Expanding EV Charging Networks Considering Transportation Pattern and Power Supply Limit

Abstract

For large-scale adoption of electric vehicles (EVs), a charging network is needed near the road. The charging station planning and deployment problem should consider the increasing penetration ratio of EVs over a long period of time, and the highly dynamic and location-dependent demands and power grid constraints. This paper focuses on the dynamic charging network design, i.e., how to optimize the charging station locations and the number of chargers in each station at different time stages with an increasing EV penetration ratio. For each candidate location, we first model its coverage area to estimate the dynamic EV charging requirements. Then, we formulate the problem at each time stage as profit maximization, which is a mixed-integer optimization problem. To make it tractable, we investigate the profitability of candidate locations and derive their upper and lower bounds on the expected profit. Then we take two steps to transform and relax the problem to convex optimization. A fastconverging search algorithm, named RMCL-E, is proposed. Using real vehicle traces, simulation results show that the proposed algorithm can make a good trade-off between the service blocking probability and the construction cost to maximize the total profit, which is attractive for charging service providers.