Electric Theft Behavior Detection Method Based on Power Customer Data Analysis

This paper presents a detection method of electricity theft based on SOM neural network and K-means clustering algorithm. This method combines the advantages of SOM neural network which can automatically classify and K-means clustering algorithm which has good application effect. In addition, considering that a single similarity index may affect the detection accuracy, in order to accurately determine the similarity of load curve, an improved algorithm of abnormal degree of load curve is proposed by weighting the Euclidean distance and cosine distance, and a diagnostic process for suspected users of electricity theft is established. The simulation results based on the measured data of the power grid show that the proposed method can accurately determine the theft behaviour.