Dissertations / Theses on the topic 'PERTURBED PROBLEM'

Nous sommes intéressés dans cette thèse par l'analyse de la diffraction directe et inverse des ondes par des couches infinies périodiques localement perturbées à une fréquence fixe. Ce problème a des connexions avec le contrôle non destructif des structures périodiques telles que des structures photoniques, des fibres optiques, des réseaux, etc. Nous analysons d'abord le problème direct et établissons certaines conditions sur l'indice de réfraction pour lesquelles il n'existe pas de modes guidés. Ce type de résultat est important car il montre les cas pour lesquels les mesures peuvent être effectuées par exemple sur une couche au dessus de la structure périodique sans perdre des informations importantes dans la partie propagative de l'onde. Nous proposons ensuite une méthode numérique pour résoudre le problème de diffraction basée sur l'utilisation de la transformée de Floquet-Bloch dans les directions de périodicité. Nous discrétisons le problème de manière uniforme dans la variable de Floquet-Bloch et utilisons une méthode spectrale dans la discrétisation spatiale. La discrétisation en espace exploite une reformulation volumétrique du problème dans une cellule (équation intégrale de Lippmann-Schwinger) et une périodisation du noyau dans la direction perpendiculaire à la périodicité. Cette dernière transformation permet d'utiliser des techniques de type FFT pour accélérer le produit matrice-vecteur dans une méthode itérative pour résoudre le système linéaire. On aboutit à un système d'équations intégrales couplées (à cause de la perturbation locale) qui peuvent être résolues en utilisant une décomposition de Jacobi. L'analyse de la convergence est faite seulement dans le cas avec absorption et la validation numérique est réalisées sur des exemples 2D. Pour le problème inverse, nous étendons l'utilisation de trois méthodes d'échantillonnage pour résoudre le problème de la reconstruction de la géométrie du défaut à partir de la connaissance de données mutistatiques associées à des ondes incidentes planes en champ proche (c.à.d incluant certains modes évanescents). Nous analysons ces méthodes pour le problème semi-discrétisée dans la variable Floquet-Bloch. Nous proposons ensuite une nouvelle méthode d'imagerie capable de visualiser directement la géométrie du défaut sans savoir ni les propriétés physiques du milieux périodique, ni les propriétés physiques du défaut. Cette méthode que l'on appelle imagerie-différentielle est basée sur l'analyse des méthodes d'échantillonnage pour un seul mode de Floquet-Bloch et la relation avec les solutions de problèmes de transmission intérieurs d'un type nouveau. Les études théoriques sont corroborées par des expérimentations numériques sur des données synthétiques. Notre analyse est faite d'abord pour l'équation d'onde scalaire où le contraste est sur le terme d'ordre inférieur de l'opérateur de Helmholtz. Nous esquissons ensuite l'extension aux cas où la le contraste est également présent dans l'opérateur principal. Nous complémentons notre travail par deux résultats sur l'analyse du problème de diffraction pour des matériaux périodiques ayant des indices négatifs. Nous établissons en premier le caractère bien posé du problème en 2D dans le cas d'un contraste est égal à -1. Nous montrons également le caractère Fredholm de la formulation Lipmann-Schwinger du problème en utilisant l'approche de T-coercivité dans le cas d'un contraste différent de -1
We are interested in this thesis by the analysis of scattering and inverse scattering problems for locally perturbed periodic infinite layers at a fixed frequency. This problem has connexions with non destructive testings of periodic media like photonics structures, optical fibers, gratings, etc. We first analyze the forward scattering problem and establish some conditions under which there exist no guided modes. This type of conditions is important as it shows that measurements can be done on a layer above the structure without loosing substantial informations in the propagative part of the wave. We then propose a numerical method that solves the direct scattering problem based on Floquet-Bloch transform in the periodicity directions of the background media. We discretize the problem uniformly in the Floquet-Bloch variable and use a spectral method in the space variable. The discretization in space exploits a volumetric reformulation of the problem in a cell (Lippmann-Schwinger integral equation) and a periodization of the kernel in the direction orthogonal to the periodicity. The latter allows the use of FFT techniques to speed up Matrix-Vector product in an iterative to solve the linear system. One ends up with a system of coupled integral equations that can be solved using a Jacobi decomposition. The convergence analysis is done for the case with absorption and numerical validating results are conducted in 2D. For the inverse problem we extend the use of three sampling methods to solve the problem of retrieving the defect from the knowledge of mutistatic data associated with incident near field plane waves. We analyze these methods for the semi-discretized problem in the Floquet-Bloch variable. We then propose a new method capable of retrieving directly the defect without knowing either the background material properties nor the defect properties. This so-called differential-imaging functional that we propose is based on the analysis of sampling methods for a single Floquet-Bloch mode and the relation with solutions toso-called interior transmission problems. The theoretical investigations are corroborated with numerical experiments on synthetic data. Our analysis is done first for the scalar wave equation where the contrast is the lower order term of the Helmholtz operator. We then sketch the extension to the cases where the contrast is also present in the main operator. We complement our thesis with two results on the analysis of the scattering problem for periodic materials with negative indices. Weestablish the well posedness of the problem in 2D in the case of a contrast equals -1. We also show the Fredholm properties of the volume potential formulation of the problem using the T-coercivity approach in the case of a contrast different from -1

Singularly perturbed problems often yield solutions ith strong directional features, e.g. with boundary layers. Such anisotropic solutions lend themselves to adapted, anisotropic discretizations. The quality of the corresponding numerical solution is a key issue in any computational simulation. To this end we present a new robust error estimator for a singularly perturbed reaction-diffusion problem. In contrast to conventional estimators, our proposal is suitable for anisotropic finite element meshes. The estimator is based on the solution of a local problem, and yields error bounds uniformly in the small perturbation parameter. The error estimation is efficient, i.e. a lower error bound holds. The error estimator is also reliable, i.e. an upper error bound holds, provided that the anisotropic mesh discretizes the problem sufficiently well. A numerical example supports the analysis of our anisotropic error estimator.

A singularly perturbed reaction-diffusion model problem is considered, and the choice of an appropriate norm is discussed. Particular emphasis is given to the energy norm. Certain prejudices against this norm are investigated and disproved. Moreover, an adaptive finite element algorithm is presented which exhibits an optimal error decrease in the energy norm in some simple numerical experiments. This underlines the suitability of the energy norm.

This thesis presents a numerical investigation of a problem on a semi infinite waveguide. The domain considered here is of a much more general form than those that have been considered using classical techniques. The motivation for this work originates from the work in 28, where unlike here, a perturbation technique was used to solve a simpler problem.

Singularly perturbed reaction-diffusion problems exhibit in general solutions with anisotropic features, e.g. strong boundary and/or interior layers. This anisotropy is reflected in the discretization by using meshes with anisotropic elements. The quality of the numerical solution rests on the robustness of the a posteriori error estimator with respect to both the perturbation parameters of the problem and the anisotropy of the mesh. An estimator that has shown to be one of the most reliable for reaction-diffusion problem is the equilibrated residual method and its modification done by Ainsworth and Babuška for singularly perturbed problem. However, even the modified method is not robust in the case of anisotropic meshes. The present work modifies the equilibrated residual method for anisotropic meshes. The resulting error estimator is equivalent to the equilibrated residual method in the case of isotropic meshes and is proved to be robust on anisotropic meshes as well. A numerical example confirms the theory.

The paper deals with a singularly perturbed reaction diffusion model problem. The focus is on reliable a posteriori error estimators for the H^1 seminorm that can be applied to anisotropic finite element meshes. A residual error estimator and a local problem error estimator are proposed and rigorously analysed. They are locally equivalent, and both bound the error reliably. Furthermore three modifications of these estimators are introduced and discussed. Numerical experiments for all estimators complement and confirm the theoretical results.

The aim of this thesis is the study of the dynamics of small perturbations of the spatial central motion problem. Our main result consists in proving that if the central potential is analytic, then, except for the Harmonic and the Keplerian case, the unperturbed system written in action angle variables is quasiconvex. Thus, when it is perturbed, one can apply a Nekhoroshev type theorem ensuring the stability over exponentially long times of the modulus of the angular momentum and of the energy of the unperturbed system. Being a \emph{superintegrable} system, namely, a system which admits a number of independent integrals of motion larger than the number of degrees of freedom, the version of Nekhoroshev theorem provided here is the one for superintegrable systems. We also give a complete proof (à la Lochak) of this result.

