基于局部异常点检测算法的电能表飞走异常智能分析  

The intelligent analysis on the trend anomaly of the electric energy meter based on LOF algorithm

作　　者：李亦非[1] 宋玮琼[1] 彭放 丁宁[1] 王芳[1] 

机构地区：[1]国网北京市电力公司电力科学研究院,北京100045

出　　处：《电测与仪表》2016年第18期69-73,共5页Electrical Measurement & Instrumentation

摘　　要：针对用电信息采集系统在计量设备在线监控业务应用的实际情况,文中介绍了一种基于LOF局部异常点检测算法的电能表飞走异常分析方法,该方法可以通过计算电能表示值异常因子,有效判断电能表示值异常属于偶然异常还是趋势性异常,在排除偶然数据的影响后,进行电能表飞走异常的二次判断。该方法可以有效提高该类异常判断的准确性,避免故障误判,提高了现场故障处理的工作效率。Aiming at the actual situation of the on-line monitoring and controlling system of the electric information acquisition system,this paper introduces a method of detecting the abnormal point of the electric energy meter based on the LOF（ loacl outlier factor） detection algorithm,which can be used to detect the occasional anomaly or the trend anomaly,and conducts the secondary judgment of trend anomaly after eliminating the influence of accidental data. This method can effectively improve the accuracy of this kind of anomaly judgment,avoid the misjudgment of the fault,and improve the working efficiency of the on-site fault treatment.

关 键 词：智能电能表 用电信息采集系统 LOF 异常点 

分 类 号：TM933[电气工程—电力电子与电力传动]