The first part of this thesis deals with detonations in gaseous reactive mixtures. Various technological applications have been proposed involving detonations, particularly in the field of propulsion. However, it has been confirmed experimentally that detonations generally exhibit an unstable behaviour, leading to complicated flow structures. A thorough understanding of the evolution of detonation waves is needed before they can be used for propulsion purposes. Herein, we present the first detailed numerical study of three-dimensional structures in gaseous detonations. This study is based on a parallelized, unsplit, shock-capturing algorithm. We show that we can reproduce all types of detonations that have been observed experimentally. The advancements in the field of gaseous compressible reactive flows paved the way for the study of the significantly more complex phenomena that occur in the flow of two-phase, heterogeneous compressible reactive mixtures. In the second part of this thesis, we develop a new shock-capturing algorithm for the study of these flows. We first present a new numerical procedure for solving exactly the Riemann problem of compressible two-phase flow models containing non-conservative products. We then examine the accuracy and robustness of three known methods for the integration of the non-conservative products. The issue of existence and uniqueness of solutions to the Riemann problem is also discussed. Due to the ill-posedness of the Riemann problem of standard two-phase models, we present and analyze, in the third and last part of this work, a conservative approximation to reduced one-pressure one-velocity models for compressible two-phase flows that contain non-conservative products. Herein, we develop an exact Riemann solver for the proposed reduced model. Further, we investigate the structure of the steady two-phase detonation waves admitted by this model. Finally, we report on numerical simulations of the transmission of a purely gaseous detonation to heterogeneous mixtures. The effect of the solid particles on the structure of the resulting two-phase detonation is discussed in detail.

In this thesis we study the central limit theorem (CLT) for nonuniformly hyperbolic dynamical systems. We examine cases in which polynomial decay of correlations leads to a CLT with a non-standard scaling factor of √ n ln n. We also formulate an explicit expression for the the diffusion constant σ in situations where a return time function on the system is a certain class of supermartingale. We then demonstrate applications by exhibiting the CLT for the return time function in four classes of dynamical billiards, including one previously unproven case, the skewed stadium, as well as for the linked twist map. Finally, we introduce a new class of billiards which we conjecture are ergodic, and we provide numerical evidence to support that claim.

Dans cette thèse nous avons étudié différents aspects pour le contrôle de systèmes hyperboliques.Tout d'abord, nous nous sommes intéressés à des systèmes hyperboliques à commutations. Cela signifie qu'il existe une interaction entre une dynamique continue et une dynamique discrète. Autrement dit, il existe différents modes dans lesquels peut évoluer la dynamique continue: ces modes sont dictés par la dynamique discrète. Ce changement de mode peut être contrôlé (dans le cas d'une boucle fermée), ou non-contrôlé (dans le cas d'une boucle ouverte). Nous nous sommes intéressés au premier cas. Par une analyse Lyapunov nous avons construit trois règles de commutations capables de stabiliser le système. Nous avons montré comment modifier deux d'entre elles pour obtenir des propriétés de robustesse et de stabilité entrée-état. Ces règles de commutations ont été testées numériquement.Ensuite, nous avons considéré la génération de trajectoire pour des systèmes hyperboliques linéaires 2x2 par backstepping. L'étape suivante a été de considérer une action Proportionnelle-Intégrale pour stabiliser la solution du système autour de la trajectoire de référence. Pour cela nous avons construit une fonction Lyapunov non-diagonale. Nous avons montré que l'action intégrale est capable de rejeter des erreurs distribuées et frontières.Enfin, nous avons considéré des aspects numériques pour l'analyse Lyapunov. Les conditions pour la stabilité et la conception de contrôleurs obtenues par des fonctions de Lyapunov quadratiques font intervenir une infinité d'inégalités matricielles. Nous avons montré que cette complexité peut être réduite en considérant une sur-approximation polytopique de ces contraintes.Les résultats obtenus ont été illustrés par des exemples académiques et des systèmes dynamiques physiques (comme les équations de Saint-Venant et les équations de Aw-Rascle-Zhang). / In this thesis we have considered different aspects for the control of hyperbolic systems.First, we have studied switched hyperbolic systems. They contain an interaction between a continuous and a discrete dynamics. Thus, the continuous dynamics may evolve in different modes: these modes are imposed by the discrete dynamics. The change in the mode may be controlled (in case of a closed-loop system), or may be uncontrolled (in case of an open-loop system). We have focused our interest on the former case. We procedeed with a Lyapunov analysis, and construct three switching rules. We have shown how to modify them to get robustness and ISS properties. We have shown their effectiveness with numerical tests.Then, we have considered the trajectory generation problem for 2x2 linear hyperbolic systems. We have solved it with backstepping. Then, we have considered the tracking problem with a Proportionnal-Integral controller. We have shown that it stabilizes the error system around the reference trajectory with a new non-diagonal Lyapunov function. The integral action has been shown to be able to reject in-domain, as well as boundary disturbances.Finally, we have considered numerical aspects for the Lyapunov analysis. The conditions for the stability and design of controllers by quadratic Lyapunov functions involve an infinity of matrix inequalities. We have shown how to reduce this complexity by polytopic embeddings of the constraints.Many obtained results have been illustrated by academic examples and physically relevant dynamical systems (as Shallow-Water equations and Aw-Rascle-Zhang equations).

