电网故障诊断专家系统中基于粗糙集理论的知识获取方法的研究

提出了电网故障诊断专家系统中基于粗糙集理论的知识获取的一种新方法,即利用分层诊断中由断路器变位信息诊断出故障区域,指导决策表中条件属性的选择,最大限度地剔除多余的条件属性,通过对决策表进行属性约简和值约简,得出故障诊断需要的诊断知识即诊断规则,并以三水地区的一个电网故障为例,说明该方法的有效性。

Research on Rough-set-based Knowledge Acquisition in Power System Electric Network Fault Diagnostic Expert System

This paper proposes a new method of rough-set-based knowledge acquisition in the power system fault diagnostic expert system.The method uses the fault section that has been estimated by the tripped breakers in the hierarchical fault diagnosis system to decide which conditional attributes in the decision table will be selected to remove the redundant conditional attributes.The diagnostic rules are extracted by attribute-reducing and value-reducing the decision table.An example of fault sample in Sanshui Power System shows that the method presented in the paper has not only effective knowledge acquisition,but also excellent fault-tolerance capability.