采用粗糙集联合规则挖掘算法的分布式电网故障诊断

Distributed Fault Diagnosis of Power Networks Applying the United Rules Mining Algorithm Based on Rough Set Theory

鉴于传统的人工智能技术在大电网故障诊断应用中存在的局限性,采用分布式的思想来解决该问题,并提出一种基于粗糙集理论的联合规则挖掘算法,该算法能够有效地从分布式信息系统中提取联合规则。为了更好地解决电网故障诊断问题,构建了基于该联合规则挖掘算法的分布式电网故障诊断模型。该模型不仅能够有效地诊断各局部电网内部的故障,而且能够有效地诊断各局部电网之间联络线的故障,弥补了分布式电网故障诊断中联络线故障诊断规则不足的缺陷。对大电网进行分割,不仅能够有效识别各类复杂故障,而且也使得每个局部电网决策表的规模大为减小,同时联合规则挖掘算法也显著地降低了规则提取的复杂度,解决了粗糙集理论在大电网故障诊断中遇到的瓶颈问题。算例表明该方法简单、有效、速度快、容错性好。

Considering the limitations of traditional artificial intelligence technology in the fault diagnosis of large scale power networks, distributed method was used to solve the problem. A novel united rules mining algorithm based on rough set theory was proposed for extracting united rules from distributed information system. In order to diagnose the faults better, a model of distributed fault diagnosis of power networks was constructed based on the algorithm. The faults in local power networks and the faults of tie lines between local power networks can be efficiently diagnosed by using the model, which has made up for the deficiency of fault diagnosis rules for tie lines in distributed fault diagnosis. Various complex faults can be identified efficiently. The decision table for sub-networks is greatly simplified by dividing the large scale power network into desired number of connected sub-networks Meanwhile the united rules mining algorithm can significantly reduce the complexity of the rule extraction and solve the bottlenecks that the rough set theory encounters in the application to large power networks diagnosis. The results of tests show that the method is simple, efficient, rapid and good fault-tolerant.