关键词: 航空发动机, 静电监测, 传感器, 故障诊断, Fisher准则, 自组织映射, 神经网络

Abstract: To solve the problem of typical gas path mechanical fault diagnosis,a simulated experiment of gas path fault by electrostatic monitoring was carried out.The electrostatic signals of four typical gas path fault modes were analyzed based on the classical characteristic parameters.On this basis,several new characteristic parameters and the logic theory of fault identification were proposed,and the Fisher criterion method was used to verify the validity of new characteristic parameters.An engineering application method of gas path fault diagnosis based on self-organizing map neural network was proposed,and the fault diagnosis algorithm was verified by an example.The results showed that the method of gas path fault diagnosis based on electrostatic signal and self-organizing map neural network could identify different failure modes well and had good visual expression effect,which had provided a good engineering method for gas path fault diagnosis.

Key words: aero-engine, electrostatic monitoring, sensor, fault diagnosis, Fisher criterion, self-organizing map, neural network