On désigne habituellement par retrait-gonflement des sols argileux les variations de volume qu’un massif d’argile peut subir sous l’effet des variations de la pression interstitielle négative (succion) de l’eau, lors d’une humidification ou d’un séchage. L’objectif de ce travail est de compléter les connaissances expérimentales sur le comportement hydromécanique des sols gonflants rencontrés dans les travaux de génie civil, de géotechnique et de géotechnique de l’environnement afin de mieux en tenir compte dans les modèles de calcul. De nombreux travaux ont montré l'intérêt de réaliser des essais œdométriques pour caractériser le potentiel et la cinétique de gonflement des sols. Ils ont révélé également l’influence de l’humidification ou du séchage sur le gonflement ou le retrait, avec des pertes ou des augmentations de rigidité des éprouvettes. En revanche, l’influence de plusieurs cycles de séchage/humidification sur le comportement mécanique n’a été que très rarement analysée. Cette thèse rapporte les résultats d’études expérimentales effectuées à l’œdomètre avec imposition de succion par la méthode osmotique sur des matériaux gonflants compactés lâches et naturels denses. Plusieurs cycles de séchage/humidification ont été appliqués sur ces matériaux sous trois faibles charges mécaniques constantes. Pendant ces cycles, les éprouvettes manifestent un retrait cumulé pour le sol lâche et un gonflement cumulé pour le sol dense. Les résultats montrent que dans les deux cas, les déformations volumiques convergent vers un état d’équilibre où le sol présente un comportement réversible. A la fin des cycles de succion, un cycle de chargement/déchargement a été effectué sous les succions constantes. Les valeurs de la pression de préconsolidation p0(s), de l’indice de compression vierge [lambda(s]] et de l’indice de compression élastique [kappa] dépendent directement des chemins de contrainte suivis. L’ensemble des résultats expérimentaux permet de déterminer les surfaces de charge : la limite de séparation de micro/macro (Lm/M) ; la surface de chargement-effondrement (LC : Loading Collapse du modèle BBM) et la surface de comportement saturé (SCS). La succion limite entre la micro- et la macrostructure (Lm/M) dépend parfaitement de la structure interne et du diamètre qui délimite les deux familles de pores. L’évolution de la pression de préconsolidation en fonction de la succion imposée est présentée par la surface LC. Les courbes de compressibilité sous différentes succions convergent vers la courbe correspondant à l’état saturé sous de fortes contraintes appliquées. La pression à partir de laquelle, le sol continue son chemin sur la courbe du comportement normalement consolidé est appelée la pression de saturation (Psat). Plus la succion imposée est élevée, plus la charge nécessaire pour atteindre cette pression de saturation est importante. La surface SCS présente la variation de la pression de saturation en fonction de la succion imposée. Nous pouvons considérer que les surfaces de charge SCS et LC sont uniques pour les sols denses cependant elles se superposent à la fin des cycles de succion pour les sols lâches. Les cycles hydriques augmentent aussi la limite (Lm/M) entre la micro- et la macrostructure pour les deux sols / We usually define the swelling-shrinkage of the swelling soils by the volume variation of a clayey layer exposed to the negative water pore pressure (suction) variations during the wetting and drying periods. In this research, we try to complete our experimental knowledge on the hydromechanical behaviour of the swelling soils used in civil engineering, geotechnical engineering and geoenvironmental engineering, to better modelize these soils in our numerical calculations. Several authors have used the oedometer tests to characterize the capacity and the process of the soil expansion. They have also studied the influence of the wetting and drying on the soil swelling or shrinkage which can decrease or increase the soil rigidity. However, the influence of several hydraulic cycles on the mechanical behaviour has been rarely studied. This thesis presents an experimental study performed on compacted loose and natural dense expansive soils using osmotic oedometers. Several successive cycles were applied under three different low constant vertical net stresses. The loose soil presents a significant shrinkage accumulation while the dense one