Os processos de conversão de energia tendem a considerar cada vez mais restrições econômicas e ambientais, tornando-se necessário o entendimento da interação entre combustão e turbulência. Este trabalho tem como objetivo o desenvolvimento de soluções analíticas para a fração de mistura de uma chama difusiva, sob forma de um jato turbulento axissimétrico. Foi desenvolvida, também, uma metodologia analíticonumérica para a determinação das frações mássicas dos componentes, considerando uma reação de combustão de dois passos. Os resultados foram comparados com dados experimentais encontrados na literatura para uma chama de hidrogênio H2=N2 (50/50% em volume). De modo geral, os resultados obtidos foram satisfatórios frente aos dados experimentais, sendo a principal limitação o fato das expressões analíticas obtidas não serem capazes de representar o jato próximo à saída do bocal, sendo válidas a partir de, aproximadamente, x=d > 10, onde x é a coordenada ao longo do comprimento do jato. A principal vantagem do método empregado neste trabalho é a diminuição da complexidade do sistema de equações a ser resolvido numericamente. / The energy conversion processes tend to consider even more economical and environmental constraints, making it necessary to understand the interaction between combustion and turbulence. This study aims at the development of analytical solutions for the mixture fraction of a diffusive flame in the form of an axisymmetric turbulent jet. It was also considered an analytical-numerical approach for the determination of the mass fractions of the compounds, for a two-step reaction. The results were compared with data found in literature for a hydrogen flame H2=N2 (50/50 % by volume). Overall, the results were satisfactory when compared with the experimental data, however the principal limitation was the fact that the analytical expressions were not able to represent the jet near the nozzle exit, being the solution valid from, approximately, x=d > 10, where x is the coordinate along the jet length. The main advantage of the method employed in this work is the decrease in the complexity of the equations system to be solved numerically.

Neste trabalho apresenta-se a modelagem de chamas difusivas na forma de jato, para baixo número de Mach e elevado número de Damköhler. O modelo é baseado na solução das equações na forma flamelet para a parte química e na fração de mistura para o fluxo. Este modelo descreve bem o comportamento de chamas difusivas, exceto na sua extremidade (ponta), onde geralmente surgem instabilidades. Resultados numéricos são apresentados para uma cinética química de uma e multietapas, utilizando a técnica LES (Large-Eddy Simulation) com o modelo de Smagorinsky para a viscosidade turbulenta. A discretização das equações governantes é feita em diferenças finitas, com a aplicação da técnica TVD (Total Variation Diminishing). Além disso, apresentamse mecanismos reduzidos multietapas para o metano, o metanol e o etanol, visando obter resultados realistas. A modelagem de chamas de metanol e etanol diferencia-se da modelagem de chamas de metano por ocorrer uma mudança de fase antes da combustão. Modela-se o efeito global das gotas usando uma descrição Lagrangeana que é incorporada à descrição Euleriana do escoamento, via termos fonte. Testes numéricos foram realizados para chamas difusivas de metano, metanol e etanol, e os resultados estão em concordância com os dados encontrados na literatura. / This work presents a model for a jet diffusion flame, for low Mach and high Damköhler numbers. The model is based on the solution of the flamelet equations for the chemistry and on the mixture fraction for the flow. This model describes well the behavior of diffusion flames, except at the flame tip, where instabilities can often occur. Numerical results are presented for an one-step and multi-step chemical kinetic models, using the LES (Large-Eddy Simulation) technique with the Smagorinsky model for the turbulent viscosity. The discretization of the governing equations follows the finite difference method, with the application of the TVD (Total Variation Diminishing) technique. Besides, multi-step reduced mechanisms for the methane, the methanol and the ethanol are employed, obtaining realistic results. The flame modeling of methanol and ethanol differs from the modeling of methane flames because of a phase change occurs before the combustion. The droplets global effect is modeled based on a Lagrangian description, which is incorporated into the Eulerian description of the flow through source terms. Numerical tests were carried out for methane, methanol and ethanol diffusion flames, and the results compare well with data in the literature.

