Despite the advantages of using Bayesian networks for probabilistic risk assessment, adoption in practice has been limited due to the lack of realistic, facility-scale studies. Scaling up from systems to facility-level safety assessments poses challenges in (i) integrating external hazards and their cascading effects, and (ii) resolving non-homogeneity of various technical and human reliability models. The novelty of the study is in formalising risk integration using Bayesian networks, at facility scale, and demonstrating its effectiveness in addressing associated challenges. A Bayesian network-based multi-hazard risk framework is introduced and demonstrated for a nuclear power plant subject to flooding and earthquake hazards, capturing dependencies among hazards and consequences. Individual reliability models – conventionally extraneous to facility-wide risk models – are included as subnetworks by using Bayesian network-based surrogate models for technical systems and a Bayesian networks approach for human reliability modelling. Two approaches are used for subnetwork integration – object-oriented and unified Bayesian networks. The unified approach allows for prediction, diagnostics and inter-causal reasoning since Bayesian inference is bi-directional. Conversely, in the object-oriented approach, diagnostics are limited to within individual subnetworks and as a consequence the model can potentially neglect dependencies between objects. However, the object-oriented model requires only 50 % of the computational memory and consumes less than 25% of the runtime as the unified network, while improving visual clarity of the risk model. The model reveals key insights – for example, variations in operator stress or available response time during a hazard event can result in up to a 77 % change in top event probability – demonstrating its effectiveness in capturing critical relationships in complex, facility-scale risk scenarios. These findings can be used to suitably allocate resources towards risk mitigation and plant safety management.

The anisotropic behaviour of sands, which is associated with their grain-scale microstructural characteristics such as the distribution of voids and the spatial orientation of particles, can lead to significant variations in macro-scale predictions. In this paper, a bounding surface plasticity based anisotropic semi-micromechanical constitutive model is developed, within the multilaminate framework, to describe the effects of fabric on the cyclic behaviour of sands. A novel plastic strain driven semi-micromechanical fabric evolution framework fulfilling the premises of anisotropic critical state theory is proposed. Rather than using a single scalar-valued fabric anisotropic variable, which is the general practise in anisotropic critical state theory based models, independently evolving fabric anisotropic variables are employed at so-called sampling planes. In addition, the semifluidised state concept is utilised at low mean effective stresses to realistically capture post-liquefaction responses, including large shear deformations and accumulative plastic strains during flow liquefaction and cyclic mobility types of behaviour. The procedure for calibrating model parameters is briefly described and the prediction capabilities of the proposed model under drained and undrained monotonic and cyclic loading conditions at different stress states, relative densities and loading orientations are demonstrated by simulating experimental data for Toyoura sand using a single set of parameters.

Gas-induced fracturing in liquid-saturated clay-rich materials presents challenges in understanding and predicting fracture behaviour, due to the complex mechanical and transport properties of clays and the compressibility of gas. This paper introduces a novel experimental device for visualising fluid-driven cracks in clays. The device allows for the induction and observation of two-dimensional cracks in clay-rich, low-permeability materials through the injection of gas or water. The experimental setup comprises precision instrumentation for measuring compression forces, displacement, and fluid pressure, along with high-resolution imaging capabilities. Preliminary tests with Helium gas injection into Boom clay samples demonstrate the device's ability to track fracture evolution. This innovative experimental tool offers insights into the mechanisms governing fluid-driven fractures in clay-rich materials and provides a means to validate numerical models.

Geological materials exhibit spatial variability in their properties as a result of their formation. Many studies have focussed on how to characterise this spatial variation by means of the correlation length θ. Such a characterisation has been applied in the geotechnical design of geostructures at numerous sites where cone penetration tests (CPTs) were available, because θ can be relatively easily estimated from this in-situ information. However, the CPTs available at a given site are often part of the initial site investigation, and hence carried out before the application of any ground improvement technique. This raises the question of how (and by how much) the estimated θ is affected by subsequent stages of the construction project and, more specifically, by the application of ground improvement processes intended to alter the initially poor mechanical condition of the in-situ soil. This paper investigates in-situ data from three trial test sites, where CPT data before and after application of vibro-compaction are available. In addition to the expected overall mechanical improvement of the soil, the application of vibro-compaction has a significant impact on soil heterogeneity, with a substantial reduction in the coefficient of variation and θ of the cone tip resistance and sleeve friction.

