Le séchage d’organogels, gels composés d’un organogélateur de type acide aminé dissous dans un solvant, conduit à la formation d’aérogels, solides très légers et très poreux. Ces aérogels constituent entre autres de très bons isolants thermiques. La préparation des aérogels nécessitent plusieurs étapes. Ce procédé supercritique est composé de trois étapes : préparation de CO2 supercritique, séchage d’organogel dans l’autoclave en le balayant par un courant de CO2 et séparation de CO2-solvant au sein d’une cascade de trois séparateurs cyclones. Dans cette étude, les deux dernières étapes ont été abordées afin d’optimiser les paramètres du procédé. Les solvants utilisés sont la tétraline et le toluène. La cinétique de séchage au sein de l’autoclave a été étudiée, trois approches différentes ont été explorées : le transport de tétraline dans le CO2 supercritique au sein du gel est gouverné uniquement par le phénomène de diffusion dans les deux premières, il est régi par les phénomènes de diffusion et de convection dans la troisième approche. Les équations de transfert de matière ont été résolues à l’aide de deux logiciels numériques Matlab et ANSYS-Fluent. La troisième approche semble donner les résultats les plus cohérents avec les résultats expérimentaux. La compréhension des raisons du dysfonctionnement des séparateurs cycloniques en aval de l’autoclave pour le cas du toluène a nécessité deux études : une étude hydrodynamique MFN a permis de simuler le transport des fluides au sein du premier séparateur cyclone, à l’aide du logiciel ANSYS-Fluent. Cette étude a montré que le toluène liquide n’était pas entraîné dans la sortie gaz du cyclone, indiquant ainsi que l’hydrodynamique favorise la séparation. Afin d’expliquer ce dysfonctionnement, une étude thermodynamique portant sur les deux solvants, toluène et tétraline a été réalisée. La séparation cyclonique a été considérée comme un simple étage théorique et simulée avec le logiciel PRO/II. Cette étude a bien expliqué les résultats expérimentaux pour les deux solvants et a permis d’optimiser les paramètres du procédé / Drying of organogels, gels composed of an organogelator such as amin acids dissolved in a solvent, leads to the formation of aerogels, very light solids with high porosity. These aerogels are expected to be very good heat insulating materials. Preparation of aerogels includes several steps. This process consists of three steps: preparation of supercritical CO2, drying of the gel in an autoclave with a CO2 flow and CO2-solvent separation carried out in a cascade of three separators. In this study both the last steps were discussed in order to optimize the process parameters. The used solvents are tetralin and toluene. The drying kinetic in the autoclave was studied, three different approaches have been exploited. Tetralin transport in supercritical CO2 within the gel is governed only by the diffusion phenomenon in the first two approaches, although, it is governed by diffusion and convection phenomena in the third approach. The mass transfer equations were solved by two numerical software Matlab and ANSYS-Fluent. It seemed that the third approach gave the most consistent results with experimental results. The understanding of the malfunction reasons of cyclonic separators downstream from the autoclave for the case of toluene required two studies: A CFD hydrodynamic study was used to simulate the fluids transport within the first cyclone separator, using the ANSYS-Fluent software. This study showed that the liquid toluene was not carried over into the gas outlet of the cyclone, indicating that hydrodynamic promotes the separation. For this reason, a thermodynamic study of both solvents, toluene and tetralin was performed. Cyclonic separation was regarded as a simple theoretical stage and simulated with PRO / II software. This study has explained very well the experimental results for both solvents and allowed to optimize the process parameters

This thesis presents an investigation of reactive transport of species in porous media, with the aim to understand better and predict the fate of radionuclide in engineered and natural barriers of future deep geological disposal facilities for nuclear waste. The work involves developments of several pore-scale models for simulating reactive transport by coupling convective, adsorptive and diffusive processes. Pore network models (PNM) are amongst the appealing approaches that provide a suitable description for dealing with mutable pore space structures. Such models have been used to describe conservative as well as reactive transport in saturated and unsaturated porous media. In the present thesis, pore network models based on a regular tessellation of truncated octahedral cells are proposed and developed to simulate mass transport in porous media with incomplete pore space information due to limitation of existing characterisation techniques. Bentonite and Opalinus Clay are selected to illustrate the methodology. The micro- and meso-structure of these clays and their effects on the transport behaviour are investigated. The research shows that the clays are anisotropic and heterogeneous with fast diffusion parallel to the bedding plane and slow diffusion perpendicular to the bedding plane. In addition, different types of species have different accessible porosity and macroscopic diffusion coefficients. The anisotropy and heterogeneity of clays are achieved by different length scales and percentage of pores in different directions in the pore network models. The transport behaviour of various species, including sorption and anion exclusion, is simulated and analyzed. The effect of sorption is simulated via changing the pore radii by a coarse grained mathematical formula or by a formula directly in each pore. The results are in good agreement with experimentally measured macroscopic (bulk) diffusivities for the materials studied, including anisotropic diffusion coefficients. This lends strong support to the physical realism of the proposed models. The developed methodology can be used for any micro and meso-porous material with known distribution of pore sizes. It can be extended to other pore space changing mechanisms, in addition to sorption, to derive mechanism-based evolution laws for the transport parameters of porous media.

No description available.