Rapid climate change and the corresponding estimated sea level rise can affect the performance of the coastal defense structures such as breakwaters, seawalls, and dikes. In order to improve these coastal defenses, a detailed understanding of the processes which contribute to wave run-up and overtopping over the coastal defenses needs to be established. Following the exponential growth of computing capacity around 1970’s, a wide variety of computational models were developed to study fluid flow. Traditionally, three computational paradigms have existed in order to study wave transformation and surf zone hydrodynamics: phase averaged models, phase resolving models, and Computational Fluid Dynamics (CFD) models. Limitations posed by the underlying linear wave theory in phase averaged and other simplifications in the phase resolving models, may not provide sufficient detail in wave breaking, wave energy dissipation, wave run-up, wave overtopping, and potentially other detailed hydrodynamic processes. This lack of resolution in depth averaged models for wave-breaking, wave run-up, and wave overtopping processes motivates a detailed investigation using CFD based models, which can correctly mimic wave-breaking and other hydrodynamic processes.

The recent growth in available computational capacity has greatly improved the applicability of CFD based models for large scale transient flows such as waves near a coast. Additionally, the developments in wave generation and wave absorption boundary conditions by Jacobsen et al. [2012] in the open-source CFD toolbox OpenFOAM, have facilitated the use of OpenFOAM in coastal engineering applications. This encourages investigating the coastal environment using relatively complex models, thus providing insights into fundamental processes which contribute to coastal safety. To that end, this thesis focuses on investigating wave overtopping and the underlying processes which contribute to the aforementioned hydrodynamic aspects.

Overtopping demands accurate capture of the free surface (interface between water and air). The waveFoam solver suffers from numerical diffusion of the interface, consequently requiring a different approach to mimic the sharp interface. In order to cater to this deficiency, a new solver which combines the capabilities of waveFoam [Jacobsen et al., 2012] and isoAdvection [Røenby et al., 2016] which has the ability to capture sharp interfaces by means of a sub-grid approach has been integrated (waveFlow) and used in this study. In addition to the new solver, a new set of Reynolds Averaged Navier-Stokes (RANS) closures developed by Larsen and Fuhrman [2018] for wave modeling applications have been employed to correctly capture turbulence levels under breaking waves. The preliminary steps include calibrating and assessment of the newly integrated waveFlow solver. Using a relatively simple conceptual test case, a comparison of the free surface behavior and overtopping discharge was carried out. This calibration test was followed by a comparison of numerical results with the experimental investigations carried out by Ting and Kirby [1994]. Following this benchmarking study, experimental studies carried out by Flanders Hydraulics investigating wave overtopping over dikes in shallow foreshore environments was validated. A comparison of waveFlow and waveFoam was made to assess the qualitative and quantitative differences between the two interface capture methods on overtopping. Using this new solver, OpenFOAM was able to reproduce the surface elevation and significant improvement in the overtopping results were obtained for identical model setup in comparison to the waveFoam solver. A coupled approach using a potential flow solver named OceanWave3D aided simulation of large domain wave propagation and helped to cut down the computational time.

Coastal facilities such as desalination plant or a nuclear power plant need a continuous discharge of salty water to carry out their functions. Hereby an intake structure, such as a velocity cap, can be used to take in water from the sea. These intakes are open seafloor founded constructions mounted at a water depth ranging between 10 and 20 meters. In order to design such an intake cap in an offshore environment, the action of waves and in general of unsteady flows is the most important design load to be accounted for. This thesis has the aim to investigate the nature of the forces and turning moment acting on the structure in presence of waves and to provide tools, in the form of hydrodynamic coefficients, that can be used to compute the design loads. The analysis is based at first on experimental records collected during a previous campaign including force measurements and PIV recordings. The measurements are then used to validate a CFD model in OpenFOAM. Structure-induced turbulence is shown to be fundamental to define the total loads on the structure in the numerical model. The use of a turbulence closure is in fact observed to be needed in order to come to a validation of the CFD model. Even if the flow separation around the cap is not always accurately predicted, the peak of the inline force is estimated by the model with an error of 8%. In the case of the vertical load the error observed reaches up to 15% but part of the mismatch is attributed to a bias in the experimental records. The CFD tool is then used to generate additional test cases on solitary waves and regular waves in order to expand the scope of this research. The hydrodynamic force coefficients are defined fitting both the experimental records and the load estimates of the numerical model by means of the weighted least squares method. The inline force characterization follows the theory of the Morison equation which is shown to provide a good fit in all analyzed cases. Good agreement is found between numerical and experimental results with regards to the estimate of the inertia coefficient while the numerical estimate of the drag coefficient is up to 19 % lower than the experimental estimate. Vertical force and overturning moment signals are originally fitted with the most common formulas used in literature which however produced poor fits to the force and moment signals. New equations are suggested therefore to come to a better characterization of these loads. The best fits for the vertical force are found with an equation that includes the effect of the horizontal drag, the vertical drag and of the vertical inertia, while in the case of the turning moment the best fits are obtained with a combination of horizontal drag, vertical drag, horizontal inertia and vertical inertia.

In the Rhine-Meuse Delta in the Netherlands, deep scour holes can develop very suddenly. The holes may form a potential risk for the stability of surrounding riverbanks, dikes, bridges, tunnels and buildings. A scour hole can mostly be the result of two conditions: local changes in hydrodynamical conditions or local changes in erosive capacity of the bed. Heterogeneity of the subsoil in the Rhine-Meuse Delta causes differences in erosive capacity, resulting the formation of deep scour holes.

A distinct method to predict the development of scour holes in heterogeneous subsoil is not found yet. In this research a numerical model that simulates the hydrodynamics in a scour hole is used to gain more insight in the flow processes that are present in such a hole. The dependencies of the flow processes on the upstream water depth, upstream flow velocity and depth of the scour hole are investigated by changing these parameters systematically in the model simulations. The processes that are included in the analysis are the flow separation and recirculation zone, the contraction of flow, the horseshoe vortex and the vortices downstream of the hole.

Changes in the depth of the scour hole showed some interesting phenomena. It is found that three-dimensional processes suppress the formation of the recirculation zone, as flow separation happened for milder slopes and shallower holes in two-dimensional simulations. The contraction of flow is not necessarily stronger for larger holes, which was hypothesized.

The analysis of the bed shear stress in the model results shows a constant upstream slope angle, as the critical value for bed shear stress is not reached. In the downstream half of the hole, this value is reached and therefore erosion and sedimentation are expected. The bed shear stress in the model results shows strips of zero bed shear stress with large gradients at both sides of the hole. These strips have an angle of 1:8, which is also found in the laboratory experiments by Koopmans (2017). The downstream expansion of the scour hole with a poorly erodible top layer had the same widening angle.