Fuzzy Model Identification Method Based On The Dga Transformer Running State

Power transformer is one of the most important electrical equipments in the electric system. Its operating state attaches importance to system's operating safety directly. Dissolved gases analysis (DGA) is an important method to diagnose the internal fault of transformer; it has been proved in practice that DGA technology is very effective to diagnose the transformer fault, transformer fault is the result of the long-term accumulation and the cooperation of the transformer itself and its application environment, so the symptoms of transformer fault are variety, and the relation of the fault symptom and fault mechanisms is complicate, then its very difficult to build up a common transformer fault diagnosis method. This paper will use dissolved gases in oil as the character for fault diagnosis; adopt fuzzy model and fuzzy reasoning technology; and study the running condition recognition of transformer.The composition of current large number dissolved gas analysis is the real samples, in order to enhance the application ability of the fuzzy model in the running condition recognition of transformer, first the problem of obtaining the fuzzy model for real samples need to be resolved, based on the obtaining process and the main obtain technology of the variety of fuzzy reasoning system when the systems analysis in the face of real samples data, the TSK fuzzy model based on GA-BP hybrid intelligence learning strategy is proposed. Some problems related to a species coding means for the model structure, evolution and fitness evaluation strategy are discussed. The error back propagation algorithm (BP) for training the antecedent and consequent parameters during the process of evolution is inferred. The validity of the method has been demonstrated by the example of function approximation and the problem of typical class, and the strategy has been applied in the transformer condition classification.This paper analysis various diagnostic methods based on the DGA data of transformer, proposes a method of predicting the DGA data of transformer and a strategy of recognizing the running condition of transformer based on TSK fuzzy, and builds up a model of transformer condition recognition. The model behaves well in accuracy and generalization. The transformer condition features and their previous handling way are analyzed during the building up of the running condition model of transformer. The validity of the DGA single-step predictive has been demonstrated by the DGA data predictive model and the DGA single-step predictive data as the capability of the transformer preventive failure diagnosing.