摘要
通过利用删失双参数指数分布考虑性能退化数据的自然次序,提出了一种新的退化可靠性建模方法。极大似然估计(MIE)被用来提供一种估计模型参数的一步法,得到了具有时间独立位置参数的删失双参数指数分布的似然函数的一个封闭形式表达式。提出了一种用于具有时间相关位置参数的删失双参数指数分布的半数字方法。通过对仿真退化数据的研究表明,提出的方法提供了合理的估计,甚至是小子样情况。
A new degradation reliability modeling approach is presented by considering the natural ordering of performance degradation data us- ing a truncated exponential distribution.Maximum likelihood estimation is used to provide a one-step method to estimate the model's parame- ters.A closed form expression of the likelihood function is derived for a two-parameter truncated exponential distribution with time-independent location parameter.A semi-numerical method is presented for the truncated exponential distribution with a time-dependent location parameter. Numerical studies of generated data show that the proposed approach provides reasonable estimates even for small sample sizes.
出处
《控制工程》
CSCD
2007年第S3期119-121,共3页
Control Engineering of China
关键词
极大似然估计
双参数指数分布
退化数据
可靠性建模
maximum likelihood estimation
two-parameter exponential distribution
degradation data
reliability modeling
