Gelfill (GF) is made of tailings, water, binder and chemical additives (Fillset, sodium silicate gel). The components of GF are combined and mixed on the surface and transported (by gravity and/or pumping) to the underground mine workings, where the GF can be used for both underground mine support and tailings storage. Thermal (T), hydraulic (H), and mechanical (M) properties are important performance criteria of GF. The understanding of these engineering properties and their evolution with time are still limited due to the fact that GF is a new cemented backfill material. In this thesis, the evolution of the thermal, hydraulic, mechanical, and microstructural properties of small GF samples are determined. Various binder contents of Portland cement type I (PCI) are used. The GF is cured for 3, 7, 28, 90, and 120 days. It is found that the thermal, hydraulic and mechanical properties are time-dependent or affected by the degree of binder hydration index. Furthermore, a relationship is found between the compressive strength and the saturated hydraulic conductivity of the GF samples. The unsaturated hydraulic properties of GF samples have also been investigated. The outcomes show that unsaturated hydraulic conductivity is influenced by the degree of binder hydration index and binder content, especially at low suction ranges. Simple functions are proposed to predict the evolution of air-entry values (AEVs), residual water content, and fitting parameters from the van Genuchten model with the degree of hydration index (α). Furthermore, two columns are built to simulate the coupled thermo-hydro-mechanical (THM) behaviour of GF under drained and undrained conditions. The obtained results from the GF columns are compared with the small samples. It is observed that the mechanical properties, hydraulic properties (suction and water content), and temperature development are strongly coupled. The magnitude of these THM coupling factors is affected by the size of the GF. The findings also show that the mechanical, hydraulic and thermal properties of the GF columns are different from samples cured in plastic moulds.

Le stockage géologique du CO2 dans des réservoirs aquifères de type calcaire et grès, du charbon non exploité est une des solutions envisagées pour réduire les émissions de gaz à effet de serre dans l’atmosphère. Cependant, l’injection de CO2 peut perturber les propriétés pétrophysiques (porosité et perméabilité), minéralogiques (transformations) et mécaniques (déformations, résistance à la rupture) des roches réservoir (calcaire, grès, charbon). Dans le cas du charbon, l’injection de CO2 peut également se traduire par des phénomènes de gonflement de la matrice liés au processus d’adsorption. L’objectif de ce travail de thèse est de traduire en termes de modèles phénoménologiques les comportements et les couplages chimio-poromécaniques des roches réservoir de type charbon. Dans ce travail, nous nous sommes focalisés en particulier sur l’étude de l’injection de CO2 dans le charbon. Pour cela, deux modèles homogénéisés de porosité du charbon ont été développés avec la prise en compte du phénomène d’adsorption, connu pour être le principal mécanisme de production ou de séquestration de CO2 dans de nombreux réservoir de charbon. Le premier modèle permet d’étudier le comportement poro-élastique du charbon pour une injection simple de CO2 et le second permet d’étudier le comportement poro-élastique du charbon pour une injection de CO2 avec une récupération assistée de méthane CH4. Le processus d’adsorption est classiquement modélisé à l’aide de l’isotherme d’adsorption de Langmuir (pour un gaz dans le premier modèle et pour deux gaz dans le second modèle). L’implantation de ces modèles dans le Code_Aster (code d’analyse de calcul de structures entièrement couplé THM, développé par Electricité De France - EDF) nous a permis de faire des simulations numériques de stockage de CO2 dans le charbon. Pour une injection simple du CO2 dans le charbon (premier modèle), la matrice du charbon s’est comportée de deux façons différentes : elle gonfle (ce qui induit une diminution de la porosité du charbon) avec la prise en compte du phénomène d’adsorption et se contracte (ce qui induit une augmentation de la porosité du charbon) dans le cas contraire. Etant en bon accord avec les résultats de la littérature spécialisée, cela montre la capacité du modèle à prédire le comportement poro-élastique du charbon durant l’injection de CO2. Toujours avec le premier modèle, nous avons en particulier étudié l’influence des propriétés hydro-mécaniques du charbon (coefficient de Biot, module de Young/module d’incompressibilité), les paramètres d’adsorption de Langmuir et la pression initiale du liquide interstitiel dans le charbon, sur la réponse du charbon à l’injection du CO2. Dans le cas d’une récupération assistée du méthane CH4 (le second modèle), un couplage du Code_Aster et un code de transport réactif HYTEC (HYdrological Transport coupled with Equilibrium Chemistry, développé par MINES Paris Tech) était nécessaire pour gérer surtout le calcul des pressions partielles des deux gaz (CO2 et CH4) à chaque pas de temps. Un travail de développement numérique sur les deux codes de calcul était alors nécessaire. Ce travail de thèse a proposé une méthode de couplage entre les deux codes (Code_Aster et HYTEC) dont les techniques sont largement décrites dans le manuscrit. / The geological storage of CO2 in aquifers reservoirs such as limestone and sandstone, coal is a possible way to reduce greenhouse gas emission into the atmosphere. However, the injection of CO2 may modify petrophysical (porosity and permeability), mineralogical (transformations) and mechanical (deformations, strength) properties of reservoir rocks (limestone, sandstone, coal). In the case of coal, the injection of CO2 can also induce matrix swelling due to adsorption processes. The focus of this thesis is to translate in terms of phenomenological models, the behaviors and chemo-poromechanical coupling of reservoir rocks of coal type. In this work, we focused particularly on the study of CO2 injection into coal. For this, two models of homogenized coal porosity have been developed by taking into account the adsorption phenomenon, known to be the main mechanism of production or sequestration of CO2 in many coal reservoirs. The first model allows the study of the poroelastic behavior of coal in the case of a single injection of CO2, and the second model allows the study of the poroelastic behavior of coal in the case of an injection of CO2 with methane CH4 recovery. The adsorption process