<p>Many geotechnical engineering problems involve combining two or more independent physical processes as a coupled solution of seepage, volume change and shear strength. For any physical process being modeled, it is desirous to be able to compute any of the volume-mass soil properties. When the volume-mass soil properties are combined with the stress state of the soil, the result is a volume-mass constitutive relationship. Three volume-mass constitutive relationships (i.e., void ratio, water content and degree of saturation) are generally viewed as being the most fundamental; however, only two of the relations are independent. The unsaturated soil properties associated with seepage, volume change and shear strength problems are also related to the volume-mass soil properties. While the unsaturated soil properties are often estimated as simply being a function of the soil-water characteristic curve, it is more accurate to define the properties in a more rigorous manner in terms of the volume-mass soil properties. The advancement of computing capabilities means that it is quite easy to formulate constitutive relations for shear strength and permeability, for example, in terms of all volume-mass properties of the unsaturated soil.</p><p>The objectives of this dissertation include: i) the development of volume-mass constitutive models for unsaturated soils; ii) the further study and verification of the volume-mass constitutive behavior of unsaturated soils; and iii) the development of techniques for visualization of volume-mass constitutive surfaces for unsaturated soils. To achieve these objectives, the present research study was conducted from both theoretical and experimental bases.</p><p>The theoretical program commenced with a comprehensive literature review of the volume-mass constitutive relationships for unsaturated soils. A new, more rigorous volume-mass constitutive model was then proposed. Appropriate terminology was introduced for the development of the model, followed by an outline of the assumptions used and the mathematical derivation. The proposed model requires conventionally obtainable soil properties for its calibration. The model is capable of predicting both the void ratio and water content constitutive relationships for various unsaturated soils, taking into account elastic and plastic volume changes. Various stress paths can be simulated and hysteresis associated with the soil-water characteristic curve can be taken into account. </p><p>Two closed-form equations for the volume-mass constitutive relationships were derived. A computer software program was written based on the theory of the proposed volume-mass constitutive model. Techniques for the visualization of the volume-mass constitutive surfaces were then presented. An experimental program was conducted in the laboratory. The experimental program involved the verification of a new testing apparatus. Several soils were selected for testing purposes and appropriate testing procedures were established (i.e., soil specimens were initially slurry). The testing stress paths followed in the experimental program were different from most research programs conducted in the past and reported in the research literature. Conclusions regarding the compressibility, stress path dependency, and hysteretic nature of the soil-water characteristic curve of an unsaturated soil were presented.</p><p>A considerable number of test results (i.e., from both the experimental program and the research literature) were used in the verification of the new volume-mass constitutive model. This model has proven to be effective in predicting both collapse and expansion of a soil. The volume-mass constitutive model appears to predict behaviour in a satisfactory manner for a wide range of soils; however, the predictions appear to be superior for certain soils. In all cases the volume-mass predictions of the model appear to be satisfactory for geotechnical engineering practice.

<p>Many geotechnical engineering problems involve combining two or more independent physical processes as a coupled solution of seepage, volume change and shear strength. For any physical process being modeled, it is desirous to be able to compute any of the volume-mass soil properties. When the volume-mass soil properties are combined with the stress state of the soil, the result is a volume-mass constitutive relationship. Three volume-mass constitutive relationships (i.e., void ratio, water content and degree of saturation) are generally viewed as being the most fundamental; however, only two of the relations are independent. The unsaturated soil properties associated with seepage, volume change and shear strength problems are also related to the volume-mass soil properties. While the unsaturated soil properties are often estimated as simply being a function of the soil-water characteristic curve, it is more accurate to define the properties in a more rigorous manner in terms of the volume-mass soil properties. The advancement of computing capabilities means that it is quite easy to formulate constitutive relations for shear strength and permeability, for example, in terms of all volume-mass properties of the unsaturated soil.</p><p>The objectives of this dissertation include: i) the development of volume-mass constitutive models for unsaturated soils; ii) the further study and verification of the volume-mass constitutive behavior of unsaturated soils; and iii) the development of techniques for visualization of volume-mass constitutive surfaces for unsaturated soils. To achieve these objectives, the present research study was conducted from both theoretical and experimental bases.</p><p>The theoretical program commenced with a comprehensive literature review of the volume-mass constitutive relationships for unsaturated soils. A new, more rigorous volume-mass constitutive model was then proposed. Appropriate terminology was introduced for the development of the model, followed by an outline of the assumptions used and the mathematical derivation. The proposed model requires conventionally obtainable soil properties for its calibration. The model is capable of predicting both the void ratio and water content constitutive relationships for various unsaturated soils, taking into account elastic and plastic volume changes. Various stress paths can be simulated and hysteresis associated with the soil-water characteristic curve can be taken into account. </p><p>Two closed-form equations for the volume-mass constitutive relationships were derived. A computer software program was written based on the theory of the proposed volume-mass constitutive model. Techniques for the visualization of the volume-mass constitutive surfaces were then presented. An experimental program was conducted in the laboratory. The experimental program involved the verification of a new testing apparatus. Several soils were selected for testing purposes and appropriate testing procedures were established (i.e., soil specimens were initially slurry). The testing stress paths followed in the experimental program were different from most research programs conducted in the past and reported in the research literature. Conclusions regarding the compressibility, stress path dependency, and hysteretic nature of the soil-water characteristic curve of an unsaturated soil were presented.</p><p>A considerable number of test results (i.e., from both the experimental program and the research literature) were used in the verification of the new volume-mass constitutive model. This model has proven to be effective in predicting both collapse and expansion of a soil. The volume-mass constitutive model appears to predict behaviour in a satisfactory manner for a wide range of soils; however, the predictions appear to be superior for certain soils. In all cases the volume-mass predictions of the model appear to be satisfactory for geotechnical engineering practice.

