Site characterization is indispensable in the design phase of geotechnical engineering projects. As a key factor in site characterization, the characterization of soil undrained shear strength (S_u) is always in the spotlight. Various methods, including laboratory and in-situ tests, have been developed to measure S_u. Nevertheless, these measurements are usually sparse at a specific site due to limited time and budget. To enhance S_u characterization, other relevant geotechnical investigation data (e.g., cone penetration test data), can be transformed into S_u through empirical correlations (referred to as transformation models) to provide more information on S_u. Considering this process introduces the transformation uncertainty and a developed transformation model may not be fully applicable to a local site, probabilistic transformation models (PTMs) have been developed to characterize soil parameters in a site-specific way and quantify the uncertainty to augment engineers’ judgement. However, few PTMs incorporate the spatial correlation of soil parameters, especially in the horizontal direction. This limitation hampers the ability to probabilistically characterize S_u in 2D/3D space, which is significant in practice. Moreover, estimating the horizontal spatial correlation from pure geotechnical data is challenging because they are typically sparse. In light of these circumstances, this thesis first proposes a PTM-based scheme to probabilistically characterize S_u in 2D. Then it is proposed to integrate geophysical data into the scheme. Compared to typical geotechnical investigations, geophysical surveys provide abundant 2D/3D measurement data, which are often correlated with geotechnical data. The fusion of these two data sources benefits characterizing geotechnical data including S_u. Particularly the horizontal spatial correlation of S_u 2D domain can be estimated from the abundant geophysical data. To be specific, a well-established PTM, MUSIC-X, by which measured S_u and other relevant soil parameters can be used to preliminarily characterize S_u, is first adopted. In this case, characterization specifically refers to simulating 1D vertical profiles of S_u. It is then combined with the intrinsic collocated co-kriging (ICCK) model, by which primary data (i.e., S_u) in 2D or theoretically 3D space can be estimated through linearly combining the preliminarily characterized S_u from MUSIC-X modelling and observed secondary data (i.e., geophysical data). The secondary parameter considered in this study is interval velocity (V_int). The scheme, to combine the MUSIC-X and ICCK model to estimate S_u in 2D space by the fusion of geotechnical and geophysical data, is applied to a real case study at Hollandse Kust (west) wind farm zone to demonstrate its effectiveness. The results indicate that such a scheme can robustly estimate a 2D cross section of S_u with quantified uncertainty. A comparative analysis is conducted between the proposed scheme and two alternatives, one lacking preliminary S_u characterization (i.e., without MUSIC-X modelling) and one lacking geophysical data, confirming that the proposed scheme has a relatively high accuracy in the estimated cross section. The research reveals it is sensible to combine MUSIC-X and ICCK for 2D S_u characterization and brings a new perspective that integrating geotechnical and geophysical data is promising to characterize soil parameters in higher dimensional space.

The Netherlands is a country that is being threatened by water, both from the rivers and from the sea. The Dutch have built dikes to keep their lands from inundation. To ensure the strength and stability of these dikes, they are being assessed on the basis of several failure mechanisms. One of these failure mechanisms is Backward Erosion Piping, or piping for short. In piping, the current underneath a dike is strong enough to take soil particles with it. Tests on piping in tidal subsoil were conducted in the summer of 2021, where a pipe was found to have grown at greater depth than expected The occurrence of this deeper piping has rarely been seen before, let alone described. This lack of knowledge poses a potential safety risk, as it may underestimate the vulnerability of certain subsoil configurations. Therefore, the objective of this thesis is to develop a comprehensive understanding of deeper piping and identify the key parameters influencing its formation. To achieve this objective, a definition of deeper piping and its differentiation from conventional piping is established. Sub-mechanisms governing deeper piping are examined by analysing the forces responsible for grain movement and the forces that maintain grain stability. A Finite Element Model of the subsoil is constructed to quantify the driving forces within the subsoil, which, when combined with resisting forces, enables the determination of whether deeper piping can occur in a given subsoil configuration. To investigate the factors contributing to deeper piping, a series of simulations are conducted using this Finite Element Model. By varying the parameter values while keeping other factors constant, the influence of each parameter on the occurrence of deeper piping was examined. The analysis revealed that several key parameters significantly affect deeper piping formation, including cohesion force (P), cohesion anisotropy (?P ), permeability and thickness of the top layer (X0 and 70, respectively), permeability of underlying layer (X1), permeability anisotropy (?X ) and representative grain diameter Q_R]. Also, it was found that the entrance configuration plays a large role in deeper pipe formation. These findings provide valuable insights into the mechanisms underlying deeper piping and enhance our ability to identify subsoil configurations that are prone to this phenomenon. These findings enhance the identification of subsoil configurations prone to deeper piping, thereby improving risk assessment and mitigation strategies associated with this failure mechanism.

