Resumo: O objetivo deste trabalho é a análise numérica e experimental de escoamento viscoso, incompressível, permanente, com transferência de calor, em um canal estreito contendo um arranjo de obstáculos retangulares. A análise experimental envolveu determinação de coeficiente de transferência de calor médio bem como o número de Nusselt médio e medidas de temperatura em esteira térmica para comparação com os resultados obtidos por simulação numérica. Para a análise numérica usamos o programa comercial de mecânica dos fluidos e transferência de calor computacional ICEPAK®. Verificamos que quanto mais adentro o obstáculo estiver no arranjo maior é a transferência de calor por convecção forçada. Determinamos coeficientes de transferência de calor médio e número de Nusselt médio (com incerteza entre 6 e 15%) e verificamos que o efeito da posição diminui à medida que a velocidade aumenta. Concluímos também que ambos os modelos de turbulência utilizados, k-ε padrão e k-ε RNG, foram incapazes de predizer o efeito da posição apropriadamente. Entretanto, o modelo k-ε RNG apresentou melhor comportamento, pois o seu uso resultou em soluções com valores de temperatura intermediários aos experimentais / Abstract: The purpose of this work is the study of the numerical and experimental viscous incompressible steady flow with heat transfer into a narrow channel containing a rectangular array of obstacles. The experimental approach involves determining the coefficient of heat transfer and temperature measurements in thermal wake for comparison with the results obtained in numerical simulations. For the numerical analysis we use the commercial program of fluid mechanics and heat transfer computational ICEPAK™. We confirmed that in the last lines of the array the biggest is the heat transfer by forced convection. We determined the average heat transfer coefficients (with uncertainty between 6 and 15%) and found that the effect of the position decreases as flow speed increases. We use in the simulations the k-ε turbulence model and the k-ε RNG turbulence model. We conclude that both turbulence models used were unable to predict the effect of the position properly. However, the k-ε RNG model showed better behavior. The numerical temperatures with this model were consistent to the experimental temperature / Orientador: João Batista Campos Silva / Coorientador: Amarildo Tabone Paschoalini / Coorientador: Marcio Antonio Bazani / Banca: Ricardo Alan Verdú Ramos / Banca: Marcio Higa / Mestre

An average heat transfer coefficient, h_bar, is often used to solve heat transfer problems. It should be understood that this is an approximation and may provide inaccurate results, especially when the temperature field is of interest. The proper method to solve heat transfer problems is with a conjugate approach. However, there seems to be a lack of clear explanations of conjugate heat transfer in literature. The objective of this work is to provide a clear explanation of conjugate heat transfer and to determine the discrepancy in the temperature field when the interface boundary condition is approximated using h_bar compared to a local, or variable, heat transfer coefficient, h(x). Simple one-dimensional problems are presented and solved analytically using both h(x) and h_bar. Due to the one-dimensional assumption, h(x) appears in the governing equation for which the common methods to solve the differential equations with an average coefficient are no longer valid. Two methods, the integral equation and generalized Bessel methods are presented to handle the variable coefficient. The generalized Bessel method has previously only been used with homogeneous governing equations. This work extends the use of the generalized Bessel method to non-homogeneous problems by developing a relation for the Wronskian of the general solution to the generalized Bessel equation. The solution methods are applied to three problems: an external flow past a flat plate, a conjugate interface between two solids and a conjugate interface between a fluid and a solid. The main parameter that is varied is a combination of the Biot number and a geometric aspect ratio, A_1^2 = Bi*L^2/d_1^2. The Biot number is assumed small since the problems are one-dimensional and thus variation in A_1^2 is mostly due to a change in the aspect ratio. A large A_1^2 represents a long and thin solid whereas a small A_1^2 represents a short and thick solid. It is found that a larger A_1^2 leads to less problem conjugation. This means that use of h_bar has a lesser effect on the temperature field for a long and thin solid. Also, use of ¯ over h(x) tends to generally under predict the solid temperature. In addition is was found that A_2^2, the A^2 value for the second subdomain, tends to have more effect on the shape of the temperature profile of solid 1 and A_1^2 has a greater effect on the magnitude of the difference in temperature profiles between the use of h(x) and h_bar. In general increasing the A^2 values reduced conjugation.

Made available in DSpace on 2014-06-11T19:23:38Z (GMT). No. of bitstreams: 0 Previous issue date: 2010-12-20Bitstream added on 2014-06-13T19:47:44Z : No. of bitstreams: 1 souza_eg_me_ilha.pdf: 959550 bytes, checksum: b4784dbcc883b1be2d0c6b7cce83f54b (MD5) / Fundação de Ensino Pesquisa e Extensão de Ilha Solteira (FEPISA) / O objetivo deste trabalho é a análise numérica e experimental de escoamento viscoso, incompressível, permanente, com transferência de calor, em um canal estreito contendo um arranjo de obstáculos retangulares. A análise experimental envolveu determinação de coeficiente de transferência de calor médio bem como o número de Nusselt médio e medidas de temperatura em esteira térmica para comparação com os resultados obtidos por simulação numérica. Para a análise numérica usamos o programa comercial de mecânica dos fluidos e transferência de calor computacional ICEPAK®. Verificamos que quanto mais adentro o obstáculo estiver no arranjo maior é a transferência de calor por convecção forçada. Determinamos coeficientes de transferência de calor médio e número de Nusselt médio (com incerteza entre 6 e 15%) e verificamos que o efeito da posição diminui à medida que a velocidade aumenta. Concluímos também que ambos os modelos de turbulência utilizados, k-ε padrão e k-ε RNG, foram incapazes de predizer o efeito da posição apropriadamente. Entretanto, o modelo k-ε RNG apresentou melhor comportamento, pois o seu uso resultou em soluções com valores de temperatura intermediários aos experimentais / The purpose of this work is the study of the numerical and experimental viscous incompressible steady flow with heat transfer into a narrow channel containing a rectangular array of obstacles. The experimental approach involves determining the coefficient of heat transfer and temperature measurements in thermal wake for comparison with the results obtained in numerical simulations. For the numerical analysis we use the commercial program of fluid mechanics and heat transfer computational ICEPAK™. We confirmed that in the last lines of the array the biggest is the heat transfer by forced convection. We determined the average heat transfer coefficients (with uncertainty between 6 and 15%) and found that the effect of the position decreases as flow speed increases. We use in the simulations the k-ε turbulence model and the k-ε RNG turbulence model. We conclude that both turbulence models used were unable to predict the effect of the position properly. However, the k-ε RNG model showed better behavior. The numerical temperatures with this model were consistent to the experimental temperature