An intelligent approach for supervisory control of grinding product particle size

Grinding product particle size (GPPS) of grinding circuit (GC) is an important performance index directly related to the product concentrate grade and metal recovery rate. However, it is hard to control effectively with conventional process control strategies due to its complex characteristics. In this paper, an intelligent supervisory control (ISC) approach of GC is developed by employed intelligent techniques, such as fuzzy and artificial neural network (ANN). This DCS-based ISC system consists of a fuzzy adjuster, an ANN-based GPPS prediction module and an expert interface, and is used to supervise the grinding system and to adjust the setpoints of lower level control loop automatically. The outputs of these loops can therefore track their renewed setpoints so that a desired and optimized GPPS can been achieved. Industrial experiments show the effectiveness of the proposed ISC approach.