La circulation sanguine joue un rôle vital en microcirculation, et ce pour le transport de l'oxygène, le dioxide de carbone et d'autres nutriments. Les globules rouges (GR) constituent la majorité des cellules du sang, c'est pourquoi par "écoulement sanguin", nous entendrons "écoulement d'une suspension de GR". Pendant longtemps l'écoulement sanguin était vu comme un phénomène passif où les GR sont considérés comme des cargos d'oxygène. La vision moderne est tout autre: l'écoulement sanguin est bel et bien un phénomène actif. Les GR ainsi que les cellules endothéliales (qui tapissent les faces internes des vaisseaux sanguins) sont impliquées dans un grand nombre de signalisations biochimiques induites par les contraintes hydrodynamiques, la route vers des régulations vasomotrices sans l'intervention du système nerveux. Par exemple, les GR ne transportent pas que l'oxygène, mais également de l'ATP (adenosine triphosphate), qui est libérée suite à des changements de conformation de protéines membranaires induite par les contraintes hydrodynamiques. Cette thèse est dédiée à la circulation sanguine et son couplage avec la signalisation biochimique ayant lieu en microcirculation. Plus précisément, les questions traités dans cette thèse sont i) la dynamique des GR, ii) le problème de la diffiusion-advection d'espèces chimiques au sein des écoulements sanguins, et iii) le rôle de la géométrie des réseaux vasculaires dans le processus de la signalisation biochimique mentionnés plus haut. Dans un premier temps nous analysons la dynamique de GR dans un écoulement de Poiseuille en présence de valeurs réalistes de contraste de viscosité. Dans un deuxième temps nous développons un modèle de diffusion-advection et le couplons aux écoulements sanguins en adoptant la méthode de Boltzmann sur réseaux; nous exploitons ensuite formulation en l'appliquant au problème de la libération de l'ATP par les GR sous écoulement. Enfin nous présentons des résultats préliminaires pour la problématique générale de l'écoulement sanguin mettant en jeu l'ATP libéré par les GR et la signalisation de calcium par les cellules endothéliales. Cette étude constitue un premier pas vers le problème général et ambitieux de la régulation locale mechano-biochimique impliquée dans la microcirculation. / Blood flow in microcirculation is vital for oxygen, carbon dioxide and nutrients transport. Most of blood cells are red blood cells (RBCs), so that by blood flow we mean flow of a suspension of RBCs. For long time blood flow has been mainly considered as a passive phenomenon, in which RBCs are viewed as passive carriers of oxygen. The modern view is completely different: blood flow is more active than we thought. The RBCs as well as vascular endothelial cells covering the internal walls of blood vessels are involved in a number of biochemical signaling processes that are triggered by shear stress eliciting a number of biochemical events, and ultimately resulting into vasomotor regulation without participation of the nerve system. For example, RBCs do not only carry oxygen but also ATP (adenosine triphosphate) , the release of which occurs thanks to changes of RBC membrane protein conformations caused by shear stress. Released ATP reacts with some endothelial membrane receptors leading to vasodilation. This thesis is devoted to blood flow and its coupling to biochemical signaling. More precisely, we investigate i) the dynamics of RBCs, ii) the advection diffusion of chemicals in blood flow and the role of iii) the geometry of vessel networks, in the mentioned signaling processes in microcirculations. Firstly, we study the RBC dynamics in a pipe flow with realistic viscosity contrast values, where a link between shape dynamics and rheology is established. Secondly, we develop an advection-diffusion solver that can handle general moving curved boundaries based on lattice-Boltzmann method (LBM); we then implement it for the study of the problem of ATP release from RBCs under shear flow. Membrane tension and deformation induced by shear stress together with vessel network geometry contribute to ATP release. Finally we demonstrate the capability of applying our model and our numerical tool to the complete problem of blood under flow involving ATP release from RBCs and endothelial calcium signaling as a preliminary step to the ambitious task of mechano-involved local regulation events in microcirculation.