is classically modelled using Langmuir’s isotherm (for one gas in the first model and for two gases in the second model). The implementation of these models in Code_Aster (a fully coupled Thermo-Hydro-Mechanical analysis code for structures calculations, developed by Electricity of France - EDF) allowed us to make numerical simulations of CO2 storage in coal. For a single injection of CO2 into coal (first model), the coal matrix behaved in two different ways: it swells (resulting in the decrease of coal porosity) when the adsorption phenomenon is taken into account and shrinks (resulting in the increase of coal porosity) otherwise. Being in good agreement with the results in specialized literature in this field, it shows the ability of the model to predict the poroelastic behaviour of coal to CO2 injection. Also with the first model, we studied particularly through numerical simulations the influence of coal’s hydro-mechanical properties (Biot’s coefficient, bulk modulus), Langmuir’s adsorption parameters and the initial liquid pressure in rock mass during CO2 injection in coal. In the case of methane recovery (second model), a coupling of Code_Aster and a reactive transport code, HYTEC (Hydrological Transport coupled with Equilibrium Chemistry, developed by Mines Paris Tech) was needed to handle the above calculation of partial pressures of the two gases (CO2 and CH4) at each time step. Digital development work on the two computers codes (Code_Aster and HYTEC) was then necessary. This thesis proposed a method of coupling between the two codes whose techniques are widely described in the manuscript.

Durant les différentes phases du stockage profond des déchets radio-actifs exothermiques (excavation, exploitation) jusqu'à la fermeture définitive, la roche hôte sera soumise à des sollicitations mécaniques, hydriques et thermiques couplées. Afin de connaître et de modéliser le comportement à court et long terme des dispositifs de stockage, une investigation approfondie du comportement de la roche est nécessaire afin de compléter les données existantes. C'est dans ce but que cette étude du comportement thermo-hydro-mécanique de la formation argileuse du Callovo-Oxfordien (COx) considérée par l'ANDRA comme roche hôte potentielle, a été développée. Dans un premier temps le comportement en compression–gonflement de l'argilite du COx a été étudié par la réalisation d'un programme d'essais oedométriques haute pression. Les résultats, interprétés en termes de couplage endommagement-gonflement, ont montré que l'ampleur du gonflement était lié à la densité de fissuration engendrée lors de la compression. Dans un second temps, le comportement hydromécanique et thermo-hydro-mécanique de l'argilite saturée sous une contrainte moyenne proche de l'in-situ a été étudié à l'aide de cellules à faible chemin de drainage (10 mm), dont une cellule isotrope et une cellule triaxiale à cylindre creux avec mesures locales d'un type nouveau. Ces appareils ont permis de résoudre deux problèmes difficiles typiques des argilites de très faible perméabilité : i) une saturation préalable correcte, attestée par de bonnes valeurs du coefficient de Skempton et ii) de bonnes conditions de drainage. Les paramètres caractéristiques du comportement de l'argilite à température ambiante (coefficients de Skempton et de Biot, compressibilité drainée et non-drainée) ont été déterminés à partir d'essais de compression isotrope qui ont également confirmé l'isotropie transverse du matériau. La cohérence des paramètres obtenus a été vérifiée dans un cadre poro-élastique saturé. Deux aspects du comportement thermo-hydro-mécanique de l'argilite du COx ont été étudiés à l'aide d'essais de chauffage et de compression volumique en température (80°C) : les effets de la température sur le comportement intrinsèque de l'argilite et le phénomène de pressurisation thermique. Un essai de chauffage drainé sous contrainte in-situ a mis en évidence, apparemment pour la première fois, un comportement plastique contractant de l'argilite (comme les argiles normalement consolidées), non pris en compte dans les modélisations thermo-élastiques actuelles des systèmes de stockage profond. Un autre élément nouveau et important observé est l'augmentation de la compressibilité avec la température, à la différence des argiles. L'étude de la pressurisation thermique (engendrée par la faible perméabilité de l'argilite et la forte différence entre les coefficients de dilatation thermique de l'eau et de la phase solide), a été réalisée à l'aide d'essais de chauffage non drainés, suite à une analyse détaillée des effets perturbateurs du système de mesure lors de variations de pression et de température (peut-être à considérer également dans les mesures in-situ). Le coefficient de pressurisation thermique s'est révélé être très sensible aux variations de température et de contrainte, il diminue de 0,14 à 0,1 MPa/°C entre 25 et 80°C. La nature des différentes réponses hydro-mécaniques et thermo-hydro-mécaniques obtenues au cours de ce travail permettront une interprétation et une modélisation plus précises du comportement du massif d'argilite autour des galeries, dans des zones qui sont pour la plupart saturées, sauf à proximité immédiate des galeries (quelques décimètres) / During the different phases of the exothermic radioactive waste deep disposal (excavation, operation) and after permanent closure, the host rock is submitted to various coupled mechanical, hydraulic and thermal phenomena. Hence, a thorough investigation of the thermo-hydro-mechanical behaviour of the rock is necessary to complete existing data and to better understand and model the short and long term behaviour of the Callovo-Oxfordian (COx) clay formation in Bure (Meuse/Haute-Marne - M/HM), considered by ANDRA as a potential host rock in France.In this work, the compression – swelling behaviour of the COx claystone was first investigated by carrying out a series of high-pressure oedometric tests. The results, interpreted in terms of coupling between damage and swelling, showed that the magnitude of swelling was linked to the density of the fissures created during compression. In a second step, the hydromechanical and thermo-hydro-mechanical behaviour of the saturated claystone under a mean stress close to the in situ one were investigated by using two devices with short drainage path (10 mm), namely a isotropic cell and a newly designed hollow cylinder triaxial cell with local displacement measurements. These devices helped to solve two majors problems related to testing very low permeability materials : i) a satisfactory