Gearbox fault diagnosis method based on MVMD-MOMEDA

Abstract

Aiming at the problem that the early weak fault diagnosis of gearbox vibration signal is difficult due to the influence of complex transmission path and strong background noise, a gearbox fault diagnosis method based on multivariate variational mode decomposition (MVMD) and multipoint optimal minimum entropy deconvolution adjusted (MOMEDA) was proposed. Firstly, MVMD was used to decompose the multi-channel vibration signals after fusion， and a series of IMF components representing the local characteristics of the signals were obtained. Secondly, the kurtosis value (Ku) was introduced to select the best mode for signal reconstruction, and the IMF with high noise content was eliminated. Finally, the fault frequency was identified by MOMEDA feature extraction to the reconstructed signal to achieve the purpose of fault diagnosis. The results show that the proposed fault diagnosis method can effectively eliminate the interference of the noise components, identify the fault impact components in the signal and its frequency doubling, and then determine the fault type. The MVMD-MOMEDA method solves the problems such as the inability to deal with multi-source signals in a single channel and the difficulty in extracting early weak fault features, which provides reference for fault diagnosis and multi-source signal processing.

Keywords

• data processing; gearbox; multivariate variational mode decomposition; multipoint optimal minimum entropy deconvolution adjusted; feature extraction; fault diagnosis