本論文討論了具有非平凡旋度的穩態亞音速流體的適定性問題。 / 首先，我們研究了通過無限長週期管道的二維亞音速流禮。當管道某一週期位置伯努利函數擾動很小，且質量數介於與適當的範圍時，有且僅有唯一的亞音速流禮。特別地，對於伯努利函數為常值的情形，我們還通過結構緊性的方法證明了亞音速-音速流體的存在性。此時，質量數可以達到一臨界值。謝春景和辛周平在處理二維司壓歐拉方程時曾引入了一個重要的處理方法一一流函數表達式。然而，對於週期流體的問題，伯努利函數和流函數的相互關係是無法事先確定的。為此，我們建立了一個關於流函數的非線性映射。該映射的不動點給出了相應歐拉方程的解。 / 其次，對於二維亞音速流體通過對稱障礙物的問題，當來流的伯努利函數關於y方向對稱，且擾動很小時，我們給出了流体的存在性和唯一性的証明。这里，我們利用歐拉方程的流函數方法，得到了對應于流函數的二階方程的解。能量方法以及動量場與來流動量場之差的L2可積性給出了流函數的漸進行為。這一漸進行為結合障礙物外無駐點的事實說明了流函數表示與原先歐拉方程是相容的。 / 最后，我們研究了當給定管道壁上法向动量時，三維穩態流體通過方體管道的問題。如果入口處伯努利函數的擾動和旋度的法向分量為零，則當邊界的法向動量不超過一臨界值時，無旋的亞音速流體存在。對於一般情形，若伯努利函數的擾動和旋度的法向分量很小時，我們利用將速度均分解均無旋部分和旋度部分的方法給出了流體存在性的證明。這裡，我們通過求解一加權的散旋系統得到了旋度部份的解:而無旋部份則由一擬線性橢圓方程的解給出。 / In this thesis, the wellposedness theory of steady subsonic flows with nontrivial vorticities is studied in various aspects. / First, we study 2-D subsonic flows through infinitely long periodic nozzles. It is showed that when mass flux lies in a suitable regime and the variation of Bernoulli's function at some given section is sufficiently small, there exists a unique global subsonic flow in the periodic nozzle. In particular, if Bernoulli's function is a constant, the existence of subsonic flow is also obtained when mass flux takes the critical number by a compensated compactness framework. One of the main tools to handle 2-D compressible Euler equations is the stream function formulation first established by Xie and Xin. The main difficulty in adapting this formulation in periodic nozzles is that the relation between Bernoulli's function and stream function cannot be fixed. We resolve this difficulty via setting up a nonlinear map from stream function at the given section to itself. The fixed point of this map induces a solution of corresponding Euler equations. / Second, the existence and uniqueness of 2-D subsonic flows past a symmetric body are established under the assumption that Bernoulli's function is given symmetrically in the upstream with small variation. By the stream function formulation for 2-D compressible Euler equations, one obtains the solution of the Euler equations via solving a quasilinear second order equation for stream function. This is achieved with the help of the theory of elliptic equations of two variables. Asymptotic behavior for the stream function is obtained via energy method and L²-integral of the difference between the momentum and its asymptotic behavior in the upstream. The asymptotic behavior, together with the property that stagnation points are absent outside the body, yields that the stream function formulation is consistent with the original Euler system. / Finally, we study the existence of 3-D steady subsonic flows in rectangular nozzles when prescribing the normal component of the momentum on the boundary. If, in addition, the normal component of the voriticity and the variation of Bernoulli's function at the exit vanish, then there exists a unique subsonic potential flow when the magnitude of the normal component of the momentum is less than a critical number. In general, if the normal component of the vorticity and the variation of Bernoulli's function are both sufficiently small, we prove the existence of Euler flows by decomposing the velocity into the vortical part and the potential part. A div-curl system with given weighted function is used to obtain the vortical part and the potential part is induced by the solution to a quasilinear elliptic equation. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Chen, Chao. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2012. / Includes bibliographical references (leaves 111-120). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract also in Chinese. / Abstract --- p.i / Acknowledgement --- p.iv / Chapter 1 --- Introduction --- p.1 / Chapter 2 --- Preliminaries --- p.12 / Chapter 3 --- 2-D subsonic flows through in finitely long periodic nozzles --- p.23 / Chapter 3.1 --- Introduction and main result --- p.23 / Chapter 3.2 --- Stream function formulation of potential flows --- p.27 / Chapter 3.2.1 --- Bernoulli's law and stream function formulation --- p.27 / Chapter 3.2.2 --- Potential flows and proof of Theorem 3.1.1 --- p.30 / Chapter 3.3 --- Analysis of the well-posedness of Euler flows --- p.32 / Chapter 3.3.1 --- Existence, uniqueness, and periodicity of truncated flows --- p.34 / Chapter 3.3.2 --- Existence and uniqueness of Euler flflows --- p.41 / Chapter 4 --- 2-D subsonic flows past a symmetric body --- p.47 / Chapter 4.1 --- Motivation and mathematical formulation --- p.47 / Chapter 4.2 --- Truncated problem --- p.53 / Chapter 4.3 --- Asymptotic behavior at upstream and downstream --- p.59 / Chapter 4.4 --- Existence and uniqueness of Euler flflows --- p.61 / Chapter 5 --- 3-D subsonic Euler flows through nitely long nozzles --- p.67 / Chapter 5.1 --- Mathematical formulation and main results --- p.67 / Chapter 5.2 --- Some preliminaries --- p.71 / Chapter 5.3 --- 3-D potential flows --- p.76 / Chapter 5.3.1 --- Apriori estimates for truncated potential flows --- p.77 / Chapter 5.3.2 --- Existence and uniqueness of potential flows --- p.91 / Chapter 5.4 --- General 3-D steady Euler systems --- p.94 / Chapter 6 --- Further discussions and future work --- p.109 / Bibliography --- p.111

