On Ls-svm-based Nonlinear Model Predictive Control

Model predictive control (MPC) has become one of the most successful advanced control techniques in the chemical and petrochemical industry recently. Model predictive control refers to a class of control strategies, in which a dynamic model is established to predict the future behavior of the unknown system and the system performance is further optimized according to the model. Research on linear predictive control has become mature and linear MPC has gained wide application in industrial processes. However, most processes in industry are nonlinear, time-variant and bear uncertainty, thus research on nonlinear predictive control has become an important issue in control field. As the expansion of production scale and the complex degree increase, mechanism-modeling approach has become more and more difficult. As a result, it is necessary to establish the identification model by using the available experimental or manufacturing data. In recent years, least squares support vector machine (LS-SVM) which based on statistical learning theory has been successfully applied in pattern recognition, system identification. According to the controlled object which has nonlinear behavior, the LS-SVM method of intelligent modeling and the optimization strategy of the corroding controller is studied. The main research contents are summarized as follows:(1)Based on deeply understanding of the LS-SVM regression theory and algorithm, LS-SVM model identification is investigated.simulation results show the superiority of the algorithm.(2)A nonlinear predictive control algorithm based on LS-SVM model is proposed. The model of the nonlinear system is obtained by LS-SVM, chaotic optimization algorithm as online optimization. The simulation results show the effectiveness of the nonlinear predictive controller.(3)A nonlinear predictive function control algorithm based on LS-SVM model is proposed. The model of the nonlinear system is obtained by LS-SVM, the offline model is linearize at each sampling instant and uses linear predictive function control methods to obtain the control law. The simulation results show the algorithm has good control performance and anti-interference capability.