Identification method of abnormal data of electric energy measurement based on BP neural network algorithm

Abstract

The problem of unclear power load type exists in the conventional abnormal data identification method of electric energy metering, which leads to a high error percentage. A method of abnormal data identification of electric energy metering based on BP neural network algorithm is designed. In this paper, the data of electric energy measurement are obtained and used to calculate the active power loss. It adjusts the connection weights and thresholds between neurons according to the error function of sample size. In this paper, BP neural network algorithm is used to identify the type of power load, and the time balance data set is constructed. According to the data set discriminant analysis of key behavior pattern characteristics, the abnormal data recognition of power measurement is completed. The experimental results show that the average error percentage of the abnormal data identification method of electric energy measurement in this paper is 0.0233, 0.0530, 0.0539 and 0.0524, respectively, compared with the other three abnormal data identification methods of electric energy measurement. The abnormal data identification method of electric energy measurement in a plain text has higher reliability.

Research on the strategy of locating abnormal data in internet of things management platform based on improved modified particle swarm optimization convolutional neural network algorithm

“…Literature [24] suggests a method based on data mining and neural networks for monitoring and assessing anomalous data. Literature [25] developed a strategy for identifying abnormal electrical energy metering data based on a BP neural network algorithm. In this study, electrical energy measurement data are collected and employed to quantify active power losses.…”