Three-dimensional and spatial variability effects on slope failure processes are investigated for an idealised slope stability problem with the random material point method (RMPM). A 45 degree slope is brought to failure by either its own weight or by a combination of its own weight and an additional surface load applied at the crest. The ultimate failure load and potential failure processes are studied for various (heterogeneous) material strength profiles. In 3D, failures tend to spread sideways and backwards. For the slope geometry considered, the resistance to initial and secondary failures in 3D simulations tends to be higher than in 2D simulations, probably due to the additional resistance from the ends of the failure surfaces. The failure behaviour changes when a depth trend in the material strength is introduced. A depth trend in the material strength triggers a flow-like failure process, instead of distinct (approximately) circular failure surfaces which are encountered in a material without a depth trend. The flow-like behaviour causes an expansion in the failure zone in all directions while avoiding (where possible) local strong zones.

Free-Field (FF) boundaries have previously been developed to replicate the (infinite) far-free-field domain in the simulation of earthquake loading problems. Although they can yield accurate results under certain conditions, it has been observed that significant problems can occur if the behaviour of the material near the boundaries is highly non-linear or incorporates cyclic attributes, or if the boundaries are located close to the domain of interest. To address the inaccuracies caused by the use of traditional FF boundaries, a novel technique is proposed: Tied Free-Fields. This technique combines the principles of both standard earthquake boundary conditions (that is, Tied-Degree (TD) and FF boundaries) to accommodate earthquake loading at the domain boundaries in a direct and economical fashion. The proposed solution has been tested using one and two-dimensional benchmarks and an advanced constitutive model. The results show a significant improvement in accuracy over traditional FF boundaries in the modelling of surface settlements and computed energy released, as well as a significant improvement in computational efficiency over TD boundaries in the modelling of asymmetric problem domains.

Capturing the spatial variability in soil is crucial for ground response analyses in the context of seismic hazard mitigation. The lateral variability in thickness and properties of the different soil layers is one of the main factors that determines the variability of the ground motion spectrum from one location to another. The absence of such lateral variability information in the subsoil in between the locations of Cone Penetration Tests (CPTs) may be compensated by the use of more densely sampled seismic data. In this research we aim to derive a shear-wave velocity field through seismic full-waveform inversion that yields a model resolution approaching that of high-resolution seismic CPT surveys. Following this, a datadriven correlation between geophysical and geotechnical information is attempted through the application of new machine-learning-based approaches.

Stratification identification and spatial interpolation play a fundamental role in geotechnical site characterization. A unified approach is needed to perform these two tasks simultaneously to reduce overall uncertainty in site characterization. This paper explores the applicability of the Mixture of Gaussian Processes (MoGP) to address this gap, with a specific focus on characterizing and completing missing CPT data. The investigation encompasses both synthetic and real-world field CPT datasets and includes a comparison of the MoGP's interpolation accuracy with the use of a single GP for entire datasets. Additionally, the study examines the sensitivity of the model's performance with respect to the number of training data points. Although the interpolation performance of the MoGP model is promising with synthetic data, limitations appear in its application to real-site CPT data.

This paper investigates the implementation of a nonlocal regularisation of the material point method to mitigate mesh-dependency issues for the simulation of large deformation problems in brittle soils. The adopted constitutive description corresponds to a simple elastoplastic model with nonlinear strain softening. A number of benchmark simulations, assuming static and dynamic conditions, were performed to show the importance of regularisation, as well as to assess the performance and robustness of the implemented nonlocal approach. The relevance of addressing stress oscillation issues, due to material points crossing element boundaries, is also demonstrated. The obtained results provide relevant insights into brittle materials undergoing large deformations within the MPM framework.

As the Material Point Method (MPM) uses both a mesh and a point discretisation scheme, the application of boundary conditions is difficult, currently limiting the flexibility of the method. While many boundary condition options have been used in the literature, the accuracy of Neumann boundary condition options has not yet been studied. Four options have here been evaluated for 1D and 2D benchmarks, although none of the options were found to be both accurate and generally applicable in MPM. However, for the generalised interpolation material point method (GIMP), the application of surface tractions on support domain boundaries or on a detected surface are valid options. Large differences between these two accurate options and the application of tractions at surface material points, a method regularly used in the literature, have been observed.

