Optimizing grinding mill performance using extremum seeking control

Abstract

A semi-autogenous grinding mill is simulated with an extremum seeking controller to maximize the performance of the mill using grind curves. Grind curves map the essential performance measures of a grinding mill to the manipulable variables. The curves vary with the changes in the feed ore characteristic but show generic parabolic features with extremums. The extremum seeking controller searches along the unknown input-output map using periodic perturbations to steer the process towards an unknown optimum. In this study the controller searches along the grind curves to either optimize the mill throughput or grind by means of the mill feed or rotational speed. The proposed controller will reduce the need for a plant operator to select the optimal operating region. Since the controller is agnostic to the process model, it will also eliminate the tedious process of developing an accurate plant model for model-based control.