T-S fuzzy modeling based on compatible relation and its application in power plant

Abstract

The characteristics of industrial process control systems are nonlinearity, time variability and uncertainties during variable operations. For complex nonlinear systems, a T-S fuzzy system modeling based on compatible relation is presented in the paper. The cluster analysis algorithm based on compatible relation can calculate the best cluster subset and number of the subset at the same time. The algorithm can take full advantage of experimental data, and greatly improve the recognition accuracy. The parameters of the proposed algorithm do not need to be adjusted, which reduces computation and shorten identification time. The consequent parameters of rules are determined by used of the recursive least squares method. The proposed modeling approach can be illustrated based on practical application. The models of coordinated control system of 160MW fuel unit and 1000MW ultra-supercritical power unit are established by using the proposed method. Simulation results show that the proposed method can achieve higher recognition accuracy, and it is very effective to coordinated control system in power plant.