254 / 2021-04-15 11:19:23

Application of integrated spectral coherence in railway axle-box bearing fault diagnosis

Vibration,Fault diagnosis,Weighted combined envelope spectrum,Spectral coherence,Railway bearing

Fault diagnosis of axle-box bearings is essential to ensure the safe operation of railway trains. Enhanced envelope spectrum generated by integrating the spectral coherence over the full spectral frequency band is an effective tool for rolling bearing fault diagnosis, but it cannot effectively reveal the bearing fault features under strong interference noise. Improved envelope spectrum obtained by the integration of the spectral coherence over an informative spectral frequency band can reduce the interference noise and improve the fault detection capability, but it can only extract the fault information in one resonance frequency band and cannot effectively integrate the fault components distributed in multiple resonance frequency bands. To solve these problems, the weighted combined envelope spectrum from the spectral coherence normalized with feature energy is proposed in this paper. First, normalized feature energy is designed to quantify the fault information in each envelope spectrum slice of the spectral coherence. Second, a threshold is introduced into the normalized feature energy to construct a weighting function. Finally, the spectral coherence normalized with the weighting function is used to generate the weighted combined envelope spectrum for bearing fault diagnosis. The proposed envelope spectrum can integrate the fault information distributed in multiple resonance frequency bands and eliminate the fault-unrelated components. The results of experimental data validate the effectiveness of the proposed method in railway axle-box bearing fault diagnosis, and the comparison with the enhanced envelope spectrum and improved envelope spectrum demonstrate the superiority of the proposed method.