Orientadores: Margarida Pinheiro Mello, Maria Aparecida Diniz Ehrhardt / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matemática Estatística e Computação Científica / Made available in DSpace on 2018-08-22T15:49:27Z (GMT). No. of bitstreams: 1 Ferreira_DaianeGoncalves_M.pdf: 1441315 bytes, checksum: 1196d8b21c6254dbdd0e0d68266fa707 (MD5) Previous issue date: 2013 / Resumo: Métodos de Otimização de terceira ordem, embora de longa tradição, eram considerados, até passado recente, impraticáveis, devido à taxa com que o esforço computacional cresce em função da dimensão do problema. Avanços no desenvolvimento de estruturas de dados, rotinas que trabalham com estas estruturas e a exploração da esparsidade de grande parte dos problemas encontrados na prática já permitem implementações destes métodos que podem torná-los competitivos com métodos de segunda ordem. O objeto desta dissertação é a apresentação do método de Halley, um método de terceira ordem, sua implementação em MATLAB e a realização de testes computacionais, visando uma comparação empírica de sua eficiência frente ao método de Newton, o método de segunda ordem mais empregado na atualidade / Abstract: Higher order optimization methods, though of long-standing tradition, until recently have been deemed impractical, due to the rate of increase of the computational effort as a function of the size of the problem. Advances in the development of data structures, routines that work with these structures and the use of the sparsity of a vast range of practical problems have led to implementations of these methods that are competitive with second order methods. The object of this dissertation is the study of Halley's method, a thirdorder method, the development of a MATLAB implementation thereof and its testing, aiming at an empirical comparison of its efficiency against that of Newton's method, the second-order method most widely used today / Mestrado / Matematica Aplicada / Mestra em Matemática Aplicada

Les solveurs de calcul en mécanique des fluides numérique (solveurs CFD) ont atteint leur maturité en termes de précision et d'efficacité de calcul. Toutefois, des progrès restent à faire pour les écoulements instationnaires surtout lorsqu'ils sont régis par de grandes structures cohérentes. Pour ces écoulements, les solveurs CFD actuels n'apportent pas de solutions assez précises à moins d'utiliser des maillages très fins. De plus, la haute précision est une caractéristique cruciale pour l'application des stratégies avancées de simulation de turbulence, comme la Simulation des Grandes Echelles (LES). Afin d'appliquer les méthodes d'ordre élevé pour les écoulements instationnaires complexes plusieurs points doivent être abordés dont la robustesse numérique et la capacité à gérer des géométries complexes.Dans cette thèse, nous étudions une famille d'approximations compactes qui offrent une grande précision non pour chaque dérivée spatiale traitée séparement mais pour le résidu r complet, c'est à dire la somme de tous les termes des équations considérées. Pour des problèmes stationnaires résolus par avancement temporelle, r est le résidu à l'état stationnaire ne comprenant que des dérivées spatiales; pour des problèmes instationnaires r comprend également la dérivée temporelle. Ce type de schémas sont appelés schémas Compacts Basés sur le Résidu (RBC). Plus précisément, nous développons des schémas RBC d'ordre élevé pour des écoulements instationnaires compressibles, et menons une étude approfondie de leurs propriétés de dissipation. Nous analysons ensuite les erreurs de dissipation et la dispersion introduites par les schémas RBC afin de quantifier leur capacité à résoudre une longueur d'onde donnée en utilisant un nombre minimal de points de maillage. Les capacités de la dissipation de RBC à drainer seulement l'énergie aux petites échelles sous-résolues sont également examinées en vue de l'application des schémas RBC pour des simulations LES implicites (ILES). Enfin, les schémas RBC sont étendus à la formulation de type volumes finis (FV) afin de gérer des géométries complexes. Une formulation FV des schémas RBC d'ordre trois préservant une précision d'ordre élevé sur des maillages irréguliers est présentée et analysée. Des applications numériques, dont la simulation d'écoulements instationnaires complexes de turbomachines régis par les équations de Navier-Stokes moyennées et des simulations ILES d'écoulements turbulents dominés par des structures cohérentes dynamiques ou en décroissance, confirment les résultats théoriques. / Computational Fluid Dynamics (CFD) solvers have reached maturity in terms of solution accuracy as well as computational efficiency. However, progress remains to be done for unsteady flows especially when governed by large, coherent structures. For these flows, current CFD solvers do not provide accurate solutions unless very fine mesh are used. Moreover, high-accuracy is a crucial feature for the application of advanced turbulence simulation strategies, like Large Eddy Simulation (LES). In order to apply high-order methods to complex unsteady flows several issues needs to be addressed among which numerical robustness and the capability of handling complex geometries.In the present work, we study a family of compact approximations that provide high accuracy not for each space derivative treated apart but for the complete residual r, i.e. the sum of all of the terms in the governing equations. For steady problems solved by time marching, r is the residual at steady state and it involves space derivatives only; for unsteady problems, r also includes the time derivative. Schemes of this type are referred-to as Residual-Based Compact (RBC). Precisely, we design high-order finite difference RBC schemes for unsteady compressible flows, and provide a comprehensive study of their dissipation properties. The dissipation and dispersion errors introduced by RBC schemes are investigated to quantify their capability of resolving a given wave length using a minimal number of grid-points. The capabilities of RBC dissipation to drain energy only at small, ill-resolved scales are also discussed in view of the application of RBC schemes to implicit LES (ILES) simulations. Finally, RBC schemes are extended to the Finite Volume (FV) framework in order to handle complex geometries. A high-order accuracy preserving FV formulation of the third-order RBC scheme for general irregular grids is presented and analysed. Numerical applications, including complex Reynolds-Averaged Navier-Stokes unsteady simulation of turbomachinery flows and ILES simulations of turbulent flows dominated by coherent structure dynamics or decay, support the theoretical results.

