Ce manuscrit se décompose en deux parties principales, associées à deux types d'applications assez différentes. Dans la première partie qui comprend les deux premiers chapitres, je m'intéresse à des systèmes issus de problèmes de contrôle et d'estimation en physique quantique; dans la deuxième partie (troisième chapitre du manuscrit), j'étudie la propagation d'ondes électriques le long des fils classiques dans un réseau de lignes de transmission et je considère certains problèmes d'estimation de paramètres. Dans le premier chapitre nous étudions le problème de la planification de trajectoires pour des systèmes quantiques fermés modélisés par des équations de Schrödinger bilinéaire. Nous démontrons alors des résultats de la stabilisation approchée pour le cas d'une boite quantique infinie ainsi que pour le cas d'un potentiel décroissant. Dans les deux cas, le manque de pré-compacité des trajectoires dans des espaces fonctionnels appropriés nous oblige à proposer des méthodes de Lyapunov qui évitent des phénomènes de perte de masse à l'infini. Dans le deuxième chapitre nous étudions le problème de stabilisation de systèmes quantiques en observation. Cette observation nécessite l'ouverture du système à son environnement. Les modèles pertinents pour l'évolution de ce type de systèmes sont des modèles stochastiques basés sur des trajectoires de Monte-Carlo quantiques. Nous étudions alors certains problèmes de stabilisation qui parviennent de vraies expériences physiques. Enfin, dans le chapitre 3 nous considérons le problème d'estimation de paramètres pour un réseau de fils de câblage électrique. Dans ce but, nous étudions deux approches : l'approche temporelle et l'approche fréquentielle. Dans l'approche temporelle, nous considérons le réseau le plus simple qui consiste d'une seule ligne de transmission et nous proposons un algorithme d'identification pour l'équation d'onde associé qui est basé sur l'application des observateurs asymptotiques. Dans l'approche fréquentielle, nous considérons un réseau plus compliqué de la forme étoile. Nous proposons alors des résultats d'identifiabilité basés sur des techniques de l'inverse scattering.

A systematic research on aeroacoustics conducted around the world for the last few decades has revealed various inherent characteristics of the jet noise radiation. However, a lot more needs to be done for the theoretical as well as experimental predictions of various jet noise features based on actual flow details. The work reported in the present thesis is an attempt in this direction. A critical study of existing literature on jet noise shows that none of the general wave equations lends itself easily for predictions of all the jet noise features. It is shown that while LighthilPs classical acoustic analogy approach, with some reasonable approximations, can be used to yield most of the information needed by the engineers, the convected wave equations of Phillips and Lilley are required to study the acoustic radiation in what has come to be known as "Refraction valley" or "Cone of relative silence". The characteristics of the sound field of underexpanded cold jet impingement flows were studied by measuring the noise emanating from two convergent nozzles of throat diameter 2.5 mm and 5 mm each and a convergent - divergent nozzle of exit diameter of 6.49 mm, when the jet impinges on a flat plate kept perpendicular to the direction of the jet. The measurements were conducted upstream of the nozzle over an extensive envelope of jet operating conditions such as chamber stagnation pressure, mass flow rate through the nozzle and diameter of the nozzle. The source strength at the jet boundary was obtained by measuring acoustic pressure amplitude close to the jet contour assuming it as locally cylindrical. Particular attention was focussed on backward projection of the sound field on to a cylindrical surface. This is the application of acoustic holography to study the sound radiation in the audio frequency region. With the help of FFT and software developed for this purpose, the theoretical predictions using data from several cylindrical surfaces were compared. A detailed analysis of noise radiation from a cold sonic and supersonic free jet was also carried out. The experimental work involved the measurement of noise field from a 2.5 mm, 5 mm convergent and a convergent - divergent nozzle of exit diameter of 6.49 mm and area ratio 1.687 for designed Mach number of two. The experimental setup consisted essentially of a pressure chamber made of mild steel, designed to withstand 50 bar pressure. This chamber is a cylinder with dia 0.421 m and length 0.85 m. The nozzles were made of mild steel. Compressed air approximately at room temperature is supplied to the nozzle via a control valve. The measuring and recording instruments consists of B & K Microphones, Preamplifiers, Conditioning amplifier and a Mediator, which measure a Sound Pressure Level at a point. The nozzles were operated at pressure ratio upto 25 bar. The noise signal was processed through 12 channel data acquisition system. Acoustic pressure and SPL were" calculated using theoretical relations and software developed. Using this software Fast Fourier Transformations of raw signal was obtained from 20 Hz to 20 kHz. Also constant SPL contour graphs were obtained. Source strength distribution at the jet boundary has been obtained by the principle of acoustic holography. Experimental values are closely matching with the results obtained by acoustic holography. The percentage error for acoustic pressure and SPL were less than 12%. The experimental results were used to obtain the source distribution in terms of gross jet parameters.

