A new model for prediction of classification performance in closed circuit ball mill systems

The traditional model of closed circuit ball mill systems has been used for several decades, however, if the classifier of the closed circuit ball mill system performs the duties of both pre-classification and checkclassification, the characterization error of the traditional model is large. To address this problem, a new model is proposed by modifying the traditional one. The results show that the new model characterizes the relative capacity of the ball mill more accurately, with a concentration of data at 65 % (classification efficiency) compared to the concentration of data at 50 % in the traditional model. The circulating load calculated by the new model is 360.81 %, and the corresponding slurry level in the ball mill is about 50 %, which is more consistent with the actual level. The new model has a higher accuracy than the traditional model in characterizing the production status of the grinding system, which is of some significance for industrial production.