Fault Diagnosis of Railway Rolling Bearing based on Wavelet Packet and Elman Neural Network

G. Cai, L. Jia, J. Yang, and D. Yao (PRC)


Railway rolling bearing, FIR, Wavelet packet, Neural network, and Fault diagnosis


A new railway rolling bearings faults diagnose of wavelet packet transform and neural network is presented. This method uses integrate patterns of denosing, and energy eigenvector extraction. This procedure can be implemented in a multi-layer wavelet packet decompose. Firstly, rolling bearing signal is denoised based on FIR. Then, three-layer wavelet packet is adopted to decompose the signal and reconstruct energy eigenvector. Last, fault samples of wavelet packet energy eigenvectors are used as neural network input parameters to realize intelligent fault diagnosis. Examples with real data demonstrate the excellent performance of this diagnose scheme as compared to existing techniques.

Important Links:

Go Back