The dissertation is made of two chapters. The first chapter is dedicated to the investigation of some properties of the layer potentials of a constant coefficient elliptic partial differential operator. In the second chapter, we focus our attention to the Lamè equations, which are related to the physic of an isotropic homogeneous elastic body. In particular, in the first chapter, we investigate the dependence of the single layer potential upon perturbation of the density, the support and the coefficients of the corresponding operator. Under some more restrictive assumptions on the operator, we prove a real analyticity theorem for the single layer potential and its derivatives. As a first step, we introduce a particular fundamental solution of a given constant coefficient partial differential operator. For this purpose, we exploite the construction of a fundamental solution given by John (1955). We have verified that, if the coefficients of the operator are constrained to a bounded set, then there exist a particular fundamental solution which is a sum of functions which depend real analytically on the coefficients of the operator. Such a result resembles the results of Mantlik (1991, 1992) (see also Tréves (1962)), where more general assumptions on the operator are considered. We observe that it is not a corollary. Indeed, we need a suitably detailed expression for the fundamental solution, which cannot be deduced by Mantlik's results. The next step is to introduce the support of our single layer potentials. It will be a compact sub-manifold of the the n-dimensional euclidean space parametrized by a suitable diffeomorphism defined on the boundary of a fixed domain. Then, we will be ready to state in Theorem 1.7 the main result of this chapter, which is a real analyticity result in the frame of Shauder spaces. The main idea of the proof stems from the papers of Lanza de Cristoforis & Preciso (1999) and by Lanza de Cristoforis & Rossi (2004, 2005) and exploits the Implicit Mapping Theorem for real analytic functions. Indeed, our main Theorem 1.7 is in some sense a natural extension of theorems obtained by Lanza de Cristoforis & Preciso (1999) and by Lanza de Cristoforis & Rossi (2004, 2005), for the Cauchy integral and for the Laplace and Helmholtz operators, respectively. Here we confine our attention to elliptic operators which can be factorized with operators of order 2. In the last section of the first chapter, we consider some applications of Theorem 1.7. In particular, we deduce a real analyticity theorem for the single and double layer potential which arise in the analysis of the boundary value problems for the Lamè equations and for the Stokes system. In the second chapter, we focus our attention to the Lamè equations. We consider some boundary value problems defined in a domain with a small hole. For each of them, we investigate the behavior of the solution and of the corresponding energy integral as the hole shrinks to a point. This kind of problem is not new at all and has been long investigated by the techniques of asymptotic analysis. It is perhaps difficult to give a complete list of contributions. Here we mention the work of Keller, Kozlov, Movchan, Maz'ya, Nazarov, Plamenewskii, Ozawa and Ward. The results that we present are in accordance with the behavior one would expect by looking at the above mentioned literature, but we adopt a different approach proposed by Lanza de Cristoforis (2001, 2002, 2005, 2007.) To do so, we exploit the real analyticity results for the elastic layer potentials obtained in the first chapter. We now briefly outline the main difference between our approach and the one of asymptotic analysis. Let d>0 be a parameter which is proportional to the diameter of the hole, so that the singularity of the domain appears when d=0. By the approach of the asymptotic analysis, we can expect to obtain results which are expressed by means of known functions of d plus an unknown term which is smaller than a positive power of d. Whereas, our results are expressed by means of real analytic functions of d defined in a whole open neighborhood of d=0 and by, possibly singular, but completely known functions of d, such as d^(2-n) or log d. Moreover, not only we can consider the dependence upon d, we can also investigate the dependence of the solution and the corresponding energy integral upon perturbations of the coefficients of the operator, and of the point where the hole is situated, and of the shape of the hole, and of the shape of the outer domain, and of the boundary data on the boundary of the hole, and of the boundary data on the boundary of the outer domain, and of the interior data. Also in this case we obtain results expressed by means of real analytic functions and completely known functions such as d^(2-n) and log d. The first boundary value problem we have studied is a Dirichlet boundary value problem with homogeneous data in the interior. Then, we turned to investigate a Robin boundary value problem with homogeneous data in the interior. In this case we have also described the behavior of the solution and the corresponding energy integral when both the domain and the boundary data display a singularity for d=0. Finally, we have studied a Dirichlet boundary value problem with non-homogeneous data in the interior.

We consider a singularly perturbed reaction-diffusion problem and derive and rigorously analyse an a posteriori residual error estimator that can be applied to anisotropic finite element meshes. The quotient of the upper and lower error bounds is the so-called matching function which depends on the anisotropy (of the mesh and the solution) but not on the small perturbation parameter. This matching function measures how well the anisotropic finite element mesh corresponds to the anisotropic problem. Provided this correspondence is sufficiently good, the matching function is O(1). Hence one obtains tight error bounds, i.e. the error estimator is reliable and efficient as well as robust with respect to the small perturbation parameter. A numerical example supports the anisotropic error analysis.

A singularly perturbed convection-diffusion problem in two and three space dimensions is discretized using the streamline upwind Petrov Galerkin (SUPG) variant of the finite element method. The dominant convection frequently gives rise to solutions with layers; hence anisotropic finite elements can be applied advantageously. The main focus is on a posteriori energy norm error estimation that is robust in the perturbation parameter and with respect to the mesh anisotropy. A residual error estimator and a local problem error estimator are proposed and investigated. The analysis reveals that the upper error bound depends on the alignment of the anisotropies of the mesh and of the solution. Hence reliable error estimation is possible for suitable anisotropic meshes. The lower error bound depends on the problem data via a local mesh Peclet number. Thus efficient error estimation is achieved for small mesh Peclet numbers. Altogether, error estimation approaches for isotropic meshes are successfully extended to anisotropic elements. Several numerical experiments support the analysis.

Thesis (Ph.D.)--Indiana University, Dept. of Mathematics, 2006.
Source: Dissertation Abstracts International, Volume: 67-04, Section: B, page: 2029. Adviser: Roger Temam. "Title from dissertation home page (viewed June 20, 2007)."

Magister Scientiae - MSc
Many chemical and physical problems are mathematically described by partial differential equations (PDEs). These PDEs are often highly nonlinear and therefore have no closed form solutions. Thus, it is necessary to recourse to numerical approaches to determine suitable approximations to the solution of such equations. For solutions possessing sharp spatial transitions (such as boundary or interior layers), standard numerical methods have shown limitations as they fail to capture large gradients. The method of lines (MOL) is one of the numerical methods used to solve PDEs. It proceeds by the discretization of all but one dimension leading to systems of ordinary di erential equations. In the case of time-dependent PDEs, the MOL consists of discretizing the spatial derivatives only leaving the time variable continuous. The process results in a system to which a numerical method for initial value problems can be applied. In this project we consider various types of singularly perturbed time-dependent PDEs. For each type, using the MOL, the spatial dimensions will be discretized in many different ways following fitted numerical approaches. Each discretisation will be analysed for stability and convergence. Extensive experiments will be conducted to confirm the analyses.

In this thesis singularly perturbed convection-diffusion equations in the unit square are considered. Due to the presence of a small perturbation parameter the solutions of those problems exhibit an exponential layer near the outflow boundary and two parabolic layers near the characteristic boundaries. Discretisation of such problems on standard meshes and with standard methods leads to numerical solutions with unphysical oscillations, unless the mesh size is of order of the perturbation parameter which is impracticable. Instead we aim at uniformly convergent methods using layer-adapted meshes combined with standard methods. The meshes considered here are S-type meshes--generalisations of the standard Shishkin mesh. The domain is dissected in a non-layer part and layer parts. Inside the layer parts, the mesh might be anisotropic and non-uniform, depending on a mesh-generating function. We show, that the unstabilised Galerkin finite element method with bilinear elements on an S-type mesh is uniformly convergent in the energy norm of order (almost) one. Moreover, the numerical solution shows a supercloseness property, i.e. the numerical solution is closer to the nodal bilinear interpolant than to the exact solution in the given norm. Unfortunately, the Galerkin method lacks stability resulting in linear systems that are hard to solve. To overcome this drawback, stabilisation methods are used. We analyse different stabilisation techniques with respect to the supercloseness property. For the residual-based methods Streamline Diffusion FEM and Galerkin Least Squares FEM, the choice of parameters is addressed additionally. The modern stabilisation technique Continuous Interior Penalty FEM--penalisation of jumps of derivatives--is considered too. All those methods are proved to possess convergence and supercloseness properties similar to the standard Galerkin FEM. With a suitable postprocessing operator, the supercloseness property can be used to enhance the accuracy of the numerical solution and superconvergence of order (almost) two can be proved. We compare different postprocessing methods and prove superconvergence of above numerical methods on S-type meshes. To recover the exact solution, we apply continuous biquadratic interpolation on a macro mesh, a discontinuous biquadratic projection on a macro mesh and two methods to recover the gradient of the exact solution. Special attentions is payed to the effects of non-uniformity due to the S-type meshes. Numerical simulations illustrate the theoretical results.