A abordagem modulacional fornece uma estrutura simplificada para descrever a interação não-linear de três ondas (tripleto) em regimes onde a freqüência dos portadores é muito maior do que a freqüência modulacional. No primeira parte deste trabalho investigamos a quebra da tradicional aproximação modulacional na interação de tripleto, que é precisa quando o acoplamento das três ondas é fraco. Examinamos os tipos de dinâmicas decorrentes quando o acoplamento se incrementa desde valores muito pequenos até valores grandes e detectamos uma transição abrupta para valores grandes onde as excursões de amplitude limitada do regime modulacional alcançam regiões muito maiores do espaço de fases. Na segunda parte, estudamos uma extensão multi modo da interação não-linear de três ondas e vemos que os limites da aproximação modulacional podem ser de fato muito restritivos. Para acoplamentos muito pequenos mostramos que todos os modos exibem modulações lentas na amplitude, porém a medida que incrementamos o acoplamento uma transição acontece e os modos mudam para um novo regime dinâmico onde nenhum deles pode mais ser visto como portadores harmônicos de alta freqüência lentamente modulados. As estimativas para o acoplamento crítico e para os tempos de relaxação foram obtidos baseados em uma análise apropriada do tripleto mais instável. / The modulational approach provides a simplified structure to describe the nonlinear interaction of three waves (triplet) in regimes where the carriers frequencies are much larger than the modulational frequency. In the first part of this work we investigate the breakdown of the traditional modulational approximation in the triplet interaction, which is accurate when the coupling between the three waves is weak. We examine the types of dynamics arising when the coupling is increased from very small values to larger values and we detect an abrupt transition for large values where the excursions of the limited amplitude of the modulational regime reach larger regions of the phase space. In the second part, we study a multimode extension of the non-linear interaction of three waves and we see that the limits of the modulational approximation can be indeed very restrictive. For very small couplings we show that all modes display slow amplitude modulations, but as we increase the coupling a transition happens and the modes change to a new dynamical regime where none of them can be seen as a slowly modulated high-frequency harmonic carrier. Estimates for the critical coupling and relax times have been obtained based on proper analysis of the most unstable triplet.

