No description available.

The present thesis deals with the fundamental aspects of turbulent mixing and non-premixed combustion in the wall-jet flow, which has a close resemblance to many industrial applications. Direct numerical simulations (DNS) of turbulent wall-jets with isothermal and exothermic reactions are performed. In the computational domain, fuel and oxidizer enter separately in a nonpremixed manner and the flow is compressible, fully turbulent and subsonic. The triple “turbulence-chemistry-wall” interactions in the wall-jet flow have been addressed first by focusing on turbulent flow effects on the isothermal reaction, and then, by concentrating on heat-release effects on both turbulence and flame characteristics in the exothermic reaction. In the former, the mixing characteristics of the flow, the key statistics for combustion and the near-wall effects in the absence of thermal effects are isolated and studied. In the latter, the main target was to identify the heat-release effects on the different mixing scales of turbulence. Key statistics such as the scalar dissipation rates, time scale ratios, two-point correlations, one and two-dimensional premultiplied spectra are used to illustrate the heat release induced modifications. Finer small mixing scales were observed in the isothermal simulations and larger vortical structures formed after adding significant amounts of heat-release. A deeper insight into the heat release effects on three-dimensional mixing and reaction characteristics of the turbulent wall-jet flow has been gained by digging in different scales of DNS datasets. In particular, attention has been paid to the anisotropy levels and intermittency of the flow by investigating the probability density functions, higher order moments of velocities and reacting scalars and anisotropy invariant maps for different reacting cases. To evaluate and isolate the Damkohler number effects on the reaction zone structure from those of the heat release a comparison between two DNS cases with different Damkohler numbers but a comparable temperature rise is performed. Furthermore, the wall effects on the flame and flow characteristics, for instance, the wall heat transfer; the near-wall combustion effects on the skin-friction, the isothermal wall cooling effects on the average burning rates and the possibility of formation of the premixed mode within the non-premixed flame are addressed. The DNS datasets are also used for a priori  analysis, focused on the heat release effects on the subgrid-scale (SGS) statistics. The findings regarding the turbulence small-scale characteristics, gained through the statistical analysis of the flow have many phenomenological parallels with those concerning the SGS statistics. Finally, a DNS of turbulent reacting wall-jet at a substantially higher Reynolds number is performed in order to extend the applicability range for the conclusions of the present study and figuring out the possible differences. / <p>QC 20150225</p>

Dans le but de prédire l’évolution d’un hypothétique accident au sein d’un réacteur nucléaire, nous nous proposons au travers de cette étude de comprendre la dynamique de bulles de gaz évoluant dans un bain stratifié constitué de deux liquides superposés. Pour ce faire, un dispositif expérimental muni de caméras à haute cadence a été construit afin d’observer en détail la dynamique de bulles d’air isolées et de trains de bulles traversant une interface séparant deux liquides newtoniens immiscibles initialement au repos. En faisant varier la taille des bulles injectées ainsi que les contrastes de viscosité entre les liquides d’un et quatre ordres de grandeur respectivement, ce dispositif a permis d’observer une grande variété de régimes d’écoulement. Dans certaines situations, les bulles de taille millimétrique restent piégées à l’interface liquide-liquide, tandis que les bulles plus grosses parviennent à traverser l’interface, entraînant une importante colonne de liquide lourd derrière elles. Après que l’influence des paramètres physiques a été qualitativement établie à la lumière de modèles simples, des simulations numériques de plusieurs situations sélectionnées ont été réalisées. Celles-ci ont été menées à partir de deux approches basées sur les équations de Navier-Stokes incompressibles, l’une utilisant une technique de capture d’interface, l’autre une description de type « interface diffuse » de Cahn-Hilliard. Les comparaisons entre les résultats expérimentaux et numériques ont confirmé la fiabilité des prédictions numériques dans la plupart des cas, mais ont également souligné le besoin d’améliorer la capture de phénomènes physiques à petite échelle, en particulier ceux liés au drainage de film. / In order to predict the evolution of a hypothetical accident in pressurized water nuclear reactors, this study aims to understand the dynamics of gas bubbles ascending in a stratified mixture made of two superimposed liquids. To this aim, an experimental device equipped with two high-speed video cameras was designed, allowing us to observe isolated air bubbles and bubble trains crossing a horizontal interface separating two Newtonian immiscible liquids initially at rest. The size of the bubbles and the viscosity contrast between the two liquids were varied by more than one and four orders of magnitude respectively, making it possible to observe a wide variety of flow regimes. In some situations, small millimetric bubbles remain trapped at the liquid-liquid interface, whereas larger bubbles succeed in crossing the interface and tow a significant column of lower fluid behind them. After the influence of the physical parameters was qualitatively established thanks to simple models, direct numerical simulations of several selected experimental situations were performed with two different approaches. These are both based on the incompressible Navier-Stokes equations, one making use of an interfacecapturing technique, the other of a diffuse Cahn-Hilliard description. Comparisons between experimental and numerical results confirmed the reliability of the computational approaches in most situations but also highlighted the need for improvements to capture small-scale physical phenomena especially those related to film drainage.

