Cette thèse se concentre sur l’étude des instabilités de combustion dans un brûleur prémélangé. Les instabilités sont généralement issues d’un couplage entre la combustion et les modes propres du système. La mise en résonance qui en résulte peut avoir des conséquences qui sont souvent dommageables, entraînant des vibrations, une fatigue des matériaux soumis à des charges acoustiques élevées et une intensification des flux de chaleur vers les parois de la chambre. Un premier objectif de cette thèse est de poursuivre le développement de méthodes de prévision des instabilités et des phénomènes non-linéaires qui en résultent comme par exemple le développement de cycles limites, les processus de déclenchement (“triggering”), la commutation de modes. Le cadre général adopté est celui de «°l’équivalent harmonique » bien connu dans le domaine du contrôle et qui a été exploré dans le domaine des instabilités de combustion dans des travaux récents du laboratoire EM2C, CNRS. Par le biais de ce concept il est possible de tenir compte de l’´evolution de la réponse de la flamme suivant l’amplitude à laquelle elle est soumise. Cette réponse de flamme en fréquence et amplitude généralise la notion de fonction de transfert et elle est désignée sous le nom de “Flame Describing Function” (FDF). Le système est ouvert à son extrémité aval. Cette géométrie permet de simplifier l’analyse et d’obtenir une large gamme de configurations au moyen d’une variation continue de la longueur du conduit d’alimentation qui est limité en amont par un piston. On peut aussi échanger le tube à flamme et utiliser des longueurs différentes de cet élément. Une étude exhaustive est réalisée pour répertorier les oscillations observées et déduire leurs propriétés. On montre que les cycles limites qui possèdent une amplitude constante sont bien décrits par la méthode unifiée fondée sur la FDF. Pour certaines configurations l’expérience fait apparaître des cycles limites dont l’amplitude et la fréquence ne se stabilisent pas au cours du temps. On observe notamment des oscillations plus complexes couplées par plusieurs modes pouvant soit donner lieu à des variations régulières ou à des fluctuations plus irrégulières avec un caractère “galopant” dans le temps. Pour ces oscillations particulières, la FDF fournit des indications sur les domaines d’apparition mais n’est pas en mesure de décrire complètement ces cycles limites complexes. Il faut dans ce cas recourir à une représentation temporelle qui n’est pas développée dans ce document. La base de données expérimentales pourra être utilisée pour guider ultérieurement ce type d’analyse. Le deuxième grand objectif de cette thèse est de rechercher des méthodes de contrôle des instabilités. On considère plus particulièrement des systèmes dynamiques utilisant des plaques perforées polarisées par un écoulement (BFP : “bias flow perforate”). Ces systèmes sont particulièrement intéressants pour atténuer les oscillations basse fréquence qui sont difficiles à réduire par des systèmes passifs. La conception de ces BFPs est fondée sur des travaux récents menés au laboratoire EM2C, CNRS avec notamment l’objectif de robustesse, c’est-à-dire la possibilité de couvrir une large bande de fréquences. L’´etude expérimentale et les calculs fondés sur la FDF menés en parallèle permettent de voir les possibilités de tels systèmes et de comprendre les conditions nécessaires à leur efficacité. Cette étude peut permettre de guider les applications qui pourraient être envisagées en pratique. / This thesis is concerned with an investigation of combustion instabilities in premixed combustors. This problem has been the subject of a continuous effort in relation with the many issues encountered in practical systems like those used in propulsion and energy production. Combustion instabilities usually arise from the coupling between combustion and acoustic eigenmodes of the system. In most cases such resonances lead to vibrations, structural fatigue and intensified heat fluxes to the chamber walls. The first part of this thesis pursues the development of prediction methods for combustion instabilities and the associated nonlinear phenomena such as limit cycles establishment, triggering, mode switching and hysteresis. The aim is to delineate physical mechanisms and develop analytical methods dedicated to prediction. The theoretical framework relies on the “harmonic balance” formalism well known in the domain of control and which has been adopted more recently in combustion instability studies carried out at EM2C, CNRS laboratory. Through this concept, it is possible to take into account the evolution of the flame response as a function of amplitude. This flame response, depending on frequency and amplitude, extends