多源数据聚类下电力二次系统故障检测方法  

Fault Detection Method of Power Secondary System Based on Multi Source Data Clustering

作　　者：陈禹 高翔[2] 徐建航 梁天 梁兴保 CHEN Yu;GAO Xiang;XU Jian-hang;LIANG Tian;LIANG Xing-bao(State Grid Tianjin Xintong Company,Tianjin 300140 China;State Grid Tianjin Electric Power Company,Tianjin 300010 China;Beijing Kedong Power Control System Co.,Ltd.,Beijing 100085 China)

机构地区：[1]国网天津信通公司,天津300140 [2]国网天津市电力公司,天津300010 [3]北京科东电力控制系统有限责任公司,北京100085

出　　处：《自动化技术与应用》2023年第9期45-48,共4页Techniques of Automation and Applications

基　　金：国网天津市电力公司科技项目(KJ21-1-5)。

摘　　要：传统方法不能有效地考虑二次设备数据间的复杂关联,难以及时发现较多潜在故障,为此提出多源数据聚类下电力二次系统故障检测方法。采用小波变换探测电力信号的瞬态成分,展开信号中各频率成分,以大数据技术计算获得数据信息的变化规律和内在联系,采用聚类方法简化多源数据间联系,实现在线检测出多源数据序列内异常度,完成系统故障检测全过程。实验条件下,以准确率和召回率作为衡量指标验证方法的有效性,结果表明所提方法能够有效地检测出二次设备的存在异常状态,可以为故障维修提供决策依据。Traditional methods can not effectively consider the complex association between secondary equipment data,which makes it difficult to find many potential faults in time.The wavelet transform is used to detect the transient components of power signal,and the frequency components of the signal are expanded.The big data technology is used to calculate the change rule and internal relationship of data information.The clustering method is used to simplify the relationship between multi-source data,realize the on-line detection of abnormal degree in multi-source data sequence,and complete the whole process of system fault detection.The experimental results show that the proposed method can effectively detect the abnormal state of secondary equipment,and can provide decision-making basis for fault maintenance.

关 键 词：小波变换 电力信号瞬态成分 大数据技术 多源数据 序列异常度计算 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]