Methods for automatic control, observation, and optimization in mineral processing plants

For controlling strongly disturbed, poorly modeled, and difficult to measure processes, such as those involved in the mineral processing industry, the peripheral tools of the control loop (fault detection and isolation system, data reconciliation procedure, observers, soft sensors, optimizers, model parameter tuners) are as important as the controller itself. The paper briefly describes each element of this generalized control loop, while putting emphasis on mineral processing specific cases.