A finite difference based solution method is derived for the velocity-pressure formulation of the two-dimensional incompressible Navier-Stokes equations. The method is proven stable using the energy method, facilitated by SBP operators, for characteristic and Dirichlet boundary condition implemented using the SAT technique. The numerical experiments show the utility of high-order finite difference methods as well as emphasize the problem of pressure boundary conditions. Furthermore, we demonstrate that a discretely divergence free solution can be obtained by use of the projection method.

A microfluídica juntamente com a intensificação de processos são duas áreas de pesquisa interessadas no estudo e desenvolvimento de processos em escala micrométrica capazes de manipular diminutas quantidades de reagentes. Para tanto, estes devem contar com dispositivos de pequena escala de tamanho e ao mesmo tempo serem tão confiáveis e eficientes quanto os de escala industrial. Uma das operações unitárias envolvidas nesses processos é a agitação. Em função da ordem de grandeza dos equipamentos e dos materiais em que são fabricados, grandes diferenciais de pressão não podem ser aplicados nos mesmos e como conseqüência no interior dos micromisturadores, como são conhecidos tais equipamentos, o escoamento se dá em regime laminar, sob está condição o processo de mistura é controlado pela difusão entre os componentes. Uma maneira de superar esta dificuldade é gerar no interior do micromisturador o aparecimento de um escoamento caótico. Para tal, podem-se utilizar fontes de energia externa (micromisturadores ativos) ou a própria energia do escoamento (micromisturadores passivos) através da construção de geometrias especiais. O desenvolvimento em laboratório destes equipamentos demanda tempo e geralmente é oneroso. A principal alternativa para este trabalho é a dinâmica dos fluidos computacional (CFD), ferramenta aplicada no presente estudo para analisar três geometrias diferentes propostas e analisadas experimentalmente no trabalho de Cunha (2007). Para caracterizar o funcionamento dos mesmos foram testadas quatro vazões distintas, com as quais foi possível levantar os perfis de pressão, velocidade e fração mássica de dois componentes que eram misturados. Com o intuito de demonstrar a eficiência dos equipamentos dois parâmetros foram analisados: o avanço da qualidade da mistura e a perda de carga para as diferentes condições operacionais. Apesar da limitação da malha e de não ter-se obtido resultados independentes da malha, foi possível se fazer uma comparação entre as três geometrias e identificouse que os micromisturadores M2 e M3 são os que apresentam o melhor desempenho para a faixa de vazão simulada (120 < Re < 1200). / Microfluidics and process intensification are two research areas interested in the study and development of new micrometric-scale devices capable of manipulating and processing small quantities of reagents. These processes have to deal with small scale equipment and at the same time be as reliable and efficient as the large-scale one. Because of the scale of this equipment and the material it is made of, large pressure differential is not possible, as a consequence in the interior of the micromixers, as they are known; a laminar flow develops, under those circumstances the mixing process is controlled by the diffusion mechanism between the two components. One way to suppress this deficiency is to generate a chaotic flow on the micromixer, which can be done by using external energy (active micromixer) or its own flow energy (passive micromixer) through special geometry construction. The experimental development of such microdevices demands time and, generally, is very expensive. The main alternative for this activity is the use of computational fluid dynamics; this tool was employed on this work with the aim of studying three geometries proposed by Cunha (2007). To characterize their working process, four different volumetric flows were simulated and analyzed the pressure, velocity and mass fraction profiles. Two parameters were calculated in order to characterize their efficiency: the mixture quality along the micromixers cross sections and the pressure drop for different operational conditions. Although we have mesh size limitations and a mesh independent results were not obtained it was possible to compare the three micromixers geometries and it was found out that both M2 and M3 micromixers had the best performance under operational conditions tested (120 < Re < 1200).

