Proposal for Benchmark Problems to Assess Spillway and Overtopped Dams Structural Stability Using CFD

Extreme floods are one of the most significant threats to the structural stability of hydraulic gravity structures. The water flow conditions, including the related pressure fields and resultant forces, of overtopped gravity dams and spillways are difficult to accurately quantify. Recent advances in CFD (Computational Fluid Dynamics) have made this approach an attractive alternative to physical models. However, prior to using CFD with confidence one has to establish its accuracy in terms of (i) the computational domain and boundary conditions, (ii) the mesh type and refinement, and (iii) the turbulence model. Four benchmarks are proposed herein to evaluate the accuracy of CFD models to assess the condition of overtopped hydraulic structures. The first benchmark compares the nappe trajectories and pressure fields with empirical data from standard ogee spillways. A minimum level of mesh refinement is necessary to achieve convergence of the vertical and horizontal components of the hydrodynamic loads. The second benchmark uses a rectangular crest spillway to study the aeration of a free-falling nappe. Two methods to achieve aeration are compared: adding a pressure outlet to the downstream wall of the spillway in a 2D model and broadening the domain on the downstream portion of a 3D model. Both strategies accurately capture the trajectory, streamlines, and pressure fields compared to experimental results. However, the 2D strategy, while being less computationally expensive, requires prior knowledge of the aeration boundary location. The third benchmark evaluates the accuracy in modelling the profile of a free-falling water stream by integrating the volume fraction of a control volume to obtain its total water mass. That is then compared against an analytical solution. An adaptative refinement around the interface between water and air is required to converge to the analytical solution. The fourth benchmark combines the lessons learned from the previous three into the problem of an overtopped gravity dam section. The numerical results are compared against analytical and empirical estimates. Adequate mesh refinement and boundary conditions are necessary to capture the hydrodynamic forces. By solving these four benchmarks, one would be able to adjust their CFD models to confidently simulate overtopping of hydraulic structures.