Pour des raisons de sécurité, les moteurs aéronautiques doivent pouvoir redémarrer en vol sur toute leur plage d'opération. Mais les contraintes sur les émissions polluantes nécessitent le développement de nouvelles chambres de combustion dont la conception peut détériorer les capacités d'allumage du moteur. Afin d'améliorer la compréhension du processus d'allumage et d'aider à l'optimisation de la conception, les recherches actuelles combinent les études expérimentales de plus en plus complexes et les simulation numériques hautes fidélités. Dans ce travail, l'étude numérique du processus d'allumage des chambres de combustion aéronautiques, de l'étincelle à la propagation azimutale de la flamme, est conduite avec plusieurs objectifs: améliorer la robustesse et la confiance de l'outil LES pour l'étude de l'allumage, étudier les mécanismes qui affectent l'allumage dans des conditions représentatives des conditions réelles et enfin améliorer les méthodes bas-ordre pour la prédiction des performances d'allumage. Dans une première partie, la SGE d'un monobruleur installé au CORIA permet de mettre en évidence les bons résultats de la LES et de construire une base de données pour l'analyses des mécanismes d'extinction. Ces données sont aussi utilisées pour développer une méthodologie permettant de prédire les performances d'allumage à bas coût en utilisant les résultats d'une SGE non-réactive. Dans une seconde partie, la propagation inter-secteur est investiguée par l'étude de deux cas expérimentaux et la SGE est capable de reproduire les modes de propagation mais aussi les temps d'allumage avec précision. Sur la bases de ces bons résultats, une analyse plus fine de la simulation permet d'identifier les mécanismes qui contrôlent la propagation de la flamme. / For safety reasons, in-flight relight of the engine must be guaranteed over a wide range of operating conditions but the increasing stringency of pollutant emission constraints requires the development of new aero-engine combustor whose design might be detrimental to the ignition capability. To improve the knowledge of the ignition process in aeronautical gas turbines and better combine conflicting technological solutions, current research relies on both complex experimental investigation and high fidelity numerical simulations. In this work, numerical study of the ignition process in gas turbines from the energy deposit to the light-around is performed with several objectives: increase the level of confidence of Large Eddy Simulations tool for the analysis of the ignition process, investigate the mechanisms controlling ignition in conditions representative of realistic aeronautical gas turbine flows and improve the low-order methodologies for the prediction of ignition performance. In a first part, LES of the single burner installed at CORIA (France) is carried out and allows to highlight the LES accuracy and to build a database on which the main mechanisms controlling the ignition success are identified. Based on these results, a methodology is developed to predict the ignition performance at a low computational cost using the non-reacting flow statistics only. In a second part, the light-around process is studied on two experimental set-ups and the very good agreement of the LES results with experiments is the starting point from an analysis of the mechanisms driving the flame propagation process.

A interação turbulência-radiação (TRI, do inglês Turbulence-Radiation Interaction) resulta do acoplamento altamente não linear entre flutuações da intensidade de radiação e flutuações da temperatura e da composição química do meio, e tem-se demonstrado experimentalmente, teoricamente e numericamente que este é um fenômeno relevante em diversas aplicações envolvendo altas temperaturas, especialmente em problemas reativos. Neste trabalho, o TRI é analisado em um escoamento interno não reativo de um gás participante que se desenvolve em um duto de seção transversal quadrada, para diferentes intensidades de turbulência do escoamento e considerando duas espécies distintas para a composição do fluido de trabalho (dióxido de carbono e vapor de água). O objetivo central é avaliar como a inclusão ou não da variação espectral das propriedades radiativas do meio no cálculo influencia a magnitude do TRI. Isso é feito através de simulações numéricas no código de dinâmica dos fluidos computacional Fire Dynamics Simulator (FDS), que resolve, através do método dos volumes finitos, as equações fundamentais que regem o problema – isto é, os balanços de massa, de quantidade de movimento e de energia e a equação de estado – em uma formulação adequada para baixos números de Mach, utilizando um algoritmo de solução explícito e de