Power Transformer Fault Diagnosis Based On The Combination Of Expert System And Artifical Neural Network

The paper, combining the development of power transformer fault diagnostics system, further investigates the fault diagnostic method, the construction of expert system and the knowledge acquisition.According to the features of transformers, a new diagnostic method, integrated expert system with artificial neural network, is presented in this paper. In the first place, the expert system diagnoses the system in high-level one (the thick grain size) and goes on expert reasoning, for the sake of reducing the failure probability. And then under the primary system harmonizing uniformly, this method transfers the corresponding neural network to recognize subsystem to identify and calculate the low-level one (the detailed grain size) .The architecture of fault diagnostics expert system of power transformer. Knowledge acquisition in the expert system is studied in depth, and on the foundation of taking full advantage of the existing three means of knowledge acquisition, a new technique for knowledge acquisition based on the data mining is introduced, which well resolves the problem of the acquisition of system knowledge.On account of the information processing traits of associative memory and strong capability to recognize and classify the input samples, the neural network can solve some problem effectively, which can't be resolved by the traditional patternrecognition methods. Through the practical application of artificial neural network to diagnose fault of equipments, it is proved that the method for the transformer fault diagnosis presented in this paper is available. The paper does some exploration and attempt in this aspect.