基于离群点算法的低压窃电行为主动辨识系统  

Active identification system based on outlier algorithm against low-voltage electricity stealing behavior

作　　者：黄荣国 陆春光 姚力 玄鑫 HUANG Rongguo;LU Chunguang;YAO Li;XUAN Xin(Electric Power Research Institute,State Grid Zhejiang Electric Power Corporation,Hangzhou 310000,China;School of Electric and Information,China Three Gorges University,Yichang 443002,China;Beijing China-Power Information Technology Co.Ltd.,State Grid Information&Telecommunication Group,Beijing 100008,China)

机构地区：[1]国网浙江省电力有限公司电力科学研究院,杭州310000 [2]三峡大学电气信息学院,湖北宜昌443002 [3]国网信通产业集团北京中电普华信息技术有限公司,北京100008

出　　处：《沈阳工业大学学报》2023年第5期486-490,共5页Journal of Shenyang University of Technology

基　　金：国家自然科学基金项目(11957203);国家电网公司总部科技项目(5400-201925177A-0-0-00).

摘　　要：为了更有效地打击窃电行为、提高对窃电行为的辨识能力,基于离群点算法设计了一种低压窃电行为主动辨识系统,用以实现对窃电现象的全面分析.根据距离离群点检测算法对原始数据进行预处理,利用判别规则筛选出离群点,再通过数据采集器直接采集用户侧用电信息,作为判断窃电的依据.基于此,通过管理电力负荷来监控电路数据信息,并引入椭圆曲线密码体制,确保用电信息的安全传输,从而构建低压窃电行为主动辨识系统.实验结果表明:该系统能够准确、安全地获取用电信息,及时辨识窃电行为,增强反窃电能力.In order to combat electricity stealing behavior more effectively and improve the identification ability of this behavior,an active identification system based on outlier point algorithm against low-voltage electricity stealing was designed to realize a comprehensive analysis of electricity stealing phenomenon.According to the distance outlier detection algorithm,the original data was preprocessed,and the outliers were screened out by the discriminant rules.Then the data collector directly collected the users’side electricity information,which was used as the basis for the judgement of electricity stealing behavior.On account of this,the circuit data information was monitored by power load management,and the elliptic curve cryptosystem was introduced to ensure the safe transmission of electricity information,in order to construct the active identification system against low-voltage electricity stealing behavior.The experimental results show that the system can accurately and safely obtain electricity information,identify electricity stealing behavior in time,and enhance the ability of anti-electricity stealing behavior.

关 键 词：离群点算法 低压窃电行为 数据采集器 主动辨识 椭圆曲线密码体制 数据预处理 电力负荷 用电信息采集 

分 类 号：TM761[电气工程—电力系统及自动化]