Reliability quantitative analysis for rotor vibration based on WCFSE

In this paper, in order to improve the precision of rotor vibration fault diagnosis with nonlinear and nonstationary signals, the wavelet correlation feature scale entropy (WCFSE) method is proposed to quantitatively analyze and diagnose the rotor vibration fault feature by fusing wavelet correlation filter method and information entropy theory. Firstly, the formation features of weak faults are picked up from rotor vibration fault signals by wavelet transform correlation filter approach to get high signal to noise (S/R, SNR) scales wavelet coefficients. Secondly, the WCFSE values of vibration signals are extracted, which are able to not only describe the degraded process of vibration faults but also quantitatively diagnose the faults. Finally, through the quantitative analysis and diagnosis of rotor vibration faults based on rotor vibration simulation experiments, the results demonstrate that the proposed WCFSE method is effective and feasible for the evaluation of rotor vibration faults and a promising way for the fault diagnosis with nonlinear and abnormal signals.