produces the swelling accumulation during the suction cycles. The suction cycles induced an equilibrium stage which indicates an elastic behaviour of the samples. At the end of suction cycles, a loading/unloading test was performed at the constant suctions for both materials. The mechanical parameters, i.e. the virgin compression index [lambda(s)], the apparent preconsolidation stress p0(s) and the elastic compression index values [kappa] ?are completely dependent on the followed stress paths. The whole experimental results made it possible to define the yielding surfaces: suction limit between micro and macrostructure (Lm/M), loading collapse (LC) and saturation curve (SCS). The suction limit (Lm/M) depends completely to the soil fabrics and to the diameter separating the micro- and macrostructure. The preconsolidation stress variation with suction is represented by the LC surface. The compression curves at different imposed suctions converge towards the saturated state for the high applied vertical stresses. We consider the saturation pressure (Psat) as the necessary pressure to reach the saturated state for an imposed suction. The higher the suction, the higher the saturation pressure. The yielding surface representing this pressure as a function of suction is called the saturation curve (SCS). Generally we can state that the suction cycles unified the LC and SC surfaces and increased the (Lm/M) up to a higher value

[pt] A necessidade da determinação do comportamento tensão-deformação de solos sob carregamento para fins de projeto de Engenharia exige o conhecimento prévio do estado natural de tensões ao que o solo está imposto. Quando se trata de solos residuais tropicais, este desafio amplia-se ainda mais devido à ação do intemperismo e variação de umidade. O estado natural do solo é usualmente expresso pelo coeficiente de empuxo no repouso (K0). Para a determinação de K0 três classes de métodos são utilizadas: métodos de correlação, de determinação in situ e em laboratório. Neste estudo, é utilizada uma célula rígida com sistema de controle de sucção para a obtenção de K0 em laboratório. Foram analisadas três amostras de solo residual de Gnaisse com diferentes graus de intemperismo, provenientes de São Conrado, Rio de Janeiro, denominadas: solo residual jovem (SRJ); solo residual de transição (SRT); e solo residual maduro (SRM). Para a análise da influência da sucção no valor do coeficiente de empuxo, cada amostra de solo foi ensaiada sob três diferentes condições de umidade, que tentaram representar as seguintes condições: mais úmida, com sucção de 10 kPa; intermediária, com 40 kPa de sucção; e mais seca, com 100 kPa. Os valores de K0 obtidos nos ensaios variaram de 0,50 a 0,65. Os resultados seguiram, em parte, a tendência esperada. Entretanto, devido à heterogeneidade do material estudado, é necessário realizar repetições dos ensaios a fim de obter dados representativos. / [en] The necessity to determine the stress-strain behavior of soils under loading for the purpose of engineering projects requires early knowledge of the natural state of stress that acting on the mass of soil. When it comes to tropical residual soils, this challenge is even greater due to weathering and moisture variation. The natural state of the soil is usually expressed by the coefficient of earth pressure at rest (K0). For the determination of K0 three classes of methods can be used: correlation meth-ods, in situ determination and laboratory tests. On this research, one rigid cell with system of suction control is used to obtain values of K0 in laboratory tests. Were analyzed three samples of Gneiss residual soil with different degrees of weathering. The samples were provenient to São Conrado, Rio de Janeiro and namely: young residual soil (SRJ), transition residual soil (SRT) and mature residual soil (SRM). To evaluate the influence of suction on the value of coefficient of earth pressure at rest, each sample of soil were tested into three different levels of moisture which correspond of following conditions: more humid, with suction of 10 kPa; interme-diate, with 40 kPa of suction; and more dry, with 100 kPa. The K0 values obtained in the tests ranged from 0,50 to 0,65. In general, the results attended the expected trend. However, due to the great heterogeneity of the studied material, it is neces-sary to perform repetitions of tests in order to obtain representative data.

[pt] A presente dissertação tem por objeto a análise teórica - comparativa, implementação computacional e avaliação da capacidade de simulação dos modelos constitutivos para solos, com ênfase nos solos não saturados. Primeiramente, são analisados os modelos básicos, clássicos e avançados formulados para solos saturados ou secos, assim como os modelos desenvolvidos para solos não saturados, incluindo também alguns modelos para solos estruturados. A seguir, e para testar a sua capacidade de modelagem, são implementadas as matrizes constitutivas de alguns dos modelos analisados para a retroanálise e previsão de comportamento de três solos residuais na condição saturada e não saturada. Como complemento da pesquisa, um dos modelos é utilizado na simulação de provas de carga com placa em solo residual pelo MEF, para os casos de saturação e umidade natural. Finalmente, são apresentadas organizadamente as conclusões gerais e particulares obtidas, assim como sugestões para futuras pesquisas no tema do presente trabalho. / [en] The objective of this thesis is to present a critical review of the main constitutive soil models published in the literature, discussing and evaluating their abilities to simulate the mechanical behavior of geological materials under the theoretical and numerical points of view, with emphasis on unsaturated soils. Firstly, some basic, classical and advanced soil models are presented and discussed, as well as a few other recent constitutive relationships specially developed for unsaturated and structured materials. In the sequence, in order to assess the efficiency and overall performance of some selected models, the respective constitutive matrices were numerically implemented for prediction and back-analysis of the response of three residual soils, in saturated and unsaturated conditions. As a complement of this research, one of the soil models is used to simulate the results of some plate load tests on a residual soil deposit through the finite element method. Finally, the general conclusions of this work are presented and some suggestions are made to aid future research in the area of soil modeling.

This study tested a dynamic modelling approach based on salt application, meteorological data and generic descriptions of hydrogeological environments for describing the spread of deicing salt to the surroundings and the corresponding increase in chloride storage in soil. Both the amount of chloride storage and the annual variation pattern were significantly altered due to deicing salt application and spread to the roadside environment. Data from field investigations comprising different hydrogeological environments and different methods of measurement were used to examine the variability of the salt deposition pattern in the vicinity of the road, and to test the performance of the model with respect to different soils and vegetation types. The use of typical hydrogeological environments to represent inputs to the model was shown to be useful to demonstrate the importance of soils, vegetation type and groundwater conditions for modelling the impact of deicing salt on soil water and the response to environmental changes in the vadose zone. However, the use of hydrogeological environment could also be misleading in view of the high degree of variability at the field scale. The different methods of measurements and simulations represented different spatial and temporal scales that were shown to be complementary useful to quantify the different pathways of deicing salt in the roadside environment. Continuous simulations complemented with selected field monitoring should therefore be promoted. / QC 20100526

Several hydraulic loadings impose earthen levees to time-dependent variably saturated seepage conditions. The main objective of this study is to improve the analysis of levees under transient seepage with the use of unsaturated soil mechanics. An extensive set of laboratory testing, field monitoring and numerical modeling are performed to analyze a silty sand setback levee located near Seattle, WA. In-situ data obtained from field monitoring are used to monitor suction and effective stress within the levee’s embankment and foundation over the past two years. Soil samples taken from the site are used to perform index, water retention, and unsaturated multi-stage triaxial tests in the laboratory. A finite element model of transient seepage under saturated-unsaturated conditions is then developed and calibrated to reasonably match the field data. The results highlight the need to consider unsaturated soil mechanics along with climatic variables and soil-atmosphere interaction when analyzing levees under transient seepage conditions.

