In this thesis we study hybrid numerical methods to be used in computational electromagnetics. The purpose is to address a wide frequency range relative to a given geometry. We also focus on efficient and robust numerical algorithms for computing the so called Smooth Surface Diffraction predicted by Geometrical Theory of Diffraction (GTD). We restrict the presentation to frequency domain scattering problems. The hybrid methods consist in combinations of Boundary Element Methods and asymptotic methods. Three hybrids will be presented. One of them has been developed from a theoretical idea to an industrial code. The two other hybrids will be presented mainly from a theoretical perspective. To be able to compute the Smooth Surface Diffracted field we introduce a numerical method that is to be used with surface curvature sensitive meshing, complemented with auxiliary data taken from a geometry database. By using two geometry representations we can show first order convergence and we then achieve an efficient and robust numerical algorithm. This numerical algorithm may be an essential part of an GTD implementation which in its turn is a component in the hybrid methods. As a background to our new techiniques we will also give short introductions to the Boundary Element Method and the Geometrical Theory of Diffraction from a theoretical and implementational point of view.

<p>In this study we propose a fast hybrid spectral-element time-domain (SETD) / finite-element time-domain (FETD) method for transient analysis of multiscale electromagnetic problems, where electrically fine structures with details much smaller than a typical wavelength and electrically coarse structures comparable to or larger than a typical wavelength coexist.</p><p>Simulations of multiscale electromagnetic problems, such as electromagnetic interference (EMI), electromagnetic compatibility (EMC), and electronic packaging, can be very challenging for conventional numerical methods. In terms of spatial discretization, conventional methods use a single mesh for the whole structure, thus a high discretization density required to capture the geometric characteristics of electrically fine structures will inevitably lead to a large number of wasted unknowns in the electrically coarse parts. This issue will become especially severe for orthogonal grids used by the popular finite-difference time-domain (FDTD) method. In terms of temporal integration, dense meshes in electrically fine domains will make the time step size extremely small for numerical methods with explicit time-stepping schemes. Implicit schemes can surpass stability criterion limited by the Courant-Friedrichs-Levy (CFL) condition. However, due to the large system matrices generated by conventional methods, it is almost impossible to employ implicit schemes to the whole structure for time-stepping.</p><p>To address these challenges, we propose an efficient hybrid SETD/FETD method for transient electromagnetic simulations by taking advantages of the strengths of these two methods while avoiding their weaknesses in multiscale problems. More specifically, a multiscale structure is divided into several subdomains based on the electrical size of each part, and a hybrid spectral-element / finite-element scheme is proposed for spatial discretization. The hexahedron-based spectral elements with higher interpolation degrees are efficient in modeling electrically coarse structures, and the tetrahedron-based finite elements with lower interpolation degrees are flexible in discretizing electrically fine structures with complex shapes. A non-spurious finite element method (FEM) as well as a non-spurious spectral element method (SEM) is proposed to make the hybrid SEM/FEM discretization work. For time integration we employ hybrid implicit / explicit (IMEX) time-stepping schemes, where explicit schemes are used for electrically coarse subdomains discretized by coarse spectral element meshes, and implicit schemes are used to overcome the CFL limit for electrically fine subdomains discretized by dense finite element meshes. Numerical examples show that the proposed hybrid SETD/FETD method is free of spurious modes, is flexible in discretizing sophisticated structure, and is more efficient than conventional methods for multiscale electromagnetic simulations.</p> / Dissertation

