Grass-covered dikes are a widely used measure for flood prevention. Overtopping waves can cause erosion that poses significant risks for flood safety, especially with rising sea levels and increasing storm intensities. During storm events, which involves frequent wave overtopping, erosion occurs progressively as multiple individual waves flow over the dike. This causes the dike profile to change over time as erosion holes begin to appear, which can influence the erosion potential of subsequent overtopping waves.

Current methods to estimate forces and erosion from overtopping waves rely on numerical models or analytical and empirical formulas. Numerical models are generally applicable and provide detailed spatial and temporal results, but are computationally too expensive to simulate an entire storm whilst updating the bed profile. Formulas offer quick results but lack general applicability and fail to account for the effects of erosion holes. No current method efficiently combines computational speed, detailed results and broad applicability, making it challenging to study the effects of gradually increasing erosion on overtopping waves.

This research focuses on the development of a new method that provides rapid and reliable wave overtopping simulations that can integrate erosion. This is done through surrogate modelling, which aims to create a computationally cheap model that emulates a detailed model. The foundation of this surrogate model is a Transformer-based deep learning architecture, which has proven superior in handling spatiotemporal processes.

The surrogate model is created by adapting the Vision Transformer model into a new model that can perform next-frame prediction, which involves the prediction of a next frame based on a sequence of input frames. The developed model can take an input sequence, recognize spatial and temporal patterns, and project them into a predicted future timestep. The model is trained on a dataset of overtopping wave simulations produced by the CFD software OpenFOAM. A masked loss function is applied to enhance the training process by forcing the model to focus on improving the relevant errors.

Using the trained model to generate wave overtopping simulations showed that it can accurately replicate the original CFD simulations. The surrogate model is validated against the original CFD simulations by comparing maximum values and time series for flow velocity and water depth at four different locations along the dike. The results generally show good agreement in capturing the maximum values, as well as the time of arrival and the overtopping duration. Simulating wave overtopping over an eroded dike profile showed promising results, though performance could improve with a larger and more diverse training dataset.

This research demonstrates that a Transformer-based surrogate model can effectively emulate wave overtopping simulations produced by CFD software. The surrogate model's speed and simplicity enables the simulation of a storm and updating of the bed profile. The model has not reached its full potential due to limitations in the training dataset and nuances in the simulation technique that require further refinement. However, it serves as a proof of concept that this surrogate model can provide a new tool in wave overtopping modelling, creating possibilities for new studies such as the influence of erosion on overtopping waves during storm conditions, or for probabilistic calculations that require a large number of simulations.

The research, "An Integrated Approach to Identify Sustainable Materials and Strategies to Achieve a Carbon-Neutral Breakwater Design," aims to design breakwaters that maintain structural integrity and functional performance while minimizing environmental impact. The research addresses the urgent need to reduce carbon emissions in coastal engineering, aligning with global climate change mitigation goals.
The primary objective is to develop a comprehensive framework for designing carbon-neutral breakwaters...

Geotextiles are widely used in civil engineering, particularly in hydraulic engineering where they serve as filters and protect sand and clay from erosion. Their primary function is to prevent the washout of fine materials from the subsoil. Geotextiles are permeable fabrics and are usually made from plastics like polypropylene and polyethylene. However, due to environmental concerns surrounding synthetic geotextiles, there is a growing interest in moving away from plastics towards natural fibers. To facilitate this transition, understanding the engineering properties of these materials is essential. This study focuses on sandtightness, aiming to make applications more feasible.

Geotextiles can reduce the total thickness of a granular filter and thereby the costs. Because a geotextile can replace multiple granular filter layers, less material is required, which also results in lower emissions. Filters can be distinguished into two categories: geometrically open and geometrically closed filters. Sand particles cannot pass through a geometrically closed filter. The sandtightness is independent of the magnitude of the hydraulic load. Open filters prevent the erosion of the base material up to a certain hydraulic load. Open geotextile filters are not common used in hydraulic engineering.

Previous research has primarily focused on synthetic geotextiles, developing formulas to predict critical filter velocities based on their properties ((Van Der Knaap et al., 1986) and ( (Klein Breteler, 1988)). In contrast to these synthetic-focused studies, a study by Lemmens in 1996 on natural geotextiles, such as jute cloth, revealed a critical hydraulic gradient of 0.26, suggesting jute's potential as an alternative to synthetic materials. Despite these findings, stability criteria for natural geotextiles remain undefined, and newly developed geotextiles from natural materials have not yet been tested.
To address this gap, experiments were conducted in a flume set-up at the Fluid Mechanics Laboratory at TU Delft, using both woven and non-woven geotextiles made from natural materials such as jute, hemp, and wool. In the flow flume, a steady current was applied to a one meter long stretch of rock, closed at the top, with a geotextile-covered sand bed beneath it, allowing hydraulic gradients up to i = 1 to be exerted. This set-up enabled the determination of the critical load.

A total of 19 tests were conducted with 11 different configurations, involving variations with 7 different types of geotextile (4 woven jute geotextiles and 3 non-woven geotextiles made of hemp, jute, and wool). During all test steps of the 19 tests in the flow flume, measurements of water levels, pressures, and discharge were taken. Two endoscopes were installed at different locations in the filter layer, with holes drilled into two stones to secure the endoscopes. This allowed the camera to be positioned at the top of a pore, providing a view of the geotextile and detecting passing sand grains.