Les fluctuations du taux de dégagement de chaleur sont souvent responsables d’intensification des flux thermiques aux parois, de vibrations et d’émissions sonores qui peuvent éventuellement dégénérer en instabilités thermo-acoustiques auto-entretenues. Ces phénomènes instationnaires dégradent les performances des foyers, provoquent un vieillissement prématuré de certains éléments de la chambre de combustion, voire des dégâts plus importants sur l’installation. Ces perturbations sont cependant difficiles à mesurer dans les foyers car il n’existe pas de diagnostic qui permette d'accéder directement au taux de dégagement de chaleur. L’objectif de ce travail est d'explorer deux alternatives aux solutions existantes pour accéder aux fluctuations du taux de dégagement de chaleur avec une bonne résolution temporelle. Ces nouvelles méthodes sont testées dans des configurations génériques parfaitement prémélangées pour des écoulements laminaires. La première méthode est une technique acoustique, qui repose sur la détermination du temps de vol d’ondes ultrasonores qui traversent l’écoulement. Les fluctuations du temps de vol de ces ondes sont utilisées pour détecter des perturbations de la largeur des gaz brûlés le long du chemin acoustique. Cette information permet de reconstituer les fluctuations du taux de dégagement de chaleur dans des flammes prémélangées. Les premières validations de cette méthode sont présentées pour des flammes en l'absence de perturbation externe lorsqu'elles présentent une instabilité de type Kelvin-Helmholtz pilotée par les phénomènes de flottabilité du panache des gaz brûlés. Des mesures sont ensuite conduites pour des flammes soumises à des modulations harmoniques de l'écoulement. Les données obtenues dans ces configurations sont comparées à des mesures optiques ainsi qu'à des prévisions analytiques. La seconde méthode est une technique optique utilisant un système d’interférométrie laser Doppler permettant de déterminer les fluctuations de densité intégrées le long du chemin optique. On montre dans un premier temps que les perturbations de densité sont principalement causées par des fluctuations du taux de dégagement de chaleur lorsque les flammes sont confinées. Un lien est établi pour reconstituer les perturbations du taux de dégagement de chaleur exploitant le signal de l'interféromètre. La technique est validée pour des flammes pulsées pour différentes richesses et débits. Les données obtenues sont comparées à des mesures reposant sur la chimiluminescence de la flamme. Un bon accord est obtenu pour des modulations harmoniques de l'écoulement à différentes fréquences et niveaux de perturbation. Ce travail permet de valider le principe de ces deux techniques pour détecter les perturbations du taux de dégagement de chaleur lorsque l'accès optique à la zone de combustion est réduit et lorsque des informations quantitatives résolues temporellement sont nécessaires. / Heat release rate disturbances are the sources of additional thermal stresses, direct and indirect combustion noise and undesirable vibrations. In extreme cases, these perturbations may even cause destructive combustion instabilities. These quantities are difficult to measure in practical burners. The objective of this work is to develop two alternative diagnostics to measure heat release rate fluctuations in unsteady flames. These techniques are validated in generic configurations for perfectly premixed laminar flames. The first method is an acoustic technique, which is based on the measurement of the travel time of ultrasonic waves through the flames. Fluctuations of the sound propagation time transmission through unsteady flames are used to estimate perturbations in the burned gases width along the acoustic path. This information is then used to reconstruct heat release rate fluctuations. This technique is validated in the cases of unstable laminar premixed flames driven by buoyancy forces and for flames submitted to harmonic flow velocity modulations. Analytical expressions are derived linking fluctuations in heat release rate and disturbances of the sound travel time. Measurements made with this acoustic technique are compared with optical detections based on the flame chemiluminescence and with predictions from an analytical model. Good agreements are obtained between these different methods validating the proposed technique. The second method envisaged is an optical technique based on a Laser Interferometric Vibrometer used to measure integrated density perturbations along the optical path of a laser beam. It is shown that density disturbances along this path result mainly from heat release rate fluctuations when the flames are confined. A link is established to reconstruct heat release rate disturbances from the signal of the interferometer. The technique is validated in the case of pulsated laminar premixed flames. Measurements are compared to line-of-sight integrated chemiluminescence emission measurements. A good agreement is obtained for harmonic flow modulations at different forcing frequencies and perturbation levels for flames operating at different flow conditions. This work validates the principle of this alternative technique for detecting heat release rate perturbations.

