• Entrar
    Ver item 
    •   Página inicial
    • BIBLIOTECA DIGITAL: Teses & Dissertações
    • 40001016030P0 Programa de Pós-Graduação em Métodos Numéricos em Engenharia
    • Dissertações
    • Ver item
    •   Página inicial
    • BIBLIOTECA DIGITAL: Teses & Dissertações
    • 40001016030P0 Programa de Pós-Graduação em Métodos Numéricos em Engenharia
    • Dissertações
    • Ver item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Estimativa de erros de iteraçao em dinâmica dos fluidos computacional

    Thumbnail
    Visualizar/Abrir
    D - D - MARCIO ANDRE MARTINS.pdf (6.079Mb)
    Data
    2002
    Autor
    Martins, Márcio André
    Metadata
    Mostrar registro completo
    Resumo
    Resumo: A principal motivação para o desenvolvimento deste trabalho consiste no aperfeiçoamento dos critérios adotados para interromper a execução de um processo iterativo. Com este fim, analisa-se o desempenho do estimador empírico. Este estimador fornece uma estimativa do erro de iteração, também denominada de incerteza (U), com base na taxa de convergência da variável de interesse. O erro (E) de iteração pode ser definido pela diferença entre a solução numérica exata e a solução numérica em uma determinada iteração. Através da efetividade do estimador de erro, isto é, da razão entre incerteza e erro, analisa-se a eficiência do estimador empírico quanto à sua acurácia (U/E ~1) e confiabilidade (U/E >1). Para tanto, são resolvidos sistemas de equações gerados a partir da aplicação dos métodos de diferenças finitas e volumes finitos sobre malhas unidimensionais e bidimensionais uniformes, de problemas de transferência de calor e de mecânica dos fluidos. Estes sistemas de equações foram resolvidos com a utilização de vários métodos iterativos. O desempenho do estimador empírico pode ser dividido em três intervalos: nas iterações iniciais, em geral a acurácia é baixa; quando o número de iterações é muito elevado, os erros de arredondamento predominam sobre os erros de iteração mas, mesmo assim, a acurácia é relativamente boa; no intervalo entre esses dois extremos, a acurácia tende a melhorar à medida que se aumenta o número de iterações. Não se recomenda o uso do estimador empírico em processos iterativos que envolvem a aplicação da técnica multigrid pois, em geral, sua acurácia é baixa. Constatou-se que existem relações diretas entre o estimador empírico e alguns estimadores disponíveis na literatura
     
    Abstract: The main motivation to develop this work consists in improve the rales or approaches adopted to interrapt the execution of an iterative process. It's proposed to calculate the uncertainty (U) of numerical Solutions, where the uncertainty is defined as an estimated iterative error (E). To a variable of any interest, this error is the difference between the exact numerical solution and the calculated numerical solution. It is analyzed the efficiency of an empirical error estimator to iterative processes. This estimator calculates the numerical uncertainty (£/) on basis on the convergence rate of the interest variable. Through the effectiveness of the error estimator, that is, of the reason between uncertainty and error, the error estimator efficiency is verified on its accuracy (U/E ) and reliability (U/E ). Therefore, are resolved linear systems generated from the application of finite differences and finite volume methods, on uniform unidimensional and bidimensional meshes, in heat transfer problems and and fluid mechanics. These systems of equations were resolved with the use of iterative methods. The performance of the empirical estimator can be divided in ihree intervals: In the "initial" iterations the accuracy is low, in general; when the number of iterations is very high, the round-off affect the accuracy, that is good; in the interval among those two limits, the accuracy tends to be big as it increases the iteration number. It was verified that, in iterative processes that involve the application of the multigrid method, the empirical estimator behavior doesn't seems to be good; because in the accomplished simulations it was obtained results with low accuracy and not very reliable. It was also verified that there are direct relations between the studied error estimator and some existent estimators in the literature
     
    URI
    https://hdl.handle.net/1884/100268
    Collections
    • Dissertações [161]

    DSpace software copyright © 2002-2022  LYRASIS
    Entre em contato | Deixe sua opinião
    Theme by 
    Atmire NV
     

     

    Navegar

    Todo o repositórioComunidades e ColeçõesPor data do documentoAutoresTítulosAssuntosTipoEsta coleçãoPor data do documentoAutoresTítulosAssuntosTipo

    Minha conta

    EntrarCadastro

    Estatística

    Ver as estatísticas de uso

    DSpace software copyright © 2002-2022  LYRASIS
    Entre em contato | Deixe sua opinião
    Theme by 
    Atmire NV