A methodology for comparing mean, fluctuating and peak wind loads of buildings

This paper proposes a methodology to distinguish true measurement uncertainty from aerodynamic effects when comparing load coefficients from different atmospheric boundary layer wind tunnels. It considers the similarity of the wind field through profiles of mean velocity, turbulence intensities, and gust factor, along with the distribution of fluctuating flow properties, especially at small scales of turbulence. To ensure consistency, peak wind velocities and responses are estimated from time-histories matched in full-scale sampling time, hence longer records are truncated to match shorter ones. A test case involving a pressure model of a medium-rise building is proposed. It was independently tested by RWDI, CPP, and Western University under five different conditions. Time-histories of wind velocity and integrated aerodynamic base shear force, overturning, and torsional moments are analyzed and compared for nominally similar exposures. The trends in two comparisons are qualitatively consistent, with discrepancies in mean and peak coefficients not exceeding 7 % and 14 %, respectively. The analysis of the alongwind response reveals even smaller differences, especially in the mean coefficients, even across all five conditions. These findings suggest that current wind tunnel testing standards could potentially be relaxed, particularly by incorporating Partial Turbulence Simulation concepts, without compromising the reliability of aerodynamic load predictions.