Neste trabalho propõe-se uma análise da dispersão de material particulado gerado por fontes estacionárias, aplicando ferramentas de Fluidodinâmica Computacional (em inglês Computatonal Fluid Dynamics, CFD), usando um modelo Euleriano para o escoamento e Lagrangeano para as partículas, em estado estacionário no software Fluent. A verificação da modelagem é apresentada em duas sessões, a primeira compreende o estudo do escoamento atmosférico em condições de estabilidade neutra incluindo os efeitos das heterogeneidades do terreno, no caso, um monte de 126m de altura em escala real. Foram comparados diferentes modelos de turbulência: padrão, RNG e parametrizado para o caso atmosférico. Os três modelos apresentaram desempenho semelhante e descrevem satisfatoriamente as tendências dos dados experimentais. A segunda, o modelo lagrangeano baseado no tempo de vida dos turbilhões (em inglês Discrete Random Walk, DRW) foi utilizado para representar a distribuição de concentrações de material particulado em um ambiente confinado. Os resultados numéricos descrevem satisfatoriamente os perfis de concentrações das partículas, porém subestimam os valores na região próxima à parede, o que indicaria que uma melhor aproximação da deposição das partículas deve ser considerada. Após estes estudos, esta abordagem euleriana-lagrangeana foi aplicada ao caso da dispersão de material particulado em condições de atmosfera neutra em uma região do Polo Cerâmico de Santa Gertrudes-SP, local responsável pela maior produção de revestimentos cerâmicos do Brasil. O trânsito de caminhões nas estradas não pavimentadas foi identificado como fonte de material particulado que atinge o perímetro urbano da cidade. Estudaram-se também as contribuições das atividades referentes à exploração das minas de argila e a preparação do solo para uso agrícola, apesar desta última não soma uma carga relevante ao ser comparada com as duas anteriores. Além da influência das fontes mencionadas, fontes de origem industriais localizadas dentro da cidade também foram consideradas no cenário. / In this study an analysis of the dispersion of particulate matter generated by stationary sources is proposed, by applying Computational Fluid Dynamics (CFD) tools, using an Eulerian model for the flow and a Lagrangean model for the particles, under steady-state conditions in the Fluent software. The model validation is presented in two sessions, the first comprises the study of the atmospheric flow under conditions of neutral stability, including the effects of the heterogeneities of the terrain. In this case, the air flow past a 126m high hill in real scale was adopted. Different turbulence models were compared: standard k-?, k-? RNG and parametrized k-?. The three models presented similar performance and described satisfactorily the trends of the experimental data. Subsequently, the Lagrangean model based on the Eddy life time (DRW - Discrete Random Walk) was used to represent the distribution of particulate matter concentrations in a confined environment. The numerical results satisfactorily describe the particle concentration profiles, but underestimate the values in the region near the wall, which would indicate that a better approximation of the particle deposition should be considered. After these studies, this Eulerian- Lagrangean approach was applied to the case of dispersion of particulate material under neutral atmosphere conditions in a region around the city of Santa Gertrudes-SP, where the Ceramic Complex is responsible for the largest production of ceramic tiles in Brazil. Transit of trucks on unpaved roads has been identified as a source of particulate matter that reaches the urban perimeter of the city. The contributions of the activities related to mining and the preparation of the soil for agricultural use were also studied, although this last one does not add a significant load when compared with the two previous ones. In addition to the influence of the mentioned sources, industrial sources located inside the city also contribute to the scenario.