[pt] Este trabalho estuda um modelo acoplado de deformação e fluxo, baseado na teoria de Biot, para solos saturados e não saturados, porém levando-se em consideração a não linearidade física do solo. Na condição não saturada o fluxo do fluido pode ser considerado bifásico ou monofásico. No caso de fluxo bifásico o modelo calcula simultaneamente a pressão de água e do ar, possibilitando assim, a avaliação dos valores da sucção. O grau de saturação é então determinado a partir dos valores obtidos de sucção. Uma vez conhecido o grau de saturação é possível avaliar a permeabilidade relativa do solo para o fluxo de água e ar, tornando o problema não linear em relação à permeabilidade. No problema de fluxo monofásico a pressão de ar é mantida constante, assumindo-se o valor da pressão atmosférica. O comportamento não linear da relação tensão-deformação em solos não saturados é avaliado pelo modelo básico de Barcelona (BBM), baseado na teoria da plasticidade e conceitos de comportamento de solos no estado crítico. Este modelo pode ser compreendido como uma generalização do modelo Cam Clay freqüentemente usado para representação do comportamento mecânico de solos saturados. A influência da sucção na resistência ao cisalhamento de solos não saturados é verificada por meio de algumas aplicações numéricas. Neste trabalho, a não linearidade física de solos secos e saturados também é simulada através do modelo elastoplástico de Lade & Kim. / [en] This dissertation studies a coupled flow and deformation model for saturated and non-saturated soils, based on Biot`s theory but taking into account the nonlinear material behavior of soil. In the non-saturated condition the fluid flow can be considered either as a two-phase or an one-phase problem. In the two-phase case, the model calculates the pressure of air and water simultaneously, thus permitting the evaluation of the pore suction values and the degree of saturation. Once known, the degree of saturation is used for the determination of the relative permeability of the soil considering the flow of water and air, solving a problem that is nonlinear with respect to permeability. In the one-phase flow, the pressure of air is kept constant and equal to the atmospheric pressure. The stress-strain nonlinear behavior of non-saturated soils is modeled through the basic Barcelona model (BBM model) based on plasticity theory and concepts from the critical state behavior of soils. This model may be understood as an extension of the modified Cam Clay model frequently used for representation of the mechanical behavior of saturated soils. The suction influence on the shear resistance of non-saturated soils is examined through some numerical applications. In this work, the material nonlinearity of dry or saturated soils is also simulated through the Lade & Kim elastoplastic model.

One of the most important steps in the numerical simulation of a hydrogeological system is the precise definition of initial and boundary conditions. The better these are characterized, the more efficient the calculation and the more accurate are the simulation result. In case of simulating processes in the unsaturated soil zone, the water retention curve, the relationship between volumetric water content and matric potential, is of great importance. However, the retention parameters determined locally by different standard methods often do not represent the whole soil system under consideration due to heterogeneities in the soil body caused by variability or different compaction of the soil. Resulting over- or underestimation of the parameters is leading to a worse performance of simulations of the water balance including to a higher calibration effort. Therefore, it is more favorable to identify these soil parameters by a method representing the whole soil system to avoid uncertainties. For this reason, a dike experiment was performed to investigate how soil parameters determined locally and globally can represent the properties of the whole soil system. When comparing the simulation results of the numerical models, a better agreement of measured and simulated water contents as well as a lower effort for calibration is observed by using the soil parameters determined globally.