The paper is concerned with the finite element resolution of layers appearing in singularly perturbed problems. A special anisotropic grid of Shishkin type is constructed for reaction diffusion problems. Estimates of the finite element error in the energy norm are derived for two methods, namely the standard Galerkin method and a stabilized Galerkin method. The estimates are uniformly valid with respect to the (small) diffusion parameter. One ingredient is a pointwise description of derivatives of the continuous solution. A numerical example supports the result. Another key ingredient for the error analysis is a refined estimate for (higher) derivatives of the interpolation error. The assumptions on admissible anisotropic finite elements are formulated in terms of geometrical conditions for triangles and tetrahedra. The application of these estimates is not restricted to the special problem considered in this paper.

The paper is concerned with the Nitsche mortaring in the framework of domain decomposition where non-matching meshes and weak continuity of the finite element approximation at the interface are admitted. The approach is applied to singularly perturbed reaction-diffusion problems in 2D. Non-matching meshes of triangles being anisotropic in the boundary layers are applied. Some properties as well as error estimates of the Nitsche mortar finite element schemes are proved. In particular, using a suitable degree of anisotropy of triangles in the boundary layers of a rectangle, we derive convergence rates as known for the conforming finite element method in presence of regular solutions. Numerical examples illustrate the approach and the results.

In this thesis we consider the numerical solution of singularly perturbed two-point boundary value problems in ordinary differential equations. We examine implementation methods for general purpose solvers of first order linear systems. The basic difference scheme is collocation at Gauss points, with a new symmetric Runge-Kutta implementation. Adaptive mesh selection is based on localized error estimates at the collocation points. These methods are implemented as modifications to the successful collocation code, COLSYS (Ascher, Christiansen & Russell), which was designed for mildly stiff problems only. Efficient high order approximations to extremely stiff problems are obtained, and comparisons to COLSYS show that the modifications work much better as the singular perturbation parameter gets small (i.e. the problem gets stiff), for both boundary layer and turning point problems.
Science, Faculty of
Computer Science, Department of
Graduate

Thesis (Ph. D. in Computational and Applied Mathematics)--Southern Methodist University.
Title from PDF title page (viewed July 13, 2007). Source: Dissertation Abstracts International, Volume: 66-11, Section: B, page: 6015. Adviser: Ian Gladwell. Includes bibliographical references.

This thesis is concerned with various aspects of perturbed conformal field theory and the methods used to calculate finite-size effects of integrable quantum field theories. Nonlinear integral equations are the main tools to find the exact ground-state energy of a quantum field theory. The thermodyamic Bethe ansatz (TBA) equations are a set of examples and are known for a large number of models. However, it is also an interesting question to find exact equations describing the excited states of integrable models. The first part of this thesis uses analytical continuation in a continuous parameter to find TBA like equations describing the spin-zero excited states of the sine-Gordon model at coupling β(^2) = 16π/3. Comparisons are then made with a further type of nonlinear integral equation which also predicts the excited state energies. Relations between the two types of equation are studied using a set of functional relations recently introduced in integrable quantum field theory. A relevant perturbation of a conformal field theory results in either a massive quantum field theory such as the sine-Gordon model, or a different massless conformal field theory. The second part of this thesis investigates flows between conformal field theories using a nonlinear integral equation. New families of flows are found which exhibit a rather unexpected behaviour. The final part of this thesis begins with a review of a connection between integrable quantum field theory and properties of certain ordinary differential equations of second- and third-order. The connection is based on functional relations which appear on both sides of the correspondence; for the second-order case these are exactly the functional relations mentioned above. The results are extended to include a correspondence between n(^th) order differential equations and Bethe ansatz system of SU(n) type. A set of nonlinear integral equations are derived to check the results.

The limits of the current micro-scale electronics technology have been approaching rapidly. At nano-scale, however, the physical phenomena involved are fundamentally different than in micro-scale. Classical and semi-classical physical principles are no longer powerful enough or even valid to describe the phenomena involved. The rich and powerful concepts in quantum mechanics have become indispensable. There are several commercial software packages already available for modeling and simulation of the electrical, magnetic, and mechanical characteristics and properties of the nano-scale devices. However, our objective here is to go one step further and create a physics-based problem-adapted solution methodology. We carry out computation for eigenfunctions of canonical and the associated perturbed quantum systems and utilize them as co-ordinate functions for solving more complex problems. We have profoundly worked with the infinite quantum potential-well problem, since they have closed-form solutions and therefore are analytically known eigenfunctions. Perturbation of the infinite quantum potential-well was done through a single box function, multiple box functions, and with a triangular function. The proposed solution concept utilizes the notion of

This thesis is concerned with the numerical approximation of problems of fluid flow, in particular the stationary advection diffusion reaction equations and the time dependent, coupled equations of incompressible miscible displacement in a porous medium. We begin by introducing the continuous discontinuous Galerkin method for the singularly perturbed advection diffusion reaction problem. This is a method which coincides with the continuous Galerkin method away from internal and boundary layers and with a discontinuous Galerkin method in the vicinity of layers. We prove that this consistent method is stable in the streamline diffusion norm if the convection field flows non-characteristically from the region of the continuous Galerkin to the region of the discontinuous Galerkin method. We then turn our attention to the equations of incompressible miscible displacement for the concentration, pressure and velocity of one fluid in a porous medium being displaced by another. We show a reliable a posteriori error estimator for the time dependent, coupled equations in the case where the solution has sufficient regularity and the velocity is bounded. We remark that these conditions may not be attained in physically realistic geometries. We therefore present an abstract approach to the stationary problem of miscible displacement and investigate an a posteriori error estimator using weighted spaces that relies on lower regularity requirements for the true solution. We then return to the continuous discontinuous Galerkin method. We prove in an abstract setting that standard (continuous) Galerkin finite element approximations are the limit of interior penalty discontinuous Galerkin approximations as the penalty parameter tends to infinity. We then show that by varying the penalization parameter on only a subset of the domain we reach the continuous discontinuous method in the limit. We present numerical experiments illustrating this approach both for equations of non-negative characteristic form (closely related to advection diffusion reaction equations) and to the problem of incompressible miscible displacement. We show that we may practically determine appropriate discontinuous and continuous regions, resulting in a significant reduction of the number of degrees of freedom required to approximate a solution, by using the properties of the discontinuous Galerkin approximation to the advection diffusion reaction equation. We finally present novel code for implementing the continuous discontinuous Galerkin method in C++.

The question of how to optimally control a large scale system is widely considered to be difficult to solve due to the size of the problem. This difficulty is further compounded when a system exhibits a two time-scale structure where some components evolve slowly and others evolve quickly. When this occurs, the optimal control problem is regarded as singularly perturbed with a perturbation parameter epsilon representing the ratio of the slow time-scale to the fast time-scale. As epsilon goes to zero, the system becomes stiff resulting in a computationally intractable problem. In this thesis, we propose an analytic method for constructing bounds on the minimum cost of a singularly perturbed, linear-quadratic optimal control problem that hold for any arbitrary value of epsilon.