High-order accurate finite difference methods have been applied to the acoustic wave equation in discontinuous media and curvilinear geometries, using the SBP-SAT method. Strict stability is shown for the 2-D wave equation with general boundary conditions. The fourth-order accurate method for the 3-D wave equation has been implemented in C and parallelized using MPI. The implementation has been verified against an analytical solution and runs efficiently on a large number of processors.

The National Education Department of South Africa has mandated a policy of outcomes-based education for all learners and educators in this country. Two of the most important principles of outcomes-based education are collaborative work in groups and continuous assessment by the teacher and peers. In Computer Studies, taken on the higher grade, learners are expected to construct algorithms and programs by themselves. In the real world such algorithms and programs would be constructed by groups of people working together. The researcher’s purpose of conducting this study was to breach the gap that exists between what is done in accordance with the outcomes-based curriculum in schools - and what is expected in the real world where collaborative work is the norm. The researcher used Bloom’s high-order thinking skills as his point of departure for this study and examined the implications of how they contribute to real-world situations in the school environment. To evaluate the South African curriculum for Computer Studies on the higher grade, the researcher compared the South African curriculum was the curriculum used in Australia for learners of the same age group. The results led to an intervention in which South African learners were examined on high-order thinking skills and programming in the real world. / Dissertation (MEd (CIE))--University of Pretoria, 2005. / Curriculum Studies / unrestricted

Wave phenomena appear in many fields of science such as acoustics, geophysics, and quantum mechanics. They can often be described by partial differential equations (PDEs). As PDEs typically are too difficult to solve by hand, the only option is to compute approximate solutions by implementing numerical methods on computers. Ideally, the numerical methods should produce accurate solutions at low computational cost. For wave propagation problems, high-order finite difference methods are known to be computationally cheap, but historically it has been difficult to construct stable methods. Thus, they have not been guaranteed to produce reasonable results. In this thesis we consider finite difference methods on summation-by-parts (SBP) form. To impose boundary and interface conditions we use the simultaneous approximation term (SAT) method. The SBP-SAT technique is designed such that the numerical solution mimics the energy estimates satisfied by the true solution. Hence, SBP-SAT schemes are energy-stable by construction and guaranteed to converge to the true solution of well-posed linear PDE. The SBP-SAT framework provides a means to derive high-order methods without jeopardizing stability. Thus, they overcome most of the drawbacks historically associated with finite difference methods. This thesis consists of three parts. The first part is devoted to improving existing SBP-SAT methods. In Papers I and II, we derive schemes with improved accuracy compared to standard schemes. In Paper III, we present an embedded boundary method that makes it easier to cope with complex geometries. The second part of the thesis shows how to apply the SBP-SAT method to wave propagation problems in acoustics (Paper IV) and quantum mechanics (Papers V and VI). The third part of the thesis, consisting of Paper VII, presents an efficient, fully explicit time-integration scheme well suited for locally refined meshes.

