Application of bayesian method in assessing reliability parameters of mechanical product

When assessing the reliability parameters, the traditional method is to deal with a large quantity of data obtained through test and to get the result with high confidence. But small sample size and the lack of test data of mechanical items make the traditional method have weakness. If the Bayesian method can be used properly, accurate assessment will be derived even if it is lack of data. What is discussed in this paper are: 1) determination of prior distribution, 2) reliability assessment based on pass/fail data with Beta distribution, Gamma distribution and cut-tail normal distribution as prior distribution, and 3) reliability assessment based on time-to-failure with Weibull distribution and normal distribution as prior distribution.