中文摘要：

测试性D矩阵包含系统故障与测试的所有信息,是测试性分析的核心。正确处理D矩阵能极大地提高故障诊断的效率。在继承D矩阵传统处理算法优点的基础上,引入局部信息熵算式,提出一种准信息熵测试性D矩阵故障诊断新算法。新算法结合全局和局部两类寻优算法的特点,具有全局寻优和局部寻优的能力,并与传统算法具有相同的适应性。通过两个实例验证了新算法的有效性,表明新测试性D矩阵故障诊断算法具有减少诊断步数、诊断时间和诊断费用等优点。

英文摘要：

Testability D matrix that is the core of testability analysis includes all information of faults and test of system. Exact manipulation of D matrix could greatly increase the efficiency of fault diagnosis. A local information entropy formula is given,and a new algorithm,called quasi information entropy,is proposed based on the analysis of the D matrix traditional processing methods. The proposed algorithm combines the characteristics of traditional global and local optimization algorithms,so it has the ability of global and local optimization and the same applicability as the traditional deposing methods. The availability of the fault diagnosis algorithm is verified by two examples,which show that the proposed algorithm of manipulating the testability D matrix has prominent advantages,such as reduced diagnostic procedure,shorted diagnostic time and reduced cost.