Numerical simulation is an important tool in understanding the electromagnetic field and how it interacts with the environment. Different topics for time-domain finite-difference (FDTD) and finite-element (FETD) methods for Maxwell's equations are treated in this thesis. Subcell models are of vital importance for the efficient modeling of small objects that are not resolved by the grid. A novel model for thin sheets using shell elements is proposed. This approach has the advantage of taking into account discontinuities in the normal component of the electric field, unlike previous models based on impedance boundary conditions (IBCs). Several results are presented to illustrate the capabilities of the shell element approach. Waveguides are of fundamental importance in many microwave applications, for example in antenna feeds. The key issues of excitation and truncation of waveguides are addressed. A complex frequency shifted form of the uniaxial perfectly matched layer (UPML) absorbing boundary condition (ABC) in FETD is developed. Prism elements are used to promote automatic grid generation and enhance the performance. Results are presented where reflection errors below -70dB are obtained for different types of waveguides, including inhomogeneous cases. Excitation and analysis via the scattering parameters are achieved using waveguide modes computed by a general frequency-domain mode solver for the vector Helmholtz equation. Huygens surfaces are used in both FDTD and FETD for excitation in waveguide ports. Inverse problems have received an increased interest due to the availability of powerful computers. An important application is non-destructive evaluation of material. A time-domain, minimization approach is presented where exact gradients are computed using the adjoint problem. The approach is applied to a general form of Maxwell's equations including dispersive media and UPML. Successful reconstruction examples are presented both using synthetic and experimental measurement data. Parameter reduction of complex geometries using simplified models is an interesting topic that leads to an inverse problem. Gradients for subcell parameters are derived and a successful reconstruction example is presented for a combined dielectric sheet and slot geometry.

This thesis describes research on Efficient Discrete Modelling of Axisymmetric Radiating Structures . Investigating the possibilities of surmounting the inherent limitation in the Cartesian rectangular Transmission Line Modelling (TLM) method due to staircase approximation by efficiently implementing the 3D cylindrical TLM mesh led to the development of a numerical model for simulating axisymmetric radiating structures such as cylindrical and conical monopole antennas. Following a brief introduction to the TLM method, potential applications of the method are presented. Cubic and cylindrical TLM models have been implemented in MATLAB and the code has been validated against microwave cavity benchmark problems. The results are compared to analytical results and the results obtained from the use of commercial cubic model (CST) in order to highlight the benefit of using a cylindrical model over its cubic counterpart. A cylindrical TLM mesh has not previously been used in the modelling of axisymmetric 3D radiating structures. In this thesis, it has been applied to the modelling of both cylindrical monopole and the conical monopole. The technique can also be applied to any radiating structure with axisymmetric cylindrical shape. The application of the method also led to the development of a novel conical antenna with periodic slot loading. Prototype antennas have been fabricated and measured to validate the simulated results for the antennas.

This work derives a nonlinear optical spatio-temporal instability. It is a perturbative analysis that begins from Maxwell’s equations and its constituent relations to derive a vectorial nonlinear wave equation. In fact, it is a new theoretical method that has been developed that builds on previous aspects of nonlinear optics in a more general way. The perturbation in the wave equation derived is coupled with its complex conjugate which has been taken for granted so far. Once decoupled it gives rise to a second-order equation and thus a true instability regime because the wavevector can become complex. The solution obtained for the perturbation that co-propagates with the driving laser is a generalization to modulation and filamentation instability, extending beyond the nonlinear Schrodinger and nonlinear transverse diffusion equations[1][2]. As a result of this new mechanism, new phenomena can be explored. For example, the Kerr Nonlinear Instability can lead to exponential growth, and hence amplification. This can occur even at wavelengths that are typically hard to operate at, such as into far infrared wave- lengths. This provides a mechanism for obtaining amplification in the far infrared from a small seed pulse without the need for population inversion. The analysis provides the basic framework that can be extended to many different avenues. This will be the subject of future work, as outlined in the conclusion of this thesis.

