Approximation and interpolation employing radial basis functions has found important applications since the early 1980's in areas such as signal processing, medical imaging, as well as neural networks. Several applications demand that certain physical properties be fulfilled, such as a function being divergence free. No such class of radial basis functions that reflects these physical properties was known until 1994, when Narcowich and Ward introduced a family of matrix-valued radial basis functions that are divergence free. They also obtained error bounds and stability estimates for interpolation by means of these functions. These divergence-free functions are very smooth, and have unbounded support. In this thesis we introduce a new class of matrix-valued radial basis functions that are divergence free as well as compactly supported. This leads to the possibility of applying fast solvers for inverting interpolation matrices, as these matrices are not only symmetric and positive definite, but also sparse because of this compact support. We develop error bounds and stability estimates which hold for a broad class of functions. We conclude with applications to the numerical solution of the Navier-Stokes equation for certain incompressible fluid flows.

We consider the application of least-squares variational principles and the finite element method to the numerical solution of boundary value problems arising in the fields of solidand fluidmechanics.For manyof these problems least-squares principles offer many theoretical and computational advantages in the implementation of the corresponding finite element model that are not present in the traditional weak form Galerkin finite element model.Most notably, the use of least-squares principles leads to a variational unconstrained minimization problem where stability conditions such as inf-sup conditions (typically arising in mixed methods using weak form Galerkin finite element formulations) never arise. In addition, the least-squares based finite elementmodelalways yields a discrete system ofequations witha symmetric positive definite coeffcientmatrix.These attributes, amongst manyothers highlightedand detailed in this work, allow the developmentofrobust andeffcient finite elementmodels for problems of practical importance. The research documented herein encompasses least-squares based formulations for incompressible and compressible viscous fluid flow, the bending of thin and thick plates, and for the analysis of shear-deformable shell structures.

[pt] Apresenta-se aqui uma aplicação do método de continuação, baseado no algoritmo de Euler-Quase Newton, a problemas de equilíbrio de materiais hiperelásticos incompressíveis sujeitos a grandes deformações. Discretiza-se o problema misto estado deformado-campo de pressão pela utilização do método dos elementos finitos, prevendo-se a compatibilidade LBB entre os espaços envolvidos. Propõe-se a utilização de uma função densidade de energia de deformação para o material de Mooney-Rivlin distinta daquela apresentada na literatura clássica, devido ao mal comportamento do Hessiano associado à formulação original. / [en] The application of a continuation method based on the Euler-Chord algorithm to equilibrium problems of incompressible, hyperelastic materials subjected to large deformations is here presented. The mixed strained state-pressure field problem is discretized by means of the finite element method, taking into account the LBB compatibility condition between the involved spaces. The utilization of a strain energy density function diverse from the one presented in the classical literature, is proposed, due to the ill behavior of the Hessian associated with the original formulation.

Non-Newtonian fluids are of great scientific interest due to their range of physical properties, which arise from the characteristic shear stress-shear rate relation for each fluid. The applications of non-Newtonian fluids are widespread and occur in many industrial (e.g. lubricants, suspensions, paints, etc.) and environmental (e.g. mud, ice, blood, etc.) problems, often involving multiple fluids. In this study, the novel technique of Incompressible Smoothed Particle Hydrodynamics (ISPH) with shifting (Lind et al., J. Comput. Phys., 231(4):1499-1523, 2012), is extended beyond the state-of-the-art to model non-Newtonian and multi-phase flows. The method is used to investigate important problems of both environmental and industrial interest. The proposed methodology is based on a recent ISPH algorithm with shifting with the introduction of an appropriate stress formulation. The new method is validated both for Newtonian and non-Newtonian fluids, in closed-channel and free-surface flows. Applications in complex moulding flows are conducted and compared to previously published results. Validation includes comparison with other computational techniques such as weakly compressible SPH (WCSPH) and the Control Volume Finite Element method. Importantly, the proposed method offers improved pressure results over state-of-the-art WCSPH methods, while retaining accurate prediction of the flow patterns. Having validated the single-phase non-Newtonian ISPH algorithm, this develops a new extension to multi-phase flows. The method is applied to both Newtonian/Newtonian and Newtonian/non-Newtonian problems. Validations against a novel semi-analytical solution of a two-phase Poiseuille Newtonian/non-Newtonian flow, the Rayleigh-Taylor instability, and a submarine landslide are considered. It is shown that the proposed method can offer improvements in the description of interfaces and in the prediction of the flow fields of demanding multi-phase flows with both environmental and industrial application. Finally, the Lituya Bay landslide and tsunami is examined. The problem is approached initially on the real length-scales and compared with state-of-the-art computational techniques. Moreover, a detailed investigation is carried out aiming at the full reproduction of the experimental findings. With the introduction of a k-ε turbulence model, a simple saturation model and correct experimental initial conditions, significant improvements over the state-of-the-art are shown, managing an accurate representation of both the landslide as well as the wave run-up. The computational method proposed in this thesis is an entirely novel ISPH algorithm capable of modelling highly deforming non-Newtonian and multi-phase flows, and in many cases shows improved accuracy and experimental agreement compared with the current state-of-the-art WCSPH and ISPH methodologies. The variety of problems examined in this work show that the proposed method is robust and can be applied to a wide range of applications with potentially high societal and economical impact.