River dikes in the eastern part of the Netherlands are characterized by a relatively low daily water table. In the unsaturated zone above the water table, a negative pore pressure (matric suction) is present that binds the soil particles together. Suction can be especially large in soils with small grains, such as clay, and increases the strength of a soil due to higher effective stress. High suction in soils has been proven to significantly increase the stability of dikes. However, due to water level variations and other climatic influences such as precipitation and evapotranspiration, suction in a dike will vary over the year and throughout the cross section. The research described in this report aims at evaluating whether there are parts of a cross-section of a dike where soil suction remains throughout a high water event in a river such that an increase in the strength of the soil can be used in stability analyses. The analysis is centered around a case study of an existing dike in the east of the Netherlands and uses time-dependent unsaturated groundwater flow models to simulate the spatially and temporally varying suction for a number of different initial conditions and climatic influences defined by scenarios. The results show that it is very unlikely that suction in the dike will be lost during the summer months. Due to high evapotranspiration, the water content in the dike is low, which results in a lower water table during a high water event and less infiltration in the dike occurs during precipitation. During winter, the probability that suction is lost throughout the dike cross-section is larger and will occur after heavy precipitation with a long duration where the return period is 1 – 10 years. Although in most scenarios some suction will remain present, the amount of suction during the winter is generally low, between 0.5 and 1 meter, which results in an apparent cohesion of 2 – 4 kPa. A simple stability calculation illustrates that the increase in strength in the dike due to suction does not result in a significant increase in the stability of the dike during high water conditions.

The need of shear strength measurements of soil in the design phase of geotechnical engineering is almost indispensable. Many methods have been applied to estimate the shear strength of soil, including various laboratory test, in-situ test and analytical methods. As an in-situ test method, cone penetration test (CPT) is a powerful and cost-effective tool for the investigation of subsoil conditions. CPT data is usually complemented by the laboratory test data for verification. The laboratory-based studies of subsoil, however, can be not only a complex but also tedious and expensive task for large projects involving large amount of data. Therefore, new approaches for estimating the soil shear strength are demanded. Having demonstrated superior predictive ability for many material properties compared to traditional methods, machine learning methods have been increasingly popular and widely used. This thesis focus on the prediction of soil undrained shear strength through cone penetration test data. The major objectives of this master thesis include testing how machine learning could help us lower the need for laboratory test data. At first, the research starts with a literature review of various methods used to evaluate the soil shear strength. Comparing to the machine learning methods, the laboratory and in-situ test methods are relatively more time-consuming, costly and labour-intensive. And the analytical methods are considered lacking in precision. Then the training dataset which consists of 526 samples is introduced. In each sample, there are four input variables obtained from cone penetration test, namely the effective stress (σ′v ), cone tip resistance (qt − σv), effective cone tip resistance (qt − u2) and the excess pore pressure (u2 − u0). The undrained shear strength obtained from laboratory test is taken as the output variable. Next, the training dataset is fed to five machine learning techniques, namely the artificial neural network, support vector machine, Gaussian process regression, random forest and XGBoost, to train models. The hyperparameters are tuned with k-fold and group k-fold cross-validation strategies in the validation process. After that, the testing dataset which consists of 20 samples is established. Cone penetration test data that are in close vicinity to the location of the samples are processed by Gaussian process regression to obtain representative cone penetration test data at the sample location, which is taken as the inputs in the testing dataset. The undrained shear strengths of the samples are measured by Consolidated-Undrained shear test and are taken as the outputs of the testing dataset. Finally, the five machine learning models are tested on the testing dataset. The crossvalidation results, together with the prediction results of the models on the training and testing dataset are evaluated, gathered and compared by various statistic metrics to show the relative performance of the models. XGBoost appears to be the most accurate of all the tested algorithms on this dataset. And Gaussian process regression is chosen as the second option due to its ability to capture uncertainties. The robustness of these two models are then validated from a statistical point of view by applying Monte Carlo analysis. The importance of the input parameters in this study is evaluated by applying random forest for the sensitivity analysis. The results from random forest indicate that the excess pore pressure and the cone tip resistance - total vertical stress are the most influential inputs to the undrained shear strength