Le syngas est un mélange gazeux de CO et H2 qui constitue un intermédiaire important dans l’industrie pétrochimique. Plusieurs approches sont utilisées pour le produire. L’oxydation partielle non catalytique (POX) et le reformage à la vapeur (SMR) en font partie. Le reformage auto thermique du méthane (ATR) combine quant à lui ces deux procédés au sein d’un même réacteur. L’amélioration du rendement global du procédé ATR requiert une meilleure caractérisation du comportement des gaz au sein de la chambre. La simulation numérique apparaît comme un outil efficace pour y parvenir. Pour réduire le coût CPU, c'est généralement l'approche RANS (Reynolds Average Numerical Simulation) qui est privilégiée pour la simulation complète de la chambre. Cette approche repose sur l'utilisation de modèles, parmi lesquels le modèle de combustion turbulente, qui a pour objectif de représenter les interactions entre la turbulence et la réaction chimique au sein du mélange. Plusieurs stratégies ont été proposées pour le calculer, qui bénéficient globalement d'une large expérience pour les systèmes classiques mettant en jeu la combustion. Cependant, les flammes observées dans les réacteurs ATR présentent des propriétés assez différentes de ces configurations classiques. La validité des modèles de combustion turbulente classiques doit donc y être vérifiée. L'objectif de cette thèse est de répondre à ce besoin, en testant la validité de différents modèles de combustion turbulente. La première partie du travail a consisté à analyser les propriétés des flammes CH4/O2 enrichies en vapeur d'eau à haute pression, et a notamment permis le développement d’une méthode d’évaluation des temps caractéristiques d’un système chimique. Dans un deuxième temps, une expérience numérique à l’aide d’un code DNS a été réalisée, afin de servir de référence pour tester a priori sur des configurations ATR plusieurs modèles RANS de combustion turbulente couramment utilisés dans le milieu industriel. / Syngas is a gaseous mixture mainly composed of CO and H2, which constitutes a major feedstock in petrochemical industry. Several industrial approaches are commonly used to produce it. Non catalytic Partial Oxidation (POX) and Steam Methane Reforming (SMR) are two of them. Autothermal Reforming (ATR) is a third process that combines both POX and SMR in the same reactor. A better knowledge of the reactive flow properties inside the chamber is required in order to improve the ATR process efficiency. Numerical simulation appears as an efficient tool to reach this goal. Because of the high CPU cost required for these simulations, RANS (Reynolds Average Numerical Simulation) formulation is usually preferred for the simulation of the whole chamber. This approach relies on the use of models, like the turbulent combustion model that aims at describing the interactions between turbulence and chemical reactions. Several approaches have been proposed to compute it, which benefit from a relatively wide experience for the simulation of classical combustion systems. However, ATR flames have some specific properties that make them quite different from these classical configurations, especially because of high pressure, reactants dilution with water and high global equivalence ratio. The validity of classical turbulent combustion models therefore requires to be assessed in ATR configurations. The objective of this thesis is to meet this need by testing the validity of several turbulent combustion models. The first part of this work has been to analyze water-enriched CH4/O2 flames properties at high pressure. In particular, a strategy for evaluating characteristic chemical time scales of a reactive system has been proposed within this context. In a second part, a DNS numerical experiment has been performed. Its results are then used as a benchmark for a priori testing several turbulent combustion models in the context of ATR reactor RANS simulations.

