Un des outils numériques les plus utilisés en ingénierie est la méthode des éléments finis, qui peut être mise en o euvre grâce à l'utilisation de nombreux codes de calcul. Toutefois, une difficulté apparaît lors de l'utilisation de la méthode des éléments finis, spécialement en géotechnique, lorsque la structure étudiée est en interaction avec un domaine de dimensions infinies. L'usage courant en ingénierie est alors de réaliser les calculs sur des domaines bornés, mais la définition de la frontière de tels domaines bornés pose de sérieux problèmes. Pour traiter convenablement les problèmes comportant des frontières à l'infini, l'utilisation d'éléments discrets "infinis" est maintenant souvent délaissée au profit de la méthode des équations intégrales ou "méthode des éléments de frontière" qui permet de résoudre un système d'équations aux dérivées partielles linéaire dans un domaine infini en ne maillant que la frontière du domaine à distance finie. La mise en oeuvre du couplage entre la méthode des éléments finis et la méthode des éléments de frontière apparaît donc comme particulièrement intéressante car elle permet de bénéficier de la flexibilité des codes de calcul par éléments finis tout en permettant de représenter les domaines infinis à l'aide de la méthode des éléments de frontière. La méthode est basée sur la construction de la "matrice de raideur" du domaine infini grâce à l'utilisation de la méthode des équations intégrales. Il suffit alors d'assembler la matrice de raideur du domaine infini avec la matrice de raideur du domaine fini représenté par éléments finis. L'utilisation de la méthode la plus simple de traitement des équations intégrales, dite méthode de « collocation » conduit à une matrice de raideur non-symétrique. Par ailleurs, la méthode dite «Singular Galerkin» conduit à une formulation symétrique, mais au prix du calcul d'intégrales hypersingulières. La thèse porte sur une nouvelle formulation permettant d'obtenir une matrice de raideur symétrique sans intégrales hypersingulières, dans le cas de problèmes plans. Quelques applications numériques sont abordées pour des problèmes courants rencontrés en géotechnique / One of the most used numerical tools in engineering is the finite element method, which can be implemented through the use of many computer codes. However, a difficulty arises when using the finite element method, especially in geotechnical engineering, where the structure is studied in interaction with a field of infinite dimensions. The commonly used in engineering is then performming the calculations on bounded domains, but the definition of the border of the domain also poses serious problems. To properly solve the problems which have the boundary at infinity, the use of discrete elements "infinite" is now often neglected in favor of the integral equations method or "boundary element method", which allows to solve a linear partial differential equations system in an infinite domain by the discretization of the only boundary of the domain at finite distance. The implementation of coupling between the finite element method and boundary element method is therefore particularly interesting because it allows to benefit the flexibility of computer codes by the finite element method, while the infinite domains is represented by the help of the integral equations method. It is sufficient to assemble the stiffness matrix of infinite domain with the stiffness matrix of finite domain represented by finite elements. Using the simplest method of treatment of integral equations, known as method of "collocation" leads to a non-symmetric stiffness matrix. Furthermore, a method known “Galerkin Singular” leads to a symmetric formulation, but it is at the cost of computing hypersingular integrals. The thesis focuses on a new formulation to obtain a symmetric stiffness matrix without full hypersingular, in the case of plane problems. Some numerical applications are discussed for common problems encountered in geotechnical engineering

Cette thèse considère la solution numérique d'un problème aux valeurs propres de grandes dimensions, dans lequel l'opérateur est dérivé d'un problème de transfert radiatif. Ainsi, cette thèse étudie l'utilisation de matrices hiérarchiques, une représentation efficace de tableaux, très intéressante pour une utilisation avec des problèmes de grandes dimensions. Les matrices sont des représentations hiérarchiques de structures de données efficaces pour les matrices denses, l'idée de base étant la division d'une matrice en une hiérarchie de blocs et l´approximation de certains blocs par une matrice de petite caractéristique. Son utilisation permet de diminuer la mémoire nécessaire tout en réduisant les coûts informatiques. L'application de l'utilisation de matrices hiérarchique est analysée dans le contexte de la solution numérique d'un problème aux valeurs propres de grandes dimensions résultant de la discrétisation d'un opérateur intégral. L'opérateur est de convolution et est défini par la première fonction exponentielle intégrale, donc faiblement singulière. Pour le calcul informatique, nous avons accès à HLIB (Hierarchical matrices LIBrary) qui fournit des routines pour la construction de la structure hiérarchique des matrices et des algorithmes pour les opérations approximative avec ces matrices. Nous incorporons certaines routines comme la multiplication matrice-vecteur ou la decomposition LU, en SLEPc (Hierarchical matrices LIBrary) pour explorer les algorithmes existants afin de résoudre les problèmes