Several problems in physics and engineering are modeled by reaction-advection-diffusion (RAD) equations. However, when the diffusive terms are small compared with the other ones, these problems can become difficult to solve numerically. Besides, formulating the unsteady version of these models in a semi-discrete fashion, it can be interpreted that the overall diffusivity gets smaller as the time step decreases. To overcome these drawbacks, this thesis considers the development of Galerkin (or Petrov-Galerkin) finite element methods based on approximation spaces enriched by residual-free bubbles (RFB) or multiscale functions. Beginning with the unsteady reaction-diffusion problem, new methods using multiscale functions are presented which improve the solutions in the reaction-dominated regime and/or when small time steps are adopted. They also give rise to a general concept of stabilizing unsteady problems differently along the time. In the following, it is shown that switching RFB by suitable multiscale functions in the elements connected to the outflow boundaries of the domain increases the accuracy of the solutions in this region for RAD problems with advection. Next, this methodology is further studied for systems of RAD equations. In a final contribution, an extension of the RFB method is introduced for the shallow waters equations. All these methods are tested through benchmark problems and compared with stabilized methods presenting stable and accurate results. / A modelagem de vários problemas físicos e de engenharia envolve a solução de problemas de transporte do tipo reação-advecção-difusão (RAD), porém, estes podem tornar-se singularmente perturbados quando os termos difusivos são pequenos comparados aos demais. Além disso, ao adotar formulações semi-discretas em problemas transientes, observa-se que diminuir o passo de tempo tem um efeito de redução da componente difusiva. Para superar estas dificuldades, esta tese considera o desenvolvimento de métodos de elementos finitos de Galerkin (ou Petrov-Galerkin) baseados em espaços de aproximação enriquecidos por funções bolhas livres do resíduo (RFB) ou funções multiescala. Começando pelo problema de reação-difusão transiente, novos métodos utilizando funções multiescala são apresentados, os quais melhoram as soluções no regime reativo-dominante e/ou quando pequenos passos de tempo são adotados. Com estes métodos, discute-se também o conceito de estabilização variável ao longo do tempo para problemas transientes. Na seqüência, verifica-se que utilizar funções multiescala nos elementos conectados às fronteiras de saída de fluxo do domínio e RFB nos demais elementos aumenta a precisão das soluções nesta região em problemas de RAD com advecção dominante. A seguir, esta metodologia é estudada para sistemas de RAD. Como contribuição final, estende-se o método RFB para o modelo de águas rasas. Todos estes métodos são submetidos a testes de robustez e comparados com métodos estabilizados, apresentando resultados estáveis e precisos.

This work aims at developing computational models capable of furnishing more realistic and less costly computationally for the problem of electrokinetic remediation of polluted clayey soils. Innovative results are obtained by improving the multiscale models previously developed by Lima and co-workers through the construction of perturbations of the local microscopic problems in conjuction with more realistic boundary conditions at the electrodes and with the development of precise estimates for the assymptotic behavior of the macroscopic solution. Considering the aliance of such techniques within the framework of the homogenization method of periodic structures we discretize the macroscopic model by the finite element method numerical simulations of an electroosmose experiment capable of predicting more realistic scenarios of electrokinetic remediation. / Este trabalho objetiva o desenvolvimento de modelos computacionais capazes de construir simulações numéricas mais realistas e menos custosas computacionalmente para o problema de descontaminação de solos argilosos por técnicas de eletrocinética. Resultados inovadores são obtidos aprimorando-se os modelos multi-escala desenvolvidos anteriormente por Lima e colaboradores via construção de soluções perturbativas dos problemas locais microscópicos aliada à condições de contorno mais realistas nos eletrodos e ao desenvolvimento de estimativas precisas para o comportamento assintótico da solução macroscópica. Por intermédio da conjunção destas técnicas imersas no contexto da teoria de homogeneização de estruturas periódicas discretizamos o modelo macroscópico pelo método dos elementos finitos e construimos simulações numéricas de um experimento de eletroosmose capazes de predizer cenários mais realistas em eletrorremediação de solos.

