Aero-engine Fault Diagnosis Based on Kernel Principal Component Analysis and Wavelet Neural Network

Abstract

As a complex high-speed mechanical system, the aero-engine is a typical fault-prone system in long-term high-altitude environments such as high temperature, high pressure, strong corrosion and high-density capacity release. It is extremely difficult to accurately diagnose it. To this end, this paper proposes an aero-engine fault diagnosis method based on kernel principal component analysis and wavelet neural network. The nuclear principal component analysis method is used to process the aero-engine original parameter data, extract its principal component features, reduce the parameter dimension, and construct the health state and fault state data sample set with the extracted principal component feature data. It is divided into training sample set and test sample set. The wavelet neural network fault diagnosis model is built by using the training feature data sample set. The diagnostic neural network fault diagnosis model is diagnosed and analyzed by using the test feature data sample set. At the same time, BP neural network is used to diagnose the same feature data sample set. In addition, the wavelet neural network fault diagnosis model is used to study the fault diagnosis technology of the original data. The research results show that the diagnosis results of the aero-engine fault diagnosis model based on kernel principal component analysis and wavelet neural network are obviously better than the diagnostic results of other methods used in this paper, and have good practical application value.