Os processos de conversão de energia tendem a considerar cada vez mais restrições econômicas e ambientais, tornando-se necessário o entendimento da interação entre combustão e turbulência. Este trabalho tem como objetivo o desenvolvimento de soluções analíticas para a fração de mistura de uma chama difusiva, sob forma de um jato turbulento axissimétrico. Foi desenvolvida, também, uma metodologia analíticonumérica para a determinação das frações mássicas dos componentes, considerando uma reação de combustão de dois passos. Os resultados foram comparados com dados experimentais encontrados na literatura para uma chama de hidrogênio H2=N2 (50/50% em volume). De modo geral, os resultados obtidos foram satisfatórios frente aos dados experimentais, sendo a principal limitação o fato das expressões analíticas obtidas não serem capazes de representar o jato próximo à saída do bocal, sendo válidas a partir de, aproximadamente, x=d > 10, onde x é a coordenada ao longo do comprimento do jato. A principal vantagem do método empregado neste trabalho é a diminuição da complexidade do sistema de equações a ser resolvido numericamente. / The energy conversion processes tend to consider even more economical and environmental constraints, making it necessary to understand the interaction between combustion and turbulence. This study aims at the development of analytical solutions for the mixture fraction of a diffusive flame in the form of an axisymmetric turbulent jet. It was also considered an analytical-numerical approach for the determination of the mass fractions of the compounds, for a two-step reaction. The results were compared with data found in literature for a hydrogen flame H2=N2 (50/50 % by volume). Overall, the results were satisfactory when compared with the experimental data, however the principal limitation was the fact that the analytical expressions were not able to represent the jet near the nozzle exit, being the solution valid from, approximately, x=d > 10, where x is the coordinate along the jet length. The main advantage of the method employed in this work is the decrease in the complexity of the equations system to be solved numerically.

Les suies se présentent sous la forme de fines particules carbonées de diamètres compris entre quelques dizaines de nanomètres à quelques micromètres. Dans l’atmosphère, elles entraînent des enjeux climatiques, de par leurs propriétés radiatives, mais aussi des enjeux sanitaires, du fait de leur faible taille : elles pénètrent facilement dans le système respiratoire et même, pour les plus fines, dans le système sanguin. L’objectif est de parfaire les connaissances sur les propriétés physiques des suies produites par différents systèmes de combustion. C’est dans le but de mieux comprendre l’influence des systèmes de combustion, faisant intervenir des temps de séjours différents, des propriétés de turbulence, d’oxydation et de pression distinctes que nous avons choisi d’étudier trois types de combustion spécifiques : d’une part, des flammes de diffusion laminaires à pression atmosphérique, initiées dans un brûleur développé au cours de ces travaux ; d’autre part, une flamme de diffusion laminaire sous atmosphère pressurisée (3 à 5 bars) ; enfin, une flamme turbulente produite par une chambre tubulaire, elle aussi sous atmosphère pressurisée (1.2 à 3 bar). Un autre enjeu de ce travail était d’approfondir les informations relatives à la combustion de carburants liquides, à savoir le kérosène et le diester. Les travaux effectués visent à déterminer les caractéristiques morphologiques (dimension fractale, diamètre des monomères...) et l’indice complexe m* des suies issues des différents systèmes de combustion. La technique employée pour la mesure de l’indice complexe de réfraction des suies, repose sur l’analyse d’une partie des fumées produites par les flammes. Ces fumées sont acheminées dans un banc d’analyse permettant la mesure de signaux d’extinction et de diffusion, ainsi que de distributions de taille des suies. Par ailleurs, des analyses de clichés obtenus par microscope en transmission d’électrons (TEM) permettent l’obtention d’informations sur la morphologie des agrégats de suies. L’utilisation de la théorie de la diffusion de la lumière pour des agrégats fractals dans la limite de Rayleigh (RDG-FA) permet d’estimer à partir de ces données deux fonctions de l’indice complexe E(m) et F(m), et ainsi de retrouver m*. / Soot are carbonaceous fine particles, which diameters are ranged from a few nanometres to a few micrometers. They have an impact on climate, due to