No / Two numerical particle models, the Smoothed Particle Hydrodynamics (SPH) and Moving Particle Semi-implicit (MPS) methods, coupled with a sub-particle scale (SPS) turbulence model, are presented to simulate free surface flows. Both SPH and MPS methods have the advantages in that the governing Navier¿Stokes equations are solved by Lagrangian approach and no grid is needed in the computation. Thus the free surface can be easily and accurately tracked by particles without numerical diffusion. In this paper different particle interaction models for SPH and MPS methods are summarized and compared. The robustness of two models is validated through experimental data of a dam-break flow. In addition, a series of numerical runs are carried out to investigate the order of convergence of the models with regard to the time step and particle spacing. Finally the efficiency of the incorporated SPS model is further demonstrated by the computed turbulence patterns from a breaking wave. It is shown that both SPH and MPS models provide a useful tool for simulating free surface flows

京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第22047号 / 工博第4628号 / 新制||工||1722(附属図書館) / 京都大学大学院工学研究科社会基盤工学専攻 / (主査)教授 後藤 仁志, 教授 細田 尚, 准教授 KHAYYER,Abbas / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DGAM

Two numerical particle models, the Smoothed Particle Hydrodynamics (SPH) and Moving Particle Semi-implicit (MPS) methods, coupled with a sub-particle scale (SPS) turbulence model, are presented to simulate free surface flows. Both SPH and MPS methods have the advantages in that the governing Navier¿Stokes equations are solved by Lagrangian approach and no grid is needed in the computation. Thus the free surface can be easily and accurately tracked by particles without numerical diffusion. In this paper different particle interaction models for SPH and MPS methods are summarized and compared. The robustness of two models is validated through experimental data of a dam-break flow. In addition, a series of numerical runs are carried out to investigate the order of convergence of the models with regard to the time step and particle spacing. Finally the efficiency of the incorporated SPS model is further demonstrated by the computed turbulence patterns from a breaking wave. It is shown that both SPH and MPS models provide a useful tool for simulating free surface flows.

O concreto é um dos materiais de construção civil mais versáteis, podendo se adaptar a formas diversas quando em seu estado fresco e resistindo a grandes cargas de compressão em seu estado rígido. No entanto, em estruturas mais densamente armadas ou com geometrias mais complexas, pode-se apresentar dificuldades para a moldagem, causando falhas no preenchimento da forma, o que reduz a capacidade resistente da peça e sua vida útil. Neste trabalho foi utilizado o método de partículas lagrangeanas Moving Particle Semi-Implicit (MPS) como base para um simulador para o estudo do comportamento do escoamento de pastas e argamassas cujo comportamento pode se aproximado por modelos reológicos como Bingham e Herschel-Bulkley. Foram propostos módulos para a simulação da viscosidade não-newtoniana, variação térmica no processo de cura e modelagem de turbulência. Para modelar a variação de viscosidade de um fluido não-newtoniano foi utilizado o modelo de Herschel-Bulkley. O modelo de Herschel-Bulkley possui uma singularidade para taxas de deformação muito pequenas, que resulta em valores de viscosidade infinitas, dificuldade contornada pela solução proposta por Papanastasiou (1987). Na modelagem térmica foram analisados dois modelos de dissipação, sendo um original do método e outro calculado por meio do divergente do gradiente utilizando os modelos de partículas e que teria melhores resultados para o cálculo da dissipação térmica. Também foi modelada a convecção térmica, utilizando para isso a hipótese de Boussinesq. A reação de hidratação do concreto foi modelada utilizando uma equação do tipo Hill para representar a elevação de temperatura obtida por meio um ensaio adiabático. Para complementar as simulações, foi utilizado o modelo de turbulência Detached Eddy Simulation (DES), baseado no método Large Eddy Simulation (LES), que utiliza um modelo de parede para simular a interação do fluido. Para a implementação deste modelo de turbulência foi proposto um algoritmo para o cálculo da distância da partícula de fluido à parede. Este algoritmo utiliza estruturas de dados já existentes no método de partículas de modo que sua execução requer muito menos recursos que a busca binária. Apesar do trabalho ter se guiado pela simulação do concreto fresco, que é um material particularmente complexo, outros escoamentos encontrados na engenharia civil também podem ser beneficiados pelo método, como os estudos do escoamento em sistemas prediais de água e esgoto, do escoamento e prevenção de erosão em rios e córregos, do escorregamento de encostas, dos reatores para depuração de esgotos, entre outros. / The concrete is one of the more versatile civil construction materials, it can adapt to various forms when in its fresh state while resisting to high compression loads in its rigid state. However, in some cases like densely reinforced concrete structures and complex geometry concrete structures can present issues to the casting, and failure in proper form filling can occur. These failures can reduce the resistance and the lifetime of the structure. This work used a simulator based on the lagrangean particle method called Moving Particle Semi-Implicit (MPS) to study the concrete behavior in its distinctive characteristics. Also, this work proposed modules to simulate the non-Newtonian viscosity, the thermal process of concrete cure and to model the turbulent flow. To model the variation of viscosity of a non-Newtonian fluid, the Herschel-Bulkley model, which relates the shear stress with the strain rate, was applied. The Herschel-Bulkley model has a singularity at low strain rates, which results in infinite viscosities. To overcome this issue, Papanastasiou (1987) proposed a modification in the model in order to eliminate the singularity. For the thermal modeling, two models for thermal dissipation were compared, the original method and other calculated from the