Censored data oriented estimation of the change point in the infant period of a bathtub curve by maximum likelihood estimation

The change point evaluation and monitoring of failure rate is an inevitable task for product reliability assurance. Since the fact that the evolution mechanism of product infant failure is fuzzy due to its complexity, it is infeasible to determine the accurate burn-in time in advance when applying burn-in testing for electronic product, the bottleneck lies in the accurate identification of the change point for infant failure. Accordingly, in this paper, the connotation of the change point of infant failure in product early lifetime is discussed, and the processing method of the censored life testing data from burn-in testing is presented. Then, based on the principle of maximum likelihood estimation(MLE) for censored life testing data, the convergent change point identification technique for Weibull distribution is proposed to be more applicative for the analysis of censored characteristics. Finally, the validity of the proposed method is verified by a change point identification example for a control board reliability analysis.