To determine the critical test step where the movement of base material begins, an innovative method of analyzing the erosion state was employed using endoscope images. These endoscopes are capable of detecting sand grains, allowing for the observation of erosion dynamics. In 4 of the 19 tests, the eroded sand was also suctioned after each test step to gain insight into the amount sediment transport. This indicates that even before the critical threshold level, there are small amounts of erosion immediately after increasing the hydraulic load. If the critical threshold level has not yet been reached, this erosion will return to zero during the rest of the test step. When determining the critical test step, two values were used for the critical load: a non-erosion criterion (Ho, 2007) and a criterion of 0.2 gr/s/m² (Klein Breteler, 1988), as used in previous studies to determine the start of movement. The non-erosion criterion is qualitative, whereas the Klein Breteler criterion is quantitative. It was found that these criteria are not equivalent, with a factor of 2 to 3 difference in the critical filter velocity.

This study concluded that the critical load for sediment transport is primarily determined by the filter velocity for open natural geotextiles. This contradicts previous studies that suggest a critical hydraulic gradient for open natural geotextiles. The critical filter velocity is influenced by the geotextile's characteristics, including thickness, opening size, and water permeability. Notably, the structure of the geotextile itself significantly affects the critical filter velocity. In this study, two woven jute geotextiles (J4: 422 gr/m² and O90 = 516.1 μm, J5: 518 gr/m² and O90 = 819.0 μm) with different structures were tested. The difference in opening size between the two geotextiles is approximately a factor of 1.5. However, the geotextile with much larger openings had a 1.75 times greater critical filter velocity with a start of movement criterion of 0.2 gr/s/m². Additionally, while the grading and size of the filter layer's grains influenced the critical hydraulic gradient, they had little to no effect on the critical filter velocity, emphasizing the significance of geotextile properties in determining their performance and stability under hydraulic load.

Ultimately, based on the tests, it can be concluded that the newly developed non-wovens could be a good alternative to synthetic geotextiles when considering sandtightness under parallel flow. All three non-wovens are reasonably stable at a hydraulic gradient of i ≈ 1, which is the maximum gradient that could be applied in our test set-up.

Recommendations for future research include testing with coarser sand, longer test durations, and synthetic counterparts for comparative analysis. Suggested improvements to the test set-up involve detailed height measurements of the sand bed and filter layer, and using high-resolution cameras for better sediment tracking. This sediment tracking by an endoscope can be improved by using a fixed pore for the endoscope for all tests. This can ensure that the pore volume is consistent across all tests and the size of the geotextile section captured in the image is also consistent. There is potential to use endoscope images for quantitative analysis of sediment transport. Further, the study emphasizes examining different flow conditions (perpendicular, and non-stationary) to enhance geotextile designs, particularly in dynamic environments like coastal revetments.

Coastal structures such as breakwaters can mitigate the erosive effects of sea level rise by protecting shorelines from wave impact. The Reef Enhancing Breakwater (REB) is a modular, permeable structure designed to dissipate wave energy and boost marine biodiversity by providing suitable habitats. This research investigates the hydraulic stability of the REB under wave loading using a physical model.

 

Experiments were conducted in the Scheldt flume of Deltares with a 1:20 scale model. Both 2 and 3 level structures with simple and complex forms were tested on gravel underlayers of varying roughness. Irregular waves with significant heights up to 12.5 cm and wave steepness ranging from 2% to 4% were used, with varying water depths. As failure was not reached on the horizontal foreshore (main location), the structure was moved closer to the transition slope (alternative location) where plunging breakers can emerge. Smart ReefBlocks with integrated mobility sensors measured accelerations and angular velocities induced by wave motion.

 

Five main types of movements were observed: sliding, rocking, shaking, tilting, and lifting. Sliding was the most common, occurring mainly in the top layer, with a maximum of 4 sliding events per block per 1000 waves. Three limit states were defined: start of motion, start of damage, and failure. Start of motion is when blocks move for the first time, damage is when a block loses interlocking, and failure is defined by a damage level Nod=0.4Nod=0.4.

 

Expected and characteristic values of the stability number (Hm0/ΔDnHm0/ΔDn) were determined for each limit state. Only one displacement occurred at the main location, while failures were reached near the transition slope due to plunging breakers. The 2 and 3 level structures were stable on bed slopes without plunging waves. For non-plunging waves, the start of movement limit state (Ns,mNs,m) ranged from 0.39 to 0.51. When plunging waves occurred, the start of damage limit state (Ns,dNs,d) was 0.94, and the failure limit state (Ns,fNs,f) ranged from 0.94 to 1.11. The start of motion was not determined for setups at the alternative location.

 

For setups at the main location, sliding was analyzed in relation to wave height, wave steepness, water level, and structure height. Higher water levels allowed for higher waves, leading to more instabilities, but excessive submersion reduced movements. Waves with 2% steepness caused more movements than 4% steepness due to higher energy in longer waves. Other factors like underlayer irregularity and increased drag from epifauna were also considered. Most sliding movements occurred for Hm0/ΔDn>0.8Hm0/ΔDn>0.8. The model blocks, made of PLA with lower friction than concrete, started sliding under smaller forces, making the model conservative for sliding movements.

 

The impact of sliding movements on the REB's structural performance was investigated by estimating stresses in the protrusions from data collected by smart ReefBlocks. A conservative model was used to calculate impact forces and check for protrusion rupture. The resulting tensile stresses exceeded the concrete strength, but the model's representativeness is questionable as it differs from the actual protrusion. Further research is needed for a definitive answer.

