FEED
stage was mainly focused on engineering detail and minor
equipment sizing rather than process design. Plant survey
work was performed during this stage to improve the
mass balance which was used to confirm pump and piping
specifications. The initial flow diagram showing the additional
equipment planned to be installed as part of the IRP in red is
shown in Figure 25.
Front-end engineering
GR Engineering Services (GRES) was selected as the FEED
engineer for the IRP. During the FEED phase a number of
details in the plant design were resolved which resulted in
minor changes to the process flow sheet. The most significant
change was the change to a second parallel bank instead of
additional series cells to the first cleaners. This was changed
mainly due to limited dart capacity in the first cleaner cells
as well as the cost associated with upgrading the first cleaner
feed pumps. The change resulted in a lower project capital cost
as capital was moved from debottlenecking the existing first
cleaner cells to installing a parallel bank of cells. The process
flow sheet was also simplified due to the change. The change
also reduced the associated tie-in time for construction as the
new equipment could be added in discrete modules, and did
not impact on the existing plant which was a key learning
from the PKU for brownfields plant expansions.
In parallel with the FEED phase, additional metallurgical
test work was performed in order to assess the effect of the
expected finer concentrate size on the settling rate of the
existing concentrate thickener. The settling tests performed
on concentrate produced from the variability trials showed
that the existing 15 m diameter concentrate thickener would
be able to handle the concentrate produced from the IRP,
however settling efficiency could be improved over current
operation by reducing the amount of entrained air in the feed.
Site investigation determined that air was being entrained
through the trash screen collection box, so the decision was
made to add a de-aeration box prior to the thickener to
improve thickener performance of the thickener.
Challenges in brownfields plant upgrades were highlighted
during the FEED phase with the marriage of new equipment
with existing equipment which cannot be relocated. Specific
design principles were included to minimise capital and
operating cost based on the learnings from the original plant
design and PKU. The design was developed with a focus on
the use of gravity flow wherever possible and gravity lines
were used in most tie-in locations despite the additional
engineering design required. This design philosophy
increased the complexity of the design effort, however
increased operability was achieved.
The existing plant concentrate production capacity was
analysed, with the existing filter able to produce concentrate
below the transport moisture limit at a rate of approximately
330 000 t/a. This was determined to be marginal at the increased
concentrate production rates as shown in Figure 26. This was
a key risk for the project as the filtration rate was expected
to decrease due to the finer concentrate size distribution. To
eliminate this risk a second filter was included in the design.
Although the second filter had fewer plates than the original
filter, the same plate size was maintained to allow common
spares with the existing filter.
FIG 25 – Initial process flow diagram for the Increased Recovery Project.
220
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
FIG 26 – Filter performance against increased recovery project concentrate production targets.
A design review was conducted during the FEED to look
at installing a second filter in the existing shed without
disruption of the existing filter operations. The review
determined that disruption would be significant, so an
additional smaller storage shed was included in the design to
allow for additional storage capacity during the wet season.
The design of the new storage shed and filtration area was
leveraged from the existing engineering designs to reduce
costs for installation. Some minor modifications were made
during the FEED phase to shorten the length of the shed in
order to reduce storage capacity in line with the additional
concentrate production.
An additional item of risk which was identified during
the FEED was the second cleaner flotation cell tailings darts
capacity not having sufficient margin above maximum
flows predicted for the IRP. Due to the lack of mitigation
measures available, contingency was allowed for in the
estimate to install inter-cell bypasses should the flow rate
significantly exceed darts capacity. It was anticipated that
the second cleaner would become the circuit bottleneck in the
IRP circuit, and if liberation was not achieved in the regrind
mills, a recirculating load of middlings particles would buildup between the second cleaner and third cleaner banks.
This would result in recovery losses or below specification
concentrate grade.
The FEED phase was completed on schedule with the capital
cost reduced by $10 M due to additional savings identified
during the design, and a reduction in the project contingency
due to better project definition. The reduction in capital
was achieved despite an increase in the scope of the project
in order to minimise risk. As the project was brownfields
integration, additional capital was allowed in the estimate for
resolution of legacy pumping and spillage issues which had
affected the Phu Kham concentrator since the original design
and the PKU. Following completion of the FEED phase the
final equipment list was confirmed along with the changes
to the process flow diagram (shown in red) in Figure 27 to
optimise both cost and design:
•• twelve 400 mm diameter regrind cyclone pack
•• M10000 IsaMill
•• seven Outotec TC70 tank cells
•• thickener de-aeration box
•• concentrate filter feed tank
We are metallurgists, not magicians
•• 40 plate Ishigaki filter
•• concentrate storage shed.
