Les travaux de recherche présentés dans ce manuscrit ont été menés dans le cadre d’une convention CIFRE avec le Groupe Dacquin. Le sous-sol rennais est composé, en majorité, de schistes datés du Briovérien. Ces terrains, qui peuvent être rencontrés sur une grande partie du massif armoricain, présentent des degrés d’altération et de fracturation extrêmement variables à l’échelle du bassin rennais. Cette variabilité est très difficile à prendre en compte dans le cadre de la réalisation de soutènement, tant au niveau de la conception qu’au niveau de l’exécution. Dans un premier temps, des modélisations basés sur la méthode des éléments finis (logiciel PLAXIS 2D ©) ont été entreprises afin d’évaluer l’influence de la prise en compte d’une anisotropie du terrain sur le comportement de la paroi, en termes de moments et de déplacements. Le modèle utilisé permettait de prendre en compte une anisotropie élastique combinée à un critère orienté de plasticité de type Mohr-Coulomb. Les résultats obtenus avec ce modèle, pour différentes orientations du critère de plasticité, ont été comparés à ceux obtenus avec un modèle élastique isotrope et critère de plasticité de type Mohr-Coulomb. La même étude a été menée avec un second logiciel éléments finis (CESAR-LCPC) afin de comparer les résultats obtenus. Dans les deux cas une influence importante de l’orientation du critère de plasticité a été observée pouvant conduire à des résultats plus défavorables, selon l’orientation, que ceux obtenus avec le modèle isotrope. Ces résultats ont confirmé la nécessité de mieux caractériser l’anisotropie dans des matériaux tels que les schistes Briovériens. La caractérisation du matériau a donc été menée dans un second temps sur des prélèvements répartis sur la ville de Rennes et ses alentours (lors de travaux d’excavations ou sur des talus). Elle a permis de balayer les caractéristiques des schistes de l’échelle microscopique jusqu’à l’échelle macroscopique. La caractérisation microscopique, basée sur des observations au microscope à balayage électronique, des essais de diffraction des rayons X et des observations de lames minces, a permis de mettre en évidence des variations dans la pétrographie des échantillons (schistosité, granulométrie, teneur en argiles) permettant de les différencier en deux types : grès et siltites. La caractérisation à l’échelle macroscopique a été divisée en trois axes : la caractérisation de la matrice rocheuse et de son anisotropie (essais de compression uniaxiale, mesures de vitesses de propagation d’ondes et essais au scléromètre), la caractérisation de discontinuités exprimées (cisaillement direct) et la caractérisation de la résistance à l’usure (essais de broyabilité et Micro-Deval modifiés). Les siltites se sont révélées être extrêmement sensibles à l’eau (forte diminution des caractéristiques mécaniques mesurées) tandis que les échantillons gréseux l’étaient dans une moindre mesure. L’ensemble des essais réalisés a permis d’établir des ordres de grandeurs de variations des propriétés mécaniques des schistes du Briovérien. Enfin la classification des schistes briovériens élaborés par des géotechniciens pour la seconde ligne de métro de Rennes a été complétée en vue de mieux définir les classes intermédiaires de résistance. / Research work presented in this phD thesis was financed by a CIFRE grant given to Groupe Dacquin. Ground from the city of Rennes is, for the majority of it, composed of schists from the Brioverian era. These schists which can be encountered in the entire Armorican massif, present a weathering and a fracturing degree highly variable. This variability is hardly considered for the construction of retaining structure whether during the design step or the execution step. First, modellings using the finite elements method (PLAXIS 2D ©) have been carried out in order to evaluate the influence of ground anisotropy on the retaining wall behaviour (horizontal displacements and bending moments). The model used combined elastic anisotropy with an oriented plasticity criterion (type Mohr-Coulomb). Results obtained with this model, for different orientations of the plasticity criterion, were compared with the ones from the elastic isotropic model (with a Mohr-Coulomb plastic criterion). The same study was carried out with another finite elements software (CESAR-LCPC) in order to compare the results between the two sofwares. In both cases, an important influence of the orientation of the plasticity criterion was noted and lead to higher efforts in the wall, for certain orientations, compared to the isotropic model. These results confirmed the need of a better characterization of the anisotropy in materials such as Brioverian schists. In a second time, the characterization of the material was carried out on samplings spread over the city of Rennes and its area (during excavation work or on natural banks). Characterization was conducted from the microscopic to the macroscopic scale. Microscopic characterization based on observations with a scanning electron microscope, XRD analyses and thin section observations showed variations in the petrography of the samples (schistosity, granulometry, clay content) that lead to the differentiation of two types: sandstone and siltstone. The characterization at the macroscopic scale was divided into three parts: characterization of the rock matrix and its anisotropy (uniaxial compression tests, ultrasonic wave velocities and Schmidt hammer test), characterization of opened discontinuities (direct shear test) and the characterization of wear resistance (grindability and Micro-Deval tests). Siltstone samples were extremely sensitive to the water content (important decrease of the mechanical characteristics measured) whereas sandstones were less sensitive. The tests carried out lead to the definition of ranges for the mechanical properties of the Brioverian schists. Classification of Brioverian schists established by geotechnical engineers for the second subway line of Rennes was completed in order to better define the transitional states of strength of the material.

