Digital twin-based prediction of vortex-induced vibration of a twin-box bridge deck within the lock-in region

Abstract

Wind tunnel tests and computational fluid dynamics (CFD) simulations are the two most common methods for investigating vortex-induced vibration (VIV) of a bridge deck at its design stage. However, current practice and research indicate that the two methods have their own uncertainties and drawbacks. This study thus integrates the two methods to develop global digital twins for investigating the VIV of a bridge deck section within its lock-in region. A twin-box bridge deck is first tested in a wind tunnel as a physical entity, and the relevant VIV information is collected. A virtual model for the physical entity is established using CFD simulation. The virtual model is then mapped with the physical entity through fusing the measurement data from the physical model with an optimization process, thereby establishing a local digital twin of the deck section for a given wind speed. Global digital twins are subsequently created based on local digital twins and the Kriging interpolation. The local and global digital twins are employed to investigate and eliminate blockage effects, and to predict the VIV of the deck section at different damping levels and wind speeds within the lock-in region. The results obtained demonstrate that the developed digital twins can provide more information and more accurate prediction of VIV of the deck section, compared with that provided by wind tunnel tests or CFD simulations alone.