证据理论和粗集在变电站故障诊断中的应用

Application of Evidence Theory and Rough Sets in Substation Fault Diagnosis

由于证据理论存在如基本概率人为给定而导致强烈主观性等固有缺陷,文中首先分析了粗集理论与证据理论的关系,确定了粗集是证据理论的基础,进而描述了基于粗集和证据理论的变电站故障诊断的特殊过程,并研究了基于上述理论的变电站故障诊断推理算法。通过在变电站故障诊断专家系统的不确定知识决策规则中应用该算法,表明此算法能增强推理过程的稳定性,并具有较强的鲁棒性,是一种处理不确定和不完整信息的有效方法。

In view of the inherent defects of Evidence Theory such as that the basic probability assignments （bpa） given out by people would lead to strong subjectivities, the paper analyzes the relation between rough sets （RS） and Evidence theory firstly, and make it certain that RS is the basis of Evidence Theory. Then the paper presents the special process of substations fault diagnosis based on RS and Evidence Theory, and investigates the reasoning algorithm for a substation fault diagnosis. By using the algorithm in the uncertainty knowledge decision-making rules of the fault diagnosis expert system for the substation, the results show that the algorithm can strengthen stability of reasoning and own strong robustness, therefore it is an effective information disposal method for uncertain and incomplete information.