Density clustering analysis of fuzzy neural network initialization for grinding capability prediction of power plant ball mill

Abstract

Ball mill of thermal power plant has high energy consumption and the grinding capability is usually used for representing the efficiency of ball mill. This paper proposes a density clustering analysis method of fuzzy neural network initialization for grinding capability prediction of power plant ball mill. The proposed method integrates the density clustering algorithm and the fuzzy neural network to predict grinding capability, where the density clustering algorithm is used to initialize the rules base of the fuzzy neural network. Furthermore, two parameters of the density clustering analysis can be determined by calculation formula, and the structure of the proposed model could be optimized by the training capability of neural network. The experiments are performed on two datasets obtained from the thermal power plant under the stable conditions. The experiments results verify that the proposed model has higher effectiveness. In addition, the proposed model has been put into practice and the field operation curve proves that the grinding capability could be predicted correctly.