Validation of multi-fidelity urban flow numerical model for innovative air mobility against wind tunnel experiments in two facilities

The secure integration of Unmanned Aerial Systems (UAS) in urban environments highlights the need
for precise characterization of wind patterns within realistic complex models. This study evaluates the
differences in validation data between two wind tunnels for numerical models that support Urban Air
Mobility (UAM). Two independent datasets in a LiDAR/DTM-derived complex urban scenario supply
velocity and surface-pressure time-averaged measurements to validate multi-fidelity numerical results.
Results show that LES improves the RANS time-averaged velocity results, reducing the normalized
root mean square error from 24% to less than 10%, minimizing the maximum local error to 30%
and outperforming RANS in over 70% of measurement locations. Although LES resolves unsteady
wake structures and downstream peaks in turbulent kinetic energy, the employed numerical setup does
not fully reproduce experimental velocity fluctuations. Despite inflow sensitivities and differences in
turbulence and blockage ratios, normalized surface pressures are highly reproducible across facilities,
also showing strong agreement between experiments and CFD simulations. Flow diagnostics reveal
distinct zones with different risk signatures depending on local urban geometry and wake interactions.
Areas with high-rise, low-density buildings produce strong shear and increase peak velocity and
turbulent kinetic energy, which can create hazardous conditions for UAS operations. Results suggest
that future UAM studies should validate numerical turbulence spectra and integral length scales at the
inflow to ensure realistic reproduction of facility conditions. The obtained dataset enables follow-up
work analyses targeting specific UAS mission requirements.