Research On The RBF Fuzzy-Neural Network And Application On The Unit Coordination Control System

Coordination Control System is an important topic on power plant automatic control. It is a highly complicated multiple-variable control system. With the development of the modern control theories, more and more intelligent control ideas are applied to the realization of the multiple-variable coordination control system. The article mainly concentrates on the following points:(1) With the case of homemade 300MW unit power set, the principle, mathematic-model and dynamic characteristic of the unit power set have been analyzed.(2) A fuzzy-neural network multiple-variable control algorithm has been brought forward. It separately controls the two loops at the side of the boiler and at the side of the turbine, eliminates or reduces the coupling of each loop by the single-layer neural network. At the same time, it is adjusted on-line that the central value and width of the network membership function according to system error.(3) Its anti-interference characteristic and robustness have been validated by a lot of simulation experiments and comparing it with traditional PID control algorithm. As a result of it, the control algorithm in the article enhances the dynamic quality of large-delay object, and excels PID control in the anti-interference characteristic and robustness.