关键词: 可拓神经网络, 变压器, 故障诊断, 可拓理论

Abstract: Extension Neural Network（ENN） is a new kind of artificial neural network，which is a combination of extension theory and neural network.At present，there are many successful applications in some fields，such as pattern recognition，fault diagnosis，classification，clustering analysis and so on.Based on the characteristics of transformer fault diagnosis，a new power transformer fault diagnosis method based on ENN is discussed in this paper.The outline of ENN is introduced，the fault diagnosis model and algorithm based on ENN are designed，and its application to the diagnosis and recognition of power transformer is presented.Simulation experiment is conducted and the comparative results indicate that the proposed method is not only simpler to application，but also has lower training error and less convergence time.This method has certain value of application and expansion.

Key words: Extension Neural Network（ENN）, transformer, fault diagnosis, extension theory