Surfactants are amphiphilic molecules which self-assemble at the interface in oil-water-surfactant mixtures such that the hydrophobic part, called tail, stays in oil and the remaining part, called head, resides in hydrophilic en-vironment. Depending upon concentration of individual components, these mixtures form several microphases, such as bilayers, micelles, columnar and lamellar phases. A lamellar phase, at equilibrium, is made up of alternat-ing layers of water and oil separated by surfactants, or of alternate layers of water and surfactant bilayers such that the hydrophilic heads are in contact with water. This equilibrium state is rarely achieved in macroscopic samples due to thermodynamic and kinetic constraints; instead, a lamellar fluid is usually disordered with a large number of defects. These defects have significant effect on the flow behaviour of the lamellar mesophase systems. They are known to alter the flow field, resulting stresses and in turn could get distorted or annihilated by the flow. In present work, we analyse this two way coupling between lamellar structure and flow field. The structural and rheological evolution of an initially disordered lamellar phase system under a shear flow is examined using a mesoscale model based on a free energy functional for the concentration field, which is the scaled difference in the concentration between the hydrophilic and hydrophobic components. Two distinct modes of structural evolution are observed depending only on Peclet number, which ratio of inertial forces to mass diffusivity, in-dependent of system size. At low Peclet number, local domains are formed which are then rotated and stretched by shear. A balance between defect creation and annihilation is reached due to which the system never reaches the equilibrium layer configuration. In the opposite limit, partially formed layers break and reform so as to form a nearly aligned lamellar phase con-figuration with residual defects. Viscosity of lamellar phase system increases with layer moduli, differences in viscosity of individual components, fluidity of the lamellae due to shear banding and defect pinning. These factors however, do not have any effect on alignment mechanism.

Numerical solution of differential equations having multitude of scales in the solution field is one of the most challenging research areas, but highly demanded in scientific and industrial applications. One of the natural approaches for handling such problems is to separate the scales and approximate the solution of the segregated scales with appropriate numerical method. Variational multiscale method (VMS) is a predominant method in the paradigm of scale separation schemes. In our work we have used the VMS technique to develop a numerical scheme for computations of turbulent flows in time-dependent domains. VMS allows separation of the entire range of scales in the flow field into two or three groups, thereby enabling a different numerical treatment for the different groups. In the context of computational fluid dynamics(CFD), VMS is a significant new improvement over the classical large eddy simulation (LES). VMS does away with the commutation errors arising due to filtering in LES. Further, in a three-scale VMS approach the model for the subgrid scale can be contained to only a part of the resolved scales instead of effecting the entire range of resolved scales. The projection based VMS scheme that we have developed gives a robust and efficient method for solving problems of turbulent fluid flows in deforming domains, governed by incompressible Navier {Stokes equations. In addition to the existing challenges due to turbulence, the computational complexity of the problem increases further when the considered domain is time-dependent. In this work, we have used an arbitrary Lagrangian-Eulerian (ALE) based VMS scheme to account for the domain deformation. In the proposed scheme, the large scales are represented by an additional tensor valued space. The resolved large and small scales are computed in a single unified equation, and the effect of unresolved scales is confined only to the resolved small scales, by using a projection operator. The popular Smagorinsky eddy viscosity model is used to approximate the effects of unresolved scales. The used ALE approach consists of an elastic mesh update technique. Moreover, a computationally efficient scheme is obtained by the choice of orthogonal finite element basis function for the resolved large scales, which allows to reformulate the ALE-VMS system matrix into the standard form of the NSE system matrix. Thus, any existing Navier{Stokes solver can be utilized for this scheme, with modifications. Further, the stability and error estimates of the scheme using a linear model of the NSE are also derived. Finally, the proposed scheme has been validated by a number of numerical examples over a wide range of problems.

