基于大数据的变结构BP神经网络反窃电技术研究  

Research on Anti-electric Larceny Technology Based on Big Data Variable Structure BP Neural Network

作　　者：强浩[1] 戴巧云 吴柯 杜健 殷新博 陈琛 QIANG Hao;DAI Xiao-yun;WU Ke;DU Jian;YIN Xin-bo;CHEN Chen(School of Urban Rail Transit,Changzhou University,Changzhou 213164,China;State Grid ChangzhouPower Supply Company,Changzhou 213000,China;State Grid Jiangsu Electric Power CO.,Ltd.Economic Research Institute,Nanjing 210000,China)

机构地区：[1]常州大学城市轨道交通学院,江苏常州213164 [2]国网江苏省电力有限公司常州供电分公司,江苏常州213000 [3]国网江苏省电力有限公司经济技术研究院,江苏南京210000

出　　处：《江苏理工学院学报》2019年第2期10-14,共5页Journal of Jiangsu University of Technology

基　　金：江苏省高等学校大学生创新创业训练计划项目“基于大数据的反窃电智能平台研究”(20181029020Z)

摘　　要：随着经济的发展,人们对电力的需求逐渐增大,用户的窃电行为成为不容忽视的问题。基于大数据,开展了变结构BP神经网络窃电用户识别技术的研究:首先,根据用户所在线路参数、用户用电参数等建立用户用电评价指标体系;其次,基于不同线路用户的实际采集数据构建BP神经网络,实现对该用户的用电评价;最后,通过算例验证该方法能有效提高窃电疑似用户的识别效率。With the development of the economy,the demand for electric power increases day by day,and the problem of electricity theft has become more and more prominent.Based on big data,this paper carried on the research on the user identification technology of power theft in variable structure BP neural network.First,according to the line parameters and the electricity consumption parameters,the user electricity evaluation index system is established.Then based on the system,a variable structure BP neural network is presented to evaluate the behavior of user electricity consumption.Finally,a numerical example is given to demonstrate the effectiveness of the proposed method.

关 键 词：大数据 BP神经网络 变结构 防窃电 

分 类 号：TM73[电气工程—电力系统及自动化]