With unit tending to large capacity and high parameter, it is difficult to build coordinated control system model of unit using conventional method. Based on the analysis of simplified unit model, a kind of fuzzy identification method based on improved T-S model is proposed. This paper research and develop identification method, input data space was partitioned into some local regions based on entropy clustering and competitive learning algorithm, and then the T-S model was built with weighted recursive least-square algorithm. Parameters setting are particularly analyzed and efficiency is validated by simulation. In this paper, coordinated control system of conventional unit and ultra supercritical unit are established by the proposed method. Simulation results validate the efficiency of the proposed fuzzy identification algorithm applying for modeling coordinated control system of unit.
|