Abstract
The spatial concentration of large-scale electric vehicle charging demand at local hotspots limits the flexibility of clusters, while transformer capacity constraints further restrict their actual dispatchable boundaries. To address this, this paper proposes a spatio-temporal coordinated dispatch framework for electric vehicle clusters based on dynamic congestion pricing. In the spatial domain, a congestion penalty multiplier linked to the real-time number of vehicles at a station is introduced and embedded into a generalised charging station selection cost function to naturally disperse demand toward to stations with lower loads; In the temporal domain, vehicles are flexibly classified into three tiers based on parking duration and state of charge; following rigid decoupling, flexible clusters are aggregated into generalised energy storage via Minkowski summation to precisely extract the dispatchable boundary. The coordinated optimization model incorporates DistFlow distribution network constraints, establishing a multi-objective function that accounts for economic arbitrage, load smoothing and multi-asset depreciation. Coupled simulations based on the Sioux Falls road network and the IEEE 33-node distribution network demonstrate that, compared to time-only coordination, the proposed spatio-temporal coordination strategy reduces total electricity procurement costs by 12.7% and lowers the inter-station Gini coefficient at peak times by 18.7%. For the specific Sioux Falls test system, the best system performance was observed at a dispatch radius of 1.0 km among the tested values. The proposed strategy minimises the system’s total operating cost over time and effectively reduces the system’s global peak-to-off-peak ratio; spatially, it reduces the inter-station Gini coefficient at peak times, achieves local load balancing, and unlocks the flexibility potential of vehicle-grid interaction.