Discontinuities have an important influence on the bulk behaviour of geomaterials. Discontinuities can be physical openings like fractures and joints, interfaces between different material types, or zones with different material behaviour like mortar or reinforcement material. Bulk material behaviour is often determined by the behaviour of the discontinuities by providing preferential pathways for flow or by acting as failure planes.
FEM simulations with continuum elements are typically not suitable for modelling discontinuities with no width, due to the inability to model discontinuous fields. One solution is to represent discontinuities using zero-thickness interface elements. Inserting zero-thickness interface elements allows for discretely modelling discontinuous behaviour between neighbouring continuum elements. The continuum elements are then used to represent the bulk behaviour and the zero-thickness interface elements discretely model the discontinuity.
A new constitutive law for zero-thickness interface elements that expands the cohesive zone model with friction is formulated in this thesis. The tangential stress response of the interface elements consists of a cohesive and frictional contribution. The cohesive component of the constitutive law is based on the Crisfield’s cohesive elasto-damage model, which is characterised by a linear elastic response to increased relative displacement, followed by degradation of the cohesive strength and stiffness. The elasto-plastic frictional component is based on the Dahl friction model and implemented in parallel to the cohesive component.
The implemented new constitutive law is verified in a large variety of loading conditions, including tangential loading, tangential loading with a reversion of loading direction and combinations of mixed-mode loading. The constitutive laws are validated based on shear box experiments performed on London clay where interface elements are used to model a predefined shearing plane.
The cohesive and frictional constitutive law can represent confinement-dependent tangential stress and residual stress at large relative displacements. These features, observed in the shear box experiments, could not be reproduced by the purely cohesive constitutive law. This new constitutive law can be used to better model discontinuities in geomaterials represented by zero-thickness interface elements.

Risk-based and probabilistic reliability design methods are starting to be integrated into the daily geotechnical engineering practice, as shown by the current draft of EN1997 (Eurocode 7), which explicitly allows Reliability-based verification of limit states. The semi-probabilistic design methods in EN1997 are a compromise between ease of use and accuracy of the achieved reliability levels. In contrast, EN1990 (Eurocode 0) requires accurate full probabilistic design targets. An assessment to investigate if the semi probabilistic methods from EN1997 comply with the EN1990 reliability targets is therefore necessary since the link between the norms is not explicitly established.
In this study, pile foundations serve as an example to assess compatibility between the EN1990 and EN1997. The deterministic Dutch pile design approach by van Mierlo & Koppejan (1956), CPT based design method, is adapted to the semi probabilistic model pile recipe in the EN1997 draft. This thesis assesses what reliability levels are achieved by the model pile design method in draft EN1997 considering the axial bearing capacity and how these compare to the reliability targets in EN1990.
Methodology: To assess the performance of the model pile method from EN1997, the design outcome from the semi-probabilistic model pile method is compared to fully probabilistic quantification methods of the resistances and loads. Two probabilistic quantifications are used in the assessment, a Bayesian and a Student-T model. The achieved reliabilities are assessed with both the First Order Reliability Method (FORM) and Monte Carlo Simulation (MCS). Additional insight into the influence of different parameters is provided with sensitivity analysis and the calibration of partial factors. The assessment was applied in two case studies located in the Netherlands.
In both case studies, the EN1997 designs resulted in reliability levels in good agreement with the reliability targets stated in EN 1990. This suggests an agreement between the semi-probabilistic and the full probabilistic models if sufficient CPTs are used. In both case studies, EN1997 designs accounted for more model uncertainty than the probabilistic models suggest, thus partial model factor (γRd) in EN1997 may be conservative. EN1997 covers spatial variability and uncertainty due to limited observations with the correlation factor (ζ, a factor that transforms calculated to representative values, and the partial resistance factor (γRc). The coverage of uncertainty with these two factors seems to be rather low, especially for situations with few CPTs available (less than 10), although the results seem to depend strongly on the degree of spatial variability in the CPT field. In the case of homogeneous soil conditions between observations, EN1997 design methods led to over-conservative designs for low numbers of observations. If the variability between limited observations (less than 10 CPTs) was high, designs are assessed to be less reliable than EN1990 requires.
Overall, the results suggest good agreement of the semi-probabilistic design methods from EN1997 with the reliability targets defined in EN1990, for designs based on sufficient observations. The sensitivity analysis and isolation of the resistance uncertainty showed for the two cases that the uncertainty of the resistance has the dominating influence on the reliability of the piles, suggesting a low impact of complex stochastic load models. The latter finding suggests that probabilistic treatment of the resistance may be sufficient for assessing pile reliability in practice, while design values could be used for the loads.

This report provides a flood risk assessment of the Isiolo River Basin, in collaboration with Nelen & Schuurmans (3Di, FIS Tool) and Zephyr Consulting (SLAMDAM). This flood risk assessment includes a study of the current flood risk management in Kenya, and in the Isiolo River Basin in particular, because the need for proper flood management is urgent: various climate studies predict an increase in rainfall and an increase in flood risk as a result of the effects of climate change.

Current flood risk management is inadequate. Kenya has defined 21 flood-prone areas whereof one of them is Isiolo Town. Isiolo Town is located in the ENN basin which is, relatively, the most prone to the effects of climate change compared to the other basins. Furthermore, the ENN basin currently has the highest poverty rate and avoidance of further enlargement in poverty rate is important, so there is a need to mitigate flood risks. Since Isiolo Town is located in the Isiolo River Basin, this basin has been chosen for an in-depth study.

