This thesis proposes a model reduction technique for nonlinear dynamical systems based upon combining Proper Orthogonal Decomposition (POD) and a new method, called the Discrete Empirical Interpolation Method (DEIM). The popular method of Galerkin projection with POD basis reduces dimension in the sense that far fewer variables are present, but the complexity of evaluating the nonlinear term generally remains that of the original problem. DEIM, a discrete variant of the approach from [11], is introduced and shown to effectively overcome this complexity issue. State space error estimates for POD-DEIM reduced systems are also derived. These [Special characters omitted.] error estimates reflect the POD approximation property through the decay of certain singular values and explain how the DEIM approximation error involving the nonlinear term comes into play. An application to the simulation of nonlinear miscible flow in a 2-D porous medium shows that the dynamics of a complex full-order system of dimension 15000 can be captured accurately by the POD-DEIM reduced system of dimension 40 with a factor of [Special characters omitted.] (1000) reduction in computational time.

We investigate the limits of localized linear control in spatially extended, nonlinear systems. Spatially extended, nonlinear systems can be found in virtually every field of engineering and science. An important category of such systems are fluid flows. Fluid flows play an important role in many commercial applications, for instance in the chemical, pharmaceutical and food-processing industries. Other important fluid flows include air- or water flows around cars, planes or ships. In all these systems, it is highly desirable to control the flow of the respective fluid. For instance control of the air flow around an airplane or car leads to better fuel-economy and reduced noise production. Usually, it is impossible to apply control everywhere. Consider an airplane: It would not be feasibly to cover the whole body of the plane with control units. Instead, one can place the control units at localized regions, such as points along the edge of the wings, spaced as far apart from each other as possible. These considerations lead to an important question: For a given system, what is the minimum number of localized controllers that still ensures successful control? Too few controllers will not achieve control, while using too many leads to unnecessary expenses and wastes resources. To answer this question, we study localized control in a class of model equations. These model equations are good representations of many real fluid flows. Using these equations, we show how one can design localized control that renders the system stable. We study the properties of the control and derive several expressions that allow us to determine the limits of successful control. We show how the number of controllers that are needed for successful control depends on the size and type of the system, as well as the way control is implemented. We find that especially the nonlinearities and the amount of noise present in the system play a crucial role. This analysis allows us to determine under which circumstances a given number of controllers can successfully stabilize a given system.

Thesis (M.S.)--University of North Carolina at Wilmington, 2003. / Includes bibliographical references (leaf : [37]).

Nonlinear mixing of two collinear, initially monochromatic, Rayleigh waves propagating in the same direction in an isotropic, nonlinear elastic solid is investigated: analytically, by finite element method simulations and experimentally. In the analytical part, it is shown that only collinear mixing in the same direction fulfills the phase matching condition based on Jones and Kobett 1963 for the resonant generation of the second harmonics, as well as the sum and difference frequency components caused by the interaction of the two fundamental waves. Next, a coupled system of ordinary differential equations is derived based on the Lagrange equations of the second kind for the varying amplitudes of the higher harmonic and combination frequency components of the fundamentals waves. Numerical results of the evolution of the amplitudes of these frequency components over the propagation distance are provided for different ratios of the fundamental wave frequencies. It is shown that the energy transfer is larger for higher frequencies, and that the oscillation of the energy between the different frequency components depends on the amplitudes and frequencies of the fundamental waves. Furthermore, it is illustrated that the horizontal velocity component forms a shock wave while the vertical velocity component forms a pulse in the case of low attenuation. This behavior is independent of the two fundamental frequencies and amplitudes that are mixed. The analytical model is then extended by implementing diffraction effects in the parabolic approximation. To be able to quantify the acoustic nonlinearity parameter, β, general relations based on the plane wave assumption are derived. With these relations a β is expressed, that is analog to the β for longitudinal waves, in terms of the second harmonics and the sum and the difference frequencies. As a next step, frequency and amplitude ratios of the fundamental frequencies are identified, which provide a maximum amplitude of one of the second harmonics as well as the sum or difference frequency components to enhance experimental results. Subsequently, the results of the analytical model are compared to those of finite element method simulations. Two dimensional simulations for small propagation distances gave similar results for analytical and finite element simulations. Consquently. this shows the validity of the analytical model for this setup. In order to demonstrate the feasibility of the mixing technique and of the models, experiments were conducted using a wedge transducer to excite mixed Rayleigh waves and an air-coupled transducer to detect the fundamentals, second harmonics and the sum frequency. Thus, these experiments yield more physical information compared to the case of using a single fundamental wave. Further experiments were conducted that confirm the modeled dependence on the amplitudes of the generated waves. In conclusion, the results of this research show that it is possible to measure the acoustic nonlinearity parameter β to quantify material damage by mixing Rayleigh waves on up to four ways.

Thesis (M. Phil.)--University of Hong Kong, 1994. / Includes bibliographical references (leaves 102-120).

Thesis (Ph.D.) - University of Queensland, 2003. / Includes bibliography.