the flame transfer function principle and is designated as the “Flame Describing Function” (FDF). The development of the FDF framework is pursued in the present study. The experimental setup which exemplifies combustion instabilities and serves to validate the method has generic features as it comprises in an idealized version, all the parts found in practical systems : a feeding manifold delivering a mixture of methane and air, a multipoint injector made of a perforated plate anchoring a collection of small laminar conical flames and a flame tube made of quartz which confines the combustion zone. The downstream boundary of the system is open. This device allows a simplified analysis and provides a wide variety of configurations through the continuous modification of the feeding manifold length which is bounded by a piston on the upstream and through changes of the flame tube lengths. Systematic comparison between theoretical results and well controlled experiments is undertaken. Depending on the geometry, the setup exhibits a large variety of unstable modes which are classified in terms of their limit cycle behavior using tools from dynamical system theory. It is shown that limit cycles with constant amplitude are well predicted by the unified FDF methodology. For some configurations, the experiment reveals limit cycles characterized by time variable amplitude and frequency. One finds situations where the oscillation is coupled by multiple modes leading either to regular amplitude variations or more irregular evolutions with a “galloping” pattern as a function of time. For this special type of limit cycle, the FDF indicates the range of the onset, but is not able to fully describe these complex limit cycles. These oscillations require a time domain state space analysis which is not addressed in this manuscript. The experimental database may be of value for further work in this direction. The second part of this thesis deals with control methods for instabilities. One specifically considers damping systems relying on perforated plates biased by a flow (BFP : “Bias Flow Perforate”). These systems are particularly interesting because they can be used to cancel low frequency oscillations which are otherwise difficult to reduce through passive control methods. This BFP design relies on recent work carried out at EM2C, CNRS laboratory which extends the frequency range where the system is effective. The experimental study and the associated FDF calculations are used to delineate the possibilities of such systems and uncover conditions required for an effective damping of oscillations. This study provides indications on the practical application of BFPs.

Les instabilités thermo-acoustiques causent des problèmes récurrents dans les chambres de combustion pour une large gamme d'applications industrielles, allant des chaudières domestiques aux turbines à gaz, en passant par les moteurs fusées. Ces phénomènes résultent d’un couplage résonant entre la dynamique de la combustion et les modes acoustiques du foyer, et peuvent donner lieu à de fortes vibrations, un vieillissement prématuré des composants de la chambre, voire des dommages structurels. Les mécanismes physiques mis en jeu sont complexes et difficiles à modéliser, ainsi les oscillations thermo-acoustiques ne sont pas facilement prévisibles au stade de la conception d’une chambre de combustion. Dans de nombreux foyers, des systèmes d’amortissement passifs sont installés pour augmenter la dissipation d’énergie acoustique et empêcher le développement de ces instabilités. Dans ce travail, des systèmes d’amortissement basés sur des plaques perforées couplées à une cavité résonante et traversées par un écoulement moyen sont analysés. Les principaux objectifs sont : (i) d’améliorer et de simplifier la conception de systèmes d’amortissement robustes en maximisant leurs propriétés d’absorption acoustique en régime linéaire, (ii) d’analyser l’effet de l’amplitude des ondes sonores incidentes sur la réponse acoustique des plaques perforées et (iii) de développer des modèles capables de reproduire cette réponse aux hautes amplitudes. Tout d’abord, deux régimes asymptotiques intéressants sont identifiés où le système fonctionne à faibles et forts nombres de Strouhal respectivement. Dans ces régimes la conception d’un système d’amortissement maximisant l’absorption acoustique est grandement simplifiée, puisque les calculs de la vitesse optimale de l’écoulement et de la taille de la cavité sont découplés. Il est démontré qu’à faible nombre de Strouhal le système se comporte comme un résonateur quart d’onde, et dispose d’une bande d’absorption très large. À fort nombre de Strouhal, le système fonctionne comme un résonateur de