A Rolling Bearing Performance Degradation Evaluation Method Based on Statistical Correlation Superposition

Accurate and reliable methods for evaluating the health condition of rotating machinery can improve equipment efficiency and reduce safety hazards. This article proposes a correlation superposition method for evaluating rolling bearing performance degradation. The method is based on frequency domain (FD) probability distribution estimation and Pearson’s correlation coefficient. First, a new health indicator describing the bearing degradation process is obtained by estimating the correlation of the FD probability distribution. Second, an evaluation indicator that considers both monotonicity and trendability provides a more precise evaluation of the regression quality. Finally, the effectiveness of the proposed method is verified by using the public datasets of Xi’an Jiaotong University (XJTU) and a performance degradation evaluation model. The results confirm the performance of our method, which is improved by 18% and 10% compared to the current common ones.