Probabilistic Thermal Design for Electric Vehicles Operating at High Altitude: Guard-Bands and Validation Using Multi-City Reanalysis

Thermal performance of air-cooled electric-vehicle (EV) packs and inverters deteriorates at altitude, yet design guidance remains calibrated to sea-level conditions. This study quantifies the altitude penalty on convective cooling and translates it into practical, probability-based design rules. Hourly meteorology (temperature, pressure, wind) from ERA5/ERA5-Land was extracted for Andean and Pacific/Caribbean cities and coupled to a property-aware 0D/1D thermal model using modern flat-plate Nusselt correlations. A physics-informed scaling for the convective coefficient yields a compact altitude penalty law, ϕ, and a sea-level equivalence mappingUeqthat preservesh. The framework is validated multi-city and embedded in uncertainty (Sobol’) and extreme-value analyses, with dependence-preserving copula sampling. Results show systematically higher thermal severity in the Andes: city-level P95(ΔT) and exceedance of ΔT > 20 K increase, and median guard-band factors for required heat-exchange areaA\*are 20.8 % higher than at the coast. The Ueq mapping reproduces altitude–sea-level equivalence with ~1–3 % mean relative error, enabling sea-level bench testing of altitude operation. Practical design charts reportA\*by city and vehicle archetype (hatch–bus), with linear slopes of +0.04 to +0.14A\*per km. EVT return levels confirm that altitude amplifies not only percentiles but also rare thermal peaks. A fan-sizing layer linksAflow, RPM, and power to maintain ΔT targets across sites. The study delivers reproducible Python tools, validated mappings, and city-resolved guard-bands that convert altitude physics into deployable requirements for EV thermal management and inverter cooling.