Cette thèse de doctorat porte sur la modélisation numérique du comportement des sables à l'aide de la loi de comportement NorSand-aUL. Cette loi, développée dans le cadre de ce projet de recherche, est une évolution du modèle NorSand spécifiquement adaptée à la modélisation du comportement des sables soumis à des sollicitations cycliques. La révision des performances de NorSand sous différents types de sollicitations (triaxial compression, cisaillement simple statique et cyclique, cisaillement à direction et ratio des contraintes principales contrôlés) a permis l'identification de certains points faibles du modèle, rendant ses modélisations du comportement des sables sous sollicitations cycliques insatisfaisantes. Afin de corriger ces lacunes, deux groupes de modifications ont été proposés. D'abord, la théorie de l'état critique anisotrope a été implémentée dans NorSand, afin d'y créer une dépendance vis-à-vis du type et de la direction des chargements. Cette modification a grandement amélioré les performances du modèle pour la modélisation du comportement encisaillement simple statique. La deuxième modification proposée s'est articulée autour de la formulation d'une nouvelle mécanique de génération de la plasticité lors du déchargement des sables. L'existence d'une deuxième surface de plasticité, nichée à l'intérieur de la surface de plasticité originale de NorSand, a été postulée. Les capacités de modélisation d'une version de NorSand incorporant cette nouvelle mécanique ont été confirmées grâce à la modélisation d'essais triaxiaux drainés et non drainés comprenant des phases de déchargement. Les deux modifications proposées au modèle NorSand au cours de ce projet de recherche ont finalement été agrégées pour mener à la formulation de NorSand-aUL. Ce nouveau modèle a été utilisé pour la modélisation d'essais de cisaillement simple cyclique effectués sur deux sables. Ces modélisations ont démontré les gains réalisés, particulièrement pour les essais effectués sur sable lâche, par rapport aux performances antérieurement obtenues à l'aide de la version originale de NorSand. Des axes de recherche future ont été proposés afin d'améliorer les performances de NorSand-aUL, notamment pour la modélisation du comportement des sables denses soumis à des sollicitations cycliques. / This PhD thesis deals with the numerical modelling of sand behaviour using the NorSandaUL constitutive law. This law, developed as part of this research project, is an evolution ofthe NorSand model specifically adapted to the modelling of sand behaviour under cyclicloading. The review of NorSand's performance under various types of loading (triaxial compression, static and cyclic simple shear, fixed principal stress direction and fixed principal stress ratio) has allowed the identification of certain weak points in the model, making its behaviour modellings under cyclic loading unsatisfactory. In order to address these short comings, two groups of modifications were proposed. First, the anisotropic criticalstate theory was implemented in NorSand, in order to create a dependency on the type and direction of loading. This modification greatly improved the performance of the model for predicting static simple shear behavior. The second proposed modification focused on the formulation of a new mechanics for the generation of plasticity during unloading. The existence of a second yield surface, nested within the original NorSand yield surface, was postulated. The modelling capabilities of a version of NorSand incorporating this new mechanics were confirmed by satisfactory modelling results of drained and undrained triaxialtests which included unloading phases. The two modifications to NorSand proposed over the course this research project were eventually aggregated to lead to the formulation of NorSand-aUL. This new model was used to model cyclic simple shear tests for two sands. These modelling results demonstrated the gains made, particularly for tests conducted on loose sands, compared to the performance previously obtained using the original version of NorSand. Areas for future research were proposed to improve the performance of NorSandaUL, particularly in modelling the behaviour of dense sands under cyclic loading.

