基于计量自动化系统的电能计量异常诊断技术研究

Research on Abnormal Diagnosis Technology of Electric Energy Metering Based on Metering Automation System

为及时发现电能计量异常,减少电力企业和用户损失,提出基于计量自动化系统的电能计量异常诊断技术。利用面向服务的体系架构设计对应计量自动化系统,将系统分为数据采集、信息交换、业务处理等层次,建立整体架构;设置系统档案管理、预警管理、实时抄表等功能;在系统中引入均值聚类算法,根据采集的大量历史计量信息,建立时间序列重构计量信息,填补缺失数据;使用相对邻域信息熵算法,结合目标样本权重,计算异常数据在全域中的比例,通过离群度诊断电能计量是否异常。仿真实验表明,所提技术可以有效填补计量缺失数据,无论对于设备异常还是用户异常均有较高的诊断精度。

In order to find out abnormalities of the electric energy measurement in time and reduce the losses of power enterprises and users,an abnormal diagnosis technology of the electric energy measurement based on measurement automation system is proposed in this paper.The corresponding measurement automation system is designed by using the service-oriented architecture.The system is divided into data collection,information exchange,business processing and other layers;the overall architecture is established,and the system file management,early warning management,real-time meter reading and other functions are set.The mean clustering algorithm is introduced into the system and according to a large amount of historical measurement information,a time series is established to reconstruct the measurement information and fill the missing dat.Using the relative neighborhood information entropy algorithm,combined with the target sample weight,the proportion of abnormal data in the whole domain is calculated,and the outlier is used to diagnose whether the electric energy metering is abnormal.The simulation results show that the proposed technology can effectively fill in the missing measurement data,and has high diagnosis accuracy for equipment abnormalities and user abnormalities.