Rolling bearing early weak fault detection and performance degradation assessment based on VMD and SVDD

In order to detect the early weak faults and monitor the fault condition of rolling bearings, a performance degradation evaluation model based on variational mode decomposition (VMD) and support vector data description (SVDD) was proposed. The vibration signal was decomposed by VMD, and the intrinsic mode component which was sensitive to the performance degradation was selected to extract its singular value.Combining the singular value with the time domain feature and the complexity feature vector matrix, a comprehensive feature index of the rolling bearing was coustituted. A SVDD evaluation model was constructed taking the comprehensive characteristics of the normal state of the bearing, as the training sample and the rolling bearing's whole life test data were used to verify the degradation assessment model. The experimental results show that the evaluation model can accurately detect the early stage weak failure of rolling bearings. At the same time, the performance degradation assessment of rolling bearings can be effectively achieved. The evaluation effect is superior to that of the fuzzy C-means clustering (FCM) method. © 2019, Editorial Office of Journal of Vibration and Shock. All right reserved.