Abstract:The reliability evaluation of the rolling bearing is crucially important in improving the reliability of mechanical equipment and saving maintenance cost. A novel reliability evaluation method is proposed based on improved logistic regression model (ILRM) to solve the problem that the reliability of rolling bearing is difficult to estimate. First, high relative feature set is constructed by selecting the effective features through extracting the time domain, frequency domain and time-frequency domain features of lifetime bearing. Second, the principal components which can accurately reflect the performance degradation process are obtained by principal component analysis. Then, the principle components are used as the covariates of ILRM. Finally, the covariates are brought into ILRM to obtain the reliability of the rolling bearing. The method can be used to extract the effective characteristics of bearing degradation, and can reflect the state of bearing, and eliminate the influence of random fluctuation of signal on the reliability evaluation. The results verified by intelligent maintenance systems full life test of rolling bearing show that the method can accurately evaluate the reliability of rolling bearings.