Abstract:  Aiming at the lack of quantitative evaluation criteria for performance degradation of rotating equipment, a real-time health evaluation model of rolling bearing based on relative features was constructed. Firstly, based on the monotonicity principle, the relative mean absolute value and relative root mean square value were selected as the health index, and the fuzzy Cmeans clustering method was used to build a datadriven realtime health assessment model of rolling bearing, and a knowledge base of health assessment criteria. Then, using the nearest neighbor principle and the designed logic judgment correction algorithm, the health status of the bearing to be tested was identified in real time. Finally, taking the second group of bearing experimental data of University of Cincinnati intelligent maintenance system (IMS) center as the model training data, the vibration data of P3409A centrifugal pump bearing “running to failure” of a petrochemical company in China was selected as the test data, to verify the health assessment model. The verification results show that the model can effectively represent the real-time performance degradation state of rotating equipment, and realize online realtime evaluation. It only needs the normal data of the equipment to be evaluated, does not rely on the prior knowledge of external experts, and has good generalization.

Key words:  rolling bearings; health condition assessment; relative characteristics; fuzzy C-mean clustering; evaluation criteria