Coastal areas face increasing threat from erosion. The use of coastal protection structures is imperative to counter this. Concrete “armouring” is widely used due to durability and cost-effectiveness. Conventional methods of “armouring” result in habitat degradation and loss of biodiversity. ECOncrete developed Coastalock, an armour unit designed to provide coastal protection but also to create marine habitats.

Initial tests conducted on impermeable slopes in deep water conditions revealed, that tightly placed units experienced pressure gradients across the top layer, leading to failure. The aim of this study is to investigate the hydraulic performance of Coastalock, both with and without modification, within permeable breakwater structures. The study examines the influence of toe berms on different surfaces, assessing the armour layer’s susceptibility to sliding.

The research aims to bridge existing knowledge gaps regarding Coastalock behavior under varying wave conditions through literature review and physical model tests conducted in the 2D wave flume at TU Delft.

Structure from motion photogrammetry enabled the creation of 3D models of the armour layer after wave attack, facilitating the tracking of armour layer deformation of selected test series. The research includes measurements of overtopping discharge and reflection coefficient.

The findings shed light on failure mechanisms observed in Coastalock armour layers on permeable core slopes, attributed to built-up pressures exceeding self-weight and interlocking capabilities during wave run-down. 'Breathing' involves upward movement perpendicular to the slope during wave run-down and downward movement during wave run-up. Friction and partial interlocking contribute to the formation of a bulge in the armour layer, growing in size and magnitude, leading to extraction.

Observing increased 'breathing' and extraction thresholds with larger inter-unit void sizes, was confirmed for permeable cores. No stability increase compared to impermeable core was found, attributed to reduced maximum run-down levels. A lowering of overtopping and reflection was found for the permeable core.

The protrusions implementation necessitated a new configuration, termed the 'Protrusion Optimized' configuration, with an orientation change from cavity upwards to downwards at SWL. 10% protrusions reduced the stability number threshold for 'breathing', while 22.5% protrusions prevented filter layer migration and reached up 𝑁𝑠=4.2 without 'breathing' or extraction. A stability increase was found in 𝑠0𝑝=0.02 conditions, attributed to the ‘reservoir effect’. Incorporating protrusions and transitioning to the 'Protrusion Optimized' configuration increased reflection due to increased surface area and reduced permeability.

Changing the orientation location towards the midpoint between SWL and the bottom row or facing all units upwards resulted in increased stability in terms of ‘breathing’ and extraction, as well as a downslope shift of the damage location. Upwards-oriented units led to smoother slopes and higher reflection coefficients, both attributed to the water retaining properties of the cavity.

The presence of toe berms, showed no significant impact on damage progression or the location. Downslope movement well below the threshold indicative of near extraction was observed.

Recommendations for advancing 22.5% protrusions are proposed, advocating a Ns=2.6 for surging waves in deep water conditions. This design offers notable overtopping reduction, orientation flexibility, and reduced concrete usage and project duration.

Wave overtopping of coastal structures is generally expressed in terms of average discharge and maximum overtopping volume. While substantial research can be found on the relationship of such variables with incident spectral characteristics and other geometrical dimensions, limited research has been done to identify the conditions for which severe individual overtopping events occur. The objective of the present work is to further investigate the statistical dependencies between individual overtopping volumes and incident wave characteristics and to identify the conditions which result in maximum overtopping events. To this aim, we statistically analyse timeseries of waves and overtopping events for the case of a vertical wall breakwater on a horizontal bed under non-breaking wave conditions. The timeseries are generated using a validated numerical model for 6 sets of boundary conditions. The analysis of the data shows that, among the examined variables, the incident wave crest level at the toe of the structure has the highest rank correlation with the individual overtopping volumes. Conversely, the incident wave heights show a lower rank correlation while the wave periods show no correlation at all. Additionally, by leveraging the tool of copulas we find that the incident wave crest levels have strong upper tail dependence with the individual overtopping volumes while the wave heights show no tail dependence. Subsequently, we examine possible dependencies with the wave group periods and wave group energy. We find that a moderate correlation exists between wave group periods and overtopping volumes but no upper tail dependence. Moreover, our results show that the fraction between the wave group energy and the group period has a high correlation and upper tail dependence with the overtopping volumes. Finally, we discuss possible practical applications of our findings as well as points for future research.

The uncertainty paired with the effects of climate change impacts the design aspects of hydraulic structures. To comprehend the uncertainty of the evolving conditions, methods and models that differ from the traditional ones could provide a better understanding of the uncertainty and potential risks associated with a design.
Hydraulic structures are designed to withstand loading conditions related to extreme environmental scenarios. Such natural hazards are characterized by the shared dependence of several random variables. Multivariate probabilistic design clears the way for exploring dependencies between design variables. A multivariate approach models the joint occurrence of a combined condition, revealing new insights e.g., the correlation between wave height, wind speed and water level. Next to incorporating new models that estimate the uncertainty and changing conditions related to climate change, there could be a leap forward could in design methodology. Adaptive design strategies can provide a way to mitigate over-designing hydraulic structures. To cope with the uncertainty of sea level rise, this methodology aims at handling the uncertainty in long-term decisions by emphasizing adaptivity in a design.
This thesis aims to develop a multivariate probabilistic tool to provide insight into incorporating adaptive design strategies, and to gain extra information that contributes to assessing the feasibility of the Wide Green Dike concept given the uncertainty of sea level rise.
A multivariate model using a vine copula is created to approximate the joint behaviour of the extreme hydraulic boundary conditions of the Wide Green Dike project. The results showed that by accounting for the joint behaviour of the extremes and the uncertainty and variability of design conditions, significant reductions in the design could be realised compared to designing in a deterministic manner. The deterministic design increased the cross-sectional area of the dike by 150% compared to the multivariate probabilistic design with a sight year of 2150.
To explore the adaptive design strategies, several strategies are created using the multivariate probabilistic design approach. The strategies consist of an initial design with a certain sight year and one or multiple adaptions during its lifetime up to 2150. The strategies are assessed based on the business case, up-scaling of the clay refinery in the Ems-Dollard estuary, the impact on the Natura 2000 area and feasibility. It was found that incorporating an adaptive strategy could promote the business case and help mitigate the risk of over-designing.

