Spatiotemporal charging demand models for electric vehicles considering user strategies

As the number of urban electric vehicles continues to increase, accurate prediction vehicle (EV) spatial and temporal distribution charging demand is great importance for safely operating power grid. Due uncertainty variability EV user discharging strategies, strategic factors behind behavior become key influencing whether results are reasonable. a result, this paper proposes model based on real-time data from Baidu map that can interpret driving strategies strategy learning capability generative adversarial imitation learning. This first analyzes correlation between SOC in data, then describes establishing 24-hour single vehicle, finally discusses building spatiotemporal region basis. The demonstrate that, while it be combined with traffic method has better accuracy robustness compared current mainstream methods high application value.