de valeur propre. Nous développons aussi des expressions analytiques pour l'approximation des noyaux dégénérés utilisés dans la thèse et déduire ainsi les limites supérieures d'erreur pour ces approximations. Les résultats numériques obtenus avec d'autres techniques pour résoudre le problème en question sont utilisés pour la comparaison avec ceux obtenus avec la nouvelle technique, illustrant l'efficacité de ce dernier / In this thesis, we consider the numerical solution of a large eigenvalue problem in which the integral operator comes from a radiative transfer problem. It is considered the use of hierarchical matrices, an efficient data-sparse representation of matrices, especially useful for large dimensional problems. It consists on low-rank subblocks leading to low memory requirements as well as cheap computational costs. We discuss the use of the hierarchical matrix technique in the numerical solution of a large scale eigenvalue problem arising from a finite rank discretization of an integral operator. The operator is of convolution type, it is defined through the first exponential-integral function and hence it is weakly singular. We access HLIB (Hierarchical matrices LIBrary) that provides, among others, routines for the construction of hierarchical matrix structures and arithmetic algorithms to perform approximative matrix operations. Moreover, it is incorporated the matrix-vector multiply routines from HLIB, as well as LU factorization for preconditioning, into SLEPc (Scalable Library for Eigenvalue Problem Computations) in order to exploit the available algorithms to solve eigenvalue problems. It is also developed analytical expressions for the approximate degenerate kernels and deducted error upper bounds for these approximations. The numerical results obtained with other approaches to solve the problem are used to compare with the ones obtained with this technique, illustrating the efficiency of the techniques developed and implemented in this work

<p> In this thesis, both direct and inverse elastic scattering problems are considered. For a given incident wave, the direct problem is to determine the displacement of wave field from the known structure, which could be an obstacle or a surface in this thesis; The inverse problem is to determine the structure from the measurement of displacement on an artificial boundary.</p><p>In the second chapter, we consider the scattering of an elastic plane wave by a rigid obstacle, which is immersed in a homogeneous and isotropic elastic medium in two dimensions. Based on a Dirichlet-to-Neumann (DtN) operator, an exact transparent boundary condition is introduced and the scattering problem is formulated as a boundary value problem of the elastic wave equation in a bounded domain. By developing a new duality argument, an a posteriori error estimate is derived for the discrete problem by using the finite element method with the truncated DtN operator. The a posteriori error estimate consists of the finite element approximation error and the truncation error of the DtN operator which decays exponentially with respect to the truncation parameter. An adaptive finite element algorithm is proposed to solve the elastic obstacle scattering problem, where the truncation parameter is determined through the truncation error and the mesh elements for local refinements are chosen through the finite element discretization error.<br></p><p>In chapter 3, we extend the argument developed in chapter 2 to elastic surface grating problem, where the surface is assumed to be periodic and elastic rigid; Then, we treat the obstacle scattering in three dimensional space; The direct problem is shown to have a unique weak solution by examining its variational formulation. The domain derivative is studied and a frequency continuation method is developed for the inverse problem. Finally, in chapter 4, a rigorous mathematical model and an efficient computational method are proposed to solve the inverse elastic surface scattering problem which arises from the near-field imaging of periodic structures. The surface is assumed to be a small and smooth perturbation of an elastically rigid plane. By placing a rectangle slab of a homogeneous and isotropic elastic medium with larger mass density above the surface, more propagating wave modes can be utilized from the far-field data which contributes to the reconstruction resolution. Requiring only a single illumination, the method begins with the far-to-near field data conversion and utilized the transformed field expansion to derive an analytic solution for the direct problem, which leads to an explicit inversion formula for the inverse problem; Moreover, a nonlinear correction scheme is developed to improve the accuracy of the reconstruction; Numerical examples are presented to demonstrate the effectiveness of the proposed methods for solving the questions mentioned above.<br></p>

The thesis presents a comprehensive study of a posteriori error estimation in the adaptive solution to some classical elliptic partial differential equations. Several new error estimators are proposed for diffusion problems with discontinuous coefficients and for convection-reaction-diffusion problems with dominated convection/reaction. The robustness of the new estimators is justified theoretically. Extensive numerical results demonstrate the robustness of the new estimators for challenging problems and indicate that, compared to the well-known residual-type estimators, the new estimators are much more accurate.