The existence and -if exists- the nature of the solutions of the scalar wave equation in spacetimes with closed timelike curves are investigated. The general properties of the solutions on some class of spacetimes are obtained. Global monochromatic solutions of the scalar wave equation are obtained in flat wormholes of dimensions 2+1 and 3+1. The solutions are in the form of infinite series involving cylindirical and spherical wave functions and they are elucidated by the multiple scattering method. Explicit solutions for some limiting cases are illustrated as well. The results of 2+1 dimensions are verified by using numerical methods.

Viele Problemstellungen der Naturwissenschaften führen zur Betrachtung von nichtlinearen Wellengleichungen. Dabei ist von großem Interesse, ob zu vorgegebenen kleinen Daten Lösungen eindeutig existieren und ob diese stetig von den Daten abhängen. Hilfsmittel für diese Probleme sind Aussagen über lineare Wellengleichungen. In der vorliegenden Arbeit werden lineare Klein-Gordon Gleichungen, also Wellengleichungen mit Potentialterm, mit zeitabhängiger Masse bzgl. des Verhaltens ihrer Lösungen untersucht. Von speziellem Interesse sind Resultate mit Bezug auf verallgemeinerte Energieerhaltung und sogenannte Lp – Lq decay-Abschätzungen. Aus der Arbeit geht hervor, dass man eine Klassifizierung für Gleichungen mit fallendem Masseterm finden kann. Für Gleichungen vom Wellentyp ist der Einfluss des Potentialterms gering und die Lösungen verhalten sich wie Lösungen der Wellengleichung. Dem gegenüber stehen Gleichungen vom Klein-Gordon-Typ mit erkennbarem Einfluss des Masseterms. Ausgangspunkt für die Klassifizierung ist das kritische Verhalten der Lösungen einer skaleninvarianten Gleichung mit speziellem Masseterm.

Cette thèse porte sur l’étude d’équations d’évolution sur certains groupes hyperboliques, en particulier, nous étudions l’équation de la chaleur, l’équation de Schrödinger et l’équation des ondes modifiée, d’abord sur les arbres homogènes, ensuite sur des graphes symétriques. Sur les arbres homogènes, nous montrons que, sous une hypothèse d’invariance de jauge, on a existence globale des solutions de l’équation de Schrödinger ainsi qu’un phénomène de ’scattering’ pour des données arbitraires dans l’espace des fonctions de carré intégrable sans restriction sur le degré de la non-linéarité, contrairement au cas euclidien ou au cas hyperbolique. Nous généralisons ensuite ce résultat sur les graphes symétriques de degré (k − 1)(r − 1) sous la condition k < r. Un de nos principaux résultats sur les graphes symétriques est l’estimation du noyau de la chaleur associé au laplacien combinatoire. Pour finir, nous établissons une expression explicite des solutions de l’équation des ondes modifiée sur les graphes symétriques. / This thesis focuses on the study of evolution equations on certain hyperbolic groups, in particular, we study the heat equation, the Schrödinger equation and the modified wave equation first on homogeneous trees then on symmetric graphs. In the homogeneous trees case, we show that under a gauge invariance condition, we have global existence of solutions of the Schrödinger equation and scattering for arbitrary data in the space of square integrable functions without any restriction on the degree of the nonlinearity, in contrast to the euclidean and hyperbolic space cases. We then generalize this result on symmetric graphs of degree (k − 1)(r − 1) under the condition k < r . One of our main results on symmetric graphs is the estimate of the heat kernel associated to the combinatorial laplacian. Finally, we establish an explicit expression of solutions of the modified wave equation on symmetric graphs.

