IABSE TG 3.1 Benchmark: Nonlinear Buffeting Simulations vs Full-Scale Hardanger Bridge Measurements

This paper completes the IABSE Task Group 3.1 benchmark framework by validating state-of-the-art numerical aeroelastic simulations against full-scale measurements from the Hardanger Bridge (main span 1310 m). A 5-minute non-synoptic wind record from Storm Tor (29 January 2016) is selected because it exhibits pronounced low-frequency variations of wind speed and angle of attack, conditions under which aerodynamic nonlinearities become important. Three independent time-domain numerical procedures are applied using a common set of structural and aerodynamic input data, including static force coefficients, flutter derivatives, and aerodynamic admittance functions. All approaches, in a general sense, adopt a band-superposition strategy: the low-frequency response is governed by a corrected nonlinear quasi-steady formulation, while the high-frequency response is obtained from a linearized unsteady model about the evolving low-frequency configuration. The turbulent wind field is reconstructed from bridge-mounted anemometers with corrections for deck-induced flow interference and structural motion. Simulated midspan accelerations and derived displacements are compared to measurements through time-history evaluation and quantitative metrics. The study identifies key drivers of agreement, including treatment of low-frequency wind fluctuations, sensitivity to modal frequencies, and uncertainties in vertical wind reconstruction, while showing limited influence of structural damping at high wind speeds.