Uncertainty quantification involves sampling the reservoir parameters correctly from a posterior probability function that is conditioned to both static and dynamic data. Rigorous sampling methods like Markov Chain Monte Carlo (MCMC) are known to sample from the distribution but can be computationally prohibitive for high resolution reservoir models. Approximate sampling methods are more efficient but less rigorous for nonlinear inverse problems. There is a need for an efficient and rigorous approach to uncertainty quantification for the nonlinear inverse problems. First, we propose a two-stage MCMC approach using sensitivities for quantifying uncertainty in history matching geological models. In the first stage, we compute the acceptance probability for a proposed change in reservoir parameters based on a linearized approximation to flow simulation in a small neighborhood of the previously computed dynamic data. In the second stage, those proposals that passed a selected criterion of the first stage are assessed by running full flow simulations to assure the rigorousness. Second, we propose a two-stage MCMC approach using response surface models for quantifying uncertainty. The formulation allows us to history match three-phase flow simultaneously. The built response exists independently of expensive flow simulation, and provides efficient samples for the reservoir simulation and MCMC in the second stage. Third, we propose a two-stage MCMC approach using upscaling and non-parametric regressions for quantifying uncertainty. A coarse grid model acts as a surrogate for the fine grid model by flow-based upscaling. The response correction of the coarse-scale model is performed by error modeling via the non-parametric regression to approximate the response of the computationally expensive fine-scale model. Our proposed two-stage sampling approaches are computationally efficient and rigorous with a significantly higher acceptance rate compared to traditional MCMC algorithms. Finally, we developed a coarsening algorithm to determine an optimal reservoir simulation grid by grouping fine scale layers in such a way that the heterogeneity measure of a defined static property is minimized within the layers. The optimal number of layers is then selected based on a statistical analysis. The power and utility of our approaches have been demonstrated using both synthetic and field examples.

Technological advances in horizontal drilling and hydraulic fracturing have paved the way for the exploration and production of shale gas and shale oil, the fastest growing energy sector globally. The imaging and quantification of the geometry, sizes, network and distribution of extremely fine-grain minerals, organic matter and pores are a significant component for the macroscopic and microscopic characterisation of shale reservoirs but is also highly challenging. X-ray computed tomography (XCT) combined with 3D Electron Microscopy (EM) are used to address this challenge and give us information in 3D from multiple length scales over 3 orders of magnitudes: mesoscale (R1), microscale (R2), submicron-scale (R3), low-resolution nanoscale (R4) and high-resolution nanoscale (R5) with spatial resolutions of ~10micro metre, ~1micro metre, ~130 nm, ~50nm and ~5nm, respectively. The multi-scale imaging and quantification method was initially applied here to the Carboniferous Bowland Shale, the largest potential shale gas play in the UK. The appropriate length scales (both field of view and voxel size) of specified phases such as pores, organic matter, clay minerals and non-clay minerals were analysed. The low connectivity of pores and high connectivity of organic matter suggests that the 20 nm and larger pores imaged did not form connected flow paths, demonstrating that porous gas flow through this sample cannot be the main transport mechanism and diffusive transport through the organic matter and clay minerals must also be considered. Then, the variation of organic matter and pore distribution along a TOC gradient were analysis on Lublin gas-mature shale samples in Poland and Baltic oil-mature shale samples in Lithuania. The results show intergranular pores dominated in this series of samples, including organic interface pores and inter-mineral pores, which further confirm that organic matter is not the primary influencing factor for porosity, but the clay minerals. Finally, a novel multi-stage workflow of pore system is proposed relying on both image quantification and numerical modelling of geological features with studies in Jurassic Haynesville shale in the US. Three stages are divided according to pore variation, mineral variation and microfacies variation across four distinct length scales (R1-R4/R5), and permeability was simulated based on the upscaled pore system. The final computed porosity and permeability shows acceptable errors when compared with the helium porosity and press decay permeability. Beyond the lab measurements, the pore occurrence and size distribution were computed in the upscaling process. The combining of XCT and 3D-EM provides a powerful tool for the multi-scale imaging and quantification of microstructural information in shales, allowing the visualization of pores, organic matter and inorganic mineral phases over a range of scales over three orders of magnitude (~ 10 micro metre to ~ 5 nm), and the volume fraction of each phases shows a reasonable correlation to traditional physical and chemistry quantification data. The further studies, such as the variation of organic matter and pores, upscaling of porosity and permeability presented in this study, has verified the feasibility of the proposed multi-scale method and promises a bit potential for reservoir prediction and other challenges in geological studies.

