The main purpose of the dissertation is to contribute to the development of numerical techniques for computational heat transfer and fluid flow, suitable for low cost (loosely coupled) parallel computers. It is focused on implicit integration schemes, using finite control volumes with multigrid (MG) algorithms.Natural convection in closed cavities is used as a problem model to introduce different aspects related with the integration of the incompressible Navier-Stokes equations, such as the solution of the pressure correction (or similar) equations that is the bottleneck of the algorithms for parallel computers. The main goal of the dissertation has been to develop new algorithms to advance in the solution of this problem rather than to implement a complete parallel CFD code. An overview of different sequential multigrid algorithms is presented, pointing out the difference between geometric and algebraic multigrid. A detailed description of segregated ACM is given. The direct simulation of a turbulent natural convection flow is presented as an application example. A short description of the coupled ACM variant is given.Background information of parallel computing technology is provided and the the key aspects for its efficient use in CFD are discussed. The limitations of low cost, loosely coupled cost parallel computers (high latency and low bandwidth) are introduced. An overview of different control-volume based PCFD and linear equation solvers is done. As an example, a code to solve reactive flows using Schwartz Alternating Method that runs particularly well on Beowulf clusters is given.Different alternatives for latency-tolerant parallel multigrid are examined, mainly the DDV cycle proposed by Brandt and Diskin in a theoretical paper. One of its main features is that, supressing pre-smoothing, it allows to reduce the each-to-neighbours communications to one per MG iteration. In the dissertation, the cycle is extended to two-dimensional domain decompositions. The effect of each of its features is separately analyzed, concluding that the use of a direct solver for the coarsest level and the overlapping areas are important aspects. The conclusion is not so clear respect to the suppression of the pre-smoothing iterations.A very efficient direct method to solve the coarser MG level is needed for efficient parallel MG. In this work, variant of the Schur complement algorithm, specific for relatively small, constant matrices has been developed. It is based on the implicit solution of the interfaces of the processors subdomains. In the implementation proposed in this work, a parallel evaluation and storage of the inverse of the interface matrix is used. The inner nodes of each domain are also solved with a direct algorithm. The resulting algorithm, after a pre-processing stage, allows a very efficient solution of pressure correction equations of incompressible flows in loosely coupled parallel computers.Finally, all the elements presented in the work are combined in the DDACM algorithm, an algebraic MG equivalent to the DDV cycle, that is as a combination of a parallel ACM algorithm with BILU smoothing and a specific version of the Schur complement direct solver. It can be treated as a black-box linear solver and tailored to different parallel architectures.The parallel algorithms analysed (different variants of V cycle and DDV) and developed in the work (a specific version of the Schur complement algorithm and the DDACM multigrid algorithm) are benchmarked using a cluster of 16 PCs with a switched 100 Mbits/s network.The general conclusion is that the algorithms developed are suitable options to solve the pressure correction equation, that is the main bottleneck for the solution of implicit flows on loosely coupled parallel computers.

Deep knowledge of combustion phenomenon is of great scientific and technological interest due to its presence in a wide range of industrial processes and equipment. Being the most important worldwide energy support provided by combustion of fossil fuels, the goal of developing more efficient and cleaner systems or equipment is clearly justified. In the last decade, the importance of the reduction of pollutant emissions has increased considerably due to both environmental consciousness and to governmental policies, being one of the most important aspects to assure the competitiveness of combustion-related industries.Traditionally, a high number of experimental studies based on trial-and-error analysis were necessary on the optimisation of thermal equipment, where heat and mass transfer and fluid flow have a dominant role. In the last decades, in agreement with the development of computational capabilities, CFD simulations have become a worthful complement to experimental investigations, reducing in this sense production costs and time to market. However, the considerable complexity of combustion phenomena and the strong feedback between the flow and the chemistry, makes more difficult the task of develop accurate, computationally capable and robust numerical codes for combustion phenomena in industrial applications. This goal remains a promising challenge today and for the foreseeable future.The work developed in this thesis contributes to this objective. Rather than assuming less accurate mathematical approaches and consider their application to engineering problems, our main intention has been centred to the development of numerical tools that enable the feasible resolution of combustion problems with the highest level of detail.On the detailed numerical simulation of combustion problems, main difficulties arise from the stiffness of the governing equations, the presence of flame fronts, and the huge number of species and reactions involved in the reaction mechanisms. In order to overcome these numerical difficulties, a parallel multiblock algorithm able to work efficiently with loosely coupled computers, has been developed. The employment of numerical strategies to deal with the commented stiffness, and the use of multiblock techniques to optimise the discretisation and to parallelise the code, are the main attributes that can be pointed out. An excellent ratio between computational time and resources have been obtained.In the analysis of the numerical solutions, special attention is given to their verification. The accuracy of the results has been analysed providing uncertainty estimations. The numerical methodology employed in this thesis to simulate reactive flows is one of the most relevant contributions presented.The numerical infrastructure developed has been applied to the numerical analysis of laminar flames. Although combustion nearly always takes place within a turbulence flow field to increase the mixing process and thereby enhance combustion, laminar flames are considered as an illustrative example of combustion phenomenon and its experimental and detailed numerical analysis is a basic ingredient on the modelling of turbulent combustion processes as well as for pollutant formation. Special attention has been given to co-flow non-premixed and partially premixed methane-air laminar flames. The wide application of these flames in house-hold and industrial heating systems due to both their intense combustion process and the relatively clean nature of natural gas (composed mainly by methane), has motivated extensive research on the experimental and numerical modelling of such flames.Detailed numerical simulations have been performed to analyse fundamental aspects of these flames, and the adequacy of several mathematical approaches employed on their modelling. Available experimental data have been taken into account both on the analysis of the influence of partially premixing to main flame properties and on the mathematical approaches comparison. Special attention has also been given to pollutant formation, NOx and CO emission indexes.

