Data-driven methodology for identifying vehicle-to-grid parking regions in urban areas

The transition of mobility to electrification resulted in a surge of electric vehicles (EVs) and an imbalance in demand and supply of electricity. In remedy, vehicle-to-grid (V2G) technology has been developed to balance and optimize energy use by integrating EVs and the electric grid. However, its implementation requires understanding the spatial and temporal patterns of demand and aggregation of EVs. In light of that, this study devises a multistage methodology that implements a data-driven approach to identify potential V2G parking zones and develops a localized Poisson model to characterize influence factors of parking duration. A real-case scenario is investigated, and a potential region for V2G is identified based on aggregated parking characteristics through the use of floating car data (FCD). The characterization, in average terms, shows that a unit increment in building footprint area, population, and area of existing parking lots entails a positive but insignificant effect on the expected parking duration. Similarly, an increase in duration is found to be related to the number of: hospitality, entertainment, food and beverage activities; retail, service, and financial centers; education, science, and research centers. On the other hand, the density of roads, a higher number of cultural places, community, and medical facilities are found to reduce the aggregated duration. Overall, the findings are expected to offer a valuable resource for parking infrastructure planners and grid operators, and contribute to the small pool of literature on V2G vehicle aggregation.