Electric Vehicles Charging Load Forecasting and Impact Analysis on Distribution Network under Vehicle-to-Grid Mode

In response to the "peak-on-peak" phenomenon in distribution network caused by the disorderly integration of large-scale Electric Vehicles (EVs) into the grid, a charging load forecasting method is proposed under Vehicle-to-Grid (V2G) mode. Firstly, a node coupling model is established based on the spatial location relationship between the power grid and the traffic grid. Secondly, the optimal driving path of the Origin-Destination (OD) matrix is determined using Dijkstra's algorithm to simulate the travel process and power consumption of EVs. The M/M/C queuing model has also been introduced to consider the charging queuing time of EVs. Lastly, in view of the charging demand of different types of EVs, such as electric private cars, electric buses, and electric taxis being the research objects, a spatial and temporal forecasting model is established for the charging load of multi-type EVs. The proposed EVs charging load forecasting model and its impact are analyzed using typical scenarios in the IEEE-33 bus system and the actual power distribution area. The effectiveness of the proposed model is verified through simulations under V2G mode.