Les propriétés mécaniques et fonctionnelles des matériaux métalliques sont conditionnées par leurs microstructures, qui sont elles-mêmes héritées des traitements thermomécaniques subis. Etre capable de prévoir et simuler la microstructure et ses hétérogénéités lors des procédés de mise en forme complexes est récemment devenu crucial dans l'industrie métallurgique. C'est également un véritable challenge d'un point de vue numérique qui met en évidence l'importance des matériaux numériques dans les nouvelles méthodes de modélisation. Dans ce travail, nous nous intéressons à un modèle en champ complet récent basé sur la méthode level set (LS) dans un cadre éléments finis (EF) pour la modélisation des mécanismes de recristallisation.Les points forts de cette approche par rapport à l'état de l'art ont motivé le développement d'un logiciel appelé DIGIMU® par la société TRANSVALOR avec le soutien de grandes entreprises industrielles. Toutefois, le principal inconvénient de cette approche, commun aux autres méthodes en champ complet utilisant des maillages EF non structurés, reste son coût numérique important.Le principal objectif de ce travail a donc été d'améliorer considérablement le coût numérique de la formulation LS utilisée dans le contexte de maillages EF non structurés. De nouveaux développements génériques ont été réalisés pour améliorer l'efficacité globale du modèle. La formulation 2D LS existante, déjà utilisée pour modéliser la croissance de grains, la recristallisation statique et l'effet d'ancrage de Smith-Zener, a été étendue et améliorée afin de modéliser ces mécanismes en 3D pour des polycristaux à grand nombre de grains en des temps de calcul raisonnables. / Mechanical and functional properties of metallic materials are strongly related to their microstructures, which are themselves inherited from thermal and mechanical processing. Being able to accurately predict and simulate the microstructure and its heterogeneities after complex forming paths recently became crucial for the metallurgy industry. This is also a real challenge from a numerical point of view which highlights the importance of digital materials in new modeling techniques. In this work, we focus on a recent front-capturing full field model based on the level set (LS) method within a finite element (FE) framework to model recrystallization mechanisms.The strengths of this approach comparatively to the state of the art have motivated the development of a software package called DIGIMU® by the company TRANSVALOR with the support of major industrial companies. However, the main drawback of this approach, common with other front-capturing full field approaches working on unstructured FE meshes, is its important computational cost, especially in 3D.Main purpose of this work was finally to drastically improve the numerical cost of the considered LS-FE formulation in context of unstructured FE meshes. New generic numerical developments have been proposed to improve the global efficiency of the model. The existing 2D LS formulation, already used to model grain growth, static recrystallization and the Smith-Zener pinning effect, has been extended and improved in order to model these mechanisms in 3D for large-scale polycrystals with reasonable computational costs.

O presente trabalho trata da formulação e implementação de modelos numéricos baseados no Método dos Elementos de Contorno (MEC). Inspirando-se em problemas de engenharia, uma abordagem multidisciplinar é proposta como meio de representação numérica mais realista. Há materiais de uso corrente na engenharia que possuem resposta dependente do tempo. Nesta tese os fenômenos dependentes do tempo são abordados por meio da Mecânica Viscoelástica Linear associada a modelos reológicos. Neste trabalho, se apresenta a dedução do modelo constitutivo de Maxwell para ser utilizado via MEC. As equações deduzidas são verificadas em problemas de referência. Os resultados mostram que a formulação deduzida pode ser utilizada para representar estruturas compostas, mesmo em casos envolvendo uma junção entre materiais viscoelásticos e não viscoelásticos. Adicionalmente as formulações apresentadas se mantém estáveis na presença de fissuras de domínio e bordo. Verifica-se que a formulação clássica dual pode ser utilizada para simular o comportamento de fissuras com resposta dependente do tempo. Essa constatação serve de base para maiores investigações no campo da Mecânica da Fratura de materiais viscoelásticos. Na sequência, mostra-se como o MEC pode ser aliado a conceitos probabilísticos para fazer estimativas de comportamentos a longo prazo. Estas estimativas incluem as incertezas inerentes nos processos de engenharia. As incertezas envolvem os parâmetros materiais, de carregamento e de geometria. Por meio do conceito de