Le développement d’aéronefs « plus » voire « tout » électriques a pour conséquence la conception d’architectures électriques embarquées de plus en plus complexes entraînant une très nette augmentation du nombre de câbles électriques à déployer au sein des véhicules. Parmi les contraintes rencontrées lors des phases de définition et d’intégration des réseaux de câblages, les aspects de compatibilité électromagnétique et de gestion des échauffements thermiques deviennent de plus en plus critiques. Ainsi, ces travaux de thèse sont dédiés au développement d’une méthodologie permettant la prédiction d’une part, des courants induits par et sur les réseaux de câblages et d’autre part, de leur niveaux d’échauffement. En particulier, l’analyse bibliographique effectuée à cette occasion montre que les phénomènes électrostatiques (à la base de la théorie des lignes de transmission) et de conduction stationnaire de la chaleur sont strictement analogues, ce qui autorise une résolution simultanée de ces deux problèmes pour les réseaux de câblages considérés. Les présents travaux démontrent que le calcul des paramètres électriques primaires (p.u.l) du réseau et de la distribution de température dans le plan transverse peut se faire de manière totalement couplé à l’aide d’un schéma numérique basé sur la Méthode des Moments(MoM). Le choix de l’utilisation des équations intégrales pour la résolution de ce problème de potentiel se fonde sur plusieurs avantages tels qu’une utilisation optimisée des ressources de calcul et l’utilisation d’algorithmes efficaces de résolution, de surcroît naturellement parallélisables pour de futurs développements. Les outils de calculs thermiques développés dans le cadre de cette thèse, et qui ont vocation à être intégrés dans la suite logicielle CRIPTE de l’ONERA, ont fait l’objet d’une validation expérimentale pour plusieurs configurations de harnais électriques. Les comparaisons simulations-mesures présentent de bons accords bien que les expérimentations menées aient montré la difficulté d’obtenir précisément des valeurs du coefficient d’échange thermique,même dans des conditions parfaitement maîtrisées. Les travaux ouvrent enfin des perspectives nouvelles sur l’optimisation en terme de masse des réseaux de câblage (EWIS). / The on-going development of “more” or “all” electrical aircraft leads to the design of ever-complex embeddedelectrical networks, which causes a significant increase of electrical cables to be used within these innovativevehicles. Among the constraints encountered during the definition and integration phases of the network, thoserelated to the electromagnetic compatibility between equipment as well as the management of thermal heatingby Joule’s effect become more and more stringent. Thus, this thesis is dedicated to the development of anoriginal methodology enabling the prediction of both induced and crosstalk currents as well as the heating upstate in complex bundles of cables. Indeed, literature review explicitly shows that electrostatic and stationaryheat transfer phenomena are, from a mathematical standpoint, strictly the same which allows the simultaneouscomputation of these two problems for an arbitrary network. This research work demonstrates that the determinationof primary electrical parameters (per unit length) and the temperature distribution within a givencross-section can be handled with the numerical Method Of Moment (MoM). This choice is motivated by theseveral inherent advantages of the method like an optimized use of the computer resource and the naturalparallelization of the algorithms. The developed numerical tools, intended to be fully integrated in the in-housesoftware suite CRIPTE, has been validated during an experimental campaign that has involved several typesof bundles. Although the comparisons between experimental and simulated results comply with each other,experiments reveal the hard task of getting a precise estimation of the heat transfer coefficients, even in awell-controlled environment. Finally, these works open new and very promising perspectives for future EWIS(Electrical Wiring Interconnection System) in term of mass optimization.