The first contribution of this thesis is a new regularity theorem for time harmonic Maxwell's equations with less than Lipschitz complex anisotropic coefficients. By using the L^p theory for elliptic equations, it is possible to prove H¹ and Hölder regularity results, provided that the coefficients are W^1,p for some p = 3. This improves previous regularity results, where the assumption W^1,∞ for the coefficients was believed to be optimal. The method can be easily extended to the case of bi-anisotropic materials, for which a separate approach turns out to be unnecessary. The second focus of this work is the boundary control of the Helmholtz and Maxwell equations to enforce local constraints inside the domain. More precisely, we look for suitable boundary conditions such that the corresponding solutions and their derivatives satisfy certain local non-zero constraints. Complex geometric optics solutions can be used to construct such illuminations, but are impractical for several reasons. We propose a constructive approach to this problem based on the use of multiple frequencies. The suitable boundary conditions are explicitly constructed and give the desired constraints, provided that a finite number of frequencies, given a priori, are chosen in a fixed range. This method is based on the holomorphicity of the solutions with respect to the frequency and on the regularity theory for the PDE under consideration. This theory finds applications to several hybrid imaging inverse problems, where the unknown coefficients have to be imaged from internal measurements. In order to perform the reconstruction, we often need to find suitable boundary conditions such that the corresponding solutions satisfy certain non-zero constraints, depending on the particular problem under consideration. The multiple frequency approach introduced in this thesis represents a valid alternative to the use of complex geometric optics solutions to construct such boundary conditions. Several examples are discussed.

Le problème de diffraction électromagnétique gouverné par les équations de Maxwell admet une formulation équivalente par une équation intégrale volumique fortement singulière. Cette thèse a pour but d'examiner l'opérateur intégral qui décrit cette équation. La première partie de ce manuscrit porte sur l'étude de son spectre essentiel. Cette analyse est intéressante en vue d'obtenir les conditions nécessaires et suffisantes pour avoir l'unicité de solutions du problème surtout quand il s'agirait de la diffraction des ondes par des matériaux négatifs où les techniques classiques perdent leurs utilité. Après avoir justifié le bon choix du cadre fonctionnel, nous étudions tout d'abord le cas où les paramètres caractéristiques du milieu à savoir la permittivité électrique et la perméabilité magnétique sont constants par morceaux avec discontinuité au travers du bord de la cible. Dans ce cadre, nous donnons une réponse complète à la question pour les domaines réguliers et Lipschitziens. Ensuite, et à l'aide d'une technique de localisation, nous donnons une extension de ces résultats dans le cas des paramètres réguliers par morceaux pour deux opérateurs intégraux, l'un qui correspond à la version diélectrique du problème et l'autre pour sa version magnétique. Nous terminons cette thèse par l'étude de la dérivée de forme des opérateurs diélectrique et magnétique et nous en déduisons une nouvelle caractérisation de la dérivée de forme des solutions des deux problèmes de diffraction. / The electromagnetic diffraction problem which is governed by the Maxwell equations admits an equivalent formulation in terms of a strongly singular volume integral equation. This thesis aims to examine the integral operator that describes this equation. The first part of this document focuses on the study of its essential spectrum. This analysis is interesting to get the necessary and sufficient conditions of solution uniqueness of the problem especially when we consider the diffraction of waves by negative materials where classic tools lose their usefulness. After justifying the adequate choice of the functional framework, we first study the case where the characteristics parameters of the medium like the electric permittivity and magnetic permeability are piecewise constant with discontinuity across the boundary of the target. In this context, we give a full answer to the question for smooth and Lipschitz domains. Then, by using a localization technique, we give an extension of those results in the case of piecewise regular parameters for two integrals operators, one which corresponds to the dielectric version of the problem and the other for its magnetic version. We end this thesis by the study of the shape derivative of the dielectric and magnetic operators and we derive a new characterization of the shape derivative of the two diffraction problems solution.