We present an Embedded Boundary with Adaptive Mesh Refinement technique for solving the compressible Navier Stokes equations in arbitrary complex domains; followed by a numerical studies for the effect of circular cylinders on the transient dynamics of the Richtmyer-Meshkov Instability. A PDE multidimensional extrapolation approach is used to reconstruct the solution in the ghost-fluid regions and imposing boundary conditions on the fluid-solid interface, coupled with a multi-dimensional algebraic interpolation for freshly cleared cells. The Navier Stokes equations are numerically solved by the second order multidimensional upwind method. Block-structured AMR, implemented with the Chombo framework, is utilized to reduce the computational cost while keeping high resolution mesh around the Embedded Boundary and regions of high gradient solutions. The versatility of the method is demonstrated via several numerical examples, in both static and moving geometry, ranging from low Mach number nearly incompressible to supersonic flows. Our simulation results are extensively verified against other numerical results and validated against available experimental results where applicable. The effects on the transient dynamics of the Richtmyer-Meshkov instability due to small scale perturbations introduced on the shock-wave or the material interface by a single or set of solid circular cylinders were computationally investigated using the developed technique. First, we discuss the RMI initiated on a flat interface by a rippled shock-wave that is disturbed by a single circular cylinder. Then, we study the effect of introducing a number of circular cylinders on the interface. The arrangement of the cylinders set mimic (in a two dimensional domain) the presence of the solid supporting grid wires used in the formation of the material interface in the experimental setup. We analyzed their effects on the mixing layer growth and the mixedness level, and qualitatively demonstrate the cylinders' perturbation effects on the mixing layer structure. We modeled the cylinders' influence based on their diameters; and showed the model ability to predict the variation of the mixing layer growth for different flow parameters.

A Direct Numerical Simulation of an iced Natural Laminar Flow NLF-0414 airfoil is carried out using a high-order spectral element method for low chord Reynolds numbers (O(10^5)). This study aims to advance the state-of-the-art for accurate computational modeling of transition, iced airfoil aerodynamics, and irregular surface spectral element method Direct Numerical Simulation. Ice accretion over an aircraft, ranging from light to severe, changes the aerodynamic profile of the airfoil and alters the overall performance. The literature presents simulations that have been carried out with a range of turbulence models which fail to accurately capture the complex physics of these flows. The iced profiles being studied, Run 606 and 622-2D, were obtained from a Technical Publication by NASA on iced airfoils including the NLF-0414, and were selected as they are relatively lightly iced profiles of the NLF-0414. The largest bottleneck with the current advancement in High Performance Computing is the computation time required for Direct Numerical Simulation. Results such as lift, drag, pressure, and skin friction coefficients, for a clean NLF-0414 and two lightly iced NLF-0414 airfoils at chord Reynolds numbers of Rec = 1 x 10^5 and Rec = 2 x 10^5 are visualized and discussed, showing the degradation of the natural laminar flow due to ice accretion. Turbulence statistics are calculated to study the effective contributions of turbulent fluctuations in the flow to further understand the flow physics near transition. The detailed study of these six cases has led us to 1) further understand the complexities of the transition process on iced airfoils, 2) observe and explain the sometimes unexpected changes in aerodynamic performance due to varying iced geometries, and 3) establish a methodology for spectral element method Direct Numerical Simulations.

La réduction des émissions polluantes des turboréacteurs nécessite une plus grande maîtrise du dimensionnement du système d’injection du mélange air-carburant au sein de la chambre de combustion.L’objectif de la thèse est d’améliorer la compréhension de la dynamique des écoulements swirlés, rencontrés dans les chambres aéronautiques. La simulation aux grandes échelles, qui exploite les super-calculateurs les plus puissants, est devenue un outil d’analyse incontournable. Cependant, la taille des simulations et le volume de données générées rendent difficile l’extraction des phénomènes à grande échelle. A cette fin, de nouvelles méthodes de post-traitement parallèles qui permettent d’accéder à l’évolution temporelle des structures tourbillonnaires dans des géométries complexes sont proposées.Ces méthodes sont appliquées à l’étude de la dynamique de flammes swirlées diphasiques dans lesquelles les structures cohérentes interagissent avec la zone réactive et le brouillard de gouttes. / The reduction of pollutant emissions of aeronautical devices requires to optimize the design of the injection systems in the combustion chamber. The objective of this work is to improve the understandingof the flow dynamics in swirl stabilized burners. Large Eddy Simulation has become a major tool for the analysis of such flows. The steady increase in computational power enables to perform high-fidelity simulations, that generates a large amount of data, making it difficult to extract relevant information regarding the large scale phenomena. To this aim, massively parallel post-processing methods, suited for complex geometries, were developed in order to extract large-scale structures in turbulent flows. These methods were applied to simulations of spray flames in swirl burners, to get a better insight of how the large scale structures interact with the flame topology and the spray dynamics.