[EN] This thesis addresses the treatment of perturbed problems with splitting methods. After motivating these problems in Chapter 1, we give a thorough introduction in Chapter 2, which includes the objectives, several basic techniques and already existing methods. In Chapter 3, we consider the numerical integration of non-autonomous separable parabolic equations using high order splitting methods with complex coefficients (methods with real coefficients of order greater than two necessarily have negative coefficients). We propose to consider a class of methods that allows us to evaluate all time dependent operators at real values of the time, leading to schemes which are stable and simple to implement. If the system can be considered as the perturbation of an exactly solvable problem and the flow of the dominant part is advanced using real coefficients, it is possible to build highly efficient methods for these problems. We show the performance of this class of methods for several numerical examples and present some new improved schemes. In Chapter 4, we propose splitting methods for the computation of the exponential of perturbed matrices which can be written as the sum A = D+epsilon*B of a sparse and efficiently exponentiable matrix D with sparse exponential exp(D) and a dense matrix epsilon*B which is of small norm in comparison with D. The predominant algorithm is based on scaling the large matrix A by a small number 2^(-s) , which is then exponentiated by efficient Padé or Taylor methods and finally squared in order to obtain an approximation for the full exponential. In this setting, the main portion of the computational cost arises from dense-matrix multiplications and we present a modified squaring which takes advantage of the smallness of the perturbation matrix B in order to reduce the number of squarings necessary. Theoretical results on local error and error propagation for splitting methods are complemented with numerical experiments and show a clear improvement over existing methods when medium precision is sought. In Chapter 5, we consider the numerical integration of the perturbed Hill's equation. Parametric resonances can appear and this property is of great interest in many different physical applications. Usually, the Hill's equations originate from a Hamiltonian function and the fundamental matrix solution is a symplectic matrix. This is a very important property to be preserved by the numerical integrators. In this chapter we present new sixth-and eighth-order symplectic exponential integrators that are tailored to the Hill's equation. The methods are based on an efficient symplectic approximation to the exponential of high dimensional coupled autonomous harmonic oscillators and yield accurate results for oscillatory problems at a low computational cost. Several numerical examples illustrate the performance of the new methods. Conclusions and pointers to further research are detailed in Chapter 6.
[ES] Esta tesis aborda el tratamiento de problemas perturbados con métodos de escisión (splitting). Tras motivar el origen de este tipo de problemas en el capítulo 1, introducimos los objetivos, varias técnicas básicas y métodos existentes en capítulo 2. En el capítulo 3 consideramos la integración numérica de ecuaciones no autónomas separables y parabólicas usando métodos de splitting de orden mayor que dos usando coeficientes complejos (métodos con coeficientes reales de orden mayor de dos necesariamente tienen coeficientes negativos). Proponemos una clase de métodos que permite evaluar todos los operadores con dependencia temporal en valores reales del tiempo lo cual genera esquemas estables y fáciles de implementar. Si el sistema se puede considerar como una perturbación de un problema resoluble de forma exacta y si el flujo de la parte dominante se avanza usando coeficientes reales, es posible construir métodos altamente eficientes para este tipo de problemas. Demostramos la eficiencia de estos métodos en varios ejemplos numéricos. En el capítulo 4 proponemos métodos de splitting para el cálculo de la exponencial de matrices perturbadas que se pueden escribir como suma A = D + epsilon*B de una matriz dispersa y eficientemente exponenciable con exponencial dispersa exp(D) y una matriz densa epsilon*B de noma pequeña. El algoritmo predominante se basa en escalar la matriz grande con un número pequeño 2^(-s) para poder exponenciar el resultado con métodos eficientes de Padé o Taylor y finalmente obtener la aproximación a la exponencial elevando al cuadrado repetidamente. En este contexto, el coste computacional proviene de las multiplicaciones de matrices densas y presentamos una cuadratura modificada aprovechando la estructura perturbada para reducir el número de productos. Resultados teóricos sobre errores locales y propagación de error para métodos de splitting son complementados con experimentos numéricos y muestran una clara mejora sobre métodos existentes a precisión media. En el capítulo 5, consideramos la integración numérica de la ecuación de Hill perturbada. Resonancias paramétricas pueden aparecer y esta propiedad es de gran interés en muchas aplicaciones físicas. Habitualmente, las ecuaciones de Hill provienen de una función hamiltoniana y la solución fundamental es una matriz simpléctica, una propiedad muy importante que preservar con los integradores numéricos. Presentamos nuevos integradores simplécticos exponenciales de orden seis y ocho tallados a la ecuación de Hills. Estos métodos se basan en una aproximación simpléctica eficiente a la exponencial de osciladores armónicos acoplados de dimensión alta y dan lugar a resultados precisos para problemas oscilatorios a un coste computacional bajo y varios ejemplos numéricos ilustran su rendimiento. Conclusiones e indicadores para futuros estudios se detallan en el capítulo 6.
[CAT] La present tesi està enfocada al tractament de problemes perturbats utilitzant, entre altres, mètodes d'escisió (splitting). Comencem motivant l'oritge d'aquest tipus de problems al capítol 1, i a continuació introduïm el objectius, diferents tècniques bàsiques i alguns mètodes existents al capítol 2. Al capítol 3, consideram la integració numèrica d'equacions no autònomes separables i parabòliques utilitzant mètodes d'splitting d'ordre major que dos utilitzant coeficients complexos (mètodes amb coeficients reials d'ordre major que dos necesariament tenen coeficients negatius). Proposem una clase de mètodes que permeten evaluar tots els operadors amb dependència temporal explícita amb valors reials del temps. Esta forma de procedir genera esquemes estables i fàcils d'implementar. Si el sistema es pot considerar com una perturbació d'un problema exactament resoluble, i la part dominant s'avança utilitzant coeficients reials, es posible construir mètodes altament eficients per aquest tipus de problemes Demostrem la eficiència d'estos mètodes per a diferents exemples numèrics. Al capítol 4, proposem mètodes d'splitting per al càcul de la exponencial de matrius pertorbades que es poden escriure com suma A = D + epsilon*B (una matriu que es pot exponenciar fàcilment i eficientemente, com es el cas d'algunes matrius disperses exp(D), i una matriu densa epsilon*B de norma menuda). L'algorisme predominant es basa en escalar la matriu gran amb un nombre menut 2^(-s) per a poder exponenciar el resultat amb mètodes eficients de Padé o Taylor i finalment obtindre la aproximació a la exponencial elevant al quadrat repetidament. En este context, el cost computacional prové de les multiplicacions de matrius denses i presentem una quadratura modificada aprofitant la estructura de matriu pertorbada per reduir el nombre de productes. Resultats teòrics sobre errors locals i propagació d'error per a mètodes d'splitting son analitzats i corroborats amb experiments numèrics, mostrant una clara millora respecte a mètodes existens quan es busca una precisió moderada. Al capítol 5, considerem la integració numèrica de l'ecuació de Hill pertorbada. En este tipus d'equacions poden apareixer resonàncies paramètriques i esta propietat es de gran interés en moltes aplicacions físiques. Habitualment, les equacions de Hill provenen d'una función hamiltoniana i la solució fonamental es una matriu simplèctica, siguent esta una propietat molt important a preservar pels integradors numèrics. Presentams nous integradors simplèctics exponencials d'orden sis i huit construits especialmente per resoldre l'ecuació de Hill. Estos mètodes es basen en una aproxmiació simplèctica eficient a la exponencial d'osciladors harmònics acoplats de dimensió alta i donen lloc a resultats precisos per a problemas oscilatoris a un cost computacional baix. La eficiencia dels mètodes s'il.lustra en diferents exemples numèrics. Conclusions i indicadors per a futurs estudis es detallen al capítol 6.
Seydaoglu, M. (2016). Splitting methods for autonomous and non-autonomous perturbed equations [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/71358
TESIS

A recent method for solving singular perturbation problems is examined. It is designed for the applied mathematician or engineer who needs a convenient, useful tool that requires little preparation and can be readily implemented using little more than an industry-standard software package for spreadsheets. In this paper, we shall examine singularly perturbed two point boundary value problems with the boundary layer at one end point. An initial-value technique is used for its solution by replacing the problem with an asymptotically equivalent first order problem, which is, in turn, solved as an initial value problem by using cubic splines. Numerical examples are provided to show that the method presented provides a fine approximation of the exact solution. The first chapter provides some background material to the cubic spline and boundary value problems. The works of several authors and a comparison of different solution methods are also discussed. Finally, some background into the specific singularly perturbed boundary value problems is introduced. The second chapter contains calculations and derivations necessary for the cubic spline and the initial value technique which are used in the solutions to the boundary value problems. The third chapter contains some worked numerical examples and the numerical data obtained along with most of the tables and figures that describe the solutions. The thesis concludes with some reflections on the results obtained and some discussion of the error bounds on the calculated approximations to the exact solutions for the numeric examples discussed.
ID: 029051011; System requirements: World Wide Web browser and PDF reader.; Mode of access: World Wide Web.; Thesis (M.S.)--University of Central Florida, 2010.; Includes bibliographical references (p. 48-50).
M.S.
Masters
Department of Mathematics
Sciences