previous sample saturation (indicated by good Skempton values) and ii) satisfactory drainage conditions. Some typical constitutive parameters (Skempton and Biot's coefficients, drained and undrained compressibility coefficients) have been determined at ambient temperature through isotropic compression tests that also confirmed the transverse isotropy of the claystone. The consistency of the obtained parameters has been checked in a saturated poroelastic framework. Two aspects of the thermo-hydro-mechanical behaviour of the COx claystone have then been investigated through different heating tests and through drained and undrained isotropic compression tests at elevated temperature (80°C) : the effects of temperature on the behaviour of claystone and thermal pressurization. A drained heating test under in-situ stress conditions evidenced, probably for first time, a plastic contractant response of the claystone (like normally consolidated clays), a feature not considered in the presently conducted numerical modelling of deep disposal systems. Another new important observed feature is the increase in drained compressibility of the COx claystone with temperature, not observed in clays. The investigation of thermal pressurization (caused by the low claystone permeability and by the significant difference in thermal expansion between water and the solid phase) has been carried out by means of undrained heating tests, after a detailed analysis of the major effects of the measurement system (which should perhaps be also analyzed when performing in-situ measurements). The thermal pressurization coefficient appeared to be quite sensitive to changes in temperature and stress, it decreased between 0.14 and 0.1 MPa/°C between 25 and 80°C. It is believed that the different thermo-hydro-mechanical volumetric responses obtained here allow a better interpretation and modelling of the behaviour of the claystone formation around the galleries in areas that are mostly saturated, except close to the galleries (a few decimetres)

L'étude du comportement du béton en température est un problème majeur qui vise in fine à évaluer le niveau de sécurité des structures sous des sollicitations thermiques sévères, lors d'incendies par exemple. Pour cela, de nombreux modèles sont développés dans un cadre de couplage thermo-hydro-mécanique (THM), pour prendre en compte les différents processus physico-chimiques et mécaniques mobilisés par ces sollicitations et conditionnant la stabilité du matériau béton.Cependant, ces modélisations ne prennent souvent pas en compte explicitement la nature hétérogène du matériau. En effet, le béton est un matériau constitué d'inclusions noyées dans une matrice cimentaire pouvant être schématisée comme un milieu poreux ouvert partiellement saturé en eau. Les inclusions sont caractérisées par leurs natures minéralogiques, leurs morphologies et leurs tailles classées en fuseaux granulaires. Cette hétérogénéité introduit une complexité supplémentaire : la nécessité de prendre en compte la microstructure pour quantifier l'effet de l'incompatibilité (thermique, hydrique et mécanique) inclusion-matrice sur le comportement THM du béton. Ce travail constitue une première avancée dans ce sens. A ce titre, une modélisation élément fini multi-échelle tridimensionnelle (3D) est développée permettant d'affecter des comportements spécifiques à la matrice et aux inclusions. Pour la matrice, siège des transports de masse dans son réseau poreux connecté, une approche thermo-hydrique à trois fluides (eau, vapeur et air sec) est adoptée et est couplée à une loi de comportement poro-mécanique endommageable. Les inclusions (granulats) sont considérées hydriquement inertes, une approche thermo-mécanique avec endommagement est alors retenue.Une analyse, par simulations numériques, des effets de la nature minéralogique des granulats (calcaires ou silico-calcaires) de leurs distributions et de leurs morphologies a été menée sur des configurations 2D et 3D. Les effets étudiés ont notamment concerné l'influence de ces paramètres sur les fluctuations locales des champs de température, de pression de gaz et d'endommagement au regard de la dispersion des mesures expérimentales. L'analyse est limitée à l'échelle mésoscopique, celle où les granulats de taille caractéristique supérieure à 5 mm sont considérés, le reste (stable et pâte de ciment) étant une matrice homogène. Enfin, cette analyse a mis en évidence le besoin de mettre en œuvre une approche expérimentale cohérente avec une analyse mutli-échelle, à la fois pour la caractérisation des propriétés (thermiques, hydriques et mécaniques) de chaque constituant et pour l'étude des évolutions des champs lors des changements d'échelles. Un protocole expérimental a été définit et des premiers résultats d'essais sont présentés et analysés au travers de résultats obtenus dans la partie modélisation / The investigation of the behavior of heated concrete is a major research topic which concerns the assessment of safety level of structures when exposed to high temperatures, for instance during a fire. For this purpose, several modeling approaches were developed within thermo-hydro-mechanical (THM) frameworks in order to take into account the involved physic-chemical and mechanical processes that affect stability of heated concrete. However, existing models often do note account explicitly for the heterogeneity of the material : concrete is composite material that may be schematized as an assembly of inclusions (aggregates) embedded in a cementitious matrix (cement paste). This latter may be described as a partially saturated open porous medium. The aggregates are characterized by their mineralogical nature together with their morphology and size distribution. The material heterogeneity bring an additional complexity : the need to take into account the microstructure in order to quantify the effect of matrix-inclusion thermal, hygral and mechanical incompatibilities on the THM behavior of concrete. This work is a first step in this direction. For this purpose, a three-dimensional (3D) multi-scale finite element model is developed. It allows affecting specific behaviors to matrix and inclusions. For the former, where mass transports occur within the connected porous network, a three-fluids approach (liquid water, vapor and dry air) is adopted and is coupled to a poro-mechanical damage based approach. For inclusions (aggregates) no hygral component arises a pure thermo-mechanical model is considered. The developed model is then used to investigate, either by 2D or 3D numerical simulations, effects of mineralogical