O estudo do transporte de partículas tem uma importância fundamental em diversas áreas de pesquisas como, por exemplo, na formação de gelo em uma aeronave pois pode afetar a sua sustentação e estabilidade. Tamanha é a preocupação com a segurança de vôo que diversos estudos têm sido realizados, resultando em códigos computacionais como o LEWICE nos Estados Unidos, TRAJICE no Reino Unido, ONERA na França e CANICE no Canadá. No Brasil, um dos estudo é feito pela EMBRAER em parceria com algumas instituições. O objetivo deste trabalho é desenvolver um algoritmo que possa ser empregado na trajetória das partículas, utilizando uma formulação euleriana que elimina a dificuldade da semeadura de partículas específica da formulação lagrangiana na determinação da fração de volume da partícula. O método empregado é dos volumes finitos em malhas não-estruturadas cuja principal chave está na reconstrução mínimos quadrados de alta ordem com restrição nos contornos. O desenvolvimento do trabalho engloba 3 etapas: definição da geometria e geração das malhas; utilização de um solver para o tratamento do escoamento do ar e obtenção do campo de velocidade; implementação e utilização do esquema numérico com reconstrução mínimos quadrados de alta ordem para simular o cálculo da fração de volume com imposição de condições limites apropriadas no contorno do corpo. Os resultados dos testes realizados mostram que o esquema numérico com reconstrução mínimos quadrados pode ser empregado na resolução de equações que apresentam uma região de descontinuidade, como é o caso da região de sombra, reduzindo a largura da banda de difusão numérica e overshoots. / The particle transport study has a fundamental importance in diverse research area like in the icing accretion on an aircraft because that can affect its sustentation and stability. The concern is so big that many researches have been carried through, resulting in computational codes like the LEWICE in the United States, TRAJICE in the United Kingdom, ONERA in France and CANICE in Canada. In Brazil, one of the study has been made by the EMBRAER with some institutes. The goal of this work is to develop an algorithm that can be used in the particles trajectory study, using an Eulerian method that eliminates the difficulty particle sowing, particular of the Lagrangian method, in the determination of the droplet fraction volume. This is made by the finite volume method on unstructured meshes whose main key is the high order reconstruction with restriction on the boundary. The development of the work involves 3 stages: geometry definition and mesh generation; using code for the treatment of the air flow and obtained flow velocity; use of the high order numerical scheme least square reconstruction to simulate the droplet fraction volume result with imposition of appropriate limit conditions in the body contour. The realized simulations shown that Least Square method can be used in problem resolution that present descontinuos region like is shadow region reducing numerical diffusion and overshoots.

Dans de nombreux réservoirs géothermiques profonds, l'échangeur thermique est un massif de roches chaudes fracturées. Comment la morphologie des fractures influence-t-elle le flux hydraulique et le champ de température lorsque de l'eau froide est injectée dans ce milieu ? Cette question est abordée à l'échelle de la fracture, de manière numérique, à l'aide de deux méthodes : par différences finies, en 2D, et par méthode de Boltzmann sur réseau, en 3D. Les calculs en différences finies montrent qu'une ouverture auto-affine induit une variabilité spatiale du flux hydraulique et du champ de température. Dans certains cas, il se crée une chenalisation du fluide, qui réduit l'efficacité du transfert thermique entre la roche et le fluide. Ces caractéristiques hydro-thermiques sont essentiellement engendrées par les plus grandes longueurs d'ondes de ces ouvertures auto-affines. Une première application à grande échelle, dimensionnée suivant les caractéristiques du site géothermique de Soultz-sous-Forêts (Alsace, France), est proposée. Une deuxième application consiste à évaluer la perméabilité du sous-sol de Draix (Alpes, France). Pour cela, la géométrie de fractures présentes dans le sous-sol est caractérisée précisément à partir de carottes géologiques. La topographie des fractures, mesurée avec un laser, est utilisée pour reconstituer la morphologie de l'ouverture des fractures. Ces données sont utilisées pour quantifier la perméabilité, ici élevée -- de l'ordre de 10^-9-10^-8 m^2. À l'aide de méthodes de Boltzmann sur réseau, la modélisation du comportement hydro-thermique dans des fractures présentant une topographie avec de fortes pentes est abordée. Au voisinage d'une aspérité de forme simple, le comportement hydro-thermique modélisé est très différent (recirculation) de celui dans le reste de la fracture. L'ensemble de ces modélisations suggère qu'il est nécessaire de prendre en compte la morphologie des fractures pour estimer le comportement hydro-thermique en géothermie.