Cette étude porte sur le transfert thermique pariétal dans un canal plan turbulent. L'étude est théorique et numérique. Nos simulations directes (DNS) sont effectuées avec le code de calcul Incompact3d. On a porté un intérêt particulier aux grandeurs que l'on trouve dans les bilans des flux thermiques turbulents et de la variance de la température : ces données permettent de valider les modèles de type RANS. On analyse également nos simulation à l'aune de statistiques plus fines, telles que les corrélations en 2 points. On distingue 2 traitements de la thermique dans le cas du canal plan turbulent : avec ou sans prise en compte du transfert thermique conjugué (couplage thermique fluide/solide). Pour les cas avec transfert thermique conjugué, on a mis en évidence une condition de compatibilité dans l'espace spectral entre la température et le flux de chaleur à l'interface fluide-solide. En l'absence de transfert thermique conjugué, notre étude se borne aux conditions limites qui sont une combinaison linéaire à coefficients constants de la température et du flux de chaleur à la paroi (Dirichlet, Neumann, Robin). Pour ces conditions aux limites simples, on met en évidence une condition de compatibilité entre les valeurs pariétales de la variance de la température et la partie normale de la dissipation associée. D'une part, cette relation souligne les limites des simulations avec une température ou un flux imposé à la paroi. D'autre part, elle permet de construire des conditions de type Robin sur-mesure qui donnent des résultats proches de ceux obtenus avec transfert thermique conjugué pour la configuration du canal plan turbulent. / This study focuses on the turbulent heat transfer in the turbulent channel flow configuration. Our Direct Numerical Simulations are performed using the open-source code Incompact3d. As our target is to produce data for RANS models validation, the budgets of the turbulent heat fluxes and of the temperature variance are extracted. Two-point correlations for the temperature and wall-normal heat flux are also presented to deepen our analysis. Regarding the thermal field, 2 configurations are considered: with and without conjugate heat transfer (thermal coupling between the fluid and solid domains). For conjugate heat transfer cases, a novel compatiblity condition, expressed in the spectral space, connects the temperature and wall-normal heat flux at the fluid-solid interface. For non-conjugate cases, our study is limited to boundary conditions that impose a linear combination of the temperature and wall-normal heat flux at the wall using constant coefficients (Dirichlet, Neumann, Robin). For such simple boundary conditions, a novel compatibility condition is obtained which connects the wall-value of the temperature variance and the wall-normal part of the associated dissipation rate. On one hand, this condition highlights the limitations of an imposed temperature or heat-flux at the wall. On the other, it allows us to build tailored Robin boundary conditions able to reproduce satisfactorily present conjugate heat-transfer results in the channel flow configuration.

Turbulent gas-particle suspensions are of great practical importance in many naturally phenomena, such as dust storms and snow avalanches, as well as in industrial applications such as fluidised, circulating bed reactors and pneumatic transport. Due to the difference in mass density of about three orders of magnitude between solids and gases, the mass loading is large, but the volume fraction of the particles is usually small. Since the length scale of these flows ranges from tens of centimeters to hundreds of meters, the Reynolds number based on the flow dimension and velocity is usually large. Due to this, these flows are almost always in the turbulent regime, and the fluid velocity fluctuations are significant. The particle sizes are typically small in these applications, of the order of 100 m or less. Due to this, the Reynolds number (based on the particle size and velocity) is usually low. This implies that the fluid inertia is not important, and the flow dynamics is dominated by fluid viscosity at the particle scale. At the same time, due to the density contrast between the particles and fluid, the Stokes number (ratio of particle inertia and fluid viscosity) is large. The inertia is sufficiently large that the particles cross the fluid streamlines. In this situation, there is a two-way coupling between the fluid turbulence and the particle dynamics. The turbulent fluid velocity fluctuations result in particle velocity fluctuations due to the drag force exerted by the particles on the fluid. In turbulent gas-particle suspensions, the fluctuating velocity of the particles results in a force on the fluid, which could either enhance or dampen the turbulent velocity fluctuations. The finite size of the particles could also result in fluid velocity effects which can not be captured by considering the particles as point masses. The dynamics of turbulent particle suspensions is analysed in the limit of low particle Reynolds number and high particle Stokes number, where there is a balance between particle inertia and fluid viscosity. The turbulent gas flow in a channel is considered for definiteness, in order to analyse the effect of turbulent fluctuations, as well as the effect of cross-stream variations in the turbulent statistics. The particle size is considered to be comparable to the Kolmogorov scales, which are the smallest scales in the turbulent flow. In addition, the fluid inertia at the particle scale is neglected, and the particles are dynamics is modeled using the Stokes equations. However, inertial effects are included at the macroscopic scale, where the Navier-Stokes equations are solved by Direct Numerical Simulations (DNS) using Chebyshev-Fourier spectral techniques. There are three important objectives in the present analysis. 