Upload
votuong
View
219
Download
0
Embed Size (px)
Citation preview
1
Xstrata Process Support (XPS)
Falconbridge, OntarioCOM 2009 Geometallurgy Short Course August 22 – 23, 2009
Peter J. Whittaker, Ph.D., P.Geo.
2
Process Mineralogy at Xstrata Process Support: Outline
•Introduction
•Applications and Enabling Concepts
•Geometallurgy•Data Review
•Sampling
•Mineralogical Input•Metallurgical Input
•Examples / Case Studies•Conclusions
3
XPS Process Mineralogy
Process
Mineralogy
Sampling & Statistics
4
XPS Process Mineralogy
• XPS Process Mineralogy is an integrated multi-disciplinary study of geology, sampling, mineralogy, and
mineral processing that links the orebody to the concentrator process, providing improved mineral separation
• Designs more suitable flowsheets to treat the orebody
more efficiently
The orebody consists of minerals not assays
5
Process Engineers &
Geoscientists
• Work closely together as integrated project teams
• Ore mineralogy data feeds into process design
[ can be concentrators, hydrometallurgy, pyrometallurgy, environment ]
• Product mineralogy leads to process optimization
• QA/QC embedded in all procedures
6
Three Applications
On-site Support &
Plant Optimization
( Statistical Benchmark Surveys, Troubleshooting, Lab
Testwork )
New Orebodies, New Concentrators
(Exploration, Plant Startup)
Replacement Orebodies
at Existing Mines
(Brownfields)
7
Enabling Technologies & Concepts
• Samples not Specimens• Drill-Core Sampling Models• Map Entire Orebody – Life-of-Mine View• Statistical Benchmark Surveying
• Minerals not Assays• QEMSCAN• Probe
• Geometallurgical Unit Definition• High-Confidence Flotation Testing• Distribution Modelling• Mini-Pilot Plant
Methods and Procedures Developed
to Support Enabling Technologies at
the 95% Confidence Level
Outcome:
Accurate Plant Design and Scale-up, Right Capex, No Surprises
8
Geometallurgy
Ore type or group of ore types
that possess a unique set of
textural and compositional
properties from which it can be
predicted they will have similar
metallurgical performance
A Geological and Metallurgical Definition
gabbro footwall
massive sulphides
net-textured sulphides
disseminated sulphides
1m
9
XPS Flowsheet Development
Drill Core
Data Review
Sample Plan
GeoMet Unit Definition
Virtual Flowsheet
Grain Sizes
Modal Analysis
Mineral Compositions
Summary Modal Mineralogy
Averaged by Blends
0% 20% 40% 60% 80% 100%
Blend 1
Blend 2
Blend 3a
Blend 3b
Blend 4
Blend 5
Blend 6
Pyrite
Chalcopyrite
Chalco_Text
Tennantite/Enargite
Sphalerite
Molybdenite
Quartz
Orthoclase
Plagioclase
Qtz-Plag-Orth/Text
Garnet
Garnet-Textures
Diopside
Amphibole
Muscovite
Biotite
Calcite
Fe Oxides
Alteration Minerals
Other Sulphides
Other Silicates
Accessory Minerals
Other
QEMSCAN
MicroProbe
PTS or
Coarse
Composites
Results
and Review
Mini Pilot Plant
Bankable
Design Criteria
Grade/Rec for
EconomicsFlotation
Testing/DOE/
Optimization
10
Ore Type AssessmentZone Variability – Data Review
11
Proposed Zones for Geomet Unit Samples
EG
N
C
12
Sampling
Aim is to produce a representative sample of a given ore type
Composite samples from drill core enable this
Smaller sample masses can result
Ore type composite – average mineralogy and metallurgical results
for the given ore type
Variability composites – test variability within a given ore type
