Health Indicator Similarity Analysis-Based Adaptive Degradation Trend Detection for Bearing Time-to-Failure Prediction

Time-to-failure (TTF) prediction of bearings is vital to the prognostic and health management rotating machines. Owing shifty degradation trends (DTs) bearings, it still difficult obtain accurate TTF results. To solve this problem, paper proposes an online, continuously updated method based on indicator (HI) similarity analysis DT detection. First, multiple features are extracted fused construct principal component HI by using dynamic analysis. Next, exponential models fitted values for future state prediction. By regarding several as a tested segment, detected analyzing segment curve. Finally, predicted extrapolating hit estimated failure threshold. Two case studies public bearing datasets demonstrate superiority proposed approach over state-of-the-art methods.