Research on soft-sensor based on support vector regression for particle size of grinding and classification process

Particle size is the key technical index of grinding process, but it is difficult to be measured online directly, and the lab analysis has large-time delay. Combined with the advantage of soft-sensor modeling based on support vector regression (SVR), a soft-sensor approach for particle size of grinding process with SVR is proposed. Firstly, the empirical formulae are used to roughly determine the range of the model parameters of the SVR, and then the particle swarm optimization (PSO) algorithm with improved inertia weights is adopted to obtain the optimal parameters of the SVR model to enhance the generalization capability. In this paper the empirical model and BP neural network model are established in order to analysis and prove the superiority of the proposed SVR soft-sensor. Finally, a software package based on the proposed method is developed. The application results show that the proposed method is easy to use and can realize the precise measurement of particle size online.