The Isiolo River Basin is an Arid Semi-Arid Land region which is often prone to flash floods. Isiolo Town is a flat area located downstream of mountainous area, the rain which falls upstream flows fast downstream and converges into town, often resulting in inundation. Many hazards, both natural and others, are increasing the flood risk in the basin and specifically Isiolo Town.

This flood risk demands flood risk mitigation measures. One possible measure is the SLAMDAM. The SLAMDAM is a movable water-filled flood-barrier. One dam has a length of 5 meters and a height of 1 meter and the dams can be connected to a desired length. The water stored in the dam can be used afterwards for irrigation or other uses.
To recommend effective areas to implement the SLAMDAM, 3Di and the FIS Tool are used. 3Di is a hydrodynamic model and it creates flood maps for different rain events. These flood maps are used as input for the FIS Tool. The FIS Tool calculates the benefits for deploying the SLAMDAM at a certain location for a particular length. The locations which result in the highest benefit are recommended to deploy the SLAMDAM in case of particular rain events. However, a site visit is required to see whether the modelled situation aligns with the real-life situation and to see whether boundary conditions are met.
The SLAMDAM is also compared to other flood risk mitigation measures. Some were analysed using the FIS Tool, whereas others are evaluated based on five self-formulated ranking criteria. These criteria form the base of a scoring matrix where each relevant mitigation measure is scored on.

The performed research has shown the SLAMDAM to rank the best compared to other mitigation measures, both when using the scoring matrix and when using the FIS Tool. However, it is highly recommended to use the SLAMDAM in combination with a Flood Early Warning System. In this way the community downstream can be warned in time to deploy the SLAMDAM. The FIS Tool is found to be especially valuable in finding proper locations for deployment and the dam can be stored close to these locations, enabling fast deployment of the dam in case of need.

For certain geotechnical structures, such as dikes and levees, the design quality is heavily reliant on the sampling quality. In addition, the extent of consequences on geotechnical design of using parameters from samples with a certain degree of disturbance in laboratory tests is unclear. While the available literature has been mainly focussed on disturbance experienced by clay samples, the same empirical findings are not always applicable to organic soils such as peat and organic clay. The global objective of this thesis was to gain an understanding of sampling disturbance in organic soils using both established and non-destructive experimental methods. In order to achieve this, the traditional Dutch Ackerman sampling technique was compared to that of a recently developed sampler, the Deltares Large Diameter Sampler (DLDS). Non-destructive methods and 1D consolidation parameters were used to quantify disturbance in organic soil samples.

The results from existing databases showed that organic clay has a clear linear trend between w and RR and peat showed no definitive correlation, although the peat sampled by the DLDS had significantly higher water contents than the peats from the Ackerman sampler. Moreover, 37% of organic clay and peat Ackerman samples had an ∆e/e₀ below 7%. Meanwhile, all the DLDS results from the Deltares laboratory had an ∆e/e₀ of less than 4%. The use of this parameter is still questionable for peats, as the higher e0 found in DLDS peats is the cause for low disturbance measurements. However, the ∆e/e₀ index may be used in conjunction with other destructive and non-destructive disturbance parameters. Furthermore, the trend showed it becomes increasingly difficult to produce high quality samples with increasing in-situ vertical stress.

Using differences in shear wave velocity as a non-destructive disturbance determination technique, it was concluded that the increased shear wave velocity in Ackerman samples relative to the DLDS samples, although slight, was due to densification of the outermost volume of the sample being compressed by the sampling procedure. This effect was further observed in the other non-destructive method use in this investigation, CT scanning. Analysis of micro-CT scans showed that all specimens had higher greyvalue intensities at the perimeter than at the centre of the sample, however the observed effect was 7.4 times higher with the Ackerman specimens than with the DLDS specimens.

The main recommendation is to continue research on the effectiveness of using non-destructive methods to quantify sample disturbance in peats. For the sampling industry, it is advised to increase the diameter of the Ackerman sampling tubes to between 70 and 120 mm when sampling in peat. In addition, in light of the lesser disturbance caused by the DLDS, it is advised to deploy it in shallow peats instead of the Ackerman sampler with the current inner diameter of 67 mm. Furthermore, it is recommended to deploy both types of samplers side-by-side in sampling projects in order to diminish the effect of site-specific variables.

