Auxiliary Particle Filter-Based Remaining Useful Life Prediction of Rolling Bearing

Rolling bearing's running state has an important influence on the health condition of rotate machinery. This work focuses on the remaining useful life prediction of the rolling bearing. An auxiliary particle filter-based predictor for rolling bearing is presented. The energy spectrum feature of vibration signal is selected as the representation of system degraded states. The wavelet packet decomposition is employed to extract the energy spectrum feature of rolling bearing. The process of prediction includes two stages: the energy spectrum feature is extracted firstly; and then auxiliary particle filter (APF) is trained by using degraded energy spectrum states; APF-based predictor is constructed in the end. Based on an experiment platform of rolling bearing life, a whole life cycle data set of vibration signal is collected for evaluating the performance of the presented APF-based predictor. The classical particle filter (PF) is utilized for comparing with the proposed method. The investigation results indicate that the proposed method can efficiently forecast the remaining useful life of the rolling bearing and has higher accuracy than PF-based predictor.