基于关联规则数据挖掘的台区线损异常检测方法  

A anomaly detection method for line loss based on association rule data mining

作　　者：辛保江 赵晓丽 XIN Bao-jiang;ZHAO Xiao-li(State Grid Weifang Power Supply Company,Weifang 261000,Shandong Province,China)

机构地区：[1]国网潍坊供电公司,山东潍坊261000

出　　处：《信息技术》2025年第4期180-184,192,共6页Information Technology

摘　　要：为确保电力系统的安全稳定运行,提出了基于关联规则数据挖掘的台区线损异常检测方法。利用基于粗糙集的数据归约算法对台区线损数据进行归约处理,去除台区线损冗余数据。通过模糊关联规则,构建台区线损数据间的关联规则,并设定支持度和置信度阈值,对满足阈值要求的关联规则进行挖掘,获取台区线损数据的特征信息。采用K-means聚类法对台区线损数据特征进行分类,与核心点距离最远的数据集即为台区线损异常数据,实现台区线损异常检测。实验结果表明,该方法的数据挖掘性能好、异常检测效率高。To ensure the safe and stable operation of the power system,a line loss anomaly detection method based on association rule data mining is proposed.A data reduction algorithm is utilized based on rough sets to reduce the line loss data in the substation area and remove redundant data in the line loss.By using fuzzy association rules,association rules between substation line loss data are established,and support and confidence thresholds are set.Association rules that meet the threshold requirements are mined to obtain feature information of substation line loss data.K-means clustering method is used to classify the characteristics of substation line loss data,and the data set which is furthest from the core point is the line loss anomaly data,achieving line loss anomaly detection.The experiment results show that this method has good data mining performance and high anomaly detection efficiency.

关 键 词：关联规则 数据挖掘 台区线损 异常检测 K-MEANS聚类 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]