Nos últimos anos, as legislações ambientais têm se tornado mais rigorosas em relação aos limites de emissão de dióxido de enxofre, provenientes das refinarias de petróleo e das industrias químicas. A recuperação de enxofre, através do processo Claus, é o padrão da industria na redução das emissões de SO2 de efluentes gasosos ricos em H2S. Basicamente o processo consiste de duas etapas em série: uma térmica e outra catalítica. A cinética do processo é limitada, devido à natureza de suas reações principais que tendem ao equilíbrio, por isto, uma unidade de recuperação de enxofre com três reatores catalíticos, por exemplo, tem uma capacidade teórica de recuperar 98% de enxofre do gás ácido. A fase térmica do Processo Claus, é responsável por 60 a 70% da conversão total de enxofre, mas apesar disto, tem sido pouco estudada e ainda hoje a maior parte dos modelos disponíveis para projeto de novas plantas ou otimização das existentes, são baseados em relações de equilíbrio ou em equações empíricas. Esses modelos têm se mostrados insuficientes para atender às novas exigências de eficiência das unidades de recuperação de enxofre, uma vez que não consideram fatores fluidodinâmicos e as limitações cinéticas das reações envolvidas nos processo de combustão. O objetivo deste trabalho é o estudo da fase térmica do Processo Claus, utilizando como técnica de modelagem a fluidodinâmica computacional (CFD), através do programa CFX-5.6. Também são incorporados os mais recentes avanços na modelagem cinética das reações: Claus, oxidação e pirólise do H2S, formação de COS e CS2. Para isto, são modelados 4 tipos diferentes de queimadores, operando em câmaras de combustão de mesmo tamanho e nas mesmas condições de processo. / Over the last years the environmental legislations have become more severe with respect to the limits of SO2 emission that are produced by the petroleum refining and chemical industries. The sulphur recovery, through the Claus process, is the industrial standard reducing the SO2 emission in the effluents gases that are rich in H2S . Basically, the process is composed by two sequential phases: one is thermal, and the other one is catalytic. The process kinetics are limited, due to the intrinsic nature of the main reactions to tend to the equilibrium, therefore, a sulphur recovery unit with three catalytic reactors, for instance, has the theoretical capacity to recover 98% of the sulphur from the acid gas. The thermal phase of the Claus process is responsible for 60 to 70% of the total sulphur conversion, however, it has not been fully investigated, and up to now, most of the existent models for designing new plants or optimisation of the current ones, are based in equilibrium reactions or empirical equations. These models have been insufficient to meet the new efficiency requirements of the sulphur recovery plants, as far as fluid dynamic factors and kinetic limitations of the reactions involved are not considered in the combustion process. The objective of this work is to study the thermal phase of the Claus process, using computational fluid dynamics (CFD) as a modelling technique through CFX-5.6 software. In addition, the most recent advances in kinetics modelling of the reactions are also incorporated: Claus reaction, oxidation and pyrolysis of the H2S, formation of COS and CS2. Four different types of burners are simulated, with a combustion chambers of the same size and even processing conditions.

Adverse health effects attributable to both short- and long-term exposure to air pollution have turned the focus on different microenvironments. The interior of vehicles is of relevance as road traffic emissions and re-suspension of road dust are major sources of pollutants associated with adverse health effects. Hence, the air quality inside vehicles deserves attention regarding human health. This thesis presents a new virtual methodology, using CFD, to study the distribution of fine particulate matter, PM2.5, inside a car cabin. In the CFD model, unsteady RANS and Lagrangian particle tracking were used to simulate particles entering from the exterior. In this study, a practical measurement of interior particle concentrations was also carried out as a first attempt to validate the CFD model. The objective was to find positions inside the cabin where elevated concentrations of PM2.5 are present. The results from the CFD simulations showed that significantly higher concentrations are present at head height in the front row. Due to a discrepancy in the investigated positions in the CFD model and the practical measurement, the simulation results could not be validated. Nevertheless, the simulation results in this study have provided guidelines for future measurements of interior particle concentrations.

A microfluídica juntamente com a intensificação de processos são duas áreas de pesquisa interessadas no estudo e desenvolvimento de processos em escala micrométrica capazes de manipular diminutas quantidades de reagentes. Para tanto, estes devem contar com dispositivos de pequena escala de tamanho e ao mesmo tempo serem tão confiáveis e eficientes quanto os de escala industrial. Uma das operações unitárias envolvidas nesses processos é a agitação. Em função da ordem de grandeza dos equipamentos e dos materiais em que são fabricados, grandes diferenciais de pressão não podem ser aplicados nos mesmos e como conseqüência no interior dos micromisturadores, como são conhecidos tais equipamentos, o escoamento se dá em regime laminar, sob está condição o processo de mistura é controlado pela difusão entre os componentes. Uma maneira de superar esta dificuldade é gerar no interior do micromisturador o aparecimento de um escoamento caótico. Para tal, podem-se utilizar fontes de energia externa (micromisturadores ativos) ou a própria energia do escoamento (micromisturadores passivos) através