This thesis develops two families of numerical methods, based upon rational approximations having distinct real poles, for solving first- and second-order parabolic/ hyperbolic partial differential equations. These methods are thirdand fourth-order accurate in space and time, and do not require the use of complex arithmetic. In these methods first- and second-order spatial derivatives are approximated by finite-difference approximations which produce systems of ordinary differential equations expressible in vector-matrix forms. Solutions of these systems satisfy recurrence relations which lead to the development of parallel algorithms suitable for computer architectures consisting of three or four processors. Finally, the methods are tested on advection, advection-diffusion and wave equations with constant coefficients.

The production of long, thin polymeric fibers is a main objective of the textile industry. Melt-spinning is a particularly simple and effective technique. In this work, we shall discuss the equations of melt-spinning in viscous and viscoelastic flow. These quasilinear hyperbolic equations model the uniaxial extension of a fluid thread before its solidification. We will address the following topics: first we shall prove existence, uniqueness, and regularity of solutions. Our solution strategy will be developed in detail for the viscous case. For non-Newtonian and isothermal flows, we shall outline the general ideas. Our solution technique consists of energy estimates and fixed-point arguments in appropriate Banach spaces. The existence result for a simple transport equation is the key to understanding the quasilinear case. The second issue of this exposition will be the stability of the unforced frost line formation. We will give a rigorous justification that, in the viscous regime, the linearized equations obey the ``Principle of Linear Stability''. As a consequence, we are allowed to relate the stability of the associated strongly continuous semigroup to the numerical resolution of the spectrum of its generator. By using a spectral collocation method, we shall derive numerical results on the eigenvalue distribution, thereby confirming prior results on the stability of the steady-state solution. / Ph. D.

The main objective of this research work is to apply the discontinuous Galerkin method to a classical partial differential equation to investigate the properties of the numerical solution and compare the numerical solution to the analytical solution by using discontinuous Galerkin method. This scheme is applied to 1-D non-linear conservation equation (Burgers equation) in which the governing differential equation is simplified model of the inviscid Navier-stokes equations. In this work three cases are studied. They are sinusoidal wave profile, initial shock discontinuity and initial linear distribution. A grid and time step refinement is performed. Riemann fluxes at each element interfaces are calculated. This scheme is applied to forward differentiation method (Euler's method) and to second order Runge-kutta method of this work.

Under adequate conditions on $g$, we show the density in $L^2((0,\pi),(0,2\pi))$ of the set of functions $p$ for which \begin{equation*} u_{tt}(x,t)-u_{xx}(x,t)= g(u(x,t)) + p(x,t) \end{equation*} has a weak solution subject to \begin{equation*} \begin{aligned} u(x,t)&=u(x,t+2\pi)\\ u(0,t)&=u(\pi,t)=0. \end{aligned} \end{equation*} To achieve this, we prove a Saddle Point Principle by means of a refined variant of the deformation lemma of Rabinowitz. Generally, inf-sup techniques allow the characterization of critical values by taking the minimum of the maximae on some particular class of sets. In this version of the Saddle Point Principle, we introduce sufficient conditions for the existence of a saddle-structure which is not restricted to finite-dimensional subspaces.

No description available.

