Research On Generalized Predictive Control Of The Gas Compression Systems Based On BP Neural Networks

With the successful application in the process industry, predictive control is one of the most promising control methods in academia and engineering. It has received considerable attention in the last decade. Today,it is universal phenomenon for non-minimum phase nonlinear system, multivariable nonlinear system during many industrial processes. Much effort has been made to extend the predictive control algorithm to nonlinear systems. Over the past several years,neural networks have been widely applied to identification and control of nonlinear system. Theoertical works have been proved that, even with one hidden layer, neural networks can uniformly approximate any continuous function over a compat domain. Artificial neural networks based on predictive control has attracted more and more attention.In this context, Generalized Predictive Control(GPC) for a class of nonlinear system is studied in this paper. The main results are concluded as follows:(1) The method that a class of nonlinear can be substituted by a time-varying linear system is given and proved. So theories of linear systems can be used for nonlinear directly. Least square identification and orthogonal polynomials are used for the time-varying parameters identification.(2) The thesis studies nonlinear system identification based on BP neural networks, and brings in available metric method in order to overcome BP networks'invariable defects of slow convergence, local extreme minimum and bad identification inaccuracy.(3) In the thesis, a sort of neural network based on nonlinear model predictive control strategy is put forward. The process is modeled by neural network, and the manipulated variable is also calculated through neural network based on controller.(4) The proposed nonlinear model prediction control strategy is executed in MPCE-1000 to control gas compress, with satisfying control results obtained.