基于多知识库与粗糙集理论的变电站故障诊断方法

Substation fault diagnosis based on multi-knowledge base and rough set theory

由于变电站故障诊断信息中含有大量不确定、噪声数据,利用粗糙集理论对大量的诊断征兆信息进行分类,从而获得简约的规则,并结合多知识库原理,从单个决策表中获得多个决策数据库,分别简约后再利用数据融合来综合这些简约的规则进行故障的诊断,从而提高诊断的效率和准确性。从最后实例仿真比较的结果中可以看出,该方法对含有噪声的数据也能获得较好的准确率。

The rough set theory was applied to classify the numerous fault diagnosis information which contains much uncertain noisy data when it is transferred from the substation, and the reduced rules were obtained consequently. Meanwhile, the multi-knowledge base was used together with the rough set theory to reduce decision databases obtained from separate decision tables. Based on data fusion, the reduced databases and rules were then used to diagnose faults, improving diagnosis efficiency and precision. Simulation results proves the method effective.