基于多特征量的D矩阵模型的建立与分析  

Establishment and analysis of D matrix model based on multi-feature quantity

作　　者：孙萌 景博[1] 黄以锋[1] 焦晓璇 徐光跃[1] 

机构地区：[1]空军工程大学航空航天工程学院,西安710038

出　　处：《电子测量与仪器学报》2017年第11期1731-1736,共6页Journal of Electronic Measurement and Instrumentation

基　　金：航空科学基金(20142896022)资助项目

摘　　要：针对基于单特征量的D矩阵模型会损失测试信号的大量信息的问题,提出了一种基于多特征量的可测性分析模型。首先,在研究D矩阵模型的基础上,对测试信号进行特征提取和多值编码,建立了基于多特征量的D矩阵模型。其次,基于信息熵制定诊断策略和生成故障诊断树,并对模型进行测试性分析,评估其诊断能力。结果表明,与单特征量的D矩阵模型相比,基于多特征量的D矩阵模型选用的测试点数和模糊组数小,故障隔离率高,诊断测试步骤数少,平均测试费用低。Aiming at the problem that the D matrix model based on single feature quantity can lose a lot of information about the test signal, a testability analysis model based on multi-feature quantity is proposed. Firstly, based on the study of D matrix model, the feature extraction and muhi-value coding of the test signal are carried out, and a D matrix model based on multi-feature quantity is established. Then, the diagnostic strategies are developed based on the information entropy, the fault diagnosis tree is generated, and the model is analyzed to assess its diagnostic capabilities. The results show that the D matrix model based on multi-feature quantity adopts a smaller number of test points and fuzzy group, and has a higher fault isolation rate, less diagnostic test step, and lower average test cost compared with the D matrix model based on the single feature quantity.

关 键 词：相关性模型 D矩阵模型 多特征量 信息熵 

分 类 号：TN707[电子电信—电路与系统] TP273[自动化与计算机技术—检测技术与自动化装置]