多约简融合的数字化变电站故障诊断  

Fault Diagnosis of Digital Substation with Multi-reduction Fusion

作　　者：张炳达[1] 冯鑫[1] 黄杰[1] 郭凯[1] 赵紫昆 

机构地区：[1]天津大学智能电网教育部重点实验室,天津300072

出　　处：《电力系统及其自动化学报》2016年第4期1-5,共5页Proceedings of the CSU-EPSA

基　　金：天津市科技计划项目(13TXSYJC40400);国家自然科学基金资助项目(51477114)

摘　　要：针对变电站故障诊断中告警信息的不确定性和多源性,提出了一种基于粗糙集和信息融合技术的数字化变电站故障诊断方法。将故障区域和故障装置作为决策对象,对电压、电流条件属性值做适当的离散,形成比较详细的变电站故障诊断原始决策表,并给出了基于Apriori算法的约简枚举方法。在定义事例相似度的基础上,采用信息融合技术确定故障区域可信度和故障装置可信度,并通过层次分析法实施故障诊断。实践表明,该方法充分利用了数字化变电站GOOSE报文和SAV报文,在信息传输有误的情况下也有较好的诊断效果。A fault diagnosis method of digital substation based on rough set and information fusion technique is pro- posed for the uncertainty and multi-source factors of fault information in a substation. In this paper, decision attributes, composed of fault section and fault device, as well as condition attributes, discretized voltage and current, form a more detailed original decision table. Then the Apriori algorithm is applied to enumerate all reductions of the decision table. By defining the similarity of instances, the information fusion technique determines the creditability of fault section and fault device, meanwhile, the analytic hierarchy process performs fault diagnosis. The proposed method, which has a good diagnosis result in the case of wrong information transmission, makes full use of GOOSE message and SAV message in the digital substation.

关 键 词：数字化变电站 故障诊断 决策表 信息融合 层次分析法 

分 类 号：TM63[电气工程—电力系统及自动化]