Research In The Nonlinear System Fuzzy Identification

In the thermal process, the dynamic behavior of plants shows a characteristic of great delay, big inertia, time variance and non-linearity, which make the modeling very difficult, and so the whole optimal control for thermal processes is impossible. According to the objective of building accurate nonlinear model for thermal process, an adaptive neural-fuzzy inference system modeling method based on T-S model is proposed.And then this paper proposes a kind of method based on improved T-S model. The heuristic information and the multiplex nonlinear optimization are combined to configure the structure of fuzzy model, and the input data space is partitioned into some local regions based on entropy clusting and competitive learning algorithm, then the T-S model for thermal process is built with weighted recursive least-square algorithm.According to the mature algorithm, a kind of fuzzy identification software is developed by the Matcom and VC++. The results show that it has good efficiency when the software is applied to the fuzzy modeling.