Outcome can be that different ore types may have similar metallurgicalresponse and can be grouped, or a problematic ore type may be identified
and isolated
13
Stratified Sampling
frequency distribution
grade
Zone X
A B
C
A wide variance will require a largersample mass
Tighter variance for each unit can be Represented by smaller masses
14
Capable of processing up to 150kg/hr
dry solids
Drillcore or 6 inch rock feed size
Any product (mesh) size down to
1.7mm can be achieved
Revolutionary blending technology
Primary Jaw Crusher
Rocklabs Boyd Crusher
Vibrating screens
Spinning rifflers and load cells ensure
replicate samples are produced
Commissioned October 2008
Cru
shin
g &
Ble
nd
ing
Pla
nt F
eatu
res
Stratified Sampling – Sample Preparation
15
Sampling Case Study
•Powerful Sampling Model
By GU
Reference Distributions
•Produces Representative Samples for
Testing
Spatially Representative
Porphyry Cu, Chile
Error of <1% on Cu grades gives confidence in samples to be used for metallurgical test work
0
0.2
0.4
0.6
0.8
1
1.2
HHG HLG WES SES TUF
Block Model
Sample
16
Stratified Sampling
•Benefit of stratified sampling applies to smaller
and representative masses
•Mineralogical work is more representative
•Metallurgical test work is tied to the same sampleas mineralogy
•Impact of variability on metallurgical performance
can be tested
•Provides a basis for modeling different blends of mill
feed or different life of mine blends
17
Mineralogical Input
Data from automated mineralogy: QEMSCAN, MLA
modal %
grain size averages and distributionsmineral associations
mineral liberation
texture %
Data from quantitative mineral analyses; EPMAmineral compositions, particulary trace concentrations
Combined data gives metal deportment to recoverable and
non-recoverable minerals, identifies problematic minerals and texturesfor metallurgy
18
Quantitative Mineralogy: Minerals not Assays
•QEMSCAN & Microprobe
19
Impact of Texture on Mill
Particles (+106um)Lost to Tailing
1.7 mm
20
Textures can create metallurgical problems –
quantification of textures can become critical
•Mineralogical measurements can be grouped into proportions of textural and/or
mineralogical populations within each composite which may have processing implications.
21
Modal mineralogy of gangue as
critical as that of pay minerals
22
Ni Distribution by Mineral
0
10
20
30
40
50
60
70
80
90
100
Ni
% D
istr
ibu
tio
n
Ni (Pn) Ni (Po)
Main North Shallow North Deep UM1A
Ni in Pn - ~31%
Co in Pn - ~2%
23
Mineralogical data from testing of deposit variability plus knowledge of current best
practice leads to preliminary flowsheet
Zinc 1st Clnr Scav
Tailings
Zinc 1st Clnr
Scav Conc
Copper Rougher
Primary Grind
MS Ball Mill
p80 = 45 microns
Copper Regrind
Ball Mill
p80 = 20 microns
Zinc Rougher
Final Tailings
Zinc Regrind
Ball Mill
p80 = 20 microns
Zinc 1st Cleaner
Zinc 3rd Cleaner
Zinc 1st Cleaner
Scavenger
Zinc 3rd Cleaner
Concentrate
Zinc 2nd Cleaner
Tailings
Copper 2nd Cleaner
Concentrate
Zinc 3rd Cleaner
Tailings
To Final Tailings
Talc Rougher
Talc Cleaner
Copper 1st
CleanerCopper 1st
Cleaner Scav
Copper 2nd
Cleaner
Zinc 2nd Cleaner
Final Tailings
24
Flotation products – test by GU or
variability composite for performance diagnostics
-420/+
212
-212/+
106
-106/+
53
CS1-2
CS3-4
CS5
CS6
Calc
CS7
Locked
Middling
Liberated
0
5
10
15
20
25
30
35
% C
halc
op
yrit
e