Helmholtz, comportant une cavité de taille plus réduite, mais une bande d’absorption beaucoup plus étroite que dans le régime précédent. Ces prévisions sont confirmées par des mesures réalisées dans les différents régimes identifiés sur un dispositif expérimental dédié. L’évolution des propriétés acoustiques d’une plaque perforée lorsque l’amplitude de forçage augmente est ensuite examinée par le biais de simulations directes. Il est montré que la transition du régime linéaire au régime non linéaire se produit lorsque l’amplitude de la vitesse acoustique dans l’orifice est comparable à la vitesse de l’écoulement moyen dans les trous. Pour des amplitudes élevées, une inversion périodique de l’écoulement est observée dans l’orifice. Des anneaux tourbillonnaires sont alternativement éjectés en amont et en aval de l’orifice à une vitesse de convection qui augmente avec l’amplitude de la perturbation acoustique. Ces mécanismes influencent profondément l’absorption acoustique des plaques perforées dans le régime non linéaire. Deux nouveaux modèles décrivant la réponse non linéaire de ces systèmes sont ensuite développés en exploitant la trajectoire des vortex (modèle VC), et une approche quasi-stationnaire (modèle IDF). Les prévisions de ces modèles sont confrontées à des mesures effectuées dans le tube à impédance et aux résultats de simulations directes. Les résultats obtenus au cours de ces travaux peuvent être utilisés pour guider la conception de systèmes d’absorption robustes, capables de fonctionner dans des environnements difficiles avec des niveaux sonores élevés, comme ceux rencontrés lors d’instabilités thermo-acoustiques. / Thermo-acoustic instabilities are of primary concern in combustion chambers for a wide range of industrial applications, from domestic boiler to gas turbines or rocket engines. They are the consequence of a resonant coupling between the flame dynamics and the acoustic modes of the combustor, and can result in strong vibrations, early aging of combustor components and structural damage. The physical mechanisms involved are complex and difficult to model, thus thermo-acoustic oscillations are not easily predictable at the design stage of a combustor. In many combustors, passive dampers are implemented to increase the acoustic energy dissipation of the system and to hinder detrimental flame-acoustics interactions. In the present work, passive damping systems based on perforated screens backed by a resonant cavity and traversed by a bias flow are investigated. The main objectives are: (i) to improve and simplify the design of these dampers by maximizing their acoustic absorption properties in the linear regime, (ii) to analyze the effect of the sound wave amplitude on the acoustic response of these systems and (iii) to develop models capable of capturing absorption at high oscillation amplitudes. First, two interesting asymptotic regimes are identified where the plate operates at low and high Strouhal numbers respectively. In these regimes the design of a damper maximizing absorption is greatly simplified, since the choice of the optimal bias flow velocity and back cavity size can be decoupled. It is shown that at low Strouhal numbers the damper behaves as a quarter-wave resonator, and features a wide absorption bandwidth. At high Strouhal numbers, the system operates as a Helmholtz resonator, featuring shorter optimal back cavity sizes but narrower absorption bandwidths. These predictions are compared to measurements in a dedicated experimental setup for the different operating regimes identified. The dependence of the acoustic properties of a perforated plate on the forcing amplitude is then examined by means of direct numerical simulations. It is shown that transition from linear to nonlinear regimes occurs when the acoustic velocity amplitude in the orifice is comparable to the mean bias flow velocity. At high amplitudes, periodic flow reversal is observed within the perforation, vortex rings are alternatively shed upstream and downstream of the hole and convected away at a velocity which is increasing with the forcing amplitude. These mechanisms greatly influence the acoustic absorption of the perforate in the nonlinear regime. Two novel models capturing this nonlinear response are then inferred based on an analysis of the vortex trajectory (VC model), and on a quasi-steady description of the flow (IDF model). Their predictions are finally compared to measurements conducted in an impedance tube, and to results from numerical simulations. The results obtained in this work can be used to ease the design of robust dampers capable of operating in harsh environments with high sound levels, such as those found during self-sustained thermo-acoustic instabilities.

