Grinding granularity prediction based on improved particle swarm optimization and gray RBF neural network model

Abstract

Grinding technology is an important part of mineral processing, in which grinding particle size is an important production index in grinding classification operation, which is directly related to concentrate grade and metal recovery rate in mineral processing production. therefore, controlling grinding particle size is the key to the whole grinding process. However, grinding process is a rather complex physical and chemical process with many interference factors, large inertia and serious non-linearity, mutual constraints, making the grinding particle size prediction difficult. By analyzing the structure and characteristics of grey dynamic prediction model and RBF neural network prediction, grey neural network prediction model is used to predict grinding particle size, which realizes the complementary advantages of grey prediction and neural network, and using the particle swarm algorithm to optimize the grey neural network model. The experimental results show that the grey neural network has the advantages of small amount of data needed for prediction, small amount of calculation, can complement the advantages of the two, achieve better precision of grinding particle size prediction, and adjust the parameters of the grey neural network by using the improved particle swarm optimization algorithm, so as to obtain the optimal training parameters, and greatly improve the grinding particle size prediction accuracy.