Les incendies sont complexes et mettent en jeu une multitude de phénomènes. Afin de les étudier, l’approche multiéchelle permet de séparer les processus.Parmi ceux-ci, la décomposition thermique des solides joue un rôle très important. Terme source, elle traduit la quantité, le débit et la nature des composés volatils émis. Sa description numérique est donc capitale. Les modèles utilisés aujourd’hui sont formés d’une loi de variation de la vitesse de forme « Arrhenius », couplée à une fonction de conversion de la masse pour chaque espèce étudiée et d’un mécanisme réactionnel organisant les réactions entre elles. Toutefois, ce modèle s’appuie sur les théories utilisées dans la phase gazeuse et de sérieux doutes peuvent être émis sur sa représentativité pour une application dans la phase condensée.Les travaux de thèse exposés dans ce rapport se focalisent sur le développement d’un modèle en partant des réactions et des processus les plus fondamentaux dans la phase condensée afin de permettre la simulation de la cinétique de décomposition des matériaux solides. Le second aspect concerne l’étude de ce modèle pour déterminer la méthode de résolution et d’optimisation la plus adéquate, le rôle de chacun des paramètres, les éventuels mécanismes de compensation et l’unicité de la solution.Finalement, l’ensemble de la démarche est appliquée à un matériau complexe, le bois. Les résultats obtenus, à partir d’une nouvelle démarche développée lors de ces travaux de thèse, montrent une amélioration significative du modèle aux aspects physiques et chimiques de la dégradation thermique des matériaux solides. / Fires are complex and a variety of phenomena are involved. In order to study them, the up-scaling approach separates all the processes.Among them, the solid thermal decomposition has an important role to play. Source term, it reflects the amount, rate and nature of volatile compounds emitted and its numerical description is essential. The models used currently are formed by a law of variable speed (the Arrhenius law) coupled with a conversion function of mass for each species and a kinetic mechanism organizing all reactions between them. However, this model is based on the theories used in the gas phase and serious doubts might be raised with regard to the representativeness for application in the condensed phase.The thesis works exposed in this report are focused on the model development departing from the reactions and the processes the more fundamental in the condensed phase in order to permit the simulation of the solid kinetic decomposition. The second aspect concerns the study of this model to determine the resolution and the optimization method the most appropriate, the role of each parameter, the possible compensation mechanisms and the uniqueness of the solution.Finally, the entire process is applied to a complex material, the wood. The results obtained from a new approach developed in this work, show a significant improvement of the model to the physical and chemical aspects of the thermal degradation of solid materials.

Simuler les réservoirs souterrains permet d’optimiser la production d’hydrocarbures. Les réservoirs naturellement ou hydrauliquement fracturés détiennent une part importante des réserves et exhibent un degré élevé d’hétérogénéité : les fractures, difficiles à détecter, impactent fortement la production via des réseaux préférentiels d’écoulement. Une modélisation précise de ces forts contrastes permettrait d’optimiser l’exploitation des ressources tout en maîtrisant mieux les risques environnementaux. L’enjeu est de prédire les processus d’écoulement multi échelles par un modèle simplement paramétrable. Une stratégie de simulations, qui améliore la fiabilité et les temps de calculs est mise au point dans cette thèse. Elle permet de simuler numériquement ou analytiquement la complexité d’un réservoir fracturé à grande échelle. Ces techniques dont l’intérêt est démontré sur un réservoir de roche mère trouvent des applications en géothermie ou dans la gestion des ressources en eau. / Fluid flow simulation is used to optimize oil and gas production. Naturally or hydraulically fractured reservoirs hold a significant part of reserves, difficult to assess. Fractures may create preferential flow paths heavily impacting fluid flow. Accurate modeling of fractured media accounting for strong contrasts would allow operators to optimize resources exploitation while better controlling environmental risks. Integrating sparse available data, we aim at predicting fluid flow processes occurring in the earth’s subsurface accounting for multi-scale fractures with a simply parameterized model. Improving the computational time and results reliability, we propose a full integrated strategy suitable for fractured reservoir specificities by simulating the fractures complexity on large scales. The techniques developed in this thesis, whose interest is demonstrated in an unconventional field case study, can find other applications in geothermal engineering and water resources management