Cette thèse vise à développer et à évaluer une méthode de décomposition de domaine avec recouvrement dans le cadre de la modélisation du contrôle non destructif (CND) par courants de Foucault (CF). L'objectif d'une telle approche consiste à éviter le remaillage systématique de l'intégralité du domaine d'étude lors du déplacement de l'un de ses éléments constitutifs(par exemple, déplacement de la sonde CF au dessus de la pièce contrôlée). Plus précisément, il s'agit de concevoir une méthode de décomposition de domaine avec recouvrement qui s'appuie sur la théorie apportée par la méthode des éléments finis avec joints. En plus de s'affranchir de la contrainte d'une interface d'échange invariante avec le mouvement, la technique décrite dans ce travail réalise des transferts d'information réciproques entre les domaines. Cette étude présente les résultats théoriques ainsi que numériques liés à la simulation magnétodynamique. Par ailleurs, l'intérêt d'une telle méthode est illustré par des applications sur des configurations bidimensionnelles de CND par CF. / This thesis aims at studying and developing a domain decomposition method with overlapping subdomains for the modeling in eddy current (EC) non-Destructive testing (NDT). The idea behind such an approach is the possibility to avoid the systematic remeshes of the whole studying domain when some of its components are modified (for example the displacement of the coil above the conductor). More precisely, this work aims at designing a domain decomposition method with overlapping based on the theory of the mortar finite element method. In addition to remove the constraint owing to an coupling interface which is invariant with the displacement, the technique described, in this work, realizes reciprocal transfers of information between subdomains. This study presents the theoretical and numerical results attached to the magnetodynamic simulation. Moreover, the interest of such a method is illustrated by applications in some 2D modeling cases of EC NDT.

This thesis investigates the possibility of enhancing the efficiency of thin film solar cells by including periodic material nonhomogeneities in combination with periodically corrugated back reflectors. Two different types of solar cell are investigated; p-i-n junctions solar cells made from alloys of hydrogenated amorphous silicon (a-Si:H) (containing either carbon or germanium), and Schottky barrier junction solar cells made from alloys of indium gallium nitride (InξGa1-ξN). Material nonhomogeneities are produced by varying the fractions of the constituent elements of the alloys. For example, by varying the content of carbon or germanium in the a-Si:H alloys, semiconductors with bandgaps ranging from 1:3 eV to 1:95 eV can be produced. Changing the bandgap alters both the optical and electrical properties of the material so this necessitates the use of coupled optical and electrical models. To date, the majority of solar cell simulations either prioritise the electrical portion of the simulation or they prioritise the optical portion of the simulation. In this thesis, a coupled optoelectronic model, developed using COMSOL Multiphysics®, was used to simulate solar cells: a two-dimensional finite-element optical model, which solved Maxwell's equations throughout the solar cells, was used to calculate the absorption of incident sunlight; and a finite-element electrical drift-diffusion transport model, either one- or two-dimensional depending on the symmetries of the problem, was used to calculate the steady state current densities throughout the solar cells under external voltage biases. It is shown that a periodically corrugated back reflector made from silver can increase efficiency of an a-Si:H alloy single p-i-n junction solar cell by 9:9% compared to a baseline design, while for a triple junction the improvement is a relatively meagre 1:8%. It is subsequently shown that the efficiency of these single p-i-n junction solar cells with a back reflector can be further increased by the inclusion of material nonhomogeneities, and that increasing the nonhomogeneity progressively increases efficiency, especially in thicker solar cells. In the case of InξGa1-ξN Schottky barrier junction solar cells, the gains are shown to be even greater. An overall increase in efficiency of up to 26:8% over a baseline design is reported.