their radiative properties, as well as on health, due to their small size. That’s why particulate matter is an important concern. In order to gain a better understanding of the influence of the combustion devices, which implies specific residence time and also specific turbulence, oxidation and pressure properties, we studied three specific kinds of combustion : first, laminar diffusion flames at atmospheric pressure ; then, a laminar diffusion flame a high pressures (3 to 5 bar) ; finally, a turbulent flame produced in a combustor at high pressures (1,2 to 3 bar). Another objective of this work was to improve the knowledge about soot produced by the combustion of liquid fuels, namely kerosene and biofuel. We studied morphological properties (fractal dimension, primary particle size…) and the refractive index m* of soot produced by these combustion systems. The technique employed to characterize the soot refractive index is based on the analysis of a part of smokes produced by flames. These are transported towards two optical cells, so that extinction and scattering coefficients can be measured, in addition to soot size distributions. Furthermore, a morphological characterization of the aggregates is conducted, using transmission electron microscopy (TEM) photographs. Rayleigh-Debye-Gans theory for fractal aggregates is used to determine two functions of the refractive index E(m) and F(m), so that m* can be deduced.

Microgravité éthylène laminaires couche limite flammes de diffusion générés par un brûleur poreux plat et caractérisées par les vitesses d'injection de carburant de 3 et 4 mm / s et une vitesse d'oxydation de 250 mm / s ont été simulées en utilisant un modèle de rayonnement précis, un mécanisme cinétique complète et un modèle de suie constitué de lancement par suite de la collision de deux molécules de pyrène, l'évolution de la surface hétérogène et oxydation suivant l'abstraction d'un atome d'hydrogène addition d'acétylène (HACA) mécanisme, la coagulation à particules de suie, et la condensation de la surface de l'HTAP. La distance d'écartement et la production de suie sont améliorées lorsque la vitesse du carburant augmente. H et des radicaux OH, responsables de la de-hydrogénation des sites dans le processus HACA, et le pyrène, de l'espèce pour la création de la suie et des processus de condensation HAP, se trouvent à être situé dans une région qui suit la distance stand-off. La suie est ensuite produite dans cette région et est transporté à l'intérieur de la couche limite par convection et thermophorèse. Perte radiatif est sensiblement plus élevé dans ces flammes que dans flammes de diffusion de gravité normales dues à beaucoup plus longues durées de séjour. Calculs effectués par négliger le rayonnement de la suie et en utilisant l'approximation optiquement mince (OTA) a révélé que la suie domine le transfert de chaleur par rayonnement dans ces flammes et que l'utilisation de l'OTA donne lieu à des écarts significatifs dans la fraction du volume température et la suie. / Microgravity ethylene laminar boundary layer diffusion flames generated by a flat porous burner and characterized by the fuel injection velocities of 3 and 4 mm/s and an oxidizer velocity of 250 mm/s have been simulated by using an accurate radiation model, a comprehensive kinetic mechanism, and a soot model consisting of inception as a result of the collision of two pyrene molecules, heterogeneous surface growth and oxidation following the hydrogen abstraction acetylene addition (HACA) mechanism, soot particle coagulation, and PAH surface condensation. Model predictions are in reasonable agreement with the experimental data in terms of the stand-off distance and soot volume fraction. The stand-off distance and soot production are enhanced as the fuel velocity increases. H and OH radicals, responsible of the de-hydrogenation of sites in the HACA process, and pyrene, of the species for soot inception and PAH condensation processes, are found to be located in a region that follows the stand-off distance. Soot is then produced in this region and is transported inside the boundary layer by convection and thermophoresis. Radiative loss is substantially higher in these flames than in normal gravity diffusion flames owing to much longer residence times. Calculations carried out by neglecting soot radiation and by using the optically-thin approximation (OTA) revealed that soot dominates the radiative heat transfer in these flames and that the use of OTA gives rise to significant discrepancies in temperature and soot volume fraction.