Increased Recovery Project COMMISSIONING AND RAMP-UP
Plant tie-ins were performed progressively as equipment was
delivered to site and in line with the project schedules which
were integrated with the planned shutdown schedules. The
majority of construction and tie-in activities were completed
by March 2013.
IsaMill commissioning
Commissioning of the new IsaMill concluded with a 72 hour
acceptance test which commenced on the 26 March 2013 with
both the new and old mills run in parallel for the first time.
Operational conditions during the period were identical for
each of the mills. During the period, the new IsaMill was run
over a range of power draws, with the ultimate design power
of 2500 kW being achieved. During the test, no external signs
of any problems were detected, however upon completion of
the test period, the mill was inspected and damage to the shell
liner (Figure 28), feed end liner and three discs was identified.
The root cause of the damage to the mill liners was
overheating of the mill slurry due to the high recycle of
IsaMill discharge to the feed. The high recycle was caused
by insufficient flow from the regrind cyclones which were
exacerbated by a biased flow from the underflow splitter box
to the existing IsaMill. The new regrind cyclones had also
been modified by operations due to the need to run them prior
to IRP commissioning and did not have the design spigots
installed prior to start-up. The high temperature in the IsaMill
would have normally been detected by the process control
system however the resistance temperature device (RTDs),
which were installed to protect the mill in such instances of
low feed, were found to be providing an inaccurate output.
The RTDs were replaced along with the shell liner, and the
discs and feed end liner were reused.
The mill acceptance test was re-run with additional wear
identified in the regrind mill due to media compression. The
media compression had been caused by a high slurry feed
density which was attributed to an incorrect and high solids
specific gravity input into the nuclear density gauge.
In addition, the output value of the density gauge had been
mistakenly assumed to be a calculation of per cent solids.
221
D Bennett et al
FIG 27 – Simplified Phu Kham process flow diagram including Increased Recovery Project expansion.
extensive. Existing equipment, particularly slurry pumps,
were utilised within the new circuit.
FIG 28 – Damage to the shell liner during the first 72 hour acceptance test.
This meant that the density in the mill was approximately
eight per cent higher than the actual density target, with
the subsequent high viscosity in the mill causing the media
compression. In order to ensure that this didn’t recur, the
target pulp density was decreased to 45 per cent solids
and the viscosity was monitored at hourly intervals with a
Marsh Funnel.
Once the control measures were put into place the decision
was made on the 18 April to run the mill for a 24-hour test.
At the completion of this test, the mill was inspected and no
further wear detected on the liners. The final acceptance test
was then performed for 48 hours at power draws of up to
2600 kW. On the 21 April, the mill was inspected for the final
time. No wear had occurred over the final trial period and
the mill was handed over by Xstrata Technology personnel to
PanAust operations.
Plant ramp-up
Even with the modular design of the IRP which aided
constructability, the integration with the existing circuit was
222
During the ramp-up phase a number of issues with slurry
pumping were encountered. These were mainly confined
to original pumps which had been upgraded or had duty
changed as part of the IRP or earlier improvement projects.
The IRP construction and precommissioning phases had
not generally considered the condition or status of existing
equipment and as it had been previously operating
satisfactorily, and no program of inspection was initiated
prior to IRP commissioning. This resulted in delays in
reaching design capacity.
As an example, the rougher concentrate pump system
had the common discharge pipeline upgraded to match
the higher IRP design flow rates, with the existing duty
and standby pumps and motors assessed as suitable for the
higher design duty. During the ramp-up phase the pumps
were not able to reach design flow rate which resulted in both
the duty and standby pumps being run to meet design flow
rate. Investigation of the pump variable speed drive systems
discovered that the original overload and frequency limit
settings had been retained for the original motors, although
the motors had been upgraded prior to the IRP. These settings
were limiting the pump speed and power draw to a level well
below maximum operating capability. Once the limit was
removed the pumps were found to be easily able to meet
design flow rate.