Dans cette thèse, nous étudions la convergence de méthode de type éléments finis pour les équations de Maxwell en régime harmonique avec condition au bord d'impédance et l'équation de Helmholtz avec une couche parfaitement absorbante(PML). On étudie en premier, la formulation régularisée de l'équation de Maxwell en régime harmonique avec condition au bord d'impédance (qui consiste à ajouter le term ∇ div à l'équation originale pour avoir un problème elliptique) et on garde la condition d'impédance comme une condition au bord essentielle. Pour des domaines à bord régulier, le caractère bien posé de cette formulation est bien connu mais cela n'est pas le cas pour des domaines polyédraux convexes. On commence alors le premier chapitre par la preuve du caractère bien posé dans le cas du polyèdre convexe, qui est basé sur le fait que l'espace variationnel est inclus dans H¹. Dans le but d'avoir des estimations explicites en le nombre d'onde k de ce problème, il est obligatoire d'avoir des résultats de stabilité explicites en ce nombre d'onde. C'est aussi proposé, pour quelques situations particulières, dans ce chapitre. Dans le second chapitre on décrit les singularités d'arêtes et de coins pour notre problème. On peut alors déduire la régularité de la solution du problème original, ainsi que de son adjoint. On a tous les ingrédients pour proposer une analyse de convergence explicite en k pour une méthode d'éléments finis avec éléments de Lagrange. Dans le troisième chapitre, on considère une méthode d'éléments finis hp non conforme pour un domaine à bord régulier. Pour obtenir des estimations explicites en k, on introduit un résultat de décomposition, qui sépare la solution du problème original (ou de son adjoint) en une partie régulière mais fortement oscillante et une partie moins régulière mais peu oscillante. Ce résultat permet de montrer des estimations explicites en k. Le dernier chapitre est dédié à l'équation de Helmholtz avec une PML. L'équation de Helmholtz dans l'espace entier est souvent utilisée pour modéliser la diffraction d'onde acoustique (en régime harmonique), avec la condition de radiation à l'infini de Sommerfeld. L'ajout d'une PML est une façon pour passer d'un domaine infini à un domaine fini, elle correspond à l'ajout d'une couche autour du domaine de calcul qui absorbe très vite toutes les ondes sortantes. On propose en premier un résultat de stabilité explicite en k. On propose alors deux schémas numériques, une méthode d'éléments finis hp et une méthode multi- échelle basée sur un sous-espace local de correction. Le résultat de stabilité est utilisé pour mettre en relation de choix des paramètres des méthodes numériques considérées avec k. Nous montrons aussi des estimations d'erreur a priori. A la fin de ces chapitres, des tests numériques sont proposés pour confirmer nos résultats théoriques. / In this thesis, we propose wavenumber explicit convergence analyses of some finite element methods for time-harmonic Maxwell's equations with impedance boundary condition and for the Helmholtz equation with Perfectly Matched Layer (PML). We first study the regularized formulation of time-harmonic Maxwell's equations with impedance boundary conditions (where we add a ∇ div-term to the original equation to have an elliptic problem) and keep the impedance boundary condition as an essential boundary condition. For a smooth domain, the wellposedness of this formulation is well-known. But the well-posedness for convex polyhedral domain has been not yet investigated. Hence, we start the first chapter with the proof of the well-posedness in this case, which is based on the fact that the variational space is embedded in H¹. In order to perform a wavenumber explicit error analysis of our problem, a wavenumber explicit stability estimate is mandatory. We then prove such an estimate for some particular configurations. In the second chapter, we describe the corner and edge singularities for such problem. Then we deduce the regularity of the solution of the original and the adjoint problem, thus we have all ingredients to propose a explicit wavenumber convergence analysis for h-FEM with Lagrange element. In the third chapter, we consider a non conforming hp-finite element approximation for domains with a smooth boundary. To perform a wavenumber explicit error analysis, we split the solution of the original problem (or its adjoint) into a regular but oscillating part and a rough component that behaves nicely for large frequencies. This result allows to prove convergence analysis for our FEM, again explicit in the wavenumber. The last chapter is dedicated to the Helmholtz equation with PML. The Helmholtz equation in full space is often used to model time harmonic acoustic scattering problems, with Sommerfeld radiation condition at infinity. Adding a PML is a way to reduce the infinite domain to a finite one. It corresponds to add an artificial absorbing layer surrounding a computational domain, in which scattered wave will decrease very quickly. We first propose a wavenumber explicit stability result for such problem. Then, we propose two numerical discretizations: an hp-FEM and a multiscale method based on local subspace correction. The stability result is used to relate the choice of the parameters in the numerical methods to the wavenumber. A priori error estimates are shown. At the end of each chapter, we perform numerical tests to confirm our theoritical results.