La plupart des méthodes pour la modélisation des réservoirs fracturés reposent sur le modèle de Warren et Root (1963). Mais ce modèle reste limité par : l'hypothèse d'un volume élémentaire représentatif, l'évaluation des transferts matrice-fractures, l’idéalisation du système fracturé, l'emploi de conditions aux limites locales. La méthode développée répond à ces quatre points. Un réseau de fractures et une grille de matrice sont discrétisés conjointement, puis un changement d'échelle est réalisé. Un ensemble de nœuds représentatifs est sélectionné, et un système simplifié équivalent est construit par décimation des autres nœuds en assurant la conservation des pressions et des débits sans imposer de conditions aux limites. Enfin le nombre de connexions est réduit et les transmissibilités restantes sont calculées par une procédure d'optimisation. Ces systèmes simplifiés ont été résolus plus rapidement lors de simulations d’écoulement tout en reproduisant le comportement du modèle fin / Most methods for modeling fractured reservoirs rely on the model of Warren and Root (1963). But this model is limited by: the assumption of a representative elementary volume, the evaluation of matrix-fracture transfers, the idealization of the fractured system, the use of local boundary conditions. The developed method overcomes these four points. A fracture network and a matrix grid are jointly discretized, then an upscaling is carried out. A set of representative nodes is selected, and an equivalent simplified system is built by decimating the other nodes while ensuring the preservation of pressure and flow rate and without imposing any boundary conditions. Finally the number of connexions is reduced and the remaining transmissibilities are evaluated by an optimization procedure. These simplified systems have been solved more quickly by the flow simulator while reproducing the fine model behavior

Modéliser les phénomènes de transport de solutés organiques en milieux poreux colonisés par des populations bactériennes se développant sous forme de biofilms est un domaine de recherche important pour un certain nombre d’applications environnementales, comme par exemple pour les méthodes de bioremédiation des sols et des eaux contaminés par des polluants organiques (biosparging, bio-barrières …). Les biofilms, qui sont composés principalement de bactéries et de substances polymériques extracellulaires, peuvent se développer sur les parois de grains d’un milieu poreux. Le métabolisme bactérien dégrade les solutés organiques et contribue ainsi à la diminution de la contamination. Le transport bio-réactif de composés organiques dans un milieu poreux incluant un biofilm est un problème fortement multi-échelle (depuis l’échelle de la bactérie jusqu’à l’échelle de l’aquifère) et fortement couplé (avec des phénomènes hydrodynamiques, physico-chimiques et biochimiques). Le soluté organique est transporté par convection et diffusion dans la phase fluide et diffuse dans la phase biofilm, où il est dégradé par le métabolisme bactérien. Le but de ce travail est de développer des modèles de transport bio-réactif définis à l’échelle de Darcy à partir des données disponibles à l’échelle du pore, en adoptant la méthode de changement d’échelle dite de prise de moyenne volumique. Dans le cas général, une telle approche conduit à un modèle macroscopique de transport à deux équations couplées (une équation par phase de transport). En considérant les relations entre les concentrations moyennées dans chaque phase, plusieurs régimes de transport permettant de dégénérer ce modèle en modèle à une seule équation peuvent être identifiés. L’hypothèse d’équilibre de masse local conduit à un tel modèle simplifié. En condition de non-équilibre, deux cas limites permettent également de développer des modèles de transport à une équation : le cas où le taux de biodégradation est contrôlé par le transfert de masse externe et le cas ou il est contrôlé par la cinétique de réaction. L’utilisation de ces quatre modèles implique la résolution numérique de problèmes de fermeture, afin d’évaluer les paramètres macroscopiques de transports (tenseur de dispersion, taux de dégradation …). Des calculs de coefficients effectifs ont été effectués dans différentes conditions de transport afin d’étudier leur comportement. Les résultats de ces modèles ont été comparés avec ceux obtenus par simulations directe à l’échelle microscopique pour une géométrie de pore bidimensionnelle stratifiée. À partir de ces comparaisons, les domaines de validité de chaque modèle ont été identifiés en termes de conditions hydrodynamique et biochimique de transport. (i.e. le nombre de Péclet et le nombre de Damköhler). Le développement d’un modèle expérimental de transport en milieux poreux incluant un biofilm a également été entamé, afin d’une part d’effectuer une validation expérimentale des modèles numériques préalablement développés et d’autre part de fournir un outil supplémentaire pour l’étude des phénomènes considérés / Modeling transport in porous media of organic chemical solute in presence of a bacterial population growing as biofilms is an important area of research for environmental applications, for example for remediation of groundwater contaminated by organic pollutants (biosparging, bio-barriers …). Biofilms, which are composed of bacteria and extracellular organic substances, grow on the pore walls of the porous medium. Bacteria degrade the organic solute by their metabolism and thus may contribute to pollution decrease. Bio-reactive transport of an organic solute in a porous medium including a biofilm phase is a strongly multi-scale (from the bacteria scale to the heterogeneity scale of the aquifer) and coupled (involving hydrodynamic, physicochemical and biochemical phenomena) process. The organic solute is transported by convection and diffusion in the fluid phase and diffuses into the biofilm phase, where it is degraded by bacterial metabolism. The goal of this work is to develop macroscopic models of bio-reactive transport at the Darcy-scale through volume averaging based on the data available at pore-scale. In the general case, the macroscopic system obtained by averaging pore-scale equations is a two coupled equations system (one equation for each phase), called two-equation model. By considering the relation between averaged concentration in the fluid phase and averaged concentration in the biofilm phase, several regimes of transport can be found which allow simplifying this system into a one equation system. The local mass equilibrium assumption leads to such a simplified model. When an equilibrium relationship between phases cannot be considered, a one equation model may though be developed if the biodegration rate is limited by external mass transfer or by the kinetics of bacterial metabolism. The use of these models implies the numerical solving of closure problems, in order to set up the values of the macroscopic transport parameters (dispersion tensor, interfacial flux …). Computations of these effective coefficients have been performed in different situations of mass transport in porous medium in order to study their behaviour. The results of these models have then been compared with direct simulations performed on a simplified geometry representative of a two-dimensional porous medium including a biofilm phase. Based on these comparisons, the validity domain of this model has been identified in terms of hydrodynamic and biochemical conditions of transport (i.e. the Péclet number and the Damköhler number). The set up of an experimental model of transport in a porous media including a biofilm phase has also been started, in order to make experimental validations of the previously developed numerical models and to build up an additional tool to study the considered phenomena

Recent studies have shown that permafrost soils in the northern circumpolar region store almost twice as much carbon as the atmosphere. Since soil organic carbon (SOC) pools have large regional and landscape-level variability, detailed SOC inventories from across the northern permafrost region are needed to assess potential remobilization of SOC with permafrost degradation and to quantify the permafrost carbon-climate feedback on global warming. This thesis provides high-resolution data on SOC storage in five study areas located in undersampled regions of the continuous permafrost zone (Zackenberg in NE Greenland; Shalaurovo and Cherskiy in NE Siberia; Ary-Mas and Logata in Taymyr Peninsula). The emphasis throughout the five different study areas is put on SOC partitioning within the landscape and soil horizon levels as well as on soil forming processes under periglacial conditions. Our results indicate large differences in mean SOC 0–100 cm storage among study areas, ranging from 4.8 to 30.0 kg C m-2, highlighting the need to consider numerous factors as topography, geomorphology, land cover, soil texture, soil moisture, etc. in the assessment of landscape-level and regional SOC stock estimates. In the high arctic mountainous area of Zackenberg, the mean SOC storage is low due to the high proportion of bare grounds. The geomorphology based upscaling resulted in a c. 40% lower estimate compared to a land cover based upscaling (4.8 vs 8.3 kg C m-2, respectively). A landform approach provides a better tool for identifying hotspots of SOC burial in the landscape, which in this area corresponds to alluvial fan deposits in the foothills of the mountains. SOC burial by cryoturbation was much more limited and largely restricted to soils in the lower central valley. In the lowland permafrost study areas of Russia the mean SOC 0–100 cm storage ranged from 14.8 to 30.0 kg C m-2. Cryoturbation is the main burial process of SOC, storing on average c. 30% of the total landscape SOC 0–100 cm in deeper C-enriched pockets in all study areas. In Taymyr Peninsula, the mean SOC storage between the Ary-Mas and Logata study areas differed by c. 40% (14.8 vs 20.8 kg C m-2, respectively). We ascribe this mainly to the finer soil texture in the latter study area. Grain size analyses show that cryoturbation is most prominent in silt loam soils with high coarse silt to very fine sand fractions. However, in profiles and samples not affected by C-enrichment, C concentrations and densities were higher in silt loam soils with higher clay to medium silt fractions. / <p>At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 3: Manuscript. Paper 4: Manuscript.</p>

