关口电能计量装置电压回路故障自动预警方法  

Automatic Early Warning Method for Voltage Circuit Fault of Gateway Electric Energy Metering Device

作　　者：范秀云 何佳美 侯欣怡 杨彩月 FAN Xiuyun;HE Jiamei;HOU Xinyi;YANG Caiyue(Marketing Service Center(Fund Intensive Control Center,Metrology Center),State Grid Jibei Electric Power Co.,Ltd.,Beijing 100045,China)

机构地区：[1]国网冀北电力有限公司营销服务中心(资金集约中心、计量中心),北京100045

出　　处：《自动化与仪表》2023年第9期104-109,共6页Automation & Instrumentation

摘　　要：关口电能计量装置校验依赖人工现场检定,工作量大且效率低下,不能及时发现并处理其产生的电压回路故障等问题,为此提出基于深度置信网络的关口电能计量装置电压回路故障自动预警方法。首先采用标准化法与插值法预处理电压数据,以此为基础搭建电压数据预测框架。应用XGBoost算法选择并提取预测电压数据特征。基于支持向量机原理构建最大化类间间隔的分类面,判定预测当前关口电能计量装置电压数据的状态,实现关口电能计量装置电压回路故障自动预警。实验结果表明,应用所提方法获得的异常电压检测比例数值与实际数值更接近,电压回路故障预警成功概率最大值为96%,应用性能更佳。The verification of gateway electric energy metering device relies on manual on-site verification,which is heavy in workload and inefficient,and can not timely find and deal with the problems such as voltage circuit faults.Therefore,an automatic early warning method for voltage circuit faults of gateway electric energy metering device based on deep confidence network is proposed.First,the standardization method and interpolation method are used to preprocess the voltage data,based on which a voltage data prediction framework is built.XGBoost algorithm is applied to select and extract the features of predicted voltage data.Based on the principle of support vector machine,a classification surface is constructed to maximize the interval between classes,judge and predict the current state of the voltage data of gateway electric energy metering device,and realize automatic early warning of the voltage circuit fault of gateway electric energy metering device.The experimental results show that the ratio of abnormal voltage detection obtained by the proposed method is closer to the actual value,and the maximum probability of successful early warning of voltage circuit fault is 96%,so the application performance is better.

关 键 词：关口电能计量装置 电压回路故障 深度置信网络 故障预警 电压预测 

分 类 号：TH89[机械工程—仪器科学与技术] TM72[机械工程—精密仪器及机械]