The construction of compact high-order Residual Distribution schemes for the discretizationof steady multidimensional advection-diffusion problems on unstructuredgrids is presented. Linear and non-linear scheme are considered. A piecewise continuouspolynomial approximation of the solution is adopted and a gradient reconstructionprocedure is used in order to have a continuous representation of both thenumerical solution and its gradient. It is shown that the gradient must be reconstructedwith the same accuracy of the solution, otherwise the formal accuracy ofthe numerical scheme is lost in applications in which diffusive effects prevail overthe advective ones, and when advection and diffusion are equally important. Thenthe method is extended to systems of equations, with particular emphasis on theNavier-Stokes and RANS equations. The accuracy, efficiency, and robustness of theimplicit RD solver is demonstrated using a variety of challenging aerodynamic testproblems.

Na presente tese é apresentada uma nova solução analítica para a equação de ad-vecção-difusão bidimensional transiente para simular a dispersão de poluentes na atmosfera. Para tanto, a equação de advecção-difusão é resolvida pela combinação da transformada de Laplace e da técnica GILTT (Generalized Integral Laplace Transform Technique). O fechamento da turbulência para os casos Fickiano e não-Fickiano é considerado. É investigado o problema de modelagem da dispersão de poluentes em condições de ventos fortes e fracos considerando, para o caso de ventos fracos, a difusão longitudinal na equação de advecção-difusão. Além disso, foi incluída no modelo a velocidade vertical e avaliada sua influência considerando-se o campo de velocidades constante e também geradas via LES (Large Eddy Simulation), para poder simular uma camada limite turbulenta mais realística. Os resultados obtidos por essa metodologia são validados com resultados experimentais disponíveis na literatura. / In the present thesis it is presented a new analytical solution for the transient two- dimensional advection-diffusion equation to simulate the pollutant dispersion in atmosphere. For that, the advection-diffusion equation is solved combining the Laplace transform and the GILTT (Generalized Integral Laplace Transform Technique) techniques. The turbulence closure for Fickian and non-Fickian cases is considered. It is investigated the problem of modeling the pollutant dispersion in strong and weak winds considering, for the case of low wind conditions, the longitudinal diffusion in the advection-diffusion equation. Moreover, it was considered in the model the vertical velocity and its influence was evaluated considering velocities field constant and also generated by means of LES (Large Eddy Simulation), to simulate a more realistic turbulent boundary layer. The results attained by this methodology are validated with experimental results available in literature.

Na presente tese é apresentada uma nova solução analítica para a equação de ad-vecção-difusão bidimensional transiente para simular a dispersão de poluentes na atmosfera. Para tanto, a equação de advecção-difusão é resolvida pela combinação da transformada de Laplace e da técnica GILTT (Generalized Integral Laplace Transform Technique). O fechamento da turbulência para os casos Fickiano e não-Fickiano é considerado. É investigado o problema de modelagem da dispersão de poluentes em condições de ventos fortes e fracos considerando, para o caso de ventos fracos, a difusão longitudinal na equação de advecção-difusão. Além disso, foi incluída no modelo a velocidade vertical e avaliada sua influência considerando-se o campo de velocidades constante e também geradas via LES (Large Eddy Simulation), para poder simular uma camada limite turbulenta mais realística. Os resultados obtidos por essa metodologia são validados com resultados experimentais disponíveis na literatura. / In the present thesis it is presented a new analytical solution for the transient two- dimensional advection-diffusion equation to simulate the pollutant dispersion in atmosphere. For that, the advection-diffusion equation is solved combining the Laplace transform and the GILTT (Generalized Integral Laplace Transform Technique) techniques. The turbulence closure for Fickian and non-Fickian cases is considered. It is investigated the problem of modeling the pollutant dispersion in strong and weak winds considering, for the case of low wind conditions, the longitudinal diffusion in the advection-diffusion equation. Moreover, it was considered in the model the vertical velocity and its influence was evaluated considering velocities field constant and also generated by means of LES (Large Eddy Simulation), to simulate a more realistic turbulent boundary layer. The results attained by this methodology are validated with experimental results available in literature.

No description available.