Coral reefs are considered to be a crucial ecosystem and form of natural coastal protection. They owe their effectiveness to their high structural complexity. This complexity, which positively influences aspects like species richness, coral recovery, live coral cover, and coastal protection, is often quantified using parameters such as porosity and rugosity as has been done in this thesis. However, coral reefs are currently facing decline due to both natural processes and human activities. To address this issue, artificial reefs (ARs) have been proposed as a potential solution. Coastruction, a start-up specializing in 3D printing, has developed a 3D printer capable of creating artificial reefs in any shape. In this thesis, biomimicry-inspired complex ARs were designed and printed to examine the impact of structural complexity on wave attenuation.

Data was collected through physical model tests in a wave flume at Deltares, using Froude scaling with a length scale of nL=4. The experiments involved 1 to 2-meter-long reefs composed of three distinct structures with different geometrical complexity levels: design 1 (medium complexity), design 2 (maximum complexity), and cubes (low complexity). These structures were tested under similar, primarily non-breaking wave conditions. The aim was to determine if the dissipation observed resulted from friction rather than wave breaking.

The results of the conducted experiments showed that structural complexity does not significantly affect wave transmission. Overall, high transmission coefficient were observed across the different structures, ranging from 0.96-0.99. Additionally, no significant differences in dissipation were found among the three structure types, with dissipation ranging from 5.5% to 13.5% of the incoming wave energy. However, increased reef coverage positively impacts transmission reduction and spectral wave dissipation, exhibiting a linear decrease in transmission coefficients and a simultaneous increase in dissipation.

In summary, the various artificial reefs examined in this study demonstrated minimal effectiveness in wave attenuation, with negligible differences among reefs featuring different degrees of structural complexity. Despite the similar wave attenuation effects, the environmental benefits of a highly structurally complex reef suggest that it should be considered a viable solution.

In this study loading characteristics of rubble mound breakwater crown walls are investigated. This is done using a 2D physical model, testing for overall stability (through sliding) and temporal & spatial distributions of pressure along the face and base of the crown wall. A variety of wave conditions and geometries is used, with special attention for the influence of foundation level on the loading characteristics. This study shows the importance of this parameter and its link to the time lag between the maximum horizontal force and maximum vertical force in a wave impact. Using a new analysis method based on the factor of safety, this time lag is quantified and a reduction factor for the vertical forces is introduced.

Living Breakwaters have evolved from traditional breakwaters to create multi-purpose structures that provide environmental and social benefits. These breakwaters promote a healthy habitat for flora and fauna, while achieving the same structural capabilities as conventional structures. Reefy has developed a modular living breakwater that can serve both as an artificial reef and a stable breakwater. This study aims to model the impact of a Reefy breakwater on the transmitted wave height, using a process-based numerical model, namely XBeach.

This study builds upon a previous one conducted by van den Brekel [2021], where scaled experiments were performed in a physical wave flume, to investigate the hydrodynamic and ecological functionalities of the Reefy breakwater. In the wave flume experiments a total of 15 structure configurations together with 35 wave conditions, both regular and irregular, were tested. The first 7 of the configurations were 2D structures with a relatively simple shape and a porosity of 20%, and the rest were complex 3D structures with a porosity above 20%. The wave flume was recreated within the numerical model. The influence of the breakwaters on the wave heights are examined through the comparison between the calculated transmission coefficient of the wave flume and the numerical model. However, XBeach, and more specifically XBeach non hydrostatic+ mode which was used, does not resolve the complex interaction between waves, friction and porosity of the structures. As a result, a simple method had to be found to compensate for the loss of these phenomena, while successfully calculating the transmission coefficient.

The findings of this analysis demonstrate that treating the breakwater as an impermeable change in bathymetry, coupled with a 15% reduction in structure height, produces the desired outcomes. From the 35 wave conditions only the 8 irregular wave conditions were chosen to be of interest due to time constraints. Among these 8 wave conditions that were used as an input in XBeach, only 5 of them were satisfying the criterion that describes the range for which the numerical model should produce accurate results (kd<2). The transmission coefficient was defined as the ratio of the transmitted wave height behind the breakwater, and incident wave height in front of the structure. The wave decomposition method used was a modified Guza split method. In pursuit of finding a valid solution, the breakwaters were modeled firstly as impermeable structures with a decreased height, secondly as impermeable structures with a decreased width, thirdly as vegetation with the help of the vegetation module, and lastly the maximum wave steepness criterion as applied in the model was increased (maxbrsteep up to 1.4), without modifying the structure height.

