Analysis of experiment result and fault diagnosis for aeroengine rotating shaft

To increase the accuracy of applying traditional fault diagnosis method to aeroengine vibrant faults, a novel approach based on wavelet neural network is proposed. The effective signal features are acquired by wavelet transform with multi-resolution analysis. These feature vectors then are applied to the neural network for training and testing. The synthesized method of recursive orthogonal least squares algorithm is used to fulfill the network structure and parameter initialization. By means of choosing enough practical samples to verify the proposed network performance, the information representing the faults is inputted into the trained network. According to the output result the fault pattern can be determined. The simulation results and actual applications show that the method can effectively diagnose and analyze the vibrant fault patterns of aeroengine and the diagnosis result is correct.