Na presente tese é apresentada uma nova solução analítica para a equação de ad-vecção-difusão bidimensional transiente para simular a dispersão de poluentes na atmosfera. Para tanto, a equação de advecção-difusão é resolvida pela combinação da transformada de Laplace e da técnica GILTT (Generalized Integral Laplace Transform Technique). O fechamento da turbulência para os casos Fickiano e não-Fickiano é considerado. É investigado o problema de modelagem da dispersão de poluentes em condições de ventos fortes e fracos considerando, para o caso de ventos fracos, a difusão longitudinal na equação de advecção-difusão. Além disso, foi incluída no modelo a velocidade vertical e avaliada sua influência considerando-se o campo de velocidades constante e também geradas via LES (Large Eddy Simulation), para poder simular uma camada limite turbulenta mais realística. Os resultados obtidos por essa metodologia são validados com resultados experimentais disponíveis na literatura. / In the present thesis it is presented a new analytical solution for the transient two- dimensional advection-diffusion equation to simulate the pollutant dispersion in atmosphere. For that, the advection-diffusion equation is solved combining the Laplace transform and the GILTT (Generalized Integral Laplace Transform Technique) techniques. The turbulence closure for Fickian and non-Fickian cases is considered. It is investigated the problem of modeling the pollutant dispersion in strong and weak winds considering, for the case of low wind conditions, the longitudinal diffusion in the advection-diffusion equation. Moreover, it was considered in the model the vertical velocity and its influence was evaluated considering velocities field constant and also generated by means of LES (Large Eddy Simulation), to simulate a more realistic turbulent boundary layer. The results attained by this methodology are validated with experimental results available in literature.

No description available.

Concrete is a popular construction material for bridges, due to its high durability and energy efficiency. An important concern for concrete bridges is the possible occurrence of chloride- induced corrosion in prestressing strands and reinforcing bars, which may substantially impact the service life of such structures. Chloride- induced corrosion is a complicated electrochemical process which is affected by heat transfer, moisture flow and transport of chemical species through the concrete pore network. Reliable and robust analytical tools are required to allow multi-physics simulations of steel corrosion. This study has developed a nonlinear finite element analysis program, called VT-MultiPhys, to enable multi-physics simulations, including analyses of chloride-induced corrosion. The program includes constitutive laws, element formulations and global solution schemes to allow the analysis of steady-state (static) and time-dependent (dynamic) problems, involving multiple, coupled processes such as mechanical deformation, heat transfer, mass flow and chemical reactions combined with advective/diffusive transport of the various species. Special analysis schemes, based on the streamline-upwind Petrov-Galerkin (SUPG) method, have also been implemented to address the spatial instabilities which characterize analyses of advection-dominated transport. The finite element modeling scheme, constitutive laws and boundary conditions for analysis of chloride-induced corrosion are described in detail. The constitutive laws can be combined with inelastic material models to capture the damage (e.g., cracking) due to chloride-induced corrosion. A set of verification analyses is presented, to demonstrate the capabilities of VT-MultiPhys to conduct different types of simulations and reproduce the closed-form analytical solutions of simple cases. Validation analyses for heat conduction, moisture flow and chloride transport, using data from experimental tests in the literature, are also presented. / Master of Science / The deterioration of concrete structures and infrastructures due to the chloride-induced corrosion in prestressing strands and reinforcing bars may substantially impact the service life of such structures. Chloride-induced corrosion is a complicated electrochemical process which is initiated and proceeds due to the chloride attacks at the surfaces of concrete structures and ends in the volume expansion, cracking and spalling of concrete. Due to the lack of comprehensive modeling tool, which can simultaneously comprise the influential factors in chloride-induced corrosion, the realistic estimation of the service life of reinforced concrete structures is still challenging. Reliable and robust analytical tools are required to allow multi-physics simulations of steel corrosion. This study has developed a comprehensive finite element analysis program, called VT-MultiPhys, for calculating and monitoring the contribution of chloride ions to chloride-induced corrosion during service life of concrete structures. The present analysis program enables modeling of the coupled physical process including heat transfer, moisture flow and transport of chemical species through the concrete pore network. Also, by modeling the influence of flexural cracks on chloride transport in concrete, the analysis program is able to predict the rate of steel corrosion in cracked concrete structures. A set of verification analyses is presented, to demonstrate the capabilities of VT-MultiPhys to conduct different types of simulations of heat conduction, moisture flow and chloride transport and the comparison is found to be satisfactory. The element formulations and solution algorithms in VT-MultiPhys also allow the investigation of other long-term deterioration mechanisms, such as carbonation-induced corrosion, alkali-silika reaction (ASR) and sulfate attack. The present contribution will hopefully enable and facilitate future research in these topics, through the formulation and implementation of proper constitutive laws and chemical reaction equations.