Probabilistic life prediction of small sample rolling bearing based on ratio distribution
To address the issues caused by limited historical failure data of rolling bearings,discretization of life data,and working environment factors with uncertainty in large complex equipment and special devices,a probabilistic life prediction model based on small sample failure data expanding was proposed.The failure data with small samples was expanded by the Bootstrap method and the dual parameters of the Weibull distribution for the expanded data were estimated by the maximum likelihood estimation(MLE)method.Based on stress-strength interference model,the algorithm for the probabilistic reliability of the bivariate normal ratio distribution was constructed using the Monte Carlo sampling technique,and the fatigue life was predicted by considering the actual working conditions.The fitting results on the XJTU-SY bearing dataset demonstrate that the probabilistic life prediction model can obtain a life prediction curve close to the standard life curve,with a mean square error of 9.71％.
rolling bearingfatigue lifeprobabilistic reliabilityMonte Carlo methodratio distribution
万方数据
维普