Dans cette thèse, nous proposons de nouvelles méthodes permettant de résoudre numériquement des problèmes de diffraction harmoniques et tridimensionnels, aussi bien acoustiques qu'électromagnétiques, pour lesquels l'objet diffractant est pénétrable et inhomogène. La résolution de tels problèmes est centrale pour des calculs de surfaces équivalentes sonar et radar (SES et SER). Elle est toutefois connue pour être difficile car elle requiert de discrétiser des équations aux dérivées partielles posées dans un domaine extérieur. Étant infini, ce domaine ne peut pas être maillé en vue d'une résolution par la méthode des éléments finis volumiques. Deux approches classiques permettent de contourner cette difficulté. La première consiste à tronquer le domaine extérieur et rend alors possible une résolution par la méthode des éléments finis volumiques. Étant donné qu'elles approximent les problèmes de diffraction au niveau continu, les méthodes de troncature de domaine peuvent toutefois manquer de précision pour des calculs de SES et de SER. Les problèmes de diffraction harmoniques, pénétrables et inhomogènes peuvent également être résolus en couplant une formulation variationnelle volumique associée à l'objet diffractant et des équations intégrales surfaciques rattachées au domaine extérieur. Nous parlons de couplages FEM-BEM (Finite Element Method-Boundary Element Method). L'intérêt de cette approche réside dans le fait qu'elle est exacte au niveau continu. Les couplages FEM-BEM classiques sont dits forts car ils couplent la formulation variationnelle volumique et les équations intégrales surfaciques au sein d'une même formulation. Ils ne sont toutefois pas adaptés à la résolution de problèmes à haute fréquence. Pour pallier cette limitation, d'autres couplages FEM-BEM, dits faibles, ont été proposés. Ils correspondent concrètement à des algorithmes de décomposition de domaine itérant entre l'objet diffractant et le domaine extérieur. Dans cette thèse, nous introduisons de nouveaux couplages faibles FEM-BEM acoustiques et électromagnétiques basés sur des approximations de Padé récemment développées pour les opérateurs Dirichlet-to-Neumann et Magnetic-to-Electric. Le nombre d'itérations nécessaires à la résolution de ces couplages ne dépend que faiblement de la fréquence et du raffinement du maillage. Les couplages faibles FEM-BEM que nous proposons sont donc adaptés pour des calculs précis de SES et de SER à haute fréquence / In this doctoral dissertation, we propose new methods for solving acoustic and electromagnetic three-dimensional harmonic scattering problems for which the scatterer is penetrable and inhomogeneous. The resolution of such problems is key in the computation of sonar and radar cross sections (SCS and RCS). However, this task is known to be difficult because it requires discretizing partial differential equations set in an exterior domain. Being unbounded, this domain cannot be meshed thus hindering a volume finite element resolution. There are two standard approaches to overcome this difficulty. The first one consists in truncating the exterior domain and renders possible a volume finite element resolution. Given that they approximate the scattering problems at the continuous level, truncation methods may however not be accurate enough for SCS and RCS computations. Inhomogeneous penetrable harmonic scattering problems can also be solved by coupling a volume variational formulation associated with the scatterer and surface integral equations related to the exterior domain. This approach is known as FEM-BEM coupling (Finite Element Method-Boundary Element Method). It is of great interest because it is exact at the continuous level. Classical FEM-BEM couplings are qualified as strong because they couple the volume variational formulation and the surface integral equations within one unique formulation. They are however not suited for solving high-frequency problems. To remedy this drawback, other FEM-BEM couplings, said to be weak, have been proposed. These couplings are actually domain decomposition algorithms iterating between the scatterer and the exterior domain. In this thesis, we introduce new acoustic and electromagnetic weak FEM-BEM couplings based on recently developed Padé approximations of Dirichlet-to-Neumann and Magnetic-to-Electric operators. The number of iterations required to solve these couplings is only slightly dependent on the frequency and the mesh refinement. The weak FEM-BEM couplings that we propose are therefore suited to accurate SCS and RCS computations at high frequencies

