Iron ore beneficiation

In an iron ore concentrator the binding constraint is rarely the mill. It is throughput lost to a bottleneck that moves with the ore, and water that has to be recovered before the plant can take more feed. Both are control problems before they are capital problems.

$Dark fractured iron ore mass breaking into fragments and dust$

What closed-loop control addresses

A bottleneck that will not stay still

As ore hardness and feed grade change through the day, the circuit that limits the plant changes with them. Operators protect the plant by backing feed off the constraint, which is the right call and an expensive one. A controller that sees the whole chain can hold the plant closer to whichever constraint is currently binding.

Thickener performance and water recovery

When the tailings thickener approaches its limit on rake torque or bed condition, feed comes down. Controlling thickener behaviour and the classification feeding it together, rather than separately, removes a constraint that otherwise caps the whole plant.

Recovery and concentrate grade stability

Grade and recovery trade against each other, and the trade moves with the ore. Holding the plant at a chosen point on that curve, rather than letting it wander, is worth more than chasing either number alone.

Results

Two published case studies come from an iron ore concentrator, and they are about exactly the two constraints above.

Removing a tailings thickener bottleneck: the vessel spent about an hour a day less in the states that forced a feed cut, at lower bed levels and on less flocculant, with bed pressure unchanged.
Classification variability more than halved: steadier separation size cut and feed density, which is what made the thickener's load predictable.

Together the two controllers raised mean plant feed rate by 1.6 to 1.8% over the test period, measured against the ten most comparable historical periods. Both pages state what their source does and does not support.