Novel data-driven performance degradation state monitoring of rolling bearing

A performance degradation state monitoring approach is introduced for rolling bearings (RB). Firstly, the time-domain features (TFs), frequency-domain features (FFs) and time-frequency-domain features (TFFs) of vibration signals (VS) are extracted to describe the performance degradation trend. Then, the performance degradation features are selected via the Spearman rank correlation coefficient (SRCC). Then, these selected features are fused to obtain health index (HI) using growing self-organizing map (GSOM), Finally, the identification of the degradation state is realized by the analysis of the monotonicity, recession and robustness of the HI and experimental results demonstrate the superiority of proposed method compared with other famous method.