The Intelligent-Learning Control for Overflow Concentration of Grinding-Classification Operation System

摘要

Algorithm of intelligent-learning controller and a prediction method are used to control the overflow concentration of grinding-classification system. Uncertainty factor and non-linear, time-delay, however, prevail; they introduce difficulties to mathematical modeling. By applying the knowledge system with mineral property to optimize the controller setting value. Practical production operation proved that the ore feeding of ball mill improved 6.0%, reclaiming rate added 2.0% and the average stable state error of overflow density is less than 3.0% in addition to the whole control system has good robustness and stability.