Cette thèse propose des contributions sur la commande événementielle pour des réseaux modélisés par une classe des systèmes de dimension infinie. Premièrement nous nous focalisons sur la modélisation et contrôle frontière des réseaux qui sont décrits par des systèmes hyperboliques de lois de conservation. En nous inspirant de modèles macroscopiques dans le cadre des réseaux de communications, nous traitons des systèmes couplés EDP-EDO, dont les noeuds (les serveurs) sont modélisés par des EDO non-linéaires alors que des lignes de transmission sont décrites par des systèmes hyperboliques lorsque des retards peuvent être pris en compte. Pour le système linéarisé resultant, autour d'un point d'équilibre optimal, on effectue aussi bien une analyse de stabilité "Input-to-state stable" que de la synthèse du contrôle pour le gain asymptotique grâce à une analyse de fonction de Lyapunov et une formulation LMI.Ensuite, nous considérons des aspects théoriques de la commande évènementielle aux frontières pour les systèmes hyperboliques. D'un côté, avec cette stratégie de contrôle, nous ciblons la réduction de la consommation d' énergie en traitant les contraintes de communication et de calcul. D' autre part, nous utilisons cette stratégie comme une manière rigoureuse pour échantillonner temporellement lorsqu' on a besoin de mettre en oeuvre les contrôleurs continus sur une plateforme numérique. Une étude mathématique sur l'existence et l' unicité des solutions ainsi que sur les aspects de stabilité est réalisée. / This thesis provides contributions on event-based control of networks model by a class of infinite dimensional systems. We first focus on the modeling and boundary control of networks described by hyperbolic systems of conservation laws. Highly inspired by macroscopic models in communication networks, we deal with a coupled PDE-ODE, where the nodes (servers) are modeled by nonlinear ODEs whereas transmission lines are described by hyperbolic equations when communication delays may be taken into account. For the resulting linearized system around an optimal equilibrium point, Input-to state stability (ISS) analysis as well as asymptotic gain control synthesis are carried out by means of Lyapunov techniques and LMI formulation.We then address some theoretical aspects of event-based boundary control of hyperbolic systems. One one hand, with this computer control strategy, we intend to reduce energy consumption when dealing with communication and computational constraints. On the other hand, we use this strategy as a rigorous way of sampling in time when implementation of continuous time controllers on a digital platform is required. A mathematical study regarding well-posedness of the solutions as well as stability issues is conducted.