La dissolution de cavités karstiques appelle à une description multi-échelle. A partir d'une discussion des hypothèses les plus fréquemment utilisées, un modèle à l'échelle du pore (ou micro-échelle) est développé pour des schémas réactifs géochimiques simples. L'impact du choix de conditions aux limites réactives ou équilibre thermodynamique est discuté. Ce modèle à l'échelle du pore est ensuite utilisé pour le développement de modèles aux échelles supérieures. Le premier problème traité considère le transport sur une surface chimiquement hétérogène et rugueuse, caractérisée par une condition mixte pour le transfert de masse. Le modèle résultant est un modèle de surface effective (ESCM). Le concept de surface effective est développé à l'aide d'une méthode de décomposition de domaine. Dans ce contexte, vitesse, pression et concentration à la petite échelle près de la surface sont estimées par une méthode de développement asymptotique par rapport aux champs loin de la surface. Des problèmes de fermeture sont alors obtenus qui sont utilisés pour définir la position de la surface effective et les conditions aux limites effectives associées. L'effet sur les propriétés effectives de la position de la surface, des nombres sans dimension est étudié. Une comparaison entre des résultats numériques à petite échelle avec ceux obtenus par le modèle effectif montre un bon accord. Dans le cas du transport dans un milieu poreux, le deuxième problème de changement d'échelle étudié, une méthode de changement d'échelle basée sur la prise de moyenne spatiale est proposée (PMM) à partir du problème à l'échelle du pore avec des conditions aux limites d'équilibre thermodynamique ou réactives non-linéaires. Une expression générale du modèle macroscopique est obtenue impliquant plusieurs propriétés effectives qui sont données par la résolution de problèmes de fermeture à l'échelle du pore. Pour une cellule unitaire représentative stratifiée, les paramètres effectifs sont obtenus analytiquement ou numériquement, alors que les propriétés pour des cellules plus complexes 2D/3D sont obtenus numériquement. L'impact sur les paramètres effectifs des propriétés physiques à l'échelle du pore (en terme de nombre de Péclet, Damköhler et ordre de la réaction) est étudié pour des cellules unitaires 1D, 2D ou 3D. Un exemple d'application du modèle macroscopique est présentée en mettant l'accent sur l'apport potentiel des termes additionnels non-classiques sur la précision des prédictions. Le modèle macroscopique de dissolution de milieu poreux est aussi utilisé comme un modèle à interface diffuse (DIM) pour décrire la dissolution d'une cavité à grande échelle, une cavité de gypse dans l'illustration traitée dans la thèse. Le modèle est basé sur l'approximation de pseudo-constituant, avec une condition d'équilibre à l'échelle du pore sur l'interface fluide-solide. Une méthodologie numérique est proposée pour choisir correctement les paramètres effectifs du DIM de façon à reproduire avec suffisamment de précision les flux et la vitesse de récession de l'interface. Une étude spécifique est effectuée sur l'impact du choix du modèle de bilan de quantité de mouvement macroscopique. De manière intéressante, les résultats numériques ne suggèrent pas un impact très important de ce choix dans le cas des problèmes aux limites traités. Des calculs ont aussi été effectués, dans le cadre d'une approximation de Boussinesq, pour évaluer l'impact éventuel de mouvements de convection naturelle. Le potentiel de la méthode est illustré dans deux cas: un correspondant à une lentille de gypse dans un aquifère, l'autre au cas d'un pilier isolé dans une carrière souterraine. Les conséquences de la dissolution sur la stabilité mécanique sont étudiées à l'aide d'un modèle géomécanique simplifié. Enfin, un cas test est étudié montrant la possibilité d'utiliser le modèle dans le cas de dissolution d'une cavité saline, matériau plus soluble que le gypse. / The karstic cavity dissolution problems are often studied from a hierarchical point of view. Based on a discussion of the frequently adopted assumptions, a pore-scale model is first developed for a simple geochemistry scheme. The impact of implementing reactive or thermodynamic equilibrium boundary condition at the dissolving surface is discussed. Such a pore-scale model is subsequently used as a basis for developing models at higher scale levels. The first problem deals with transport from a heterogeneous and rough surface characterized by a mixed boundary condition. The resulting macro-scale model takes the form of an effective surface theory. In the homogenized model developed with the effective surface concept (denote ESCM), the original rough surface is replaced locally by a homogeneous and smooth surface, where effective boundary conditions are prescribed. To develop the concept of effective surface, a multi-domain decomposition approach is applied. In this framework the velocity, pressure and concentration are estimated at the micro-scale with an asymptotic expansion of deviation terms with respect to macro-scale velocity and concentration