Being part of a research group on functional differential equations (FDEs, for short), due to my formation in non-absolute integration theory and because certain kinds of FDEs can be expressed as integral equations, I was motivated to investigate the latter. The purpose of this work, therefore, is to develop the theory of integral equations, when the integrals involved are in the sense of Kurzweil- Henstock or Kurzweil-Henstock-Stieltjes, through the correspondence between solutions of integral equations and solutions of generalized ordinary differential equations (we write generalized ODEs, for short). In order to be able to obtain results for integral equations, we propose extensions of both the Kurzweil integral and the generalized ODEs (found in [36]). We develop the fundamental properties of this new generalized ODE, such as existence and uniqueness of solutions results, and we propose stability concepts for the solutions of our new class of equations. We, then, apply these results to a class of nonlinear Volterra integral equations of the second kind. Finally, we consider a model of population growth (found in [4]) that can be expressed as an integral equation that belongs to this class of nonlinear Volterra integral equations. / Sendo parte de um grupo de pesquisa em equações diferenciais funcionais (escrevemos EDFs), por causa de minha formação em teoria de integração não absoluta e porque certos tipos de EDFs podem ser escritas como equações integrais, decidi estudar esse último tipo de equações. O objetivo desse trabalho, portanto, é desenvolver a teoria de equações integrais, quando as integrais envolvidas são no sentido de Kurzweil-Henstock ou Kurzweil-Henstock-Stieltjes, através da correspondência entre soluções de equações integrais e soluções de equações diferenciais ordinárias generalizadas (ou EDOs generalizadas). A fim de obter resultados para estas equações integrais, propomos extensões de ambas a integral de Kurzweil e as EDOs generalizadas (encontradas em [36]). Desenvolvemos propriedades fundamentais dessa nova EDO generalizada, como resultados de existência e unicidade de solução, e propomos conceitos de estabilidade para as soluções de nossa nova classe de equações. Nós, então, aplicamos esses resultados a uma classe de equações integrais de Volterra não lineares de segunda espécie. Finalmente, consideramos um modelo de crescimento de populações (encontrado em [4]) que pode ser escrito como uma equação integral pertencente a essa classe de equações integrais de Volterra não lineares.

Sabe-se que a relaxa&ccedil;&atilde;o de l&iacute;quidos em meios porosos envolve tr&ecirc;s mecanismos principais: relaxa&ccedil;&atilde;o bulk, relaxa&ccedil;&atilde;o de superf&iacute;cie e difus&atilde;o. Muitas vezes, os processos de relaxa&ccedil;&atilde;o de l&iacute;quidos confinados em meios porosos s&atilde;o dominados pelo processo de relaxa&ccedil;&atilde;o de superf&iacute;cie e difus&atilde;o do flu&iacute;do. No chamado regime de difus&atilde;o r&aacute;pida, a relaxa&ccedil;&atilde;o de um &uacute;nico poro &eacute; comandada por uma fun&ccedil;&atilde;o mono exponencial que depende, principalmente, da rela&ccedil;&atilde;o superf&iacute;cie-volume do poro, de modo que em um material poroso, isto &eacute;, contendo uma distribui&ccedil;&atilde;o ampla de tamanho de poros, o sinal de decaimento de magnetiza&ccedil;&atilde;o obtido por meio da resson&acirc;ncia magn&eacute;tica nuclear &eacute; formado pela soma de exponenciais com diferentes tempos de relaxa&ccedil;&atilde;o. O problema-chave abordado neste trabalho consiste, portanto, em obter por meio desse sinal de magnetiza&ccedil;&atilde;o a distribui&ccedil;&atilde;o dos tempos de relaxa&ccedil;&atilde;o que controlam o decaimento das fun&ccedil;&otilde;es mono-exponenciais. Matematicamente, esse sinal de decaimento de magnetiza&ccedil;&atilde;o pode ser descrito na forma geral de uma equa&ccedil;&atilde;o integral de Fredholm do primeiro tipo, cuja solu&ccedil;&atilde;o &eacute; um reconhecido problema inverso mal-posto. As abordagens utilizadas na tentativa de solucionar o problema s&atilde;o oriundas de uma &aacute;rea conhecida como processamento digital de sinais, e os seguintes m&eacute;todos s&atilde;o analisados e comparados neste trabalho: algoritmo dos m&iacute;nimos quadrados m&eacute;dios com restri&ccedil;&atilde;o de n&atilde;o negatividade (LMS-NN), algoritmo dos m&iacute;nimos quadrados m&eacute;dios com restri&ccedil;&atilde;o de n&atilde;o negatividade e regularizado (LMS-RNN), redes recorrentes de Hopfield e o j&aacute; bem conhecido na solu&ccedil;&atilde;o de problemas inversos mal-postos, o algoritmo dos m&iacute;nimos quadrados regularizado (LS-R). Os resultados obtidos no trabalho s&atilde;o bastante positivos, demonstrando que, al&eacute;m do LS-R, existem outras alternativas na solu&ccedil;&atilde;o do problema, que principalmente, permitem atestar as solu&ccedil;&otilde;es obtidas por qualquer um dos algoritmos. / It is known that the relaxation of liquids in porous media involves three principal