No description available.

L'objectif de cette thèse est de mettre en œuvre une description multi-échelle adaptée aux écoulements de fluides dans des micro-/nano-conduites. Cette approche doit permettre de décrire, aussi bien les petites échelles relatives aux interactions du fluide avec les atomes du mur, que les grandes échelles de l’écoulement engendrées par les conditions aux limites d'entrée/sortie du canal. Pour cela, nous avons développé une méthode qui couple une modélisation continue des écoulements et des transferts de chaleur dans le cœur du canal avec une modélisation discrète proche des parois, basée sur une représentation atomistique du fluide et du mur.Les équations de Navier-Stokes et de l’énergie, couplées à une équation d’état, sont approximées par une méthode de volumes finis dans le cœur de l’écoulement alors que des simulations de dynamique moléculaire sont utilisées pour représenter finement les interactions entre le fluide et la paroi. Cette approche hybride nécessite la transmission d’informations entre les modélisations : les grandeurs moyennées moléculaires sont imposées comme conditions aux limites pour le modèle continu, et la dynamique sous contrainte, couplée à un thermostat de Langevin, est utilisée pour piloter l’échelle moléculaire. Une représentation par des plots moléculaires locaux de petite taille, intelligemment répartis le long de l’interface entre le fluide et le mur, permet de traiter des écoulements et des transferts dans des canaux de très grands allongements, pour des coûts de calcul raisonnables.Après une partie de validation, des simulations hybrides multi-échelles d’écoulements dans des canaux constitués de parois en platine ont été menées pour de l’argon en phase liquide (incompressible) ou gazeuse (compressible), en tenant compte éventuellement du changement de phase au voisinage de la paroi / The main objective of this thesis is to model the multi-scale heat and fluid flows in micro-/nano channels. This method must be able of capturing at the same time the fluid/solid interaction at the small scale but also the flows induced by the inlet/outlet boundary conditions at the large scale. To this aim, we have adopted an approach coupling the continuum model in the bulks of the channel and the discrete model at the vicinity of the wall, based on an atomistic representation of the fluid and the solid.The Navier-Stokes and energy equations, coupled with an equation of state, are approximated by a finite volume method and the molecular dynamics simulations are used to finely represent the interaction between the fluid and the solid. This hybrid method requires information transmission between the former two regions: averaged quantity in molecular dynamics simulations are imposed as boundary conditions for the continuous model and constrained dynamics, coupled with a thermostat Langevin, is used to control in the molecular level. A set of small molecular dynamics blocks, smartly distributed all along the wall/fluid interface, allows to treat flow and heat transfers in a long micro/nano-channel with a reasonable computational cost.After a validation step, the hybrid multi-scale simulations of complex fluid flows in the channel composed of the platinum wall have been conducted for argon in incompressible liquid or compressible gaseous phase with and without phase change in the vicinity of the wall

L’objectif de ce travail est de qualifier l’utilisation du collage structural en remplacement des solutions d’assemblage conventionnel dans les applications lanceur spatiaux. Si les gains potentiels tant en terme opérationnel qu’en terme de performance ne sont plus à démontrer il reste à mettre en place une démarche de dimensionnement qui garantisse le niveau important de fiabilité requis. Deux adhésifs structuraux de types époxydes ont été étudiés : l’EA9394 et le SW2216. Un travail important d’analyse physico-chimique des surfaces a été mené pour mettre en place un procédé d’assemblage stable et performant. L’analyse mécanique des liaisons collées a montré que l’adhésif subit en situation des sollicitations combinées de traction et de cisaillement. Un dispositif dérivé des essais Arcan à été proposé qui permet d’obtenir la réponse du film de colle sous forme de courbe contrainte déplacement relatif des substrats. La résistance mécanique mesurée au moyen des essais Arcan mais également sur le matériau adhésif montre une forte dispersion liée à un taux de porosité important mais également à une microstructure très hétérogène constituée d’une matrice époxyde et