Made available in DSpace on 2014-10-09T12:50:37Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T13:58:52Z (GMT). No. of bitstreams: 1 00621.pdf: 16649575 bytes, checksum: 34a61ca5ff67945244a79765b98ddb45 (MD5) / Tese (Doutoramento) / IEA/T / University of London

Made available in DSpace on 2014-10-09T12:50:37Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T13:58:52Z (GMT). No. of bitstreams: 1 00621.pdf: 16649575 bytes, checksum: 34a61ca5ff67945244a79765b98ddb45 (MD5) / Tese (Doutoramento) / IEA/T / University of London

Problemas de otimização se fazem cada vez mais presentes nos mais diversos ramos da Engenharia. Encontrar configurações ótimas para um determinado problema significa, por exemplo, melhorar desempenho, reduzir custos entre outros ganhos. Existem hoje diversas maneiras de atacar um problema de otimização, cada qual com suas particularidades, vantagens e desvantagens. Dentre os métodos de otimização que utilizam gradientes de sensibilidade, o cálculo numérico dos mesmos consiste em uma importante etapa do projeto que, dependendo do problema, pode acarretar em custos computacionais muito elevados inviabilizando a abordagem escolhida. Este trabalho visa desenvolver e apresentar uma nova metodologia para o cálculo desses gradientes de sensibilidade, com base no Método Adjunto. O Método Adjunto é um método amplamente estudado e com diversas aplicações principalmente em Engenharia Aeronáutica. Nesse trabalho, todo o conhecimento prévio é utilizado para a derivação do método para aplicá-lo a escoamentos viscosos e incompressíveis. É desenvolvido também o cálculo do gradiente de sensibilidade com respeito a parâmetros geométricos e não geométricos. Para validar a metodologia proposta são feitas simulações numéricas das equações governantes do escoamento e adjuntas utilizando dois códigos computacionais distintos, SEMTEX e FreeFem++, o primeiro baseado no Método dos Elementos Espectrais e o segundo no Método dos Elementos Finitos, mostrando assim a independência do Método Adjunto na sua formulação contínua em relação a métodos computacionais. Para a validação são cujos gradientes possam ser calculados de outras formas permitindo comparações para calibrar e aperfeiçoar o cálculo do gradiente de sensibilidade. / Optimization problems are widely present in differents fields of Engineering. Finding optimal configurations in a problem means, for example, improving performance, reducing costs, among other achievements. There are several wellknown ways to tackle an optimization problem, each one has its own advantages and disadvantages. Considering the gradient-based optimization methods, the step of their numerical calculation is extremely important, as it may result in huge computational costs, thus making the chosen method impracticable. This work aims to develop and present a new methodology to compute these sensitivity gradients based on the Adjoint Method. The Adjoint Method is a widely studied method with several applications chiefly in A eronautical Engineering. In the present work, all the previous knowledge will be used to derive the equations of the method in order to apply them to viscous incompressible flows. The calculation of the sensitivity gradient, with respect to both geometric and non-geometric paramatersm will be developed as well. To validate the proposed methodology, numerical simulations of the governing and adjoint equations are carried out, using two computational codes called SEMTEX and FreeFem++, the former is based on the Spectral Element Method and the later, on the Finite Element Method, thus showing that the Adjoint Method, in its continuous formulation, is independent of the particular numerical method that is used. In order to validate the algorithm, simple problems are chosen, for which the gradients can be computed by other methods. This choice admits comparison between numerical values of gradients in order to calibrate and improve the methodology proposed.