[pt] Para o conhecimento do estado de tensão natural do solo, expresso pelo coeficiente de empuxo no repouso (k0), emprega-se a realização de ensaios de laboratório em amostras indeformadas. A importância da determinação desse parâmetro para solos coluvionares diz respeito à aplicação em obras de engenharia que envolvam empuxo de terra e fundações. Existem correlações na literatura para a determinação de k0 que provém de pesquisas desenvolvidas considerando solos sedimentares, tipicamente de regiões temperadas ou frias. Ao lidar com solos coluvionares e residuais, típicos de regiões tropicais, não é condizente utilizar dessas correlações, visto que o k0 é afetado por vários fatores, tais como a sucção, grau de intemperismo, processo de laterização, entre outros que também afetam solos sedimentares saturados. Esse estudo analisa as amostras de solo coluvionar provenientes do talude da Pontifícia Universidade Católica do Rio de Janeiro (PUC-Rio). Para determinar o coeficiente de empuxo no repouso do solo coluvionar na condição não saturada, é utilizado um equipamento de parede flexível, com sistema de controle de sucção. Para avaliar a influência da sucção no valor do coeficiente de empuxo no repouso, cada amostra foi ensaiada sob diferentes teores de umidade: mais úmida, com sucção de 10 kPa; intermediária com sucção de 40 kPa; e mais seca, com 100 kPa. Os valores de k0 obtidos nos ensaios variaram de 0,19 a 0,24. O coeficiente de empuxo no repouso do solo coluvionar na condição saturada (sucção nula) também foi determinado usando um equipamento triaxial servo-controlado tipo Bishop-Wesley. O valor de k0 obtido para o solo coluvionar na condição saturada foi de 0,36. Os resultados obtidos seguiram, em parte, a tendência esperada. / [en] For the knowledge of the natural stress state of the soil, expressed by the coefficient of earth pressure at rest (k0), a laboratory test is used in undisturbed samples. The importance of the determination of this parameter for colluvium soils relates to the application in engineering works that involve buoyancy of earth and foundations. There are correlations in the literature for determination k0 that comes from researches developed considering sedimentary soils, typically from temperate or cold regions. In dealing with residual and colluvium soils, typical of tropical regions, it is not appropriate to use these correlations, since k0 is affected by several factors, such as degree of weathering, laterization processes and suction, among others that also affect saturated sedimentary soils. This research analyzes samples of a colluvium soil from the Pontifícia Universidade Católica do Rio de Janeiro (PUC-Rio). For the determination of the coefficient of earth pressure at rest of the unsaturated colluvium soil, a flexible wall equipment, with a system of suction control was used. In order to evaluate the influence of suction on the value of the coeficiente of erath pressure at rest, each sample was tested under different moisture contents: wetter, with suction of 10 kPa; intermediate with suction of 40 kPa; and drier with 100 kPa. The k0 values obtained in the tests ranged from 0,19 to 0,24. The coefficient of earth pressure at saturated colluvium (null suction) was also determined using Bishop and Wesley type, servo-controlled triaxial equipment. The value of k0 obtained for the colluvium soil in the saturated condition was equal to 0.36. The results in both trials followed, in part, the expected trend.

[pt] Muitos dos movimentos de massa gravitacionais estão fortemente associados a uma perda de sucção na resistência ao cisalhamento do solo, decorrente de variações de umidade de origem natural (variações atmosféricas) ou de origem artificial (atividade humana). Dentro deste contexto, no presente trabalho analisou-se a influência da sucção na resistência ao cisalhamento, determinaram-se as envoltórias de resistência ao cisalhamento na condição não saturada e avaliou-se a influência de ciclos de umedecimento e secagem na resistência dos solos. Dois tipos de solos da região de Tinguá/RJ foram selecionados para estudo: um solo Residual Jovem e um Colúvio. Estes solos foram submetidos a um programa experimental que consistiu da caracterização granulométrica e mineralógica do material, seguido da obtenção da curva de retenção de água pelo método do papel filtro. Os ensaios de resistência ao cisalhamento envolveram ensaios de cisalhamento direto na condição saturada ou inundada e, para a condição não saturada, utilizou-se o equipamento de cisalhamento direto com sucção controlada (CDSC) da PUC-Rio, projetado por de Campos (1988) e implementado por Fonseca (1991) e Delgado (1993). O equipamento está baseado na técnica de translação de eixos. Para a determinação da velocidade de cisalhamento na condição não saturada utilizou-se a metodologia proposta por Ho e Fredlund (1982). A partir dos resultados obtidos foi possível obter os parâmetros da envoltória do solo não saturado, parâmetros