Innovation en géométallurgie et

comminution

Frédéric Couët, ing., Ph. D.

© COREM (2018)

© COREM (2018)

Comminution Crushing and grinding

2

Energy bandwidth study (DOE, 2010)

4%

© COREM (2018)

Geometallurgy Spatial model and process model

3

© COREM (2018)

COREM

© COREM (2018) From Metso.com

© COREM (2018)

Drop weight test

© COREM (2018)

GeM Comminution index

Geometallurgical mapping

7

Limitations

Low correlation with A x b

Variable feeding

Large variations in shapes

Closed side setting (CSS)

drift

Use of a power meter to

quantify the energy

consumed in crushing

Kojovic et al. (2010) XXV International mineral processing congress.

© COREM (2018)

AG/SAG mill testing Procedures available to characterize grindability

8

Verret, F.-O., Chiasson, G. & McKen, A., 2011. SAG mill testing - An overview of the test procedures available to characterize ore grindability. SGS.

❶

❷ (SVT)

© COREM (2018)

SAGDesign SAG variability test (SVT)

9

Uses a bench scale mill

similar to SAG mill

operation.

Is similar to Bond

methodology.

Gives SAG grindability

in kWh/t.

© COREM (2018)

Crusher indicator versus SVT Set-up and methodology

10

10

❶ ❷

Some samples were tested at two different closed size settings.

© COREM (2018)

Linear regression model SVT ~ b0 + b1 F70 + b2 P70 + b3 P70 Em

11

(5 times 5-fold cross-validation)

© COREM (2018)

Repeatability 16 repetitions on a reference material

12

© COREM (2018)

Robustness Closed side setting drift

13

❶

© COREM (2018)

Effect of iron grade on rock hardness

14

Composition is correlated to Sag Variability Tests

for drill cores.

𝑊𝑜𝑖 ↑ ~𝐼𝑟𝑜𝑛% ↓

Pérez-Barnuevo L. et al. (2016) IMPC, Paper # 304 -

1.0

2.0

3.0

4.0

5.0

58.0 60.0 62.0 64.0 66.0 68.0

SVT

(kw

h/t

)

%FeT

© COREM (2018)

Characterization A typical characterization campaign scenario

15

50: SVT (±10%)

500: Crusher index (±17%)

All: Chemical analysis

Models

– SVT ~ Crusher index (with energy)

– SVT ~ Crusher index (without energy)

– SVT ~ Chemical analysis

Grindability for all samples but with different

confidence intervals

© COREM (2018)

Sequential spatial sampling

SVT Crusher indicator

Couet, F. et al., 2015. A new methodology for geometallurgical mapping of ore hardness. Proceedings of SAG conference, Vancouver.

© COREM (2018)

Real-time control and reconciliation Online hardness indicator at plant scale

17

© COREM (2018) 18

Segmentation of bimodal size

distributions

© COREM (2018) 19

3D vision system Plant data acquisition and analysis

Estimate the conveyor

geometry.

Measure the speed of

the conveyor (real-time).

© COREM (2018) 20

3D vision system Plant data acquisition and analysis

Region of interest is

discretized in bands.

Compute minimum of each

band to estimate the

amount of fines.

Make an histogram of

everything above the

minimum to describe the

shape of size distribution.

Calibrate α to estimate how

much coarse particles float

on fines.

© COREM (2018) 21

3D vision system Plant data acquisition and analysis

Use volume and speed

to estimate the volume

flow rate.

Possibility to estimate

bulk density and iron

ore content.

R² = 0.964

0

500

1000

1500

2000

2500

3000

3500

0 1000 2000 3000

Ton

nag

e f

rom

cam

era

(t/

h)

Nuclear scale (t/h)

Tonnage validation (AMEM)

© COREM (2018)

3D vision system

Plant data acquisition and analysis

© COREM (2018)

Real-time control and reconciliation Online hardness indicator at plant scale

23

© COREM (2018)

Real-time control and reconciliation

Reconciliation of the block model

24

Wambeke, T. & Benndorf, J., 2017. Mathematical Geosciences, 49(1).

Information from the

concentrator is used

to update locally the

block model.

© COREM (2018)

Traceability – Silo model To

ns

© COREM (2018)

Traceability – From crusher to AG mill

26

© COREM (2018)

Traceability – From mine to AG mill

27

corem.qc.ca

SAGTools

© COREM (2018)

What is SAGTools?

Mill filling to optimize throughput

29

Powell & Mainza (2006) Extended grinding curves are essential to the comparison of milling performance.

© COREM (2018)

Charge volume soft sensor

30

Austin SAG Model

https://fredcou.shinyapps.io/demo_power_model/

F(DEM) SAG Tools

Calibration process

© COREM (2018) 31

Pilot plant mill instrumentation

Installation of 3D cameras in a 6’x2’ AG/SAG mill

© COREM (2018)

What is SAG Tools? Discrete element method

Quebec Mine 2015

© COREM (2018)

What is SAG Tools? Inputs and outputs

33

AG or SAG

PLC

SAG Tools

Web browser

Inputs

Power

RPM

% solids

Pressure (overload model)

Tonnage (wear model)

Geometry, media, wear, etc.

Outputs

Charge volume

Overload risk

Media wear

Liners wear

Angles

© COREM (2018)

What is SAG Tools?

User interface

34

Shell feed Shell middle Shell discharge

Quebec Mine 2015

corem.qc.ca

Mineralogy

© COREM (2018)

Mineralogical tools

36

Lévesque, S. et al., 2016. Mineralogical tools for ore characterization – key data at all steps of iron ore concentration. IMPC.

© COREM (2018)

Bulk mineralogical analyser

© COREM (2018)

Bulk mineral analyzer

38

IMPC 2016

© COREM (2018)

Drill core textures

© COREM (2018)

Drill core textures

© COREM (2018)

Drill core textures

© COREM (2018)

Geometallurgy Spatial model and process model

42

© COREM (2018)

Forward

Integrated - Horizontally (Mine-to-mill) - Vertically (Variable resolution) Native support - Multicomponent - Optimization - Machine learning - Stochastic Mineralogy