The Condrey Mountain Schist (CMS) occupies a window through Late Triassic amphibolite facies melange in the north central Klamath Mountains in northern California and southwest Oregon. The schists owe their present level of exposure to a large structural dome centered on the Condrey Mountain Window. Transitional blueschist-greenschist facies assemblages are widespread in mafic schists in the structurally lowest levels of the window; structurally higher CMS near the window margins contains medium- to high-pressure greenschist facies parageneses. An ⁴⁰Ar/³⁹Ar crossite age indicates a late Middle Jurassic age of metamorphism. All subunits of the CMS contain evidence of progressive, polyphase deformational and metamorphic histories. The styles and geometries of minor structures in the central part of the window suggest that early folding and transposition was the result of noncoaxial deformation, and that rotational strains were replaced by irrotational flattening strains with time. Rotational strains were accompanied by the development of epidote-crossite assemblages and the growth of deerite in meta-ironstones; irrotational flattening strains were accompanied and followed by the growth of albite, actinolite, spessartine, and the Ba-silicate, cymrite. Pressure-temperature estimates, the relative ages of mineral growth and deformation, and strain geometries are consistent with, but not restricted to, a subduction zone environment. High shear strains may reflect descent and burial, whereas flattening and late, static mineral growth occur during uplift. Pressure-temperature estimates for the overlying CMS greenschists suggest temperatures similar to those in the central part of the window, but at slightly lower pressures. Thrusting of the overlying amphibolites at 150-156 Ma occurred while the amphibolites were above about 500°C. Stretching lineations indicate a movement vector of about N45W. Comparisons of the sequence and timing of metamorphic and structural events, radiometric ages, and movement directions during thrusting indicate the CMS does not represent an inlier of Klamath Western Jurassic Belt flysch but is instead an older, isolated thrust plate. Similarities with the age of metamorphism and plutonism in the overlying amphibolites suggest the two plates may be remnants of the same Middle Jurassic paired metamorphic belt. / text

The Koongarra uranium deposit is hosted by quartz-chlorite schists. A conceptual model for the hydrogeology of the deposit is proposed on the basis of lithologic criteria and limited hydraulic testing. Water-level and aquifer-test data are presented that indicate the deposit lies within a partially confined, heterogeneous, anisotropic fractured-rock aquifer. The aquifer is dynamic with annual, diurnal, and semidiurnal water-level fluctuations. The results of aquifer tests indicate a high degree of connectivity in the aquifer. Fracture-dominated flow is observed in some tests, but the overall aquifer response appears to be that of an equivalent porous medium. A homogeneous, anisotropic model is used to estimate the transmissivity tensor for subregions of the aquifer. Anisotropy is well-developed with north- to east-northeast-oriented principal transmissivities. Northeast directions represent large-scale drawdown patterns and are subparallel to bedrock structure and the Koongarra fault. Northerly directions are localized and may reflect a less extensive fracture fabric or a flexure in the bedrock foliation.

O alto custo do carvão ativado tem motivado a busca por materiais adsorvedores de baixo custo, como os subprodutos industriais. Neste sentido, o uso dos subprodutos industriais de xisto: finos de xisto (XC), xisto retortado (XR) e xisto retortado com pneus (XRP), proveniente do processo PETROSIX/PETROBRAS, e o catalisador exaurido (CAT), da unidade de FCC (Craqueamento Catalítico em Leito Fluidizado), foram caracterizados e utilizados neste trabalho na adsorção de compostos orgânicos de efluente líquido industrial. O objetivo geral deste estudo é modelar matematicamente o processo de adsorção de compostos orgânicos em xisto, catalisador exaurido de FCC e carvão ativado em pó, utilizando o modelo HSDM (Modelo de Difusão por Superfície Homogênea), e o comportamento hidráulico do sistema adsorvedor. A caracterização estrutural e química de várias amostras de xisto (finos de xisto, xisto retortado e xisto retortado com pneus) e do catalisador exaurido de FCC, além do desenvolvimento de um modelo computacional para a simulação do processo de adsorção são contribuições adicionais ao trabalho. Os resultados de caracterização dos adsorventes de xisto mostram que estes materiais são basicamente macroporosos e com área superficial em torno de 0,51 a 3,36 2 . 1 m g , além de apresentarem as mesmas estruturas cristalinas, e micrografias características dos argilominerais. Já o adsorvente CAT mostrou-se formado basicamente por faujasita, sílica e alumina, apresentando-se na forma de grânulos esféricos, irregulares e microporosos, com área superficial característica de materiais zeolíticos (148 e 155 2 . 1 m g ). Os ensaios de adsorção realizados nos efluentes sintéticos e industriais (Fenólico e Petroquímico) mostraram que existe potencial de aplicação dos subprodutos industriais de xisto e CAT na remoção de compostos orgânicos (corantes, fenóis e COT) destes efluentes. Os resultados das simulações obtidos do modelo proposto mostraram que o tratamento unicamente por adsorção em xisto ou CAT, tendo em vista os padrões ambientais de descarte de efluente (resolução CONAMA 357), é inviável. Isto por que a quantidade de adsorvente requerida é muito elevada. Entretanto, tais adsorventes podem ser utilizados na redução da carga orgânica de ambos os efluentes. Por questões de transporte, é indicado que cada subproduto industrial seja utilizado na própria indústria que lhe deu origem. Ou seja, o CAT é indicado para o tratamento do efluente Petroquímico e os adsorventes de xisto para o Fenólico. / The high cost of the activated carbon has motivated the search of low cost adsorbents such as industrial by-products. In this sense, the use of industrial by-products of oil shale: Oil Shale (XC), Pirolized Oil Shale (XR) and Pirolized Oil Shale with Tires (XRP), from PETROSIX/PETROBRAS, and the spent catalyst (CAT), from FCC (Fluid Catalytic Cracking) unit were characterized and used in this work in the adsorption of organics compounds of industrial liquid effluent. The main contribution of this thesis is to propose a mathematical model to the adsorption process of organic compounds in oil shale, spent catalyst of FCC and powdered activated carbon. This model is based on HSDM (Homogeneous Surface Diffusion Model) model and on the hydraulic behavior of the adsorbent system. Other contribution is the structural and chemical characterization of several samples of oil shale (oil shale, Pirolized oil shale and Pirolized oil shale with Tires) and of the spent catalyst of FCC. A computational model to simulate the adsorption process of these materials is also developed and can be considered an additional contribution of this work. Experimental and simulated results allow characterization of the oil shale adsorbent as basically macroporous and with superficial area about 0.51 to 3.36 m2.g-1. In addition, they present the same crystal structures and clay micrografies characteristics. The adsorbent CAT is composed basically by Faujasite, silica and alumina. They present spherical beads, irregular forms and micropores, with superficial area characteristics of zeolitic materials (148 and 155 m2.g-1). The adsorption tests realized in the synthetic and industrial effluents (Phenolic and Petrochemical) showed the potential of application of the industrial by-products of oil shale and CAT in the removal of organics compounds (dyes, Phenol and COT) of these effluents. In respect of environmental standards to effluents disposal (CONAMA resolution 357), simulations results, obtained with the proposed model, has demonstrated that the use of only oil shale or CAT is not viable, because the high quantity of adsorbents required. However, such adsorbents can be used in the reduction of organic loads in both effluents, when combined with other processes. Each adsorbent showed be used where it is produced because of transportation cost. The CAT is indicated for the treatment of Petrochemical effluent while the others adsorbents may be applied in the phenolic effluent treatment.

O alto custo do carvão ativado tem motivado a busca por materiais adsorvedores de baixo custo, como os subprodutos industriais. Neste sentido, o uso dos subprodutos industriais de xisto: finos de xisto (XC), xisto retortado (XR) e xisto retortado com pneus (XRP), proveniente do processo PETROSIX/PETROBRAS, e o catalisador exaurido (CAT), da unidade de FCC (Craqueamento Catalítico em Leito Fluidizado), foram caracterizados e utilizados neste trabalho na adsorção de compostos orgânicos de efluente líquido industrial. O objetivo geral deste estudo é modelar matematicamente o processo de adsorção de compostos orgânicos em xisto, catalisador exaurido de FCC e carvão ativado em pó, utilizando o modelo HSDM (Modelo de Difusão por Superfície Homogênea), e o comportamento hidráulico do sistema adsorvedor. A caracterização estrutural e química de várias amostras de xisto (finos de xisto, xisto retortado e xisto retortado com pneus) e do catalisador exaurido de FCC, além do desenvolvimento de um modelo computacional para a simulação do processo de adsorção são contribuições adicionais ao trabalho. Os resultados de caracterização dos adsorventes de xisto mostram que estes materiais são basicamente macroporosos e com área superficial em torno de 0,51 a 3,36 2 . 1 m g , além de apresentarem as mesmas estruturas cristalinas, e micrografias características dos argilominerais. Já o adsorvente CAT mostrou-se formado basicamente por faujasita, sílica e alumina, apresentando-se na forma de grânulos esféricos, irregulares e microporosos, com área superficial característica de materiais zeolíticos (148 e 155 2 . 1 m g ). Os ensaios de adsorção realizados nos efluentes sintéticos e industriais (Fenólico e Petroquímico) mostraram que existe potencial de aplicação dos subprodutos industriais de xisto e CAT na remoção de compostos orgânicos (corantes, fenóis e COT) destes efluentes. Os resultados das simulações obtidos do modelo proposto mostraram que o tratamento unicamente por adsorção em xisto ou CAT, tendo em vista os padrões ambientais de descarte de efluente (resolução CONAMA 357), é inviável. Isto por que a quantidade de adsorvente requerida é muito elevada. Entretanto, tais adsorventes podem ser utilizados na redução da carga orgânica de ambos os efluentes. Por questões de transporte, é indicado que cada subproduto industrial seja utilizado na própria indústria que lhe deu origem. Ou seja, o CAT é indicado para o tratamento do efluente Petroquímico e os adsorventes de xisto para o Fenólico. / The high cost of the activated carbon has motivated the search of low cost adsorbents such as industrial by-products. In this sense, the use of industrial by-products of oil shale: Oil Shale (XC), Pirolized Oil Shale (XR) and Pirolized Oil Shale with Tires (XRP), from PETROSIX/PETROBRAS, and the spent catalyst (CAT), from FCC (Fluid Catalytic Cracking) unit were characterized and used in this work in the adsorption of organics compounds of industrial liquid effluent. The main contribution of this thesis is to propose a mathematical model to the adsorption process of organic compounds in oil shale, spent catalyst of FCC and powdered activated carbon. This model is based on HSDM (Homogeneous Surface Diffusion Model) model and on the hydraulic behavior of the adsorbent system. Other contribution is the structural and chemical characterization of several samples of oil shale (oil shale, Pirolized oil shale and Pirolized oil shale with Tires) and of the spent catalyst of FCC. A computational model to simulate the adsorption process of these materials is also developed and can be considered an additional contribution of this work. Experimental and simulated results allow characterization of the oil shale adsorbent as basically macroporous and with superficial area about 0.51 to 3.36 m2.g-1. In addition, they present the same crystal structures and clay micrografies characteristics. The adsorbent CAT is composed basically by Faujasite, silica and alumina. They present spherical beads, irregular forms and micropores, with superficial area characteristics of zeolitic materials (148 and 155 m2.g-1). The adsorption tests realized in the synthetic and industrial effluents (Phenolic and Petrochemical) showed the potential of application of the industrial by-products of oil shale and CAT in the removal of organics compounds (dyes, Phenol and COT) of these effluents. In respect of environmental standards to effluents disposal (CONAMA resolution 357), simulations results, obtained with the proposed model, has demonstrated that the use of only oil shale or CAT is not viable, because the high quantity of adsorbents required. However, such adsorbents can be used in the reduction of organic loads in both effluents, when combined with other processes. Each adsorbent showed be used where it is produced because of transportation cost. The CAT is indicated for the treatment of Petrochemical effluent while the others adsorbents may be applied in the phenolic effluent treatment.