probabilidade de falha, os resultados mostram que as incertezas relacionadas às estimativas das cargas atuantes apresentam maior impacto no desempenho esperado a longo prazo. Esta constatação serve para realizar estudos que colaborem para a melhoria dos processos de concepção estrutural. Outro aspecto de interesse desta tese é a busca de formas otimizadas, por meio da Otimização Topológica. Neste trabalho, um algoritmo alternativo de otimização topológica é proposto. O algoritmo é baseado no acoplamento entre o Método Level Set (MLS) e o MEC. A diferença entre o algoritmo aqui proposto, e os demais presentes na literatura, é forma de obtenção do campo de velocidades. Nesta tese, os campos normais de velocidades são obtidos por meio da sensibilidade à forma. Esta mudança torna o algoritmo propício a ser tratado pelo MEC, pois as informações necessárias para o cálculo das sensibilidades residem exclusivamente no contorno. Verifica-se que o algoritmo necessita de uma extensão particular de velocidades para o domínio a fim de manter a estabilidade. Limitando-se a casos bidimensionais, o algoritmo é capaz de obter os conhecidos casos de referência reportados pela literatura. O último aspecto tratado nesta tese retrata a maneira pela qual as incertezas geométricas podem influenciar na determinação das estruturas otimizadas. Utilizando o MEC, propõe-se um critério probabilístico que permite embasar escolhas levando em consideração a sensibilidade geométrica. Os resultados mostram que os critérios deterministas, nem sempre, conduzem às escolhas mais adequadas sob o ponto de vista de engenharia. Assim, este trabalho contribui para a expansão e difusão das aplicações do MEC em problemas de engenharia de estruturas. / The present work deals with the formulation and implementation of numerical models based on the Boundary Element Method (BEM). Inspired by engineering problems, a multidisciplinary combination is proposed as a more realistic approach. There are common engineering materials that have time-dependent response. In this thesis, time-dependent phenomena are approached through the Linear Viscoelastic Mechanics associated with rheological models. In this work, the formulation of Maxwell\'s constitutive model is presented to be used via MEC. The resultant equations are checked on reference problems. The results show that the presented formulation can be used to represent composite structures, even in cases involving a junction between viscoelastic and non-viscoelastic materials. Additionally the formulations presented remain stable in the presence of cracks. It is found that the classical DUAL-BEM formulation can be used to simulate cracks with time-dependent behaviour. This result serves as the basis for further investigations in the field of Fracture Mechanics of viscoelastic materials. In the sequence, it is shown how the BEM can be associated with probabilistic concepts to make predictions of long-term behaviour. These predictions include the inherent uncertainties in engineering processes. The uncertainties involve the material, loading and geometry parameters. Using the concept of probability of failure, the results show that the uncertainties related to the estimations of loads have important impact on the long-term expected performance. This finding serves to carry out studies that collaborate for the improvement of structural design processes. Another aspect of interest of this thesis is the search for optimized forms through Topological Optimization. In this work, an alternative topological optimization algorithm is proposed. The algorithm is based on the coupling between the Level Set Method (LSM) and BEM. The difference between the algorithm proposed here, and the others present in the literature, is a way of obtaining the velocity field. In this thesis, the normal fields of velocities are obtained by means of shape sensitivity. This change makes the algorithm adequate to be treated by the BEM, since the information necessary for the calculation of the sensitivities resides exclusively in the contour. It is found that the algorithm requires a particular velocity extension in order to maintain stability. Limiting to two-dimensional cases, the algorithm is able to obtain the known benchmark cases reported in the literature. The last aspect addressed in this thesis involves the way in which geometric uncertainties can influence the determination of optimized structures. Using the BEM, it is proposed a probabilistic criterion that takes into consideration the geometric sensitivity. The results show that deterministic criteria do not always lead to the most appropriate choices from an engineering point of view. In summary, this work contributes to the expansion and diffusion of MEC applications in structural engineering problems.

