Rolling bearing weak fault diagnosis utilizing successive variational mode decomposition with sparsity index reconstructing strategy

Abstract

By introducing much stricter criteria into VMD, the successive VMD (SVMD) originating from VMD is proposed, which is more suitable for extracting fault feature of faulty rolling bearing than VMD, especially the inner race failure or cage failure, because the frequency spectrum structure of the above two kinds of failures take on the compact frequency spectrum distribution characteristic of the center frequency with sidebands, which is compliant with the criteria of SVMD. In addition, the optimal number of modes does not need to be determined in advance for SVMD, and SVMD could decompose and extract the optimal modes adaptively. So SVMD is used in the paper for modes extraction of the faulty signal of rolling bearing. Unfortunately, the impulse characteristic components might be distributed in each mode more or less in the decomposition processes of SVMD, and the fault features is impossible to be extracted effectively only based on a single mode. Accordingly, a mode regrouping strategy based on sparsity index is proposed to regroup the modes containing useful fault information. At last, the regrouped signal is analyzed by using envelope spectral and satisfactory fault features are extracted. Through the verification of simulation, SVMD has better comprehensive performance compared with other time-frequency analysis methods such as VMD, VME and EEDD. Effectiveness of the proposed method is verified by one experimental and one engineering vibration data respectively. Besides, the advantage of the proposed method compared with the advanced analysis method, that is, Mkurtogram is also verified through comparison.