Multiobjective Electric Vehicle Charging Station Siting Under Traffic and Grid Capacity Constraints

ABSTRACT Siting choices for electric‐vehicle charging stations propagate through both the road network that delivers vehicles to each candidate node and the distribution feeder that supplies the load, so a plan optimized for one side may silently degrade the other. Formulate the siting problem as a constrained multiobjective program that minimizes the annualized cost, unserved flow‐captured demand and one‐sided undervoltage, and enforces incremental link saturation along with substation and voltage limits in a single optimization. The traffic side was resolved through a published user‐equilibrium background with charging‐trip detour, pile‐capacity, queue‐cost, and link‐time feedback, the grid side through six‐block backward–forward sweep on a radial feeder, and the two layers were coupled through a node‐to‐bus correspondence and time‐block load and modular‐storage balances. NSGA‐II and NSGA‐III returned a pooled Pareto front on coupled IEEE 33‐bus and Sioux Falls benchmarks, from which a normalized ideal‐point compromise was drawn. The pooled front had 200 designs; the compromise installed 12 piles at three sites, served 64.39% at $897 158 year −1 , maintained a minimum voltage of 0.9640 p.u., and loaded the substation to 5.5536 MVA. In a matched high‐service layout, 600 kWh of storage reduced peak loading from 6.0085 to 5.8977 MVA and restored feasibility. The traffic guard remained inactive in the base case and became active only under extreme road‐capacity stress.