This thesis concerns two topics: the kinematics of Pacific cross-equatorial flow ??? the location, timing and magnitude of the flow; and their dynamics???what are the driving forces controlling the flow? Despite extensive observations in the central and eastern Pacific, observations of these flows remain contradictory. We use output from an Ocean General Circulation Model (OGCM) viewed from a Lagrangian framework on density layers. This addresses the problem of high variability due to features such as Tropical Instability Waves. The annual mean flow is found to be southward nearly everywhere, east of 140??W. Flow becomes stronger in the second half of the year due to a bolus transport of very light surface water, introduced by Tropical Instability Waves. A Tropical Cell pattern occurs along the equator that does not require diapycnal downwelling. From 160??E to 160??W the annual mean flow is northward, occurring mostly in the mixed layer, appearing to originate partly from the Equatorial Undercurrent surfacing in the east. The northward flow is strongest in March and becomes southward in September. The wind stress and nonlinear terms are shown to be the key driving features, with a prescribed biharmonic Smagorinsky horizontal friction scheme having negligible impact. From 160??E to 160??W, the flow is partly accounted for by an Ekman forcing, with the curl of the nonlinear term providing a crucial additional torque, more than doubling the magnitude in some instances. From 160??W to 120??W the wind stress curl provides a weak southward flow of about 1 Sv, which increases by the nonlinear addition to around 5 Sv. The curl of the steady component of the nonlinear term, derived from annual mean currents, is similar in structure to the total nonlinear term, but higher in magnitude. The structure of the variable term, which was mostly of opposite sign to the steady term, suggests damping occurs in place of friction. While our study is limited to an examination of the model's characteristics, our results provide important clues to the observed flow patterns not resolved by present-day measurements. This study also highlights the importance of time-space variability and both horizontal and vertical density structure in controlling the flow and its feedback on the system.

Thesis (M. Phil.)--University of Hong Kong, 2006. / Title proper from title frame. Also available in printed format.

Le travail présenté ici traite de l'interaction évolutive en temps entre un fluide incompressible et une structure élastique et s'attache à construire une modélisation mathématique rigoureuse qui conduit à une mise en oeuvre numérique efficace même dans le cas tridimensionnel. Le fluide est modélisé par l'équation évolutive de Stokes et la structure est supposée linéairement élastique. Deux modèles mathématiques pour la résolution découplée du problème fluide structure sont présentés. Ces modèles sont bien posés et par l'intermédiaire des éléments finis mixtes pour la discrétisation en espace et des différences finies pour la discrétisation en temps permettent l'écriture d'un algorithme de résolution d'implémentation relativement aisée fournissant le déplacement et la vitesse de la structure, la vitesse d'écoulement, la pression du fluide et les forces d'interface. Les résultats numériques sont très satisfaisants.

Le cadre des systèmes dynamiques lagrangiens non-réguliers est issu de<br />l'analyse des contacts non-permanents entre des solides parfaitement<br />rigides. Il nous amène à travailler avec des outils mathématiques<br />inhabituels en automatique, comme des vitesses à variations localement<br />bornées, ou des équations différentielles de mesures. L'automatique de<br />ces systèmes dynamiques commence tout juste à apparaître et les<br />théories élémentaires, comme celle de la stabilité au sens de<br />Lyapunov, nécessitent encore d'être établies.<br /><br />Dans ce travail nous proposons donc d'établir les premières bases<br />permettant l'analyse de la stabilité des systèmes dynamiques<br />non-réguliers. Nous montrons qu'il est possible, sous réserve parfois<br />d'hypothèses supplémentaires, d'étendre certains résultats classiques.<br />Nous proposons par exemple un théorème de stabilité au sens de<br />Lyapunov et une extension du théorème de LaSalle pour des systèmes<br />dynamiques décrits par des flots pouvant subir des discontinuités.<br /><br />Dans la cas des systèmes dynamiques lagrangiens non-réguliers, ces<br />résultats de stabilité peuvent s'écrire sous la forme d'un théorème de<br />Lagrange-Dirichlet, en montrant que leur énergie correspond<br />naturellement à une fonction de Lyapunov. Ces résultats sont ensuite<br />appliqués pour l'étude de la stabilité d'une régulation en position et<br />en force d'un bras manipulateur et d'un robot marcheur sans aucune<br />supposition sur l'état des contacts. Nous soulignons également<br />l'intérêt des commandes basées sur la passivité pour les systèmes<br />dynamiques lagrangiens non-réguliers

