PSI - Issue 62

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

627

3

a height of 3.7 m, with jersey barriers 1.0 m high at the extremities. The net height below the deck varies due to the irregular topography. The real bridge shape has been used in the study; however, it must be stressed that the bridge deck is taken 6.5 m lower than in the real case to illustrate the effects of both the free-surface and the pressure flow conditions with realistic values of flow discharge. Accordingly, in the modelled setup, the net height below the deck ranges from 7 m to 17 m. Simulations are performed with a flow discharge of Q = 14,485 m³/s, which corresponds to an event with return period of 500 years. In this condition, a PF regime occurs beneath the bridge deck. An additional simulation has been conducted with the same discharge but removing the bridge deck, so as to simulate the corresponding free surface ( FS ) condition. This allows to investigate the differences associated with the change in flow regime from FS to PF conditions. The STAR-CCM+ CFD suite has been used to solve the Navier-Stokes equations with a finite-volume method. The two-phase Volume of Fluid (VoF) method is used to track the free surface deformation (Hirt and Nichols, 1981), together with a high-resolution interface capturing scheme (HRIC) with sharpening factor equal to 1 (see e.g., Lazzarin et al., 2023c). The model uses the eddy-resolving DES approach for turbulence (see e.g., Lazzarin et al., 2023a) to couple the eddy-resolving LES method away from solid boundaries, and the RANS method near the walls, where the SST k -  turbulence model is used (Menter et al., 2003). An additional simulation has been performed using the RANS method with the purpose of comparing the two numerical approaches. The computational domain includes a 4’300 m long reach of the Po river, in which the bridge is placed at ~ 1’900 m downstream the inlet section (Fig. 1b). Besides the active riverbed, the model includes the floodplains up to the main levees. The ground elevation is taken from the 2015 LiDAR survey, and the riverbed bathymetry from the 2004 multibeam survey. Fixed water levels at the outlet section are obtained from a preliminary simulation performed on a longer reach with a 2D depth-averaged model (Lazzarin et al., 2023b; Lazzarin and Viero, 2023; Viero et al., 2013). A logarithmic velocity profile is assigned for the streamwise water velocity, and a hydrostatic pressure distribution is prescribed at the outlet section. The riverbed is modelled as a rough wall with roughness height of 0.1 m, whereas the bridge structure and the lateral boundaries are modelled as smooth walls. The computational mesh, generated within the STAR-CCM+ framework, is made of ~19 · 10 ⁶ cells. The hexahedral structure of the grid allows for an easy adaptation to the irregular boundaries, and multiple local refinements provide a proper mesh resolution in the bridge region. The mesh size is of about 10 m far from the bridge, and of about 0.6 m near the bridge piers. Close to the solid boundaries, prism layers allowed to further refine the mesh in the wall-normal direction.

Fig. 1. (a) location of the case study in the North of Italy; (b) plan view with topography; (c) sketch of the bridge piers and deck.