fields. Closure problems for the deviations are obtained and used to define the effective surface position and the corresponding boundary conditions. The evolution of some effective properties and the impact of surface geometry and some dimensionless numbers are investigated. A comparison between the numerical results obtained with this effective model and those from direct numerical simulations with the original rough surface shows good agreements. In the case corresponding to mass transport in porous media, upscaling is carried out with the method of volume averaging to develop a macro-scale porous medium model (denote PMM), starting from a pore-scale transport problem involving thermodynamic equilibrium or nonlinear reactive boundary conditions. A general expression to describe the macro-scale mass transport is obtained involving several effective parameters which are given by specific closure problems. The impact on the effective parameters of the fluid properties, in terms of pore-scale Péclet number (Pe), and the process chemical properties, in terms of pore-scale Damköhler number (Da) and reaction order (n), is studied for periodic stratified, 2D and 3D unit cells. An example of the application of the macro-scale model is presented with the emphasis on the potential impact of additional, non-traditional effective parameters appearing in the theoretical development on the improvement of the accuracy of the macro-scale model. The above developed PMM is also used as a Diffuse Interface Model (DIM) to describe the evolution of a gypsum cavity formation induced by dissolution. The method is based upon the assumption of a pseudo-component dissolving with a thermodynamic equilibrium boundary condition. A methodology is proposed in order to choose suitable parameters for the DIM model and hence predict the correct dissolution fluxes and surface recession velocity. Additional simulations are performed to check which type of momentum balance equation should be used. Calculations with a variable density and Boussinesq approximation were also performed to evaluate the potential for natural convection. The results showed that the impact of density driven flows were negligible in the cases under investigation. The potential of the methodology is illustrated on two large-scale configurations: one corresponding to a gypsum lens contained within a porous rock layer and the other to an isolated pillar in a flooded gypsum quarry. Geomechanical consequences of the dissolution in terms of mechanical stability is evaluated with the help of a simplified geomechanical model. A final case is also studied in which gypsum is replaced by salt to show the applicability of the proposed methodology to a rapidly dissolving material

Among the different pulmonary fungal infections, aspergillosis, and in particular invasive pulmonary aspergillosis (IPA), are becoming the most worrying diseases in immunocompromised patients. This is due to their high incidence and mortality. Indeed, invasive aspergillosis manifests as invasive pulmonary disease accounting for 50/60% of all cases, with a mortality of 50-90% in severely immunocompromised patients. Triazoles act by inhibiting 14-α demethylase, a fungal cytochrome P450 enzyme implicated in the synthesis of ergosterol, an essential constituent of fungal cell walls. Moreover, they interact with the same cytochrome present in large quantities in the human liver, inducing possible drug-drug interactions in IPA patients. Consequently, interactions resulting from inhibitors, inductors, or substrates of cytochromes can modify the plasmatic concentrations of triazoles or other drugs administered concomitantly. To overcome these important issues, pulmonary delivery of triazoles could be an interesting alternative to conventional routes.The aim of this work was to develop triazole-based dry powders for inhalation able to be deposited adequately in the lungs, with a release of drug and a lung retention that can optimize its pharmacological action. This work focused on two active pharmaceutical ingredients (API): itraconazole (ITZ), for which improved solubility was needed, and voriconazole (VCZ), for which slow release was required.Concerning ITZ, solid dispersions for inhalation (SDIs) comprising ITZ and mannitol were previously developed in our laboratory. The selected SDI showed interesting results in terms of improved dissolution and lung retention in vivo in mice during a pharmacokinetic study. Therefore, this SDI was tested in a murine preclinical model of IPA and showed promising results in terms of prophylaxis efficacy. One aim of this work was to continue the pharmaceutical development of this promising SDI by making a scaling-up study. These methods were intended to improve the SDI’s ecological footprint and productivity by increasing the production yield and decreasing the amount of solvents and time used in its manufacture. During the first step of this study, the obtained SDI showed interesting results obtaining similar powder characteristics (i.e. amorphous content, aerodynamic performance, and dissolution profiles) from concentrated solutions using a laboratory-scale