Among the suggested solutions and after being validated for 13 experiments with both simple and complex structures, decreasing the width did not achieve the desired results while the vegetation module failed to produce consistent outcomes. As far as transmission coefficient is of interest, a 15% decrease in the structure height had the smallest mean absolute percentage error of almost 10% and a root mean square error of 8.2%. On the other hand, choosing to increase the maximum wave steepness (to a value of 1.4) is not a valid option for the complex 3D structures, but illustrates even better behavior than a change in the structure height when simple structures with 20% porosity are concerned. To summarize, this thesis provides alternatives to replicate in a simplified way the impact of porous structures, such as Reefy breakwaters, without the need of an extensive and time expensive numerical model.

Placed block revetments on dikes are used for the protection of the structure against wave loading. The dimensions of the revetment are determined with empirical relations and design tools. The stability of the design solution needs to be checked by physical flume tests to ensure the safety of the structure. However, these tests can be costly and time-consuming. Research has to show whether computational fluid dynamics (CFD) can be used to improve the design process of the revetment and reduce the number of required physical flume tests. The most important load on a placed block revetment is an uplifting force that results from a (porous media) flow through the top and filter layer of the structure during wave-structure interaction. This study aimed to determine whether CFD modelling can accurately predict the uplift of the top layer of a placed block revetment and identify the optimal approach for modelling this type of revetment.

In order to reach this goal, an existing CFD model is used and validated with a physical test performed in the Delta Flume. First, the data from the physical test is extensively analysed to check how the results relate to the physical processes described in the literature. Next, a two-dimensional CFD model has been set up with the open-source software package OpenFOAM. To mimic the physical test, the model is created in 1:1 scale with all the structure dimensions similar to the Delta Flume set-up. The top layer in the physical test consisted of C-star blocks that have been schematized in order to model the structure as a two-dimensional model. Two different schematizations are proposed and tested on their performance. In the first schematization, the top layer is modelled as a continuum in which the resistance is determined by the Darcy-Forchheimer equations, just as in the filter layer. In the second schematization, the top layer is modelled by rectangular blocks alternated by small gaps. The CFD model results and the physical test are compared based on trends and statistics.

The model performance has been evaluated for three consecutive processes. First, the wave generation and propagation are analysed by comparing the wave spectra of the incoming waves and monitoring the wave conditions along the numerical flume. The model could reproduce a wave spectrum accurately, and the wave properties remained constant during propagation. The second process that has been evaluated is the wave-breaking that results in wave loads on the slope. It has been shown that the quasi-static wave loads are predicted with reasonable accuracy (i.e. measured values deviate less than 10%) while the dynamic wave loads deviate up to 35% from the physical test results. It is concluded that there is room for improvement in modelling the wave-breaking. The last evaluated process is the porous media flow that leads to pressure differences over the top layer. The optimal schematization of the top layer was the rectangular blocks alternated with gaps. Just as in the physical test, only the dynamic wave loads lead to upward forces. These predicted forces deviate less than 20%.

To model the flow through the top layer, the local inertia, advective inertia and resistance terms are important in the conservation of momentum equation. For the top layer schematized with a continuum layer, the advective inertia term is underestimated by orders of magnitude leading to inaccurate results. The top layer schematized with rectangular blocks led to load predictions with reasonable accuracy. The resistance term was underestimated because no shear stress in the gaps between the blocks was included. However, the resistance term is small compared to the local inertia and advective inertia.

In future research, additional attention should be given to determining the optimal grid resolution. Results have shown that decreasing the grid sizes leads to increased predicted pressures. It could also be beneficial to further investigate the modelling strategy of the top layer. Schematization 2 has led to the most accurate model results. However, this model requires a fine grid resolution to resolve the flow through the gaps, leading to high computational times. A CFD model in which the top layer is schematized as a continuum would be beneficial in terms of computational times. Increasing the fictional porosity and grain size increase the contribution of the inertia term in the momentum balance, leading to better results. Investigating a method for obtaining these values may be of added value.

Crownwalls are often placed on top of vertical composite breakwaters to reduce overtopping. Adding a seawards facing overhang at the top of crownwalls has successfully reduced overtopping even further without increasing the freeboard of the crownwall. Crownwalls with a fully curved (FC) face are most often used in seawalls where they are subjected to wave load by breaking waves. However, it has recently become of interest to explore the possibility of using the fully curved shape for vertical breakwater crownwalls instead of a crownwall with a recurve (R). This thesis aimed to investigate further the wave loading by non-breaking waves and the dynamic response of a FC crownwall on a vertical breakwater through an offline one-way coupling of a CFD-generated pressure-time series and a FEM model in Diana FEA. The results were then compared to those of other studies on the wave load and dynamic response of a R crownwall. Pressure-time series from CFD numerical simulations of three wave states (W5, W6 and W7) were analysed to determine how the structure was loaded. The analysis of the pressure-time series showed that for the smallest wave state, W5, the pressure distribution at the moment of maximum total force was trapezoidal. However, for the larger wave states, W6 and W7, there was a pressure peak at the top of the curve of the crownwall due to the C-CI phenomenon. Waves with impulsive pressure and force impacts were found in all wave states, and the pressure and force impulses were much less variable than the maximum pressure and force values. Dynamic nonlinear analysis showed the fully curved crownwall cracks under wave load by wave states 6 and 7. Cracking occurs at the centre of the curve, where the largest tensile stresses are located. Comparing the results to those of the recurved crownwall, it was concluded that the fully curved crownwall was less favourable than the recurved crownwall. For the same wave states, the FC crownwall was subjected to larger wave forces than the R crownwall, which led to a larger bending moment exerted on the structure. The FC crownwall is also larger than the R crownwall, with the same freeboard, and therefore requires more use of concrete. Due to the shape of the FC crownwall, the maximum tensile stresses are 2.5-3 times higher than those in the R crownwall. More use of steel reinforcement bars is therefore also needed. It would, however, be interesting to investigate further FC crownwalls on vertical breakwaters with different geometries, e.g. radius of the curve, to see if their design can be improved. The applicability of the extended Goda method for calculating static wave load on recurved crownwalls was also investigated for the FC crownwall. It was found that for waves with wave steepness less than 6% and vertical wave velocity less than 2.4 m/s, the calculation method could be used for a preliminary design. However, the dynamic response of the structure must be addressed in the final design, e.g. with a dynamic analysis or the DAF method.