Two types of calculations are currently used to assess dikes for macro stability. These are the semi-probabilistic calculation and the probabilistic calculation. In practice, the semi-probabilistic calculation is currently mostly used. However, the probabilistic calculation is the most accurate calculation of the two methods resulting in the most realistic results. Performing a probabilistic calculation is on the other hand less simple and a lot more time consuming. Therefore it would be more efficient to still regularly apply the semi-probabilistic calculation to save time and also find a way to make this calculation more accurate. Although the probabilistic calculation results directly in a reliability index, this is not the case with the semi-probabilistic calculation. To link the target reliability index to the factor of safety of the semi-probabilistic calculation, a calibration formula is used. This formula was made using a set of probabilistic calculations from the calibration study. The formula is designed in such a way that in reality 80% of dike profiles are calculated too conservatively with the semi-probabilistic calculation and 20% not strict enough. This means that this formula results in a lower reliability index for the majority of dike profiles than what the realistic value would be. The result of this conservative formula is that in some cases actual reliability of the dike profile can deviate reasonably from the probabilistic reliability index. Some profiles can therefore be made far too robust, while a few others may be approved unfairly. There could be a lot of profit to be made in both safety and in the amount of reinforcement required if the existing calibration formula can be further elaborated. This leads to the main question of this report: “How can the current calibration equation for the assessment of a dike on macro stability be elaborated by using the results of probabilistic calculations from the calibration study and SAFE?” As emerges in the research question, the probabilistic calculations of two projects are used for the research; the calibration study and the project SAFE. The calibration study is the project from Deltares on which the current calibration formula is based. The set of dikes in this project consists of a mix of local and regional dike profiles around the Netherlands. The set is arranged in such a way that almost every type of dike in the Netherlands appears in it so that it can be applied as widely as possible. In contrast to the dataset from the calibration study, the dataset from SAFE consists of dike profiles of only two dike sections at the Lek. The dike profiles belonging to SAFE have been compiled by the water authority and calculated probabilistically by Sweco and Arcadis. The data set from safe is composed in such a way that the critical dike locations of these 2 sections on the Lek are included. Several ways to see if the current calibration formula can be made more accurate are investigated. Initially, SAFE's location-specific set is compared to the calibration formula. It is checked to which extent SAFE corresponds to this and what the reasons can be for existing differences. It is also examined whether the current calibration formula can be improved on the basis of different dike characteristics. A distinction is made of several general , soil and geometric characteristics. The research shows that SAFE's dike profiles require a slightly different calibration line than the line from the calibration study. Instead of 80%, only 64% of the dike profiles lie below the calibration line of the calibration study from Deltares. Especially at the lower reliability indexes, the profiles from SAFE are not calculated safely enough. It is not demonstrable that the profiles from SAFE lead to different results due to the fairly similar type of dikes from the project or the specific location. The difference will therefore mainly be due to the different calculation methods of both projects. In addition, there were also two dike cases/characteristics that has an influence on the difference in reliability index; the design profiles and the berm. With Both, it seems useful to draw up a special calibration line that takes this into account. Especially a calibration line for only the design profiles produces a much more accurate result. This new calibration line results in 30% more accurate reliability indexes. The conclusion of the report is therefore that it would be effective to apply a separate calibration line for SAFE and for the design profiles. For SAFE, the difference with the current calibration line could also be solved with a better calibrated schematization factor, using frequently probabilistic calculations or by renewing the existing calibration study because the difference that occurs seems to be mainly due to the calculation method. Moreover, a separate calibration line for the design profiles thus produces a lot more accurate results. However, the consequence of these more accurate results is that fewer profiles will be on the positive side of the calibration line. The separate calibration line for the dike profiles without designs is 20% less accurate but resulted in safer calculations. When both formulas for designs and normal profiles are used, the complete data set gets a little more accurate with 8% and also even safer than the current calibration formula. So there is certainly something to be improved from the current calibration formula by distinguishing between dike profiles without design and with design.

