• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

计算机工程与科学

• 人工智能与数据挖掘 • 上一篇    下一篇

故障样本的多指标集成加权分配方法

刘磊,宋家友   

  1. (郑州大学信息工程学院,河南 郑州 450001)
  • 收稿日期:2017-01-16 修回日期:2017-03-29 出版日期:2019-04-25 发布日期:2019-04-25
  • 基金资助:
     
     

A multiple index integrated weighted
method of failure samples allocation

LIU Lei,SONG Jiayou   

  1. (School of Information Engineering,Zhengzhou University,Zhengzhou 450001,China)
  • Received:2017-01-16 Revised:2017-03-29 Online:2019-04-25 Published:2019-04-25

摘要:

针对测试性验证实验中故障样本分配结果置信度比较低的问题,提出了一种多指标集成加权的分配方法。首先,综合分析故障属性与环境因子对故障样本分配结果置信度的影响,在此基础上,定量计算影响指标的值;其次,采用集成加权法确定各个指标的权重;最后,利用归一化后的指标比值进行故障样本分配。该方法改进了故障率、严酷度、危害度及指标权重的计算方法,并提出了故障样本分配中故障-环境关联度的计算方法。在某型装备故障样本分配中的应用结果表明,该方法的分配结果更合理,置信度更高,更具有实际应用价值。
 
 

关键词: 故障样本分配, 故障属性, 环境因子, 集成加权, 故障-环境关联度

Abstract:

Aiming at the problem that the result of failure samples allocation is imprecise in testability demonstration test, a multiple index integrated weighted method is proposed. Firstly, on the basis of analyzing the effect of failure attribute and environment factor on failure samples allocation comprehensively, the value of each factor is calculated. Secondly, the weight of each factor is determined by the integrated weighted method. Finally, the ratio of the normalized indicators is applied to allocate the failure samples. The proposed method improves the calculation method of failure rate, severity degree, damage degree and the weight. A calculation method of correlation degree between failure and environment is proposed. Through the application of fault sample allocation in a certain type of equipment, the result shows that the proposed method is more reasonable and credible, and has higher practical value

Key words: