Foram estudados dois sistemas referentes à células a combustível de membrana de troca protônica através de modelos computacionais. Um protótipo de célula foi utilizado para o estudo da dinâmica de fluidos, tendo sido utilizados resultados cedidos para a validação da formulação do escoamento e posterior quantificação da contribuição do mesmo no transporte de espécies nos meios porosos. Este mesmo sistema foi explorado quanto à geometria dos canais de escoamento, sendo possível estabelecer uma hierarquia de perda por sobrepotencial de concentração em função da estequiometria do reagente. Obteve-se que, para as estequiometrias recomendadas em uma célula real, os melhores desempenhos seriam atingidos por geometrias de serpentina simples e serpentina dupla, ou tripla, para o ânodo e cátodo, respectivamente. Uma célula completa também fora estudada, sendo comparada a resultados experimentais obtidos localmente. Foi observado que o modelo utilizado apresenta boa concordância com os aspectos globais do dispositivo real, enquanto que os aspectos locais são largamente mal representados pelo modelo computacional. Este modelo também fora utilizado para quantificar a contribuição do transporte convectivo na distribuição de reagentes, corroborando e estendendo os estudos realizados com o protótipo de célula. Em suma, são sugeridas as possibilidades e desafios de se utilizar modelos computacionais para células a combustível e sistemas similares, neste caso estabelecendo a conexão entre um protótipo e um dispositivo real. / Two systems related to proton-exchange membrane fuel cells were studied through computational models. A cell prototype was used to study fluid dynamics, where the fluid flow formulation was validated against provided experimental results, being later used to quantify its contribution to species transport in the porous media. This same systems was explored regarding its flow channel geometry, where it was possible to establish a hierarchy relative to the concentration overpotential loss as a function of the reactant stoichiometry. It was obtained that, for the recommended stoichiometries for a real cell, the best performance would be attained by a single serpentine and a double, or triple, serpentine geometries for the anode and cathode, respectively. A full cell was also studied, being compared against experimental results obtained locally. It was observed that the model shows good agreement with the global aspects of the real device, while the local aspects are widely misrepresented. This model was also used to quantify the contribution of convection to reactant distribution, corroborating and extending the studies performed with the cell prototype. In all, the possibilities and challenges concerning the use of computational models with fuel cells and similar systems are suggested, particularly by establishing the connection between a prototype and a real device.

The aim of this dissertation is to appraise and critically reflect on the physical pertinence of governing equations and closure relations often used for the modeling of gas-phase transport phenomena in high-temperature solid-oxide fuel cells (SOFCs). More precisely, this work conducts a critical literature review on the concentration-induced voltage losses (i.e., concentration polarization) resulting from mass transfer limitations. Thus, the overall object of this work was to stress awareness of the limits of mathematical models studied and developed in the SOFC literature to date, and which are specifically related to concentration polarization processes. To a great extent, the design of SOFC porous layers is likened to that of optimizing the transport of multicomponent gas mixtures in structured porous catalysts, for which diffusional and flow limitations are of cardinal importance. In both cases, severe inconsistencies in mass transport models cannot be simply ignored and the main uncertainties in utilizing such models should be clarified. It is hoped that the information herein will serve usefully to support future developments of more consistent theoretical frameworks, thereby improving the confidence on the results of numerical simulations. The critical literature review has been carried out so to identify a number of physical inconsistences, ill-defined approximations, and misleading mathematical derivations. Along the review, it is argued that the choice (or, more properly, the lack of conceptual refinement) of a particular mathematical model can significantly impair the \"prediction\" of transport processes relevant to concentration-induced power losses in SOFCs. One of the keystones of this work was therefore to re-interpret and thus to reassess the frequently contradictory literature related to certain classes of gas-phase transport models pertinent to the evaluation of concentration polarization. With this revisionary approach, it is expected that one could reduce confusion, clear up apparent contradictions, and improve the possibility of gaining new insights. / Esta dissertação tem o objetivo de avaliar e refletir criticamente sobre a pertinência física de equações de conservação e de relações de fechamento, frequentemente utilizadas na modelagem multiescala de fenômenos de transporte em células a combustível de óxido-sólido (SOFC). Dêu-se atenção especial ao escoamento em \"microescala\" de misturas gasosas multicomponentes, dentro de meios porosos quimicamente reativos. Em outras palavras, esta monografia busca ressaltar quais os limites para aplicação de certas classes de modelos matemáticos, os quais têm sido desenvolvidos e utilizados na literatura de SOFCs até o presente momento. O projeto de camadas porosas de SOFCs assemelha-se à tarefa de otimizar processos de transporte em catalisadores estruturados, para os quais a existência de limitações de transporte por difusão e por escoamento tem importância primordial. Por esta razão, inconsistências originadas em modelos de fenômenos de transporte não podem ser, simplesmente, negligenciadas e, portanto, as principais incertezas ao se utilizar tais modelos devem ser devidamente esclarecidas. Espera-se, com efeito, que as informações contidas neste trabalho sejam úteis para futuros desenvolvimentos teóricos mais consistentes, de forma a aumentar a confiabilidade no uso de resultados obtidos por simulações numéricas.