Work here presented is the result of basic research in key aspects of the currently available engineering tools and methodologies for the design, optimisation and development of thermal systems and equipment: turbulence modelling, high performance computing and quality tests and procedures so as to assess credibility to the numerical solutions (verification and validation). The thesis comprises six main chapters written in a paper format. Two of them have already been published in international journals, one in the proceedings of a Spanish conference and two in proceedings of international conferences on Computational Fluid Dynamics and heat transfer. The last chapter has recently been submitted for publication to an international journal. Therefore, all the chapters are written so as to be self-contained, complete and concise. As a consequence, some contents of the chapters such those describing the governing equations, or the verification procedure used to assess the credibility of the numerical solutions, are repeated in several of them. Furthermore, as only minor changes have been introduced in the chapters respect to the original papers, each of them reflects the know-how of the CTTC (Heat and Mass Transfer Technological Centre were the research has been carried out) when they were published.Papers presented in chapters 1 and 2 deal with turbulence modelling. A general overview is given on the formulation and numerical techniques of the different levels of turbulence modelling: Direct Numerical Simulation (DNS), Large Eddy Simulation (LES) and Reynolds Averaged Navier-Stokes Simulation (RANS). Main attention is focussed on the eddy viscosity two-equation RANS models. Their formulation is presented in more detail, and numerical solutions of the most extended. Benchmark problems on turbulence modelling are given compared to the available experimental data.Chapters 3 and 4 focus on the use of the multiblock method (domain decomposition method), as a numerical technique that combined with the parallel computing may allow reducing the demanding computational time and memory (high performance computing). The multiblock approach used is based on the conservation of all the physical quantities (fully conservative method) and on an explicit information exchange between the different blocks of the domain. The goal of the work presented in these two chapters is to verify that such a multiblock approach does not introduce additional uncertainty in the numerical solutions.Chapter 5 presents a tool that has been developed at the CTTC for the verification of finite volume computations. In fact, this tool is also partially used and described in the results presented in the previous chapters. Here, it is described and discussed in detail and it is applied to a set of different CFD and heat transfer problems in two and three dimensions, with free and forced convection, with reactive and non-reactive flows and with laminar and turbulent flows.The last chapter shows a complete study for the development of a credible heat transfer relation for the heat evacuated from a ventilation channel. Such study comprises all the different steps that have to be accomplished so as to develop credible and applicable results in mechanical engineering. It comprises a description of the mathematical model to represent the physical phenomena in the channel, the numerical model to solve the set of coupled differential equations of the mathematical model, the construction and testing of an ad-hoc experimental set-up, and a verification and validation (V&V) test that guarantees that the numerical solution is an accurate enough approximation of the mathematical model (verification), and that it properly predicts the reality (validation).

