Waves and circulation processes interact in daily wind and tide driven flows as well as in more extreme events such as hurricanes. Currents and water levels affect wave propagation and the location of wave-breaking zones, while wave forces induce setup and currents. Despite this interaction, waves and circulation processes are modeled separately using different approaches. Circulation processes are represented by the shallow water equations, which conserve mass and momentum. This approach for wind-generated waves is impractical for large geographic scales due to the fine resolution that would be required. Therefore, wind-waves are instead represented in a spectral sense, governed by the action balance equation, which propagates action density through both geographic and spectral space. Even though wind-waves and circulation are modeled separately, it is important to account for their interactions by coupling their respective models. In this dissertation we use discontinuous-Galerkin (DG) methods to couple spectral wave and circulation models to model wave-current interactions. We first develop, implement, verify and validate a DG spectral wave model, which allows for the implementation of unstructured meshes in geographic space and the utility of adaptive, higher-order approximations in both geographic and spectral space. We then couple the DG spectral wave model to an existing DG circulation model, which is run on the same geographic mesh and allows for higher order information to be passed between the two models. We verify and validate coupled wave/circulation model as well as analyzing the error of the coupled wave/circulation model. / text

NR 20140805

Automation of finite element analysis based on a fully adaptive <i>p</i>-refinement procedure can reduce the magnitude of discretization error to the desired accuracy with minimum computational cost and computer resources. This study aims to 1) develop an efficient <i>p</i>-refinement procedure with a non-uniform <i>p</i> analysis capability for solving 2-D and 3-D elastostatic mechanics problems, and 2) introduce a stress error estimate. An element-by-element algorithm that decides the appropriate order for each element, where element orders can range from 1 to 8, is described. Global and element data structures that manage the complex data generated during the refinement process are introduced. These data structures are designed to match the concept of object-oriented programming where data and functions are managed and organized simultaneously. The stress error indicator introduced is found to be more reliable and to converge faster than the error indicator measured in an energy norm called the residual method. The use of the stress error indicator results in approximately 20% fewer degrees of freedom than the residual method. Agreement of the calculated stress error values and the stress error indicator values confirms the convergence of final stresses to the analyst. The error order of the stress error estimate is postulated to be one order higher than the error order of the error estimate using the residual method. The mapping of a curved boundary element in the working coordinate system from a square-shape element in the natural coordinate system results in a significant improvement in the accuracy of stress results. Numerical examples demonstrate that refinement using non-uniform <i>p</i> analysis is superior to uniform <i>p</i> analysis in the convergence rates of output stresses or related terms. Non-uniform <i>p</i> analysis uses approximately 50% to 80% less computational time than uniform <i>p</i> analysis in solving the selected stress concentration and stress intensity problems. More importantly, the non-uniform <i>p</i> refinement procedure scales the number of equations down by 1/2 to 3/4. Therefore, a small scale computer can be used to solve equation systems generated using high order <i>p</i>-elements. In the calculation of the stress intensity factor of a semi-elliptical surface crack in a finite-thickness plate, non-uniform <i>p</i> analysis used fewer degrees of freedom than a conventional <i>h</i>-type element analysis found in the literature. / Ph. D.

Many physical problems lead to boundary value problems for partial differential equations, which can be solved with the finite element method. In order to construct adaptive solution algorithms or to measure the error one aims at reliable a posteriori error estimators. Many such estimators are known, as well as their theoretical foundation. Some boundary value problems yield so-called anisotropic solutions (e.g. with boundary layers). Then anisotropic finite element meshes can be advantageous. However, the common error estimators for isotropic meshes fail when applied to anisotropic meshes, or they were not investigated yet. For rectangular or cuboidal anisotropic meshes a modified error estimator had already been derived. In this paper error estimators for anisotropic tetrahedral or triangular meshes are considered. Such meshes offer a greater geometrical flexibility. For the Poisson equation we introduce a residual error estimator, an estimator based on a local problem, several Zienkiewicz-Zhu estimators, and an L_2 error estimator, respectively. A corresponding mathematical theory is given.For a singularly perturbed reaction-diffusion equation a residual error estimator is derived as well. The numerical examples demonstrate that reliable and efficient error estimation is possible on anisotropic meshes. The analysis basically relies on two important tools, namely anisotropic interpolation error estimates and the so-called bubble functions. Moreover, the correspondence of an anisotropic mesh with an anisotropic solution plays a vital role. AMS(MOS): 65N30, 65N15, 35B25

