Electric Larceny Recognition Method Based on High Dimensional Random Matrix Analysis

Electric larceny is hard to be checked and is important for power check. Based on big data theories, high dimensional random matrix can be built with parameters collected from the power grid as elements. Through characterizing statistical properties of the matrix, electric larceny recognition method based on big data analysis is proposed, which solves such problems existed in the traditional electric larceny recognition method as high cost of manpower, time lag and low accuracy. Therefore, an efficient anti-electric larceny is realized. Taking a 33-bus power grid as anexample, occurrence recognizing, the time determining, the precise locating and the type of electric larceny are realized based on such operation data collected from the grid as voltage and current that vary with time.