无失效数据下一种新分布的可靠性分析方法  

Study on Reliability Analysis Method for New Distribution under Zero-failure Data

作　　者：姜祥周[1] 师义民[1] 沈政[1] 

机构地区：[1]西北工业大学应用数学系,西安710072

出　　处：《科学技术与工程》2008年第3期600-603,共4页Science Technology and Engineering

基　　金：国家自然科学基金(70471057);陕西省教育厅自然科学基金(03Jk065)项目资助

摘　　要：根据无失效数据推导出分布失效率的Bayes估计,然后运用Virene算法和非线性回归最小二乘法,对分布的可靠性进行评估。最后,利用Monte-Carlo方法对所得结论进行了分析验证。结果表明,在无失效数据下,将非线性回归最小二乘法与Bayes方法相结合,可以很好地评估指数-威布尔分布的可靠性。The Bayes estimation of the distributed failure probability based on zero-failure data is deduced. Then the estimations of the distributed reliability were assessed by using the Virene algorithm and using the least square mother of nonlinear regression. At last, an analysis was given by Monte-Carlo method to prove the conclusion. The computational results indicate that the least square mother of nonlinear regression combined with the Bayes estimation is suitable for the reliability analysis of exponentialted-weibull distribution in the case of zero-failure data.

关 键 词：Virene算法 BAYES估计 可靠性 无失效 非线性回归 最小二乘法 

分 类 号：O213.2[理学—概率论与数理统计]