1. The first is to examine the turbulence modification due to the reverse force of the particles. There are two models used for the reverse force of the particles on the fluid. The first is a point force, which is modeled as a delta function in real space. Instead of using smoothing functions for the delta function, we prefer to incorporate the point force in the momentum conservation equation in spectral space. A more complicated representation proposed here involves the inclusion of the symmetric and anti-symmetric force moments, calculated from the solution of Stokes equations for the flow around the sphere. These are represented as gradients of delta functions, and are also included in the momentum conservation equations in the spectral co-ordinates. 2. The second objective is to examine the effect of particle rotation and collisions on the flow dynamics. While particle rotation is usually included in the analysis of granular flows, this is not normally included in the treatment of particle collisions. 3. The third objective is to examine whether the effect of the fluid turbulence can be modeled as a fluctuating force. When the viscous relaxation time of the particles is larger than the integral time for the fluid velocity fluctuations, the fluid velocity fluctuations can be considered as delta function correlated in time, and the effect of these fluctuations can be incorporated using a Langevin description. In this case, the diffusion coeffcients in the Langevin equation for the particles is calculated from the correlation in the fluid velocity fluctuations. The new objective here is to include both the drag force and the torque exerted on the particles in the presence of particle rotation, and to examine whether these are sufficient to capture the effect of ff fluid turbulence on the particle phase. The Direct Numerical Simulations show that there is a significant attenuation of the turbulent velocity fluctuations when the reverse force exerted by the particles is added in the fluid momentum equations, and the particles are considered to be smooth. This turbulence attenuation is greater when the particle volume fraction increases, and when the particle mass density increases. However, when particle rotation is included, the turbulent velocity fluctuations are significantly larger than those without rotation, and in come cases are close the fluctuation levels when the reverse force is included. Thus, the particle rotation has a significant enhancement on the turbulent velocity fluctuations. The attenuation in the fluid turbulence is also reflected in the magnitude of the particle fluctuating velocities. The particle fluctuating velocities are higher when the effect of particle rotation is included. The reason for this is that there is particle rotation induced due to mean fluid shear, and this rotational energy gets transformed into translational energy in inter-particle collisions. The effect of inclusion of the symmetric and anti-symmetric force moments does not result in a significant change in the turbulence intensities for the range of volume fractions and mass densities considered here. There is a small but discernible increase in the turbulence for the largest volume fraction and mass density considered here, but this increase is much smaller than the significant turbulence attenuation due to the inclusion of particle rotation. Systematic trends are also observed in the particle linear and angular velocity distributions. The particle stream-wise linear velocity distribution, and the span-wise angular velocity distribution are broader than a Gaussian distribution near the zero, and exhibit steep decrease at larger velocity. They are also asymmetric, and the distribution depends on the location across the channel. The distribution of the cross-stream and span-wise linear velocity and the stream-wise and cross-stream angular velocity, is narrower than a Gaussian distribution at the center, and exhibits long tails for high velocities. Thus, there are systematic variations in the distribution functions for both the linear and angular velocities, which need to be included in kinetic theory descriptions for the particle phase. The fluctuating force model has also been simulated, where particle dynamics is explicitly simulated, the fluid velocity fields are not simulated, but are modeled as fluctuating forces and torques acting on the particles. The variance in the fluctuating force and torque are determined from the correlations in the fluid velocity and the vorticity fields, and these are modified to include the turbulence attenuation due to the reverse force exerted by the particles. The fluctuating force simulations do accurately capture the trends observed in the mean and fluctuating velocities. They are also able to capture the non-Gaussian nature of the linear and angular velocity distributions of the particles, even though the random forcing is considered to be a Gaussian function. Thus, the fluctuating force formulation can be used to accurately capture the effect of the fluid on the particles, only if the forces are modified to include the effect of turbulence attenuation due to the reverse force exerted by the particles.

