摘要
对于长寿命、高可靠性产品,失效样本很少或根本没有失效样本,如何准确地评价它们的时变可靠性,一直是一个很难解决的问题。本文提出一种基于退化数据,采用残差自回归模型结合可靠性理论的分析方法。首先采用残差自回归模型对样本的退化数据进行处理,然后结合可靠性理论,针对产品退化量服从正态分布和Weibull分布的情况,利用时间序列的特点估计分布参数,进行产品退化过程的时变可靠性分析。最后,通过与大样本的可靠性评估方法进行对比,证明了该方法的有效性。
For long-life, high reliability products, failure samples have few or no failure samples, It has always been a difficult problem for evaluating accurately their time-varying reliability. This paper presents an approach to using residual autoregressive model combined with reliability theory based on degradation data. Firstly, the residual autoregressive model is used to process the degradation data of the samples, and then time series is used to estimate distribution parameter of degradation data submitted to normal distribution and Weibull distribution. Moreover, the reliability theory is used to analyze the time-varying reliability of the degradation data series. Finally, the validity of the approach is proved by comparing with the reliability evaluation method of large samples.
出处
《强度与环境》
2017年第6期57-64,共8页
Structure & Environment Engineering
基金
黑龙江省自然科学基金项目:水下超空泡航行体动力稳定可靠性分析(A201311)资助
关键词
时变可靠性
退化数据
残差自回归模型
时间序列
小样本
可靠性
time-variant reliability
degradation data
residual autoregressive model
time series
smallsample
reliability