mechanisms: bulk relaxation, surface relaxation, and diffusion. Relaxation processes of confined fluids in porous media are often controlled by surface relaxation process and diffusion. In the so-called fast diffusion regime, the relaxation of a single pore is governed by a mono-exponential function that depends primarily on the relation surface-volume of the pore, so that in a porous medium, i.e, in a medium which contains a wide distribution of pore sizes, the signal of magnetization decay obtained by nuclear magnetic resonance is composed by a sum of exponentials controlled by different relaxation times. The main issue discussed in this work consists in obtaining the distribution of relaxation times that controls the decay of the mono-exponential functions that comprise the magnetization signal. Mathematically this signal of magnetization decay can be generally described as a Fredholm integral equation of the first kind, whose solution is a recognized ill-posed inverse problem. The approaches adopted to try to solve the problem come from an area known as digital signal processing, and the following methods analyzed and compared are: non-negative least mean square algorithm (NN-LMS), regularized and nonnegative nleast mean square algorithm (RNN-LMS), recurrent Hopfield networks and regularized least square algorithm (R-LS), acknowledged in the solution of ill-posed inverse problems. The results obtained are very positive, and show that in addition to R-LS there are other alternatives in the solution of the problem, which mainly allow to attest the results achieved through any of the algorithms.

In this dissertation, a Method of Moments (MoM) Volume Integral Equation (VIE)-based modeling approach suitable for a patch or slot antenna on a thin finite dielectric substrate is developed and validated. Two new key features of this method are the use of proper dielectric basis functions and proper VIE conditioning, close to the metal surface, where the surface boundary condition of the zero tangential-component must be extended into adjacent tetrahedra. The extended boundary condition is the exact result for the piecewise-constant dielectric basis functions. The latter operation allows one to achieve a good accuracy with one layer of tetrahedra for a thin dielectric substrate and thereby greatly reduces computational cost. The use of low-order basis functions also implies the use of low-order integration schemes and faster filling of the impedance matrix. For some common patch/slot antennas, the VIE-based modeling approach is found to give an error of about 1% or less in the resonant frequency for one-layer tetrahedral meshes with a relatively small number of unknowns. This error is obtained by comparison with fine finite- element method (FEM) simulations, or with measurements, or with the analytical mode matching approach. Hence it is competitive with both the method of moments surface integral equation approach and with the FEM approach for the printed antennas on thin dielectric substrates. Along with the MoM development, the dissertation also presents the models and design procedures for a number of practical antenna configurations. They in particular include: i. a compact linearly polarized broadband planar inverted-F antenna (PIFA); ii. a circularly polarized turnstile bowtie antenna. Both the antennas are designed to operate in the low UHF band and used for indoor positioning/indoor geolocation.

The purpose of fisheries management is to achieve a sustainable development of the activity, so that future generations can also benefit from the resource. However, the optimal harvesting strategy usually maximizes an economically important objective function formed by the harvester which can lead to the extinction of the resource population. Therefore, sustainability has been far more difficult to achieve than is commonly thought; fish populations are becoming increasingly limited and catches are declining due to overexploitation. The aim of this research is to determine an optimal harvesting strategy which fulfills the economic objective of the harvester while maintaining the population density over a pre-specified minimum viable level throughout the harvest. We develop and investigate the harvesting model in both deterministic and stochastic settings. We first employ the Expected Net Present Value approach and determine the optimal harvesting policy using various optimization techniques including optimal control theory and dynamic programming. Next we use real options theory, model fish harvesting as a real option, and compute the value of the harvesting opportunity which also yields the optimal harvesting strategy. We further extend the stochastic problem to include price elasticity of demand and present results for di¤erent values of the coefficient of elasticity.