La presente Tesis centra la atención en la simulación numérica y validación experimental del comportamiento térmico y fluido-dinámico de flujos bifásicos (líquido-vapor), así como su aplicación en sistemas y equipos térmicos, propios del campo de la refrigeración y del aire acondicionado. Consecuencia del estudio realizado ha sido la elaboración de códigos informáticos para la simulación numérica de condensadores, evaporadores y dispositivos de expansión del tipo capilar, en base a la formulación de las ecuaciones gobernantes sobre volúmenes de control finitos. La Tesis también aborda la simulación de sistemas frigoríficos completos en base a algoritmos que acoplan la resolución de los elementos que lo integran.El flujo bifásico se ha modelizado en base a la resolución de las ecuaciones gobernantes sobre volúmenes de control finitos, escritas en forma unidimensional y estado permanente o transitorio. Las ecuaciones gobernantes se resuelven mediante esquemas numéricos del tipo implícito que avanzan tramo a tramo en la dirección del flujo o bien esquemas numéricos basados en corrección de presiones. La formulación implementada requiere de información empírica para evaluar la interacción entre el fluido y los contornos sólidos limitantes, así como la estructura del flujo bifásico. Las formulaciones matemáticas implementadas han sido suficientemente contrastadas experimentalmente en base a resultados presentados en la literatura científica. El modelo implementado para la simulación del flujo a través de tubos capilares, utiliza la modelización del flujo bifásico, resolviéndose en base a un algoritmo de tipo tramo a tramo. El caudal másico que circula por el capilar, o alternativamente la presión de entrada, se evalúan en base a un algoritmo Newton-Raphson, tanto para las condiciones de flujo crítico (estrangulado) como de flujo no crítico. La comparación realizada entre resultados numéricos y experimentales de diferentes autores muestra un notable grado de correlación. El modelo implementado para la simulación de condensadores y evaporadores de tipo tubular concéntricos, se ha estructurado en base a un algoritmo que en forma segregada e iterativa resuelve cada una de las partes del dominio: flujo en el interior del tubo, flujo en el ánulo, y los sólidos (tubo interior, tubo exterior, aislante). Para las zonas de flujo de fluido se utiliza el modelo desarrollado para flujo bifásico. Para los sólidos, la ecuación de conducción de calor se ha discretizado en base a un esquema numérico implícito del tipo central difference, resolviéndose el sistema de ecuaciones algebraicas en base a un algoritmo Gauss-Seidel apoyado en un TDMA en las direcciones radiales y axiales alternativamente. El tubo aislado ha recibido un tratamiento tridimensional axialsimétrico, mientras que el tubo interior recibe un tratamiento unidimensional.El estudio de evaporadores y condensadores de tipo compacto (tubos y aletas). La simulación desarrollada del flujo bifásico se ha implementado en un algoritmo numérico detallado de resolución del flujo del aire exterior y de los sólidos. Así de forma acoplada se resuelve el flujo en el interior de los conductos, el flujo de aire exterior y la transferencia de calor por conducción a través de los elementos sólidos (tubos y aletas). Se presentan distintos resultados experimentales con fines a la contrastación y validación de las subrutinas desarrolladas. La integración de los diferentes equipos anteriores mencionados en sistemas de refrigeración por compresión de vapor se realiza mediante un modelo numérico para la resolución tanto de sistemas de refrigeración por compresión de una sola etapa, como de sistemas de refrigeración con sobrealimentación de líquido. Ambos algoritmos numéricos se basan en un programa principal el cual llama de manera secuencial a diferentes subrutinas que evalúan los diferentes equipos y partes del sistema hasta que se alcanza la convergencia deseada. Se detalla la infraestructura experimental desarrollada con el objeto de validar las simulaciones numéricas para cada uno de los elementos integrantes del sistema y del sistema en su conjunto. Se presenta la contrastación entre resultados experimentales y numéricos para los sistemas de refrigeración por compresión de una sola etapa, así como resultados numéricos obtenidos por la simulación para el caso del sistema con sobrealimentación de líquido.