Submitted by Jorge Silva (jorgelmsilva@ufrrj.br) on 2017-10-17T16:30:00Z No. of bitstreams: 1 2016 - Maur?cio de Carvalho.pdf: 8945176 bytes, checksum: 4b68e7d4395ffbaee8b4d17639d8a28e (MD5) / Made available in DSpace on 2017-10-17T16:30:00Z (GMT). No. of bitstreams: 1 2016 - Maur?cio de Carvalho.pdf: 8945176 bytes, checksum: 4b68e7d4395ffbaee8b4d17639d8a28e (MD5) Previous issue date: 2016-04-12 / CAPES / In this work we consider the injection of water with dissolved ions into a linear horizontal porous rock cylinder with constant porosity and absolute permeability initially containing oil and water in several proportions. The water is assumed to have low salinity concentration, where some ions are dissolved. We disregard that there is in the rocks some possible minerals that can dissolve or precipitate in water phase. There are two chemical fluid components as well as two immiscible phases: water and oil, (w, o). The dissolved ions are: positive divalent ions: calcium ions, Ca2+ and magnesium ions, Mg2+; negative divalent ions: sulphate ions, SO42?; positive monovalent ions: sodium ions, Na+; negative monovalent ions: cloride ions, Cl?. The cations are modeled to be involved in fast ion exchange process with a surface negative X? which can absorb the positive ions, Ca2+, Mg2+ and Na+. We use simple mixing rules and we disregard any heat of precipitation/dissolution of substance reactions or ion desorption. Moreover we disregard any volume contraction efects resulting from mixing and reactions in any phase. We are going to solve in this work, the Riemann problem and we are going to discuss some features about the studied model. / Neste trabalho consideramos a inje??o de ?gua com ?ons dissolvidos em um meio po-roso linear horizontal cil?ndrico com porosidade e permeabilidade absoluta constantes, inicialmente, contendo ?leo e ?gua em v?rias propor??es. A ?gua ? assumida ter baixa concentra??o de sais, onde alguns ?ons est?o dissolvidos. Desconsideramos a exist?ncia de alguns poss?veis minerais na rocha que possam dissolver ou precipitar na fase da ?gua. Existem dois componentes qu?micos fluidos assim como duas imisc?veis fases: ?gua e ?leo,(w, o). Os ?ons dissolvidos s?o: ?ons divalentes positivos: ?ons c?lcio, Ca2+ e ?ons magn?sio, Mg2+; ?ons negativos divalentes: ?ons sulfato, SO42?; ?ons positivos monovalentes: ?ons s?dio, Na+; ?ons negativos monovalentes: ?ons cloro, Cl?. Os c?tions est?o envolvidos em um processo r?pido de troca de ?ons com a superf?cie do meio poroso carregada eletronega-tivamente X?, onde o meio absorver? os ?ons positivos Ca2+, Mg2+ e Na+. Usando regras simples de misturas e desconsiderando qualquer calor de precipita??o ou dissolu??o de rea??es de subst?ncias ou dessor??o de ?ons. Al?m disso, desconsideramos quaisquer efeitos de contra??o de volume resultante das misturas e rea??es em qualquer fase. Resolveremos neste trabalho, o Problema de Riemann e discutiremos algumas caracter?sticas do modelo estudado.

Les mélanges multiphasiques en déséquilibre de vitesse sont habituellement modélisés à l'aide d'un modèle à 6 ou 7 équations (Baer and Nunziato, 1986). Ces modèles sont très efficaces pour traiter des mélanges avec effets d'interpénétration. Ils peuvent aussi être utilisés pour traiter des problèmes à interface dans lesquels il est nécessaire de respecter les conditions d'interface (continuité de la vitesse normale et de la pression). Ceci est réalisé à l'aide de solveurs de relaxation mécanique (Saurel and Abgrall, 1999). Une autre méthode consiste à utiliser un modèle à une vitesse et une pression (Kapila et al., 2001). Cependant, de nombreuses applications font intervenir des interfaces instables entre fluides. On traite habituellement ces zones de mélanges turbulents en utilisant un modèle à une vitesse et en résolvant spatialement les diverses instabilités. Dans de nombreuses applications cela devient impossible en raison du trop grand nombre de « jets » et de « bulles ». De plus, on rencontre des difficultés numériques y compris pour le calcul d'une instabilité isolée (Liska and Wendroff, 2004). Dans ce manuscrit, nous abordons le problème de la modélisation des zones de mélange avec des modèles multiphasiques. Cela pose un sérieux problème de modélisation pour des écoulements évoluant d'une situation où l'interface est bien définie (une seule vitesse) vers une configuration de mélange de fluides à plusieurs vitesses. Cette question a été abordée par Besnard and Harlow (1988), Youngs et al. (1989), Chen et al. (1996), Glimm et al. (1999), Saurel et al. (2003) par exemple. / Multiphase mixtures with velocity disequilibrium are usually modelled with 6 or 7 equations models (Baer and Nunziato, 1986). These models are very efficient to model mixtures with velocity drift effects. They can also be used to model interfacial flows where the respect of interface conditions (continuous normal velocity and pressure) is mandatory. Such aim is usually achieved with the help of stiff mechanical relaxation solvers (Saurel and Abgrall, 1999). Another option is to use single pressure and single velocity models (Kapila et al., 2001). However, many applications involve unstable fluid-fluid interfaces for which flow conditions range from well separated fluids to fully mixed ones. The usual way to deal with these turbulent mixing zones is to use a single velocity flow model and to resolve spatially the various instabilities. However, spatial resolution of these instabilities in many applications is impossible as too many ‘jets' and ‘bubbles' are present. Also, numerical difficulties and large inaccuracies are present even for an isolated instability computation (Liska and Wendroff, 2004). In this work, we address the issue of mixing zone modelling with multiphase flow models. This poses the serious difficulty of model derivation for flows conditions ranging from well defined interfaces (single velocity) to fluid mixtures evolving with several velocities. This issue has been addressed by Besnard and Harlow (1988), Youngs et al. (1989), Chen et al. (1996), Glimm et al. (1999), Saurel et al. (2003) to cite a few. In Saurel et al. (2010) an extension of the Kapila et al. (2001) model was done to deal with permeation effects through material interfaces.