divergence of gradient using the differential operators for the particle model and that could present improved results for the thermal dissipation calculation. Also, the thermal convection was modeled using the Boussinesq hypothesis. The hydration reaction of the concrete was modeled using a Hill type equation to reproduce the temperature elevation. In addition, a Detached Eddy Simulation (DES) based turbulence model with a simple wall model in the interaction of wall and fluid was applied in the simulations. To improve the turbulence model, an algorithm to calculate the distance between fluid and the nearest wall particle was proposed. The algorithm uses already calculated information from particles so that the execution requires much less resources than a binary search. Although the work has been focused on to the modeling of fresh concrete simulation, which is a particularly complex material, other flows found in civil engineering can also be benefited by the method, such as studies of drainage in water and sewage systems, drainage and prevention of erosion into rivers and streams, prevention and damage mitigation of landslides, reactors for sewage treatment among many others.

O concreto é um dos materiais de construção civil mais versáteis, podendo se adaptar a formas diversas quando em seu estado fresco e resistindo a grandes cargas de compressão em seu estado rígido. No entanto, em estruturas mais densamente armadas ou com geometrias mais complexas, pode-se apresentar dificuldades para a moldagem, causando falhas no preenchimento da forma, o que reduz a capacidade resistente da peça e sua vida útil. Neste trabalho foi utilizado o método de partículas lagrangeanas Moving Particle Semi-Implicit (MPS) como base para um simulador para o estudo do comportamento do escoamento de pastas e argamassas cujo comportamento pode se aproximado por modelos reológicos como Bingham e Herschel-Bulkley. Foram propostos módulos para a simulação da viscosidade não-newtoniana, variação térmica no processo de cura e modelagem de turbulência. Para modelar a variação de viscosidade de um fluido não-newtoniano foi utilizado o modelo de Herschel-Bulkley. O modelo de Herschel-Bulkley possui uma singularidade para taxas de deformação muito pequenas, que resulta em valores de viscosidade infinitas, dificuldade contornada pela solução proposta por Papanastasiou (1987). Na modelagem térmica foram analisados dois modelos de dissipação, sendo um original do método e outro calculado por meio do divergente do gradiente utilizando os modelos de partículas e que teria melhores resultados para o cálculo da dissipação térmica. Também foi modelada a convecção térmica, utilizando para isso a hipótese de Boussinesq. A reação de hidratação do concreto foi modelada utilizando uma equação do tipo Hill para representar a elevação de temperatura obtida por meio um ensaio adiabático. Para complementar as simulações, foi utilizado o modelo de turbulência Detached Eddy Simulation (DES), baseado no método Large Eddy Simulation (LES), que utiliza um modelo de parede para simular a interação do fluido. Para a implementação deste modelo de turbulência foi proposto um algoritmo para o cálculo da distância da partícula de fluido à parede. Este algoritmo utiliza estruturas de dados já existentes no método de partículas de modo que sua execução requer muito menos recursos que a busca binária. Apesar do trabalho ter se guiado pela simulação do concreto fresco, que é um material particularmente complexo, outros escoamentos encontrados na engenharia civil também podem ser beneficiados pelo método, como os estudos do escoamento em sistemas prediais de água e esgoto, do escoamento e prevenção de erosão em rios e córregos, do escorregamento de encostas, dos reatores para depuração de esgotos, entre outros. / The concrete is one of the more versatile civil construction materials, it can adapt to various forms when in its fresh state while resisting to high compression loads in its rigid state. However, in some cases like densely reinforced concrete structures and complex geometry concrete structures can present issues to the casting, and failure in proper form filling can occur. These failures can reduce the resistance and the lifetime of the structure. This work used a simulator based on the lagrangean particle method called Moving Particle Semi-Implicit (MPS) to study the concrete behavior in its distinctive characteristics. Also, this work proposed modules to simulate the non-Newtonian viscosity, the thermal process of concrete cure and to model the turbulent flow. To model the variation of viscosity of a non-Newtonian fluid, the Herschel-Bulkley model, which relates the shear stress with the strain rate, was applied. The Herschel-Bulkley model has a singularity at low strain rates, which results in infinite viscosities. To overcome this issue, Papanastasiou (1987) proposed a modification in the model in order to eliminate the singularity. For the thermal modeling, two models for thermal dissipation were compared, the original method and other calculated from the divergence of gradient using the differential operators for the particle model and that could present improved results for the thermal dissipation calculation. Also, the thermal convection was modeled using the Boussinesq hypothesis. The hydration reaction of the concrete was modeled using a Hill type equation to reproduce the temperature elevation. In addition, a Detached Eddy Simulation (DES) based turbulence model with a simple wall model in the interaction of wall and fluid was applied in the simulations. To improve the turbulence model, an algorithm to calculate the distance between fluid and the nearest wall particle was proposed. The algorithm uses already calculated information from particles so that the execution requires much less resources than a binary search. Although the work has been focused on to the modeling of fresh concrete simulation, which is a particularly complex material, other flows found in civil engineering can also be benefited by the method, such as studies of drainage in water and sewage systems, drainage and prevention of erosion into rivers and streams, prevention and damage mitigation of landslides, reactors for sewage treatment among many others.