segunda ordem no tempo e no espaço. A turbulência é modelada através da simulação de grandes escalas (LES, do inglês Large Eddy Simulation), empregando-se o modelo de Smagorinsky dinâmico para o fechamento dos termos submalha; para a radiação térmica, o método dos volumes finitos é utilizado na discretização da equação da transferência radiativa e os modelos do gás cinza e da soma-ponderada-de-gases-cinza (WSGG, do inglês Weighted-Sum-of-Gray-Gases) são implementados como forma de desconsiderar e de incluir a dependência espectral das propriedades radiativas, respectivamente. A magnitude do TRI sobre o problema é avaliada através de diferenças entre as médias temporais dos fluxos de calor superficiais e do termo fonte radiativo obtidas em cálculos que consideram os efeitos do fenômeno e cálculos que os negligenciam. Em geral, a interação turbulência-radiação mostrou ser pouco importante em todos os casos considerados, o que concorda com resultados de outros estudos sobre o tema em escoamento não reativos. Com o modelo WSGG, as contribuições do fenômeno foram maiores do que com a hipótese do gás cinza, evidenciando que a inclusão da variação espectral na solução do problema radiativo tem um impacto sobre a magnitude dos efeitos do TRI. Além disso, é feita uma discussão, em parte inédita no contexto do TRI, sobre diferentes metodologias para a análise do fenômeno. Finalmente, é proposto um fator de correção para o termo fonte radiativo médio no modelo WSGG, que é validado através de sua implementação nos casos simulados. Em estudos futuros, uma análise de sensibilidade sobre os termos constituintes desse fator de correção pode levar a um melhor entendimento de como as flutuações de temperatura se correlacionam com o fenômeno da interação turbulência-radiação. / Turbulence-radiation interaction (TRI) results from the highly non-linear coupling between fluctuations of radiation intensity and fluctuations of temperature and chemical composition of the medium, and its relevance in a number of high-temperature problems, especially when chemical reactions are included, has been demonstrated experimentally, theoretically, and numerically. In the present study, the TRI is analyzed in a channel flow of a non-reactive participating gas for different turbulence intensities of the flow at the inlet and considering two distinct species for the medium composition (carbon dioxide and water vapor). The central objective is to evaluate how the inclusion or not of the spectral variation of the radiative properties of a participating gas in the radiative transfer calculations affects the turbulence-radiation interaction. With this purpose, numerical simulations are performed using the computational fluid dynamics Fortranbased code Fire Dynamics Simulator, that employs the finite volume method to solve a form of the fundamental equations – i.e., the mass, momentum and energy balances and the state equation – appropriate for low Mach number flows, through an explicit second-order (both in time and in space) core algorithm. Turbulence is modeled by the large eddy simulation approach (LES), using the dynamic Smagorinsky model to close the subgrid-scale terms; for the thermal radiation part of the problem, the finite volume method is used for the discretization of the radiative transfer equation and the gray gas and weighted-sum-of-gray-gases (WSGG) models are implemented as a way to omit and consider the spectral dependence of the radiative properties, respectively. The TRI magnitude in the problem is evaluated by differences between values for the time-averaged heat fluxes at the wall (convective and radiative) and for the time-averaged radiative heat source calculated accounting for and neglecting the turbulence-radiation interaction effects. In general, TRI had little importance over all the considered cases, a conclusion that agrees with results of previous studies. When using the WSGG model, the contributions of the phenomenon were greater that with the gray gas hypothesis, demonstrating that the inclusion of the spectral variance in the solution of the radiative problem has an impact in the TRI effects. Furthermore, this paper presents a discussion, partly unprecedented in the context of the turbulence-radiation interaction, about the different methodologies that can be used for the TRI analysis. Finally, a correction factor is proposed for the time-averaged radiative heat source in the WSGG model, which is then validated by its implementation in the simulated cases. In future studies, a sensibility analysis on the terms that compose this factor can lead to a better understanding of how fluctuations of temperature correlate with the turbulence-radiation interaction phenomenon.

