基于ISSA-DELM算法的CSTR系统广义预测控制研究OA北大核心CSTPCD

The continuously stirred tank reactor(CSTR),as a typical equipment used in chemical production of polymerization reaction,has strong nonlinearity,large lag and uncertainty during operation,so it is difficult to establish an accurate mathematical model with the traditional methods.In this paper,the Hammerstein-Wiener model is adopted according to a class of CSTR reaction processes,and the two nonlinear modules of the model are modeled by Gaussian radial basis function LS-SVM,and the established Hammerstein-Wiener model is used as the prediction model for generalized predictive control(GPC).For the rolling optimization of GPC,a hybrid optimization algorithm strategy based on deep extreme learning machine(DELM)which is optimized by multi-strategy improved sparrow search algorithm(ISSA)is adopted,and the superiority of the ISSA is tested by reference function.The hybrid optimization algorithm is applied to nonlinear CSTR objects.The experimental results show that the proposed hybrid optimization algorithm ISSA-DELM has a good control effect on CSTR system.The simulation results of the SSA-DELM algorithm and the DELM algorithm show that the control effect of the proposed algorithm is significantly better than those of the SSA-DELM algorithm and the traditional DELM algorithm.