The harmonic modulation bispectrum: a modulated vibration signal analysis method for compound fault diagnosis of gearboxes

In the compound fault diagnosis of gearboxes based on vibration signal, the fault feature information often exists in the complex modulation components of the signal. Using demodulation technology to extract the feature information in the signal is the basis for fault diagnosis. In this article, a modulation information detection technology called harmonic modulation bispectrum (HMB) is proposed after considering the multiple coupling relationship between the frequencies in the modulated signal. HMB can extract effective information from complex modulation signal and reduce noise interference in signal. At the same time, a diagnosis strategy suitable for the gearbox compound fault is developed based on HMB. The third-order Renyi entropy is used to realize the adaptive setting of the parameters in HMB, and the marginal spectrum trends of the HMB are calculated by combining the idea of scale space. In addition, combined with the harmonic spectral kurtosis, the effective information of each frequency band in HMBgram is evaluated quantitatively. The frequency band with the most abundant effective information is located, from which the fault feature information is extracted to realize the fault diagnosis finally. The diagnostic effectiveness of this method has been confirmed through simulation and experimental signals, and the superiority of this method has been demonstrated by comparing with other methods.