The thesis deals with the numerical solution of partial differential equati- ons describing the flow of the so-called shallow water neglecting the flow in the vertical direction. These equations are of hyperbolical type of the first or- der with a reactive term representing the bottom topology. We discretize the resulting system of equations by the implicit space-time discontinuous Ga- lerkin method (STDGM). In the literature, the explicit techniques are used most of the time. The implicit approach is suitable especially for adaptive methods, because it allows the usage of different meshes for different time niveaus. In the thesis we derive the corresponding method and an adaptive algorithm. Finally, we present usage of the method in several examples. 1

On dérive un nouveau modèle d’écoulements cisaillés et turbulents d’eau peu profonde. Les écarts de la vitesse horizontale par rapport à sa valeur moyenne sont pris en compte par une nouvelle variable appelée enstrophie, liée à la vorticité et à l’énergie turbulente. Le modèle comporte trois équations qui sont les bilans de masse, de quantité de mouvement et d’énergie. Le modèle est hyperbolique et peut être écrit sous forme conservative. L’énergie turbulente, dont l’intensité peut être importante, est produite par les ondes de choc qui apparaissent naturellement dans le modèle. Les écoulements rapidement variés étudiés sont caractérisés par l’existence d’une structure turbulente appelée rouleau dans laquelle la dissipation d’énergie turbulente joue un rôle majeur. Cette dissipation, qui détermine notamment le profil de profondeur, est modélisée par l’introduction d’un terme nouveau dans le bilan d’énergie. Le modèle comporte deux paramètres. L’un gouverne la dissipation de l’énergie turbulente du rouleau. L’autre paramètre, l’enstrophie de paroi, liée au cisaillement sur le fond, peut être considéré comme constant dans la partie rapidement variée d’un écoulement, sur laquelle il exerce une influence assez faible. Ce modèle a été appliqué avec succès aux vagues des trains de rouleaux et au ressaut hydraulique classique. Le profil de la surface libre est en très bon accord avec les résultats expérimentaux. L’étude numérique en régime non stationnaire permet notamment de prédire le régime oscillatoire du ressaut hydraulique. La fréquence d’oscillations correspondante est en accord satisfaisant avec les mesures expérimentales de la littérature. / We derive a new model of turbulent shear shallow water flows. The deviation of the horizontal velocity from its average value is taken into account by a new variable called enstrophy, which is related to the vorticity and to the turbulent energy. The model consists of three equations which are the balances of mass, momentum and energy. The model is hyperbolic and can be written in conservative form. The turbulent energy, which can be of high intensity, is produced in shock waves which appear naturally in the model. The rapidly varied flows we studied are characterized by the presence of a turbulent structure called roller in which the turbulent energy dissipation plays a major part. This dissipation, which determines, in particular, the depth profile, is modelled by the introduction of a new term in the energy balance equation. The model contains two parameters. The first one governs the dissipation of the turbulent energy of the roller. The second one, the wall enstrophy, related to the shearing at the bottom, can be considered as constant in the rapidly varied part of the flow on which it does not exert an important influence. This model was successfully applied to roll waves and to the classical hydraulic jump. The free surface profile was found in very good agreement with the experimental results. The numerical study in the non-stationary case can notably predict the oscillations of the hydraulic jump. The corresponding oscillation frequency is in good agreement with the experimental measures found in the literature.

Interaction of fluid with elastic tube is complicated issue studied by many scientific departments around the world. Object of this thesis is to analyze simplified one-dimensional model. At the beginning, used balance equations and basics of hyper-elasticity are presented. Then we review three most common materials used for the description of blood vessels and other soft tissues. For these materials we introduce a method which we use to derive a relation between tube deformation and transmural pressure (i.e. difference between inner and outer pressure). In mathematical section we give brief review of theory of nonlinear hyperbolic equations and some relatively new results in the field of existence and uniqueness of a solution of one-dimensional hyperbolic system. The "building stone" of these results is a solution of the so-called Riemann problem. We use a method for finding exact solutions to the Riemann problem to analyze studied model of fluid-tube interaction and study dependence of the qualitative behavior of the solution on the material properties of the tube wall.