Data-driven equitable placement for electric vehicle charging stations: Case study San Francisco

Abstract

California electric vehicle (EV) regulation on phasing out new gasoline vehicles by 2035 and government subsidies are important steps toward the wide adoption of EVs. However, the lack of EV charging infrastructure with their inequitable distribution is still the most severe obstacle to widespread EV uptake while the charging infrastructure continues to expand across the US. This paper proposes a novel data-driven placement and sizing of charging stations in San Francisco, by quantifying social equity access, EV charging demand coverage, and site development costs including construction, operating, and installation costs. To this end, first 200 possible EV charging stations, as an initial generation, are located in San Francisco. Then, the optimal size, charging type, and locations of charging stations are adjusted/identified resulting from the trade-off between minimized site development cost, maximized social equity access, and EV charge demand fulfillment. The placement, charging type, and sizing of EV charging stations are formulated as a multi-objective optimization (MOO) problem and solved by the Non-dominated Sorting Genetic Algorithm-II (NSGA-II). Ultimately, the Technique for Order of Preference by Similarity to the Ideal Solution (TOPSIS) is applied to obtain the optimal solution from the Pareto-optimal front. According to the balance between minimizing site development costs, maximizing EV demand coverage, and ensuring the highest level of social equity access, energy service providers and charging station owners can strategically decide to place charging stations in San Francisco's middle, west, and northwest regions.