Dans cette thèse, nous calculons la forme d’un objet immergé d’aire donnée qui minimise la résistance de vague. Le corps, considéré lisse, avance à vitesse constante sous la surface libre d’un fluide qui est supposé parfait et incompressible. La résistance de vague est la traînée, c’est-à-dire la composante horizontale de la force exercée par le fluide sur l’obstacle. Nous utilisons les équations de Neumann-Kelvin 2D, qui s’obtiennent en linéarisant les équations d’Euler irrotationnelles avec surface libre. Le problème de Neumann-Kelvin est formulé comme une équation intégrale de frontière basée sur une solution fondamentale qui intègre la condition linéarisée à la surface libre. Nous utilisons une méthode de descente de gradient pour trouver un minimiseur local du problème de résistance de vague. Un gradient par rapport à la forme est calculé par la méthode de variation de frontières. Nous utilisons une approche level-set pour calculer la résistance de vague et gérer les déplacements de la frontière de l’obstacle. Nous obtenons une grande variété de formes optimales selon la profondeur de l’objet et sa vitesse. / In this thesis, we compute the shape of a fully immersed object with a given area which minimises the wave resistance. The smooth body moves at a constant speed under the free surface of a fluid which is assumed to be inviscid and incompressible. The wave resistance is the drag, i.e. the horizontal component of the force exerted by the fluid on the obstacle. We work with the 2D Neumann-Kelvin equations, which are obtained by linearising the irrotational Euler equations with a free surface. The Neumann-Kelvin problem is formulated as a boundary integral equation based on a fundamental solution which handles the linearised free surface condition. We use a gradient descent method to find a local minimiser of the wave resistance problem. A gradient with respect to the shape is calculated by a boundary variation method. We use a level-set approach to calculate the wave-making resistance and to deal with the displacements of the boundary of the obstacle. We obtain a great variety of optimal shapes depending on the depth of the object and its velocity.

O presente trabalho tem por objetivo desenvolver uma formulação numérica baseada em Análise Isogeométrica para o estudo de problemas de interação fluido-estrutura (IFE) em aplicações envolvendo corpos rígidos submersos, onde escoamentos incompressíveis de fluidos Newtonianos com superfície livre são considerados. Propõe-se o emprego da Análise Isogeométrica por permitir a unificação entre os procedimentos de pré-processamento e análise, melhorando assim as condições de continuidade das funções de base empregadas tanto na discretização espacial do problema como na aproximação das variáveis do sistema de equações. O sistema de equações fundamentais do escoamento é formado pelas equações de Navier-Stokes e pela equação da conservação de massa, descrita segundo a hipótese de pseudo-compressibilidade, em uma formulação cinemática ALE (Arbitrary Lagrangean- Eulerian). A consideração da superfície livre no escoamento se dá tratando o fluido como um meio bifásico, através do método Level Set. O corpo rígido apresenta não linearidade na rotação e restrições representadas por vínculos elásticos e amortecedores viscosos, sendo a equação de equilíbrio dinâmico resolvida através do método de Newmark. O esquema de acoplamento sólido-fluido adotado é o particionado convencional, que impõe condições de compatibilidade cinemáticas e de equilíbrio sobre a interface sólido-fluido, analisando ambos os meios de maneira sequencial. A discretização das equações governantes é realizada através do esquema explícito de dois passos de Taylor-Galerkin, aplicado no contexto da Análise Isogeométrica. Por fim, são analisados alguns problemas da Dinâmica de Fluidos Computacional, de onde se concluiu que os resultados obtidos são bastante consistentes com os fenômenos envolvidos, com as ferramentas exclusivas da Análise Isogeométrica, como o refinamento k, melhorando a convergência dos resultados. Para escoamentos bifásicos, verificou-se que o método Level Set obteve resultados bastante promissores apresentando, entretanto, uma dissipação numérica excessiva. Propõe-se, para estudos futuros, a elaboração de esquemas numéricos que conservem melhor o volume da fase líquida do escoamento. / The present work aims to development of a numerical formulation based on Isogeometric Analysis for the study of Fluid-Structure Interaction problems in applications involving rigid bodies submerged, considering incompressible Newtonian flows with free surface. The use of the Isogeometric Analysis allows unification between the preprocessing and analysis steps, improving then the continuity of the base functions employed, both in the spatial discretization and approximation of the variables in the system of equations. The fundamental flow equations are formed by the Navier-Stokes and the mass conservation, described by de pseudo-compressibility hypothesis, in an ALE (Arbitrary Lagrangean-Eulerian) kinematic formulation. The free surface consideration of the flow is handled treating the fluid like a two- phase medium, using the Level Set method. The rigid body considers nonlinearity in rotation, and restrictions represented by elastic springs and viscous dampers, with the dynamic equilibrium equation being resolved using the Newmark’s method. The solid-fluid coupling scheme is the conventional partitioned, which imposes kinematics and equilibrium compatibility conditions on the solid-fluid interface, analyzing both mediums in a sequential manner. The governing equations are discretized using the explicit two step Taylor-Galerkin method, applied in an Isogeometric Analisys context. Finally, some Computational Fluid Dinamics problems are analysed, from which it was concluded that the results obtained are quite consistent with phenomena involved, with the unique tools of Isogeometric Analysis, such as k-refinement, improving the convergence of the results. For biphasic flows, it was verified that the Level Set method obtained very promising results, presenting, however, an excessive numerical dissipation. For future studies, it is proposed the elaboration of numerical schemes that better preserve the volume of the liquid phase of the flow.

