基于改进FCM聚类的电力计量数据异常辨识研究  

Research on anomaly identification of electricity power measurement data based on improved FCM clustering

作　　者：钟磊 姜雪娇 吴民 孙延松 徐佳隆 ZHONG Lei;JIANG Xuejiao;WU Min;SUN Yansong;XU Jialong(Hainan Power Grid Company Limited,Haikou 570010,China)

机构地区：[1]海南电网有限责任公司,海南海口570010

出　　处：《电子设计工程》2025年第4期142-145,150,共5页Electronic Design Engineering

基　　金：南方电网有限责任公司科技项目(ZN-YD-007)。

摘　　要：针对电力计量数据量大且较为复杂,导致电力计量数据异常辨识难度上升的问题,提出基于改进FCM聚类的电力计量数据异常辨识方法。根据电力异常负荷的FCM聚类条件,确定改进聚类参数的取值范围,计算异常电力数据的实时负荷量,完成电力异常数据的初步统计。针对异常负荷数据实施耦合处理,结合异常负荷行为分析条件求解异常值指标,以此实现电力计量数据异常辨识。实验结果表明,该方法电力异常负荷数据的实验结果与标准值的数值变化基本一致,二者差值接近于0,能够有效实现电力异常负荷数据聚类统计,满足实际应用需求。Aiming at the problem of increasing difficulty in identifying abnormal electricity metering data due to the large and complex amount of electricity metering data,a method for abnormal identification electricity power measurement data based on improved FCM clustering is proposed.Based on the FCM clustering conditions of abnormal power loads,determine the range of improved clustering parameters,calculate the real-time load of abnormal power data,and complete the preliminary statistics of abnormal power data.Implement coupling processing for abnormal load data,and combine abnormal load behavior analysis conditions to solve abnormal value indicators,in order to achieve abnormal identification of power metering data.The experimental results show that the experimental results of this method for abnormal power load data are basically consistent with the changes in the standard value,and the difference between the two is close to 0.It can effectively achieve clustering statistics of abnormal power load data and meet practical application needs.

关 键 词：改进FCM聚类 电力计量 数据辨识 电力负荷 数据耦合 

分 类 号：TN911[电子电信—通信与信息系统]