da construção de geometrias especiais. O desenvolvimento em laboratório destes equipamentos demanda tempo e geralmente é oneroso. A principal alternativa para este trabalho é a dinâmica dos fluidos computacional (CFD), ferramenta aplicada no presente estudo para analisar três geometrias diferentes propostas e analisadas experimentalmente no trabalho de Cunha (2007). Para caracterizar o funcionamento dos mesmos foram testadas quatro vazões distintas, com as quais foi possível levantar os perfis de pressão, velocidade e fração mássica de dois componentes que eram misturados. Com o intuito de demonstrar a eficiência dos equipamentos dois parâmetros foram analisados: o avanço da qualidade da mistura e a perda de carga para as diferentes condições operacionais. Apesar da limitação da malha e de não ter-se obtido resultados independentes da malha, foi possível se fazer uma comparação entre as três geometrias e identificouse que os micromisturadores M2 e M3 são os que apresentam o melhor desempenho para a faixa de vazão simulada (120 < Re < 1200). / Microfluidics and process intensification are two research areas interested in the study and development of new micrometric-scale devices capable of manipulating and processing small quantities of reagents. These processes have to deal with small scale equipment and at the same time be as reliable and efficient as the large-scale one. Because of the scale of this equipment and the material it is made of, large pressure differential is not possible, as a consequence in the interior of the micromixers, as they are known; a laminar flow develops, under those circumstances the mixing process is controlled by the diffusion mechanism between the two components. One way to suppress this deficiency is to generate a chaotic flow on the micromixer, which can be done by using external energy (active micromixer) or its own flow energy (passive micromixer) through special geometry construction. The experimental development of such microdevices demands time and, generally, is very expensive. The main alternative for this activity is the use of computational fluid dynamics; this tool was employed on this work with the aim of studying three geometries proposed by Cunha (2007). To characterize their working process, four different volumetric flows were simulated and analyzed the pressure, velocity and mass fraction profiles. Two parameters were calculated in order to characterize their efficiency: the mixture quality along the micromixers cross sections and the pressure drop for different operational conditions. Although we have mesh size limitations and a mesh independent results were not obtained it was possible to compare the three micromixers geometries and it was found out that both M2 and M3 micromixers had the best performance under operational conditions tested (120 < Re < 1200).

Fixed bed reactors are an essential part of the chemical industry as they are used in a wide variety of chemical processes. To better model these systems a more fundamental understanding of the processes taking place in a fixed bed is required. Fixed bed models are traditionally based on high tube-to particle diameter ratio (N) beds, where temperature and flow profile gradients are mild and can be averaged. Low-N beds are used in extremely exo- and endothermic processes on the tube side of tube and shell type reactors. In these beds, heat transfer is one of the most important aspects. The importance of accurate modeling of heat transfer and its dependence on accurate modeling of the flow features leads to the need for studying the phenomena in these low-N beds in detail. In this work a comparative study is made of the influence of spherical and cylindrical packing particle shapes, positions and orientations on the rates of heat transfer in the near-wall region in a steam reforming application. Computational Fluid Dynamics (CFD) is used as a tool for obtaining the detailed flow and temperature information in a low-N fixed bed. CFD simulation geometries of discrete particle packed beds are designed and methods for data extraction and analysis are developed. After conceptual and quantitative analysis of the simulation data it is found that few clear relations between the complex phenomena of flow and heat transfer can be easily identified. Investigated features are the orientations of the particle in the flow, and many design parameters, such as the number and size of longitudinal holes in the particle and external features on the particle. We find that many of the investigated features are related and their individual influences could not be isolated in this study. Some of the related features are, for example, the number of holes in the particle design and the particle orientation in the flow. Some general conclusions could be drawn. External features on the particles enhance the overall heat transfer properties by better mixing of the flow field. When holes are present in the cylindrical particle design, heat transfer effectiveness can be improved with fewer larger holes. After identifying the packing-related features influencing the near-wall heat transfer under steam reforming conditions, an attempt was made to incorporate the steam reforming reaction in the simulation. In the initial attempts the reaction was modeled as an energy flux at the catalyst particle surfaces. This approach was based on the abilities of the CFD code, but turned out not accurate enough. Elimination of the effects of local reactant depletion and the lack of solid energy conduction in the catalyst particles resulted in an unphysical temperature field. Several suggestions, based on the results of this study, are made for additional aspects of particle design to be investigated. Additionally, suggestions are made on how to incorporate the modeling of a reaction in fixed bed heat transfer simulations.