La segmentation d'images est un domaine important dans le traitement d'images et un grand nombre d'approches différentes ent été développées pendant ces dernières décennies. L'approche des contours actifs est un des plus populaires. Dans ce cadre, cette thèse vise à développer des algorithmes robustes, qui peuvent segmenter des images avec des inhomogénéités d'intensité. Nous nous concentrons sur l'étude des énergies externes basées région dans le cadre des ensembles de niveaux. Précisément, nous abordons la difficulté de choisir l'échelle de la fenêtre spatiale qui définit la localité. Notre contribution principale est d'avoir proposé une échelle adaptative pour les méthodes de segmentation basées sur les statistiques locales. Nous utilisons l'approche d'Intersection des Intervalles de Confiance pour définir une échelle position-dépendante pour l'estimation des statistiques image. L'échelle est optimale dans le sens où elle donne le meilleur compromis entre le biais et la variance de l'approximation polynomiale locale de l'image observée conditionnellement à la segmentation actuelle. De plus, pour le model de segmentation basé sur une interprétation Bahésienne avec deux noyaux locaux, nous suggérons de considérer leurs valeurs séparément. Notre proposition donne une segmentation plus lisse avec moins de délocalisations que la méthode originale. Des expériences comparatives de notre proposition à d'autres méthodes de segmentation basées sur des statistiques locales sont effectuées. Les résultats quantitatifs réalisés sur des images ultrasonores de simulation, montrent que la méthode proposée est plus robuste au phénomène d'atténuation. Des expériences sur des images réelles montrent également l'utilité de notre approche. / Image segmentation is an important research area in image processing and a large number of different approaches have been developed over the last few decades. The active contour approach is one of the most popular among them. Within this framework, this thesis aims at developing robust algorithms, which can segment images with intensity inhomogeneities. We focus on the study of region-based external energies within the level set framework. We study the use of local image statistics for the design of external energies. Precisely, we address the difficulty of choosing the scale of the spatial window that defines locality. Our main contribution is to propose an adaptive scale for local region-based segmen¬tation methods. We use the Intersection of Confidence Intervals approach to define this pixel-dependent scale for the estimation of local image statistics. The scale is optimal in the sense that it gives the best trade-off between the bias and the variance of a Local Polynomial Approximation of the observed image conditional on the current segmenta¬tion. Additionally, for the segmentation model based on a Bayesian interpretation with two local kernels, we suggest to consider their values separately. Our proposition gives a smoother segmentation with less mis-localisations Chan the original method.Comparative experiments of our method to other local region-based segmentation me¬thods are carried out. The quantitative results, on simulated ultrasound B-mode images, show that the proposed scale selection strategy gives a robust solution to the intensity inhomogeneity artifact of this imaging modality. More general experiments on real images also demonstrate the usefulness of our approach.

