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

Abstract

Electric Private Car (EPC) charging load forecasting 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, and rarely analyze the differences in charging loads in different day types and seasons in future years. As the coupling between operation and planning is increasing, 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 timescales. Therefore, a multi-time scale charging load forecasting method for EPCs based on an improved gravity model that considers stochastic charging behavior is proposed. First, a consumer state decision model based on the Markov chain is construct to deduce the EPC ownership, as the basic data for charging load prediction. Second, based on the travel chain model and the improved gravity model, the user travel behavior for different typical days, seasons and years in each region and each time period in the future is described. Then, a stochastic charging behavior model based on the Hybrid Choice Model (HCM) is constructed. This model is used to simulate the stochastic charging behavior that takes differentiated individual attributes and attitudinal latent variables into consideration, and 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 EPC ownership, the model proposed in this paper can effectively predict the charging load of different regions at multiple time scales during 2024–2030.