Ramp-up was also hindered by power supply continuity to
site, as there was often insufficient power able to be supplied
from the hydroelectric schemes which provide all power to
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
the Phu Bia Mining operations. This was due to low water
levels in the reservoir at the end of the dry season coinciding
with the IRP ramp-up and extended planned maintenance
periods on generating units. Operation response to power
limitation was to load-shed power across the plant with
power draw being lowered in the ball mills and IsaMills to
maintain throughput. This increased primary grind size and
regrind size lowering copper mineral liberation and reducing
plant copper recovery. The problem continued until June with
the onset of the wet season replenishing reservoir levels.
INcreased Recovery Project PERFORMANCE
Rougher recovery
Despite the described issues during ramp-up, rougher mass
recovery was quickly increased to 20 per cent of feed mass
following completion of commissioning.
At an early stage of ramp-up the operating strategy was to
effectively ‘fill up’ the cleaner circuit by recovering as much
mass as possible from the rougher banks. Cleaner circuit
performance and overall copper recovery were generally
compromised by this strategy because excessive gangue was
recovered in roughing requiring rejection in cleaning, and
the regrind product size to achieve liberation targets was not
being met. In addition, the feed ore quality was generally
challenging during ramp-up due to the requirement to mine
ore from the upper areas of the pit which had very high pyrite
and clay content, and a high proportion of easily oxidised
secondary copper sulfide species.
During this period, operational strategies were redefined
and training undertaken with site personnel to ensure that
copper recovery was maximised from the rougher banks and
the cleaning circuit was not overloaded with liberated gangue.
Back-to-basics improvements were undertaken on operating
parameters; in particular work was completed with Outotec
on optimising the rougher cell level and air addition profiles
which were then set to avoid excessive water and gangue
recovery and improve low quality coarse composite recovery.
Process control improvements were also implemented
during the ramp-up period, including commissioning of
automated rougher mass recovery control through use of
froth velocity measurements and level and air controls.
With the improvements in control, equipment performance
and operating knowledge for the new circuit, the rampup of the IRP was mainly completed by June 2013. Copper
rougher recovery increased from an average of 78 per cent to
83 per cent within the first month of operation, in line with
expectations. The increase in gold rougher recovery was very
significant from an average of 55 per cent to 78 per cent as
shown in Figure 29. Gold recovery at Phu Kham had always
been highly variable, which led to a conservative estimate
of gold recovery increase used for the IRP investment case,
however the observed increase in gold recovery exceeded
the results achieved during the plant trials and even the most
optimistic estimates. The reasons for the improvement were
likely the improved rougher recovery of lower quality copper
sulfide and pyrite composites containing gold and improved
fine gold recovery.
Once operational control was improved and selectivity
against liberated pyrite and non-sulfide gangue was improved
in the rougher circuit, the increased copper recovery was
confirmed though a consistent reduction in the quality of the
copper sulfide-gangue binary particles in the rougher tailings.
Following the implementation of the IRP, the copper sulfide
content of these particles decreased from approximately
15 per cent to seven per cent copper sulfide indicating that
the IRP target low quality copper sulfide-gangue composite
particles were being recovered (Figure 30).
A photomicrograph of typical rougher tailings is presented
in Figure 31, indicating the lean quality of copper sulfidegangue composites following the IRP.
Overall recovery
The ramp-up in overall recovery was achieved in June 2013
with 76.4 per cent average weekly recovery being achieved.
This was 3.8 per cent above the baseline PKU recovery of
72.6 per cent. In addition during this time the variability of
the overall recovery decreased with the standard deviation
of weekly recovery reducing from six per cent to 4.6 per cent
of the mean. As presented in Figure 32, the number of weeks
with recovery below 70 per cent has decreased over time as
the operational strategies implemented during the IRP have
been further reinforced.
As for copper, initial increases in the rougher gold recovery
did not correspond to a significant increase in overall gold
FIG 29 – Rougher mass, gold and copper recovery following the Phu Kham Upgrade and Increased Recovery Projects.
We are metallurgists, not magicians
223
D Bennett et al
FIG 30 – Quality of the copper sulfide gangue binary composites in the rougher tailings following the Phu Kham Upgrade and Increased Recovery Projects.
recovery. Gold losses in the cleaner tailings were investigated
during this period in order to determine the mechanisms of
gold loss from the cleaner circuit.
FIG 31 – Photomicrograph of the typical rougher tailings
following Phu Kham Upgrade and Increased Recovery Projects.
Cp – chalcopyrite; Py – pyrite; Gn – non-sulfide gangue.