Cette thèse est dédiée a l''etude de certaines propriétés des équations d' évolutions non-autonomes $u'(t)+A(t)u(t)=f(t), u(0)=x.$ Il s'agit précisément de la propriété de la régularité maximale $L^p$: étant donnée $fin L^{p}(0,tau;H)$, montrer l'existence et unicité de la solution $u in W^{1,p}(0,tau;H)$. Ce problème a 'et'e intensivement étudie dans le cas autonome, i.e., $A(t)=A$ pour tout $t$. Dans le cas non-autonome, le problème a été considéré par J.L.Lions en 1960. Nous montrons divers résultats qui étendent tout ce qui est connu sur ce problème. On suppose ici que la famille des opérateurs $(mathcal{A}(t))_{tin [0,tau]}$ est associée à des formes quasi-coercives, non autonomes $(fra(t))_{t in [0,tau]}.$ Nous considérons également le problème de régularité maximale pour les d'ordre 2 (équations des ondes). Plusieurs exemples et applications sont considérés. / This Thesis is devoted to certain properties of non-autonomous evolution equations $u'(t)+A(t)u(t)=f(t), u(0)=x.$ More precisely, we are interested in the maximal $L^p$-regularity: given $fin L^{p}(0,tau;H),$ prove existence and uniqueness of the solution $u in W^{1,p}(0,tau;H)$. This problem was intensively studied in the autonomous cas, i.e., $A(t)=A$ for all $t.$ In the non-autonomous cas, the problem was considered by J.L.Lions in 1960. We prove serval results which extend all previously known ones on this problem. Here we assume that the familly of the operators $(mathcal{A}(t))_{tin [0,tau]}$ is associated with quasi-coercive, non-autonomous forms $(fra(t))_{t in [0,tau]}.$ We also consider the problem of maximal regularity for second order equations (the wave equation). Serval examples and applications are given in this Thesis.

Made available in DSpace on 2014-06-11T19:32:09Z (GMT). No. of bitstreams: 0 Previous issue date: 2011-07-05Bitstream added on 2014-06-13T18:43:07Z : No. of bitstreams: 1 castro_lrb_dr_guara.pdf: 373502 bytes, checksum: ce90da31dd1eacd9e0839fbf879727bd (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Fazemos uma revisão detalhada de alguns fundamentos básicos do formalismo de Du n- Kemmer-Petiau (DKP). Analisamos as consequências sobre o potencial matricial V para fornecer uma quadricorrente conservada. Também analisamos o comportamento das intera ções vetoriais mínimas e não mínimas sob as transformações de paridade, conjugação de carga e reversão temporal. A ambiguidade do acoplamento eletromagnético (interação vetorial mínima) também é revisada em detalhe. Algums conceitos errados sobre hermiticidade e valores esperados na teoria de DKP difundidos na literatura são discutidos. Além disso, neste trabalho desenvolvemos uma forma alternativa de procurar soluções analíticas da equação de DKP tridimensional (setor spin-0) para o caso de acoplamentos vetoriais (mínimo e não-mínimo). Considerando potenciais com simetria esférica, o problema pode ser mapeado num problema de Sturm-Liouville (da mecânica quântica não relativística) para um dos componentes do espinor de DKP. Neste processo a quadricorrente, a condição de normalização e valores esperados também podem ser expressos em termos desse componente do espinor de DKP de uma forma simples. Como uma aplicação do método desenvolvido, consideramos uma forma linear para os acoplamentos vetoriais. / A detailed review of some basics fundamentals of the Du n-Kemmer-Petiau (DKP) formalism is made. The consequences on the potential matrix V for furnish a conserved four-current are analyzed. We also analyze the behavior of minimal and nonminimal vector interactions under parity transformation, charge conjugation and time reversal. The ambiguity of the electromagnetic coupling (minimal vector interaction) is also reviewed in detail. Some misconceptions about the hermiticity and expectation values of the DKP theory widespread in the literature are discussed. In addition, an alternative way to search for analytical solutions of the DKP equation in (3+1) dimensions (spin-0 sector) in the case of vector coupling (minimal and nonminimal) is developed. Considering potentials with spherical symmetry, the problem can be mapped into a Sturm-Liouville problem (nonrelativistic quantum mechanics) for one of the components of the DKP spinor. In this process, the four-current, normalization condition and expectation values can also be expressed in terms of that component of the DKP spinor in a simple way. As an application of the developed method, we consider a linear form for the vector couplings.

