Division of Rolling Bearing Health Stage Based on Multi-Domain Feature Fusion

The division of rolling bearing failure stages is essential in the maintenance of rotating machinery. In order to classify the bearing failure stage more accurately in different situations, a rolling bearing health stage division method based on multi-domain features is proposed. The clustering algorithm is used to extract the sensitive features of rolling bearing health at each stage, and multiple Features merge into a degenerate curve. The index adaptively divides the entire life of the bearing into multiple health states and is verified by the frequency domain components of each stage. The results show that the method proposed can validly classify the health status of bearings.