O presente trabalho tem como objetivo estudar escoamentos de tornados e sua ação sobre corpos imersos empregando ferramentas numéricas da Engenharia do Vento Computacional (EVC). Os tornados constituem-se atualmente em uma das causas de desastres naturais no Brasil, especialmente nas regiões sul e sudeste do país, como também em alguns países vizinhos. Os efeitos gerados são geralmente localizados e de curta duração, podendo ser devastadores dependendo da escala do tornado. Tais características dificultam a realização de estudos detalhados a partir de eventos reais, o que levou ao desenvolvimento de modelos experimentais e numéricos. A abordagem numérica é utilizada neste trabalho para a simulação de tornados, a qual se baseia nas equações de Navier-Stokes e na equação de conservação de massa, considerando a hipótese de pseudo-compressibilidade e condições isotérmicas. Para escoamentos com turbulência utiliza-se a Simulação Direta de Grandes Escalas com o modelo clássico de Smagorinsky para as escalas inferiores à resolução da malha (Large Eddy Simulation ou LES em inglês). A discretização das equações fundamentais do escoamento se realiza com um esquema explícito de dois passos de Taylor-Galerkin, onde o Método dos Elementos Finitos é empregado na discretização espacial utilizando-se o elemento hexaédrico trilinear isoparamétrico com um ponto de integração e controle de modos espúrios Na presença de corpos imersos que se movem para simular os deslocamentos dos tornados, o escoamento é descrito cinematicamente através de uma formulação Arbitrária Lagrangeana-Euleriana (ALE) que inclui um esquema de movimento de malha. Tornados são reproduzidos através da simulação numérica de dispositivos experimentais e do Modelo de Vórtice Combinado de Rankine (RCVM). Exemplos clássicos da Dinâmica dos Fluidos Computacional são apresentados inicialmente para a verificação das ferramentas numéricas implementadas. Finalmente, problemas envolvendo tornados móveis e estacionários são analisados, incluindo sua ação sobre corpos imersos. Nos modelos baseados em experimentos, a variação da relação de redemoinho determinou os diferentes padrões de escoamento observados no laboratório. Nos exemplos de modelo de vórtice, quando o tornado impactou o corpo imerso gerou picos de forças em todas as direções e, após a passar pelo mesmo, produziu uma alteração significativa na estrutura do vórtice. / Analyses of tornado flows and its action on immersed bodies using numerical tools of Computational Wind Engineering (CWE) are the main aims of the present work. Tornadoes are currently one of the causes of natural disasters in Brazil, occurring more frequently in the southern and southeastern regions of the country, as well as in some neighboring countries. Effects are usually localized, presenting a short time interval, which can be devastating depending on the scale of the tornado. These characteristics difficult to carry out detailed studies based on real events, leading to the development of experimental and numerical models. The numerical approach is used in this work for the simulation of tornadoes, which is based on the Navier-Stokes equations and the mass conservation equation, considering the hypothesis of pseudo-compressibility and isothermal conditions. For turbulent flows, Large Eddy Simulation (LES) is used with the classical Smagorinsky model for sub-grid scales Discretization is performed the explicit two-step Taylor-Galerkin scheme, where the Finite Element Method is used in spatial discretization using isoparametric trilinear hexahedral elements with one-point quadrature and hourglass control. In the presence of immersed bodies that are moving in order to simulate translating tornadoes, the flow is kinematically described through a Lagrangian-Eulerian Arbitrary (ALE) formulation, which includes a mesh motion scheme. Tornadoes are reproduced using numerical simulation of experimental devices and the Rankine Combined Vortex Model (RCVM). Classical examples of Computational Fluid Dynamics are presented initially for the verification of the numerical tools implemented here. Finally, problems involving moving and stationary tornadoes are analyzed, including their actions on immersed bodies. For models based on experiments, the variation of the swirl ratio determined the different flow patterns observed in the laboratory. In the vortex model examples, when the tornado impacted on the immersed body, peaks of forces were generated in all directions and, after passing over it, a significant change in the structure of the vortex was produced.

Les simulations sont utilisées pour concevoir des chambres de combustion industrielles robustes et peu polluantes. Parmi les polluants, l’émission de particules de suies constitue une question sociétale et une priorité politico- industrielle, en raison de leurs impacts néfastes sur la santé et l'environnement. La taille des particules de suies joue un rôle important sur ces effets. Il est donc important de prévoir non seulement la masse totale ou le nombre de particules générées, mais également leur distribution en taille (PSD). De plus, les suies peuvent jouer un rôle important dans le rayonnement thermique. Dans des configurations confinées, la prédiction des transferts de chaleur est une question clé pour augmenter la robustesse des chambres de combustion. Afin de déterminer correctement ces transferts, les flux radiatifs et de conducto-convectifs aux parois doivent être pris en compte. Enfin, la température pariétale est aussi contrôlée par les transferts conjugués de chaleur entre les domaines fluides et solides. L’ensemble de ces transferts thermiques impactent la stabilisation de la flamme, la formation de polluants et la production de suies elle-même. Il existe donc un couplage complexe entre ces phénomènes et la simulation d'un tel problème multiphysique est aujourd'hui reconnu comme un important défi. Ainsi, l'objectif de cette thèse est de développer une modélisation multiphysique permettant la simulation de flammes suitées turbulentes avec le rayonnement thermique et les transferts conjugués de chaleur associés aux parois. Les méthodes retenues sont basées sur la Simulation aux Grandes Échelles (LES), une description en taille des suies, des transferts conjugués et un code Monte Carlo pour le rayonnement. La combinaison de telles approches est réalisable grâce aux ressources de calcul aujourd’hui disponibles afin d’obtenir des résultats de référence. Le manuscrit est organisé en trois parties. La première partie se concentre sur le développement d'un modèle détaillé pour la description de la production de suies dans les flammes laminaires. Pour cela, la méthode sectionnelle est retenue ici car elle permet la description de la PSD. La méthode est validée sur des flammes laminaires éthylène/air. Dans la deuxième partie, un formalisme LES spécifique à la méthode sectionnelle est développé et utilisé pour étudier deux flammes turbulentes : une flamme jet non-prémélangée et une flamme swirlée pressurisée confinée. Les champs de température et de fraction volumique de suies sont comparés aux données expérimentales. De bonnes prédictions sont obtenues et l’évolution des particules de suies dans de telles flammes est analysée à travers l'étude de l’évolution de leur PSD. Dans ces premières simulations, les pertes de chaleur aux parois reposent sur des mesures expérimentales de la température aux parois, et un modèle de rayonnement simple. Dans la troisième partie, une approche Monte Carlo permettant de résoudre l'équation de transfert radiatif avec des propriétés radiatives détaillées des phases gazeuse et solide est utilisée et couplée au solveur LES. Cette approche est appliquée à l'étude de la flamme jet turbulente. La prédiction des flux thermiques est comparée aux données expérimentales et la nature des transferts radiatifs est étudiée. Ensuite, une modélisation couplée de la combustion turbulente prenant en compte la production de suies, les transferts conjugués de chaleur et le rayonnement thermique est proposée en couplant les trois codes dédiés. Cette stratégie est appliquée pour la simulation du brûleur pressurisé confiné. L'approche proposée permet à la fois de prédire la température des parois et la bonne stabilisation de la flamme. Les processus de formation de suies se révèlent être affectés par la modélisation des transferts thermiques. Ceci souligne l’importance d’une description précise de ces transferts thermiques dans les développements futurs de modèles de production de suies et leur validation. / Numerical simulations are used by engineers to design robust and clean industrial combustors. Among pollutants, soot control is an urgent societal issue and a political-industrial priority, due to its harmful impact on health and environment. Soot particles size plays an important role in its negative effect. It is therefore important to predict not only the total mass or number of emitted particles, but also their population distribution as a function of their size. In addition, soot particles can play an important role in thermal radiation. In confined configurations, controlling heat transfer related to combustion is a key issue to increase the robustness and the life cycle of combustors by avoiding wall damages. In order to correctly determine these heat losses, radiative and wall convective heat fluxes must be accounted for. They depend on the wall temperature, which is controlled by the conjugate heat transfer between the fluid and solid domains. Heat transfer impacts the flame stabilization, pollutants formation and soot production itself. Therefore, a complex coupling exists between these phenomena and the simulation of such a multi-physics problem is today recognized as an extreme challenge in combustion, especially in a turbulent flow, which is the case of most industrial combustors. Thus, the objective of this thesis is to develop a multi-physics modeling enabling the simulation of turbulent sooting flames including thermal radiation and wall heat transfer. The retained methods based on Large-Eddy Simulation (LES), a soot sectional model, conjugate heat transfer, a Monte Carlo radiation solver are combined to achieve a stateof- the-art framework. The available computational resources make nowadays affordable such simulations that will yield present-day reference results. The manuscript is organized in three parts. The first part focuses on the definition of a detailed model for the description of soot production in laminar flames. For this, the sectional method is retained here since it allows the description of the particle size distribution (PSD). The method is validated on laminar premixed and diffusion ethylene/air flames before analyzing the dynamics of pulsed diffusion flames. In the second part, an LES formalism for the sectional method is developed and used to investigate two different turbulent flames: a non-premixed jet flame and a confined pressurized swirled flame. Predicted temperature and soot volume fraction levels and topologies are compared to experimental data. Good predictions are obtained and the different soot processes in such flames are analyzed through the study of the PSD evolution. In these first simulations, wall heat losses rely on experimental measurements of walls temperature, and a coarse optically-thin radiation model. In the third part, to increase the accuracy of thermal radiation description, a Monte Carlo approach enabling to solve the Radiative Transfer Equation with detailed radiative properties of gaseous and soot phases is used and coupled to the LES solver. This coupled approach is applied for the simulation of the turbulent jet flame. Quality of radiative fluxes prediction in this flame is quantified and the nature of radiative transfers is studied. Then, a whole coupled modeling of turbulent combustion accounting for soot, conjugate heat transfer and thermal radiation is proposed by coupling three dedicated codes. This strategy is applied for a high-fidelity simulation of the confined pressurized burner. By comparing numerical results with experimental data, the proposed approach enables to predict both the wall temperature and the flame stabilization. The different simulations show that soot formation processes are impacted by the heat transfer description: a decrease of the soot volume fraction is observed with increasing heat losses. This highlights the requirement of accurate description of heat transfer for future developments of soot models and their validation.

Orientador: João Batista Campos Silva / Banca: Sérgio Said Mansur / Banca: Edson Luiz Zaparoli / Resumo: O objetivo principal deste trabalho é a simulação numérica de escoamentos viscosos, incompressíveis e transientes, com transferência de calor e massa; através do método de elementos finitos de subdomínio; usando a metodologia de simulação de grandes escalas para a modelagem da turbulência. Algumas aplicações de interesse são as simulações de escoamentos com transporte de um escalar, como nos casos de dispersão de poluentes induzida pelo movimento do ar atmosférico. O domínio é discretizado usando elementos finitos quadrilaterais de nove nós e as equações são integradas em volumes de controle ao redor dos nós dos elementos finitos. As equações governantes passam por um processo de filtragem, devido à metodologia aplicada, Simulação de Grandes Escalas (LES - Large-Eddy Simulation), e desta forma as maiores escalas são resolvidas diretamente através da solução das equações de Navier-Stokes filtradas, enquanto que as menores escalas ou escalas submalhas são modeladas, pelo modelo de viscosidade turbulenta de Smagorisnky. Alguns casos testes bidimensionais clássicos são resolvidos para validação do código e os resultados são apresentados e comparados com resultados disponíveis na literatura. Alguns poucos casos de dispersão de poluentes em geometrias que simulam cânions de ruas (urban street canyons) foram também simulados. / Abstract: The main purpose of this work is the numerical simulation of viscous, incompressible and unsteady fluid flows by a sub-domain finite element method, using the methodology of large-eddy simulation (LES) for turbulence modeling. Some applications of interest are isothermal and thermal flows with transport of scalar variable such as the pollutant dispersion in the atmosphere by airflow. The domain is discretized using nine-nodes quadrilateral finite elements and the equations are integrated into control volumes around the nodes of the finite elements. The government equations are submitted to a filtering process for application of LES methodology, in which the large scales are directly solved using the filtered Navier-Stokes equations, while the small or sub-grid scales are modeled by the eddy viscosity model of Smagorinsky. Two-dimensional benchmark problems are solved to validate the numerical code and the results are presented and compared with available results from the literature. Some cases of pollutant dispersion in geometries that simulate urban street canyons have been also simulated. / Mestre

Turbulent flow and transition are some of the most important phenomena of fluid mechanics and aerodynamics and represent a challenging engineering problem for aircraft manufacturers looking to improve aerodynamic efficiency. Laminar flow technology has the potential to provide a significant reduction to aircraft drag by manipulating the instabilities within the laminar boundary layer to achieve a delay in transition to turbulence. Currently prediction and simulation of laminar-turbulent transition is con- ducted using either a low-fidelity approach involving the stability equations or via a full Direct Numerical Simulation (DNS). The work in this thesis uses an alternative high-fidelity simulation method that aims to bridge the gap between the two simulation streams. The methodology uses an LES approach with a low-computational cost sub-grid scale model (WALE) that has inherent ability to reduce its turbulent viscosity contribution to zero in laminar regions. With careful grid spacing the laminar regions can be explicitly modelled as an unsteady Navier-Stokes simulation while the turbulent and transitional regions are simulated using LES. The methodology has been labelled as an unsteady Navier-Stokes/Large Eddy Simulation (UNS/LES) approach. Two test cases were developed to test the applicability of the method to simulate and control the crossflow instability. The first test case replicated the setup from an experiment that ran at a chord-based Reynolds number of 390, 000. Two methods were used to generate the initial disturbance for the crossflow vortices, firstly using a continuous suction hole and secondly an isolated roughness element. The results for this test case showed that the approach was capable of modelling the full transition process, from explicitly modelling the growth of the initial amplitude of the disturbances to final breakdown to turbulence. Results matched well with the available experimental data. The second test case replicated an experimental setup using a custom- designed aerofoil run at a chord-based Reynolds number of 2.4 million. The test case used Distributed Roughness Elements (DRE) to induce crossflow vortices at both a critical and a control wavelength. By forcing the crossflow vortices at a stable (control) wavelength a delay in laminar-turbulent transition can be achieved. The results showed that the UNS/LES approach was capable of capturing the initial disturbance amplitudes due to the roughness elements and their growth rates matched well with experimental data. Finally, downstream a transitional region was assessed with low-freestream turbulence provided using a modified Synthetic Eddy Method (SEM). The full laminar-turbulent transition pro- cess was simulated and results showed significant promise. In conclusion, the method employed in this thesis showed promising results and demonstrated a possible route to high-fidelity transition simulation run at more realistic flow conditions and geometries than DNS. Further work and validation is required to test the secondary instability region and the final breakdown to turbulence.

Les objectifs ambitieux, fixés aux acteurs du secteur automobile par les pouvoirs publics, en matière d'émission de polluants et de gaz à effet de serre rendent aujourd'hui indispensable une compréhension plus fine de la combustion dans les moteurs. La simulation 3D aux grandes échelles (LES) représente une voie prometteuse pour répondre à ces enjeux. Elle permet l'étude de phénomènes transitoires complexes inaccessibles avec des moyens expérimentaux ou des méthodes de calculs traditionnelles de type RANS. Ce travail de thèse est une première étape vers la simulation LES de l'injection de carburant liquide dans les moteurs à piston. Il a consisté à adapter le code de calcul aux particularités physiques de l'injection directe, technologie qui se généralise actuellement à tous les types de moteurs à piston. Dans un premier temps, et afin de s'affranchir du calcul 3D complexe en sortie d'injecteur, une méthodologie originale, consistant à initier le calcul en aval de l'injecteur, est proposée et validée sur différents cas. Pour la simulation 3D, l'approche Eulérienne mésoscopique, à laquelle est ajouté un modèle d'interaction particules-particules, est utilisée pour simuler le spray. Les simulations ont été premièrement validées par comparaison expérimentale dans des conditions proches de l'injection Diesel. De plus, une étude sur la dynamique du spray a permis de mieux comprendre son évolution et de dégager des points communs avec un jet de gaz turbulent. Des simulations complémentaires ont également montré la prédictivité de la LES sur des injections Diesel réalistes. Enfin, un premier calcul moteur à injection directe a été réalisé et a permis de valider les développements réalisés dans le cadre de cette thèse. / Car manufacturers are facing increasingly severe regulations on pollutant emissions and fuel consumption. To respect these regulations, a better understanding of combustion processes is needed. Large Eddy Simulation (LES) is becoming a promising tool for such issues as it allows the study of complex unsteady phenomena which can not be analysed with RANS simulations or experiments. The present work is a step towards the LES of liquid injection in piston engines. The numerical code has been adapted to the specifications of Direct Injection which is more and more used in industry. Firstly, in order to avoid the difficulties linked to the 3D simulation of cavitation, primary break-up and turbulence in the near-nozzle region, an original methodology, based on an injector model, has been proposed. The idea is to initiate the spray physics downstream to the injector exit. Then LES 3D simulations of spray have been conducted using the Eulerian Mesoscopic approach extended to dense dispersed sprays by the addition of a particle-particle interactions model. The simulation results have been validated by comparison with experimental data in Diesel conditions with a low injection pressure. Furthermore a study on the spray dynamics has permitted to better understand its development and to find similarities with a turbulent gaseous jet. Additional simulations on realistic Diesel injection conditions have shown the good predictivity of LES in such cases. Finally, a first simulation of a Direct Injection Engine has been been carried out to assess the developments achieved in this work.

Les turbines aéronautiques doivent satisfaire à des normes d’émissions polluantes toujours en baisse. La qualité du mélange du carburant et de l’air dans la chambre de combustion est responsable de la formation de polluants nocifs pour l’environnement. La simulation aux grandes échelles (LES) permet d’étudier les mécanismes de mélanges turbulents de l’air et du carburant. La prise en compte de l’aspect liquide du carburant injecté devient nécessaire pour prédire correctement l’apparition de vapeur de carburant au sein du foyer. Le but de cette thèse est évaluer la fiabilité des simulations LES Euler-Euler dans une configuration complexe. Les processus d’injection, et d’évaporation du carburant liquide sont analysés et modélisés dans les simulations LES car ils pilotent la formation de vapeur de carburant. Les méthodes numériques pour résoudre les équations continues de la phase dispersée doivent permettre des simulations précises et robustes dans une configuration représentative d’une chambre de combustion. Les simulations présentées dans ces travaux reproduisent l’écoulement diphasique évaporant non-réactif du banc d’essai Mercato. Ce banc est équipé d’un système d’injection d’air vrillé et d’un atomiseur pressurisé-swirlé de kérosène typiques des foyers aéronautiques réels. Dans ces travaux, le modèle pour l’injection de liquide FIM-UR a été développé pour définir les conditions limites conduisant à un spray issu d’un atomiseur préssurisé-swirlé. Le kérosène employé dans les campagnes expérimentales est modélisé dans les simulations par un composé permettant d’obtenir des temps d’évaporation réalistes. Trois stratégies numériques ont été mises en place sur la configuration Mercato. Les comparaisons des résultats numériques aux mesures expérimentales ont permis d’évaluer la stratégie numérique conduisant à la meilleure précision. L’utilisation du schéma centré TTGC associé à un opérateur de viscosité artificielle localisée par un senseur adapté est optimale lorsque l’équation sur l’énergie décorrélée des gouttes est résolue. Cette stratégie permet de contrôler la localisation et les niveaux de viscosité par rapport à un schéma décentré. Les termes sources liés au mouvement mésoscopique permettent de redistribuer l’énergie dans les zones de compression ou de détente de la phase dispersée, et d’obtenir les bonnes répartitions des fluctuations dans la chambre de combustion. La stratégie retenue est comparée aux statistiques de la dynamique du spray résolu par une approche Lagrangienne employant la même injection monodispersse. Le méthode Euler-Euler conduit à la même précision de la dynamique de la phase dispersée que la méthode Euler-Lagrange. L’accès à l’évolution instationnaire de l’écoulement permet d’identifier les mêmes mécanismes de dispersion et de mélange dans les deux simulations. Des différences sur la répartition de diamètre moyen et de carburant dans la chambre ont été mis en évidence et reliés à la polydispersion locale qui n’est pas résolue