Besides global warming, climate change is expected to result in (among others) an increased number of extreme precipitation events and a rising sea level. In the future, this can be the cause for an increased number of high river discharge events in combination with high water levels on the North Sea, which goes hand in hand with an increased risk of flooding in the Netherlands. However the rate of change in the future is hard to predict and therefore very uncertain.

One of the projects which seeks to be a (part of the) solution in climate change and flood prevention for the Netherlands is the Delta21 project. By combining a sea defence, energy storage lake, pumping station, overflow structure and storm surge barrier, the Delta21 project increases safety against flooding while providing a sustainable way for temporary energy storage.

In this thesis a design was created for the sea defence of the Delta21 project, during the design process the uncertainty of climate change was taken into account by the use of the adaptive pathway approach.

Three conceptual design variants are created which provide water safety until one meter of sea level rise occurs. When the sea level increases even more, the designs should be altered to the new circumstances. Various adaptation options are put together in the form of adaptive pathways which provide adaptation methods for up to five meters of sea level rise. The pathway schemes are evaluated on 4 evaluation criteria to select the preferred adaptation method for each of the three base design variants.

Using an analysis over the total lifetime of the sea defence measured in meters of sea level rise, the preferred conceptual design for the Delta21 sea defence is selected.

It turns out that the adaptive pathway approach can be used to integrate the uncertainty of climate change into the selection of a conceptual design variant for a sea defence. The use of the approach provides a methodological way of evaluating design options under uncertain circumstances.

Groynes in the Lower Elbe estuary in Germany are subject to attack from overflowing long-period primary ship-induced waves. Damage can occur on the crest and lee side of the structure due to the high turbulent overflowing flow velocities. An actual design tool for groynes exposed to overflowing primary ship waves is not yet present. Therefore, experimental research based on physical modelling is performed to establish a relationship between the damages and corresponding overflowing primary ship waves. This can be used as a design tool for groynes exposed to these types of wave attack.

With the advance of urbanisation in coastal regions comes an increase in armoured shorelines and the destruction of coastal marine habitats. To combat this, ECOncrete Tech Ltd has developed a new eco-friendly armour unit, the CoastaLock. With its ecological advantages proven, its hydraulic performance is still largely unknown. This thesis aims to do the first exploratory research into the stability, reflection and overtopping performance of this CoastaLock armour unit, through a literature study and physical model tests. Data is gathered during physical model tests on a deep-water impermeable slope with different wave steepnesses, armour unit spacings, underlayer thicknesses and block orientations.

The mechanism of breathing is witnessed in tests with CoastaLock armour units. This mechanism shows the armour layer repeatedly getting lifted up from the slope and placed back on it as a result of pressure differences over the top layer during maximum wave run down. This mechanism causes movement of granular material in the underlayer, with large under deformations as a result. This thesis, therefore, advocates for listing the breathing mechanism as a failure mechanism.
The CoastaLock armour units show stability numbers of approximately 2 when tested at armour spacings below 10%, and stability numbers larger than 4 when tested at armour spacings larger than 10% for a wave steepness of 3.5%. Wave steepness is shown to be of a minor influence too, with the influence of underlayer thickness and unit orientation being negligible.
The reflection coefficient is shown to be mainly dependent on wave steepness, and minorly on armour spacing. Reflection coefficients of approximately 0.8 to 0.5 are measured for wave steepnesses between 1.5% and 4.5%.
Overtopping is shown not to play a role in situations with armour spacings lower than 10%, as the layer fails before overtopping can become an issue. For the more stable spacings, those larger than 10%, the overtopping discharge is quantified and the roughness factor is shown to vary from 0.732 for 10% spacing to 0.61 for 25% spacing. Factors other than the spacing are of a minor to negligible influence.

The CoastaLock armour units show promising results for situations with a short leakage length and large armour spacings, i.e. situations where the pressure differences over the top layer are small. However, this research is the first exploratory research into these armour units and should therefore be interpreted as no more than an indication of the behaviour of CoastaLock. It is recommended to expand the research into these units, both on the units in different situations as on the failure mechanisms of the units and the solutions thereto.

Starting in the 1930’s, many research projects have been performed with the intent to establish rules, which can be used to design granular, geometrically closed filter structures for hydraulic structures. However, for the vast majority of these research projects, the filter structure was loaded by uniform flow, either in the parallel or perpendicular direction. This means that, while for filters under flow loading multiple design options and diagrams are available, the knowledge on design rules for sloped, closed filters under wave loading remained limited. These knowledge gaps mean that, in design practice, often the design criteria for closed filters under flow loading, for instance the original criterion by Terzaghi, are used for the design of closed filters under wave loading, while these have not yet been experimentally verified for this use case.