nature, morphology and distribution of aggregates. Studied effects have mainly concerned the influence of these parameters on local fluctuations of simulated temperature, gas pressure and damage fields with regard to experimentally observed dispersion. The analysis is here limited to the mesoscale, at which only aggregates with a characteristic size above 5 mm are meshed while the remaining inclusions together with the cement paste are considered to be a homogeneous matrix. Finally, the numerical analysis carries out the need to perform an experimental campaign that is consistent with a multi-scale approach of the THM behavior of concrete : an experimental campaign that allows to identify thermal, hygral and mechanical properties of each concrete constituent and that permit to assess evolution of fields during upscalling. An experimental protocol is then elaborated for this purpose and some obtained results are presented and analyzed with regard to results obtained in the modeling part

La durabilité des structures en béton est désormais intégrée dans la démarche de conception des ouvrages de Génie Civil. En effet, quel que soit le type de sollicitation (mécanique, thermique, hydrique) une fissuration est susceptible de se produire risquant d'impacter la durée de vie de l'ouvrage par la pénétration d'agents agressifs. L'aptitude au service peut elle-même être affectée pour les structures où une étanchéité est requise (enceinte de confinement de centrales nucléaires, réservoirs de gaz naturel liquéfié, barrages, stockages des déchets radioactifs ou de CO2, etc.). Dans ce contexte industriel, la prédiction du débit de fuite traversant des éléments composés de matériaux à base cimentaire est donc un enjeu scientifique et industriel majeur. Pour parvenir à cet objectif de simulation numérique, il est nécessaire de mettre en place un couplage hydro-mécanique. L'anisotropie de la fissuration induite par les sollicitations mécaniques complexes conduit à un tenseur de perméabilité macroscopique anisotrope. La détermination de ce tenseur est un enjeu important dans l'objectif de mener des calculs à l'échelle macroscopique avec des modèles phénoménologiques. De plus, les calculs de perméabilité sont un moyen de comparer les volumes fissurés obtenus par les différents modèles mécaniques. La modélisation de la fissuration pour les matériaux quasi-fragiles hétérogènes à l'échelle mésoscopique tels que le béton est complexe et suivant les approches utilisées, les résultats peuvent fortement varier. C'est pourquoi l'étude numérique proposée dans la thèse comporte une comparaison entre deux approches mécaniques : - une première basée sur une modélisation mécanique de type E-FEM (Embedded Finite Element Method) [Benkemoun et al., 2010] - - une seconde basée sur une modélisation mécanique d'endommagement [Mazars, 1984] régularisée en énergie de fissuration [Hillerborg et al., 1976]. Le travail numérique associé à cette thèse consiste donc à développer un modèle couplant de manière faible un modèle mécanique à un modèle de transfert en 3D à l'échelle mésoscopique. En se basant sur le concept de « double porosité », la perméabilité du milieu fissuré est vue comme la combinaison d'une perméabilité diffuse et isotrope (liée au réseau poreux initial du béton et à son degré de saturation) et d'une perméabilité « discrète » et orientée au sein des fissures (le calcul de cette dernière étant basé sur les ouvertures de fissures données par le modèle mécanique et sur les équations de la mécanique des Navier-Stokes en régime permanent). La comparaison des résultats obtenus sur différents résultats expérimentaux issus de la littérature (un tirant traversé par de l'eau [Desmettre et Charron, 2011] et un élément structurel traversé par de l'air sec [Nahas et al., 2014]) permet de comparer la pertinence des deux modèles mécaniques utilisés ainsi que l'approche utilisée pour estimer le débit traversant des éléments en béton fissurés. / The durability of concrete structures is nowadays fully integrated in the civil engineering constructions design process. Whatever the loading is (mechanical, thermic, hydric), cracks may appear and impact the structure lifespan by the infiltration of aggressive agents. The serviceability can be directly impacted for the structures playing an air/water tightness role (containment building nuclear power plants, liquefied natural gas storage tanks, dams, radioactive waste disposal, etc.). The prediction of the flow going through elements composed of a cementitious material is therefore a major scientific and industrial issue. To achieve this goal, a hydro-mechanical coupling must be implemented. The anisotropic cracking induced by complex mechanical loadings leads to an anisotropic macroscopic permeability tensor. This tensor computation is an important issue dealing with phenomenological models for macroscopic problems. The cracking modelling of quasi-brittle materials, heterogeneous at the mesoscopic scale like concrete, is complex and different mechanical approaches can lead to various results. Therefore, permeability calculations are an elegant way to examine cracking patterns obtained with several mechanical models. Consequently, this study compares two mechanical approaches: - the first one is based on an Embedded Finite Element Method (E-FEM) mechanical model [Benkemoun et al., 2010] - - the second one is based on a damage mechanical model [Mazars, 1984] regularised by the fracture energy of the material [Hillerborg et al., 1976]. This thesis presents a hydro-mechanical approach weakly coupling a mechanical model with a permeation model in 3D at the mesoscopic scale. This work is based on the “double porosity” concept splitting the permeability into two parts: the first one is isotropic and corresponds to flows within the porosity of the material- the second one, based upon a set of cracks with different orientations and openings, is anisotropic. For the latter, each crack is a path for mass flow according to the fluid laws considering two infinite planes. In order to check this approach relevance, numerical results are compared to experimental results extracted from the literature (an experiment where water goes through a specimen made of a steel reinforcing bar covered with concrete under load [Desmettre et Charron, 2011] and a device where dry air goes through a structural element made of reinforced concrete [Nahas et al., 2014]). The computation of the flow going to those cracked concrete elements helps to understand the presented approach efficiency and the differences between the two used mechanical models.