L’étude des écoulements souterrains et l’ingénierie réservoir partagent le même intérêt pour la simulation d’écoulement multiphasique dans des sols aux propriétés intrinsèquement hétérogènes. Elles rencontrent également les mêmes défis pour construire un modèle à l’échelle réservoir en partant de données micrométriques tout en contrôlant la perte d’informations. Ce procédé d’upscaling est utile pour rendre les simulations faisables et répétables dans un cadre stochastique. Deux processus de mise à l’échelle sont définis: l’un depuis l’échelle micrométrique jusqu’à l’échelle de Darcy, et, un autre depuis l’échelle de Darcy vers l’échelle du réservoir. Dans cette thèse, un nouvel algorithme traitant du second upscaling Finite Volume Mixed Hybrid Multiscale Method (FV-MHMM) est étudié. L’extension au diphasique est faite au moyen d’un couplage séquentiel faible entre saturation et pression grâce à une méthode de type IMPES. / The groundwater specialists and the reservoir engineers share the same interest in simulating multiphase flow in soil with heterogeneous intrinsic properties. They also both face the challenge of going from a well-modeled micrometer scale to the reservoir scale with a controlled loss of information. This upscaling process is indeed worthy to make simulation over an entire reservoir manageable and stochastically repeatable. Two upscaling steps can be defined: one from the micrometer scale to the Darcy scale, and another from the Darcy scale to the reservoir scale. In this thesis, a new second upscaling multiscale algorithm Finite Volume Mixed Hybrid Multiscale Methods (Fv-MHMM) is investigated. Extension to a two-phase flow system is done by weakly and sequentially coupling saturation and pressure via IMPES-like method.

Made available in DSpace on 2017-07-21T19:26:00Z (GMT). No. of bitstreams: 1 Jaqueline Borges.pdf: 2739563 bytes, checksum: d03b8e7308ae59a90b2d872a90cad206 (MD5) Previous issue date: 2011-03-11 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Computed tomography (CT) is a technique that provides images of sample cross sections of porous media. The technique has shown efficient to investigate different soil physical phenomena since its first use in soil science in 1982. One of the most widely investigated issues in physics applied to porous media regards the definition of representative elementary areas (REA) to measure the soil physical properties. REA is related to the minimum area of a sample needed to represent its characteristics of interest. The objective of this work was to define REAs to be adopted for density and porosity measurements in clod samples of a clay soil. The equipment used was a first generation tomograph with 241Am source, with millimetric resolution (1.1 x 1.1 mm2). The photon detection was made with a plane NaI(Tl) solid crystal scintillator. Clods (18 samples) were also submitted to the clod method (standard method) in order to determine its. In order to reconstruct tomographic images, the program Microvis developed by Embrapa/CNPDIA was used. Concentric square areas with sizes ranging from 1.2 to 1162.8 mm2 were selected. To define REA for the following criteria were adopted: average value relative deviation between the last and each of the remaining areas not superior to 5%, 4%, 3%, 2% and 1%; and 2) that at least three consecutive areas cannot differ among themselves, using the variation criterion in item 1. The measurements through the clod method presented variation coefficient of 3.8%. Thus, it was considered that the REA was reached at the 4% variation criterion. For the criterion is similar, but the relative deviation considered cannot be superior to 10%. The REA defined for measurements corresponds to 640.1 mm2. On the other hand, for the REA found was 882.1 mm2. So, in order to carry out representative measurements of these two physical properties the REA must be at least 882.1 mm2. Key-words: Representative Measurements. Tomographic Images. Porous Media. / A Tomografia Computadorizada (TC) trata-se de uma técnica que fornece imagens de seções transversais de amostras de meios porosos. A técnica tem se mostrado eficaz na investigação de diversos fenômenos físicos do solo desde sua primeira aplicação nessa área em 1982. Uma questão amplamente investigada em física aplicada a meios porosos diz respeito à definição de áreas elementares