Research On The Robust Identification Of Fuzzy Model And Its Application

Due to the complexity and uncertainty of the actual industrial process, as well as the strong-coupling condition, it is difficult to establish the precise mathematical model. Even to build its mathematical model, it is often too complex, making it difficult to achieve the traditional ideal effect of control. Zadeh proposed by the fuzzy set theory which can provide an effective method for the complex and sick system. The input and output data of the system is considered as the basis, then determining the fuzzy rules and estabishing the fuzzy model. This article focuses on the fuzzy identification of the non-linear system to discuss.Firstly, in the fuzzy modeling of the non-linear system, kinds of uncertainties, such as the noise and coupling, made the description of the relationship matrix which exists serious linear correlation among the inter-out. How to find the rules, simplify rules and reduce the fuzzy model of the input space is the key to the problem of the fuzzy modeling. Based on the fuzzy objective function to choose the model structure, a new method used to the non-linear fuzzy modeling and parameter estimation is proposed in the article and proved the method of convergence by theorem.Secondly, H_∞identification method in time domain applied to non-linear system offuzzy modeling, which makes the largest energy gain from the interference to the estimation error of to be smallest. And the LMI identify the unknown energy factorγ. It overcomes uncertainty that the value ofγusually was given by the practical engineering and the complexity that the value ofγwas given by the iterative algorithm.Finally, the fuzzy identification methods are used for the power station simulator in the turbo-generator seal oil and Simulation results show that the fuzzy modeling methods are effective ways in the establishment of the non-linear system model and have good accuracy. Compared to other modeling methods, the model has to be explanatory.