EV Fleet Charging and Transformer Thermal Limits

What the model tests

A representative 75-kVA oil-immersed service transformer serves a ten-vehicle depot. Each vehicle needs 40 kWh overnight on an 11.5-kW Level-2 charger; vehicles arrive between 17:00 and 19:00 and leave between 05:00 and 07:00, giving an overnight dwell. Four scenarios are compared. S1 is the site base load without EV charging and serves as the reference. The three charging scenarios deliver identical fleet-charging energy and differ in when that energy is drawn and in the mitigation used: unmanaged charging that begins at plug-in (S2), managed charging that charges each vehicle at the lowest constant power that completes its 40 kWh within its dwell (S3), and managed charging paired with a 30-kW / 120-kWh battery that discharges only when loading would exceed nameplate (S4).

The transformer's operating envelope is evaluated with the IEEE Std C57.91-2011 dynamic loading model, which relates per-unit load and ambient temperature to top-oil and hot-spot temperature and, through an Arrhenius relationship, to equivalent insulation loss-of-life. Because S2–S4 deliver the same charging energy, the differences in thermal stress among them come from charging timing and the mitigation used.

Results

Unmanaged charging pushes the transformer to 170.7% of nameplate and holds it above rated load for 3.75 hours a night. The hot-spot peaks at 161.4 °C, far past the 110 °C normal-aging reference, so the insulation ages about 8,373 times faster than baseline: roughly five days of transformer life used up each day. (The ratios are large because the no-charging baseline ages only about 0.014 hours per day.) This deliberately severe scenario uses near-simultaneous plug-in, no diversity beyond the arrival window, and a constant 30 °C ambient to examine high-stress conditions in the model.

Spreading the same energy across the dwell cuts the peak by 61%, from 170.7% to 66.4% of nameplate, and eliminates rated-load overload entirely. The hot-spot falls to 71.0 °C and equivalent aging drops to 0.129 hours per day, about 9× the no-charging baseline and far below the normal-life rate of 24 hours per day. At this fleet size, spreading the charging overnight is enough to keep the transformer within its thermal limits.

Adding storage changes nothing at this depot size. Because managed charging already holds loading below nameplate, the battery has no peak left to reduce. Until the fleet grows, storage is optional.

Fleet size and storage

Three fleet sizes were tested (10, 15 and 20 vehicles), changing only the fleet size. At 15 vehicles, managed charging peaks at 91.4%, still under nameplate, so the battery again stays idle. At 20 vehicles the result changes. Managed charging alone reaches 115.8% and 10.5 hours a day above rated load, which fails the overload screen. Adding storage caps loading at 100.0%, removes the time above rated load, and cuts daily loss-of-life from about 1,222× to 447× baseline. These results apply to the three tested sizes under the modeled managed-charging policy; no threshold is inferred between 15 and 20 vehicles.

Schedule charging first. Add storage only when scheduling alone can no longer keep the transformer within its limits.

Why it matters

The depot transformer is a small, fully modeled example of a wider grid problem: concentrated new load arriving faster than the infrastructure built to serve it. At the distribution edge, the binding constraint can be the transformer's loading and thermal limits. At the transmission edge, it is the time-to-power gap between an approved project and an energized one. FERC's large-load interconnection rulemaking considers faster studies for large loads that agree to reduce demand when needed. It is the same approach, applied to transmission.

A data center, an electrified plant, and an EV depot differ in size but have the same parts: a large load in one place, equipment with a physical limit, and some flexibility in when power is used. In the modeled 10- and 15-vehicle cases, spreading charging across each vehicle's dwell kept transformer loading below nameplate without using transformer temperature or loading feedback. The broader engineering implication is to coordinate flexible loads within equipment limits. Charging schedules based solely on energy prices can concentrate demand and increase transformer thermal stress.