Abstract:

Aiming at the problem of product reliability evaluation in condition of zero-failure data, a method combining the weight-least-square method with expectedBayesian (E-Bayes) reliability evaluation is put forward. According to the engineering experience, the prior distribution of product failure probability is constructed. Then the E-Bayes estimation of product failure probability is given by using the theory and method of Bayesian. In condition of Weibull distribution, the weight-east-square method is used to seek out the parameter of Weibull distribution, and obtain the reliability indexes’ point and interval estimation. At last the robustness of the model and the superiority of the method are discussed by an example which shows the value of the method in engineering application.