A l’heure actuelle, des sédiments fluviaux et lacustres se trouvent pollués par les polychlorobiphényles (PCB), à des degrés de gravité divers. Les PCB se présentent comme des micropolluants persistants dans l’environnement et nuisibles à la santé. Un programme de recherche « PCB-Axelera » est conduit afin de pencher sur cette problématique et explorer les solutions de diminution de cette pollution. En France, l'ensemble des arrêtés préfectoraux (30/10/2002 et 14/06/2000) et le décret du 18/04/2002 considèrent les sédiments pollués comme déchets et autorisent leur stockage dans les installations de stockage de déchets si leurs teneurs en PCB sont comprises entre 0,68 ppm et 50 ppm. L’objectif principal de cette étude est de concevoir une méthodologie adaptée pour une installation de déchets non-dangereux ISDND. Les géomatériaux dans ce stockage seront utilisés comme barrière d’étanchéité pour les sédiments pollués. L’étude est composée de trois parties : Dans la première partie la caractérisation hydromécanique chimique des sédiments pollués par les PCB est présentée. La classification des sédiments est effectuée. Le comportement mécanique du sédiment est déterminé à l’aide de l’œdomètre. Quant au comportement hydromécanique il est étudié à l’aide du filtre presse et du perméamètre à paroi rigide (PPR). Cette caractérisation est indispensable pour proposer la meilleure mise en œuvre des sédiments dans le centre de stockage. Dans la deuxième partie, les performances hydraulique (perméabilité) et mécanique (gonflement) des différents géomatériaux en contact avec un fluide extrait du sédiment et d’autres fluides synthétiques contenant des PCB sont explorées. L’étude du gonflement des géomatériaux est réalisée à l’échelle du laboratoire tandis que, l’étude de sa performance hydrique est réalisée sur deux échelles centimétrique et métrique. L’étude menée sur un pilote de 1,2 m3 de volume comprenant un système de pluviation automatique a permis de suivre les tassements à l’aide d’un capteur ultrason et de mesurer la perméabilité des sédiments et des géomatériaux sous conditions réelles de stockage. Dans la troisième partie, un modèle hydraulique 1D est développé afin de suivre le front d’eau au sein d’un échantillon de sédiment soumis à une charge hydraulique représentant l’expérimentation dans un PPR. La méthode VER est utilisée. Les résultats expérimentaux montrent que le degré de contamination en PCB de ces sédiments n’influence pas le comportement hydromécanique et donc la qualité d’étanchéité des géomatériaux étudiés. Les résultats numériques de simulation du comportement hydraulique du sédiment sont en concordance avec les résultats expérimentaux. La solution de stockage hors site des sédiments pollués est alors envisageable puisqu’elle confine la pollution en éliminant le risque de contamination du milieu environnant. / Currently, river and lake sediments are contaminated with polychlorinated biphenyls (PCBs), to varying degrees of severity. PCBs are persistent micropollutants in the environment and harmful to health. A research program "PCB-Axelera" is conducted to address this issue and explore solutions to eliminate or decrease the pollution. In France, all prefectural orders (30/10/2002 and 14/06/2000) and the Decree of 18/04/2002 considered contaminated sediments as waste and allow their landfilling in waste storage facility if their PCB concentrations are between 0.68 ppm and 50 ppm. The main objective of this study is to develop a methodology suitable for installation of non-hazardous waste landfill sites. Geomaterials in this storage will be used as a barrier for confinement of contaminated sediment. The study consists of three parts: 1. In the first part, hydro-chemical characterization of sediments contaminated with PCBs is presented. The sediment classification is performed. The mechanical behavior of the sediment is determined using oedometer tests. As for the hydro-mechanical behavior, it is studied using the filter press and the rigid wall permeameter (PPR). This characterization is essential to provide the best implementation of sediment in the storage center. 2. In the second part, the hydraulic performance (permeability) and mechanical (swelling) of various geomaterials in contact with a fluid extracted from the sediment and other synthetic fluids containing PCBs are being explored. The study of swelling of geomaterials is performed on a laboratory scale while studying its water performance was achieved on two centimetric and metric scales. The pilot scale study was conducted on a 1.2 m3 volume tank. The pilot test was equipped with an automatic rain system and an ultrasonic sensor to measure to measure the follow the settlements of the sediment. The permeability of sediments and geomaterials were measured using the pilot test under actual storage conditions . 3. In the third part, a 1D hydraulic model is developed to follow the waterfront in a sediment sample subjected to a hydraulic load representing a PPR experiment. The VER method is used. A water model of the sediment is proposed to reproduce the experimental results. The water retention curve of the sediment is calculated using Durner’s model (1994). The experimental results show that the degree of PCB contamination of the sediments does not affect the hydro-mechanical behavior and therefore the hydro-mechanical performance of the studied geomaterials studied. Numerical simulation results of the hydraulic behavior of sediment are consistent with the experimental results. The solution of landfilling the PCB contaminated sediment is possible since it assure the confinement of the pollution and eliminates the risk of environmental contamination.

