Research on neural network fault diagnosis method based on DGA for large transformer

By researching the fundamental of DGA (Dissolved Gas Analysis) and fault characteristic of large transformer, analyzing the disadvantage of traditional method of large transformer fault diagnosis and utilizing the ability of function approximation and classification of Radial Basis Function neural network, the paper tries to establish a large transformer fault diagnosis model using Radial Basis Function neural network based on DGA theory In this model. Compared with the traditional method, three-ratio method which lacks of fault coding and no fault types with coding existent, the method proposed here overcome the disadvantage of the Three-ratio method and the related experiment results are more corrective, so it has comparatively good practical value in the future.