In this thesis, the failures that occurred during the construction of the Jamuna Bridge Abutment in Bangladesh have been investigated. In particular, the influence of heterogeneity on slope stability has been studied using statistical methods, random field theory and the finite element method. The research is divided into three main parts: the statistical characterization of the Jamuna River Sand, based on an extensive in-situ and laboratory database available for the site; calibration of the laboratory data against a double-hardening elastoplastic soil model; and stochastic finite element slope stability analyses, using a Monte Carlo simulation, to analyse the slope failures accounting for heterogeneity. The sand state has been characterised in terms of state parameter, a meaningful quantity which can fully represent the mechanical behaviour of the soil. It was found that the site consists of predominantly loose to mildly dilative material and is very variable. Also, a Normal distribution was found to best represent the state parameter and a Lognormal distribution was found to best represent the tip resistance.The calibration of the constitutive model parameters was found to be challenging, as alternative approaches had to be adopted due to lack of appropriate test results available for the site. Single-variate random fields of state parameter were then linked to the constitutive model parameters based on the relationships found between them, and a parametric study of the abutment was then carried out by linking finite elements and random field theory within a Monte Carlo framework.It was found that, as the degree of anisotropy of the heterogeneity increases, the range of structural responses increases as well. For the isotropic cases, the range of responses was relatively smaller and tended to result in more localised failures. For the anisotropic cases, it was found that there are two different types of deformation mechanism. It was also found that, as the vertical scale of fluctuation becomes bigger, the range of possible structural responses increases and failure is more likely. Finally, it was found that the failed zones observed during the excavation of the West Guide Bund of the Jamuna Bridge Abutment could be closely predicted if heterogeneity was considered in the finite element analyses. In particular, it was found that, for such a natural deposit, a large degree of anisotropy (in the range of 20) could account for the deformation mechanisms observed on site.

Efter det dammbrott som skedde vid Brumadinho i Brasilien under januari 2019, där hundratals människor miste sina liv, har frågor kring statisk liquefaction hos anrikningssand fått stort fokus inom gruvdammsindustrin.  Statisk liquefaction kan modelleras med den konstitutiva modellen NorSand, där NorSand-parametrarna kan utvärderas från triaxialförsök. I det här arbetet har ett laboratorieprogram bestående av siktning, sedimentationsanalys, korndensitet, ödometerförsök och triaxialförsök utförts och NorSand-parametrar utvärderats. Fokus ligger på genomförande av laboratorieförsök samt utvärdering av parametrar. Målet är att redogöra för olika osäkerheter som kan påverka resultatet för att på så sätt reda ut vilka de känsligaste faktorerna är, och identifiera vilka moment under utförandet som är särskilt kritiska. Totalt har fyra dränerade och sex odränerade triaxialförsök utförts. Baserat på dessa har NorSand-parametrar utvärderats grafiskt samt genom att matchning av modellerade triaxialförsök i PLAXIS Soiltest mot de utförda försöken. Samtliga utvärderade parametrar ligger inom förväntat spann baserat på litteratur. Flertalet kritiska moment har identifierats avseende genomförandet av triaxialförsök. Vid utvärdering av NorSand-parametrar är bestämningen av portal det klart mest kritiska. Nästan samtliga NorSand-parametrar relateras till de utvärderade portalen. Portalen kan i sin tur bestämmas på olika sätt med olika osäkerheter. För att kunna värdera resultatet är det viktigt att veta vilken metod som använts och hur de känsliga momenten hanterats. På grund av hur kritiskt det är med korrekta portal så kan det vara en bra idé att använda två metoder och sedan jämföra resultaten med hänsyn till deras osäkerheter.  Samtliga NorSand-parametrar förutom Poissons tal bygger på subjektiv tolkning. Antingen utvärderas parametrarna grafiskt baserat på linjer anpassade av individen, eller så matchas modellerade försök till de utförda och då är det upp till individen att balansera försöken mot varandra. För att minimera utrymmet för individuell tolkning är det viktigt att tillräckligt många försök utförs så att potentiella uteliggare kan identifieras. Det är också viktigt att kommande försök planeras för att täcka in de spänningar och portal där det saknas information. / After the tragic failure of the Brumadinho tailings facility in Brazil during January 2019, where hundreds of people lost their life, the question of static liquefaction has received a lot of focus in the tailings industry.  Static liquefaction can be modelled using the constitutive model NorSand, where the parameters are evaluated from triaxial tests. The focus of this work has been on laboratory work and evaluation of NorSand-parameters. The laboratory work consisted of sieving, sedimentation analysis, particle density, oedometer tests, and triaxial tests from which the NorSand-parameters were evaluated. The aim is to identify uncertainties during triaxial testing and evaluation of NorSand-parameters, from which critical factors can be determined. In total, six undrained and four drained triaxial tests have been performed.  Multiple critical factors have been identified during the execution of triaxial tests. In relation to the evaluation of NorSand-parameters, the determination of void ratios is the most significant. For the resulting values to be assessable, it is important to know which method was used and how uncertainties were managed. Because of how critical correct void ratios are, it might be a good idea for them to be determined and compared using different methods. All NorSand-parameters except one are based on a certain amount of subjective interpretation. The person conducting the evaluation either manually fits lines to the test data or tries to match modelled triaxial tests to the ones performed by balancing the tests against each other. To minimize the impact of the individual it is important to carry out enough tests for potential outliers to be identified. It is also of importance to plan upcoming tests to cover stresses and void ratios with lacking data.

