基于用户行为的窃电嫌疑分析  

Analysis on Suspicion of Stealing Electricity Based on User Behavior

作　　者：玉梅 王昕 任关友[2] 高海丽 殷要红 YU Mei;WANG Xin;REN Guanyou;GAO Haili;YIN Yaohong(Yunnan Wenshan Electricity Co., Ltd., Wenshan, Yunnan 663000;Yunnan Electric Power Research Institute, Yunnan Power Grid Co.Ltd., Kunming 650217;Yunnan power grid co., LTD XiShuangbanna Menghai power supply bureau, Menghai, Yunnan 666200;Qujing Qilin Power Supply Bureau, Yunnan Power Grid Co.Ltd., Qujing, Yunnan 655000;Key Laboratory of Electric Power Measurement, China Southern Power Grid, Kunming 650217)

机构地区：[1]云南文山电力股份有限公司,云南文山663000 [2]云南电网有限责任公司电力科学研究院,昆明650217 [3]中国南方电网公司电能计量重点实验室,昆明650217 [4]云南电网有限责任公司勐海供电有限公司,云南勐海666200 [5]云南电网有限责任公司曲靖麒麟供电局,云南曲靖655000

出　　处：《云南电力技术》2018年第2期57-59,62,共4页Yunnan Electric Power

摘　　要：针对窃电问题,分析了窃电原理以及对现有窃电与防窃电方法进行了归纳与总结。利用电力计量自动化系统采集的数据和BP神经网络算法,构建了窃电嫌疑分析模型。结合实际案例对模型准确性进行评估,实验表明利用神经网络预测窃电嫌疑具有很高的可行性及可靠性。Aiming at the problem of stealing electricity, the principle of stealing electricity and the methods of stealing electricity and preventing stealing electricity are summarized and summarized. Based on the data collected from the electric power metering automation system and the BP neural network algorithm, this paper constructs a model to analyze the electric larceny. Finally, the accuracy of the model is evaluated by the actual case. The experiment shows that the use of neural network to predict the suspicion of stealing electricity is of high feasibility and reliability.

关 键 词：电能计量 防窃电 数据分析 BP神经网络 

分 类 号：TM74[电气工程—电力系统及自动化]