Modeling of fluid flow, heat transfer and reaction in fixed beds is an essential part of their design. This is especially critical for highly endothermic reactions in low tube-to-particle diameter ratio (N) tubes, such as methane steam reforming (MSR) and alkane dehydrogenation as two important commercial reactions. The modeling of fixed bed reaction is available in literatures with lots of assumptions. However, there is a need for implementing the reaction conditions with diffusion aspects on a real fixed bed reactor without assuming any pseudo conditions. Computational fluid dynamics (CFD) has been found as a suitable tool by many researchers to simulate fixed beds. CFD can simulate complex geometry of randomly-packed tubes, and provides us with more fundamental understanding of the transport and reaction phenomena in reactor tubes. CFD can be used to obtain detailed three-dimensional velocity, species and temperature fields that are needed to improve engineering approaches. Previously, the geometry of 120-degree wall segment (WS) of the whole reactor tube has been studied in our group. Previous works have introduced the coupling of gas flow and resolved species and temperature gradients inside pellets by CFD for methane steam reforming (MSR) and propane dehydrogenation (PDH) without considering deactivation. The deactivation of catalysts due to carbon formation is an important problem in industry, such as steam reforming and the catalytic dehydrogenation of alkanes, which are both strongly endothermic reactions. Many researches were carried out to study the effect of carbon formation and catalyst deactivation on the reactor performance. The local carbon deposition on catalysts can cause particle breakage and strongly decrease reaction rates. Catalyst deactivation in heated tubes removes the heat sink and can result in local hot spots that weaken the reactor tube. This is particularly a problem for a low tube-to-particle diameter ratio fixed bed reactor. A 3D resolved CFD model simulation was used to study the local details of carbon deposition in which the reactions and deactivation take place inside the catalytic solid particles. CFD simulations of flow, heat transfer, diffusion and reaction were carried out using the commercial CFD code FLUENT/ANSYS 6.3 in a 3D 120-degree periodic wall segment with N=4. The mesh used boundary layer prism cells at both the inside and outside particle surfaces and at the tube wall. These reactions were represented in the solid particles using user-defined scalars to mimic species transport and reaction, with user-defined functions supplying reaction rates. Diffusion in the particles was modeled by Fick's law using an effective diffusivity, given by Hite and Jackson's approximation of the Dusty Gas Model. The transient developments of particle internal gradients and carbon accumulation have been studied for the early stages of deactivation. Carbon concentration is initially strongest close to the surface and in the high temperature regions of the catalysts and affected by the wall heat flux. Deactivation of the endothermic reactions causes a slow increase in the average catalyst temperature. The second stage of the research was the verification of our CFD reaction model with experimental data under reacting conditions. The highly endothermic commercial methane steam reforming (MSR) reaction was studied experimentally in a fixed bed reactor. The temperature contributions inside catalyst particles were measured. The MSR reaction showed strong effects on the temperature profile along the reactor. Then, a CFD model was used to predict temperature profiles under MSR reaction conditions. Comparison of CFD and experimental data showed very good qualitative as well as quantitative agreement for temperature inside catalyst particles at different inlet gas temperatures. The last stage was to develop a fundamental energy equation without introducing new adjustable parameters to study heat transfer in fixed beds. In the past, many researchers have been carried out to simulate the heat transfer in fixed bed reactors by using kr (effective thermal conductivity) and hw (heat transfer coefficient). But the classical model with kr and hw cannot give a correct T(r) near tube wall, where deactivation is strongest. Therefore we need a better model which can represent the near wall heat transfer more accurate. CFD modeling was used to develop pseudo-continuum model for T(r) using Vr(r,z) and Vz(r). To get better temperature at the wall vicinity close to the physical reality. In this model radial thermal conductivity was obtained from Zehner-Schlünder model. The convection heat transfer was calculated in the 2D flow fluid from the CFD run. Results were obtained for Reynolds numbers in the range 240€“1900. The accuracy of the new model has been validated by analytical solution. The temperature calculated by the new velocity field pseudohomogenous energy equation showed reasonable quantitative agreement with values predicted by the CFD model.