mélangée à des particules d’aluminium (adhésif EA9394). Enfin, le collage a été étudié à une échelle plus macroscopique en tentant de proposer un dimensionnement de liaison pour le lanceur démonstrateur Arcadia. Les lanceurs étant des structures à symétrie de révolution, un modèle semi-analytique de liaison collée axisymétrique a été développé.Ce travail a permis la mise en place d’une démarche d’analyse mettant en avant les analyses à l’échelle du film, de la liaison et de la structure pour mettre en évidence les nombreux phénomènes à l’origine d’une ruine prématurée de la liaison. Cette optimisation multiéchelle doit être conduite pour maximiser la résistance spécifique de la liaison en assurant le niveau de fiabilité requis pour cette application spatiale. / The objective of this study is to qualify the use of structural bonding as a replacement for conventional assembly solutions in space launcher applications. Potential benefits in fabrication and of final performance are already known, but it remains to establish a design approach to guarantee the high level of reliability required. Two epoxy structural adhesives have been studied: EA9394 and SW2216. A study of the physico-chemistry of surfaces has been effected in order to find a stable and strong bonding process. Mechanical analysis of bonded connections revealed that the adhesives could be subjected in use to mixed loadings in tension and shear. A device derived from Arcan tests has been proposed in which it is possible to obtain adhesive layer responses as stress/relative deformation for different substrates. Mechanical strength measured with the Arcan device and also on the bulk adhesive material revealed a high variability coming from a significant porosity fraction but also due to a highly heterogeneous microstructure constituted of epoxy matrix and aluminium particles (EA9394adhesive). Bonding has been studied on a macroscopic scale proposing connection design for the demonstration Arcadia launcher. A rocket being an axially symmetrical structure, a semi-analytical model of axisymmetric bonded connections has been developed. This study resulted in the finalisation of an analysis approach pointing to scale analysis of the adhesive layer, connection and structure in order to emphasise the many phenomena which could cause early bond failure. This multiscale approach should be effected to increase specific assembly strength while assuming the required reliability level for this spatial application.

Une méthode de couplage continu-discret a été développée pour simuler les mécanismes complexes d'endommagement de la silice soumise à un choc laser de haute puissance. Dans un premier temps, une classification des méthodes numériques existantes a été faite pour choisir celles les mieux adaptées à la simulation du comportement sous choc de la silice. Comme résultat de cette classification, deux méthodes ont été retenues: la méthode des éléments discrets (DEM) et la méthode des éléments naturels contraints (CNEM). Ces méthodes sont alors couplées en se basant sur la technique dite "Arlequin". Puis, un modèle numérique permettant de tenir compte des différents phénomènes qui caractérise le comportement de la silice sous haute pression a été développé. Pour bien caractériser les mécanismes de fissuration de la silice à l’échelle microscopique, un nouveau modèle de rupture a été développé dans ce travail. Finalement, ces deux modèles, modèle de comportement et modèle de rupture, ont été intégrés dans la méthode du couplage pour simuler d'un point de vue mécanique le choc laser sur un échantillon en silice. / A discrete-continuum coupling approach has been developed to simulate the laser-induced damage in silica glass. First, a classification of the different numerical methods has been performed to select the ones that best meet the objectives of this work. Acting upon this classification, the Discrete Element Method (DEM) and the Constrained Natural Element Method (CNEM) have been retained. Subsequently, a coupling approach between these methods has been proposed. This approach is based on the Arlequin technique. In the second part, a numerical model of the silica glass mechanical behavior has been developed to better characterize the silica glass response under highly dynamic loadings and particularly loading generated by a laser beam. To correctly characterize the silica glass cracking mechanisms, a new fracture model has been proposed in this work. Finally, all these developments have been used to simulate the laser-induced damage in silica glass.