Ce travail étudie les stratégies de contrôle des limites pour l'analyse de stabilité et la stabilisation d'un système hyperbolique de premier ordre couplé à des conditions limites dynamiques non linéaires. La modélisation d'un écoulement à l'intérieur d'un tube (phénomène de transport de fluide) avec une stratégie de contrôle des limites appliquée dans une installation expérimentale physique est considérée comme une étude de cas pour évaluer les stratégies proposées. Dans le contexte des systèmes de dimension finie, des outils de contrôle classiques sont appliqués pour traiter des systèmes hyperboliques de premier ordre ayant des conditions limites données par le couplage d'un modèle dynamique de colonne de chauffage et d'un modèle statique de ventilateur. Le problème de suivi de cette dynamique complexe est abordé de manière simple en considérant des approximations linéaires, des schémas de différences finies et une action intégrale conduisant à un système linéaire à temps discret augmenté avec une dimension dépendant de la taille d'échelon de la discrétisation dans l'espace. Par conséquent, pour la contrepartie dimensionnelle infinie, deux stratégies sont proposées pour résoudre le problème de contrôle de frontière des systèmes hyperboliques de premier ordre couplé à des conditions de frontière dynamique non linéaires. Le premier se rapproche de la dynamique du système hyperbolique de premier ordre par un retard pur. La stabilité convexe et les conditions de stabilisation des systèmes quadratiques non linéaires retardés d'entrée incertaine sont proposées sur la base de la théorie de la stabilité de Lyapunov-Krasovskii (LK) qui sont formulées en termes de contraintes de l'inégalité matricielle linéaire (LMI) avec des variables supplémentaires lâches (introduites par le lemme de Finsler ). Ainsi, des fonctions strictement de Lyapunov sont utilisées pour dériver une approche basée sur LMI pour la stabilité de la frontière régionale robuste et la stabilisation des systèmes hyperboliques de premier ordre avec une condition de frontière définie au moyen d'un système dynamique non linéaire quadratique. Les conditions de stabilité et de stabilisation proposées pour LMI sont évaluées en tenant compte de plusieurs exemples universitaires et de l'écoulement à l'intérieur d'une étude de cas. / This work studies boundary control strategies for stability analysis and stabilization of first-order hyperbolic system coupled with nonlinear dynamic boundary conditions. The modeling of a flow inside a pipe (fluid transport phenomenon) with boundary control strategy applied in a physical experimental setup is considered as a case study to evaluate the proposed strategies. Firstly, in the context of finite dimension systems, classical control tools are applied to deal with first-order hyperbolic systems having boundary conditions given by the coupling of a heating column dynamical model and a ventilator static model. The tracking problem of this complex dynamics is addressed in a simple manner considering linear approximations, finite difference schemes and an integral action leading to an augmented discrete-time linear system with dimension depending on the step size of discretization in space. Hence, for the infinite dimensional counterpart, two strategies are proposed to address the boundary control problem of first-order hyperbolic systems coupled with nonlinear dynamic boundary conditions. The first one approximates the first-order hyperbolic system dynamics by a pure delay. Then, convex stability and stabilization conditions of uncertain input delayed nonlinear quadratic systems are proposed based on the Lyapunov-Krasovskii (L-K) stability theory which are formulated in terms of Linear Matrix Inequality (LMI) constraints with additional slack variables (introduced by the Finsler's lemma). Thus, strictly Lyapunov functions are used to derive an LMI based approach for the robust regional boundary stability and stabilization of first-order hyperbolic systems with a boundary condition defined by means of a nonlinear quadratic dynamic system. The proposed stability and stabilization LMI conditions are evaluated considering several academic examples and also the flow inside a pipe as case study.

