An experimental investigation on the micromechanics of non-active clays in saturated and partially saturated states

This thesis presents an experimental investigation of the micromechanical behaviour of non-active clay, both in saturated and unsaturated states.For the case of saturated clays, 1-D compression tests on kaolin saturated with fluids having different dielectric permittivity (air, acetone, and water) and pH (acid and basic aqueous solutions) were performed. Based on the Pore-Size Distribution (PSD) of samples taken on both the virgin and the unloading curves, the electrochemical and mechanical forces controlling inter-particle interaction and the particle configuration along reversible and non-reversible compression paths were inferred. Ultimately, a conceptual micromechanical model was formulated and designed in such a way that it could be potentially implemented in a Discrete Element Model (DEM) for clay geomaterials. To show the capability of the conceptual micromechanical model, an 'embryonic' 1-D DEM model was developed. It was shown to be capable of capturing, at a qualitative level, the 1-D response of clay saturated with fluids of different dielectric permittivity.Moving to the unsaturated state, an observation was initially made that the mono-modal PSDs of kaolin in reconstituted and dry states fit quite well the pore-size distribution of the micro-pores and the macro-pores respectively in compacted samples. Based on this observation, an assumption was put forward that macro-pores are filled with air ('air-saturated') and micro-pores are filled with water ('water-saturated'). The comparison of the PSD of samples compacted with water and acetone and the comparison of the PSDs of samples freeze-dried and oven-dried appeared to corroborate such an assumption.A particle-based microstructural conceptual model for unsaturated soils was then formulated. Such a model can explain the evolution of PSD observed at various compaction water contents. Besides, this microstructural conceptual model was proven to be as effective as the traditional aggregate-based microstructural model in interpreting some of the classical responses of unsaturated soils including volumetric collapse upon wetting. A particle-based microstructural model provides an alternative approach to the aggregate-based microstructural model that has (potentially) the advantage of facilitating the formulation of DEM models for unsaturated clays. Another advantage is that the conceptual microstructural model and, hence, any DEM models would remain essentially the same for saturated and unsaturated states.Following the idea that macro-pores are filled with air and micro-pores are filled with water, it was assumed that the response of unsaturated soils can be modelled by considering separately the response of the dry part (air-saturated) and wet part (water-saturated). To model 1-D compression, the void ratio of the dry part was read from the 1-D compression of dry kaolin and the void ratio of the wet part was read from the 1-D compression of the saturated reconstituted kaolin. This approach allowed modelling the water-undrained compression tests in Tarantino & De Col (2008) with excellent accuracy.This thesis presents an experimental investigation of the micromechanical behaviour of non-active clay, both in saturated and unsaturated states.For the case of saturated clays, 1-D compression tests on kaolin saturated with fluids having different dielectric permittivity (air, acetone, and water) and pH (acid and basic aqueous solutions) were performed. Based on the Pore-Size Distribution (PSD) of samples taken on both the virgin and the unloading curves, the electrochemical and mechanical forces controlling inter-particle interaction and the particle configuration along reversible and non-reversible compression paths were inferred. Ultimately, a conceptual micromechanical model was formulated and designed in such a way that it could be potentially implemented in a Discrete Element Model (DEM) for clay geomaterials. To show the capability of the conceptual micromechanical model, an 'embryonic' 1-D DEM model was developed. It was shown to be capable of capturing, at a qualitative level, the 1-D response of clay saturated with fluids of different dielectric permittivity.Moving to the unsaturated state, an observation was initially made that the mono-modal PSDs of kaolin in reconstituted and dry states fit quite well the pore-size distribution of the micro-pores and the macro-pores respectively in compacted samples. Based on this observation, an assumption was put forward that macro-pores are filled with air ('air-saturated') and micro-pores are filled with water ('water-saturated'). The comparison of the PSD of samples compacted with water and acetone and the comparison of the PSDs of samples freeze-dried and oven-dried appeared to corroborate such an assumption.A particle-based microstructural conceptual model for unsaturated soils was then formulated. Such a model can explain the evolution of PSD observed at various compaction water contents. Besides, this microstructural conceptual model was proven to be as effective as the traditional aggregate-based microstructural model in interpreting some of the classical responses of unsaturated soils including volumetric collapse upon wetting. A particle-based microstructural model provides an alternative approach to the aggregate-based microstructural model that has (potentially) the advantage of facilitating the formulation of DEM models for unsaturated clays. Another advantage is that the conceptual microstructural model and, hence, any DEM models would remain essentially the same for saturated and unsaturated states.Following the idea that macro-pores are filled with air and micro-pores are filled with water, it was assumed that the response of unsaturated soils can be modelled by considering separately the response of the dry part (air-saturated) and wet part (water-saturated). To model 1-D compression, the void ratio of the dry part was read from the 1-D compression of dry kaolin and the void ratio of the wet part was read from the 1-D compression of the saturated reconstituted kaolin. This approach allowed modelling the water-undrained compression tests in Tarantino & De Col (2008) with excellent accuracy