Les parties tournantes des turbomachines aéronautiques sont composées d’une succession de roues aubagées qui permettent le transfert de l’énergie entre l’air et le rotor. Ces roues aubagées constituent des pièces particulièrement sensibles car elles doivent répondre en termes de dimensionnement à des impératifs de performances aérodynamiques, d’aéroacoustique et de tenue mécanique à la rotation,à la température et à la charge aérodynamique. Le contact avec frottement existant au niveau des attaches aube-disque joue un rôle important sur les niveaux vibratoires.Ce travail porte sur l’étude de l’usure par fretting sous chargements dynamiques dans les interfaces frottantes. En effet, les vibrations de l’aube peuvent produire des micro-glissements en pied d’aubequi peuvent entraîner un phénomène d’usure par fretting. Les connaissances sur le comportement de l’usure sous sollicitations dynamiques sont faibles. Seuls existent des outils numériques pour modéliser l’usure dans le cas de sollicitations quasi-statiques. Nous proposons dans cette thèse des méthodes pour calculer l’évolution de l’usure au cours des cycles de chargement dynamique basées sur une approche multi-échelle en temps. La réponse vibratoire de la structure est liée à une échelle de temps rapide qui est calculée par une méthode d’équilibrage harmonique, dans laquelle les déplacements et les efforts sont projetés sur la base de Fourier. Différentes approches temps-fréquence de calcul des coefficients de Fourier des forces de contact sont présentées. La cinétique d’usure est liée à une échelle lente et différentes méthodes sont proposées pour l’intégrer. La prise en compte des géométries usées dans le modèle éléments finis se fait par l’ajout d’un vecteur des profondeurs d’usure dans le terme de pénalité des lagrangiens dynamiques. Des exemples académiques valident et illustrent les méthodes proposées. Ces méthodes sont ensuite appliquées à l’étude de l’usure par fretting en pied d’aube de soufflante. L’étude numérique met en lumière le couplage entre vibration et usure par fretting aux interfaces de contact. La modification du comportement dynamique global de la roue aubagée est aussi observée. / The rotating parts of aeronautical turbomachineries are made of bladed disks which enable the transfer of energy from the air to the rotor. These bladed disks are especially critical parts because their dimensioning has to meet strict requirements in terms of aerodynamical performance, aeroacoustics and mechanical resistance to rotation, temperature and aerodynamical loads. The frictional contact at the interface between blade and disk has an important influence on the vibratory levels.This work deals with the study of fretting-wear in frictional interfaces under dynamical loading. Indeed,the blade’s vibrations can produce micro-slidings in blade’s root which may entail fretting-wear. Wear under dynamical loading is a badly known phenomenon. Numerical tools exist for quasic-static conditions only. Here,methods are proposed to quantify the evolution of wear along dynamical loadingcycles based on a time-multiscale approach. The vibratory response of the structure is linked with a fasttime scale which is calculated by a harmonic balance method : displacements and forces are expressed through Fourier series. Different frequency-time approaches are presented to compute the Fourier coefficients of contact forces. Wear kinetics is linked with a slow time scale and different methods are proposed to integrate it.Worn geometries are taken into account in the finite elements model by a wear depth vector included in the penalty term of dynamic lagrangians. Academic examples validate and illustrate the proposed methods. These methods are then used to study fretting-wear in a fan’s bladeroot. The numerical results highlight the coupling between vibration and fretting-wear in frictional interfaces.The modification of the global dynamical behaviour of the bladed disk is also observed.

