装备故障样本优化方法研究  

Optimization Method for Equipment Fault Samples

作　　者：刘刚[1,2] 黎放[1,2] 

机构地区：[1]海军工程大学管理工程系,湖北武汉430033 [2]海军工程大学舰船综合保障工程研究室,湖北武汉430033

出　　处：《计算机仿真》2015年第1期453-457,共5页Computer Simulation

基　　金：总装技术基础课题(4314231428);海军工程大学自然科学基金项目(HGDQNEQJJ13016);军队院校2110工程三期资助项目

摘　　要：测试性验证装备的故障样本往往相互关联,全部注入费用较高、代价较大。为了降低验证试验费用,采用适当方法对故障样本进行优化分析。为提高故障检测率,提出了一种等价样本的故障样本优化方法。方法在分析故障-测试关联矩阵及其扩展、故障模式功能等价集合和故障模式测试等价集合的基础上,构建了故障样本等价集合,并进行重要度特征分析和最大熵求解,确定了最小的故障样本集合。通过对某型试验台故障样本优化实例分析,并与传统的方法进行了试验结果对比分析,使得故障样本数量、试验费用大为减少,提高了测试性验证的经济性。The fault samples are related to each other in the process of testability verification experiment. In order to reduce experiment costs, an appropriate approach should be taken for fault sample optimization. In this paper, a fault sample optimization approach based on sample equivalence is proposed. The fault-test dependency matrix and its extended matrix, fault mode function equivalence set and fault modes test equivalence set are analyzed, and fault mode sample equivalence set is proposed. Meanwhile, importance degree feature is analyzed and maximum entropy is solved, and the smallest fault sample set is obtained. With the result that fault sample optimization of some experiment table and contrastive analysis of the other methods, it is proved that this method can reduce the amount of fault samples and experiment costs, improve the economy of testability verification experiment.

关 键 词：故障 验证试验 最大熵 样本优化 

分 类 号：TP206[自动化与计算机技术—检测技术与自动化装置]