基于粗糙集的变电站故障诊断规则提取方法研究  

Rule extraction study of substation fault diagnosis based on rough set

作　　者：张延松[1] 赵英凯[1] 

机构地区：[1]南京工业大学自动化与电气工程学院,江苏南京210009

出　　处：《计算机工程与设计》2010年第14期3277-3279,3312,共4页Computer Engineering and Design

基　　金：国家科技攻关计划基金项目(2004BA204B08-03)

摘　　要：运用粗糙集理论研究了变电站故障诊断规则提取,提出了基于属性重要度和分辨矩阵的诊断规则提取算法。通过改进基于属性重要度的粗糙集属性约简算法,得到决策表的一个最小约简,然后将分辨矩阵引入值约简中得到决策规则。采用该算法对变电站故障决策表进行属性约简和值约简,抽取出变电站故障诊断的精确规则库进行故障诊断。实验结果表明了该方法的有效性。A rule extraction approach of substation fault diagnosis is studied based on rough set and the diagnosis rules extraction algorithm is proposed based on the attribute significance and the discernibility matrix. The attribute reduction algorithm of the rough set theory is improved based on the attribute significance to get a minimum reduction of the decision table,then the discernibility matrix is introduced into the value reduction to obtain the decision rules. Finally,the algorithm is applied to the substation fault decision table for the attribute reduction and the value reduction to extract the accurate rules of the fault diagnosis. The result shows the validity of the method.

关 键 词：粗糙集 属性重要度 分辨矩阵 变电站 规则提取 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]