The vast majority of oil and gas reserves are trapped in fractured carbonate reservoirs. Most carbonate reservoirs are naturally fractured, with fractures ranging from millimeter- to kilometer-scale. These fractures create complex flow behaviors which impact reservoir characterization, production performance, and, eventually, total recovery. As we know, bridging the gas from plug to near-wellbore, eventually to field scales, is a persisting challenge in modeling Naturally Fractured Reservoirs (NFRs). This dissertation will focus on assessing the fundamental flow mechanisms in fractured rocks at the plug scale, understanding the governing upscaling parameters, and ultimately, developing fit-for-purpose upscaling tools for field-scale implementation. In this dissertation, we first focus on the upscaling of rock fractures under the laminar flow regime. A novel analytical model is presented by incorporating the effects of normal aperture, roughness, and tortuosity. We then investigate the stress-dependent hydraulic behaviors of rock fractures. A new and generalized theoretical model is derived and verified by a dataset collected from public experimental resources. In addition, an efficient coupled flow-geomechanics algorithm is developed to further validate the proposed analytical model. The physics of matrix-fracture interaction and fluid leakage is modeled by a high-resolution, micro-continuum approach, called extended Darcy-Brinkman-Stokes (DBS) equations. We observe the back-flow phenomena for the first time. Machine learning is then implemented into our traditional upscaling work under complex physics (e.g., initial and Klinkenberg effects). We finally consolidate the lab-scale upscaling tools and scale them up to the field scale. We develop a fully coupled hydro-mechanical model based on the Discrete-Fracture Model (DFM) in fractured reservoirs, in which we incorporate localized effects of fracture roughness at the field-scale.

Material point method (MPM) is a numerical tool which was originally used for modelling large deformations of solid mechanics problems. Due to the particle based spatial discretiza- tion, MPM is naturally capable of handling large mass movements together with topological changes. Further, the Lagrangian particles in MPM allow an easy implementation of history dependent materials. So far, however, research on MPM has been mostly restricted to explicit dynamic formu- lations with linear approximation functions. This is because of the simplicity and the low computational cost of such explicit algorithms. Particularly in MPM analysis of geomechan- ics problems, a considerable attention is given to the standard explicit formulation to model dynamic large deformations of geomaterials. Nonetheless, several limitations exist. In the limit of incompressibility, a significantly small time step is required to ensure the stability of the explicit formulation. Time step size restriction is also present in low permeability cases in porous media analysis. Spurious pressure oscillations are another numerical instability present in nearly incompressible flow behaviours. This research considers an implicit treatment of the pressure in MPM algorithm to simu- late material incompressibility. The coupled velocity (v)-pressure (p) governing equations are solved by applying Chorin’s projection method which exhibits an inherent pressure stability. Hence, linear finite elements can be used in the MPM solver. The main purpose of this new MPM formulation is to mitigate artificial pressure oscillations and time step restrictions present in the explicit MPM approach. First, a single phase MPM solver is applied to free surface incompressible fluid flow problems. Numerical results show a better approximation of the pressure field compared to the results obtained from the explicit MPM. The proposed formulation is then extended to model fully saturated porous materials with incompress- ible constituents. A solid velocity(v S )-fluid velocity (v F )-pore pressure (p) formulation is presented within the framework of mixture theory. Comparing the numerical results for the one-dimensional consolidation problem shows that the proposed incompressible MPM algorithm provides a stable and accurate pore pressure field even without implementing damping in the solver. Finally, the coupled MPM is used to solve a two-dimensional wave propagation problem and a plain strain consolidation problem. One of the important features of the proposed hydro mechanical coupled MPM formulation is that the time step size is not dependent on the incompressibility and the permeability of the porous medium.