As the construction of sub-aerial and submarine geotechnical structures increase in amount, rate and size, so do their associated risks. Often, with use constitutive models, finite element analyses (FEAs) are performed in order to identify and mitigate these risks. NorSand, which is a consti- tutive model based on critical state soil mechanics for particulate materials (e.g., sand), is one of the first models to integrate the state parameter ψ into its constitutive framework to model dense and loose sands material with the same parameter set. Importantly, it is able to identify the liquefaction potential as it can simulate softening behaviour due to pore pressure increase of loose soils in undrained conditions. Since NorSand has recently been implement into PLAXIS, a geotechnical analysis software capable of performing FEAs, it must be verified, validated and applied with the software, which is done in this report. First, stress-path and parametric analyses were conducted at single stress points. The stress- path analyses show how the state variables evolve for different triaxial conditions. Systematically changing the input parameters to extremes found in literature helped determine their influence on the evolution of stresses and strains. The resulting figures can be used to help future calibrations to experimental lab test data. The PLAXIS implemented NorSand (PLAXIS NorSand) was verified by comparing it with an implementation written in Visual Basic for Applications (VBA NorSand) by the authors of the model Jefferies and Been. Verification in this context means determining if PLAXIS NorSand is able to produce outputs as intended by the authors. Various testing conditions, both triaxial and direct simple shear, showed overlap and agreement between the outputs of both implementations, verifying PLAXIS NorSand. Then, the model was compared to, albeit not in a traditional sense, an ’analytical solution’, which is the relationship between the mobilized friction ratio Mi and state parameter ψ in its simplest form. The mobilized friction ratio and stress ratio at peak strength of PLAXIS NorSand and the ’analytical solution’ were compared. The values between both showed less than 3% difference, further verifying PLAXIS NorSand. The constitutive model was then validated - i.e., established that PLAXIS NorSand is able to approximate soil behaviour as intended. First, by using the soil parameter set that had been derived from lab tests of Erksak sand as a baseline, the input parameters were varied until PLAXIS NorSand was calibrated to individual triaxial tests as best as possible. Then, triaxial tests of Erksak, Nerlerk and Ticino sand were approximated with PLAXIS NorSand without changing the soil parameters determined from lab test data. PLAXIS NorSand is able to follow lab test data decently well with one parameter set. And if one decides to take the time and calibrate individual lab tests, and deviate from soil parameters determined from a set of lab tests, they can be matched even better. Additionally, it showed a consistent need for activation of the softening flag (S = 1) in order to appropriately model loose soils in undrained conditions. Furthermore, NorSand exhibits indefinite hardening in dense soils during undrained loading, which can be avoided by employing a ’cavitation cut-off’. The last part of this report tested PLAXIS NorSand by applying it in FEAs of a simplified submerged landslide, which was subjected to 20 centimeters of displacement at the crest through a rigid slab in undrained conditions. First, the difference in slope behaviour due to change in soil density within NorSand was determined: dense soil resulted in the slope to be able to bear the full 20 centimeter displacement, whereas increasing the void ratio (i.e., increasing the positive value for the state parameter) gave the effect of even quicker slope collapse and a lower bearing capacity. In other words, when using NorSand, the looser soil the further the failure surface moves up and the quicker the structure fails to maintain equilibrium. Lastly, NorSand was compared to Modified Cam-Clay and Mohr-Coulomb to highlight the differences in their ability to model static liquefaction, while being triggered by unrealistic loading conditions. Even though none of the FEAs showed actual liquefaction, since it is accompanied with the fluidization and loss of structure, they still gave in indication of the liquefaction potential. NorSand, contrary to the other constitutive models, showed the expected high sensitivity to forced displacement resulting in clear shear bands resembling Prandtl-type failure mechanism and early onset soil body collapse.

The construction of a dike improvement often causes a decrease in reliability during construction. Raising a dike may give rise to excess pore water pressures in the subsoil, resulting in a temporal strength decrease. Observing the survival of such a loading condition might provide information on the strength of a dike. Therefore, this research investigates the effect of incorporating construction survival in estimating the slope reliability of dikes. Assessing slope reliability can be done by means of simulation. However, due to the high reliability of dikes, many realizations and thus model evaluations may be needed to achieve sufficient accuracy, leading to infeasible computation times. In this research, a method is developed for reliability updating using metamodelling, providing reasonable computation times while not losing accuracy. The effect of incorporating construction survival is explored by updating the inward slope reliability of a case study dike improvement. The case study shows that the current semi-probabilistic approach to assessing construction stability in practice leads to situations not effective for reliability updating. Though adjusting the construction phasing may lead to a significant update. It is shown that adapting the dike improvement design after survival of the critical construction step results in a 75% reduction of the dimensions of the berm to be constructed, while achieving the same reliability as prior to the update. The main finding of this research is that incorporating construction survival in the reliability assessment can be very effective in terms of reliability, and in making dike improvements more efficient and less space consuming.

In this master thesis the current Dutch method of combining failure probabilities at different spatial scales is researched. This method is used in the Dutch flood risk analysis of earthen dikes to assess the probability of flooding. Currently, probabilities of flooding for geotechnical failure mechanisms are obtained which are not considered to be realistic. To obtain more realistic results the following steps are researched; the assessment of cross-sections, scaling to dike sections and combining sections to a dike trajectory. The increase of failure probability over increasing length, known as the length-effect, is of importance in the current Dutch method of combining failure probabilities. This research shows the conservatism of the assessment of cross-sectional failure probabilities and offers possibilities to reduce the length-effect within and between dike sections to obtain more realistic results.

