A fault diagnosis method of planetary gearbox under variable speed condition using Vold-Kalman filter and Laplacian score

Planetary gearboxes operation under non-stationary working condition exhibit complex time-varying modulations and spectral structures, resulting in difficulty in the fault diagnosis of planetary gearbox. In order to effectively remove the influence of rotating speed and extract the fault characteristics, this paper aims to develop a fault diagnosis scheme based on Void-Kalman filter (VKF) and Laplacian score (LS). In this method, VKF is firstly adopted to remove the fault-unrelated components and give a clear representation of the fault symptoms. Second, the time-domain features and frequency-domain features are used to extract fault features. Third, LS approach is introduced to refine the fault features by sorting the scale factors. Lastly, the selected features are fed into the LSSVM to automatically complete the fault pattern identifications. The combined method is experimentally demonstrated to be able to recognize the different sun gear fault types of planetary gearboxes.