Non-stationary vibration signal analysis and fault diagnosis method of aircraft power plant using wavelet network

To improve the limitation of applying traditional fault diagnosis method to the diagnosis of multi-concurrent vibrant faults for aeroengine in aircraft, a novel approach combining the wavelet transform with self-organizing learning array system is proposed. The effective eigenvectors are acquired by binary discrete orthonormal wavelet transform based on multi-resolution analysis. These feature vectors then are applied to the proposed system for training and testing. The synthesized method of recursive orthogonal least squares algorithm is used to fulfill the combined network structure and parameter initialization. By means of choosing enough practical samples to verify the proposed network performance and the information representing the faults is inputted into the trained network, the output result the type of fault can be determined. The simulation results and actual applications show that the proposed method can effectively diagnose and analyze the fault patterns of aeroengine.