Within the general context of solid-state sintering process, this work presents a numerical modeling approach, at the particle scale, of ceramic particle packing consolidation. Typically, the sintering process triggers several mass transport paths that are thermally activated. Among those diffusion paths, the most important ones are: surface diffusion, grain boundary diffusion and volume diffusion. Including this physics into a high-performance computing framework would permit to gain precious insights about the driving mechanisms. The aim of the present work is to develop a model and a numerical strategy able to integrate the different diffusion mechanisms into continuum mechanics framework. In the cases of surface diffusion and volume diffusion, the mass flux is calculated as a function of the surface curvature Laplacian and the hydrostatic pressure gradient, respectively. The physical model describing these two transport mechanisms is first presented within the framework of continuum mechanics. Then the numerical strategy developed for the simulation of the sintering of many particles is detailed. This strategy is based on a discretization of the problem by using a finite element approach coupled with a Level-Set method used to describe the particles free surface. This versatile strategy allows us to perform simulations involving a relatively large number of particles. Furthermore, a mesh adaptation technique allows the particles surface description to be improved, while the number of mesh elements is kept reasonable. Several 3D simulations, performed in a parallel computing framework, show the changes occurring in the structure of 3D granular stacks.

The level set method offers a simple and robust approach to modeling liquid-vapor interfaces that arise in boiling and condensing flows. The current liquid-vapor phase-transition techniques used with the level set method are not able to account for different thermal conductivities and specific heats in each respective phase, nor are they able to accurately account for latent heat absorption and release. This paper presents a new level set based technique for liquid-vapor phase-transition that accounts for different material properties in each respective phase, such as thermal conductivity and specific heat, while maintaining the interface at the saturation temperature. The phase-transition technique is built on the ghost fluid framework coupled with the standard level set method. A new technique is presented for constructing ghost nodes that implicitly captures the immersed boundary conditions and is second order accurate. The method is tested against analytical solutions, and it is used to model film boiling. The new phase-transition technique will greatly assist efforts to accurately capture the physics of boiling and condensing flows. In addition to presenting a new phase transition technique, a coupled level set volume of fluid advection scheme is developed for phase transition flows. The new scheme resolves the mass loss problem associated with the level set method, and the method provides an easy way to accurately calculate the curvature of an interface, which can be difficult with the volume of fluid method. A film boiling simulation is performed to illustrate the superior performance of the coupled level set volume of fluid approach over the level set method and the volume of fluid method.

Ce travail présente une technique de simulation de feux de forêt qui utilise la méthode Level-Set. On utilise une équation aux dérivées partielles pour déformer une surface sur laquelle est imbriqué notre front de flamme. Les bases mathématiques de la méthode Level-set sont présentées. On explique ensuite une méthode de réinitialisation permettant de traiter de manière robuste des données réelles et de diminuer le temps de calcul. On étudie ensuite l’effet de la présence d’obstacles dans le domaine de propagation du feu. Finalement, la question de la recherche du point d’ignition d’un incendie est abordée. / This work presents a forest fire simulation model which uses the Level-Set method. We use a partial differential equation to deform a surface on which our flame front is inscribed. The mathematical foundations of the Level-set method are presented. We then explain a reinitialization method that allows us to treat in a robust way real data and to reduce the calculation time. The effect of the presence of barriers in the fire propagation domain is also studied. Finally, we make an attempt to find the ignition point of a forest fire.