Cette thèse consiste en l'étude expérimentale et numérique des processus de formation des particules de suie au sein des flammes laminaires non-pré-mélangées et partiellement prémélangées sous l'influence d'un champ magnétique ou d'une stimulation acoustique. Dans une premiére étape, la capacité du code CIAO à prédire la fraction volumique de suie dans une flamme axisymétrique est étudiée. Par la suite, deux flammes subissant une stimulation acoustique ont été étudiées. Les résultats peuvent être utilisés pour améliorer les modèles de suie futurs, en particulier concernant les différentes échelles temporelles de la chimie en phase gazeuse, et la formation d'hydrocarbures polyaromatiques (PAH) et de suie couplée avec des flux transitoires. Pour étudier la formation des particules de suie sous l'influence de gradients de champ magnétique, un brûleur de type Santoro est utilisé. Les techniques de mesure utilisées dans le cadre de cette thèse sont l'imagerie directe à haute cadence, la technique Background Oriented Schlieren (BOS) et la méthode d'Absorption/Emission Modulée (MAE). Une augmentation de la fraction volumique de suie intégrée a été mise en évidence lorsque le gradient de champ magnétique est ascendant. Une analyse de stabilité linéaire locale appliquée à l'écoulement non-visqueux est présentée pour une flamme sous l'influence de la perturbation magnétique envisagée. Le gradient de champ magnétique provoque alors une réduction du taux d'amplification. De fait, l'étude est complété par l'identification d'un domaine où les flammes qui oscillent naturellement peuvent être stabilisées et contrôlées par des gradients de champ magnétique. / In this thesis light is shed on the soot formation processes in laminar coflow flames influenced by magnetic field gradients and acoustic forcing. Both influences have been assessed experimentally and numerically. First, the CIAO in-house code's ability to predict soot volume fraction fields in a steady coflow flame is studied. Then, two acoustically forced cases were studied. These findings are used to improve future soot models, especially, concerning the different time scales of gas phase chemistry and the formation of polycyclic aromatic hydrocarbons (PAH) and soot coupled with unsteady flows. To investigate soot formation under magnetic field gradients, a Santoro type burner is used. The measurement techniques applied in the course of this thesis are high-speed luminosity measurements, Background Oriented Schlieren (BOS) and one- and two-color Modulated Absorption/Emission (MAE) techniques. The magnetic field impact on soot formation was first studied experimentally in steady laminar flames. A scaling of soot production similar to the increased integrated soot volume fraction with increased oxygen content in the coflow was documented. A local inviscid stability analysis is presented for an ethylene coflow flame to investigate the flame's response to small perturbations of the mean velocity, temperature, fuel, and oxygen massfraction under magnetic field exposure. The magnetic field is found to reduce the perturbations' growth rate. The magnetic field study is completed by identifying a domain where naturally oscillating flames can be stabilized and controlled by magnetic field gradients.

The search for clean and efficient combustors is motivated by the increasingly stringent emissions regulations. New gas turbine engines are designed to operate under lean conditions with inhomogeneous reactants to ensure cleanliness and stability of the combustion. This ushers in a new mode of combustion, called the inhomogeneous reactants premixed combustion. The present study investigates the effects of inhomogeneous reactants on premixed combustion, specifically on the interactions of an initially planar flame with field of inhomogeneous reactants. Unsteady and unstrained laminar methane-air flames are studied in one- and two-dimensional simulations to investigate the effects of normally and tangentially (to the flame surface) stratified reactants. A three-dimensional DNS of turbulent inhomogeneous reactants premixed combustion is performed to extend the investigation into turbulent flames. The methaneair combustion is represented by a complex chemical reaction mechanism with 18 species and 68 steps. The flame surface density (FSD) and displacement speed S_d have been used as the framework to analyse the inhomogeneous reactants premixed flame. The flames are characterised by an isosurface of reaction progress variable. The unsteady flames are compared to the steady laminar unstrained reference case. An equivalence ratio dip is observed in all simulations and it can serve as a marker for the premixed flame. The dip is attributed to the preferential diffusion of carbon- and hydrogen- containing species. Hysteresis of S_d is observed in the unsteady and unstrained laminar flames that propagate into normally stratified reactants. Stoichiometric flames propagating into lean mixture have a larger S_d than lean flames propagating into stoichiometric mixtures. The cross-dissipation term contribution to S_d is small (~~10%) but its contribution to the hysteresis of S_d is not (~~50%). Differential propagation of the flame surface is observed in the laminar flame that propagates into tangentially stratified reactants. Stretch on the flame surface is induced by the differential propagation, which in turn increases the flame surface area.