Analysis concluded that although 25 per cent of the identified
gold containing particles lost from the cleaner tailings were
liberated gold, the majority of these particles were less
than 10 µm in size. Almost 50 per cent of the particles were
liberated or high quality composites with pyrite or copper
sulfide minerals. The other 50 per cent of gold particles lost
were in extremely low quality binary composites with pyrite
(Figure 33). These composites were required to be rejected to
maintain acceptable final copper concentrate grade. As gold
particles even in binary composites were fine (Figure 34),
gold recovery to the final concentrate relies on liberation
from pyrite due to the high lime pH environment for pyrite
depression, which also reduces fine gold flotation kinetics.
FIG 32 – Overall copper recovery mean and standard deviation following implementation of the Increased Recovery Projects.
224
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
FIG 33 – Quality of the gold particles lost from the cleaner tailings.
FIG 34 – Examples of poor quality composites and fine liberated gold losses in the cleaner tailings.
With improvement in plant performance, overall gold
recovery increased, with weekly gold recovery averaging
over 50 per cent by September 2013 (Figure 35). No decrease in
the variability of the gold recovery was achieved because the
gold recovery is mainly affected by the variable mineralogy,
liberation and gold feed grades to the plant.
High gold recovery was maintained throughout the year
following the IRP even though the plant experienced a
significant change in the gold head grade from 0.3 g/t to
0.2 g/t from September 2013 to September 2014 (Figure 36).
Once head grades returned to normal levels in late 2013, overall
gold recoveries of 60 per cent were being achieved which was
almost a 20 per cent increase over the pre-IRP recoveries.
Recovery against increased recovery project investment case
The initial analysis of the plant data shows that the average
increase in copper recovery was less than the six per cent
which was originally predicted for the project and the increase
in gold recovery was above the original six per cent target.
This was despite other design targets (eg mass pull rate and
rougher recovery) within the plant being achieved. The source
of the copper recovery difference was due to a number of other
operational and ore quality factors which were impacting on
production during the period as discussed below.
It can be seen in Figure 37 that during the period of the
ramp-up, the quantity of pyrite in the ore as measured by the
sulfur to copper content fraction (known as the S/Cu ratio)
We are metallurgists, not magicians
increased from over 16 during the PKU ramp-up period to 21
during the IRP ramp-up. In addition, head grades declined
over the period which allowed the throughput to be increased
in order to maintain the copper production rates. Pivotal in
this increase in mill throughput was the implementation of
new plant control strategies (Baas, Bennett and Walker, 2014)
along with the additional rougher concentrate, regrind and
cleaner capacities provided by the IRP.
The S/Cu ratio has been used at Phu Kham in recovery
modelling as a proxy for pyrite in the ore, which is the
dominant driver of metallurgical performance. Increased
S/Cu ratios result in decreased copper recovery, due to the
requirement to be highly selective against pyrite. On a S/
Cu ratio basis, the increased copper recovery achieved by
the IRP varies from seven per cent at a S/Cu ratio of over
30 to four per cent at a S/Cu ratio of less than 20 (Figure 38).
The increase achieved by the project was determined to have
reached the six per cent goal when the ore quality schedule
was compared on a like for like basis.
FUTURE OPPORTUNITIES
Since the IRP has been implemented, a number of further
opportunities for recovery improvement have been pursued.
Major additional improvement in reducing the variation
in overall recovery has since been achieved through
implementation of a mass recovery control system which
utilises froth cameras on the rougher banks and first cleaner
225
D Bennett et al
FIG 35 – Overall gold recovery following commissioning of the Increased Recovery Projects.
FIG 36 – CUSUM chart for gold showing increase gold recovery despite declining head grades.
FIG 37 – CUSUM chart for copper recovery showing changes to operation and ore quality indicators.
226
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
FIG 38 – Copper recovery against S/Cu ratio in plant feed.
banks. This system has allowed maintenance of mass and
copper recovery based on concentrate grade targets and metal
input. The integration of second cleaner and third cleaner
control to maximise copper recovery and trim concentrate
grade with the Jameson Cell concentrate grade to achieve a final
concentrate grade target has been implemented, with a particle
size analyser used to control IsaMill regrind power draw.
