基于数据挖掘模型的高压输电线系统故障诊断  

A NEW DATA MINING APPROACH FOR FAULT DIAGNOSIS OF HIGH VOLTAGE TRANSMISSION LINE BASED ON ROUGH SET THEORY

作　　者：廖志伟[1] 孙雅明[1] 

机构地区：[1]天津大学电气自动化与能源工程学院,天津300072

出　　处：《电力系统自动化》2001年第15期15-19,共5页Automation of Electric Power Systems

基　　金：国家自然科学基金资助项目 (5 98770 16 )

摘　　要：在大多数故障诊断系统中 ,由于诊断所依据的实时信息在其形成和传递过程中都有可能产生信息的畸变 ,从而导致故障诊断结果的错误。文中提出利用基于粗糙集理论的数据挖掘模型来处理实时输入信息的畸变和实现输电线系统的故障诊断 ,它是依据粗糙集定性分析能力对知识域的数据集进行分析 ,粗糙集的约简是通过遗传算法求取。还给出了构造测试样本的理论准则 ,从而使检验故障诊断系统的容错性能具有保证和真正的实用价值。通过仿真测试证明 ,基于数据挖掘模型的故障诊断与基于神经网络模型的故障诊断相比 ,具有更高的容错性能。In most practical application of fault diagnosis system, misjudgement may be caused by real-time information distorted in the process of generation and transfer. This paper presents rough set (RS) based data mining method to deal with distorted information and to implement the fault diagnosis of HV transmission line system (HVTLS). In this approach, the qualitative analysis ability of RS is used to analyze knowledge region data set and the reductions of RS are solved by a genetic algorithm (GA). At same time. this paper proposes the criterion to build testing samples, in order to get the assurance of fault tolerance performance of tested diagnosis system and have practical application potential of studied system. The higher fault tolerance performance of the proposed approach is confirmed through the comparison with that of NN-model based fault diagnosis system.

关 键 词：高压输电线系统 故障诊断 数据挖掘 粗糙集 

分 类 号：TM723[电气工程—电力系统及自动化]