An identification method of anti-electricity theft load based on long and short-term memory network.

Aiming at the problem of many illegal individuals or units adopting various methods to steal electricity under the current power environment, this paper proposes an anti-theft load identification method based on Long Short-Term Memory (LSTM). First, screen users with electricity theft and abnormal behaviors, analyze the collected voltage, current and electric energy by the electricity theft detection equipment, and use the LSTM neural network algorithm to judge the characteristics of the screened suspect users, and initially determine users with abnormal behaviors. Then comprehensively use GPS equipment to send the comparison results of the host to the master station in real time. The master station only needs to be responsible for a small amount of post-processing and statistical analysis, and intelligent diagnosis based on the characteristics of the electricity theft, so that the user can perceive the electricity theft incident on the spot in real time to avoid the loss of electricity bills, and provide a strong theoretical basis for the intelligent operation and maintenance of power companies. [ABSTRACT FROM AUTHOR]