Electric Load Prediction of Electric Vehicle Charging Stations Based on Moving Average–Gated Recurrent Unit

Abstract

The load prediction of electric vehicle charging stations is the basis of their static safety, which directly affects the safety of operation, the rationality of planning, and the economy of supply. However, various factors lead to drastic changes in short-term power consumption, which makes the data more complicated and difficult to predict. In this paper, the moving average–gated recurrent unit method is proposed to predict the electric load of electric vehicle charging stations. A prediction model is established based on the historical data of electric load of electric vehicle charging stations to realize the accurate prediction of future electric loads. Firstly, considering the problems of noise in the historical data of electric vehicle charging stations, the moving average method is used for smoothing. Secondly, the smoothed data are modeled by the gated recurrent unit, and the future prediction results are obtained. Finally, the validity and practicability of the proposed method are proved by the research and testing of the actual electric vehicle charging station power load dataset. Compared with the classic LSTM prediction model, the proposed MA-GRU method can achieve more accurate prediction performance.