Identification Method of Electricity Theft Based on DBSCAN and Pearson Correlation Analysis

In order to solve the problems of limited application scenarios and low adaptability of existing identification methods, this paper proposes an electricity-theft identification method based on Pearson correlation analysis. Firstly, users are divided into several sets with similar voltages according to their voltage values, and each set forms a virtual node. Then, according to the principle of negative correlation between virtual node voltage and load power, the virtual node for electricity theft is identified. Finally, the users of the identified virtual node are eliminated one by one to judge whether the user steals electricity according to the increase or decrease of the negative correlation between the voltage of virtual nodes and the load power. The effectiveness of the proposed method is verified by an example based on actual electricity consumption data.