Large-scale unsteady flow structures play an influential role in the dynamics of many practical flows, such as those found in gas turbine combustion chambers. This thesis is concerned primarily with large-scale unsteady structures that arise due to self-sustained hydrodynamic oscillations, also known as global hydrodynamic instability. Direct numerical simulation (DNS) of the Navier--Stokes equations in the low Mach number limit is used to obtain a steady base flow, and the most unstable direct and adjoint global modes. These are combined, using a structural sensitivity framework, to identify the region of the flow and the feedback mechanisms that are responsible for causing the global instability. Using a Lagrangian framework, the direct and adjoint global modes are also used to identify the regions of the flow where steady and unsteady control, such as a drag force or heat input, can suppress or promote the global instability. These tools are used to study a variety of reacting and non-reacting flows to build an understanding of the physical mechanisms that are responsible for global hydrodynamic instability in swirling diffusion flames. In a non-swirling lifted jet diffusion flame, two modes of global instability are found. The first mode is a high-frequency mode caused by the instability of the low-density jet shear layer in the premixing zone. The second mode is a low-frequency mode caused by an instability of the outer shear layer of the flame. Two types of swirling diffusion flames with vortex breakdown bubbles are considered. They show qualitatively similar behaviour to the lifted jet diffusion flames. The first type of flame is unstable to a low-frequency mode, with wavemaker located at the flame base. The second type of flame is unstable to a high-frequency mode, with wavemaker located at the upstream edge of the vortex breakdown bubble. Feedback from density perturbations is found to have a strong influence on the unstable modes in the reacting flows. The wavemaker of the high-frequency mode in the reacting flows is very similar to its non-reacting counterpart. The low-frequency mode, however, is only observed in the reacting flows. The presence of reaction increases the influence of changes in the base flow mixture fraction profiles on the eigenmode. This increased influence acts through the heat release term. These results emphasize the possibility that non-reacting simulations and experiments may not always capture the important instability mechanisms of reacting flows, and highlight the importance of including heat release terms in stability analyses of reacting flows.

Orientador: João Andrade de Carvalho / Resumo: Application technologies based on the detonation cycle has proven a significant impact on the overall efficiency. However, detonation engines are not currently available on the markets due to the lack of physical and chemical knowledge of the detonation phenomenon. The present study aims to provide new insights by studying the pressure and velocity, the density gradient of the detonation wave, and the OH distribution on the reaction zone of hydrogen-air detonation. Three strategies were proposed to obtain repeatable detonation events. The strategies vary on the geometry of the obstacle and the amount of spark plug to ignite the mixture. Pressure and velocity were recorded to determine if the transition from deflagration to detonation is successful. To image the density gradient of the shock wave, the optical technique called Schlieren was adapted to the detonation test bench. The OH radical distribution was studied by the optical diagnostic technique called planar laser-induced fluorescence. The pressure trace results showed high peaks in the regimen of Chapman-Jouguet state for detonation, unlike fast flames. The velocity results showed a considerable influence of the obstacle geometry to enhance the velocity of the wave, although the repeatable detonation events and the steadiness of the velocity were not boosted. The third strategy proved that adding more energy to a transient detonation wave, enhanced the stability and the consistent production of detonation events. The S... (Resumo completo, clicar acesso eletrônico abaixo) / Mestre

