Statistical Analysis On Accelerated Life Tests For Weibull Distribution

With continuous improvement of the reliability of the product,the life tests has been widely used and developed rapidly.Accelerated life test(ALT)has always been the main method in the research of product reliability.Type I and Type II censoring have been widely applied as the criterion to terminate the life tests.The step stress and sequential stress ALT are more efficient than the constant stress ALT in that the experiment duration can be reduced sharply.Assume that the life time of the product follows a Weibull distribution,we studied the optimal design of step stress ALT under Type I censoring with random removals,and the parameter estimation of step stress ALT and sequential stress ALT under Type ? censoring with random removals.Firstly,we introduce the concepts of reliability,accelerated life test and Weibull distribution as a preparation for the whole thesis.Then the optimal design of step stress ALT with two stress levels is obtained by minimizing the asymptotic variance of the estimator of the mean lifetime.Finally,the maximum likelihood estimation of the lifetime parameters is investigated for step stress ALT and sequential stress ALT under Type II censoring with random removals.The results give some guidance in designing an ALT.