Neste trabalho foram estudadas instabilidades magnetohidrodinâmicas (MHD) utilizando um novo sistema bolométrico que foi instalado no tokamak TCABR para medidas da evolução temporal da potência irradiada. Este novo sistema conta com 24 cordas verticais, capazes de mapear toda uma secção poloidal da coluna de plasma com resolução espacial de aproximadamente 2 cm e uma resolução temporal de 20 µs. Como se sabe, as instabilidades MHD degradam o connamento do plasma e modicam a topologia das superfícies magnéticas, causando a perda da energia do plasma. Por conta disso, compreender essas instabilidades é fundamental para o sucesso dos futuros reatores de fusão nuclear. As perturbações (oscilações) causadas pelas instabilidades MHD modulam diversos parâmetros macroscópicos do plasma como a densidade, a temperatura e a potência irradiada. Então, utilizando o diagnóstico bolométrico, é possível medir as oscilações no perl de potência irradiada e, a partir deles, extrair informações importantes para determinar a origem e as características de tais instabilidades. No tokamak TCABR, as instabilidades foram caracterizadas através da análise espectral dos 24 sinais provenientes do novo sistema bolométrico. Para auxiliar a caracterização das instabilidades, um programa foi desenvolvido em Matlab para simular as medidas das perturbações no perl de potência irradiada. Através do mesmo procedimento de análise espectral, os resultados simulados foram comparados aos experimentais de forma que os parâmetros simulados, como largura e posição das ilhas magnéticas, fossem ajustados aos experimentais. Através dessa metodologia de análise, que combina simulação e experimento, foi possível caracterizar diversas instabilidades como o precursor dos dentes de serra e ilhas magnéticas de modos m = 2 e m = 3. / In this dissertation, magnetohydrodynamic (MHD) instabilities were investigated using a new bolometric system that was installed in the TCABR tokamak for radiation power measurements. This diagnostic is composed by 24 vertical chords that provide a full view of the poloidal cross section of the plasma column and provides spatial and temporal proles with approximately 2 cm space and 20 µs time resolution. As it is well known, the MHD instabilities degrade the plasma connement and modify the magnetic topology, leading to energy loss from the plasma. Therefore, the understanding of these instabilities is essential for the success of the controlled thermonuclear fusion reactors. The MHD instabilities also cause perturbations (oscillations) in various macroscopic parameters, such as plasma density, temperature, and radiated power. Therefore, the oscillations in the radiated power prole measured by the bolometric diagnostic system provide a possibility to investigate the origin and features of the instabilities. In the TCABR tokamak, the instabilities were characterized by spectral analysis of the 24 vertical chords of the bolometric signals. In addition, a Matlab program was developed to simulate the integral characteristic of the oscillations in the radiated power measured by the bolometric system. The spectral analysis of the simulated signals is then compared with the spectral analysis of the bolometric signals. The simulated parameters, island width and radial position, were then adjusted to t the experimental spectrum results. Using this method of analysis, which combines experiment and simulation, it was possible to characterize various instabilities, such as sawtooth precursor and m = 2 and m = 3 magnetic islands.

Orientador: Pedro Fernandes Tavares / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Fisica Gleb Wataghin / Made available in DSpace on 2018-08-08T03:19:38Z (GMT). No. of bitstreams: 1 Abreu_NataliaPradode_D.pdf: 5325698 bytes, checksum: b2120bdd02b9ec07d7563695665add5e (MD5) Previous issue date: 2007 / Resumo: Esta tese investiga a dinâmica longitudinal não-linear do feixe de elétrons de alta intensidade circulante no anel de armazenamento do Laboratório Nacional de Luz Síncrotron (LNLS). Em particular, analisamos os efeitos sobre os modos de oscilação coerentes instáveis e sobre as características de equilíbrio dos pacotes de elétrons decorrentes da manipulação da distribuição de densidade eletrônica no espaço de fase longitudinal, realizada por meio de dois mecanismos distintos: a modulação em fase dos campos estacionários aceleradores e o uso de cavidades harmônicas. Determinamos a viabilidade do uso de cada uma das técnicas estudadas visando melhorar as propriedades do feixe dadas as características específicas e parâmetros do anel de armazenamento Brasileiro. Observamos que a modulação em fase é um método capaz de amortecer modos de oscilação de pacotes acoplados (coupled bunch modes) do feixe de elétrons e apresentamos evidências teóricas e experimentais de que o mecanismo físico responsável por este amortecimento é o aumento da dispersão das frequências de oscilação incoerente dos elétrons dentro dos pacotes, efeito chamado de amortecimento Landau. Observamos também um aumento no tempo de vida do feixe de elétrons devido µa diluição da densidade eletrônica, o que diminui a probabilidade de espalhamento elétron-elétron dentro dos pacotes, reduzindo assim a contribuição Touschek para a taxa de perda de elétrons. Atualmente a modulação em fase é utilizada rotineiramente em turno de usuários no anel de armazenamento do LNLS. Uma análise teórica assim como simulações da ação de cavidades harmônicas passivas no anel do LNLS mostram que efeitos semelhantes são observados, no que se refere µa estabilização de oscilações do feixe e ao aumento do tempo de vida, através dos mesmos mecanismos físicos presentes ao modularmos a fase dos campos de radiofrequência (RF). Há no entanto diferenças importantes nas características dos dois métodos: apesar da efficiência de operação dessas cavidades depender fortemente da corrente total armazenada, a distribuição natural de energias do feixe de elétron não é modificada, diferentemente do que ocorre quando usamos a modulação em fase; o aumento do tempo de vida em ambos os casos é muito parecido (da ordem de 30%) mas a cavidade harmônica, para determinados