• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2016, Vol. 52 ›› Issue (2): 182-188.doi: 10.3901/JME.2016.02.182

• 交叉前沿 • 上一篇    下一篇

扫码分享

极少失效数据和无失效数据的可靠性评估

贾祥1,  王小林2,  郭波1   

  1. 1. 国防科技大学信息系统与管理学院  长沙  410073;
    2. 78020部队  昆明  655000
  • 收稿日期:2015-07-05 修回日期:2015-12-05 出版日期:2016-01-15 发布日期:2016-01-15
  • 通讯作者: 贾祥,男,1992年出生,博士研究生。主要研 究方向为系统可靠性分析。 E-mail:jiaxiang09@sina.cn
  • 基金资助:
    国家自然科学基金资助项目(71371182)

Reliability Assessment for Very Few Failure Data and Zero-failure Data

JIA Xiang1,  WANG Xiaolin2,  GUO Bo1   

  1. 1. College of Information System and Management, National University of Defense Technology, Changsha 410073;
    2. Unit 78020 of PLA, Kunming 655000
  • Received:2015-07-05 Revised:2015-12-05 Online:2016-01-15 Published:2016-01-15

摘要: 当前对于无失效数据和极少失效数据的处理方法,不能同时得到参数的点估计和置信区间估计。在具体应用中,如果基于不同方法得到点估计和置信区间估计,则造成结果的可信度降低。针对这一问题,在威布尔分布下,对当前广泛应用的配分布曲线法的各个环节进行详细分析并加以改进,使得计算结果既包含点估计,也包含置信区间估计,从而增强了结果的可信度。经过对3组不同产品的寿命数据的分析计算,并与其他方法得到的结果相对比,发现根据所提出的方法计算得到的结果更为合理。这证明了这一方法的有效性和合理性。

关键词: 极少失效, 配分布曲线法, 威布尔分布, 无失效

Abstract: For zero-failure data and very few failure data, the current methods fail to obtain both the estimate and confidence interval of parameters. If the estimate and confidence interval of parameters are obtained based on different methods, the results seem to be unconvincing. In this case, the match distribution curve (MDC) method is analyzed and improved to be able to compute both the estimate and confidence interval of parameters for Weibull distribution. And the improved method could make the results more convincing. After calculation for three examples, the estimate and confidence interval of parameters based on the proposed method are better than the results by other methods. It demonstrates the effectiveness and rationality of this method.

Key words: the match distribution curve method, very few failure data, Weibull distribution, zero-failure data

中图分类号: