Measured-Data EV Hosting Capacity Assessment andDegradation-Aware V2G Reassessment under SpatioTemporal Charging Uncertainty

Resumen

Reliable electric-vehicle (EV) hosting-capacity (HC) assessment requires the timing and location of charging demand to be represented jointly. This paper presents a probabilistic framework based on measured public charging data and uses it to reassess HC after degradation-aware vehicle-to-grid (V2G) dispatch. Station-hour records are converted into node-hour scenarios by combining empirical daily profiles with scenario-specific spatial allocations. For each scenario, power flow is solved using the backward/forward sweep method, and the results are checked against bus-voltage, transformer, and main feeder current limits. A Stackelberg V2G model is then solved iteratively, and HC is recomputed from the dispatched net load. The case study uses data from 97 representative days recorded at 92 public charging stations in Jiaxing, China. The mean HC is 1.10 MW, with P5 and P95 values of 0.89 and 1.42 MW, respectively. The measured EV peak is about 74% of the mean HC. V2G increases HC by only 2.8% because the measured public-charging profile is relatively flat; applying the same procedure to a sharp-peak benchmark yields a much larger gain. The capacity contribution of V2G therefore depends on load shape, connection location, and the binding network constraint. The framework provides a common basis for charging-station siting and feeder-reinforcement decisions.