parâmetros de operação do anel de armazenamento, poder tornar o amortecimento de modos de oscilação de pacotes acoplados 16 vezes maior, enquanto que a modulação em fase aumenta o amortecimento por um fator 1.5 / Abstract: This work investigates the nonlinear longitudinal dynamics of an intense electron beam at the Brazilian Synchrotron Light Source (LNLS electron storage ring). Our main interest lies in the effects of the longitudinal electron density manipulation (by means of RF phase modulation or through the use of harmonic cavities) on the beam Touschek lifetime, coupled bunch instabilities and the steady state characteristics of the circulating bunches. We also determine the viability of each technique and optimize their results given the specific needs and parameters of the Brazilian storage ring. We show that phase modulation is efficient in damping coupled bunch instabilities of the circulating electron beam and present theoretical and experimental evidences that an increased spread of the electrons incoherent frequency inside the bunches is the physical mechanism responsible for this damping, also called Landau damping. We show, as well, an increase in the overall beam lifetime due to the dilution of the particle density inside the bunches, which decreases the probability of electron-electron scattering thus reducing the Touschek contribution to the electron loss rate. Since 2004, phase modulation has been used routinely during users shifts at the LNLS storage ring. The effects of the presence a set of passive harmonic cavities were studied theoretically and using a simulation program. Similar effects, related to the damping of instabilities and increasing in lifetime, are also obtained when using passive harmonic cavities by the same physical mechanisms observed when using phase modulation. There are, however, some important differences between the two methods (harmonics cavities and phase modulation): even though the optimum conditions for operation of these cavities varies strongly with the stored current, this devices does not modify the natural energy distribution of the electron beam as observed when using phase modulation. Also, the increase in lifetime using both methods is similar (around 30%), but harmonics cavities can enhance the damping of coupled bunch modes by a factor of 16 while phase modulation provides an increase in damping by a factor 1.5 / Doutorado / Física Geral / Doutor em Ciências

Redes neurais artificias, tipo \"feedforward\", de duas camadas, foram utilizadas neste trabalho para fazer previsões das instabilidades de disruptura que surgem nas descargas de plasma do tokamak TEXT (E.U.A.), obtendo-se resultados bastante encorajadores. Verificou-se que uma arquitetura de rede, do tipo m:2m:m:1, onde m é dimensão de imersão do atrator do sistema dinâmico em estudo, costuma ser um bom chute inicial para a escolha da arquitetura ideal de trabalho, que costuma ser livre e, não raro, trabalhosa. Utilizando-se, em sinais de raios-X, uma rede neural artificial com arquitetura 15:30:15:1, por exemplo, conseguiu-se fazer previsões com uma antecipação de até 4 ms das instabilidades de disruptura, tempo quatro vezes maior do que o obtido utilizaudo-se sinais magnéticos das bobinas de Mirnov. Tal antecipação é bastante significativa e abre a possibilidade de, no futuro, utilizarem­se mecanismos de defesa da máquina, tais como injeção de partículas neutras (ou\"pellets\"), aplicação de campos magnéticos externos, etc, no sentido de se tentar evitar a ocorrência destas instabilidades, ou, pelo menos, minimizar os seus efeitos nocivos. Isto certamente contribuirá significativamente para a viabilização dos futuros reatores de fusão à plasma. Finalmente, o sistema de diagnóstico de raios-X de baixas energias do tokamak TCABR, que foi projetado e já se encontra em fase de instalação para fornecer sinais que servirão para alimentar a rede neural, também possibilitará a reconstrução tomográfica das regiões de mesma emissividade da coluna de plasma. A análise tomográfica, utilizando-se os sinais de dois conjuntos de detectores de raios-X moles, também será muito útil na investigação dos mecanismos físicos que dão surgimento às instabilidades de disruptura, além de permitir, ainda, a medida da temperatura eletrônica do plasma, através do método dos absorvedores. / Two-layer feedforward neural network has been used in this work to forecast the disruptive instabilities that occur in the TEXT tokamak plasma discharges. For this task, soft X-ray experimental signals were used with very promising results. It was verified that a neural net with an architecture of the type m:2m:m:1, where m is the embedding dimension of the atractor of dynamical system in focus, is usually a good initial guess in the searching process of finding the ideal architecture. A neural network with architecture 15:30:15:1 was capable, for example, to forecast the disruptive instabilities up to 4 ms in advance. This period of time is four time larger than the one obtained when magnetic signals from Mirnov coils were used. This forecasting time is quite significative and opens up the possibility of using defensive mechanisms, such as the injection of neutral particles (or pellets), the application of external magnetic fields, etc, with the objective of avoiding the occurrence of the disruptions or, at least, to minimize their harmful effects. This achievement certainly would be an important contribution to the development of the next generation fusion devices. Finally, the soft X-ray diagnostic system for the TCABR was projected and it is already being installed. This system will provide experimental signals that will be analyzed by neural networks and will be also used to identify, through tomografic image reconstructions, the regions of the plasma that have the same soft X-ray emissivity. The tomography analysis of the plasma, that will be carried out by using the signals of two soft X-ray detectors arrays, will be also very usefull for investigating the triggering mechanism of disruptions and will also allow the determination of the plasma electron temperature through the two foil absorbing method.

