Dynamic prediction model for mixed concentrate grade of mineral processing plant

摘要

A non-linear modelling approach of dynamic prediction model for mixed concentrate grade consisting of a linear part and a nonlinear part is developed. The nonlinear part is implemented using the least squares support vector machine (LS-SVM), where the problem of selecting model parameters is transformed into the probability distribution function (PDF) control of the modelling error. Both the PDF control based and minimum entropy based model parameter selection approaches are proposed. The experiment results show the effectiveness of the proposed approaches.