representativas (AER) para medidas de propriedades físicas do solo. A AER está relacionada com a área mínima de uma amostra, necessária para representar suas características de interesse. O objetivo deste trabalho foi definir AERs a serem adotadas para medidas da densidade e porosidade de amostras de torrão de um solo de textura argilosa. Para isso, utilizou-se um tomógrafo de primeira geração, equipado com fonte de 241Am com resolução da ordem de milímetros (1,1 x 1,1 mm2). A detecção dos fótons foi feita com um detector de cintilação sólida de NaI(Tl) do tipo plano. Os torrões (18 amostras) também foram submetidos ao método do torrão parafinado (TP, método padrão) para determinação da sua Para reconstrução das imagens tomográficas utilizou-se o programa Microvis desenvolvido pela Embrapa/CNPDIA. Nestas imagens foram selecionadas áreas quadrangulares concêntricas com dimensões variando de 1,2 até 1162,8 mm2. Para definição da AER para foram utilizados os seguintes critérios: 1) desvio relativo do valor médio entre a última e cada uma das demais áreas não superior a 5%, 4%, 3%, 2% e 1% e 2) que pelo menos três áreas consecutivas não devem diferir entre si, utilizando o critério de variação do item 1. As medidas de pelo método do TP apresentaram coeficiente de variação da ordem de 3,8%. Assim, considerou-se que a AER foi atingida no critério de 4% de variação. Para o critério é semelhante, mas o desvio relativo considerado não deve ser superior a 10%. A AER definida para medidas de corresponde a 640,1 mm2. Para a AER encontrada é de 882,1 mm2. Portanto, para a realização de medidas representativas destas duas propriedades físicas, a AER deve ser de pelo menos 882,1 mm2. Palavras-chave: Medidas Representativas. Imagens Tomográficas. Meios Porosos.

L'objectif est d'élaborer un nouveau procédé de fabrication de mousses métalliques par voie de fonderie en modélisant l'infiltration et la solidification d'un métal liquide dans un milieu poreux. La modélisation est faite en deux étapes.Tout d'abord, à l'échelle locale un brin de la mousse métallique est considéré comme un tube capillaire et l'infiltration et solidification d'un métal liquide dans un moule cylindrique est étudiée. Deuxièmement,le modèle macroscopique de la solidification diffusive d'un métal liquide dans un milieu poreux est obtenu par prise de moyenne volumique. Le modèle local est codée dans un outil CFD opensource et trois études paramétriques ont été faites permettant la détermination des relations de la longueur et le temps d'infiltration en fonction de paramètres de fonctionnement. La modélisation de la solidification d’un métal liquide dans un milieu poreux est simplifié en considérant que le moule est complètement saturé par un métal liquide au repos,par suite la solidification se produit par diffusion pure (pas de convection). L'équilibre thermique local (LTE) est considéré entre les phases solide et liquide du métal tandis qu'un non équilibre thermique local (LTNE) est retenue entre la phase métallique et le moule. Les problèmes de fermeture associés ainsi que le problème macroscopique ont été résolus numériquement. / The objective of this work is to elaborate a new manufacturing process of metal foams via casting by modelling the infiltration and solidification of liquid metal inside a porous medium.However, due to the complexity of this problem the study is divided into two steps. First, at local scale one strut of the metal foam is considered as a capillary tube and the infiltration and solidification of liquid metal inside a cylindrical mould is studied. Second, a macroscopic model of diffusive solidification is derived using the volume average method. The local model is coded in an open source CFD tool and three parametric studies were done where the relations between the infiltration length and time as function of the operating parameters are determined. The modelling of the solidification of liquid metal inside a porous medium is simplified by considering that the mould is fully saturated by liquid metal at rest, solidification occurs by pure diffusion. Local thermal equilibrium (LTE) is considered between the solid and liquid phases of the metal while local thermal non equilibrium (LTNE) is retained between the metallic mixture and the mould. The associated closure problems as well as the macroscopic problem were numerically solved.