La chiralité est une propriété géométrique caractérisant les objets qui ne sont pas superposables à leur image dans un miroir. Nos mains en sont un exemple emblématique, puisqu’elles existent sous deux formes différentes droite et gauche. Si la chiralité s'observe à toutes les échelles de l'univers, elle joue un rôle particulièrement important en chimie. Une molécule chirale et son image miroir peuvent réagir différemment avec leur environnement et être thérapeutiques ou toxiques. Ces effets ont évidemment d'immenses répercussions sur le règne animal et végétal. Il apparaît alors clairement qu'il est essentiel d’étudier précisément les dynamiques des réactions chimiques chirales.Dans cette thèse, nous avons étudié les dynamiques ultrarapides de molécules chirales par des sources lasers de durée femtosecondes).($10^{-15}$ s). La chiralité moléculaire étant généralement difficile à détecter, nous avons ici utilisé une technique récente, le dichroïsme circulaire de photoélectrons (PECD) qui permet de générer un signal chiral très important. Nous avons ainsi observé des dynamiques moléculaires ultrarapides jusqu'à l'échelle attoseconde ($10^{-18}$ s), et mis en avant des dynamiques de relaxation et d'ionisation encore jamais observées.Parallèlement à ces études résolues en temps, nous avons développé plusieurs expériences employant une nouvelle source laser Yb fibrée à haute cadence et grande puissance moyenne. Nous avons développé une nouvelle méthode, par extension du PECD, qui nous a permis de mesurer la compositions d'échantillons chiraux rapidement avec une grande précision. Enfin, nous avons développé une ligne de lumière XUV ultrabrève de très haute brillance ($sim 2$ mW). Cette source, couplée à un détecteur de photoélectrons et photoions en coïncidence, servira à étudier les mécanismes de reconnaissance chirale. / Chirality is a geometric property that characterizes objects that cannot be superposed on their mirror image. Our hands are an emblematic example of this, since they exist in two different forms, right and left. While chirality is observed at all scales in the universe, it plays a particularly important role in chemistry. A chiral molecule and its mirror image can react differently with their environment and be therapeutic or toxic. These effects obviously have immense repercussions on the animal and plant kingdom. It then becomes clear that it is essential to study precisely the dynamics of chiral chemical reactions.In this thesis, we studied the ultrafast dynamics of chiral molecules by laser sources of femtosecond duration ($10^{-15}$ s). Molecular chirality is generally difficult to detect, so we have used a recent technique, circular photoelectron dichroism (PECD), to generate a very important chiral signal. We have thus observed ultrafast molecular dynamics at the attosecond scale ($10^{-18}$ s), and highlighted relaxation and ionization dynamics never observed before.In parallel to these time-resolved studies, we have developed several experiments using a new high repetition rate, high mean power Yb fiber laser. We have developed a new method, by extending the PECD, that has allowed us to measure the composition of chiral samples quickly and accurately. Finally, we have developed an ultra-short XUV beamline with very high brightness ($sim 2$ mW). This source, coupled with a photoelectron and photoion coincidence detector, will be used to study chiral recognition mechanisms.

Arbitrary high order numerical methods for time-harmonic acoustic scattering problems originally defined on unbounded domains are constructed. This is done by coupling recently developed high order local absorbing boundary conditions (ABCs) with finite difference methods for the Helmholtz equation. These ABCs are based on exact representations of the outgoing waves by means of farfield expansions. The finite difference methods, which are constructed from a deferred-correction (DC) technique, approximate the Helmholtz equation and the ABCs to any desired order. As a result, high order numerical methods with an overall order of convergence equal to the order of the DC schemes are obtained. A detailed construction of these DC finite difference schemes is presented. Details and results from an extension to heterogeneous media are also included. Additionally, a rigorous proof of the consistency of the DC schemes with the Helmholtz equation and the ABCs in polar coordinates is also given. The results of several numerical experiments corroborate the high order convergence of the proposed method. A novel local high order ABC for elastic waves based on farfield expansions is constructed and preliminary results applying it to elastic scattering problems are presented.