O trabalho aborda a questão das estimativas a posteriori dos erros de discretização e particularmente a recuperação dos gradientes de soluções numéricas obtidas com o método dos elementos finitos (MEF) e com o método dos elementos finitos generalizados (MEFG). Inicialmente, apresenta-se, em relação ao MEF, um resumido estado da arte e conceitos fundamentais sobre este tema. Em seguida, descrevem-se os estimadores propostos para o MEF denominados Estimador Z e \"Superconvergent Patch Recovery\" (SPR). No âmbito do MEF propõe-se de modo original a incorporação do \"Singular Value Decomposition\" (SVD) ao SPR aqui mencionada como SPR Modificado. Já no contexto do MEFG, apresenta-se um novo estimador do erro intitulado EPMEFG, estendendo-se para aquele método as idéias do SPR Modificado. No EPMEFG, a função polinomial local que permite recuperar os valores nodais dos gradientes da solução tem por suporte nuvens (conjunto de elementos finitos que dividem um nó comum) e resulta da aplicação de um critério de aproximação por mínimos quadrados em relação aos pontos de superconvergência. O número destes pontos é definido a partir de uma análise em cada elemento que compõe a nuvem, considerando-se o grau da aproximação local do campo de deslocamentos enriquecidos. Exemplos numéricos elaborados com elementos lineares triangulares e quadrilaterais são resolvidos com o Estimador Z, o SPR Modificado e o EPMEFG para avaliar a eficiência de cada estimador. Essa avaliação é realizada mediante o cálculo dos índices de efetividade. / The paper addresses the issue of a posteriori estimates of discretization errors and particularly the recovery of gradients of numerical solutions obtained with the finite element method (FEM) and the generalized finite element method (GFEM). Initially, it is presented, for the MEF, a brief state of the art and fundamental concepts about this topic. Next, it is described the proposed estimators for the FEM called Z-Estimator and Superconvergent Patch Recovery (SPR). It is proposed, originally, in the ambit of the FEM, the incorporation of the \"Singular Value Decomposition (SVD) to SPR mentioned here as Modified SPR. On the other hand, in the context of GFEM, it is presented a new error estimator entitled EPMEFG in order to expand the ideas of Modified SPR to that method. In EPMEFG, the local polynomial function that allows to recover the nodal values of the gradients of the solution has for support clouds (set of finite elements that share a common node) and results from the applying of a criterion of least squares approximation in relation to the superconvergent points. The number of these points is defined from an analysis of each cloud\'s element, considering the degree of local approximation of the displacement field enriched. Numerical examples elaborated with linear triangular and quadrilateral elements are solved with the Z-Estimator, the Modified SPR and the EPMEFG to evaluate the efficiency of each estimator. This evaluation is done calculating the effectivity indexes.

In this paper, we use the newly developed method of particular solution (MPS) and one-stage method of fundamental solution (MFS-MPS) for solving partial differential equation (PDE). In the 1-D Poisson equation, we prove the solution of MFS-MPS is converge to Spectral Collocation Method using Polynomial, and show that the numerical solution similar to those of using the method of particular solution (MPS), Kansa's method, and Spectral Collocation Method using Polynomial (SCMP). In 2-D, we also test these results for the Poisson equation and find the error behaviors.

Multiquadric collocation method is highly efficient for solving partial differential equations due to its exponential error convergence rate. More amazingly, there are two ways to reduce the error: the traditional way of refining the grid, and the unexpected way of simply increasing the value of shape constant $c$ contained in the multiquadric basis function, $sqrt{r^2 + c^2}$. The latter is accomplished without increasing computational cost. It has been speculated that in a numerical solution without roundoff error, infinite accuracy can be achieved by letting $c ightarrow infty$. The ability to obtain infinitely accurate solution is limited only by the roundoff error induced instability of matrix solution with large condition number. Using the arbitrary precision computation capability of {it Mathematica}, this paper tests the above conjecture. A sharper error estimate than previously obtained is presented in this paper. A formula for a finite, optimal $c$ value that minimizes the solution error for a given grid size is obtained. Using residual errors, constants in error estimate and optimal $c$ formula can be obtained. These results are supported by numerical examples.

Many types of radial basis functions, such as multiquadrics, contain a free parameter called shape factor, which controls the flatness of RBFs. In the 1-D problems, Fornberg et al. [2] proved that with simple conditions on the increasingly flat radial basis function, the solutions converge to the Lagrange interpolating. In this report, we study and extend it to the 1-D Poisson equation RBFs direct solver, and observed that the interpolants converge to the Spectral Collocation Method using Polynomial. In 2-D, however, Fornberg et al. [2] observed that limit of interpolants fails to exist in cases of highly regular grid layouts. We also test this in the PDEs solver and found the error behavior is different from interpolating problem.

Domain decomposition and convex optimization play fundamental roles in current computation and analysis in many areas of science and engineering. These methods have been well developed and studied in the past thirty years, but they still require further study and improving not only in mathematics but in actual engineering computation with exponential increase of computational complexity and scale. The main goal of this thesis is to develop some efficient and powerful algorithms based on domain decomposition method and convex optimization. The topicsstudied in this thesis mainly include two classes of convex optimization problems: optimal control problems governed by time-dependent partial differential equations and general structured convex optimization problems. These problems have acquired a wide range of applications in engineering and also demand a very high computational complexity. The main contributions are as follows: In Chapter 2, the relevance of an adequate inner loop starting point (as opposed to a sufficient inner loop stopping rule) is discussed in the context of a numerical optimization algorithm consisting of nested primal-dual proximal-gradient iterations. To study the optimal control problem, we obtain second order domain decomposition methods by combining Crank-Nicolson scheme with implicit Galerkin method in the sub-domains and explicit flux approximation along inner boundaries in Chapter 3. Parallelism can be easily achieved for these explicit/implicit methods. Time step constraints are proved to be less severe than that of fully explicit Galerkin finite element method. Based on the domain decomposition method in Chapter 3, we propose an iterative algorithm to solve an optimal control problem associated with the corresponding partial differential equation with pointwise constraint for the control variable in Chapter 4. In Chapter 5, overlapping domain decomposition methods are designed for the wave equation on account of prediction-correction" strategy. A family of unit decomposition functions allow reasonable residual distribution or corrections. No iteration is needed in each time step. This dissertation also covers convergence analysis from the point of view of mathematics for each algorithm we present. The main discretization strategy we adopt is finite element method. Moreover, numerical results are provided respectivelyto verify the theory in each chapter. / Doctorat en Sciences / info:eu-repo/semantics/nonPublished

O trabalho aborda a questão das estimativas a posteriori dos erros de discretização e particularmente a recuperação dos gradientes de soluções numéricas obtidas com o método dos elementos finitos (MEF) e com o método dos elementos finitos generalizados (MEFG). Inicialmente, apresenta-se, em relação ao MEF, um resumido estado da arte e conceitos fundamentais sobre este tema. Em seguida, descrevem-se os estimadores propostos para o MEF denominados Estimador Z e \"Superconvergent Patch Recovery\" (SPR). No âmbito do MEF propõe-se de modo original a incorporação do \"Singular Value Decomposition\" (SVD) ao SPR aqui mencionada como SPR Modificado. Já no contexto do MEFG, apresenta-se um novo estimador do erro intitulado EPMEFG, estendendo-se para aquele método as idéias do SPR Modificado. No EPMEFG, a função polinomial local que permite recuperar os valores nodais dos gradientes da solução tem por suporte nuvens (conjunto de elementos finitos que dividem um nó comum) e resulta da aplicação de um critério de aproximação por mínimos quadrados em relação aos pontos de superconvergência. O número destes pontos é definido a partir de uma análise em cada elemento que compõe a nuvem, considerando-se o grau da aproximação local do campo de deslocamentos enriquecidos. Exemplos numéricos elaborados com elementos lineares triangulares e quadrilaterais são resolvidos com o Estimador Z, o SPR Modificado e o EPMEFG para avaliar a eficiência de cada estimador. Essa avaliação é realizada mediante o cálculo dos índices de efetividade. / The paper addresses the issue of a posteriori estimates of discretization errors and particularly the recovery of gradients of numerical solutions obtained with the finite element method (FEM) and the generalized finite element method (GFEM). Initially, it is presented, for the MEF, a brief state of the art and fundamental concepts about this topic. Next, it is described the proposed estimators for the FEM called Z-Estimator and Superconvergent Patch Recovery (SPR). It is proposed, originally, in the ambit of the FEM, the incorporation of the \"Singular Value Decomposition (SVD) to SPR mentioned here as Modified SPR. On the other hand, in the context of GFEM, it is presented a new error estimator entitled EPMEFG in order to expand the ideas of Modified SPR to that method. In EPMEFG, the local polynomial function that allows to recover the nodal values of the gradients of the solution has for support clouds (set of finite elements that share a common node) and results from the applying of a criterion of least squares approximation in relation to the superconvergent points. The number of these points is defined from an analysis of each cloud\'s element, considering the degree of local approximation of the displacement field enriched. Numerical examples elaborated with linear triangular and quadrilateral elements are solved with the Z-Estimator, the Modified SPR and the EPMEFG to evaluate the efficiency of each estimator. This evaluation is done calculating the effectivity indexes.