Can crossbelt sample cutters be trusted

1
Can cross-belt sample cutters be trusted?

• Geoff Robinson, Matt Sinnott, Paul ClearyCSIRO
  Mathematical and Information Sciences
• Based on presentation to Sampling 2008 conference
  held in Perth 27-29 May 2008

2
Structure of presentation

• Background knowledge
• DEM simulation results
• Description of most important bias mechanism
• Material thrown off belt
• Portion of stream appears to be over-represented
  by about 40
• What should happen next?

3
ACARP project C15072intended to have two parts

• Cross-belt samplers in practice
• Monitoring of increment masses
• High-speed video
• Physical bias testing (preferably for particle
  size)
• Simulation using Discrete Element Modelling
• Four sampling planes
• Models to be calibrated as well as possible
• Use 12mm bottom size to keep computations
  practical (about 60 by mass is really smaller)

4
Four sampling planes

• Just inside the jaws of the cutter
• Upstream classifying plane
• Classify by current horizontal position
• Classify by current vertical position
• Whether particle is in reference sample (i.e. its
  centre of gravity would pass through Sampling
  Plane 1)
• Through edge of belt to count all particle thrown
  off the belt
• Entry to sample chute (subset of plane 3)

5
Mechanisms that may cause bias

• Modelled
• Particles spill off belt without entering cutter
• Poor contact between cutter and belt
• Inadequate capacity
• Not modelled
• Timing of samples correlated with quality
• Material sitting in imperfections in the belt
• Loss through gaps in front of scrapers
• Bouncing off back of chute
• Variable cutter speed

6
Sample delimitation is correcteven if cutter
speed varies
7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
Choice of cross-belt cutter parameters

• Based on a real cutter for which we have
  drawings. Edges are portions of ellipses.
• Conveyor belt has circular cross-section with
  radius 0.68 m
• 1.2m wide belt moves at 4 m/s, cutter at 6 m/s
• Cutter 33.7 degrees from square to belt
• Belt on 20 slope (11.3 degrees), loading 1600
  tonne/hr
• Cutter aperture 150 mm (centre-to-centre)
• Cutter blades 10 mm thick with 3 mm diameter
  rounded edges

25
Particle properties

• PSD on a mass basis between 12 mm and 50 mm
• About 60 of particles taken to be 12mm
• Coefficient of restitution (e) 0.25
• Friction coefficients
• coal-coal 0.8
• coal-belt 0.9
• coal-cutter 0.5
• Density 1350 kg/m3
• Super-quadric indices uniformly from 2.5 to 6
• Aspect ratios uniform from 0.7 to 1

26
Note that material is thrownDefinition of chute
27
Old film of a cross-belt cutterNote amount of
coal thrown off even though all coal is fine
28
Classification by position on belt
29
Results for base case
30
Description of experimental program

• Details in written paper
• But have added runs with single rods instead of
  sample cutters
• Compare skew and square cutters
• Compare cutter apertures
• Investigate sensitivities
• Does contact to edge of belt matter much?

31
View of DEM cutter from downstream
32
View of DEM cutter from upstream
33
View of DEM cutter from above
34
Three regions
35
Region-specific extraction ratios
36
View of rod moving at 6mps
37
1.5mm diameter rod
38
View of rod moving at 9mps
39
View of rod moving at 4mps
40
Sample cutter correct only if blades move slowly?
41
Inclined Rod
42
Masses (kilograms)

• Base case reference sample 20.83, 20.82
• Chute sample 0.84,0.96 missed 3.42,2.91 extra
• Thrown off 0.69, 0.80 missed 6.27, 5.98 extra
• Thrown by 4m/s rod 2.47, 2.82
• Thrown by 6m/s rod 3.10, 2.93
• Thrown by 9m/s rod 3.74, 3.81
• Thrown by 1.5mm rod at 6 m/s 2.65, 2.89
• Thrown by inclined rod 2.11, 1.95

43
Future research

• Physical experiments to estimate mass thrown off
  by cutter blades (no back on cutter)
• DEM models to investigate cutters at higher
  angles
• Lower cutter speeds

44
What can we display?

• From any viewpoint
• Only show particles satisfying some condition
• Colour by speed, a component of velocity,
• Can make cutter transparent
• However
• Particles disappear at edge of control volume
• Lighting model is rudimentary

45
(No Transcript)
46
Practical implications

• Monitor extraction ratios
• Be suspicious of cross-belt samples from
  segregated flows
• Expect more sample bias for high-speed cutters
  and when belts are loaded nearly up to the edge
• Strongly prefer skew cutters to square cutters
• Slower may be better
• Delete 1.5 times belt speed clauses in
  Standards
• Slightly prefer screw-shaped cutter sides

47
(No Transcript)
48
(No Transcript)
49
Determining an estimate of maximum likely sample
bias

• When belt is carrying extremely segregated
  material, stop it.
• Put sampling frame onto belt
• Shovel 1/3 belt loading from final-high region
  into a container for assaying (A)
• Shovel remaining 2/3 into container (B)

50
Can cross-belt sample cutters be trusted?

• Sample delimitation is correct
• Sample extraction may be OK for slow cutters
• Sample extraction is biased for fast cutters, but
  the extent of the bias can be bounded
• Measure stream segregation
• Final-top region over-represented by about 50
• Physical experiment using cutter with no back is
  the next step