Deep knowledge of combustion phenomena is of great scientific and technological interest. In fact, better design of combustion equipments (furnaces, boilers, engines, etc) can contribute both in the energy efficiency and in the reduction of pollutant formation. One of the limitations to design combustion equipments, or even predict simple flames, is the resolution of the mathematical formulation. Analytical solutions are not feasible, and recently numerical techniques have received enormous interest. Even though the ever-increasing computational capacity, the numerical resolution requires large computational resources due to the inherent complexity of the phenomenon (viz. multidimensional flames, finite rate kinetics, radiation in participating media, turbulence, etc). Thus, development of capable mathematical models reducing the complexity and the stiffness as well as efficient numerical techniques are of great interest.The main contribution of the thesis is the analysis and application of the laminar flamelet concept to the numerical simulation of both laminar and turbulent non-premixed flames. Assuming a one-dimensional behavior of combustion phenomena in the normal direction to the flame front, and considering an appropriate coordinates transformation, flamelet approaches reduce the complexity of the problem.The numerical methodology employed is based on the finite volume technique and a parallel multiblock algorithm is used obtaining an excellent parallel efficiency. A post-processing verification tool is applied to assess the quality of the numerical solutions.Before dealing with flamelet approaches, a co-flow partially premixed methane/air laminar flame is studied for different levels of partial premixing. A comprehensive study is performed considering different mathematical formulations based on the full resolution of the governing equations and their validation against experimental data from the literature. Special attention is paid to the prediction of pollutant formation.After the full resolution of the governing equations, the mathematical formulation of the flamelet equations and a deep study of the hypothesis assumed are presented. The non-premixed methane/air laminar flame is considered to apply the flamelet modelling approach, comparing the results with the simulations obtained with the full resolution of the governing equations. Steady flamelets show a proper performance to predict the main flame features when differential diffusion and radiation are neglected, while unsteady flamelets are more suitable to account for these effects as well as pollutant formation. Assumptions of the flamelet equations, the scalar dissipation rate modelling, and the evaluation of the Lagrangian flamelet time for unsteady flamelets are specially analysed. For the numerical simulation of turbulent flames, the mathematical formulation based on mass-weighted time-averaging techniques, using RANS EVM two-equation models is considered. The laminar flamelet concept with a presumed PDF is taken into account. An extended Eddy Dissipation Concept model is also applied for comparison purposes. A piloted non-premixed methane/air turbulent flame is studied comparing the numerical results with experimental data from the literature. A clear improvement in the prediction of slow processes is shown when the transient term in the flamelet equations is retained. Radiation is a key aspect to properly define the thermal field and, consequently, species such as nitrogen oxides. Finally, the consideration of the round-jet anomaly is of significant importance to estimate the flame front position.In conclusion, flamelet modelling simulations are revealed to be an accurate approach for the numerical simulation of laminar and turbulent non-premixed flames. Detailed chemistry can be taken into account and the stiffness of the chemistry term is solved in a pre-processing task. Pollutant formation can be predicted considering unsteady flamelets.

DE LA TESISEs una contribución al desarrollo de un nuevo método de purificación de gas de síntesis en plantas IGCC, que posibilita un incremento de la eficiencia energética de las plantas de obtención de energía, a partir de combustible fósil, y con ello reducir la emisión de gas de efecto invernadero a la atmósfera. Se estudia un proceso nuevo por absorción regenerativa en óxidos metálicos, capaz de eliminar los contaminantes del gas a alta temperatura, aprovechando con ello el contenido energético asociado al intervalo térmico, que los procesos actuales pierden en la purificación por vía húmeda. Se enumeran a continuación los objetivos concretos alcanzados, con las fases del trabajo desarrollado:1- Estudio de la cinética de la reacción de purificación del gas de síntesis, siguiendo las reacciones de absorción no catalíticas, con deposición de sólido, realizando un estudio exhaustivo, para conseguir unos modelos de reacción gas-sólido que contemplen el mayor número de posibilidades. Se ha pretendido crear una herramienta de simulación con el menor número de restricciones a la hora de aplicarlo a cualquier reacción que involucre cambio en la estructura del sólido con la conversión del mismo, para de este modo no tener un modelo cautivo de un determinado tipo de sorbente.2- Se han obtenido los parámetros cinéticos experimentalmente, mediante análisis termogravimétrico, aplicables a los modelos desarrollados, para el caso concreto del sorbente disponible.3- Se han modelizado matemáticamente los posibles reactores que pueden emplearse en la eliminación del compuesto H2S de este gas de gasificación, especialmente los reactores de lecho fluidizado de burbujeo, ya que poseen muy buenas cualidades de homogeneidad de temperatura y concentraciones. Se ha realizado el estudio de sensibilidad a las variaciones de los parámetros de funcionamiento. 4- Se ha construido una planta experimental a pequeña escala que ha permitido probar el sorbente con garantías en los resultados, para lo que se la ha dotado de los elementos de control y análisis de gas necesarios. 