SVDD-based Evaluation of Rolling Bearing Performance Degradation

Abstract

For the problem of difficult to monitor the operating condition of bearings A model based on SVDD rolling bearing performance degradation evaluation is proposed. Firstly, signal preprocessing is performed to obtain each subcomponent by ICEEMDAN decomposition, and screening is performed by correlation coefficient cliqueness criterion. In order to characterize the bearing characteristics accurately, the time domain index and complex domain index were combined, and the characteristics of each component were obtained.And the SVDD evaluation model is completed with the comprehensive characteristic indexes under normal condition as training samples, and the evaluation model is verified by using the rolling bearing whole life test data.The results show that the model can reflect the law of early bearing failure and its performance decline.