Finally, work has progressed to improve cleaner circuit pulp
chemistry following regrind. This body of work was identified
by site metallurgists during Jameson Cell optimisation test
work where it was found that the copper sulfides in regrind
cyclone overflow exhibit significantly faster flotation kinetics
than the IsaMill regrind discharge. This has been attributed to
oxygen depletion via reaction of pyrite in the closed milling
system. Automated analysis of the pulp chemistry has been
implemented and a trial using peroxide (dosed to the regrind
discharge to modify the pulp redox potential in the cleaner
feed) is currently being advanced.
CONCLUSIONS
The installed 2008 Phu Kham 12 Mt/a concentrator was
a compromise between a high recovery but high capital
intensity design, and a lower recovery but technically low risk
and low capital intensity design suitable for the prevailing
low copper price market. The chosen selective rougher
flotation design was driven by the complex and highly
variable mineralogy particularly in the transition ore zones,
and high pyrite content. Over 90 per cent of pyrite is required
to be rejected in order to produce a final concentrate of over
23 per cent copper.
The optimisation of the Phu Kham flow sheet was been driven
by low copper and gold recoveries and the requirement to
maximise copper production by increasing plant throughput.
Flotation circuit improvements and debottlenecking
projects including additional roughing capacity and cleaner
feed scalping had increased copper metal production by
16 per cent since 2009, while maintaining copper recovery at
the original maximum design throughput of 14 Mt/a. In 2012
the operation was upgraded and debottlenecked during the
We are metallurgists, not magicians
Phu Kham Upgrade Project to process a nominal throughput
of 16 Mt/a with installation of a second 13 MW ball mill, a
33 per cent increase in rougher flotation capacity, a 40 per cent
increase in second cleaner capacity, and 33 per cent increase
in third cleaner capacity.
The Phu Kham IRP was completed under budget and
schedule in 2013, resulting in increased copper recovery
of five per cent and increased gold recovery of at least
ten per cent over the first year of operation at ore throughput
rates exceeding maximum design of 2250 t/h by up to
ten per cent. The increased recovery was achieved despite
decreasing ore quality over the period. In addition, the extra
cleaner capacity allowed throughput to be increased in order
to increase copper feed tonnes to the plant despite the lower
head grades.
The plant optimisation and development programs and
major plant designs at Phu Kham have been supported
by extensive mineralogy, mineral association and mineral
liberation data from plant monthly composites and bench
scale test products. The collection and analysis of this data
has revealed reasons for copper and gold losses and mineral
deportment, and significant opportunities for increasing
recovery of copper and gold have been identified.
Throughput Forecasting and Optimisation
Project Background and Scope
Following the completion of the PKU and IRP circuit
expansions in 2013, the processing plant capacity was
increased to a design maximum capacity of 2250 t/h,
equivalent to 18 Mt/a. As mining extends deeper into
the deposit, the operation will experience an increased
proportion of highly competent ores which have the potential
to limit plant throughput, in particular through the 13 MW
SAG mill. Insufficient comminution data in the mine block
model created a lack of confidence in the ability to predict
mill throughput, particularly in the later years of the mine life.
Phu Bia Mining commenced a throughput forecasting and
optimisation project in 2012 to evaluate how to maintain the
227
D Bennett et al
target throughput over the LOM. Metso PTI was engaged
to conduct a full process integration and optimisation (PIO)
project. The main objectives were to:
•• review the current blasting, crushing and grinding
processes, identify opportunities for increasing
throughput, and improve overall comminution circuit
performance when treating very competent ore types
•• develop a throughput prediction model based on
geometallurgical modelling for long-term planning
and optimisation
•• identify if and when secondary crushing or other process
changes will be required over the LOM to maximise
plant throughput.
Methodology overview
Metso’s PIO methodology involves the development of
integrated operating and control strategies from the mine
to the plant that maximise throughput, minimise the overall
energy consumption, cost per ton and maximise profitability.
This requires an understanding of the physical properties and
composition of the orebody, where the valuable mineral is
located within it, and what mineral associations exist between
the ore and gangue.
The process starts with ore characterisation to define
domains within the orebody that will behave similarly
throughout the blasting and comminution processes. The
SmartTag™ ore tracking system developed by Metso PTI
(La Rosa et al, 2007) is used to track the characterised ore from
the mine, through the crusher and finally into the grinding
mills. With the ore source and characteristics known, detailed
audits of the blasting and processing operations are used to
develop site-specific predictive models for each operation
(blasting, comminution, separation). Using these predictive
models, the blast design is optimised to generate optimal
run-of-mine (ROM) fragmentation for all ore types, and
downstream processes can be adjusted accordingly. The
models also allow prediction of throughput and recovery
performance for each ore domain, and when combined with
the mine plan can be used for forecasting, planning and
optimisation purposes.
