PSI - Issue 62

Tommaso Lazzarin et al. / Procedia Structural Integrity 62 (2024) 625–632 Lazzarin et al./ Structural Integrity Procedia 00 (2019) 000 – 000

629

5

Fig. 3. Mean streamwise velocity component in a vertical section in between piers P2 and P3 ( y = 0m) for the PF regime simulation with DES (a) and RANS (b), and for the corresponding simulation in FS regime (c).

Several studies showed that LES and DES simulations are more accurate than RANS simulations in predicting complex flow patterns, such as those forming at bridge piers (e.g., horseshoe vortices, wakes, etc.) and in natural, deformed bathymetries (Constantinescu et al., 2011; Ettema et al., 2017; Kang et al., 2011). In the present PF case, the mean flow field predicted by the RANS simulation is in good agreement with that provided by the DES approach. The high-velocity orifice flow beneath the deck and the recirculation regions downstream of the bridge are also in good agreement, although with some small differences (e.g., compare Fig. 3a,b). Large discrepancies emerge when considering the instantaneous flow fields. As expected, the dynamics of the vortex shedding at the piers wakes and just downstream of the deck where the flow expand are not well captured by the RANS solver. Close inspections of the horseshoe vortices developing at the piers also reveal discrepancies between DES and RANS results. It can be concluded that the RANS simulation fails in correctly reproducing smaller scale unsteadiness, especially if associated to coherent structures developed in case of strong pressure gradients and/or flow separations, as already demonstrated in literature (e.g., Paik et al., 2007). The results of the CFD simulations can be profitably used to assess the hydrodynamic actions on the bridge structure, by integrating the flow pressure computed on the different parts of the bridge. While previous studies evaluated the hydrodynamic loading on the submerged deck using laboratory experiments, simplified formulations, or CFD models applied to 2D vertical slices (see e.g., Malavasi and Guadagnini, 2003; Pregnolato et al., 2022), DES simulations explicitly simulate the turbulent eddies, thus allowing to monitor the temporal variation of the force components (e.g., see Fig. 4 that refers to the longitudinal component of the force, F x , acting on the submerged bridge deck). Besides providing information for each single structural element, the results also allow to obtain the time-varying spatial distribution of the flow action, which can be used as input values for numerical studies of the dynamic response of the bridge structure to space-and time-varying hydrodynamic loads.

Fig. 4. Longitudinal drag force component on the bridge deck for the case of flow overtopping simulated with the DES approach.