Mostrar registro simples

dc.contributor.advisorPinto, Marcio Augusto Villela, 1969-pt_BR
dc.contributor.otherMartins, Márcio Andrépt_BR
dc.contributor.otherUniversidade Federal do Paraná. Setor de Tecnologia. Programa de Pós-Graduação em Métodos Numéricos em Engenhariapt_BR
dc.creatorSabatini, Marcelo Henriquept_BR
dc.date.accessioned2024-12-16T20:28:27Z
dc.date.available2024-12-16T20:28:27Z
dc.date.issued2024pt_BR
dc.identifier.urihttps://hdl.handle.net/1884/93767
dc.descriptionOrientador: Prof. Dr. Marcio Augusto Villela Pintopt_BR
dc.descriptionCoorientador: Prof. Dr. Marcio André Martinspt_BR
dc.descriptionTese (doutorado) - Universidade Federal do Paraná, Setor de Tecnologia, Programa de Pós-Graduação em Métodos Numéricos em Engenharia. Defesa : Curitiba, 10/07/2024pt_BR
dc.descriptionInclui referênciaspt_BR
dc.description.abstractResumo: O tema do presente trabalho é a verificação numérica do problema de escoamento bifásico uni e bidimensionais em meio poroso rígido, utilizando a formulação matemática mista pressão-saturação. Neste âmbito, utiliza-se a múltipla extrapolação de Richardson (MER) com o objetivo de reduzir e estimar o erro de discretização. As soluções numéricas foram obtidas mediante o emprego do método dos volumes finitos (MVF) na dimensão espacial, o método de Euler implícito para a dimensão temporal e a linearização pelo método de Picard modificado. Na resolução do sistema linear associado, foi utilizado o método de Gauss-Seidel acoplado juntamente com o método multigrid para acelerar a convergência do processo iterativo. As variáveis de interesse analisadas foram as pressões para as fases úmida e não úmida localizadas no ponto central do domínio Os resultados obtidos apontam que o emprego da MER é promissora para a redução do erro de discretização e aumento da ordem de acurácia das soluções numéricas estudadas. Em relação à estimativa para o erro de discretização, foram analisados alguns estimadores presentes na literatura. Constatou-se que o estimador de Richardson Corrigido apresentou maior nível de acurácia e confiabilidade, representando uma ferramenta robusta para a verificação numérica do problema estudado.pt_BR
dc.description.abstractAbstract: The focus of the present work is numerical verification of the one- and two-dimensional twophase flow problem in rigid porous media using the mixed pressure-saturation mathematical formulation is considered. Within this framework, repeated Richardson extrapolation (RRE) is employed with the aim of reducing and estimating discretization error. The numerical solutions were obtained using the finite volume method (FVM) in the spatial dimension and the implicit Euler method for the temporal dimension. In this resolution, linearization was achieved through the modified Picard method, and for solving the associated linear system, the coupled Gauss-Seidel method with multigrid was utilized to accelerate the convergence of the iterative process. The variables of interest analyzed were the wetting and non-wetting pressures located at the central point of the domain. The results indicate that the use of RRE shows promise in reducing discretization error and increasing the accuracy order of the studied numerical solutions. Regarding the estimation of discretization error, some estimators from the literature were analyzed. It was found that the corrected Richardson estimator exhibited higher levels of accuracy and reliability, representing a robust tool for the numerical verification of the problem under study.pt_BR
dc.format.extent1 recurso online : PDF.pt_BR
dc.format.mimetypeapplication/pdfpt_BR
dc.languageMultilinguapt_BR
dc.languageTexto em português e inglêspt_BR
dc.languageporengpt_BR
dc.subjectTeoria da estimativapt_BR
dc.subjectEscoamento multifasicopt_BR
dc.subjectMateriais porosospt_BR
dc.subjectAnálise numéricapt_BR
dc.titleRedução e estimativa do erro de discretização para problemas de escoamento multifásico em meios porosos rígidospt_BR
dc.typeTese Digitalpt_BR


Arquivos deste item

Thumbnail

Este item aparece na(s) seguinte(s) coleção(s)

Mostrar registro simples