Detection Method for Load Stealing Behavior in Distribution Networks Based on Time Series and Abnormal Characteristics of Regional Electricity Consumption

The detection results of electricity theft based on the characteristic threshold values of electricity consumption behavior data ignore the impact of multidimensional data on the detection results, resulting in low accuracy of the detection results. Therefore, a detection method for load stealing behavior in distribution networks based on time series and regional abnormal electricity consumption characteristics is proposed. Based on time series monitoring of electricity consumption behavior data, filter the dimensions of the monitored data, calculate the stationarity of electricity consumption behavior data dimensions, reduce the impact of data dimensions, analyze the abnormal characteristics of regional electricity consumption within the distribution network area, extract electricity consumption data that does not meet the threshold of electricity theft behavior, and investigate one by one to obtain the detection results of electricity theft behavior. The experimental results show that the detection results obtained after the application of the proposed method exhibit high accuracy, with an average of 0.956, meeting the practical needs of normal operation of the distribution network.