<p>Methodologies for the analysis and computer simulations of active optimal vibration control of complex elastic structures are considered. The structures, generally represented by a large number of degrees of freedom (DOF), are to be controlled by a comparatively small number of actuators.</p><p>Various techniques presently available to solve the optimal control problems are briefly discussed. A Parametric optimization technique that is versatile enough to solve almost any type of optimization problems is found to give poor accuracy and is time consuming. More promising is the optimality equations approach, which is based on Pontryagins principle. Several new numerical procedures are developed using this approach. Most of the problems in this thesis are analysed in the modal space. Even complex structures can be approximated accurately in the modal space by using only few modes. Different techniques have been first applied to the cases where the number of modes to control was the same as the number of actuators (determined optimal control problems), then to cases in which the number of modes to control is larger than the number of actuators (overdetermined optimal control problems). </p><p>The determined optimal control problems can be solved by applying the Independent Modal Space Control (IMSC) approach. Such an approach is implemented in the Beam Analogy (BA) method that solves the problem numerically by applying the Finite Element Method (FEM). The BA, which uses the ANSYS program, is numerically very efficient. The effects of particular optimization parameters involved in BA are discussed in detail. Unsuccessful attempts have been made to modify this method in order to make it applicable for solving overdetermined or underactuated problems. </p><p>Instead, a new methodology is proposed that uses modified optimality equations. The modifications are due to the extra constraints present in the overdetermined problems. These constraints are handled by time dependent Lagrange multipliers. The modified optimality equations are solved by using symbolic differential operators. The corresponding procedure uses the MAPLE programming, which solves overdetermined problems effectively despite of the high order of differential equations involved.</p><p>The new methodology is also applied to the closed loop control problems, in which constant optimal gains are determined without using Riccatis equations.</p>

In this thesis, we investigate two expressions for symmetric group characters: Kerov?s universal character polynomials and Stanley?s character polynomials. We give a new explicit form for Kerov?s polynomials, which exactly evaluate the characters of the symmetric group scaled by degree and a constant. We use this explicit expression to obtain specific information about Kerov polynomials, including partial answers to positivity questions. We then use the expression obtained for Kerov?s polynomials to obtain results about Stanley?s character polynomials.

<p>Methodologies for the analysis and computer simulations of active optimal vibration control of complex elastic structures are considered. The structures, generally represented by a large number of degrees of freedom (DOF), are to be controlled by a comparatively small number of actuators.</p><p>Various techniques presently available to solve the optimal control problems are briefly discussed. A Parametric optimization technique that is versatile enough to solve almost any type of optimization problems is found to give poor accuracy and is time consuming. More promising is the optimality equations approach, which is based on Pontryagins principle. Several new numerical procedures are developed using this approach. Most of the problems in this thesis are analysed in the modal space. Even complex structures can be approximated accurately in the modal space by using only few modes. Different techniques have been first applied to the cases where the number of modes to control was the same as the number of actuators (determined optimal control problems), then to cases in which the number of modes to control is larger than the number of actuators (overdetermined optimal control problems). </p><p>The determined optimal control problems can be solved by applying the Independent Modal Space Control (IMSC) approach. Such an approach is implemented in the Beam Analogy (BA) method that solves the problem numerically by applying the Finite Element Method (FEM). The BA, which uses the ANSYS program, is numerically very efficient. The effects of particular optimization parameters involved in BA are discussed in detail. Unsuccessful attempts have been made to modify this method in order to make it applicable for solving overdetermined or underactuated problems. </p><p>Instead, a new methodology is proposed that uses modified optimality equations. The modifications are due to the extra constraints present in the overdetermined problems. These constraints are handled by time dependent Lagrange multipliers. The modified optimality equations are solved by using symbolic differential operators. The corresponding procedure uses the MAPLE programming, which solves overdetermined problems effectively despite of the high order of differential equations involved.</p><p>The new methodology is also applied to the closed loop control problems, in which constant optimal gains are determined without using Riccatis equations.</p>

Thèse de doctorat : sciences. Mécanique : Ecully, Ecole centrale de Lyon : 2002. / 103 réf.