The SmartTag™ system is also used to increase the accuracy
of geometallurgical modelling and update the block model
automatically. The ore is tracked from the mine through
the process with SmartTags™ and linked with the plant
control system (DCS) to provide actual plant performance
data (throughput, recovery, grade etc) for each ore type and
associated blast conditions. These data are automatically
compared with model predictions and updated in the block
model using the SmartTag™ software. Incorporation of the
actual plant data into the block model in real-time eliminates
the need for further expensive ore characterisation tests. More
accurate data in the block model improves mine planning,
and the plant receives advance notice of the ore type about to
be processed. Adjustments can then be made to blast designs
and operating conditions to optimise performance.
can be measured with laboratory tests such as Point Load
Index (PLI), Drop Weight or SMC tests (DWi), and Bond ball
mill work index (BWi). The unconfined compressive strength
(UCS) is a common measure of strength and can be estimated
from PLI values to reduce laboratory testing requirements.
In general terms, the rock structure affects the coarse end of
the ROM fragmentation, while the strength (hardness) affects
the generation of fines. Improved plant throughput can be
achieved by manipulating ROM fragmentation through
optimisation of blasting to reduce top size and increase fines,
especially for SAG mills.
At Phu Kham, ore domains were defined based on the RQD
(for structure) and PLI (for strength) values in the geotechnical
block model. This resulted in a matrix of nine ore domains as
shown in Figure 39. Ore within a domain will produce similar
ROM fragmentation for a given blast design.
An increase in harder ores is expected in the future at Phu
Kham as the pit deepens, and has the potential to reduce
throughput. Consequently, the focus of the project was on
increasing throughput and improving overall comminution
circuit performance when treating the hardest ore types.
Therefore, the hardest ore available was selected for a trial
blast and was followed by detailed auditing of the blasting
and processing operations. For the trial blast, RQD data
was obtained from the geotechnical block model, and point
load tests were conducted on stockpile samples. These
indicated that the trial blast consisted of moderately hard, but
reasonably low quality, jointed and fractured rock.
Blast modelling and simulations
One of the main objectives of the project was to develop
strategies to maximise mill throughput to maintain LOM
operational targets even when treating harder ores. This
can be achieved by improving ROM fragmentation through
optimisation of blasting practices. A reliable model of blast
fragmentation is required to determine the effect of changing
blast parameters on ROM fragmentation.
The Metso PTI blast fragmentation model, which is sensitive
to the major parameters known to affect blasting performance,
was calibrated using the ore characterisation data and design
parameters from the audited blast. Image analysis of the
ROM size distribution produced by the trial blast was used
to calibrate the coarse size fractions. A limitation of image
analysis techniques to determine ROM size distributions
is that they cannot effectively delineate fines; therefore, the
size distribution of the primary crusher product belt cut
sample was used to correct the fine portion of the curve.
A comparison between the measured and model generated
ROM size distribution is shown in Figure 40. The model
Ore characterisation
The optimisation methodology starts with ore characterisation
in terms of structure, strength and comminution properties.
These are used to define ore domains within the geotechnical
block model with similar blastability and fragmentation
properties.
Rock structure is determined by the in situ joints and
fractures and can be quantified with rock quality designation
(RQD), fracture frequency and joint mapping. Rock strength
228
FIG 39 – Definition and distribution of ore domains at Phu Kham.
We are metallurgists, not magicians
Process development and throughput forecasting at the Phu Kham copper-gold operation, Laos PDR
FIG 40 – Blast fragmentation model.
predictions correlated well with measured values at both
the coarse and fine ends of the particle size distribution. This
demonstrated that the model was quite accurate in predicting
ROM fragmentation, and suitable for simulation studies.
Simulations were conducted using the blast fragmentation
model to investigate the impact of changes in spacing,
burden, stemming length and blasthole diameter on ROM
fragmentation. The parameters of the current blast design and
five selected scenarios (with 115 mm diameter blastholes) are
provided in Table 9. The corresponding model predictions of
ROM size distribution are shown in Figure 41.
As expected, simulations indicated that tightening the
blast pattern to increase the powder factor (PF) resulted
in a significant increase in the fines generated in the blast.
