Study On Bearing Reliability Evaluation Method Using On No-failure Data

With the continuous improvement of bearing production and manufacturing technology, the service life of the bearing has been greatly extended. In bearing reliability test, due to the characteristics of high reliability and long life, it is difficult to obtain living data of bearing, only zero failure data, in this case, the traditional reliability assessment method based on life observation data for no failure data bearing reliability assessment has limitations. The combing the traditional product reliability statistical inference method based on, for zero failure data for this special case, given the bearing reliability evaluation of classical confidence limits and Bayes confidence limits and the specific contents are as follows:First, the paper detailed introduced the home and abroad research situation of bearing reliability and zero failure data analysis theory and the method, expounds the zero failure data in the analysis of the reliability of bearing the significance.Secondly, this paper introduces the theory and method of the reliability analysis, and makes a detailed introduction to the traditional bearing reliability analysis process. The analysis shows that the bearing obeys the Weibull distribution.The disadvantage of this method must obtain the bearing failure data, otherwise, it is impossible to analyze the reliability in method.Again, the analysis of the no failure data of bearing is studied, By using this method, the reliability of Weibull distribution is estimated, and the results obtained are quite accurate when the unknown parameters are known. the method was validated by Monte Carlo simulation finally, through an application example shows the validity of the method, and can get the reliability of bearings in a given time. But this method also has some disadvantages, it can not get the bearing life distribution curve.Finally, for the small sample problem of zero failure data. This paper introduces the prior distribution is determined according to the prior information, discusses the zero failure data under the condition of bearing Bayesian reliability analysis. Through the comparative study of classical method and Bayes result, shows that Bayesian method can be reliability analysis more accurate, and can be based on zero failure data has been bearing service life distribution curve.