This thesis combines model test results from two historically performed experiment programs on sloped, closed granular filters under wave loading with new experiments, which were conducted especially for this thesis. By analyzing the results from these three research projects, it was found that commonly used interface stability criteria for closed granular filters under flow loading are not valid for use in the case of a sloped, closed granular filter under wave loading. Based on the test results from the three test programs that were analyzed, a new interface stability criterion for sloped, closed granular filters under wave loading is proposed.

Wave overtopping, in which a wave washes over the crest of a dike and damages its inner slope, is one of the many mechanisms that is found to be responsible for dike failures. As these dikes are generally covered with grass, insight into the erosion resistance of these grass covers against wave overtopping is desired. To assess the performance of dikes against wave overtopping, various concepts exist, such as the widely used cumulative overload method. However, these concepts lack a thorough understanding of the physical mechanisms at play during grass cover failure as a result of wave overtopping. Furthermore, the methods are generally not time-efficient and are labor intensive, leading to high costs. Therefore, a time-efficient and predictive method is desired that is based on the physical mechanisms of grass cover failure.

The grass pull device, which is used extensively in this thesis, may serve as an alternative for the existing assessment methods. The device, which is reminiscent of a tensile test used in mechanical sciences, is able to exert various load mechanisms on the grass cover. In this thesis, the grass pull device is used to study various aspects of grass cover failure. Special attention is given to the influence of cyclic loading on the grass cover. Additionally, material properties have been derived that may serve as input for numerical grass erosion models. Furthermore, the influence of grass roots, subsoil type and pore saturation on the failure mode of grass covers was investigated.

The results of this study showed a continuous growth of deformation and a decrease in stiffness when the grass cover is loaded cyclically. The behavior of grass during cyclic loading was found to be comparable to other composite materials, such as fiber-reinforced plastics. The material properties Young's and shear modulus were derived. The Young's moduli were found to be slightly overestimated, while the shear moduli were found to be comparable to what may be expected from literature. Differences in grass cover properties on different subsoils were identified, showing that grass covers on clay are generally better at resisting deformation, while having a brittle failure mode. For grass covers on sand, a large spread was observed and the material was found to deform easily, while still providing resistance at large deformations.

Based on the findings of this study, recommendations were made to improve the grass pull device. It was found that the grass pull device was successful in providing insight into various physical processes. Whether the grass pull device will be able to capture all relevant erosion mechanisms remains questionable, but it has proven to be a successful addition to existing assessment methods.

The design of hydraulic structures like breakwaters and crest walls is often based on empirical formulations, physical models test, numerical models and a fair amount of expert judgement. Each technique has its own pros and cons. The main limitation of the empirical formulas is that often they have to be applied outside their range of validity. Physical modelling also has its own shortcomings. When breaking waves hit the structure, the location of the maximum pressures is still not well known due to the high spatial variability. By using an array with low spatial resolution, the forces estimated in the physical ume will usually underestimate to some extent the actual forces experienced by the wall (Ramachandran et al., 2013). In the last decades, numerical modelling has become an attractive alternative in simulating wave-structure interactions. Nevertheless, estimating loads on crest walls in the numerical ume is still at its early stages. On that account, the present work validates the prediction of wave induced forces on the front face of crest walls on top of rubble mound breakwaters in CoastalFOAM. A scale model of the Holyhead breakwater, located in Wales, is used. The key validation topics are the reproduction of wave-structure interaction when heavily breaking waves reached the wall, the evaluation of the ventilated boundary condition implemented by Jacobsen et al. (2018), the porous ow inside the armour layer formed by Tetrapods units and whether the simplication of not using a turbulence model, as done by Jensen et al. (2014) and Jacobsen et al. (2018), is also valid under heavily breaking waves. Four validation cases were used to test the capabilities of the numerical ume. The results conrm that it is possible to accurately reproduce the wave conditions and the wave induced forces from a physical modelling campaign. Overall, the CoastalFOAM model is able to capture the shape and the order of magnitude of the force events. A calibrated model predicts the highest wave induced forces (forces with an exceedance probability of 0.4% and 0.1%) with errors lower than 20%. Moreover, the results indicate that for practical applications it is not essential to include a turbulence model in the numerical ume to obtain reliable forces on the front face of crest walls for dierent wave conditions. Another outcome of this study is that the implementation of the ventilated boundary condition is required in the interface between water surface and structural elements to mimic accordingly the air-water mixture when the structure is subjected to a heavy wave attack. Nonetheless, there is still room for improvement in this area, where a better understanding of this boundary condition and of the air entrapment during wave-structure interaction needs further research. Despite the large computational time required by the numerical ume when large wave trains must be simulated, a CFD model during design stages of breakwaters and crest walls provides higher spatial and temporal resolution of the wave induced pressures exerted on the wall than a physical test. Therefore, a better picture of the forces and pressure distribution in the front face can be obtained.