The design and optimization of entrained flow gasifiers is conducted more and more via computational fluid dynamics (CFD). A detailed resolution of single coal particles within such simulations is nowadays not possible due to computational limitations. Therefore the coal particle conversion is often represented by simple 0-D models. For an optimization of such 0-D models a precise understanding of the physical processes at the boundary layer and within the particle is necessary. In real gasifiers the particles experience Reynolds numbers up to 10000. However in the literature the conversion of coal particles is mainly regarded under quiescent conditions. Therefore an analysis of the conversion of single particles is needed. Thereto the computational fluid dynamics can be used. For the detailed analysis of single reacting particles under flow conditions a CFD model is presented. Practice-oriented parameters as well as features of the CFD model result from CFD simulations of a Siemens 200MWentrained flow gasifier. The CFD model is validated against an analytical model as well as two experimental data-sets taken from the literature. In all cases good agreement between the CFD and the analytics/experiments is shown. The numerical model is used to study single moving solid particles under combustion conditions. The analyzed parameters are namely the Reynolds number, the ambient temperature, the particle size, the operating pressure, the particle shape, the coal type and the composition of the gas. It is shown that for a wide range of the analyzed parameter range no complete flame exists around moving particles. This is in contrast to observations made by other authors for particles in quiescent atmospheres. For high operating pressures, low Reynolds numbers, large particle diameters and high ambient temperatures a flame exists in the wake of the particle. The impact of such a flame on the conversion of the particle is low. For high steam concentrations in the gas a flame appears, which interacts with the particle and influences its conversion. Furthermore the impact of the Stefan-flow on the boundary layer of the particle is studied. It is demonstrated that the Stefan-flow can reduce the drag coefficient and the Nusselt number for several orders of magnitude. On basis of the CFD results two new correlations are presented for the drag coefficient and the Nusselt number. The comparison between the correlations and the CFD shows a significant improvement of the new correlations in comparison to archived correlations. The CFD-model is further used to study moving single porous particles under gasifying conditions. Therefore a 2-D axis-symmetric system of non-touching tori as well as a complex 3-D geometry based on the an inverted settlement of monodisperse spheres is utilized. With these geometries the influence of the Reynolds number, the ambient temperature, the porosity, the intrinsic surface and the size of the radiating surface is analyzed. The studies show, that the influence of the flow on the particle conversion is moderate. In particular the impact of the flow on the intrinsic transport and conversion processes is mainly negligible. The size of the radiating surface has a similar impact on the conversion as the flow in the regarded parameter range. On basis of the CFD calculations two 0-D models for the combustion and gasification of moving particles are presented. These models can reproduce the results predicted by the CFD sufficiently for a wide parameter range.

