Uncertainties establishment method of friction moment on rolling bearing based on Bayesian theory and robust theory

A method on basis of Bayesian theory and robust theory was proposed to solve uncertainties of time series data on unknown distribution.The method combining with median and Huber M estimate was used to analyze robust property of time series data and establish prior distribution on Bayesian theory.Uncertainties were established by inferring from prior distribution and distribution function on time series data.At the significance level 0-0.1 range, for studies of friction torque of rolling bearing, the method determined boundaries value and confidence level of friction torque, and established prior distribution on Bayesian theory.Compared with classic statistics, the method distinctly increased evaluation accuracy 0-50% under the same confidence level.Application significance of the method lied in taking robust data to establish prior distribution, and provide a establishment method of prior distribution on Bayesian theory;adopting the median and Huber M estimate method to determine confidence level and boundary value could reduce errors of subjective factors, and provide a theory to establish uncertainties on unknown distribution. © 2022, Editorial Department of Journal of Aerospace Power. All right reserved.