A condutividade hidráulica de dois solos não saturados típicos da região de São Carlos, um arenoso e outro argiloso é estudada através de ensaios de campo empregando o permeâmetro Guelph. Duas alternativas de ensaio foram empregadas e os resultados experimentais foram analisados com os modelo teóricos de REYNOLDS e ELRICK (1985) e PHILIP (1985). Paralelamente, obtiveram-se em laboratório curvas de retenção de água dos solos o que possibilitou determinar a função condutividade hidráulica através das fórmulas empíricas de VAN GENUCHTEN, (1980) e de GARDNER (1958). As características e a variação dos valores da condutividade hidráulica não saturada de campo fornecidos pelo permeâmetro Guelph, são discutidos e confrontados com aqueles obtidos a partir dos métodos indiretos provenientes de laboratório. Observou-se que as condutividades hidráulicas saturadas nos solos foram praticamente iguais, quer se considerem os resultados de campo ou laboratório. Os resultados de campo mostraram-se mais consistentes para o solo argiloso qualquer que tenha sido a técnica de ensaio. O parâmetro &#945 da equação de GARDNER (1958) obtido no campo resultou maior que o valor sugerido na literatura ou obtido em laboratório para o solo arenoso. Para o solo argiloso, todas as opções de obtenção do parâmetro tenderam a fornecer valores praticamente iguais. / The unsaturated hydraulic conductivity of two typical soils from the region of São Carlos City, sandy and clayey soils, has been studied through field tests using the Guelph permeameter. Two different alternatives of tests have been performed and the experimental data have been analyzed following the theoretical models from REYNOLDS and ELRICK (1985) and PHILIP (1985). However, soil-water characteristic curves were obtained in such a way that the hydraulic conductivity function could be evaluated using the empirical expressions from VAN GENUCHTEN (1980), and GARDNER (1958). The characteristics and the variation of the field unsaturated hydraulic conductivity values from the Guelph permeameter are discussed and compared to those from laboratory indirect methods. The field results showed to be more consistent to the clayey soil independent on the test or calculation method. The &#945 value (GARDNER, 1958) parameter obtained in the fields was Iarger than the values suggested in the literature or obtained in laboratory for the sandy soil. To the clayey soil, all the obtaining options of the parameter tended to supply values quite the same.

A condutividade hidráulica de dois solos não saturados típicos da região de São Carlos, um arenoso e outro argiloso é estudada através de ensaios de campo empregando o permeâmetro Guelph. Duas alternativas de ensaio foram empregadas e os resultados experimentais foram analisados com os modelo teóricos de REYNOLDS e ELRICK (1985) e PHILIP (1985). Paralelamente, obtiveram-se em laboratório curvas de retenção de água dos solos o que possibilitou determinar a função condutividade hidráulica através das fórmulas empíricas de VAN GENUCHTEN, (1980) e de GARDNER (1958). As características e a variação dos valores da condutividade hidráulica não saturada de campo fornecidos pelo permeâmetro Guelph, são discutidos e confrontados com aqueles obtidos a partir dos métodos indiretos provenientes de laboratório. Observou-se que as condutividades hidráulicas saturadas nos solos foram praticamente iguais, quer se considerem os resultados de campo ou laboratório. Os resultados de campo mostraram-se mais consistentes para o solo argiloso qualquer que tenha sido a técnica de ensaio. O parâmetro &#945 da equação de GARDNER (1958) obtido no campo resultou maior que o valor sugerido na literatura ou obtido em laboratório para o solo arenoso. Para o solo argiloso, todas as opções de obtenção do parâmetro tenderam a fornecer valores praticamente iguais. / The unsaturated hydraulic conductivity of two typical soils from the region of São Carlos City, sandy and clayey soils, has been studied through field tests using the Guelph permeameter. Two different alternatives of tests have been performed and the experimental data have been analyzed following the theoretical models from REYNOLDS and ELRICK (1985) and PHILIP (1985). However, soil-water characteristic curves were obtained in such a way that the hydraulic conductivity function could be evaluated using the empirical expressions from VAN GENUCHTEN (1980), and GARDNER (1958). The characteristics and the variation of the field unsaturated hydraulic conductivity values from the Guelph permeameter are discussed and compared to those from laboratory indirect methods. The field results showed to be more consistent to the clayey soil independent on the test or calculation method. The &#945 value (GARDNER, 1958) parameter obtained in the fields was Iarger than the values suggested in the literature or obtained in laboratory for the sandy soil. To the clayey soil, all the obtaining options of the parameter tended to supply values quite the same.

L'étude présentée dans cette thèse porte sur une analyse expérimentale des propriétés physiques (densités, porosité, granulométrie) et hydrodynamiques (isotherme de désorption, courbe caractéristique sol-eau, perméabilité à saturation) d'un sol type aride du Burkina Faso et une simulation numérique des transferts d'eau. La modélisation des transferts est abordée par une approche thermodynamique qui constitue un cadre général pour décrire à la fois l'état de l'eau dans le sol et les mécanismes de transferts mis en jeu : filtration de la phase liquide, diffusion de la vapeur d'eau, changement de phase liquide-vapeur. Des essais expérimentaux de transferts d'eau dans des colonnes de sol hermétiquement fermées et placées dans une enceinte régulée en température, ont permis d'établir l'évolution des profils de teneur en eau au cours du temps. La configuration expérimentale est choisie de sorte que seule la filtration de la phase liquide de l'eau dans le sol soit prépondérante. Les profils expérimentaux ont servi à l'estimation par approche inverse de la conductivité hydraulique dans les faibles teneurs en eau. La prise en compte des écoulements par films dans la modélisation du coefficient de perméabilité relative corrige les insuffisances des modèles capillaires. Les coefficients de changement de phase proche et loin de l'équilibre modélisés à partir de résultats expérimentaux permettent la mise en évidence du non-équilibre liquide/gaz dans les transferts aux faibles teneurs en eau. Il ressort que le non-équilibre dans les transferts aux faibles teneurs en eau est plus marqué avec le modèle capillaire avec un front de séchage plus franc à la surface. La prise en compte des écoulements par films prolonge la filtration de la phase liquide de l'eau jusqu'à l'état hygroscopique du sol avec un flux d'eau liquide maintenu à la surface du sol jusqu'à des temps plus longs. / The study presented in this thesis focuses on an experimental analysis of physical properties (densities, porosity, particle size distribution) and hydrodynamic properties (desorption isotherm, soil-water characteristic curve, saturated permeability) of an arid soil from Burkina Faso and numerical simulation of water transfers. The transfer modelling is addressed through thermodynamic approach which provides a general framework to describe both the water state in the soil and the transport mechanisms: filtration of liquid phase, water vapor diffusion, liquid/vapor phase change. Experimental attempts of water transfer in soil columns, sealed and placed into a controlled temperature chamber, have established the evolution of water content profiles over time. The experimental configuration is chosen so that only the filtration of water liquid phase is taken into account. These experimental profiles were used to estimate, by inverse approach, the unsaturated hydraulic conductivity at low water contents. Consideration of film flows in the relative permeability modelling corrects the shortcomings of capillary bundle models used to describe water flow from saturation to oven-dryness. The coefficients of phase change near and far from equilibrium modeled from experimental results prove non-equilibrium liquid/gas existence. It appears that the liquid/gas non-equilibrium at low water content is more pronounced with the capillary model with a frank drying front at the soil surface. Consideration of film flows extends liquid phase filtration into the soil until the hygroscopic state with a liquid water flux maintained at the soil surface for longer times.