The coastal population is growing, putting extra stress on coastal sediments and protection features, such as beach dunes. Moreover, global warming will increase the frequency of storms, and coastal dunes and other defense infrastructure will be subjected to increased erosion and scouring, endangering the people they are meant to protect. Understanding soil dynamics and fluid interaction is crucial to predict the effects of sand erosion. In particular, the study of wind erosion of sands in coastal dunes is essential due to the protective role these earthen structures have during storm events. One of the challenges about predicting wind erosion in coastal dunes is its extended spatial scale and the associated economic and logistics costs of sampling and characterizing the sediments. Because of this, in-situ testing for sediment characterization is essential. In particular, the usage of free-fall penetrometers (FFP) is appealing due to their portability and robustness. The sediment properties obtained with this type of testing can later be used to assess wind erosion susceptibility by determining, for example, the wind velocity to initiate the erosion process. FFP testing involves dropping an instrumented probe that impacts the soil and measures the kinematics or kinetics during the penetration process. For example, deceleration measurements are used to compute an equivalent quasi-static failure, which is not in line with the dynamic process characteristic of FFP testing. This preassumed failure mechanism is used to back-calculate the sand's geomechanical properties. However, soil behavior is highly complex under rapid loading, and incorporating this behavior into FFP sediment characterization models is challenging. Advanced numerical modeling can improve the understanding of the physics behind FFP testing. This thesis presents various advancements in numerical modeling and erosion models to bridge FFP in-situ testing with predicting the initiation of wind erosion of sands. First, improvements oriented to the Material Point Method (MPM) for modeling in-situ FFP testing are proposed. The numerical results show that the simulation of FFP deployment in sands is affected by strain localization and highlight the importance of considering constitutive models sensitive to different loading rates. Because of the importance of rate effects in soil behavior, the second aspect of this thesis proposes a novel consistency framework. Two constitutive models are adapted to study strain-rate sensitive non-cohesive materials: i) a strain-softening Mohr-Coulomb, and ii) a NorSand model. In addition to increased strength, the proposed framework captures increased dilatation, an early peak deviatoric stress, and relaxation. Finally, a novel sand erosion model is derived using a continuum approximation and limit equilibrium analysis. The erosion law considers geotechnical parameters, the effects of slope, and moisture suction, in a combined manner. The proposed model is theoretically consistent with existing expressions in the literature. It covers a wide range of environmental and geometrical conditions and helps to reconcile the results from FFP testing with the prediction of the initiation of wind erosion. The model was validated in a wind tunnel and is demonstrated to be a viable alternative for predicting sand erosion initiation. This thesis opens up new research prospects, such as improving the soil characterization models or the direct prediction of sand erosion using rapid, reliable, and efficient in-situ testing methods. / Doctor of Philosophy / With global warming and climate change, it is expected that the frequency and intensity of storms will increase. This increment will put extra stress on coastal sediments such as beach sand and coastal dunes, making them prone to erosion. Coastal dunes lose their ability to withstand storms as they erode, potentially making coastal flooding more frequent. In light of this, all stakeholders involved in the protection against coastal disasters must have the tools to predict, prepare for, and mitigate for situations like the ones stated above. An essential aspect of the prediction component is dependent on a successful sediment characterization, for example, determining how much wind the sand can withstand before it erodes. Free-fall penetrometers (FFP) are devices designed to conduct the characterization mentioned above. However, the procedures used to perform this characterization are mainly based on empirical or semi-empirical expressions. Computer models, capable of simulating the physics behind FFP testing, can bring more insight into the process of interaction between FFP devices, sands, and water and can be the basis to improve the characterization methods. The latter results can be utilized for instance to predict wind erosion, including several properties of the sand, such as its mineralogy and shape. This study contributes to developing the computer simulations of FFP deployment and the wind erosion prediction models. Eventually, these developments can help engineers and coastal managers to anticipate and prepare for more frequent coastal hazards.