Improved T-S fuzzy model identification approach and its application in power plants
ABSTRACT Systems in power plants often contain nonlinearity, complexity and randomicity. It is difficult to build their model by traditional methods. An improved fuzzy identification approach based on Takagi-Sugeno (T-S)model is proposed to solve the problem. In this paper, T-S model is firstly modified to make its identification easier. Following that, input vector is determined by heuristic knowledge and exponential form membership function is used to avoid conclusion can not be calculated. Then, entropy cluster algorithm is analyzed and improved to automatically determine the number of subspace and initial subspace centers. Finally, competitive learning algorithm and weighted recursive least-square algorithm are used to estimate the parameters of T-S model. Simulation results show that the proposed approach can describe nonlinear system in power plants accurately, and the relevant algorithm is simple and fast.