Size Fraction
Chalcopyrite Liberation in 2nd Cleaner Concentrate
Locked
Middling
Liberated
Locked Middling Liberated
Fraction -420/+212 0.03 0.03 0.06
-212/+106 0.15 0.39 3.32
-106/+53 0.24 1.07 18.15
CS1-2 0.30 1.53 30.96
CS3-4 0.22 1.24 20.87
CS5 0.04 0.26 6.72
CS6 0.17 0.96 9.54
Calc CS7 0.09 0.53 3.12
Combined 1.24 6.01 92.75
Strong positive flotation by liberated chalcopyrite
25
Bench scale high-confidence flotation
trials test the flowsheet
•Minimises Metal Balance Errors•Develops Reproducible Results•Sampling and Results at 95% Level of Confidence•Quality Control System
26
Flow sheet validation is carried out with a mini-pilot plant; smaller masses made possible by geomet units feed the MPP as
representative samples
• Sized for Drill-Core
• Continuous Operation
• Campaign 3-12 days (750-3000
kg ore)
• 10 kg/h nominal feed rate
• 1 Flowsheet/3 day run
• Allows for Testing by GU
• Wide range of Flowsheets
• Produces Design Level
• Mass/value Balance
• Grade/recovery Performance
• 95% Level of Confidence
– Quantitative Quality
Assurance Programme
Ability to Simulate Existing Concentrators Has Been Demonstrated – no need for pilot shaft
27
Raglan flowsheet developments tied to future oresprogram on a Geomet foundation with Process Mineralogy test
work validation
• Statistical Benchmark Surveying• Extracts Sample Suite
– at 95% Confidence Level
• QEMSCAN Measurement• Identifies Flowsheet Opportunities
Payback on Overall Investment
92% IRR
Raglan 1998-2002
NPV $12.7 million in 2002 financial terms
28
Montcalm – Predicting Startup
Timmins
29
Montcalm – Predicting Startup
Comparison of Montcalm Start-up Curve with McNulty Curves
0
20
40
60
80
100
120
0 2 4 6 8 10 12 14
Quarter after start-up
Ni
ou
tpu
t in
Ni
an
d C
u c
on
c,%
of
de
sig
n
Type 4
Type 3
Type 2
Type 1 Montcalm start-up
Montcalm start-up - October 2004
Equal to or better than Type 1
30
Montcalm Scaleup
•Successful Scale-Up
• High-Confidence Flotation Testing 2001-2003
– 82.9% Ni Recovery
• Statistical Benchmark Survey of Operations July 2005
– 84.0% Recovery
82.984
75
80
85
90
HCFT Survey
Type 1 Startup
31
XPS Process MineralogyHighlights of Kabanga Process
Development
• Scoping Phase 1 (2005)
– Ore Characterisation/Geo Met Unit Definition
• Scoping Phase 2 (2006)
– Recovery and Grade Validation/Minor Element and Self Heating assessment
• PFS (2007)
– Flowsheet Optimization
– Reagent/Circuit Configuration DOEs
– Mini Pilot Plant/Design Basis
• FS (2009)
– Self Heating Full Scale Trials
– Water Recycle Impacts
– Final QEMSCAN on Flotation Products
32
Mini Pilot Plant Results
MPP1 and 2 summary Grade/Recovery
15
16
17
18
19
20
21
22
23
60 65 70 75 80 85 90 95
Ni% Recovery
Ni %
Gra
de
MPP 1 Run 2 North
MPP 1 Run 3 North - Tembo
MPP 2 Run 1 North - Tembo
MPP 2 Run 2 North - Tembo - Main
MPP 2 Run 3 Tembo
HCFT North Opt.
HCFT Math. Blend North - Tembo
open sampling data
Tembo HCFT
33
Conclusions
•Geometallurgy is a powerful source of information to support Process Mineralogy
•Variability within a deposit in terms of mineralogy andits effect on metallurgical performance can be quantified
•Metallurgical risks can be identified and mitigated
•Flowsheet development time and validation to designcriteria and bankable feasibility level can be minimized, this represents large savings for project costs and improved ratesof return