EV Charging Load Forecasting Method Considering Time-of-Use Price and Meteorological Factors

The randomness of load fluctuations at electric vehicle (EV) charging stations presents a significant challenge for the accurate load forecasting. Considering time-of-use (TOU) price and meteorological factors, EV charging load forecasting method for EV charging station is proposed. First, the factors affecting charging load are analyzed using the Pearson correlation coefficient and the maximum information coefficient (MIC). Key factors such as TOU price, temperature, weekday type, and recent historical load data are determined. Next, convolutional neural network (CNN) is used to extract important features and capture complex nonlinear relationships from key factors. Finally, the extracted features are processed by gated recurrent unit (GRU) network to forecast charging load at different time scales. Case study is conducted using load data from EV charging station with a capacity of 1800 kW. The results demonstrate that the CNN-GRU model significantly improves forecasting accuracy at 4h, 24h, 72h, and 168h forecasting lengths compared to other models.