La pyrolyse est la décomposition thermochimique de la biomasse en gaz de synthèse valorisables en biocarburants. Cette technologie, propre et renouvelable, nécessite aujourd’hui des efforts de recherche et de développement afin de prouver sa compétitivité par rapport aux autres sources d’énergie. L’objectif de cette thèse est de développer un modèle macroscopique de la pyrolyse en utilisant la méthode de prise de moyenne volumique. Le modèle sera ensuite utilisé pour faire des études numériques afin de caractériser le procédé et améliorer les performances des réacteurs. Une approche en deux temps a été établie afin d’atteindre notre objectif. D’abord, des modèles macroscopiques d’écoulements à masse volumique variable en milieu poreux ont été développés. Ce type d’écoulements est similaire à celui mis en jeu en pyrolyse pour deux deux raisons: la masse volumique varie sous l’effet de gradients forts de température et le réacteur de pyrolyse peut être considéré comme un milieu poreux à double porosité (porosité à l’échelle du lit et porosité à l’échelle de la particule). Les résultats théoriques ont montré que les équations de conservation macroscopiques (continuité, quantité de mouvement et énergie) et les propriétés effectives (masse volumique, perméabilité et diffusivité thermique) font apparaitre de nouveaux termes résultants de la variation de densité. La forme explicite de ces termes a été établie et validée par simulations numériques. Les résultats obtenus ont été utilisés dans un deuxième temps afin de développer un modèle macroscopique de la pyrolyse. / Pyrolysis is a thermo-chemical conversion of biomass into bio-fuels. This technology has not been fully developed and its competitiveness against other sources of energy is yet to be proven. The aim of this work is to derive a macroscopic model of pyrolysis by means of volume averaging method. The obtained macroscopic model can then be used to conduct fast and low-cost numerical simulations to characterize the process and improve the reactor efficiency. To achieve our objective, a two-steps methodology has been established. First, the fundamental problem of variable density flow in porous media has been investigated. The physical phenomena in this kind of problem are very similar to those involved in pyrolysis for two reasons: the fluid density varies due to high temperature gradients and the pyrolysis reactor can be considered as a double porosity medium (porosity at the reactor scale and porosity at the biomass particle scale). The obtained macroscopic conservation equations (continuity, momentum and energy) and the effective properties (density, permeability and thermal diffusivity) contain additional terms resulting from the fluid density variation. The explicit form of these terms has been established and their components have been computed. The resulting models of the first step have then been used to develop a macroscopic model of the pyrolysis in the second part of our study.

