变点模型下Weibull分布场合步加寿命试验的极大似然估计

Maximum Likelihood Estimation for Step-Stress Accelerated Life Testing with a Weibull Change-Point Model

摘要: 加速寿命试验是一种极为有效而经济的寿命试验方法，其中步加寿命试验又是比恒加寿命试验具有失效更快与参试样品数更少的优点，因此被广泛应用。此文中解决了失效机理改变时步加寿命试验失效数据的时间折算问题，给出了失效机理产生突变的Weibull分布定时和定数场合步进应力加速寿命试验的极大似然估计。

Abstract: Accelerated life testing is an extremely effective and economical method for life testing, among which step-stress accelerated life testing offers the advantages of faster failure occurrence and fewer test samples compared to constant-stress accelerated life testing, thus being widely adopted. This paper addresses the time transformation issue for failure data in step-stress accelerated life testing when the failure mechanism changes, and provides maximum likelihood estimation for step-stress accelerated life testing under Type-I and Type-II censoring, where the failure mechanism undergoes an abrupt change and is modeled by a Weibull distribution.