Collaborative Diagnosis and Early Warning Method for Gear Faults
Aiming at the problems of gear fault diagnosis and evolution monitoring,a fast and effective collaborative diagnosis and early warning method were proposed.Firstly,wavelet transform and correlation criterion were used to select the resonance band which reflects the strong fault impact characteristics of gear.Secondly,the envelope demodulation signal was obtained by Hilbert transform,and the noise interference was removed by autocorrelation filter.Then,a series of side spectral lines on the envelope signal spectrum were simplified to a single spectral line by using cepstrum to obtain fault char-acteristics.Finally,Cepstral Amplitude Ratio (CAR) was constructed to represent the fault evolution trend.The analysis re-sults of two sets of public data with fine sampling frequency show that compared with other typical methods and indicators,the spectral peaks corresponding to the fault characteristic frequency obtained by the proposed collaborative diagnosis meth-od are clearer,and the proposed indicators can better reflect the fault evolution trend.
fault diagnosisgearmulti-method collaborationevolution monitoringfine sampling frequency
万方数据
维普