Due to continued outward expansion of industry and community development into the wildland-urban interface (WUI), the threat to life safety and property from wildland fires has become a significant problem. Such fire scenarios can be better understood through the use of computation fluid dynamics based fire-spread models. However, current physical fire models must be specifically adapted to handle the phenomena associated with WUI fires. Only then can they be reliably used as research and decision making tools to help mitigate the problem. In this research, the current standard in wildland fire modeling for representing the effect on wind flow from a porous vegetative medium is examined. The technique used employs basic correlations for object drag, and its validity with respect to real vegetation has yet to be examined in detail by the scientific community. The modeling of vegetation is studied within the framework of the existing Wildland-Urban Interface Fire Dynamics Simulator (WFDS), and the potential need for continued development is assessed. Comparisons are made to both experimental and numerical studies. Additionally, the validity of the model is considered at both the scale of an individual tree, as well as that of a whole forest canopy. Results show that as a first approximation the model is able to perform well in the latter case. At the scale of an individual tree, however, the behavior is governed by theoretical constants. The assumption of cylindrical vegetation elements performs slightly better than the commonly used spherical case, but neither adequately captures experimental tendencies. Accurate flow representation for single trees is crucial to modeling the key driving factors of fire behavior (such as combustion and heat transfer) in small scale WUI scenarios. Ultimately, this study illustrates the need for well-designed experiments, specifically to generate empirical constants which will improve the behavior of the simplified theory.

Submitted by JOSIANE SANTOS DE OLIVEIRA (josianeso) on 2018-11-07T15:29:53Z No. of bitstreams: 1 Peterson Machado de Oliveira_.pdf: 3557303 bytes, checksum: 0639bd079ad8864f61fccb73df335d8e (MD5) / Made available in DSpace on 2018-11-07T15:29:53Z (GMT). No. of bitstreams: 1 Peterson Machado de Oliveira_.pdf: 3557303 bytes, checksum: 0639bd079ad8864f61fccb73df335d8e (MD5) Previous issue date: 2018-05-11 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Durante a etapa de projeto de fornos de unidades de destilação atmosférica são definidos limites para as temperaturas dos tubos da serpentina de aquecimento de petróleo, a fim de garantir sua integridade estrutural. Como estes equipamentos operam por anos, acabam por passar por modificações no projeto original e, portanto, faz-se necessário também que hajam ferramentas capazes de se adaptar e identificar os efeitos destas mudanças. No intuito de avaliar as condições de operação de um forno de uma destas unidades de destilação, neste trabalho realiza-se uma avaliação numérica em CFD - Computational Fluid Dynamics através do software Ansys CFX, onde a modelagem do escoamento reativo, turbulento e não isotérmico é implementada para simular o processo de combustão e a transferência de calor no interior do forno. A modelagem teve como base a solução das equações de conservação de massa, de quantidade de movimento, de energia e de espécies químicas, considerando a radiação térmica em meios participantes, além da modelagem por taxas finitas de reações químicas com o modelo Eddy Breakup - Arrhenius para representar a queima do combustível em dois passos. A turbulência foi representada no escoamento através do modelo k- e o espectro de absorção dos gases de combustão na radiação tpermica foi descrito pelo modelo WSGG - Weighted Sum of Gray Gases. A solução das equações de conservação em acoplamento aos demais modelos selecionados é desenvolvida pelo Método de Volumes Finitos. Para avaliar o fluxo médio de calor para o petróleo foi utilizado o software comercial HTRI Xfh 6.0 e seus resultados foram aplicados como condições de contorno no modelo numérico no CFX. A validação da modelagem é realizada por comparação com dados experimentais e numéricos disponíveis na literatura e por comparação com dados experimentais operacionais do forno. Como principais resultados apresentam-se os campos de temperaturas no interior do forno e a influência da geometria destes equipamentos no fluxo dos gases de combustão. Além disto, foi possível verificar o perfil de temperatura ao longo de cada tubo da serpentina de aquecimento individualmente bem como a distribuição de concentração de espécies químicas e taxas de transferência de calor na câmara de combustão. A modelagem numérica apresentou resultados mais próximos aos dados operacionais quando comparada a simulação com o software HTRI Xfh, mas em ambas as simulações as temperaturas obtidas estavam dentro da mesma ordem de grandeza, apresentando diferença máxima de 12,5% entre si.