Dans cette thèse, nous nous intéressons aux équations aux dérivées partielles hyperboliques sous contraintes ; plus particulièrement aux problèmes provenant de la mécanique de la plasticité parfaite. Un bref historique de l'origine mécanique des problèmes de la plasticité parfaite ainsi que des résultats précédemment obtenus sont décrits dans le Chapitre 1. Dans le Chapitre 2, nous concentrons notre attention sur les systèmes hyperboliques avec conditions de bord. Nous développons une théorie faible pour ces problèmes et expliquons dans un cas simplifié le caractère bien posé de cette théorie. Puis, nous introduisons de manière similaire la notion de solution faible pour des systèmes hyperboliques avec condition de bord soumis à une contrainte. Nous nous dédions, dans le chapitre 3, à l'étude d'un modèle simplifié de la dynamique de la plasticité parfaite. Nous confrontons l'approche introduite au chapitre précédent avec celle, plus classique, provenant du calcul des variations qui permet d'obtenir l'existence et l'unicité des solutions pour ce modèle. Cela nous permet de mettre en évidence une nouvelle interaction entre les conditions de bord et les contraintes ainsi que d'aboutir à un théorème de régularité des solutions. Dans le chapitre 4, nous nous intéressons à l'approximation numérique des systèmes hyperboliques sous contraintes grâce à des schémas de type volumes finis. Cela nous permet d'obtenir un résultat de convergence pour les problèmes sans bord et d'illustrer numériquement les interactions entre les conditions de bord et les contraintes sur l'exemple du chapitre 3. / In this thesis, we consider constrained hyperbolic partial differential equations and more precisely mechanical problems coming from perfect plasticity. The goal of this thesis is to study these problems thanks to different approaches, to analyze the interactions between these different points of view and to confront these various analyzes to get new results. A brief review of the mechanical origin of perfect plasticity problems and also of the previous results on these topics are described in Chapter 1. In Chapter 2, we focus our attention on hyperbolic systems with boundary conditions. First, we develop a weak theory for these problems and explain, in a simplified case, why this theory is well-posed. Then, we introduce similarly a notion of weak solutions for constrained hyperbolic systems with boundary conditions. Chapter 3 is devoted to the study of the simplified model of dynamical perfect plasticity. We confront the approach introduced in the previous chapter with the one, more standard, coming from calculus of variations that allows us to obtain existence and uniqueness of the solutions for this model. It allows us to bring to light a new interaction between the boundary conditions and the constraints and to get a short-time regularity theorem. Lastly, in Chapter 4, we are interested in the numerical approximation of constrained hyperbolic systems thanks to finite volume schemes. This work allows us to get a convergence result for problems without boundary condition and to show numerically the link between boundary conditions and constraints on the example of the previous chapter.

In this dissertation we present an analysis for the discontinuous Galerkin discretization error of multi-dimensional first-order linear symmetric and symmetrizable hyperbolic systems of conservation laws. We explicitly write the leading term of the local DG error, which is spanned by Legendre polynomials of degree p and p+1 when p-th degree polynomial spaces are used for the solution. For special hyperbolic systems, where the coefficient matrices are nonsingular, we show that the leading term of the error is spanned by (p+1)-th degree Radau polynomials. We apply these asymptotic results to observe that projections of the error are pointwise O(h^p+2)-superconvergent in some cases and establish superconvergence results for some integrals of the error. We develop an efficient implicit residual-based a posteriori error estimation scheme by solving local finite element problems to compute estimates of the leading term of the discretization error. For smooth solutions we obtain error estimates that converge to the true error under mesh refinement. We first show these results for linear symmetric systems that satisfy certain assumptions, then for general linear symmetric systems. We further generalize these results to linear symmetrizable systems by considering an equivalent symmetric formulation, which requires us to make small modifications in the error estimation procedure. We also investigate the behavior of the discretization error when the Lax-Friedrichs numerical flux is used, and we construct asymptotically exact a posteriori error estimates. While no superconvergence results can be obtained for this flux, the error estimation results can be recovered in most cases. These error estimates are used to drive h- and p-adaptive algorithms and assess the numerical accuracy of the solution. We present computational results for different fluxes and several linear and nonlinear hyperbolic systems in one, two and three dimensions to validate our theory. Examples include the wave equation, Maxwell's equations, and the acoustic equation. / Ph. D.