5- Se han realizado un número de experimentos suficiente para estudiar el comportamiento del sorbente en estudio, y para la obtención de datos experimentales que han permitido la validación de los diferentes modelos desarrollados.6- Se han obtenido las siguientes conclusiones resumidas:6.1 Se ha obtenido la cinética propia de la reacción gas-sólido no catalítica del sistema estudiado, viendo la influencia de la composición del gas en la velocidad de reacción. La cinética de primer orden describe adecuadamente los procesos de absorción y regeneración, habiéndose obtenido los parámetros cinéticos correspondientes. 6.2 Utilizando los modelos cinéticos desarrollados se proponen los valores de las características estructurales que optimizarían el comportamiento del sólido, en cuanto a su capacidad de retención de contaminante sulfurado: Porosidad no superior al 30% y área superficial de alrededor de 25 m2/g (frente a los valores actuales del 50%y 8,93 m2/g).6.3 El análisis termogravimétrico ha permitido identificar y cuantificar los valores de los diferentes parámetros que intervienen en la reacción gas-sólido: el coeficiente de reacción intrínseca superficial, la difusividad del gas en la capa de sólido formado, y el coeficiente de transferencia de masa gas-sólido. Asimismo las pruebas experimentales realizadas han puesto de manifiesto la influencia de la composición del gas en la cinética de reacción. 6.4 El equipo piloto de reacción experimental a pequeña escala, capaz de operar a alta temperatura y presión, ha permitido la validación de los modelos desarrollados del sistema de desulfurización, demostrándose que son una herramienta adecuada para su estudio, tanto para diseño, como en operación y control avanzado. Como posible continuación se presentan las siguientes posibilidades en la tarea de modelización: Estudiar modelos más precisos basados en ecuaciones de continuidad multifásicas, para su aplicación en sistemas de computación de alto rendimiento, o bien desarrollar modelos basados en reglas lógicas. En cuanto al desarrollo del sorbente, se pone de manifiesto la necesidad de continuar en su desarrollo para alcanzar los valores de las características señaladas. / OF THE THESISThis thesis is a contribution to the development of a new method of synthesis gas purification in IGCC plants, that make possible a further increase of the energy efficiency, and with this to reduce the emission of greenhouse gas to the atmosphere. It is studied a regenerative absorption in metallic oxides, capable of the elimination of the pollutants from the high-temperature gas, but taking advantage of the energy contents associated with the high temperature process streams, which is lost in the purification by wet route. The objectives and conclusions reached in the development of the thesis work are summarized below:1- The study of the kinetics of the purification reactions of the synthesis gas, non catalytic absorption reactions, with solid deposition. An exhaustive study has been carried out, to obtain some gas-solid reaction models that envisage a great number of possibilities. It has been created a simulation tool with a reduced number of restrictions that makes possible its use with any reaction that involves changes in the structure of the solid thus ensuring generality and independence of a given type of sorbent.2- The kinetic parameters have been obtained by thermogravimetric analysis for the concrete case of the available sorbent to be used in the testing of the developed models. 3- Appropriated type of reactors that can be employed in the elimination of the H2S compound of the gas contemplated have been identified and modeled. Special attention has been given to the bubbling fluidized bed reactors, since they possess very good properties of temperature and concentrations homogeneity. It has been accomplished the sensibility study to the variations of the operation parameters.4- An experimental bubbling fluidized bed reactor plant at bench-scale has been built that has permitted to extract the data from the selected sorbent with guarantees in the results.5- A sufficient number of experiments to study the behaviour of the selected sorbent under study have been accomplished successfully, and the experimental data obtained have permitted the validation of the different developed models.6- The specific conclusions reached are: 6.1. The study of the no-catalytic gas-solid reaction of the system studied shows that kinetics of first order describe adequately the absorption and regeneration processes.6.2. The accomplished experimental tests have shown the influence of the composition of the gas in the kinetics mechanism. The kinetic parameters are very dependent of the composition of the gas, not only of the reactive species, specially of the components that determine the reducing power.6.3. The thermogravimetric analysis has permitted to identify the values of the different parameters that appear in the gas reaction-solid: the coefficient of superficial intrinsic reaction, the diffusivity of the gas in the formed solid layer, and the gas-solid mass transfer coefficient.6.4. Using the developed kinetic models, the values of the structural characteristics that would optimise the behaviour of the solid are proposed, concerning its capacity of retention of sulfured pollutant: Porosity not superior to 30% and superficial area of about 25 m2 / g (as compared to the current values of the 50% and 8,93 m2 / g).6.5. The experimental small scale BFB pilot plant, capable of operating at high-temperature and pressure, has permitted the validation of the developed models for the desulphurisation system. The models developed have demonstrated that they are an adequate tool for the study, design, as well as the operation and advanced control of the system under study. As possible continuation of this thesis work in the modelization task the following is indicated: To study more accurate models based on fundamental multiphasic continuity equations, or else to develop models based on logic rules. With respect to development of the sorbent, the need of continuing in its development to reach the values of the indicated characteristics is shown.