L'étude des instabilités aéro-élastiques, le flottement notamment, a été initiée au milieu du vingtième siècle, suite à de nombreux accidents de rupture d'aile d'avions ou de tabliers de ponts. L'un des accidents les plus connus est l'effondrement du pont de Tacoma, quelques mois après sa mise en service en 1940.Le flottement est une vibration synchronisée d'une structure souple se déplaçant dans un milieu fluide. Il se produit lorsque deux mouvements rythmiques réguliers coïncident de telle façon que l'un alimente l’autre, tirant l'énergie supplémentaire de l'écoulement environnant. Un cas classique de flottement d'aile d’avion consiste en la combinaison de mouvements de flexion et de torsion. Ce travail pose la problématique des instabilités par couplage fluide structure des surfaces portantes dans l’eau, les instabilités hydro-élastiques. Une différence importante par rapport aux instabilités aéroélastiques est le fait que la structure souple évolue dans un fluide lourd, ce qui implique en particulier des effets de masse ajoutée et d'amortissement fluide a priori importants. Le flottement est apparu pour la première fois sur les quilles composites des voiliers de compétition, donc dans l'eau, en 2004 :• Sur le voilier IMOCA 60 pieds POUJOULAT-ARMORLUX de Bernard Stamm, pendant la course transatlantique «The Transat» : il a perdu sa quille et chaviré.• Sur le voilier IMOCA 60 pieds SILL de Rolland Jourdain: la quille et le bateau ont été sauvés. Suite à ces problèmes - en particulier suite à la perte de la quille du voilier de Bernard Stamm, un accident qui aurait pu avoir des conséquences dramatiques pour le skipper – la société HDS, spécialisée dans la conception et le dimensionnement de structures composites, notamment dans le domaine du nautisme, s’est penchée sur le phénomène. Le flottement ne s'est produit que sur des quilles composites basculantes de voilier IMOCA 60 pieds et VOLVO 70 pieds. Les principales questions posées sont donc "Pourquoi les quilles composites sont-elles susceptibles de flotter, et est-il possible de prévoir et de prévenir ce comportement ?", puis "Une bonne estimation de la vitesse critique de flottement d'une quille peut-elle être calculée à moindre frais ?".Ce travail présente les méthodes analytiques, expérimentales et numériques mises en œuvre pour estimer la vitesse critique de flottement pour différents types d'appendices dans l'eau. Des modèles, basés sur une base modale tronquée pour les modes les plus énergétiques qui sont généralement, pour une quille à bulbe, le mode de flexion prépondérante et le mode de torsion prépondérante, sont développés et un outil de conception est proposé. Une des exigences de ce travail était, en effet, de réaliser un outil simple pour intégrer le calcul de la vitesse critique de flottement dans les premières boucles de conception d'une quille composite ou acier. Les modèles proposés, qui donnent de bons résultats pour les deux cas de flottement de quille présentés ci-dessus, sont confrontés à des résultats expérimentaux et à des simulations multiphysiques en couplage fluide structure réalisées en utilisant le logiciel ADINA. Enfin, une étude paramétrique est proposée présentant l'influence des principaux paramètres de conception sur l'apparition des instabilités. / The study of aeroelastic instabilities, including flutter, was initiated in the mid-twentieth century, following numerous accidents breaking wing aircraft or bridge decks. One of the most famous accidents is the collapse of the Tacoma Narrows Bridge, a few months after its commissioning in 1940.Flutter is a synchronized vibration which takes place in a flexible structure moving through a fluid medium. It occurs when two regular, rhythmic motions coincide in such a way that one feeds the other, drawing additional energy from surrounding flow. A classic case of wing flutter might combine wing bending with either wing twisting. This work raises the issue of instabilities in fluid-structure coupling for hydrofoils in water. An important difference from the flutter phenomenon in air is the fact that the flexible structure is evolving in heavy fluid; this implies in particular added mass effects and important fluid damping. Flutter appeared for the first time on racing yacht keels with composite fins, so in water, in 2004 :• On the IMOCA 60 feet boat POUJOULAT-ARMORLUX of Bernard Stamm during the transatlantic race 'TheTransat': he lost his keel and capsized.