We develop novel numerical methods for optimization problems subject to constraints given by nonlinear hyperbolic systems of conservation and balance laws in one space dimension. These types of control problems arise in a variety of applications, in which inverse problems for the corresponding initial value problems are to be solved. The optimization method can be seen as a block Gauss-Seidel iteration. The optimization requires one to numerically solve the hyperbolic system forward in time and the corresponding linear adjoint system backward in time. We test the optimization method on a number of control problems constrained by nonlinear hyperbolic systems of PDEs with both smooth and discontinuous prescribed terminal states. The theoretical foundation of the introduced scheme is provided in the case of scalar hyperbolic conservation laws on an unbounded domain with a strictly convex flux. In addition, we empirically demonstrate that using a higher-order temporal discretization helps to substantially improve both the efficiency and accuracy of the overall numerical method. / acase@tulane.edu

We consider linear and semilinear stochastic partial differential equations that in some sense can be viewed as being at the "endpoints" of the classical variational theory by Krylov and Rozovskii [25]. In terms of regularity of the coeffcients, the minimal assumption is boundedness and measurability, and a unique L2- valued solution is then readily available. We investigate its further properties, such as higher order integrability, boundedness, and continuity. The other class of equations considered here are the ones whose leading operators do not satisfy the strong coercivity condition, but only a degenerate version of it, and therefore are not covered by the classical theory. We derive solvability in Wmp spaces and also discuss their numerical approximation through finite different schemes.

Made available in DSpace on 2014-07-29T16:02:23Z (GMT). No. of bitstreams: 1 versaofinal_lidiane.pdf: 708631 bytes, checksum: c97f21b6c4fd093f9c0ec8f92aa22110 (MD5) Previous issue date: 2010-04-09 / In this dissertation we describe the solutions to the Riemann problem for a system of two conservation laws written in the normal from according to classification of Schaeffer-Shearer in [9]. Through changes of variables Schaeffer-Shearer determined the normal form for a nonlinear system of two conservation laws with an isolated umbilical point in state space. The normal form consists of a system of two equations, with homogeneous and quadratic functions of flow that depend only on two parameters. Also in [9] were established four distinct regions in terms of parameters, denoted by I, II, III and IV, in which varying pair of parameters in each region, the curves of waves that make up the solution of the Riemann problem have the same configuration. In this dissertation we consider the case in which the pair of parameters belongs to region IV, and in the particular case in which one of the parameters is null. In this case, the classic Lax criterion for admissibility of shocks (discontinuity solutions) generally is sufficient to obtain uniqueness of solution. Although, for some initial states, it is necessary to admit in solution also the called compressive shocks, which do not satisfy the Lax criterion. / Nesta dissertação determinamos as soluções do problema de Riemann para um sistema de duas leis de conservação escrito na forma normal segundo a classificação de Schaeffer-Shearer, em [9]. Através de mudanças de variáveis, Schaeffer-Shearer determinaram a forma normal para um sistema não linear de duas leis de conservação com um ponto umbílico isolado no espaço de estados. A forma normal consiste de um sistema de duas equações, com funções de fluxo quadráticas homogêneas que dependem apenas de dois parâmetros. Também em [9] foram determinadas quatro regiões distintas no plano dos parâmetros, denotadas por I, II, III e IV, onde, variando o par de parâmetros em cada região, as curvas de onda que compõem a solução do problema de Riemann tem a mesma configuração. Nesta dissertação consideramos o caso em que o par de parâmetros pertence a região IV, e ainda no caso particular em que um dos parâmetros é nulo. Neste caso, o clássico critério de Lax para admissibilidade dos choques (soluções descontínuas), em geral, é suficiente para se obter unicidade de solução. Embora, para alguns estados iniciais, é necessário admitir na solução também os chamados choques compressivos, que não satisfazem o critério de Lax.