It is well-known that the solution of a so-called singularly perturbed differential equation exhibits layers. These are small regions in the domain where the solution changes drastically. These layers deteriorate the convergence of standard numerical algorithms, such as the finite element method on a uniform mesh. In the past many approaches were developed to overcome this difficulty. In this context it was very helpful to understand the structure of the solution - especially to know where the layers can occur. Therefore, we have a lot of analysis in the literature concerning the properties of solutions of such problems. Nevertheless, this field is far from being understood conclusively. More recently, there is an increasing interest in the numerics of optimal control problems subject to a singularly perturbed convection-diffusion equation and box constraints for the control. However, it is not much known about the solutions of such optimal control problems. The proposed solution methods are based on the experience one has from scalar singularly perturbed differential equations, but so far, the analysis presented does not use the structure of the solution and in fact, the provided bounds are rather meaningless for solutions which exhibit boundary layers, since these bounds scale like epsilon^(-1.5) as epsilon converges to 0. In this thesis we strive to prove bounds for the solution and its derivatives of the optimal control problem. These bounds show that there is an additional layer that is weaker than the layers one expects knowing the results for scalar differential equation problems, but that weak layer deteriorates the convergence of the proposed methods. In Chapter 1 and 2 we discuss the optimal control problem for the one-dimensional case. We consider the case without control constraints and the case with control constraints separately. For the case without control constraints we develop a method to prove bounds for arbitrary derivatives of the solution, given the data is smooth enough. For the latter case we prove bounds for the derivatives up to the second order. Subsequently, we discuss several discretization methods. In this context we use special Shishkin meshes. These meshes are piecewise equidistant, but have a very fine subdivision in the region of the layers. Additionally, we consider different ways of discretizing the control constraints. The first one enforces the compliance of the constraints everywhere and the other one enforces it only in the mesh nodes. For each proposed algorithm we prove convergence estimates that are independent of the parameter epsilon. Hence, they are meaningful even for small values of epsilon. As a next step we turn to the two-dimensional case. To be able to adapt the proofs of Chapter 2 to this case we require bounds for the solution of the scalar differential equation problem for a right hand side f only in W^(1,infty). Although, a lot of results for this problem can be found in the literature but we can not apply any of them, because they require a smooth right hand side f in C^(2,alpha) for some alpha in (0,1). Therefore, we dedicate Chapter 3 to the analysis of the scalar differential equations problem only using a right hand side f that is not very smooth. In Chapter 4 we strive to prove bounds for the solution of the optimal control problem in the two dimensional case. The analysis for this problem is not complete. Especially, the characteristic layers induce subproblems that are not understood completely. Hence, we can not prove sharp bounds for all terms in the solution decomposition we construct. Nevertheless, we propose a solution method. Numerical results indicate an epsilon-independent convergence for the considered examples - although we are not able to prove this.

Magister Scientiae - MSc
Efficient numerical approaches for parameter dependent problems have been an inter- esting subject to numerical analysts and engineers over the past decades. This is due to the prominent role that these problems play in modeling many real life situations in applied sciences. Often, the choice and the e ciency of the approaches depend on the nature of the problem to solve. In this work, we consider the general linear first-order singularly perturbed Volterra integro-differential equations (SPVIDEs). These singularly perturbed problems (SPPs) are governed by integro-differential equations in which the derivative term is multiplied by a small parameter, known as "perturbation parameter". It is known that when this perturbation parameter approaches zero, the solution undergoes fast transitions across narrow regions of the domain (termed boundary or interior layer) thus affecting the convergence of the standard numerical methods. Therefore one often seeks for numerical approaches which preserve stability for all the values of the perturbation parameter, that is "numerical methods. This work seeks to investigate some "numerical methods that have been used to solve SPVIDEs. It also proposes alternative ones. The various numerical methods are composed of a fitted finite difference scheme used along with suitably chosen interpolating quadrature rules. For each method investigated or designed, we analyse its stability and convergence. Finally, numerical computations are carried out on some test examples to con rm the robustness and competitiveness of the proposed methods.

