Large-scale spatiotemporal modeling of travel behavior, electric vehicle charging demand, and flexibility based on human mobility big data

The rapid growth of electric vehicles (EVs) and renewable energy has increased the importance of detailed power system planning at multiple scales ranging from local distribution networks to macro-level generation capacity and operations. To maximize the contribution of EVs to decarbonization and system transitions, high spatiotemporal data on EV charging demand and flexibility applicable at multiple scales would be effective. Previous research has established useful modeling approaches. However, limitations to expressing macro and local characteristics simultaneously remain. To address this deficiency, this study developed a data-driven workflow to extract vehicle usage profiles from human mobility big data collected from mobile devices to quantify vehicle usage, EV charging demand, and flexibility on a large scale at a high spatiotemporal resolution. The application to entire Japan showed that if all the private vehicles are electric, uncontrolled home charging would reach a maximum of 25 GW. EVs have significant flexibility potential through smart charging and vehicle-to-grid operations. The 1 km-mesh-resolution data showed that activity patterns, demand, and flexibility vary considerably depending on the accessibility to public transportation. The established method can provide high spatiotemporal resolution EV data for various purposes and also allows for iterative studies that integrate local and macro-level power systems.