基于BP神经网络的配电网防窃电降线损研究  

Study of anti-power theft and line loss reduction for power distribution network based on BP neural network

作　　者：黄星知 杨奕纯 杨兰[3] 吕飞帆 HUANG Xing-zhi;YANG Yi-chun;YANG Lan;LV Fei-fan(Information&Communication Company,State Grid Hunan Electric Power Co.Ltd.,ChangSha 410000,China;Huaneng Wuhan Power Generation Co.Ltd.,Yang luo Power Plant,Wuhan 431400,Chian;School of Electrical&Information Engineering,Changsha University of Science and Technology,Changsha 410114,China)

机构地区：[1]国网湖南省电力有限公司信息通信分公司,湖南长沙410000 [2]华能武汉发电有限责任公司阳逻电厂,湖北武汉431400 [3]长沙理工大学电气与信息工程学院,湖南长沙410174

出　　处：《电力科学与技术学报》2019年第4期143-147,共5页Journal of Electric Power Science And Technology

基　　金：湖南省自然科学基金(13JJ6044)

摘　　要：窃电增大了管理线损,窃电手段的日趋复杂又难用某一确定的逻辑或算法识别窃电,因此提出了基于BP神经网络的窃电检测方法。以强针对性的支路线损变化率和三相电压及三相电流的不平衡率作为窃电检测指标并归一化处理,且用具有双隐含层的BP神经网络提高窃电检测的准确性。测试结果表明,该方法不仅窃电检测准确率高,而且更加智能化的变被动防窃电为主动防窃电,从而有效降低管理线损。Power theft increases the difficulty of line loss management. As the means of power theft become more and more complex, it is difficult to identify power theft with a certain logic or algorithm. Therefore, a detection method of power theft is proposed based on BP neural network. The branch line loss rate, the three-phase voltage and current unbalance rate are normalized and taken as the detection indices in corporation of the double hidden layer BP neural network to improve the accuracy of detection. The test results show that the method not only has a high detection accuracy of power theft, but also intelligently changes the passive power theft prevention into active one. Thus the method effectively reduces line loss management.

关 键 词：窃电检测 BP神经网络 配电网 管理线损 

分 类 号：TM73[电气工程—电力系统及自动化] TP274[自动化与计算机技术—检测技术与自动化装置]