Since the behaviour of damage formation on steep slopes varies from mild slopes, a comprehensive definition of damage needs to be established for different situations. Knowledge about the transition of rock stability between steep and mild slopes is lacking. This study presents a reliable and efficient design method. The goal of this study is to describe the stability of rock slopes (1:6 - 1:10). Insight into governing processes and quantified damage characteristics are gained by analyzing wave attack in the transition zone of mild and steep slopes. The method is based on physical model test series with deep water wave conditions and perpendicular wave attack, which were performed to understand the stability of slope 1:6. The slopes were measured with stereophotogrammetry and processed in a 3D model with Agisoft Metashape software. Before wave attack, an initial profile is retrieved which can be compared with cumulative damaged profiles. Automatically detectable Ground Control Points are used as reference to assign scaling information and location recognition. The quantity and location of normative damage on mild slopes are identified for six damage parameters, based on different relations regarding erosion area and depth. The tests results have led to a number of key findings. Damage formations are developed by dominant downward (bar profile) or upward transport (berm profile). In general, a thicker layer and a milder slope give more upward transport. Besides the change in transportation direction, no significant differences in stability numbers were observed for various layer thicknesses with similar wave characteristics. For an increasing slope angle, the influence on stability significantly decreases between various wave steepnesses, due to the impact of a higher amount of more damaging plunging waves. The definition of initial, intermediate and failure damage limits are specified and the test profiles are matched per limit. Initial and intermediate damage are found at a constant value of _3_.3 for a varying layer thickness and slope angle based on the definition and matching profile analysis. Failure of the slope is defined by the event in which the impermeable layer or filter layer becomes visible, which occurs for a layer thickness of 2.53=50 at a damage level of _3_.3 = 1.5. Research at 1:6 slope shows that a higher failure limit is observed for layer thickness 53=50. Due to the lack of test results of other slopes (1:8 - 1:10), specific values cannot yet be indicated for the range 1:6 - 1:10. The lack of testing is also due to limits of the flume and wave machine. In order to remain at least 13=50 over the slope, an acceptable damage limit for a design with a layer thickness of 2.53=50 is found at _3_.3 = 0.9. For a layer thickness of 53=50, an acceptable value of _3_.3 = 1.5 holds a remaining layer of 2.53=50 due to the increase of bed mobility for higher damage numbers. The design formula of mild slopes is based on _3_.3 because this gives the lowest bias error, variability and measurement errors. The design formula for mild slopes is used to estimate stability, nominal rock diameter and acceptable damage for conditions within the applicability range of performed tests. This new method for mild slope design indicates that a higher allowable stability can be used compared to the previously used (extrapolated) method of Van der Meer (1988). A displacement is not necessarily linked to erosion due to alternating upward and downward transport on mild slopes. It can be discussed to what extent a rock, after displacement, contributes to the stability of the design. This can be investigated by the ratio of mobility and erosion over the entire slope. This study has shown that the damage parameters, based on displacements, are larger than the width-averaged erosion parameters. This appears to be a useful ratio for distinguishing mild slopes from steep slopes. In the latter these parameters are considered equal. Requirements for mild slope stability can also be imposed on the basis of the ratio between mobility and erosion as described in this research.

Rocking behavior of breakwater research was conducted by Centre for Civil Engineering Research and Codes workgroup CUR C70 (1989) and after that no further validation of the research work has been done. In order to understand the rocking behavior and validate the previously conducted research CUR C70 (1989) wave flume test has been conducted in similar setup. One of the important parameter of rocking behavior is the magnitude of impact velocity and the distribution of this impact velocity stochastically and spatially. This research investigates the order of the magnitude of the impact velocity and distribution of the impact velocity comparing, validating and concluding new findings those were not incorporated in CUR C70 (1989) research.

During this research a new type of sensor Tinyduino has been used inside a cube to measure the acceleration and angular velocity of the cubes. Eight of these sensors were used in eight different cubes. All the sensors were tested properly in Deltares to check the sensors working properly and better understanding the sensors. Before this research two test programs were performed by Deltares in tetrapod in double layer using this Tinyduino sensor in stand alone mode but the data were not analyzed. This research analyzed the data for the tetrapods provided by Deltares and compared with CUR 70 (1989). During this analysis and the testing of the sensors it is found that angular velocity measurement can be much more reliable than acceleration hence the angular velocity measurement is used during data processing of instrumented cubes in double layer.

For experiments eight instrumented cubes were placed using very flexible wire in a randomly placed double layer cubes with same size and almost similar density over a permeable filter layer. The slope 1:1.5 is used for the experiment same as the CUR C70 (1989) research conducted. The cubes were placed in one constant level but the water level is varied in order to demonstrate different slope position. Three slope position Y/Dn=0, Y/ Dn =-2 and Y/ Dn =-4 is used during the research program. Three different wave heights and two wave steepness were used during this test program. Due to time limitation 18 test setups within two days were performed. The data were collected real-time using wire and saved in text file simultaneously with eight instrumented cubes stored in a laptop provided by Deltares.

After the test program all the data is processed with matlab script and analyzed. The result of the analysis showed that the order of the magnitude of impact velocities is same as the CUR C70 (1989) research. It has also been found that the impact velocities are also dependent on the wave steepness which was not included in the CUR C70 (1989) research. So it is recommended for future work to update the equation incorporating wave steepness. Another important parameter is number of collision which was assumed to be 3 times is CUR C70 (1989) but found incorrect during this thesis. It is concluded that number of collision is dependent wave height, wave steepness, position over slope and also exposure to wave attack. It is also found that the number of collision is continuous during wave attack after a certain impact velocity.

One of the finding of CUR 70 (1989) was the location of maximum impact velocity over slope. It was concluded in CUR 70 (1989) that the maximum impact velocity lies on Y/Dn=0 meaning on the water level but during this research all the sample those worked observed that the maximum impact velocity is located on Y/Dn =-2 over the slope under the water level.

For future research work it is recommended to use different types of armor unit in single layer to understand the rocking behavior and also it is recommended using the Tinyduino sensors which provides accurate data on movement of the armor units.