Reduction of stemming length also generated more fines and
reduced the top size of the rock due to the increased explosive
energy at the stemming horizon. These simulations indicated
potential to increase throughput by increasing the fines and
reducing the top size of the ROM fragmentation by optimising
the blasting parameters.
Simulations were conducted for each of the nine ore domains
(using both 115 mm and 127 mm blasthole diameters). This
allows the blast design to optimised for each of the ore domains,
and a ‘cookbook’ is generated which provides a ‘recipe’ (ie an
optimised blast design for each ore domain). Blasting according
to this cookbook provides a more consistent and optimised
feed size distribution to the downstream processes, increasing
throughput, process stability and efficiency. Following the
FIG 41 – Selected blast fragmentation simulation results.
cookbook avoids excessive blasting in softer ore domains,
thus reducing energy consumption and costs, and preventing
the excessive production of ultra-fines that can be detrimental
to some downstream processes. This has been successfully
applied by PTI at several other large open pit operations
globally (Rybinski et al, 2011; Burger et al, 2006). The final
definition of the blasting cookbook at Phu Kham is ongoing.
SmartTagTM Ore Tracking
To link the process performance with ore characterisation and
blasting outcomes, the ore from the trial blast was tracked from
the mine through the process using SmartTagTM ore tracking.
The SmartTag™ ore tracking system developed by Metso
PTI allows parcels of ore to be tracked from the mine, through
the crusher and finally into the grinding mills, as shown in
Figure 42. The SmartTags™ are built around robust passive
radio frequency (RFID) transponders. They do not have an
internal power source, so they can remain in stockpiles and
ROM pads for extended periods of time. Antennas to detect
the SmartTags™ are located at critical points in the process
ahead of the milling circuit; tags can be detected a number
of times and provide valuable information on material
movements. In particular, they make it possible to link the
spatial data associated with the ore in the mine to the timebased data of the concentrator.
At Phu Kham, SmartTag™ antennas were installed under the
crusher product and SAG mill feed conveyors. The installation
TABLE 9
Selected blast designs.
Current design
Scenario 1
Scenario 5
RQD (%)
25
UCS (MPa)
60
Bench height (m)
Scenario 9
Scenario 13
Scenario 17
10
Burden (m)
3.5
2.6
3
3.4
3.8
4.2
Spacing (m)
4
3.2
3.6
4
4.5
4.9
Subdrill (m)
1
Stemming length (m)
3
1.8
2.2
2.5
2.8
3.2
Explosive type
70% emulsion
Density (g/cc)
1.15
Powder factor (kg/m3)
0.69
1.34
0.99
0.76
0.58
0.46
Powder factor (kg/t)
0.26
0.5
0.36
0.28
0.21
0.17
Change in powder factor (%)
--
94.2
43.5
10.1
-15.9
-33.3
We are metallurgists, not magicians
229
D Bennett et al
FIG 42 – SmartTag™ ore tracking.
on the SAG feed belt is shown in Figure 43. SmartTags™ were
inserted into the stemming column of every blasthole for the
audited blast. The origin of each SmartTag™ is saved with
its unique identification number (ID). As the SmartTags™
and associated orepass the antennas in the process plant, the
system automatically records the time and tag ID, thus the
source of the ore being processed at any given time is known.
During the project at Phu Kham, this ensured that ore from the
trial blast was being fed to the concentrator during the plant
audits, and allowed correlations to be established between ore
origin and process performance.
Comminution modelling and simulations
A comprehensive grinding circuit survey was successfully
conducted on 4 February 2013 when the circuit was treating
ore from the trial blast (as determined by SmartTag™ ore
tracking). At the time of the grinding survey the plant was
operating under SAG mill limiting conditions, with a total
mill load of 29 per cent and mill power draw of 12.0 MW.
Ore samples collected during the site survey were sent for
ore characterisation tests, including drop weight (DWT) and
Bond ball mill work index (BWi) testing. These tests provide
ore parameters required for comminution modelling, and the
results from the survey sample are provided in Table 10. The
A*b and ta parameters are determined from the DWT. The
A*b value for the ore puts it in the ‘moderately soft’ category
in terms of resistance to impact breakage, and the ta value
falls into the ‘soft’ range for abrasion resistance. The BWi test
conducted at a closing sieve size of 106 µm resulted in a BWi
TABLE 10
DWT and BWi test r