Open menu Close menu
Case · Flow

Transfer box, concentrator plant

The model located the wear zone before assembly. The floor was redesigned and a shift stoppage avoided.

Mining Validated Demonstration
12.4 m/s
Peak jet velocity
Velocity field of the multiphase flow inside a transfer box
Velocity field of the multiphase flow inside a transfer box

Demonstration figures and fields: they show the format of a delivery, not an executed project. They are replaced by real cases with their validation before publishing.

The decision at stake

The box was about to be assembled with the straight floor of the original design. The question was not whether the equipment worked, but where it would wear first and whether that point could be reached during a shift maintenance window. A liner in the wrong place forces the line to stop for a change.

Geometry and mesh

The box, the inlet launder and the first metre of the receiving conveyor were modelled. The support structure was left out: it does not change the velocity field. Polyhedral mesh with refinement in the impact zone and a resolved boundary layer on the floor. The sensitivity study ran on three levels at refinement ratio r = 2.

Physics and assumptions

Multiphase flow with a free surface; slurry treated as a Newtonian fluid equivalent to the operating concentration. It was stated from the start that the model does not resolve particle segregation: that is what the Particle line is for.

Convergence

Residuals below 1e-4 in continuity and momentum, with the mass balance closed under 0.5 %. The cut-off criterion was declared before running, together with the acceptance criterion against measurement.

Validation

Three control points against site measurement: discharge jet velocity, slurry depth and flow per launder. The largest deviation was 4.5 %, within the ±5 % criterion agreed with the client.

Conclusion and limits

Redesigning the floor moved the impact zone to a face reachable from the maintenance platform. The conclusion holds for the measured flow range; outside it the impact point moves and the model has to be run again.