Fault Diagnosis of Aerospace Rolling Bearings Based on Improved Wavelet-Neural Network

In order to improve the performance of fault diagnosis systems based on a wavelet neural network,according to the frequency domain characteristics of the vibration signals of the ball bearings, a diagnosis system which based on the wavelet packet analysis for picking up character and improved wavelet neural network is proposed ,the conception of wavelet packet analysis and the basic idea of fault diagnosis of wavelet and neural network are also involved.The energy distributing of each frequency segment which is decomposed by wavelet packet is treated as the eigenvector and input the IWNN, and the recognition of the fault models of the ball bearings is completed by using improved wavelet neural network. The result of test and theory shows that circuit fault can be detected and located quickly by using this method and the training speed of wavelet neural network is dramatically accelerated.