The design and optimization of entrained flow gasifiers is conducted more and more via computational fluid dynamics (CFD). A detailed resolution of single coal particles within such simulations is nowadays not possible due to computational limitations. Therefore the coal particle conversion is often represented by simple 0-D models. For an optimization of such 0-D models a precise understanding of the physical processes at the boundary layer and within the particle is necessary. In real gasifiers the particles experience Reynolds numbers up to 10000. However in the literature the conversion of coal particles is mainly regarded under quiescent conditions. Therefore an analysis of the conversion of single particles is needed. Thereto the computational fluid dynamics can be used. For the detailed analysis of single reacting particles under flow conditions a CFD model is presented. Practice-oriented parameters as well as features of the CFD model result from CFD simulations of a Siemens 200MWentrained flow gasifier. The CFD model is validated against an analytical model as well as two experimental data-sets taken from the literature. In all cases good agreement between the CFD and the analytics/experiments is shown. The numerical model is used to study single moving solid particles under combustion conditions. The analyzed parameters are namely the Reynolds number, the ambient temperature, the particle size, the operating pressure, the particle shape, the coal type and the composition of the gas. It is shown that for a wide range of the analyzed parameter range no complete flame exists around moving particles. This is in contrast to observations made by other authors for particles in quiescent atmospheres. For high operating pressures, low Reynolds numbers, large particle diameters and high ambient temperatures a flame exists in the wake of the particle. The impact of such a flame on the conversion of the particle is low. For high steam concentrations in the gas a flame appears, which interacts with the particle and influences its conversion. Furthermore the impact of the Stefan-flow on the boundary layer of the particle is studied. It is demonstrated that the Stefan-flow can reduce the drag coefficient and the Nusselt number for several orders of magnitude. On basis of the CFD results two new correlations are presented for the drag coefficient and the Nusselt number. The comparison between the correlations and the CFD shows a significant improvement of the new correlations in comparison to archived correlations. The CFD-model is further used to study moving single porous particles under gasifying conditions. Therefore a 2-D axis-symmetric system of non-touching tori as well as a complex 3-D geometry based on the an inverted settlement of monodisperse spheres is utilized. With these geometries the influence of the Reynolds number, the ambient temperature, the porosity, the intrinsic surface and the size of the radiating surface is analyzed. The studies show, that the influence of the flow on the particle conversion is moderate. In particular the impact of the flow on the intrinsic transport and conversion processes is mainly negligible. The size of the radiating surface has a similar impact on the conversion as the flow in the regarded parameter range. On basis of the CFD calculations two 0-D models for the combustion and gasification of moving particles are presented. These models can reproduce the results predicted by the CFD sufficiently for a wide parameter range.:List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XIII Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XV Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XIX 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 1.1 State of the Art in Carbon Conversion Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.1.1 Combustion of Solid Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.1.2 Gasification of Porous Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.2 Classification of the Present Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 1.3 Overview of the Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 2 Basic Theory and Model Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.1 Geometry and Length Scales of Coal Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9 2.2 Conditions in a Siemens Like 200 MW Entrained Flow Gasifier . . . . . . . . . . . . . . . . . . . . 11 2.2.1 Velocity Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12 2.2.2 Temperature Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.2.3 Particle Volume Fraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 2.3 Time Scales of the Physical Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.4 Basic Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21 2.5 Conservation Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 2.6 Gas Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26 2.7 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.8 Numerics and Solution Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30 2.9 Mesh and Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31 3 CFD-based Oxidation Modeling of a Non-Porous Carbon Particle . . . . . . . . . . . . . . . . . . . . .37 3.1 Chemical Reaction System for Combustion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37 3.1.1 Heterogeneous Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37 3.1.2 Homogeneous Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40 3.1.3 Comparison of the Semi-Global vs. Reduced Reaction Mechanisms for the Gas Phase . .41 3.2 Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43 3.2.1 Validation Against an Analytical Solution of the Two-Film Model . . . . . . . . . . . . . . . . . .43 3.2.2 Validation Against Experiments I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 3.2.3 Validation Against Experiments II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49 3.3 Influence of Ambient Temperature and Reynolds Number . . . . . . . . . . . . . . . . . . . . . . . .51 3.4 Influence of Heterogeneous Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 3.5 Influence of Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .61 3.6 Influence of Operating Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66 3.7 Influence of Particle Diameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .70 3.8 The influence of Particle Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 3.9 Impact of Stefan Flow on the Boundary Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 3.9.1 Impact of Stefan Flow on the Drag Coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .83 3.9.2 Impact of Stefan Flow on the Nusselt and Sherwood Number . . . . . . . . . . . . . . . . . . . .85 3.10 Single-Film Sub-Model vs. CFD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 3.11 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 4 CFD-based Numerical Modeling of Partial Oxidation of a Porous Carbon Particle . . . . . . . . . .99 4.1 Chemical Reaction System for Gasification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 4.1.1 Heterogeneous Chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .100 4.1.2 Homogeneous Chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 4.2 Two-Dimensional Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 4.2.1 Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 4.2.2 Influence of Reynolds Number and Ambient Temperature . . . . . . . . . . . . . . . . . . . . . .109 4.2.3 Influence of Porosity and Internal Surface . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . 120 4.3 Comparative Three-Dimensional Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 4.3.1 Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126 4.3.2 Results of the 3-D Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 4.4 Extended Sub-Model for Gasification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .133 4.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .138 5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . .141 5.1 Summary of This Work . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . .141 5.2 Recommendations for Future Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .145 6 Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 6.1 Appendix I: Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 6.2 Appendix II: Two-Film Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 6.3 Appendix III: Sub-Model for the Combustion of Solid Particles . . . . . . . . . . . . . . . . . . . . 160 6.4 Appendix IV: Sub-Model for the Gasification of Porous Particles . . . . . . . . . . . . . . . . . . . 161