La modélisation des massifs rocheux fracturés est un problèmes important dans de nombreux secteurs industriels, y compris, mais sans s'y limiter à l'exploitation pétrolière et gazière. Dans la littérature, les roches fracturées sont reconnues comme des milieux à double porosité et double perméabilité dans lesquels le réseau de fractures fournit la perméabilité primaire et la matrice rocheuse la perméabilité secondaire. L'idée de la dissociation de l'écoulement à l'intérieur du réseau de fractures et de la matrice,la double perméabilité, est toujours contestée pour les réservoirs fracturés. De nombreuses contributions sur cette question ont été présentées dans la littérature et les méthodes utilisées pourraient être classées dans deux approches principales : approches continues et discontinues. Chaque approche a ses avantages et ses limites. Pour surmonter les limites en gardant les avantages de ces deux approches, une approche nommée Embedded Fracture Continumm Approach (EFCA) qui emprunte le concept du modèle continu et intègre également l'effet des fractures explicites est considérée dans cette thèse. L'idée principale de cette approche repose sur le concept de la « cellule fracturée » représentant un milieu poreux qui a ses propres propriétés calculées à partir des propriétés de la matrice poreuse et des fractures qui la traversent. Le code de calcul développé dans le cadre de ce travail est basé sur la bibliothèque source DEAL.II. L'exactitude de l'EFCA a été étudiée à travers de différents tests. Plusieurs applications traitées dans ce travail comme la détermination des propriétés hydro-mécaniques effectives d'un site réel, estimation de la production de puits dans laquelle les fractures sont modélisées explicitement, démontrent la performance de l'EFCA dans la modélisation des roches fracturées ainsi que l'effet de la double porosité et de la double perméabilité aux comportements des réservoirs fracturés. / Fractured rock masses modeling is a challenge issue in many field of industry including but not limited to oiland gas exploitation. In the literature, fractured rock masse are in many cases recognized as double permeability medium in which fracture network provides the primary permeability and rock matrix plays asthe second one. The idea of dissociation of flow inside the fracture network and the matrix, the double permeability, is still challenged for fractured reservoirs. Numerous contributions on this issue have been presented in the past could be cast in two main approaches: continuum media approach and discontinuous approach. Each approach has its advantages and limitations. To overcome the limitation and to take advantage of these two approaches, the Embedded Fractured Continuum Approach (EFCA) which borrows the concept of continuum models and also incorporates the effect of explicit fractures is considered in this thesis. The principal idea of this approach lies on the concept of fracture cell representing a porous medium that has their own properties calculated from the properties of porous matrix and fractures intersecting it.The development in this work was conducted by using the library source code DEAL.II. The accuracy of EFCA was investigated through different verifications. Through some applications: determination of effective hydro-mechanical properties of an actual site, estimation of well production in which necessary fractures are modeled explicitly, we demonstrate the performance of the EFCA in the modeling fracture drock masses as well as the effect of double porosity and double permeability on behaviours of fractured reservoirs.

Deep geological repository (DGR) for disposal of high-level radioactive waste (HLW) is designed to rely on successive superimposed barrier systems to isolate the waste from the biosphere. This multiple barrier system comprises the natural geological barrier provided by the repository host rock and its surrounding and an engineered barrier system (EBS). The EBS represents the synthetic, engineered materials placed within the natural barrier, comprising array of components such as waste form, waste canisters, buffer materials, backfill and seals. The buffer will enclose the waste canisters from all directions and act as a barrier between canisters and host rock of the repository. It is designed to stabilise the evolving thermo-hydro-mechanical-chemical stresses in the repository over a long period (nearly 1000 years) to retard radionuclides from reaching biosphere. Bentonite clay or bentonite-sand mix have been chosen as buffer materials in EBS design in various countries pursuing deep geological repository method. The bentonite buffer is the most important barrier among the other EBS components for a geological repository. The safety of repository depends to a large extent on proper functioning of buffer over a very long period of time during which it must remain physically, chemically and mineralogically stable. The long term stability of bentonite buffer depends on varying temperature and evolution of groundwater composition of host rocks in a complex way. The groundwater in the vicinity of deep crystalline rock is often characterized by high solute concentrations and the geotechnical engineering response of bentonite buffer could be affected by the dissolved salt concentration of the inflowing ground water. Also during the initial period, radiogenic heat produced in waste canisters would radiate into buffer and the heat generated would lead to drying and some shrinkage of bentonite buffer close to canister. This could alter the dry density, moisture content and in turn the hydro-mechanical properties of bentonite buffer in DGR conditions. India has variety of bentonite deposits in North-Western states of Rajasthan and Gujarat. Previous studies on Indian bentonites suggest that bentonite from Barmer district of Rajasthan (termed as Barmer-1 bentonite) is suitable to serve as buffer material in DGR conditions. Nuclear power agencies of several countries have identified suitable bentonites for use as buffer in DGR through laboratory experiments and large scale underground testing facilities. Physico-chemical, mineralogical and engineering properties of Kunigel VI, Kyungju, GMZ, FoCa clay, MX-80, FEBEX and Avonseal bentonites have been extensively studied by Japan, South Korea, China, Belgium, Sweden, Spain, Canada. It is hence essential to examine the suitability of Barmer-1 bentonite as potential buffer in DGR and compare its physico-chemical and hydromechanical properties with bentonite buffers identified by other countries. The significant factors that impact the long-term stability of bentonite buffer in DGR include variations in moisture content, dry density and pore water chemistry. With a view to address these issues, the hydromechanical response of 70 % Barmer-1 bentonite + 30 % river sand mix (termed bentonite enhanced sand, BES specimens) under varying moisture content, dry density and pore water salt concentration conditions have been examined. The broad scope of the work includes: 1) Characterise the physico-chemical and hydro-mechanical properties of Barmer-1 bentonite from Rajasthan, India and compare its properties with bentonite buffers reported in literature. 