Fault feature extraction of rolling bearings based on complex envelope spectrum

A fault feature extraction method of rolling bearings based on complex envelope spectrum was proposed. Firstly, the vibration signals at the same support of rolling bearings were obtained by orthogonal sampling technology, and they were composed into a complex signal. Secondly, the complex signal was decomposed into the sum of several complex intrinsic mode functions (CIMF) by using the bivariate empirical mode decomposition (BEMD). The envelope signals of real part and virtual part of CIMFs were demodulated by the Hilbert transform. The envelope signals of real part and virtual part were composed of complex envelope signals, and the spectrum was calculated by the complex Fourier transform. According to the characteristics of the complex signal Fourier transform, which can enhance the characteristic amplitude and synthesize the characteristic frequency, the fault characteristic frequency of rolling bearing was extracted according to the complex envelope signal Fourier spectrum. According to the Fourier spectrum of the complex envelope signal, the fault characteristic frequency of rolling bearing was extracted and compared with the full vector spectrum. A composite fault test case proved the feasibility and effectiveness of this method; the composite fault analysis of XJTU-SY rolling bearing data set further proved that the fault features extracted by this method are more comprehensive. © 2021, Editorial Office of Journal of Vibration and Shock. All right reserved.