Research of intelligent control based on Support Vector Machines for power plant boiler

Abstract

The power plant boiler combustion system has the features of strong coupling multivariable, nonlinear and large delay, in order to control it effectively, based on the idea of inverse control of nonlinear systems, we apply the inverse model recognition system of Support Vector Machines and support the combination of inverse control of vector machine and fuzzy control, we established a Intelligent Control system for the Boiler. In this composite control, the fuzzy controller is used to achieve the rapid response in early control, and the SVM controller plays a major role during the later period. The inverse recognizer transfers the parameters adjusted on line to the inverse controller, for eliminate the steady-state error of the controlled system. Because of increase the feedback link in this combination of control system, so the direct inverse control has more excellent performance and robustness, and it resolved effectively the problem of inverse control model accuracy requirements. In the application of boiler steam pressure control system it is proved that the strategy of intelligent composite control is effective and practical.