摘要
综合分析历史电量和负荷等数据特征,建立用电行为特征库,提出一种基于用电特征分析的无监督方式异常用电检测方法。该检测方法引入离群点查找算法,量化海量数据中不同异常用电行为,将其提取为异常用电特征序列,再构建基于局部离群因子(Local Outlier Factor,LOF)检测算法,实现疑似异常用电用户的在线快速识别与精准定位,提高现场检查的命中率,降低运营成本。
A comprehensive analysis of historical power and load data features,a power consumption behavior feature library,and an unsupervised method of abnormal power consumption detection based on power consumption feature analysis are proposed.The detection method introduces outlier point finding algorithm,quantifies different abnormal power consumption behaviors in massive data,extracts them into abnormal power consumption feature sequences,and builds a Local Outlier Factor(LOF)based detection algorithm,which realizes online fast identification and accurate positioning of suspected abnormal power users,improves the hit rate of on-site inspection,reduces operation costs.
作者
苏鹏涛
孙晨辉
SU Pengtao;SUN Chenhui(Shanghai Shineenergy Information Technology Development Co.,Ltd.,Shanghai 200025,China)
出处
《信息与电脑》
2022年第15期55-57,共3页
Information & Computer
关键词
异常用电
局部离群因子(LOF)检测
特征序列
abnormal electricity consumption
Local Outlier Factor(LOF)detection
signature sequence