Les suies de combustion sont principalement connues pour leur caractère nocif, dans le cas des feux de forêt, de fumées de cheminées ou d'émissions polluantes d'un tuyau d'échappement. Cependant, le noir de carbone, un produit industriel de combustion d'hydrocarbures largement utilisé dans notre vie quotidienne. La surface d'une particule de suies ou de noir de carbon joue un rôle important tant au niveau de son utilisation que de son effet nocif. Il est donc important de connaître la masse, le volume ainsi que la morphologie des suies. En particulier, la surface des particules est un paramètre important pour prédire leur utilisation ainsi que leur effet nocif. Les suies sont généralement des agrégats présentant une structure fractale constituée d'éléments de forme sphérique, appelés particules primaires. Il est possible de connaître la surface des agrégats à partir de la distribution en taille de particules primaires (PPSD-Primary particules size distribution). Compte tenu de l'intérêt grandissant pour la surface des particules et leurs évolutions, il est aujourd’hui nécessaire d'étendre les modèles numériques pour la prévision de la PPSD. De plus, comme la taille des la particules primaires influence les processus chimiques et les processus de collision, la prise en compte de ce paramètre peut améliorer les prévisions des modèles. Les flammes multidimensionnelles laminaires, comme les flammes de diffusion, sont moins complexes que les flammes rencontrées dans les systèmes de combustion industriels. Cependant, les processus de formation de suies sont analogues dans les deux cas, ce qui rend l'étude de ces flammes intéressante. Afin d'obtenir une description détaillée des processus chimiques ayant lieu dans ces flammes tout en maintenant le coût de calcul à un niveau abordable, l'utilisation de modèles sectionnels discrets chimiques (CDSchemical discret sectional methods) est un choix approprié. Le développement de modèles CDS est au coeur de cette thèse. D'abord, une stratégie numérique pour déterminer la taille des particules primaires est présentée dans le contexte des modéles CDS. Elle repose sur la résolution d'une équation de transport pour la densité en nombre de particules primaires pour chaque section d'agrégats considérée. Pour valider la taille des particules primaires déterminée numériquement, les résultats doivent être comparés avec des données expérimentales obtenues via la technique d'Incandescence Induite par Laser résolue temporellement (TiRe-LII). Cette comparaison, dite inverse, est affectée par les incertitudes expérimentales et les hypothèses sous-jacentes au post-traitement du signal TiRe-LII pour obtenir la PSD. Pour améliorer la stratégie de validation, une nouvelle approche, dite directe, est proposée pour la validation de la PPSD à partir des données obtenues par TiRe-LII. Elle est basée sur la reconstruction numérique de l'évolution temporelle du signal d'incandescence à partir des résultats numériques et de sa comparaison avec le signal mesuré. L'efficacité de l'approche proposée est démontrée a priori en évaluant l'erreur potentiellement évitée par la nouvelle stratégie. Le modèle proposé pour le suivi des particules primaires est ensuite validé en utilisant à la fois les approches ’directe’ et ’inverse’ sur les flammes cibles issues de l'International Sooting FlameWorkshop (ISF): une flamme pré-mélangée éthylèneair et une flamme de diffusion coflow avec deux dilutions différentes. Le caractère général du modèle est discuté en effectuant une étude de sensibilité des résultats aux paramètres du modèle même. Enfin, le modèle est utilisé pour comprendre l'effet de la dilution du combustible sur la taille des particules primaires dans les flammes de diffusion en examinant les corrélations possibles entre phase gazeuse et phase solide ainsi que l'évolution temporelle des particules le long de leur trajectoires. / An image appearing when the phrase soot is heard is the smoke emitted by an exhaust pipe. The imperfect combustion of hydrocarbon fuels is a source of this harmful pollutant. The industrially controlled combustion of hydrocarbons can provide the carbon black, an industrial product widely used in our everyday life. For both its utilization and its harming effect, the surface of these combustion generated particles plays an important role, therefore, it is of interest to possess information on the particle morphology beside its mass or volume. Soot particles were found, at various conditions, to have a fractal-like structure built up from spherical shape building blocks, socalled primary particles. This increased interest in the particle surface and its evolution gives the motivation to extend numerical models to provide related information, i.e. particle surface or primary particle size. Furthermore, as the primary particle size influences the chemical and collisional processes, accounting for this parameter can improve the model predictions. The requirements for numerical models are various depending on the purpose of the simulation. Multidimensional laminar flames, like a laminar coflow diffusion flame, are less complex than flames of industrial combustion systems. However, the soot formation processes are analogous in the two cases, therefore, the investigation of these