dans l’approche Euler-Euler monodisperse et qui apparaît naturellement dans l’approche Euler-Lagrange malgré l’injection monodisperse. / Aeronautical gas turbines are facing growing demands on emission reductions. Indeed, the quality of the air-fuel mixture directly triggers the formation of pollutants degrading the environment. Large Eddy Simulation is an accurate numerical method to predict turbulent mixing in combustors. Adding the liquid phase of the fuel in these simulations also becomes necessary to properly predict the injection process and the vaporization of the fuel in the combustion chamber. The purpose of this dissertation is to evaluate the accuracy and reliability of Euler-Euler LES in a complex combustor configuration. The injection and vaporization processes of the fuel liquid phase are both modeled in the present LES as they drive the formation of the fuel gas phase. Moreover, the numerical methods that solve the continuous equations of the disperse phase must be accurate and robust in realistic combustor configurations. The simulations shown in the present study reproduce the non-reactive two-phase flow of the ONERA Mercato test bench. The experimental set-up is equipped with an air-swirler injection system and a pressure-swirled atomizer typical of actual turboengine combustors. In the present work the FIM-UR liquid injection model has been developed. It creates boundary conditions profiles for a liquid spray produced by a pressure-swirled atomizer. Kerosene used in the experiments is modeled in the present numerical simulations by a single species leading to a good estimate of the vaporization rate. Three numerical strategies have been tested on the Mercato configuration. Comparisons between experimental and LES results help defining the most accurate numerical strategy. The use of the centered numerical scheme TTGC stabilized by a localized artificial viscosity operator is best when the random uncorrelated energy of droplets is also resolved. Unlike an upwind numerical scheme, the selected strategy allows the user to control where and how much artificial viscosity is added. The source terms coming from the mesoscopic movement redistribute the energy in the compression or expanding zones of the disperse phase, and provide the proper distribution of fluctuations in the combustion chamber. The obtained strategy is compared with the statistics provided by a Lagrangian description of the liquid spray in the same mono-disperse injection. The Euler-Euler approach leads to the same accuracy in the same spray dynamics of the disperse phase as in the Euler-Lagrange method. Both unsteady flow simulations also provide the same dispersion and mixing processes in the Mercato set-up. Differences on the mean diameter and the fuel distribution in the combustion chamber are seen and related to the local poly-dispersion that cannot be resolved in the mono-disperse Euler-Euler approach and that naturally appear in the Euler-Lagrange method despite the mono-disperse injection.

Investigating flow and turbulence structure around a barrier mounted on the ground or placed in its vicinity is a fundamental problem in wind engineering because of many practical applications related to protection against adverse effects induced by major wind storms (e.g., hurricanes) and snow events (e.g., snow fences used to reduce adverse effects of snow drifting on the roads). In this work the focus is on the case when the obstacle/barrier is porous and the shape of the obstacle is close to a high-aspect-ratio rectangular cylinder situated in the vicinity of the ground. The study employs a range of numerical and experimental techniques to achieve this goal that include 3D LES and 2D RANS numerical simulations, and RTK survey and 3D photogrammetry techniques to measure ground elevations and snow deposits in the field. In the first part of the study, high-resolution large eddy simulations are used to understand the fundamental flow physics of flow past 2D solid and porous vertical plates with a special focus on describing the unsteady wind loads on the obstacle, vortical structure of the turbulent wake, spectral content of the wake, the separated shear layers and of the characteristics of the large-scale vortex shedding behind the plate, if present. Results show that LES can accurately predict mean flow and turbulence statistics around solid/porous cylinders. Then, a detailed parametric study of flow past vertical solid and porous plates situated in the vicinity of a horizontal bed is performed for the purpose of understanding changes in the mean flow structure, turbulence statistics and dynamics of large scale coherent structures as a function of the main nondimensional geometrical parameters (bottom gap for solid and porous plates, and porosity and average hole size of porous plates) and flow variables (e.g., bed roughness) that affect the wake flow. In particular, the LES flow fields allowed clarifying how the interactions between the bottom and the top separated shear layers change with increasing bottom gap and what is the effect of the bleeding flow on the interactions between the separated shear layers that determine the coherence of the large-scale eddies at large distances from the wake. In the second part of the thesis, a novel methodology based on field monitoring of the snow deposits and RANS numerical simulations is proposed to improve the design of snow fences and in particular the design of lightweight plastic snow fences that are commonly used to protect roads in the US Midwest against the snow drifting. The goal of the design optimization procedure is to propose a snow fence design that can retain a considerable amount of snow within a shorter downwind distance compared to fences of standard design. A major contribution of the present thesis was the development of a novel non-intrusive image-based technique that can be used to quantitatively estimate the temporal evolution of the volume of snow trapped by a fence over long periods of time. This technique is based on 3-D close range photogrammetry. Results showed that this technique can produce estimations of the snow deposits of comparable accuracy to that given by commonly used methods. This is the first application of this type of techniques to measurements of the snow deposits.

