Reliability Prediction of Power Converters at Accelerated High Temperature Using Weibull Proportional Hazard Model

Power converters are extensively utilized in a variety of safety and non-safety critical applications, including aerospace, automotive, and astronomical. The failure or malfunction of converter in such applications severely impacts the mission of such expeditions. Thus, the converter reliability prediction or assessment becomes a mandatory requirement. The accelerated life testing (ALT) can be applied to predict converter reliability in a faster manner by taking the temperature as the reliability and life influencing factor. The paper proposes Cox-proportional hazard model (Cox-PHM), an alternative to the ubiquitous Arrhenius model, by taking thermal stress as a covariate influencing the time varying hazard rate. In this paper, the data of ISL8225M series DC/DC converters at accelerated temperature i.e., 100°C, 120°C,140°C is taken from the literature to model the situation. The parameters of proposed Weibull-PHM model are estimated using STATA/SE 14.2 and then reliability, hazard rate and cumulative hazard rate are obtained from the proposed model. The results obtained from proposed Weibull-PHM model are compared with the Weibull-Arrhenius model to show the supremacy of proposed model.