- / -

Orientadores: Marcia Assumpção Guimarães Scialom, Felipe Linares / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Matematica, Estatistica e Computação Cientifica / Made available in DSpace on 2018-08-10T05:23:45Z (GMT). No. of bitstreams: 1 Braganca_LucianaMariaMendonca_D.pdf: 688598 bytes, checksum: af92575e77efea7c1bdd4a624c7d70dc (MD5) Previous issue date: 2007 / Resumo: Neste trabalho estudamos o problema de Cauchy associado a um sistema de Equações do tipo Schrodinger não linear de terceira ordem. Obtemos resultados de boa colocação local para o problema, com dado inicial nos espaços de Sobolev Hs(R) x Hs(R), s '> ou =' 1/4 e no caso períodico em Hs(T)xHs(T), s '> ou =' 1/2. No caso particular'sigma' 'alfa' = 'sigma' 'beta' = 'sigma''mu' = 1 obtemos resultados de boa colocação global em Hs(R) x Hs(R), 3/5 < s '> ou = 1 e H1(T) x H1(T). Mostramos também um resultado de má colocação para o problema com dado inicial em Hs(R) x Hs(R), -1/2 < s < 1/4 / Abstract: In this work we study the Cauchy problem associated to a system of coupled third-order nonlinear Schrodinger equation. We establish local well-posedness results for the problem with data in Sobolev spaces Hs(R) x Hs(R), s '> or =' 1/4 and in the periodic case Hs(T)xHs(T), s '> or =' 1/2. In the particular case ... Note: The complete abstract is available with the full electronic digital thesis or dissertations / Doutorado / Doutor em Matemática

Orientador: Jaime Angulo Pava / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Matematica, Estatistica e Computação Cientifica / Made available in DSpace on 2018-08-08T09:23:36Z (GMT). No. of bitstreams: 1 Natali_FabioMatheusAmorin_D.pdf: 1495954 bytes, checksum: 470d5cbae62b9e11fa888d6ab8e61976 (MD5) Previous issue date: 2007 / Resumo: Nesta tese estabelecemos condições suficientes para obter a estabilidade de soluções ondas Viajantes periódicas para equações de tipo KdV-type + ut +upux -- (Mu)x = 0, p ? N, com M sendo um operador pseudo-diferencial geral, porem com características especiais. Nossa abordagem é a de usar a teoria dos operadores totalmente positivos, o Teorema do Somatório de Poisson e a teoria das funções Elípticas de Jacobi. Em particular nós obtemos a estabilidade de uma família de soluções ondas viajantes periódicas para a equação de Benjamin-Ono e a equação KdV crítica. Nossas técnicas fornecem uma nova maneira para obter a existência e a estabilidade das ondas cnoidal e dnoidal associadas as equações de Korteweg-de Vries e modificada Korteweg-de Vries respectivamente. A teoria propõe o estudo de soluções ondas viajantes periódicas para outras equações diferencias parciais por exemplo, os resultados de estabilidade e instabilidade de soluções do tipo standing waves periódicas para a equação não linear de Schrödinger crítica / Abstract: In this thesis we establish su?cient conditions to obtain the stability of periodic travelling waves solutions for equations of KdV-type ut + upux -- (Mu)x = 0, p N, with M being a general pseudo-differential operator, but this operator has special characteristics. Our approach use the theory of totally positive operators, the Poisson summation theorem and the theory of Jacobi elliptic functions. In particular we obtain the stability of a family of periodic travelling waves solutions for the Benjamin-Ono equation and critical Korteweg-de Vries equation. Our techniques give a new way to obtain the existence and stability of cnoidal and dnoidal waves solutions associated to the Korteweg-de Vries and modified Korteweg-de Vries equations respectively. The theory has prospects for the study of periodic travelling waves solutions of other partial diferential equations, for instance, the results of stability and instability of periodic standing wave solutions for the critical Schrödinger equation / Doutorado / Doutor em Matemática

In this thesis we study the behavior of a simple control system based on a delay differential equation with multiple loops of negative feedback. Numerical solutions of the delay differential equation with N delays d/dt x(t) = -x(t) + 1/N POT.N IND.i=1 / POT.n IND.i + x (t- IND.i) POT.n have been investigated as function of its parameters: n, i and i. A simple numerical method for determine the stability regions of the equilibrium points in the parameter space (i, n) is presented. The existence of a doubling period route to chaos in the equation, for N = 3, is characterized by the construction of bifurcation diagram with parameter n. A numerical method that uses the analysis of Poincaré sections of the reconstructed attractor to find aperiodic solutions in the parameter space of the equation is also presented. We apply this method for N = 2 and get evidences for the existence of chaotic solutions as result of a period doubling route to chaos (chaotic solutions for N = 2 in that equation had never been observed). Finally, we study the solutions of a piecewise constant equation that corresponds to the limit case n . / In this thesis we study the behavior of a simple control system based on a delay differential equation with multiple loops of negative feedback. Numerical solutions of the delay differential equation with N delays d/dt x(t) = -x(t) + 1/N POT.N IND.i=1 / POT.n IND.i + x (t- IND.i) POT.n have been investigated as function of its parameters: n, i and i. A simple numerical method for determine the stability regions of the equilibrium points in the parameter space (i, n) is presented. The existence of a doubling period route to chaos in the equation, for N = 3, is characterized by the construction of bifurcation diagram with parameter n. A numerical method that uses the analysis of Poincaré sections of the reconstructed attractor to find aperiodic solutions in the parameter space of the equation is also presented. We apply this method for N = 2 and get evidences for the existence of chaotic solutions as result of a period doubling route to chaos (chaotic solutions for N = 2 in that equation had never been observed). Finally, we study the solutions of a piecewise constant equation that corresponds to the limit case n .