The subject that concerns this thesis is the modelling and control of plasma equilibria in the RFX-mod device operating as shaped tokamak. The aim was to develop an overall model of the plasma-conductors-controller system of RFX-mod shaped tokamak configuration for electromagnetic control purposes, with particular focus on vertical stability. Thus, the RFX-mod device is described by models of increasing complexity and involving both theoretical and experimental data. The CREATE-L code is used to develop 2D linearized plasma response models, with simplifying assumptions on the conducting structures (axisymmetric approximations). Such models, thanks to their simplicity, have been used for feedback controller design. The CarMa0 code is used to develop linearized plasma response models, but considering a detailed 3D description of the conducting structures. These models provide useful hints on the accuracy of the simplified models and on the importance of 3D structures in the plasma dynamics. The CarMa0NL code is used to model the time evolution of plasma equilibria, by taking into account also nonlinear effects which can come into play during specific phases (e.g. disruptions, limiter-to-divertor transitions, L-H transition etc.). The activity can be divided into two main parts: the first one involves the modelling of numerically generated low-β plasmas, which are used as a reference for the design and implementation of the plasma shape and position control system; the second part is related to the results of the experimental campaigns on shaped plasmas from low-β to H-mode regime, with particular efforts on the development of a novel plasma response model for the new equilibrium regimes achieved. Several challenges and peculiarities characterize the project in both the modelling and control frameworks. Strong plasma shape and different plasma regimes (i.e. low-β to H-mode plasmas), deeply affect the modelling activity and require the development of several numerical tools and methods of analysis. From the control system point of view, non-totally observable dynamic and model order reduction requirements allowed a full application of the model based approach in order to successfully design the plasma shape and vertical stability control system. The first part is based on theoretical data generated by the MAXFEA equilibrium code and used to derive the linearized model through the CREATE-L code. Two reference models have been produced for the magnetic configurations interested in shaped operations: the lower single null (LSN) and the upper single null (USN). The CREATE-L models are the most simple in terms of modelling complexity, because the conducting structures are described within the axisymmetric approximation. On the other hand, the simple but reliable properties of the CREATE-L model led to the successful design of the RFX-mod plasma shape and control system, which has been successfully tested and used to increase plasma performances involved in the second part of the thesis. Then, an investigation on the possible 3D effects of the conducting structures on these numerically generated plasma configurations has been carried out by producing plasma linearized models with an increased level of complexity. A detailed 3D volumetric description of the conducting structures of RFX-mod has been carried out and included in the plasma linearized models through the CarMa0 code. A comparison between the accuracy of this model and the previous 2D one has been performed. The different assumptions and approximations of the various models allow a clear identification of the key phenomena ruling the evolution of the n=0 vertical instability in RFX-mod tokamak discharges, and hence, provide fundamental information in the planning and the execution of related experiments and in refining the control system design. Finally, the nonlinear evolutionary equilibrium model including 3D volumetric structures CarMa0NL has been used to model nonlinear effects by simulating a "fictitious" linear current quench. The second part involves a modelling activity strictly related to the results of the experimental campaigns. In particular, new linearized models for the experimental plasmas in USN configuration have been carried out for all the plasma regimes involved in the experimental campaign, i.e. from low-β to H-mode. An iterative procedure for the production of accurate linearized plasma response models has been realized in order to handle the experimental data. The new plasma linearized models allowed further investigations on vertical stability, including 3D wall effects, in the three different plasma regimes (i.e. low-β, intermediate-β, H-mode). Furthermore, the axisymmetric plasma linearized models (CREATE-L) have been analyzed in the framework of the control theory revealing peculiar features in terms of associated SISO transfer function for vertical stability control and in terms of full MIMO model for shaping control. The MIMO model has been used to investigate the plasma wall-gaps oscillations experimentally observed in some intermediate-β plasma shots. A non-linear time evolution of the plasma discharge for a low-β plasma has been carried out by using the evolutionary equilibrium code CarMa0NL. Finally, it was investigated the vertical instability for the experimental plasmas in terms of a possible relation between plasma parameters and the occurrence of it; for these purposes, the solution of the inverse plasma equilibrium problem for the production of numerically generated plasma equilibria with variations on the plasma parameters observed experimentally was performed. This involves a wide class of numerical methods that will be described in details. Then, statistical hypothesis test has been adopted to compare the mean values of the parameters of both experimental and numerically generated plasmas showing different behaviours in terms of vertical stability.
La presente tesi tratta la modellazione e il controllo di plasmi in equilibrio, a sezione non circolare e relativi all’esperimento RFX-mod operante come tokamak. L’obiettivo è di sviluppare un modello complessivo di RFX-mod (includendo plasmaconduttori- controllore) con finalità di controllo elettromagnetico del plasma. L’esperimento RFX-mod è stato descritto con modelli caratterizzati da un crescente livello di complessità, coinvolgendo sia dati teorici che sperimentali. Il codice CREATE-L è stato usato per lo sviluppo di modelli linearizzati di risposta di plasma, con ipotesi semplificative sulla rappresentazione delle strutture conduttrici (approssimazione assialsimmetrica). Questi modelli, grazie alla loro semplicità, sono stati utilizzati per la progettazione del sistema di controllo. Il codice CarMa0 è stato usato per sviluppare modelli analoghi ma con una rappresentazione tridimensionale delle strutture conduttrici; questi permettono di verificare l’accuratezza dei modelli semplificati e indagare l’importanza delle strutture tridimensionali sulla dinamica del sistema. Il codice CarMa0NL ha permesso la trattazione di fenomeni evolutivi nel tempo e nonlineari (e.g. disruzioni, transizioni limiter-divertor, transizioni L-H etc.). L’attività può essere suddivisa in due parti: la prima riguarda la modellizzazione di plasmi a basso β teorici, non ottenuti sperimentalmente, usati come riferimento per la progettazione e l’implementazione del sistema di controllo della forma e della posizione verticale del plasma; la seconda parte, è legata ai risultati delle campagne sperimentali sui plasmi a sezione non circolari in diversi regimi, dal basso β al modo H, con particolare attenzione allo sviluppo di un nuovo modello linearizzato di risposta di plasma per i nuovi regimi di equilibrio raggiunti. L’attività di ricerca è caratterizzata da molteplici problematiche e peculiarità sia in termini di modellazione che di controllo. La pronunciata non circolarità della forma di plasma e i diversi regimi coinvolti hanno influenzato fortemente l’attività di modellazione che ha richiesto, infatti, lo sviluppo di molteplici strumenti computazionali e di analisi dati. Per quanto concerne il controllo, la non completa osservabilità della dinamica del sistema e la necessità di ridurre l’ordine del modello sono solo alcuni degli aspetti che hanno determinato la progettazione del sistema di controllo di forma e di posizione verticale. La prima parte è basata su dati teorici generati dal codice di equilibrio MAXFEA e poi utilizzati per derivare il modello linearizzato attraverso il codice CREATE-L. In questo contesto, sono stati prodotti due modelli di riferimento per le configurazioni magnetiche relative a plasmi non circolari: il singolo nullo inferiore (LSN) e il singolo nullo superiore (USN). I modelli CREATE-L sono i più semplici in termini di complessità di modellazione, in quanto le strutture conduttive della macchina sono descritte nell’approssimazione assialsimmetrica. D’altro canto, le proprietà semplici ma affidabili del modello CREATE-L hanno portato alla progettazione del sistema di controllo di forma e posizione verticale del plasma di RFX-mod, che è stato in seguito testato e utilizzato con successo per aumentare le prestazioni del plasma. Successivamente, è stata condotta un’analisi sui possibili effetti 3D delle strutture conduttrici sulle due configurazioni di plasma di riferimento, producendo dunque modelli linearizzati caratterizzati da un sempre maggiore livello di complessità. Una dettagliata descrizione volumetrica (3D) delle strutture conduttrici di RFX-mod è stata eseguita e inclusa nei modelli linearizzati di plasma attraverso il codice CarMa0. Successivamente, è stato eseguito un confronto tra l’accuratezza di questo modello e quello precedente 2D. Le diverse ipotesi e approssimazioni dei vari modelli consentono una chiara identificazione dei fenomeni chiave che governano l’evoluzione dell’instabilità verticale n = 0 in scariche RFX-mod tokamak e quindi forniscono informazioni fondamentali nella pianificazione ed esecuzione di esperimenti correlati oltre che nella raffinazione del progetto del sistema di controllo. Infine, il modello di equilibrio evolutivo non lineare CarMa0NL, che comprende le strutture volumetriche 3D, è stato utilizzato per modellare gli effetti non lineari simulando una variazione di corrente lineare "fittizia". La seconda parte è costituita da un’attività di modellazione strettamente correlata ai risultati delle campagne sperimentali. In particolare, sono stati eseguiti nuovi modelli linearizzati per i plasmi sperimentali nella configurazione USN per tutti i regimi di plasma coinvolti, cioè dal basso β fino al modo H. È stata ideata e sviluppata una procedura iterativa per la produzione di modelli linearizzati di risposta di plasma estremamente accurati, al fine di riprodurre al meglio i dati sperimentali. I nuovi modelli hanno consentito ulteriori studi sulla stabilità verticale, inclusi gli effetti della parete 3D, nei tre diversi regimi studiati (basso β, β intermedio, modo H). I modelli linearizzati assialsimmetrici (CREATE-L) sono stati analizzati dal punto di vista della teoria dei controlli, rilevando caratteristiche peculiari in termini di funzione di trasferimento SISO associata al controllo della stabilità verticale e in termini di modello completo MIMO relativo al controllo di forma. Il modello MIMO è stato utilizzato per indagare le oscillazioni nella forma del plasma osservate sperimentalmente in alcune scariche a β intermedio. L’evoluzione temporale non lineare della scarica di plasma, per plasmi sperimentali a regimi a basso β, è stata effettuata usando il codice di equilibrio evolutivo CarMa0NL. Infine, è stata studiata l’instabilità verticale per i plasmi sperimentali in termini di un possibile rapporto tra i parametri del plasma e il suo verificarsi; a tal fine è stata eseguita la soluzione del problema inverso per la produzione di equilibri di plasma teorici di riferimento, prodotti come variazioni sui parametri dei plasmi osservati sperimentalmente, il che comporta una vasta gamma di metodi numerici descritti in dettaglio. Successivamente, è stato adottato un test di ipotesi statistica per confrontare i valori medi dei parametri di plasma, sia sperimentali che teorici, associati a due diversi comportamenti in termini di stabilità verticale.

Philosophiae Doctor - PhD
Singularly perturbed problems have been studied extensively over the past few years from different perspectives. The recent research has focussed on the problems whose solutions possess interior layers. These interior layers appear in the interior of the domain, location of which is difficult to determine a-priori and hence making it difficult to investigate these problems analytically. This explains the need for approximation methods to gain some insight into the behaviour of the solution of such problems. Keeping this in mind, in this thesis we would like to explore a special class of numerical methods, namely, fitted finite difference methods to determine reliable solutions. As far as the fitted finite difference methods are concerned, they are grouped into two categories: fitted mesh finite difference methods (FMFDMs) and the fitted operator finite difference methods (FOFDMs). The aim of this thesis is to focus on the former. To this end, we note that FMFDMs have extensively been used for singularly perturbed two-point boundary value problems (TPBVPs) whose solutions possess boundary layers. However, they are not fully explored for problems whose solutions have interior layers. Hence, in this thesis, we intend firstly to design robust FMFDMs for singularly perturbed TPBVPs whose solutions possess interior layers and to improve accuracy of these approximation methods via methods like Richardson extrapolation. Then we extend these two ideas to solve such singularly perturbed TPBVPs with variable diffusion coefficients. The overall approach is further extended to parabolic singularly perturbed problems having constant as well as variable diffusion coefficients.
2023-08-31

L'utilisation de l'holographie acoustique de champ proche dans des situations industrielles (champ perturbe) est un probleme mal pose, au sens de l'instabilite inherente de la methode en probleme inverse et de la negligence du bruit dans les traitements. Pour permettre son implantation par la methode de fourier, nous avons considere le probleme dans le cadre de la theorie de la regularisation. Les regularisations proposees ont effectue des filtrages selectifs en nombres d'onde. Ainsi, les instabilites causees par le bruit (mesure ou numerique) ont ete maitrisees. De plus, certaines approches developpees (filtrage de tikhonov avec contraintes spatiales, methodes d'inversion iterative) ont l'avantage de ne pas requerir une exploitation supervisee. Des traitements incorporant des techniques de regularisation ont ensuite ete developpes pour reduire le bruit dans les composantes spectrales et corriger ainsi les mesestimations residuelles du champ de pression complexe reconstruit a la surface de la source. Les techniques de reduction de bruit se sont appuyees sur des estimations spectrales et spatiales par la methode de capon. L'incorporation des estimations a ete effectuee dans des filtrages de wiener ou par un couplage a des techniques de restauration de phase. Une reduction du bruit a aussi ete pratiquee par identification des termes de rayonnement de la source dans un modele simple exploitant des mesures double-couche. Des mesures (en champ libre ou perturbe) autour d'une coque rayonnante cylindrique, ont permis de valider ces traitements d'imagerie acoustique avec extrapolation spatiale