During the design phase of a foundation installation aspects are easily overlooked. When this aspect is overlooked a foundation element risks not reaching its design depth or getting damaged during the installation process. As a result significant delays and/or costs could occur. A driveability study gives insight in the installation aspects of a foundation element. Driveability is the ability of a foundation element to be driven to a designated depth with a
reasonable speed and without exceeding acceptable material stress. In this report will be researched how the probability of refusal of a sheet pile installed by vibro-driving can be predicted by means of a driving speed-depth-curve (vp-z-curve). Having insight in probability of refusal improves risk assessment and decision-making on driving projects. The focus is on sheet pile installation by vibro hammers as these projects are generally executed in large numbers and under similar conditions. Therefore, these project types are very suitable for application of probability theory. Allwave-PDP is used as a basis to create a probabilistic prediction method. Parameters of the pile driving system (pile, hammer and soil) modelled in Allwave-PDP are turned into stochastic parameters in the probabilistic method. Probability of refusal is predicted with a Monte Carlo simulation. Pile and hammer parameters are modelled by deterministic parameters. Variability of soil parameters is caused by soil heterogeneity and transformation variability. The probabilistic method is validated with three well-reported case studies in Woudsend, Den Oever and Rotterdam. Predicted results show agreement with measured results. The model is a proof of concept that shows the potential of applying probability theory to driveability prediction methods. A proof of concept of the modelling (transformation and spatial variability) and the method (extension of Allwave-PDP) is given. Not any existing model is known to incorporate probability theory in driveability prediction methods. The probabilistic method can simply be extended to other pile types or impact hammers. Therefore, this method is promising to predict different pile driving system configurations.

Seismic design codes are currently moving from a force-based design approach to a performance-based design approach. For example, in a performance-based design approach it could be specified how many lanes must be available during the lifetime of a bridge given a certain earthquake intensity.The problem with this approach is that it is not specified what the probability must be that the performance criterion is satisfied. This raises the question whether the design codes are acceptably safe or not. Focus is laid on the Canadian Highway Bridge Design Code (CHBDC), in which a total resistance factor approach is used. Because the total resistance factor in the CHBDC is a multiplicative factor, lower resistance factors lead to stronger foundation designs. The goal of this thesis is to calibrate the design procedure in the CHBDC for geotechnical systems under seismic loading, by finding a relationship between resistance factors and the lifetime probabilities of failure of said systems. The resistance factor can then be fine-tuned to a lifetime probability of failure that is consistent with the lifetime probability of failure targeted in static design. As an example problem, the bearing capacity of a shallow foundation on a clay with a pseudo-dynamic earthquake load is tested. The research question that is answered in this thesis is: "What should the resistance factors for geotechnical seismic design be in order to achieve a target lifetime probability of failure that is consistent with static design targets?'' Not every possible combination of soil strengths and forces on the superstructure can be taken into account, and therefore the random finite element method is used in a Monte Carlo simulation. Thousands of realization sare performed for each resistance factor, design return period, and "actual''return period that the designed foundations are tested against. By seeing how many realizations of the Monte Carlo simulation fail given a certain earthquake intensity, the conditional probability of failure given that earthquake intensity can be estimated. The total lifetime probability of failure can then be estimated from the conditional probabilities of failure with the total probability theorem. As part of a parametric study, the lifetime probabilities of failure are estimated for six different scenarios, each of which has different sources of uncertainty.  The resulting lifetime probabilities of failure are interpolated in order to find a resistance factor that targets a lifetime probability of failure consistent with static design targets. Currently, the resistance factor that the CHBDC recommends for geotechnical systems under seismic loading are defined as the static resistance factor for that geotechnical system incremented with 0.20, meaning that compared to static design, weaker foundations are designed for seismic load cases. The resistance factor found in this thesis is closer to the resistance factor for static design than to the resistance factor for seismic design. It should therefore be considered to lower the seismic resistance factor to the value of the static resistance factor so that a sufficient lifetime reliability can be targeted.