spray-dryer B-290 (Büchi, Switzerland) before using a pilot-scale spray-dryer (GEA Niro, Denmark). Then, the upscaling was performed on the pilot spray-dryer allowing the production of SDIs with increased productivity (yield and process duration). These SDIs had similar powder characteristics than the optimized lab-scale SDIs. During the second part of this work we developed VCZ based dry powder for inhalation. The aim was to slow down the release of this highly permeable and very slightly soluble API and to prolong its lung residence. To this end, various lipidic excipients were chosen. The selection took into account the potential good pulmonary tolerance of the lipids and their hydrophobicity to evaluate their ability to slow down the VCZ release (FPFs 20-25%, slowed release up to 24h, burst effect of ± 58% of VCZ dissolved within 30min). Immediate-release SDIs were also developed to have a comparator reference for the pharmacokinetic and efficacy studies (FPFs of 40%).Then, a pharmacokinetic study in mice was performed following the pulmonary administration of one immediate-release and two sustained-release SDIs (with or without PEG excipient). With an 80-fold higher pulmonary exposure over 24 hours, the slow-release SDIs presented a real interest compared to the immediate-release SDI. Moreover, in accordance with these results, VCZ plasma exposure following the administration of the SDI with PL90-H was more than 1.5-fold higher than its pulmonary exposure (AUC0-24 of 8.70 µg.h/g in the lungs and 14.70 µg.h/mL in the plasma). The slow-release formulations presented plasma exposures at least 15 times lower than their pulmonary exposures (AUC0-24 in lung of 741.40 and 686.85 µg.h/g vs plasmatic AUC0-24 of 37.44 and 42.81 µg.h.mL, respectively with and without PEG excipient). Moreover, the presence of PEG excipient did not influence the residence time and the exposure of the VCZ within the lungs. Finally, the sustained-release SDIs administration by inhalation led to VCZ lung and plasma concentrations higher than the minimal inhibitory concentration (MIC) of VCZ against Aspergillus fumigatus (1 μg/mL) over 24 h. Finally, a murine model of IPA was developed in our lab. The immunosuppression model was fixed and performed by the intraperitoneal (IP) injection of corticosteroids to induce a neutropenia state. Then, different doses of spores (from 1.10^4 to 5.10^6 spores) were inoculated to the neutropenic mice via an endotracheal instillation and the survival rate of each group was observed. Unfortunately, the survival rate resulting from the different infections were not reproducible. Therefore, these models were not suitable to conduct the efficacy study. This underlined the link between the immunosuppressive model and the infection. Indeed, the IPA murine model should be developed according to the immune state of the animal, the Aspergillus conidia species and its concentration to be used. / Doctorat en Sciences biomédicales et pharmaceutiques (Pharmacie) / info:eu-repo/semantics/nonPublished

Rural water systems (RWS) are commonly used to provide water to households for domestic uses (drinking, cleaning, washing, and sanitation) in developing countries. Water supply practitioners often classify these systems as single-use water systems (SUS) or multiple-use water systems (MUS). Smallholder farming communities in rural western hills of Nepal typically use such systems for both domestic and income-generating productive activities (e.g., agriculture, livestock, dairy, bio-gas, Rakshi), regardless of whether they were designed for single or multiple water uses. Therefore, this research frames both systems as providing multiple-use water services that enhance the productive activity and livelihoods of small- holders. Little is known on the factors that influence the productive activity of households in the western middle hills of Nepal and the impact these activities have on the technical performance of water systems (measured by duration of system breakdowns). This research identifies the extent of water-related productive activities in rural Nepali households supported by single-use water systems (SUS) vs. multiple-use water systems (MUS), and explores the factors that enables households to engage in high-levels of productive activity. The vast majority of households were found to engage in small-scale productive activities no matter what the rural water systems were designed to support, and more than half of them earned an income from water-based activities. Households engaged in high-levels of productive activity farm as a primary occupation, use productive technologies, are motivated to pursue productive activities, have received water-related productive activity training, and have received external support related to productive activities. A multinomial regression was used to predict the factors associated with high levels of productive activities undertaken by small farms. A hierarchical regression model was then used to examine both household- and system-level variables that contribute to the breakdown of rural water systems, focusing on the duration of breakdowns. The predictors of water system