Simulation of vortex shedding and vortex induced vibrations in multibox bridge decks using CFD

Abstract

[Abstract] In long-span bridge engineering the designer must consider a wide range of economical, environmental and structural aspects in order to obtain a safe, and economical prototype that fulfils the design constraints and satisfies the needs of the potential users. These give rise to the refurbishing of existing ageing bridges and the proposal of new bridges, which thanks to the advances in design tools, methods, physical theories and models, allow to design long-span bridges in sites where a feasible solution was priorly impossible. Thus, the length of bridges has been growing consistently in the last decades, producing very flexible structures, specially for long span cable supported bridges, where the main design constraint is the wind action. Different wind related phenomena can affect a bridge structure. The most catastrophic one is the phenomenon of flutter, which can cause the collapse of the bridge, as in the famous Tacoma Narrows Bridge incident. This is one of the most studied aeroelastic phenomena, and all bridges must be designed to yield very high critical flutter velocities in order to avoid the occurrence of flutter in built structures. Another aeroelastic phenomenon, considered during the design process, is vortex induced vibration. In this case, the bridge deck is subjected to sustained self-limiting oscillations at reduced wind speed, that may cause serviceability and fatigue related issues. Traditionally the study of aeroelastic phenomena in bridge engineering has been done by means of wind tunnel tests. Nonetheless, CFD simulations are becoming more reliable in recent years. The fundamental goal of this Ph.D. thesis is to assess the feasibility of numerical approaches in industrial applications in long-span bridge design. This Ph.D. thesis is dedicated to the study of the phenomenon of vortex induced vibrations in bridge decks, particularly twin-box bridge decks, by means of CFD numerical simulations. This technique allows the consideration of a large number of design possibilities at the early design stages, as oppose to the wind tunnel tests, although they need of high quality wind tunnel data for validation, as well as important computer power to solve the very demanding associated models. This numerical methodology has been successfully applied to a fundamental case, as it is the ratio 4:1 rectangular cylinder, and the twin-box Stonecutters Bridge, considering the actual design and changes in the gap distance. For the considered application cases, the force coefficients and their standard deviations have been obtained. Also the VIV has been studied, focusing on the amplitude of the response and the lock-in range of reduced velocities. Also, the flow topology and the spanwise correlation of force coefficients and pressures have been studied. In all these cases, the numerical results have been compared with available experimental data in the literature. In this manner, the maturity of CFD techniques, and their feasibility in design-related problems in long-span bridge engineering, have been assessed.