关键词: 故障诊断, BP神经网络, 集成神经网络

Abstract: This paper analyzes the characteristics and shortages of traditional fault diagnosis methods, upon these analysis chooses BP neural network to perform fault diagnosis, discusses BP neural network model building method in detail, according to the layer structure and fault characteristics of equipment, applies neural network integrated technology to effectively overcome the shortage of single network fault diagnosis and increase its efficiency and accuracy dramatically.

Key words: fault diagnosis, BP neural network, integrated neural network