The absence of information on lateral variability in the soil is detrimental to estimating accurately the local site response in the event of an earthquake. To address this problem, the use of densely sampled seismic data together with sparsely distributed but detailed vertical soil profiles obtained from cone penetration tests (CPTs) is advantageous. This study explores the adaptation of suitable machine learning (ML) approaches to derive reliable, site- and depth-specific correlations between seismic shear-wave velocity (Vs) and cone-tip resistance (qc). Such correlation could be successfully established by combining information from seismic CPT surveys with available borehole information for the Groningen region in the Netherlands. It is found that, even over substantial distances, ML-based techniques offer site- and depth-specific correlations between Vs and qc.

Uncertainty is inevitable in the characterisation of a geotechnical site, especially due to the inherently heterogeneous nature of the ground. In this paper, a method for characterising a subsurface with limited cone penetration test (CPT) data is proposed. The method is based on integrating predictions of CPT parameters with a probabilistic approach for subsoil classification at the CPTs. The predicted stratigraphy is able to capture the spatial variability of soil measured via CPTs and takes account of the uncertainties that arise from transforming CPT measurements into soil units as well as errors in the measurements themselves. The applicability of the proposed method is demonstrated for a site in the Netherlands. The results show that the proposed approach can identify the most likely classification in the domain with good accuracy. Furthermore, the significance of considering the uncertainties in predicting the most likely classification is illustrated via finite element stability analyses of a slope cut-out in the domain.

Soil liquefaction is investigated considering a saturated soil deposit and by implementing standard techniques of random field theory to distribute initial void ratio values and assess liquefaction risk. The soil domain is represented in a 2-dimensional (2D) random finite element model for the dynamic analysis of coupled behavior. Multiple Monte Carlo realizations are subjected to a base acceleration, while cyclic and small strain soil behaviours are achieved through a hypoplastic constitutive model. This investigation demonstrates that 2D stochastic simulations converge to 2D deterministic simulations when small standard deviations and/or small scales of fluctuation are used. However, large standard deviations combined with relatively large scales of fluctuation may cause significant uncertainty in the response of the soil deposit. Finally, common techniques employed to assess soil liquefaction are evaluated based on the results of the deterministic and random field analyses.

In this paper, a state-dependent semi-micromechanical framework for anisotropic sands is proposed. A simple constitutive model based on critical state theory and bounding surface (BS) plasticity is used to describe idealized micro-level soil behaviour, and a slip theory based multilaminate framework employed to create a link between the micro and macro level responses of soil. A contact normal based second order fabric tensor is used to create a mathematical description of the anisotropic nature of sand. The proposed constitutive framework can reproduce various soil responses, stemming from both the inherent anisotropy which highly depends on the sample preparation method and induced anisotropy resulting from the applied stress path. This paper presents concise theoretical aspects of the multilaminate framework and the anisotropic elastoplastic constitutive formulation. Finally, the model's performance in predicting sand response is demonstrated under drained and undrained conditions at different stress states, relative densities and loading conditions by simulating Karlsruhe sand, and is examined through a comparison with two other sophisticated constitutive models for sand, namely the Dafalias and Manzari (2004) version of Sanisand and hypoplasticity with intergranular strain.

To protect embankments along German inland waterways against local slope sliding failure caused by ship-induced water level drawdown, they are mainly secured by bank revetments. Often, large embankment sections are designed on the basis of a limited number of field and laboratory tests. Thus, uncertainties arise with regard to the mechanical and hydraulic ground properties. Current design standards account for these uncertainties by conservative design assumptions and empirical knowledge. This paper investigates the effects of vertically non-homogeneous ground properties on the required armour layer thickness using 1D random fields and an infinite slope model, which was modified to account for ship-induced drawdowns. Within the limitations of the infinite slope assumptions, the effects of a spatially variable friction angle and hydraulic conductivity are investigated and compared to deterministic benchmark cases. The investigations show that the level of safety obtained with the deterministic design depends strongly on the choice of the characteristic values. Particularly, the hydraulic conductivity determines the reliability of the design. In some cases, the 5 % quantile of the hydraulic conductivity does not yield a conservative estimate of the required armour layer thickness. In the case of the effective friction angle, the 5 % quantile may overestimate the required armour layer thickness for permeable soils. For less permeable soils, the 5 % quantile meets the solution of the random field analyses. For the combination of random effective friction angle and random hydraulic conductivity, all investigated benchmark studies seem to ensure engineering safety, but on different reliability levels. Based on these findings, recommendations regarding site exploration and choice of characteristic values of hydraulic conductivity and effective friction angle are provided.

