Mechanical Reliability Assessment Methods Research

With the increase of reliability demands for the mechanical products in the aeronautic field and the decrease of reliability test sample size, the existing reliability assessment methods cannot meet engineering requirements gradually. On the basis of summarizing the available results of internal and oversea researches, reliability assessment methods in the case of high reliability product, extreme small sample and small sample life test are studied intensively in this dissertation.A new reliability test method and the corresponding reliability assessment method are proposed in chapter 2. First, the least sample size for high reliability aviatic mechanical products in common experiment is calculated. And then the reason for performing small sample and extreme small sample test is discussed. After pointing out the shortcomings of Bayesian estimation method that it just give qualitative analysis but not quantitative calculation when Bayesian method combined with different experimental results, a reliability test method under normal state, half-tightened state and full-tightened state is presented. Furthermore, a weighted synthesis reliability assessment method considering all the three states is given.In chapter 3, a reliability estimation method in extreme small sample circumstances based on promoted virtual expanding sample technique is proposed. At the beginning, an engineering method called semi-empirical reliability estimation method is introduced. Then, a detailed introduction of the bootstrap method, which has been widely applied to small sample reliability assessing since it was proposed in 1979, is made immediately. At last, a promoted virtual sample expanding method, that the reliability test sample size is expanded from 2-3 to 10, is proposed on the basis of predecessor's research results. And the new method associated with bootstrap method is used to assess the test life lower limit in the case of logarithmic normal distribution. In chapter 4, the theoretical foundation of small sample reliability assessment method based on t-distribution theory is introduced firstly. Then, the formulas of assessing the unilateral test life lower limit and the corresponding mission life reliability, which is more concerned in engineering, are given on the basis of existing research results. Further, considered the considerable uncertainty of current sample variance, a reliability estimation method based on t-distribution theory and empirical logarithmic standard deviation is proposed.