Intelligent Optimal Control of Grinding Circuits for Optimization of Particle Size Index

Grinding circuits are characterized by being multivariable, non-linearity, severe coupling, large-time delay and time-variance. The key technical index, i.e. the particle size, is difficult to control directly with conventional loop control strategies. In this paper, an intelligent optimal control technique of grinding circuits achieving optimization of particle size index was proposed by combining intelligent methods including NN, CBR and IF-THEN rules-based reasoning. This hierachical control system is composed by a loop control layer and a loop setting layer. The formal consists of the loop controllers of the fresh ore feed, pulp density and overflow density, the later consists of a loop setting module, a soft-sensing predictor, a feedback compensator and a feedforward compensator. The loop setting layer gives the optimal setpoints to the loop controllers. This technique has been successfully applied in grinding circuits in a mineral processing plant. The particle size index has been kept in the required ranges. The control and operation of the grinding circuits achieved optimization