A multi-time scale charging load forecasting method for electric private cars based on an improved gravity model considering stochastic charging behavior

Electric Private Car (EPC) charging load forecast is the basis of charging facility planning and long-term development of distribution network upgrading and transformation. However, most current studies focus on short-term charging load prediction, rarely analysing the differences of charging loads in different day types and seasons in future years. With the increasing coupling of operation and planning, it is not only necessary to observe the intraday charging load characteristics, but also to consider the evolutionary trend of the spatial and temporal distribution of charging loads over long time scales. Therefore, A multi-time scale charging load forecasting method for electric private cars based on an improved gravity models considering stochastic charging behavior is proposed：Firstly, combining the predicted value of EPC ownership as the basic data for charging load prediction, based on the travel chain model and the improved gravity model to portray the user travel behaviour in the future for different typical days and seasons in each region and each time period; Secondly, a stochastic charging behaviour model based on Hybrid Choice Model (HCM) is constructed to simulate the stochastic charging behaviour considering differentiated individual attributes and attitudinal latent variables; and then a simulation process for EPC charging load prediction at multiple time scales including typical days, seasons, and years is established; Finally, using the travel data of various administrative districts in Nanjing and analysing the interaction between EPC ownership and vehicle travel and charging behaviour, the model proposed in this paper can effectively predict the charging load of different regions at multiple time scales during 2024-2030.