No description available.

La turbulence tridimensionnelle se caractérise par sa capacité à transférer de l'énergie des grandes vers les petites échelles où elle est finalement dissipée. Lorsqu’elle se produit dans un plasma non-collisionnel comme le vent solaire, une modélisation cinétique semble a priori nécessaire. Toutefois, la complexité d’une telle approche limite les développements théoriques et condamne les expériences numériques à se restreindre à des nombres de Reynolds peu élevés. Dans quelles mesures un modèle mono-fluide comme la MHD Hall permet-il de rendre compte des phénomènes observés dans le vent solaire aux échelles sub-ioniques ? C’est la problématique à laquelle s’est attaquée cette thèse. L’idée directrice de ce travail est de tirer profit de la relative simplicité des modèles fluides et de la puissance algorithmique des méthodes pseudo-spectrales pour aborder la turbulence du vent solaire par des simulations numériques directes tridimensionnelles massivement parallèles à grands nombres de Reynolds. Ces simulations numériques ont permis de mettre en évidence l’existence d’une brisure spontanée de symétrie chirale en turbulence MHD Hall incompressible, ainsi que l’existence d’un nouveau régime appelé ion MHD (IMHD). Un modèle phénoménologique a été proposé pour rendre compte de ces résultats et de nouvelles prédictions ont été faites, puis confirmées numériquement. Enfin, l’étude de l’effet d’un fort champ magnétique uniforme sur la dynamique turbulente a permis de confirmer pour la première fois une ancienne conjecture. L’inertie des électrons a ensuite été prise en compte toujours dans un modèle fluide. Par une approche hydrodynamique classique, une loi universelle a été obtenue pour les fonctions de structure d’ordre trois. L’ensemble de ces résultats est qualitativement en accord avec les mesures in situ du vent solaire et remet en cause le paradigme selon lequel les raidissements successifs du spectre des fluctuations magnétiques sont provoqués nécessairement par des phénomènes d’origine cinétique. De manière plus générale, cette thèse soulève des questions fondamentales sur les processus non-collisionnels de dissipation dans les plasmas turbulents. / Three-dimensional turbulence is characterized by its capacity to transfer energy from large to small scales where it is finally dissipated. When it occurs in a non-collisional plasma like the solar wind, a kinetic modelisation is necessary a priori. The complexity of such an approach however limits the theoretical developments and forces numerical experiments to be restricted to low Reynolds numbers. To what extent does a single-fluid model such as MHD Hall account for the phenomena observed in the solar wind at ion sub-scales ? It is to this question that this thesis tries to answer. The main idea of this work is to take advantage of the relative simplicity of fluid models and of the high precision of pseudo spectral methods to tackle the problem of turbulence in solar wind by direct numerical simulations massively parallelized at high Reynolds numbers. These simulations have helped to highlight the existence of a spontaneous breaking of chiral symmetry in incompressible Hall MHD turbulence, as well as the existence of a new regime called ion MHD (HDMI). A phenomenological model has been proposed to account for these results and new predictions were made and confirmed numerically. The study of the effect of a strong uniform magnetic field on the turbulent dynamics confirmed an ancient conjecture for the first time. The inertia of the electrons was then taken into account in a still fluid model. By a classical hydrodynamic approach, a universal law has been obtained for the third order structure functions. All these results are in qualitative agreement with in situ measurements of the solar wind and challenge the paradigm according to which the successive steepening of the magnetic fluctuations spectrum is necessarily caused by phenomenon of kinetic origin. More generally, this thesis raises fundamental questions about the non-collisional dissipation process in turbulent plasmas.