Dentre todos os sistemas dinâmicos os sistemas Anosov têm atraído a atenção de muitos matemáticos. No caso de fluxo Anosov em uma variedade fechada M de dimensão três, Sérgio Fenley definiu o conceito de losangos no recobrimento universal de M e obteve resultados importantes envolvendo losangos e automorfismos do recobrimento universal. Seguindo o que foi feito por Fenley, e utilizando o conceito de losangos no espaço das órbitas do fluxo levantado (no recobrimento universal), Thierry Barbot obteve condições suficientes para que um toro incompressível numa 3-variedade fechada suportando um fluxo Anosov seja isotópico a um outro que é transverso ao fluxo. Neste trabalho consideramos ações Anosov de \'R POT. k\' sobre uma variedade fechada M de dimensão k + 2. Primeiramente, conseguimos resultados análogos aos de Fenley (sobre existência de losangos) para estas ações, e usando isso, finalmente obtemos condições suficientes para que um toro incompressível seja isotópico a um toro transverso à ação. Este último resultado é uma generalização de Barbot mencionado acima / Among all dynamical systems the Anosov systems has attracted the attention of many mathematicians. In the case of an Anosov flow in a closed manifold M of dimension three, Sérgio Fenley defined the concept of lozenges in the universal covering of M and obtained important results involving lozenges and covering automorphism. Following what was made by Fenley, and using the concept of lozenge on the orbit space of the lifted flow (in the universal covering). Thierry Barbot obtains sufficient conditions for an incompressible torus in a closed 3-manifold supporting an Anosov flow to be isotopic to another which is transverse to flow. If this work we considered Anosov of \'R POT. k\' on a closed manifold M of dimension k + 2. First, we obtain analogous results those of Fenley (about existence of lozenges) for this actions, and using this, finally we obtain sufficient conditions for an incompressible torus to be isotopic to another torus which is transverse to action. This last result is a generalization of Barbot\'s result mentioned above

Context. Simulating realistic fluid behavior in incompressible fluids for computer graphics has been pioneered with the implicit incompressible smoothed particle hydrodynamics (IISPH) solver. The algorithm converges faster than other incompressible SPH-solvers, but real-time performance (in the perspective of video games, 30 frames per second) is still an issue when the particle count increases. Objectives. This thesis aims at improving the performance of the IISPH-solver by proposing a parallel solution that runs on the GPU using CUDA. The solution should not compromise the physical accuracy of the original solution. Investigated aspects are execution time, memory usage and physical accuracy. Methods. The proposed implementation uses a fine-grained approach where each particle is calculated on a separate thread. It is compared to a sequential and a parallel OpenMP implementation running on the CPU. Results and Conclusions. It is shown that the parallel CUDA solution allow for real-time performance for approximately 19 times the amount of particles than that of the sequential implementation. For approximately 175 000 particles the simulation runs at the constraint of real-time performance, more particles are still considered interactive. The visual result of the proposed implementation deviated slightly from the ones on the CPU.

We investigate the Richtmyer-Meshkov instability (RMI) that occurs when an incident shock impulsively accelerates the interface between two different fluids. RMI is important in many technological applications such as Inertial Confinement Fusion (ICF) and astrophysical phenomena such as supernovae. We consider RMI in the presence of the magnetic field in converging geometry through both simulations and analytical means in the framework of ideal magnetohydrodynamics (MHD). In this thesis, we perform linear stability analyses via simulations in the cylindrical geometry, which is of relevance to ICF. In converging geometry, RMI is usually followed by the Rayleigh-Taylor instability (RTI). We show that the presence of a magnetic field suppresses the instabilities. We study the influence of the strength of the magnetic field, perturbation wavenumbers and other relevant parameters on the evolution of the RM and RT instabilities. First, we perform linear stability simulations for a single interface between two different fluids in which the magnetic field is normal to the direction of the average motion of the density interface. The suppression of the instabilities is most evident for large wavenumbers and relatively strong magnetic fields strengths. The mechanism of suppression is the transport of vorticity away from the density interface by two Alfv ́en fronts. Second, we examine the case of an azimuthal magnetic field at the density interface. The most evident suppression of the instability at the interface is for large wavenumbers and relatively strong magnetic fields strengths. After the shock interacts with the interface, the emerging vorticity breaks up into waves traveling parallel and anti-parallel to the magnetic field. The interference as these waves propagate with alternating phase causing the perturbation growth rate of the interface to oscillate in time. Finally, we propose incompressible models for MHD RMI in the presence of normal or azimuthal magnetic field. The linearized equations are solved numerically using inverse Laplace transform. The incompressible models show that the magnetic field suppresses the RMI, and the mechanism of this suppression depends on the orientation of the initially applied magnetic field. The incompressible model agrees reasonably well with compressible linear simulations.