The aim of this research dissertation is at studying natural and mixed convections of fluid flows, and to develop and validate numerical schemes for interface tracking in order to treat incompressible and immiscible fluid flows, later. In a first step, an original numerical method, based on Finite Volume discretizations, is developed for modeling low Mach number flows with large temperature gaps. Three physical applications on air flowing through vertical heated parallel plates were investigated. We showed that the optimum spacing corresponding to the peak heat flux transferred from an array of isothermal parallel plates cooled by mixed convection is smaller than those for natural or forced convections when the pressure drop at the outlet keeps constant. We also proved that mixed convection flows resulting from an imposed flow rate may exhibit unexpected physical solutions; alternative model based on prescribed total pressure at inlet and fixed pressure at outlet sections gives more realistic results. For channels heated by heat flux on one wall only, surface radiation tends to suppress the onset of recirculations at the outlet and to unify the walls temperature. In a second step, the mathematical model coupling the incompressible Navier-Stokes equations and the Level-Set method for interface tracking is derived. Improvements in fluid volume conservation by using high order discretization (ENO-WENO) schemes for the transport equation and variants of the signed distance equation are discussed

L'effet Leidenfrost répresente un cas particulier de caléfaction : lorsqu'une goutte de liquide est déposée sur une surface dont la température est très supérieure à la température d'ébullition du liquide, ce dernier s'évapore avant de toucher la surface et la vapeur ainsi créée forme un coussin sous la goutte qui la maintient en sustentation et l'isole de la plaque chauffante.Ce travail de thèse concerne la modélisation et la simulation de ce phénomène complexe. Dans une première partie, nous étudions un modèle avec interface raide basée sur les équations de Navier-Stokes enrichies avec des termes interfaciaux prenant en compte le changement de phase et la tension de surface. La simulation d'une couche uniforme de liquide sur un film de vapeur nous ramène à un cas unidimensionnel pour lequel on utilise la méthode ALE (Arbitrary Lagrangian Eulerian) afin de gérer la hauteur variable de chaque phase. La discrétisation du modèle est validée sur un cas test.Dans une seconde partie, on utilise la méthode de capture d'interface Level-Set dans laquelle la frontière liquide/gaz est représentée par la ligne de niveau zéro d'une fonction. Cette interface est artificiellement épaissie et les quantités thermodynamiques y sont régularisées. La tension de surface et le changement de phase sont alors introduits sous forme de termes volumiques dans nos équations. L'hypothèse d'incompressibilité de chaque phase pure nous amène alors à un fluide généralisé dont la compressibilité se manifeste uniquement dans la zone interfaciale, là où se produit le changement de phase.La troisième partie est consacrée à la discrétisation de ce modèle pour l'étude tridimensionnelle d'une goutte d'eau, immobile et symétrique par rotation, se ramenant ainsi à un problème bi-dimensionnel axisymétrique. La méthode Level-Set nécessite des choix numériques particuliers qui sont alors explicités : schéma d'advection peu diffusif, redistanciation par résolution d'une équation de Hamilton-Jacobi et correction globale du volume de la goutte, prenant en compte le changement de phase. Un algorithme de projection de type Chorin est également utilisé afin de satisfaire la contrainte sur la compressibilité de notre fluide. On présentera également un nouveau schéma aux différences finies à stencil compact pour l'approximation du gradient.La dernière partie présente et compare nos résultats numériques avec plusieurs courbes théoriques, décrivant chacune l'évolution de certains paramètres de la goutte de liquide : son volume, son rayon et la hauteur de la couche de vapeur. / The Leidenfrost effect is a special case of calefaction: when a liquid is dropped on a surface which temperature is far hotter than the liquid's boiling point, he evaporates before touching the hot plate and the produced gas forms a thin layer, under the liquid droplet, that makes it hovering and isolates it from the heat source.This thesis deals with the modeling and the numerical simulation of this complex phenomenon. In the first part, we study a sharp interface model that supplement the Navier-Stokes equations with interfacial conditions taking into account the phase change and the surface tension. Simulating an uniform liquid layer over a vaporous film reduces the problem to one dimension. The ALE method (Arbitrary Lagrangian Eulerian) is then used to deal with the variable height of each phase. The numerical code is finally validated on a test case.In the second part, the gas/liquid interface is captured by a Level-Set method. The interface is artificially thickened and inner regularization is applied to the thermodynamic quantities. Therefore, surface tension and phase change are approximated by volume terms. Each pure phase is supposed to be incompressible but, due to the phase change, the velocity field is not divergence-free in the interfacial zone.The third part focusses on the discretization of this three-dimensional model, for the simulation of a motionless and rotational symmetric droplet. The problem reduces to an axisymmetric two-dimensional setting. The use of the Level-Set method requires devoted numerical algorithms which are developed: non-diffusive efficient advection scheme, reinitialization by Hamilton-Jacobi equation with global volume correction taking into account the phase change. An adapted Chorin projection algorithm is used to ensure the prescribed compressibility constraint that holds on the interfacial fluid. In addition, we introduce a new finite difference scheme for the gradient approximation that uses a compact stencil.The last part describes and compares our simulation results with several theoretical curves based on different droplet shape simplifications, plotting the evolution of indicators like the volume and radius of the droplet, or the height of the vapor layer.