L'industrie se tourne de plus en plus vers les matériaux composites. A l'échelle de la microstructure leur comportement est fortement hétérogène mais à l'échelle de la structure ceux-ci peuvent être considérés homogène. Les méthodes multi-échelles ont été développées pour résoudre les problèmes de structure avec un temps de calcul raisonnable. Ces méthodes sont validées par comparaison avec un calcul numérique où les hétérogénéités sont entièrement maillées. Dans ce travail de thèse, une structure architecturée modèle a été créée au centre d'une plaque (homogène) mince en acier inoxydable (304L). La cellule unitaire du matériau architecturé est constitué d'un carré avec un trou au centre. L'utilisation d'une caméra à très haute résolution (270 millions de pixels) permet de suivre simultanément l'évolution des déformations aux échelles microscopique et macroscopique. La variation de l'orientation de la structure architecturée modifie les sollicitations appliquées aux cellules unitaires. Les expériences réalisées ont pour but d'analyser les cinématiques de déformation des cellules unitaires sous un chargement multi-axial. La recherche des cellules ayant une cinématique périodique est réalisée. Il est ainsi montré que les cellules avec une cinématique non périodique correspondent à la zone de transition entre le matériau architecturé et le matériau homogène. La connaissance des cinématiques des cellules permet d'investiguer les changements d'échelles dans le domaine linéaire et non-linéaire. Le passage de l'échelle macroscopique à l'échelle microscopique est particulièrement étudié avec le choix des conditions aux limites. Le remplacement des cellules ayant une cinématique périodique par un milieu homogène équivalent (MHE) est traité. La géométrie de la cellule unitaire introduit des symétries dans le comportement du MHE, celui-ci devient cubique. Les caractéristiques élastiques du MHE sont obtenus par homogénéisation à partir des résultats expérimentaux. Un critère de Tsaï-Hill est identifié dans le domaine non-linéaire. Le dernier chapitre s'intéresse à la fissuration de la zone architecturée et à l'initiation de la localisation des déformations dans les cellules. Le support de la localisation est calculé à partir du champ des déformations mesuré par CIN. La cinématique de la cellule est enrichie avec une discontinuité et le saut de déplacement normal à la localisation est identifié. Une comparaison avec le saut de déplacement calculé par corrélation d'images étendue à l'échelle macroscopique est menée afin de valider la stratégie d'identification à l'échelle microscopique. / Industry employ more and more composite materials in structures todecrease the weight. At the microstructure scale behavior is strongly heterogeneous but at the structure scale behaviour may be considered homogeneous. Multiscale methods have been developed to solve the structural problems with a reasonable calculation time. These methods are validated by comparison with a numerical calculation where heterogeneities are fully meshed. In this thesis work, an ideal architectural material was created in the center of a (homogeneous) stainless steel (304L) thin plate. The unit cell architecture material consists of a square with a hole in the center. The use of a high resolution camera (270 million pixels) allows to follow simultaneously the evolution of deformation at microscopic and macroscopic scales. The orientation of the heterogeneous structure modifies the sollicitations applied to the unit cells. The experiments are designed to analyze the kinematics of deformation of the unit cells in a multi-axial loading. Unit cells with periodic kinematics are searched. It is thus shown that the cells with a non-periodic kinematic correspond to the transition zone between the homogeneous material and the architectured material. Knowledge of the kinematic cells allows to investigate the scale changings in the linear and nonlinear range. The downscaling from the macroscopic to the microscopic scale is particularly studied with the choice of boundary conditions. An equivalent homogeneous medium (MHE) is determined as a remplacement for the cells having a periodic kinematic. The geometry of the unit cell introduced symmetries in the behavior of MHE, it becomes cubic (orthotropic with material parameters). The elastic characteristics of the MHE are obtained by homogenization from the experimental results. A criterion of Tsai-Hill is identified in the non-linear domain. The last chapter is interested in cracking of the architected zone and the initiation of strain localization in cells. The support of location is calculated from the strain field measured by correlation. The kinematics of the cell is enriched with a discontinuity and the displacement jump normal to the localization is identified. A comparison with the displacement jump calculated by extended digital image correlation at the macroscopic scale is conducted to validate the identification strategy at the microscopic level.