PNNL

Abstract

The widespread adoption of Electric Vehicles (EVs) in recent years has necessitated the development of an effective charging infrastructure that can support the commuting needs of the population. However, expanding the existing infrastructure is a multi-faceted problem that requires careful consideration of location and capacity of existing charging stations, spatiotemporal distribution of charging demands, commute pattern of EV adopters, capacity of the power-grid, budget constraints, etc. To approach this complex problem, we present a novel data-driven simulation-optimization framework, CHARGE-MAP; with a focus on ensuring meaningful charging experience by individual EV owners. CHARGE-MAP consists of an agent-based simulation module, an optimization module, and a power module. By simulating mobility patterns of EV adopters and their possible charging behaviors, we estimate the spatiotemporal distribution of charging demands. It allows us to identify the locations where the existing infrastructure is inadequate to meet demand. The optimization module determines where to build new charging stations and the capacity of the new and existing charging stations. It ensures favorable charging experience for EV adopters by minimizing the expected average detour distance for charging and the wait-time of EV-adopters, with a small number of new charging stations. Finally, the power module decides how to connect the stations to the power-grid, while maintaining grid stability. To show the effectiveness and scalability of CHARGE-MAP, we use the state of Virginia, consisting of 95 counties and 38 independent cities, as our study area. Our results show that, with 1,305 new public charging stations and 2,164 new public chargers for all of Virginia, CHARGE-MAP is able to meet the charging demand of ~198.6 thousand predicted EVs. It reduces the expected average detour distance for charging by 66% and the expected average wait time at charging stations by 72%, compared to the existing infrastructure. Moreover, our analysis of transformer capacity requirements across Virginia reveals that only 1.8% of existing residential transformers require upgrades for projected EV adoption, while over 80% of commercial charging locations can be supported with modest transformer infrastructure (25-50 kVA). This indicates that targeted rather than wholesale infrastructure investments can facilitate cost-effective EV integration. Consequently, CHARGE-MAP provides policymakers and urban planners with coordinated, data-driven insights, essential for effective expansion of the EV charging infrastructure.