In most current dike assessments only the stationary water levels are investigated in the assessment of the stability of the inner slope, while there are differences for all kind of dikes between the stationary and transient pore water pressures and therefore in the stability. This results in a conservative probability of failure, while determination of a probability of failure should not be conservative but should be as realistic as possible. When time dependency is included in a calculation, an average flood duration is used, while the flood duration is highly variable. The following research question is defined to address the problem: “What is the influence of the flood duration on slope stability and in what degree affects the flood duration the design?” The degree of influence of time dependency on the pore water pressures and slope stability depends on dike characteristics, flood wave characteristics and the delay in failure. The basis for answering the research question is the software SEEP/W to model the time dependent pore water pressures and the software SLOPE/W to calculate the safety factor for the stability of the inner slope. In the research theoretical dike are used and there is focused on the flood waves in the Rhine and Meuse. A correlation analysis is performed to get insight in the contribution of different flood wave shape variables to the safety factor. And a probabilistic analysis is performed using transient and stationary water levels to know the differences in probability of failure between taking the shape of a flood wave into account or not. In both probabilistic analyses is varied in the permeability and the strength of the material; the shape of the flood waves is varied in the transient analysis. In this way the contribution of the flood to the probability of failure can be quantified. Dike characteristics The differences in pore water pressure are especially large for dikes that consist of an impermeable material such as clay. When only the subsoil consists of clay, larger differences are expected than when only the dike body consist of clay. However, large differences in pore water pressures do not necessary lead to large differences in the safety factor. The largest differences in safety factor are obtained when uplifting of the hinterland takes place during the stationary state and/ or during the passage of a flood wave. A transient calculation is therefore most useful for dikes with an aquifer and a thin (thinner than 5 m) weak (low POP values) hinterland. Flood wave characteristics The differences in safety factor during a permanent water level and the passage of a flood wave are large when no stationary conditions are reached during the passage of a flood wave. This is the case for high and short flood waves. Both in the Rhine and Meuse, the amount of short waves (< 7days) is high, which increases the influence of a time dependent calculation. Also, the importance of a time dependent calculation increases when the response to the increased pore water pressures is delayed caused by the permeability of the material. The influence of the height of a flood wave on the stability increases when the soil is permeable. Delay in failure Time dependency causes failure of the embankment to not occur simultaneously with the maximum wave height. The flood wave is decisive for the dike failure, but the permeability and the strength of the dike determines the moment of failure. Influence on design Taking time dependency into account leads to higher safety factors and lower probabilities of failure with exception for dikes that consist completely out of sand. For these types of dikes, the probability of failure and safety factors are the same order of magnitude. This could affect the design, because the dikes are safer when time dependency is considered. The strength of the material is the largest contributor to the distribution of safety factors and therefore to the probability of failure (60-95%). Whereas the contribution of the permeability to the probability of failure is small (2-12%), the variation in the height and duration of a flood wave contribute for 2-20% to the probability of failure. In a permeable dike this contribution is mainly determined by the height of a flood wave, while in an impermeable dike the duration of a flood wave is of importance. Considering the influence of time in stability probabilities of failure, this research proved that probabilities of failure taking the duration into account differ significantly from stationary calculations. It is therefore useful to take time dependency into account when determining the correct safety factor for impermeable dikes, but it is not useful in determining the correct safety factor for permeable dikes, because a stationary calculation is sufficient. In clay dikes it is useful to take the variation in height and duration into account, while for a sand dike it is sufficient to only consider the variation in height of a flood wave. When the variation of the duration of a flood wave is not considered, it is recommended to use a representative duration of a flood wave; that results in the same total probability of failure as when the variation of the duration is included. At Lobith the duration of the representative flood wave varies from 13 - 16 days. At Borgharen the representative duration varies between the 10 – 11 days for different dike types.

The reliability assessment of hydraulic structures, in primary water defences before 2017, is characterised by exceedance probabilities and implicit uncertainty. In January of 2017, the new water law enforces a change in the assessment of primary water defences. The water law requests a probability of flooding when assessing the safety requirements. The studies of VNK1 and 2 are based on this new safety approach and provide background information and a basis for the new water law. A failure probability and explicit uncertainties provide an increase of reliability in the safety judgement of primary water defences and more specifically hydraulic structures. The aim of a more reliable assessment procedure is to some extent limited by the tools prescribed by the water law. The assessment tool Riskeer does not allow the engineers to have transparency in calculations results, which prevents insight into the model behaviour and contribution to the failure probability. To investigate the mechanics contributing to the overall failure probability for hydraulic structures, analysis of the limit states, failure mechanisms and Riskeer software were performed. The results of the analysis were incorporated into a reference model to validate the suspected mechanics. The input for the reference model was a sensitivity analysis and case study, which cover the assessment tracks of Non-Closure and Strength and Stability for hydraulic structures. The aim of this reference model was not to replace the Riskeer software but describe the mechanics in the correct way. Analysis of the schematisation manual and previous assessment software (Ring toets) provided the incorporated failure mechanisms and the specific model that was used to obtain the limit state. Additionally, a fault tree defined how the different failure mechanisms are related. The knowledge of the fault trees combined with sensitivity analysis provided information about the largest contributors to the overall failure probability, regarding resistance and solicitation variables. The dominant solicitation factor in all failure mechanisms was the water level difference. The resistance factor was dominated by the failure probability of Non-Closure in the gate and the strength of a gate element. In the comparison of the reference model with the Riskeer software, the results were similar but not accurate enough to replace one with another. The reference model was significantly influenced by the approximation in the hydraulic boundary condition, which is formed by assuming a distribution for the water level, wave height and wave period. In the testing of the Riskeer software, the software results showed a difference between two levels of the fault tree. The lower level describes the fault tree including all mechanisms and sub-mechanisms. The top level shows the top probability of the fault tree on which extra simulations of scenarios are executed. The Riskeer software showed the top level as the final result, however, the reference model only describes the lower level. The extra simulations work in a conservative capacity and cannot be assessed by the reference model. The main processes in the fault trees were validated by the reference model, as is shown by their characteristics in the sensitivity analysis. Together with the temporary calculation results, a rough validation of the input parameters was feasible in the lower level. In contrast to the top level, where numerous scenarios were simulated without any transparency to what these values entail. The sensitivity analysis also showed the difference of the two levels over the different variables, which is a (almost constant) significant difference.