Application Of EEMD-based Resonance Demodulation Technology In EMU Bearing Fault Diagnosis

With the increasing mileage of the railway,the amount of the EMU in our country is rising rapidly.Security has become the most important part in the operation process.Accurately finding and identifying incipient fault feature of axle box rolling bearing of EMU can help to decide the maintenance scheme of EMU.This paper,basing on the research of the vibration mechanism of rolling bearing faults,has introduced origin cause,basic form,vibration model and fault characteristic frequency of rolling bearing,briefly.Many methods of rolling bearing fault diagnosis have been applied to rolling bearing of locomotive and EMU,including diagnosis based parameters of time-frequency field,diagnosis based Hilbert-Huang Transform,diagnosis based Resonance Demodulation,diagnosis based wavelet package,etc.By applying ensemble empirical mode decomposition(EEMD)and Resonance Demodulation,a new solution of fault diagnosis is proposed.Firstly,the vibration signals are decomposed into mode components by EEMD so as to highlight the high frequency resonance components,and the kurtosis criterion method and correlation analysis are used to get adaptively the components of IMF and reconstruction.Secondly,wavelet packet energy spectrum and spectral kurtosis method are used to calculate the parameters of the band-pass filter,which can extract high frequency resonance components.Finally,the result of fault diagnosis was obtained by Hilbert transform and adopting envelope demodulation and the solution has been tested by analyzing actual fault signal from Bearing Data Center of Case Western Reserve University.Through analysis the real vibration signal of axle box bearing of EMU,the background noise analysising work was finished and component were explicated.Combining the result of fault characteristic analysis of rolling bearing,establishes the basis of building the emulation model.To verify the validity of the proposed strategy,using axle box bearing of CRH380BL EMU as an example respectively builds up the emulation model of outer ring,inner ring,roller and cage of rolling bearing complex background noise,which is accomplished using MATLAB,verified the effectiveness of the algorithm.