A Ressonância Magnética Nuclear (RMN) é uma importante técnica empregada nas principais áreas de conhecimento, tais como, Física, Química e Medicina. Importantes trabalhos da RMN aplicada ao estudo da dinâmica de moléculas em fluidos presentes em meios porosos permitiram que esta técnica ganhasse também notoriedade na indústria do petróleo. O presente projeto é fundamentado em alguns destes trabalhos seminais, reproduzindo, através de modelos físico-computacionais, os principais efeitos físicos da difusão e a consequente relaxação dos spins dos núcleos atômicos presentes nas moléculas dos fluidos imersos nos meios porosos. Os métodos teóricos utilizados para a interpretação dos parâmetros de relaxação transversal (T2) e longitudinal (T1), levam em consideração as propriedades macroscópicas da dinâmica de um ensemble de spins nucleares. Para a compreensão da difusão e relaxação nuclear em meios porosos, é de fundamental importância o emprego de métodos físico-computacionais que levam em consideração tanto a microestrutura quanto a difusão translacional dos spins nucleares através desses meios. Dentro dos modelos propostos nesta dissertação, foram utilizados conceitos básicos de mecânica estatística e de relaxação nuclear, resultando em programas envolvendo consideráveis recursos computacionais. Para a realização das simulações numéricas, foi necessária a construção da microestrutura digital dos meios porosos, usando como base do processo as imagens obtidas por microtomografia de raios-X 3D. Tendo disponível a microestrutura digital, o processo de difusão através do meio poroso e a consequente relaxação dos spins nucleares foram simulados através de modelos Random Walk, um exemplo clássico de Método de Monte Carlo. As simulações numéricas dos experimentos clássicos de RMN, empregados para a determinação de distribuições de tempos de relaxação transversal (conhecidas pela sigla CPMG), e mais avançados, para a determinação de taxas de troca entre diferentes sítios dos meios porosos (experimentos de troca ou Exchange), produziram importantes resultados para a compreensão da dinâmica de fluidos em meios porosos através de difusão translacional. / Nuclear Magnetic Resonance (NMR) is a versatile technique applied in many different areas of science such as Physics, Chemistry and Medicine. Several important NMR works applied in the study of the molecular dynamics of fluids inside porous media allowed this technique to become popular in the oil industry. This project is based on some of these seminal works. Using computational physics models, the main physical effects of diffusion and relaxation for the nuclear spins in an ensemble of molecules in a fluid permeating the porous media are reproduced. The theoretical methods used to interpret the transverse (T2) and longitudinal (T1) relaxation parameters take into consideration the macroscopic dynamical properties of an ensemble of nuclear spins. In order to comprehend the diffusion and relaxation in porous media, the use of computational physics methods that take into account the translational diffusion of the nuclear spins inside the media as well as its microstructure is essential. In the proposed models by this dissertation, basic statistical mechanics and nuclear relaxation concepts were used, resulting in programs that demand considerable computational resources. For the numerical simulations it was necessary to digitally construct the microstructure of the porous media and that was achieved with the use of X-ray 3D microtomography. On this digital microstructure, the diffusion process and the resulting relaxation of the nuclear spins were simulated with random walk models, an example of the Monte Carlo Method. The numerical simulation of classic NMR experiments, such as the one used to determine the distribution of relaxation times (known as CPMG), and more advanced ones, such as the one used to determine exchange rates between different sites (2D exchange experiments), produced fundamental results to the understanding of the fluid dynamics in porous media.

Le Laboratoire de Thermique, Energétique et Procédés de Pau travaille en parallèle depuis plusieurs années sur les deux thématiques suivantes, la modélisation des réacteurs multifonctionnels et la modélisation des transports en milieux poreux. Dans ce travail, un modèle général de transports en milieu poreux multiphasique multiconstituant réactif est développé puis appliqué au cas particulier de l’absorption réactive afin de décrire les phénomènes se produisant au sein du garnissage. La modélisation débute par une description classique de chacune des phases continues grâce aux équations de conservation de la masse, de la quantité de mouvement et de l’énergie. Les phases fluides sont multiconstituantes et le siège de réactions homogènes. Bien qu’à cette échelle les mécanismes soient parfaitement décrits, le passage à la simulation impose une étape d’homogénéisation par prise de moyenne. Ce changement d’échelle conduit à un système d’équations à l’échelle locale. Le modèle est appliqué à l’absorption du dioxyde de carbone dans une solution aqueuse d’hydroxyde de sodium au sein d’une colonne garnie. Les simulations bidimensionnelles instationnaires fournissent des informations locales sur le procédé. Les résultats obtenus montrent une bonne adéquation avec la connaissance classique des mécanismes d’absorption réactive. / The “Laboratoire de Thermique, Energétique et Procédés de Pau” works on the two following subjects, modeling of multifunctional reactor and modeling of transport in porous media. In this work, a general model of transport in multiphase multi component reactive porous media is developed and then applied to the specific case of reactive absorption in order to describe the phenomena which occur within the packing. Modeling begins with a classic description of each continuous phase through the equations of conservation of mass, momentum and energy. The fluid phases are multi component and reactive. Although mechanisms are perfectly described at this scale, the simulation imposes a homogenization by volume averaging. This change of scale leads to a set of equations at the local scale. The model is applied to the absorption of carbon dioxide in an aqueous solution of sodium hydroxide within a packed column. Unsteady two-dimensional simulations provide local information about the process. The results obtained show a good agreement with the classical knowledge of reactive absorption mechanisms.