flames are of interest. In order to obtain a detailed description of the chemical processes, while keeping the computational cost in these flames at an affordable level, using chemical discrete sectional models is a suitable choice. As in their current version, these models do not provide information on the primary particle size their development in this direction is of interest. Guided by the above motivation, a numerical strategy to determine the primary particle size is presented in the context of the chemical sectional models. The proposed strategy is based on solving the transport equation of the primary particle number density for each considered aggregate section. In order to validate numerical primary particle size, the comparison to experimental data is required. Due to its numerous advantages, the Time-Resolved Laser-Induced Incandescence (TiRe-LII) technique is a nowadays popular experimental method. However, the comparison of the numerically and the experimentally obtained primary particle size may be charged with uncertainties introduced by the additional measurements or assumptions of the numerous parameters required to derive primary particle size from the detected signal. In order to improve the validation strategy, an additional approach for primary particle size distribution validation with TiRe-LII is proposed. This is based on the reconstruction of the temporal evolution of incandescence from the numerical results and its comparison with the measured signal. The effectiveness of this ’forward’ method is demonstrated a priori by quantifying the errors potentially avoided by the new strategy. The validity of the proposed primary particle tracking model is tested by both the traditional ’inverse’ and the ’forward’ method on target flames of the International Sooting Flame (ISF) Workshop. In particular a laminar premixed ethylene flame is considered first. Then, two laminar coflow ethylene flames with different dilutions are put under the scope. The sensitivity to the model parameters, such as accounting for the surface rounding and the choice of smallest aggregating particle size, is explored in both the premixed flame and in the coflow flame with highest ethylene content. To understand the effect of the fuel stream dilution on the primary particle size in the coflow flame, first, the flame-flow interaction and the effect of the dilution on the flame structure is investigated. [...]

Les systèmes de combustion à prémélange pauvre (PP) sont l’une des technologies les mieux adaptées pour la réduction des émissions de polluants, mais ils sont très sensibles aux phénomènes d’extinction, aux retours de flamme (flashback) dans l’injecteur et aux instabilités de combustion. La plupart des chambres de combustion des turbines à gaz utilisent de swirleurs pour stabiliser des flammes compactes et permettre une combustion efficace et propre avec des densités de puissance élevée. Une meilleure connaissance des mécanismes de la dynamique de la combustion d’écoulements swirlés PP présente un intérêt aussi bien pratique que fondamental. Ce travail est une contribution pour atteindre ce but. Le brûleur Noisedyn, avec une geometrie modifiable, a été spécialement conçu pour répondre à cet objectif. Une analyse expérimentale a etait conduite pour examiner les paramètres qui reduisent la sensibilité des systèmes PP aux phénomènes dynamiques. Mesures de fonction de transfert de flamme (FTF), diagnostiques laser (LDV et PIV) et imagerie des flammes sont les principaux techniques utilisé dans ce travail. Large eddy simulation sont aussi utilisé pour expliquer les mécanismes derrière les observations experimentaux. / Lean premixed (LPM) combustion systems achieve low pollutant emission levels, with compact flames and high power densities, but are highly sensitive to dynamic phenomena, e.g, flashback, blowout and thermoacoustic instabilities, that hinder their practical application. Most LPM gas turbine combustors use swirling flows to stabilize compact flames for efficient and clean combustion. A better knowledge of the mechanisms of steady and unsteady combustion of lean premixed swirled mixtures is then of practical, as well as fundamental interest. This thesis is a contribute towards the achievement of this goal. A burner, made of several components with variable geometry, was specifically designed for this scope. An experimental analysis was conducted to investigate the main parameters leading to a reduction of the sensitivity of LPM systems to dynamic phenomena. The diagnostics applied include flame transfer function (FTF) measurements, laser diagnostics (LDV and PIV) and flame imaging. Large eddy simulations were also exploited to elucidate the mechanisms behind the experimental observations.