The coupling effect of initial shear stress and thermal cycles on the thermomechanical behaviour of clay concrete and sand-concrete interfaces has been studied. A set of drained monotonic direct shear tests was conducted at the soil-concrete interface level. Samples were initially sheared to half of the material's shear strength and then they were subjected to five heating/cooling cycles before being sheared to failure. The test results showed that the effect of thermal cycles on the shear strength of the materials was negligible, yet shear displacement occurred during application of thermal cycles without an increase in shear stress, confirming the coupling between the shear stress and temperature. In addition, a slight increase of stiffness due to the coupling was observed which diminished with further shearing.

Geological uncertainty can significantly influence the computed response of a geotechnical structure. For example, ignoring the presence of a weak soil layer embedded within a stronger layer and assuming a deterministic stratigraphic boundary can significantly underestimate the probability of failure. In this paper, the coupled Markov chain method has been used for modelling this form of uncertainty. A strategy for estimating the horizontal transition probability matrix with limited data has been proposed, which is one of the biggest challenges with using this method. In particular, different sampling intervals in the vertical and horizontal directions have been considered in estimating the matrix for simulating realistic field situations. The applicability of the proposed method has been demonstrated using a set of CPTs in the Netherlands. The results highlight a problem that arises due to the coupling algorithm used in this method.

This research focuses on investigating the relative performance of a range of machine learning algorithms, namely the artificial neural network, support vector machine, Gaussian process regression, random forest, and XGBoost, for predicting the undrained shear strength from cone penetration test data. This is to assess how machine learning could help us lower the need for laboratory test data. The training dataset compiles 526 data from 12 regions and the testing dataset consists of 20 data from a polder located close to Leiden in the Netherlands. In addition, k-fold and group k-fold cross-validation strategies are both applied to validate the models. The poor performance of the models during group k-fold cross-validation suggests that, while machine learning techniques can perform well when site-specific data are included during training, they struggle to generalize without site-specific data. This highlights the difficulty of capturing soil heterogeneity and suggests that either machine learning methods should be trained on specific sites for which some data are already available, or much larger training datasets are needed.

Natural soil deposits may possess a highly anisotropic nature. The fabric anisotropy of soils which is induced during the soil formation process can lead to severe variation in field scale responses. Although the influence of fabric on the response of sands is well known and several advanced constitutive models have been developed to account for it, most of the studies incorporating anisotropy have focused on element test simulations while practical boundary value problem simulations are usually omitted. In this paper, the undrained response and liquefaction resistance of anisotropic sand deposits with different inherent fabric anisotropies are numerically investigated through element test simulations and one-dimensional nonlinear effective stress site response analyses. A novel semi-micromechanical constitutive model accounting for the effect of fabric anisotropy on sand liquefaction has been incorporated into a fully coupled dynamic in-house code employing the u-p formulation. The proposed numerical framework shows that, in both element test simulations and site response analyses, the fabric effects stemming from both the inherent and induced anisotropies can significantly influence the liquefaction resistance of sands.

Experimental studies show that initial fabric and its evolution under different stress paths greatly influences soil behaviour. Even though different sample preparation methods create different inherent anisotropies and cause different material responses, the same initial fabric structure under different stress paths also results in different material behaviours. In this paper, a simple state-dependent, bounding surface-based elastoplastic constitutive model, which can simulate the anisotropic nature of sands including the effect of principal stress rotation, is described. The model is developed based on a semi-micromechanical concept within the multilaminate framework and, to include the inherent anisotropy of sand, a deviatoric fabric tensor describing the initial microstructure is introduced. In addition, a fabric evolution rule compatible with anisotropic critical state theory is employed to describe the evolving fabric structure and induced anisotropy towards the critical state. In contrast to the classical strain-driven formulation for fabric evolution, a micro-level evolution rule is proposed. This paper presents concise theoretical aspects of the multilaminate framework and the anisotropic elastoplastic constitutive formulation. The model's capability under drained and undrained monotonic loading conditions at different stress states, relative densities and principal stress orientations is demonstrated by simulating experimental data for Toyoura sand.