摘要
文章设计了一种基于深度神经网络的电能计量异常筛选方法,其通过拉格朗日插值方法对电能计量数据实施丢帧处理,分别筛选出电量异常的特征数据、用电异常的特征数据以及接线异常的特征数据,将遗传算法与深度神经网络中的BP神经网络相结合,构建电能计量异常筛选模型,实现电能计量异常筛选。测试结果表明,该方法筛选中的分类精度较高,最高可达98.52%;在实验数据中加入畸变数据后,该方法的电能计量异常筛选误差仍较低。
This paper designs a method for screening the abnormal data of electric energy measurement based on deep neural network.The Lagrangian interpolation method is used to process the frame loss of the electric energy measurement data,and the characteristic data of the electric quantity anomaly,the characteristic data of the electric consumption anomaly and the characteristic data of the wiring anomaly are screened out respectively.The genetic algorithm is combined with the BP neural network in the deep neural network to build a power metering anomaly screening model to achieve the power metering anomaly screening.The test results show that the classification accuracy of this method is high,up to 98.52%;After adding the distorted data into the experimental data,the error of abnormal screening of electric energy metering is still low.
作者
金晨
JIN Chen(Hangzhou Power Supply Company,Hangzhou 310000,China)
出处
《数字通信世界》
2023年第1期84-86,共3页
Digital Communication World
关键词
深度神经网络
电能计量
遗传算法
适应度
拉格朗日插值
异常筛选
deep neural network
electric energy metering
genetic algorithm
fitness
lagrange interpolation
exception filtering
