Fault feature extraction of rolling bearing based on VMD-SVD and time-frequency analysis

In this paper, the bearing fault signal is decomposed and denoised by using variational mode (VMD) and singular value decomposition (SVD), and a rolling bearing fault feature extraction method based on vmd-svd and time-frequency analysis is proposed. VMD is used to decompose the bearing fault signal into single component modal components, and the spectral kurtosis criterion is used to determine the optimal frequency band of signal fault characteristics; At the same time, SVD is used to reconstruct the signal, improve the signal-to-noise ratio, eliminate the high-frequency interference component, and highlight the weak fault impact characteristics; Then the processed signal can be effectively extracted by Wigner analysis of Choi Williams filter. The results of simulation and experimental data show that the frequency component of fault signal can be accurately reflected by decomposing fault signal with VMD-SVD model; at the same time, VMD-SVD decomposition method can effectively suppress mode aliasing, and the signal can effectively extract weak fault features after Wigner bispectrum analysis, so as to improve the accuracy of fault detection.