>Magister Scientiae - MSc
With the development of technology seen in the last few decades, numerous solvers have been developed to provide adequate solutions to the problems that model different aspects of science and engineering. Quite often, these solvers are tailor-made for specific classes of problems. Therefore, more of such must be developed to accompany the growing need for mathematical models that help in the understanding of the contemporary world. This thesis treats two point boundary value singularly perturbed problems. The solution to this type of problem undergoes steep changes in narrow regions (called boundary or internal layer regions) thus rendering the classical numerical procedures inappropriate. To this end, robust numerical methods such as finite difference methods, in particular fitted mesh and fitted operator methods have extensively been used. While the former consists of transforming the continuous problem into a discrete one on a non-uniform mesh, the latter involves a special discretisation of the problem on a uniform mesh and are known to be more accurate. Both classes of methods are suitably designed to accommodate the rapid change(s) in the solution. Quite often, finite difference methods on piece-wise uniform meshes (of Shishkin-type) are adopted. However, methods based on such non-uniform meshes, though layer-resolving, are not easily extendable to higher dimensions. This work aims at investigating the possibility of capitalising on the advantages of both fitted mesh and fitted operator methods. Theoretical results are confirmed by extensive numerical simulations.

En optimisation de formes, de nombreux résultats ont déjà été obtenus dans le
cas de domaines à frontière régulière et pour des perturbations régulières de ces domaines.
Par contre, l'étude de domaines non-réguliers, tels que des domaines fissurés par exemple,
et l'étude de perturbations singulières telles que la création d'un trou dans un domaine est
plus récente et plus complexe. Ce nouveau domaine de recherche est motivé par de multiples
applications, car en pratique, les hypothèses de régularité ne sont pas toujours vérifiées. Les
outils tels que la dérivée topologique permettent d'appréhender ces perturbations singulières
de domaines et leur utilisation est maintenant fréquente.

Dans la première partie, nous étudions la structure de la dérivée de forme pour des domaines fissurés. Dans le cas d'un ouvert régulier, de classe C1 ou lipschitzien par exemple,
la dérivée dépend uniquement des perturbations de la frontière du domaine en direction de
la normale. Ce théorème de structure n'est plus valable pour des domaines contenant des
fissures. On généralise ici ce théorème de structure aux domaines fissurés en dimension quelconque pour les dérivées premières et secondes. En dimension deux, on retrouve le résultat
usuel, à savoir qu'en plus du terme classique, deux nouvelles contributions apparaissent dûes
aux extrémités de la fissure. En dimension supérieure, un nouveau terme apparaît en plus du
terme classique, dû à la frontière de la variété à bord représentant la fissure.

Dans la deuxième partie, nous étudions la perturbation singulière d'un domaine et nous
modélisons cette perturbation à l'aide d'extensions auto-adjointes d'opérateurs. Nous décrivons cette modélisation, puis nous montrons comment elle peut être utilisée pour un problème
d'optimisation de forme. En définissant une fonctionnelle d'énergie approchée pour ce problème modèle, on retrouve notamment la formule de la dérivée topologique usuelle.

Dans la troisième partie, on propose une application numérique de la dérivée topologique
et de la dérivée de forme pour un problème non-linéaire. On cherche à maximiser l'énergie
associée à la solution d'un problème de Signorini dans un domaine ­. L'évolution du domaine
est représentée à l'aide d'une méthode levelset.

Philosophiae Doctor - PhD
Numerical approximations of multiscale problems of important applications in ecology are investigated. One of the class of models considered in this work are singularly perturbed (slow-fast) predator-prey systems which are characterized by the presence of a very small positive parameter representing the separation of time-scales between the fast and slow dynamics. Solution of such problems involve multiple scale phenomenon characterized by repeated switching of slow and fast motions, referred to as relaxationoscillations, which are typically challenging to approximate numerically. Granted with a priori knowledge, various time-stepping methods are developed within the framework of partitioning the full problem into fast and slow components, and then numerically treating each component differently according to their time-scales. Nonlinearities that arise as a result of the application of the implicit parts of such schemes are treated by using iterative algorithms, which are known for their superlinear convergence, such as the Jacobian-Free Newton-Krylov (JFNK) and the Anderson’s Acceleration (AA) fixed point methods.

Le sujet de cette thèse est le développement de méthodes itératives permettant la résolution degrands systèmes linéaires creux d'équations présentant plusieurs seconds membres simultanément. Ces méthodes seront en particulier utilisées dans le cadre d'une application géophysique : la migration sismique visant à simuler la propagation d'ondes sous la surface de la terre. Le problème prend la forme d'une équation d'Helmholtz dans le domaine fréquentiel en trois dimensions, discrétisée par des différences finies et donnant lieu à un système linéaire creux, complexe, non-symétrique, non-hermitien. De plus, lorsque de grands nombres d'onde sont considérés, cette matrice possède une taille élevée et est indéfinie. Du fait de ces propriétés, nous nous proposons d'étudier des méthodes de Krylov préconditionnées par des techniques hiérarchiques deux niveaux. Un tel pre-conditionnement s'est montré particulièrement efficace en deux dimensions et le but de cette thèse est de relever le défi de l'adapter au cas tridimensionel. Pour ce faire, des méthodes de Krylov sont utilisées à la fois comme lisseur et comme méthode de résolution du problème grossier. Ces derniers choix induisent l'emploi de méthodes de Krylov dites flexibles
The topic of this PhD thesis is the development of iterative methods for the solution of large sparse linear systems of equations with possibly multiple right-hand sides given at once. These methods will be used for a specific application in geophysics - seismic migration - related to the simulation of wave propagation in the subsurface of the Earth. Here the three-dimensional Helmholtz equation written in the frequency domain is considered. The finite difference discretization of the Helmholtz equation with the Perfect Matched Layer formulation produces, when high frequencies are considered, a complex linear system which is large, non-symmetric, non-Hermitian, indefinite and sparse. Thus we propose to study preconditioned flexible Krylov subspace methods, especially minimum residual norm methods, to solve this class of problems. As a preconditioner we consider multi-level techniques and especially focus on a two-level method. This twolevel preconditioner has shown efficient for two-dimensional applications and the purpose of this thesis is to extend this to the challenging three-dimensional case. This leads us to propose and analyze a perturbed two-level preconditioner for a flexible Krylov subspace method, where Krylov methods are used both as smoother and as approximate coarse grid solver

Se obtienen aproximaciones numéricas a tres leyes de conservación usando programación matemática. Para ello se aplica el algoritmo de Seneta-Steiger para obtener soluciones l1 óptimas de un sistema algebraico nx(n- 1)

Nous considérons des systèmes d'équations différentielles stochastiques faisant intervenir deux échelles de temps bien distinctes. Nous commençons par établir, dans un cadre général, des propriétés de concentration des trajectoires au voisinage des variétés lentes du système déterministe correspondant. Nous étudions ensuite la dynamique au voisinage de points de bifurcation de la variété lente, en particulier dans le cas d'une bifurcation noeud-col et d'une bifurcation fourche. Les phénomènes apparentées de la résonance stochastique et de l'hystérésis dynamique sont également étudiés en détail. Finalement, nous dérivons la loi des temps de passage à travers une orbite périodique instable, pour une famille d'équations qui ne sont pas limitées au cas d'échelles de temps distinctes.

In the field of singularly perturbed reaction- or convection-diffusion boundary value problems the research area of a priori error analysis for the finite element method, has already been thoroughly investigated. In particular, for mesh adapted methods and/or various stabilization techniques, works have been done that prove optimal rates of convergence or supercloseness uniformly in the perturbation parameter epsilon. Commonly, however, it is assumed that the exact solution behaves nicely in that it obeys certain regularity assumptions although in general, e.g. due to corner singularities, these regularity requirements are not satisfied. So far, insufficient regularity has been met by assuming compatibility conditions on the data. The present thesis originated from the question: What can be shown if these rather unrealistic additional assumptions are dropped? We are interested in epsilon-uniform a priori estimates for convergence and superconvergence that include some regularity parameter that is adjustable to the smoothness of the exact solution. A major difficulty that occurs when seeking the numerical error decay is that the exact solution is not known. Since we strive for reliable rates of convergence we want to avoid the standard approach of the "double-mesh principle". Our choice is to use reference solutions as a substitute for the exact solution. Numerical experiments are intended to confirm the theoretical results and to bring further insights into the interplay between layers and singularities. To computationally realize the thereby arising demanding practical aspects of the finite element method, a new software is developed that turns out to be particularly suited for the needs of the numerical analyst. Its design, features and implementation is described in detail in the second part of the thesis.