Nous cherchons à modéliser numériquement un jet flottant air-hélium dans une cavité avec deux ouvertures à partir de simulations aux grandes échelles (LES) et de simulations numériques directes (DNS). La configuration considérée est basée sur une étude expérimentale menée au CEA de Saclay reproduisant une fuite d'hydrogène en environnement confiné. La dimension de la cavité a été choisie pour permettre une transition laminaire-turbulent intervenant environ à la mi-hauteur de la cavité. Cette étude porte principalement sur trois points majeurs : l'influence des conditions aux limites sur le développement du jet et son interaction avec l'environnement extérieur, la validité du modèle numérique qui est analysée en comparant la distribution de vitesse obtenue numériquement aux mesures expérimentales (PIV) et, enfin, la compréhension de la distribution air-hélium et le phénomène de stratification qui s'établit à l'intérieur de la cavité. Nous observons dans un premier temps que des conditions limites de pression constante appliquées directement au ras des évents conduisent à une sous-estimation du débit volumique d'air entrant dans la cavité et donc à une surestimation de la masse de l'hélium à l'intérieur de la cavité, ce qui n'est pas acceptable dans un contexte d'évaluation du risque hydrogène. En revanche, la prise en compte, dans le domaine de calcul, d'une région extérieure à la cavité prédit correctement le flux d'air entrant. Les résultats numériques sont alors en bon accord avec les données PIV. Il a été montré que les prédictions de la DNS, par rapport à la LES, concordent mieux avec les mesures de vitesse par PIV. Le champ de concentration prédit numériquement présente une couche homogène en haut de la cavité, dont la concentration est en accord avec le modèle théorique de Linden et al. 1990. Cependant, sa position et son épaisseur ne correspondent pas au modèle. Ceci est principalement dû aux interactions directes entre le jet flottant et, d'une part, avec les limites solides de la cavité et d'autre part, avec l'environnement extérieur. L'analyse statistique concernant la production de la flottabilité de l'énergie cinétique turbulente (TKE) a permis d'identifier les limites du jet flottant. / We present numerical results from large eddy simulations (LES) and coarse direct numerical simulations (DNS) of an air-helium buoyant jet rising in a two vented cavity. The geometrical configuration mimics the helium release experimental set-up studied at CEA Saclay in the framework of security assessment of hydrogen-based systems with an indoor usage. The dimension of the enclosure was chosen to ensure a laminar-turbulent transition occurring at about the middle height of the cavity. This study focuses mainly on three key points : the influence of the boundary conditions on the jet development and its interaction with the exterior environment, the validity of the numerical model which is analyzed by comparing the numerical velocity distribution versus the measured particle image velocimetry (PIV) ones, and finally understanding the distribution of air-helium and the stratification phenomenon that takes place inside the cavity. We observe at first that applying constant pressure outlet boundary conditions directly at the vent surfaces underestimates the volumetric flow rate of air entering the enclosure and thus overestimate the helium mass inside the cavity. On the contrary, modelling an exterior region in the computational domain better predicts the air flow-rate entrance and numerical results matches better with the experimental PIV data. It has been figured out that the coarse DNS predictions match better with the velocities PIV measurements, compared to the LES. Numerical prediction of the helium field depicts a homogeneous layer formed at the top of the cavity, with a concentration in good agreement with the theoretical model of Linden et al. 1990. However, the position and the thickness of the layer do not correspond to the theory. This is mainly due to the direct interactions between the buoyant jet and both the solid boundaries of the cavity and the exterior environment. Statistical analysis regarding the buoyancy production of the turbulent kinetic energy (TKE) served to identify the limits of the buoyant jet.