2) Examine the influence of variations in dissolved salt concentration (of infiltrating solution), dry density and moisture content of compacted BES specimens on their hydro-mechanical response; the hydro-mechanical properties include, swell pressure, soil water characteristic curve (SWCC), unsaturated hydraulic conductivity, moisture diffusivity and unconfined compression strength. Organization of thesis: After the first introductory chapter, a detailed review of literature is performed to highlight the need for detailed characterisation of physico-chemical and hydromechanical properties of Barmer-1 bentonite for its possible application in DGR in the Indian context. Further, existing literature on hydro-mechanical response of bentonite buffer to changes in physical (degree of saturation/moisture content, dry density) and physico-chemical (solute concentration in pore water) is reviewed to define the scope and objectives of the present thesis in Chapter 2. Chapter 3 presents a detailed experimental programme of the study. Chapter 4 characterises Barmer-1 bentonite for physico-chemical (cation exchange capacity, pore water salinity, exchangeable sodium percentage) and hydro-mechanical properties, such as, swell pressure, saturated permeability, soil water characteristic curve (SWCC) and unconfined compression strength. The properties of Barmer-1 bentonite are compared with bentonite buffers reported in literature and generalized equations for determining swell pressure and saturated permeability coefficient of bentonite buffers are arrived at. Chapter 5 describes a method to determine solute concentrations in the inter-lamellar and free-solutions of compacted BES (bentonite enhanced sand) specimens. The solute concentrations in micro and macro pore solutions are used to examine the role of osmotic flow on swell pressures developed by compacted BES specimens (dry density 1.50-2.00 Mg/m3) inundated with distilled water and NaCl solutions (1000-5000 mg/L). The number of hydration layers developed by the compacted BES specimens on inundation with salt solutions in constant volume swell pressure tests is controlled by cation hydration/osmotic flow. The cation hydration of specimens compacted to dry density of 2.00 Mg/m3 is mainly driven by matric suction prevailing in the clay microtructure as the number of hydration layers developed at wetting equilibrium are independent of the total dissolved solids (TDS) of the wetting solution. Consequently, the swell pressures of specimens compacted to 2.00 Mg/m3 were insensitive to the salt concentration of the inundating solution. The cation hydration of specimens compacted to dry density of 1.50 Mg/m3 is driven by both matric suction (prevailing in the clay micro-structure) and osmotic flow as the number of hydration layers developed at wetting equilibrium is sensitive to the TDS of the wetting solution. Expectedly, the swell pressures of specimens compacted to 1.50 Mg/m3 responded to changes in salt concentration of the inundating solution. The 1.75 Mg/m3 specimens show behaviour that is intermediate to the 1.50 and 2.00 Mg/m3 series specimens. Chapter 6 examines the influence of initial degree of saturation on swell pressures developed by the compacted BES specimens (dry density range: 1.40- 2.00 Mg/m3) on wetting with distilled water from micro-structural considerations. The micro-structure of the bentonite specimens are examined in the compacted and wetted states by performing X-ray diffraction measurements. The initial degree of saturation is varied by adding requisite amount of distilled water to the oven-dried BES mix and compacting the moist mixes to the desired density. The montmorillonite fraction in the BES specimens is responsible for moisture absorption during compaction and development of swell pressure in the constant volume oedometer tests. Consequently, it was considered reasonable to calculate degree of saturation based on EMDD (effective montmorillonite dry density) values and correlate the developed swell pressure values with degree of saturation of montmorillonite voids (Sr,MF). XRD measurements with compacted and wetted specimens demonstrated that if specimens of density series developed similar number of hydration layers on wetting under constant volume condition they exhibited similar swell pressures, as was the case for specimens belonging to 1.40 and 1.50 Mg/m3 series. With specimens belonging to 1.75 and 2.00 Mg/m3 series, greater number of hydration layers were developed by specimens that were less saturated initially (smaller initial Sr,MF) and consequently such specimens developed larger swell pressures. When specimens developed similar number of hydration layers in the wetted state, the compaction dry density determined the swell pressure. Chapter 7 examines the influence of salt concentration of infiltrating solution (sodium chloride concentration ranges from 1000- 5000 mg/L) on SWCC relations, unsaturated permeability and moisture diffusivity of compacted BES specimens. Analysis of the experimental and Brooks and Corey best fit plots revealed that infiltration of sodium chloride solutions had progressively lesser influence on the micro-structure and consequently on the SWCC relations with increase in dry density of the compacted specimens. The micro-structure and SWCC relations of specimens compacted to 1.50 Mg/m3 were most affected, specimens compacted to 1.75 Mg/m3 were less affected, while specimens compacted to 2.00 Mg/m3 were unaffected by infiltration of sodium chloride solutions. Variations in dry density of compacted bentonite impacts the pore space available for moisture flow, while, salinity of wetting fluid impacts the pore structure from associated physico-chemical changes in clay structure. Experimental results showed that the unsaturated permeability coefficient is insensitive to variations in dry density and solute concentration of wetting liquid, while, the effective hydraulic diffusivity is impacted by variations in these parameters. Chapter 8 summarises the major findings of the study.