Made available in DSpace on 2016-06-02T20:28:26Z (GMT). No. of bitstreams: 1 3788.pdf: 684718 bytes, checksum: 743ac325dfb93fd96a6cc9b15d66467d (MD5) Previous issue date: 2011-05-27 / Financiadora de Estudos e Projetos / In this work, we study the result of global well-Posedness for the cubic wave equation @2 t u&#56256;&#56320;_u+u3 = 0 in R_R3, where the Cauchy data is in the Sobolev space Hs(R3)_ Hs&#56256;&#56320;1(R3) with 13 18 < s < 1. The proof is based on the work of T. Roy, [23], in this paper Roy propose a almost conservation law for the energy and from this he get a inequality that together with the local well-posedness theory proved by Lindbald and Sogge in [18] guarantee the global well-posedness for the problem. / Neste trabalho estudamos um resultado de boa colocação global para a equação da onda cúbica &#948;(_t^2)u-&#8710;_u+U^3=0 em R_R3, no qual os dados de Cauchy estão no espaço de Sobolev Hs(R3) x Hs&#56256;&#56320;1(R3), para 13 18 < s < 1. A prova é baseada no rabalho de T. Roy, [23], nele é estabelecido uma lei de quase conservação de energia e a partir disso se obtém uma desigualdade que aliada a teoria da boa colocação local estabelecida por Lindbald e Sogge em [18] garante a boa colocação global para o problema.

Orientador: Adalberto Spezamiglio / Banca: Juan Amadeo Soriano Palomino / Banca: Waldemar Donizete Bastos / Resumo: Um problema de controle exato na fronteira para um sistema de equações de ondas acopladas e considerado em um retângulo do plano. Obtem-se controle de quadrado integrável para estados iniciais de energia finita. / Abstract: We are concerned with a problem of exact boundary controllability for a coupled sistem of wave equations in a rectangle of the plane. We obtain square integrable control for initial state with nite energy. / Mestre

Este trabalho é dedicado ao estudo do comportamento a longo prazo de uma classe de equações viscoelásticas não lineares com memória, da forma |\'upsilon IND. t\'| POT. ho\' \'upsilon IND. tt\' - DELTA \'upsilon\' - \'DELTA upsilon IND. tt\' + \'INT. SUP. t INF. \\tau\' upsilon (t- s) \'DELTA epsilon\' (s) ds = h, \'\\tau\' > 0, definida num domínio limitado de \'R POT. N\'. Tal classe de problemas foi estudada por diversos autores desde 2001, com \'\\tau = 0. Os resultados existentes são principalmente devotados à existência de soluções globais, decaimento da energia, com ou sem dissipações adicionais, existência com dados pequenos, entre outros. Entretanto, a questão da unicidade de soluções e existência de atratores globais não foram discutidas em trabalhos anteriores. No presente trabalho, apresentamos resultados de unicidade e existência de atratores globais para essa classe de problemas num contexto mais geral, incluindo o caso em que \'\\tau\' = -\'INFINITO\'. Além disso, incluímos um problema complementar, de quarta ordem onde estudamos a existência de atratores exponenciais / This work is concerned with the long-time behaviour of a class nonlinear viscoelastic equations of the form |\'upsilon IND. t\'| POT. ho\' \'upsilon IND. tt\' - DELTA \'upsilon\' - \'DELTA upsilon IND. tt\' + \'INT. SUP. t INF. \\tau\' upsilon (t- s) \'DELTA epsilon\' (s) ds = h, \'ho\' > 0, defined in a bounded domain of \'R POT. N\'. Such class of problems was studied by several authors since 2001, with \'\\tau\' = 0. Existing results are mainly devoted to global existence, energy decay, with or without additional dampings, existence with small data, among others. However, uniqueness and existence of global attractors were not considered previously. In the present work, we establish some results on the uniqueness of solutions and existence of global attractors in a more general setting, including \'\\tau\' = - \'INFINITY\'. In addition, we have added a second problem concerned with a fourth order equation where we study the existence of exponential attractors