efetivos de resistência do solo saturado, parâmetro que quantifica o aumento na resistência devido a um aumento na sucção matricial; , e c; e plotar a superfície 3D da envoltória que relaciona a resistência do solo não saturado com a tensão normal líquida e a sucção. / [en] Many of gravitational mass movements are strongly associated with a loss of suction on shear strength of the soil due to moisture variations of natural origin (atmospheric variations) or artificial sources (human activity). Within this context, the present study examined the influence of suction on shear strength, determined whether the envelopes of shear strength in unsaturated condition and evaluated the influence of wetting and drying cycles in soil strength. Two types of soils in the region of Tinguá / RJ were selected for study: A Young Residual soil and colluvium. These soils were subjected to an experimental program that consisted of particle size and mineralogical characterization of the material, followed by obtaining the curve of water retention by the filter paper method. The shear strength tests involved direct shear tests on saturated or flooded condition and to unsaturated condition, we used the direct shear device with controlled suction (CDSC) of PUC-Rio, designed by Field (1988) and implemented by Fonseca (1991) and Delgado (1993). The device is based on the technique of translation of the axes. For the determination of the shear rate in the unsaturated condition used the methodology proposed by Ho and Fredlund (1982). From the results it was possible to obtain the parameters of the envelope of unsaturated soil, effective resistance parameters of saturated soil, b, and c; and plot the 3D surface of the envelope which relates the strength of unsaturated soil with net normal stress and suction.

[pt] O presente trabalho de pesquisa teve dois objetivos principais. O primeiro foi obter um índice geotécnico que permita avaliar o grau de intemperismo dos solos residuais. Para isto foram analisados dados correspondentes a perfis de alteração de Gnaisse característicos do município do Rio de Janeiro e executou-se a caracterização completa de um perfil específico de alteração de Gnaisse Facoidal. O segundo objetivo foi a determinação experimental das propriedades hidráulicas de solos tropicais não saturados. Como parte deste objetivo foram desenvolvidos novos instrumentos e modificados equipamentos existentes. A determinação da curva característica de sucção foi efetuada em seis solos residuais e em dois solos sedimentares. Na obtenção das curvas de retenção foram utilizadas as técnicas de translação de eixos, papel filtro e umidade relativa. Esta curva foi também definida a partir de resultados de ensaios de porosimetria de mercúrio. Avaliou-se a aplicabilidade dos modelos disponíveis na literatura no ajuste dos dados experimentais. A função de permeabilidade foi determinada em dois solos, um residual e um sedimentar, utilizando um permeâmetro de parede flexível com sucção controlada. Análises numéricas, efetuadas para avaliar o comportamento do equipamento utilizado, possibilitaram a definição de uma metodologia otimizada de interpretação dos resultados. A aplicabilidade dos modelos disponíveis na literatura na obtenção da função de permeabilidade foi analisada, tendo-se verificado a necessidade de introdução de mudanças nestas formulações para se conseguir reproduzir as medidas diretas efetuadas. / [en] The present research work had two main objectives. The first was to obtain a geotechnical index for measuring the weathering degree of gneissic residual soils. For that it were analyzed data corresponding to typical weathering profiles found in the municipality of Rio de Janeiro and executed a complete characterization of a specific weathering profile of facoidal gneiss. The second objective was to determine, experimentally, the hydraulic properties of tropical unsaturated soils. As part of that new instruments were developed and existing equipments were modified. Moisture retention curves were determined for six residual soils and two sedimentary ones. The axis- translation technique, the filter paper method and the relative humidity one were employed to obtain such curves. Data from mercury intrusion porosimetry tests were also used to define these curves. The applicability of models available in literature to adjust the obtained experimental data was evaluated. The hydraulic conductivity function was determined for two soils, one residual and one sedimentary. A suction controlled, flexible wall permeameter set up was used for that. Numerical modeling was implemented to evaluate the performance of the experimental set up. This provided means for an optimized interpretation of the testing results. The applicability of models available in the literature to obtain the permeability functions was evaluated, being found the need to introduce changes in their formulation in order to be able to reproduce the obtained experimental data.