Nous avons étudié plusieurs aspects dynamiques et photométriques des disques planétaires et protoplanétaires perturbés. Dans la première partie, à l'aide de modèle numérique simple, nous avons étudié l'évolution thermodynamique d'un disque protoplanétaire composé de planétésimaux subissant des collisions physiques inélastiques, au voisinage d'un embryon de planète géante (15 masses terrestres). Dès l'apparition de l'embryon, un transfert de chaleur se met en place dans le disque, augmentant fortement les vitesses relatives dans ce dernier, sur une région s'étendant sur plusieurs unités astronomiques. L'évolution de ce mécanisme transitoire a été étudiée sur des temps longs (un million d'années) et pour une vaste gamme de masses du perturbateur. C'est un mécanisme générique qui pourrait avoir profondément affecté le processus de formation des planètes, aussi bien telluriques que géantes. Les conséquences sur la formation de la ceinture d'asteroides sont discutées, ainsi que l'effet possible de la fragmentation des planétésimaux, qui n'a pu être prise en compte dans le modèle numérique. La deuxieme partie de cette thèse est une étude photométrique de l'anneau F de saturne, qui est perturbé par ses deux satellites gardiens. En étudiant un ensemble de plus de 300 images prises au télescope CFH, nous avons mis en évidence la présence de structures allongées, dont l'origine est toujours mal connue. En combinant nos données et celles du télescope spatial, nous avons établi avec précision une nouvelle orbite de l'anneau F, a 14006060 km, soit plus faible de 150 km par rapport à celle déterminée en 1980-81. Ceci pourrait être le signe qu'une importante restructuration radiale de l'anneau F a dû avoir lieu entre 1980 et 1995.

Submitted by Claudinei Pracidelli (cpracide@ipen.br) on 2015-03-17T10:41:16Z No. of bitstreams: 0
Made available in DSpace on 2015-03-17T10:41:16Z (GMT). No. of bitstreams: 0
Tese (Doutorado em Tecnologia Nuclear)
IPEN/T
Instituto de Pesquisas Energeticas e Nucleares - IPEN-CNEN/SP

U tezi se istražuje uniformna konvergencija Galerkinovog postupka konačnih elemenata na mrežama različitog tipa za dvoparametarske singularno perturbovane probleme.Uvedene su slojno-adaptivne mreže za probleme konvekcije-reakcije-difuzije:  Bahvalovljeva, Duran-Šiškinova i Duranova za jednodimenzionalni i Duran-Šiškinova i Duranova mreža za dvodimenzionalni problem. Za pomenute probleme na svim ovim mrežama analizirane su greške interpolacije, diskretizacije i greška u energetskoj normi i dokazana je uniformna konvergencija Galerkinovog postupka konačnih elemenata. Sva teorijska tvrđenja su potvrđena numeričkim eksperimentima. 
The thesis explores the uniform convergence for Galerkin nite elementmethod on various meshes for two parameter singularly perturbed problems.Layer-adapted meshes are introduced for convection-reaction-diusionproblems: Bakhvalov, Duran-Shishkin and Duran meshes for a one dimensionaland Duran-Shishkin and Duran meshes for a two dimensional problem.We analyze the errors of interpolation, discretization and error in the energynorm and prove the parameter uniform convergence for Galerkin nite elementmethod on mentioned meshes. Numerical experiments support theoreticalndings. 

This Dissertation is devoted to the singular perturbation and homogenization analysis of boundary value problems in the periodically perforated Euclidean space. We investigate the behaviour of the solutions of boundary value problems for the Laplace, the Poisson, and the Helmholtz equations, as parameters related to diameter of the holes or the size of the periodicity cells tend to 0. The Dissertation is organized as follows. In Chapter 1, we present two known constructions of a periodic analogue of the fundamental solution of the Laplace equation and we introduce the periodic layer and volume potentials for the Laplace equation and some basic results of periodic potential theory. Chapter 2 is devoted to singular perturbation and homogenization problems for the Laplace and the Poisson equations with Dirichlet and Neumann boundary conditions. In Chapter 3 we consider the case of (linear and nonlinear) Robin boundary value problems for the Laplace equation, while in Chapter 4 we analyze (linear and nonlinear) transmission problems. In Chapter 5 we apply the results of Chapter 4 in order to prove the real analyticity of the effective conductivity of a periodic dilute composite. Chapter 6 is dedicated to the construction of a periodic analogue of the fundamental solution of the Helmholtz equation and of the corresponding periodic layer potentials. In Chapter 7 we collect some results of spectral theory for the Laplace operator in periodically perforated domains. In Chapter 8 we investigate singular perturbation and homogenization problems for the Helmholtz equation with Neumann boundary conditions. In Chapter 9 we consider singular perturbation and homogenization problems with Dirichlet boundary conditions for the Helmholtz equation, while in Chapter 10 we study (linear and nonlinear) Robin boundary value problems. Chapter 11 is devoted to the study of periodic layer potentials for general second order differential operators with constant coefficients. At the end of the Dissertation we have enclosed some Appendices with some results that we have exploited.
Questa Tesi è dedicata all'analisi di problemi di perturbazione singolare e omogeneizzazione nello spazio Euclideo periodicamente perforato. Studiamo il comportamento delle soluzioni di problemi al contorno per le equazioni di Laplace, di Poisson e di Helmholtz al tendere a 0 di parametri legati al diametro dei buchi o alla dimensione delle celle di periodicità. La Tesi è organizzata come segue. Nel Capitolo 1, presentiamo due costruzioni note di un analogo periodico della soluzione fondamentale dell'equazione di Laplace, e introduciamo potenziali di strato e di volume periodici per l'equazione di Laplace e alcuni risultati basilari di teoria del potenziale periodica. Il Capitolo 2 è dedicato a problemi di perturbazione singolare e omogeneizzazione per le equazioni di Laplace e Poisson con condizioni al bordo di Dirichlet e Neumann. Nel Capitolo 3 consideriamo il caso di problemi al contorno di Robin (lineari e nonlineari) per l'equazione di Laplace, mentre nel Capitolo 4 analizziamo problemi di trasmissione (lineari e nonlineari). Nel Capitolo 5 applichiamo i risultati del Capitolo 4 al fine di provare l'analiticità della conduttività effettiva di un composto periodico. Il Capitolo 6 è dedicato alla costruzione di un analogo periodico della soluzione fondamentale dell'equazione di Helmholtz e dei corrispondenti potenziali di strato. Nel Capitolo 7 raccogliamo alcuni risultati di teoria spettrale per l'operatore di Laplace in domini periodicamente perforati. Nel Capitolo 8 studiamo problemi di perturbazione singolare e di omogeneizzazione per l'equazione di Helmholtz con condizioni al contorno di Neumann. Nel Capitolo 9 consideriamo problemi di perturbazione singolare e di omogeneizzazione con condizioni al contorno di Dirichlet per l'equazione di Helmholtz, mentre nel Capitolo 10 studiamo problemi al contorno di Robin (lineari e nonlineari). Il Capitolo 11 è dedicato allo studio di potenziali di strato periodici per operatori differenziali generali del secondo ordine a coefficienti costanti. Alla fine della Tesi abbiamo incluso delle Appendici con alcuni risultati utilizzati.

L'objet de cette thèse est la résolution de problèmes elliptiques dans différents domaines non bornés. Dans un premier temps, nous étudions l'opérateur de Laplace dans un domaine extérieur avec des conditions aux limites non homogènes mêlées, puis dans un domaine extérieur dans le demi-espace avec des conditions de type Dirichlet, Neumann et mêlées. Nous considérons ensuite le problème de Stokes dans trois géométries non bornées: un domaine extérieur dans le demi-espace, un demi-espace perturbé et un domaine avec ouverture. Nous donnons pour chacun de ces problèmes des résultats fondamentaux d'existence et d'unicité en théorie L^p (avec p strictement compris entre 1 et l'infini) dans le cadre fonctionnel des espaces de Sobolev avec poids. De plus, nous nous intéressons également aux cas des solutions fortes (avec en particulier des résultats de régularité) et aux cas des solutions très faibles.