breakdowns include social factors (household involvement in decision-making during water system planning and construction and a household's sense of ownership toward the water system), technical factors (the management capacity of the water user committee and activity level of the system operator), economic factors (income earned from water-based productive activities), and geographic factors (the distance from the village to the water source). Finally, a conceptual model was developed to help identify strategies for implementing and scaling-up MUS. Scaling-up strategies for MUS begin with community participation in lo- cal government planning and budgeting. Under a new Constitution that went into effect in January 2017, newly formed local governments are to be provided with the funding and budget authority to determine local service priorities and how these services will be funded, designed, and implemented. The scaling-up of MUS would require local government officials, water system users, and private actors to develop the technical and institutional capacity needed to build and manage MUS, including the many support services needed by small- holder growers to realize its benefits. This research also examines the potential approaches that could enable subsistence farmers to become viable commercial producers. While growers are typically risk-adverse producers, this research identifies the mediating factors that could expand the long-term engagement of these producers in commercial agricultural production. These factors include adequate access to year-round irrigation, the use of improved production technologies and practices, improved access to rural markets, and improved production skills. The findings of this research will also be of value to Governmental, Non-Governmental Organizations (NGOs) and private sector actors who are looking to effectively mobilize their resources and expertise in support of smallholder farming in the hills of Nepal. / Doctor of Philosophy / A vast majority of farmers in the western middle hills of Nepal are smallholders who often use family labor and follow traditional agricultural and water use practices. They have been traditionally using rural water systems to meet their multiple water needs alongside domes- tic uses (drinking, cleaning, washing, and sanitation). There is growing interest for these systems to also be used for small-scale productive activities such as growing vegetables and livestock production. Evidence shows that these activities are an important source of income for farming families. However, little is known on the conditions that are needed to expand these activities and improve livelihoods. This research identifies the conditions under which rural water systems can become productive and technically sound, and outlines the strategies that can be deployed to scale-up productive activities. The research examines a broad range of perspectives (from rural farmers to development experts) on the limited commercialization of rural agriculture in the rural middle hills of Nepal and the potential approaches to promoting agricultural growth and commercialization among small landholders. The substance farmers were found to require both the means and motivation (i.e., extensive support services such as access to markets, input suppliers, irrigation and agricultural technologies, and production training) to become commercial farmers. Second, more than 90% of households were engaged in small-scale water-based production activities and more than half of them earned an income from these activities. The research revealed the conditions that enabled these households to engage in high levels of productive activities. Further, the factors that affect water system breakdowns were investigated. Since farmers are engaged in small-scale production, the interlinkages between productive income and system performance were examined. Finally, the research explores how multiple-use water services have the potential to be scaled-up in the middle hills of Nepal and beyond. Successful scaling-up strategies begin with community participation in local government planning and budgeting. This activity needs to be supported by substantial capacity building among government officials, water system users, and private actors on the factors needed to expand the productive use of water. Broad implementation of multiple-use water systems also requires careful documentation and dissemination of their benefits to key state and non-state actors. The results from this research can be used to identify appropriate households, communities, and water systems for programs focused on expanding water and agricultural productivity. Therefore, this research will have important implications for the Nepali government with regards to what policy, capacity development, and institutional arrangements need to be addressed to implement productive and sustainable rural water systems. This research can also be of special interest to Non-Governmental Organisations (NGOs) and private sector actors looking to effectively mobilize their resources and expertise relating to a smallholder farming in Nepal.