Foam flow in porous media is important for many industrial operations such as enhanced oil recovery, remediation of contaminated aquifers and CO2 sequestration. The application of foam in these processes is due to its unique ability to reduce gas mobility and to divert gas to low permeability zones in porous media which otherwise would not be reached. To achieve optimum success with foam as a displacing fluid in oil recovery and remediation operations, it is essential to understand how different parameters influence foam flow in porous media. In this thesis, a variety of experimental techniques were used to study foam stability, foam rheology as well as the dynamics and patterns of oil displacement by foam under different boundary conditions such as surfactant formulation, oil type, foam quality (gas fraction) and porous media geometry. Bulk scale studies showed that foam stability was surfactant and oil dependant such that decreasing oil carbon number and viscosity decreased the stability of foam. However, no meaningful correlation was found between foam stability at bulk scale and the efficiency of oil displacement in porous media for the various surfactants studied in this work. Additionally, our results show that foams consisting of smaller bubbles do not necessarily correspond to higher apparent viscosity as the foam quality is also crucial. For the same foam quality decreasing bubble size resulted in higher apparent viscosity. Although in theory a higher apparent viscosity (i.e. higher foam quality) would be ideal for displacement purposes, increasing foam quality resulted in less stable foam in porous media due to formation of thin films which were less stable in the presence of oil. The effect of pore geometry on foam generation and oil displacement has also been investigated. Our findings provide new insights about the physics and complex dynamics of foam flow in porous media.

Ce travail s'intéresse a` l'identification des mécanismes physiques en jeu lors de la déshydratation assistée thermiquement. Une cellule de filtration/compression a été instrumentée avec des capteurs de pression liquide. Ces mesures permettent de suivre la formation du gâteau de filtration, sa consolidation puis les évolutions thermomécaniques induites par l’apport de chaleur. Sous l'effet du chauffage, elles montrent l'apparition d’un gradient de pression liquide, qui induit un écoulement additionnel. Un modèle mathématique est ensuite proposé. Des caractérisations en cellule de compression/perméabilité sont utilisées pour de´terminer les lois d'évolution des propriétés du milieu poreux. Le modèle permet de retrouver les évolutions des grandeurs macroscopiques et locales mais pas les cinétiques. Ceci a été attribué au modèle de déformation du milieu poreux choisi. / In the scope of thermally assisted mechanical dewatering process, this work focuses on the identification of physical mechanisms involved in the thermally assisted mechanical dewatering. Operating conditions ensure that the water is expelled in liquid phase. An experimental study on talc and cellulose saturated suspensions highlights the gains in terms of final dry solid contend involved by the thermal intensification. The filtrationcompression test cell is instrumented with pore liquid sensors along the cake thickness that allows to measures the pressure of the interstitial liquid phase. Obtained data highlight an increase of the liquid pressure in the heated side of the cake during thermally assisted mechanical dewatering. Indeed, as the temperature increases the water density decrease. As the cake consolidation restricts the flow a liquid pressure gradient reappears inside the cake. As the temperatures on the filter side of the cake increase, the thermally induced liquid pressure gradient vanishes with an additional filtrate outflow. In order to measure the porous media properties a compression-permeability cell has been build. This apparatus allows permeability measurement to be conducted for different temperature and loading on the porous media. A physical model including the thermal pressurization has been implemented in COMSOL Multiphysics in order to simulate the process. The model confirms the thermal pressurization occurring during thermally assisted mechanical dewatering, but because of deviation in the calculated mechanical behavior and temperature of the porous media, the model can’t depict the experimental additional filtrate outflow.