• On the IMOCA 60 feet boat SILL Rolland JOURDAIN: the keel and the boat were saved. Following these problems - particularly following the loss of the keel of Bernard Stamm sail-boat, accident that could have dramatic consequences for the skipper - HDS company, which is is a major player in the design of complex composite parts, especially for racing sailing yachts, focused on the phenomenon. Flutter has occurred only for canting keels with composite fins on IMOCA 60 feet and VOLVO 70 feet racing yacht. The main questions asked are “Why are composite keels susceptible to flutter, and is it possible to predict and prevent this behaviour?”, then “Can a fair indication of the flutter critical speed of the keel be given at low cost? ». This work presents the analytical, experimental and numerical methods implemented to estimate the critical speed for different types of appendages in water. Models, based on a truncated modal for most energetic modes which are generally, for a bulb keel, the lateral bending predominant mode and the torsion predominant mode, are developed and a design tool is proposed. One of the requirements of this work was to make a simple tool to integrate the calculation of the flutter critical speed in the first design loops of a composite or steel keel. The proposed models that give good results for both flutter cases on keels presented above, are confronted with experimental results and with fluid-structure multiphysic simulations performed using the software ADINA. Finally, a parametric study is proposed with the influence of the main design parameters on the occurrence of instabilities.

Richtmyer-Meshkov instability is studied in shock tube experiments with an Atwood number of 0.7. The interface is formed in a vertical shock tube using opposed gas flows, and three-dimensional random initial interface perturbations are generated by the vertical oscillation of gas column producing Faraday waves. Planar Laser Mie scattering is used for flow visualization and for measurements of the mixing process. Experimental image sequences are recorded at 6 kHz frequency and processed to obtain the time dependent variation of the integral mixing layer width. Measurements of the mixing layer width are compared with Mikaelian's [1] model in order to extract the growth exponent. where a fairly wide range of values is found varying from theta approximate to 0.2 to 0.6.

Une étude expérimentale a été menée sur le sillage d’un cylindre circulaire stratifié dans le but de décrire l’effet de l’inclinaison et de la stratification sur les instabilités tridimensionnelles d’un sillage. Principalement 4 modes instables ont été trouvés.Ces résultats mettent ainsi en évidence le fait que la stratification et l’angle d’in- clinaison modifient fortement la transition d’un écoulement 2D à 3D du sillage d’un cylindre avec la présence de nouveaux modes instables. / An experimental study has been delineated of a stratified wake of a circular cy- linder in order to describe the effect of the tilt and of the stratification on the 3D instabilities of the wake. Principally, four unstable modes have been figured out.These results highlight the fact that the stratification strongly modifies the transition from a 2D to a 3D flow in a cylinder wake, with the presence of new unstable modes. The tilt of the cylinder with respect to the vertical plays a major role, such that a study limited to a vertical or a horizontal cylinder misses a lot of the rich dynamics of the tilted wake.

Small solids embedded in gaseous protoplanetary disks are subject to strong dust-gas friction. Consequently, tightly coupled dust particles almost follow the gas flow. This near conservation of the dust-to-gas ratio along streamlines is analogous to the near conservation of entropy along flows of (dust-free) gas with weak heating and cooling. We develop this thermodynamic analogy into a framework to study dusty gas dynamics in protoplanetary disks. We show that an isothermal dusty gas behaves like an adiabatic pure gas, and that finite dust-gas coupling may be regarded as effective heating/cooling. We exploit this correspondence to deduce that (1) perfectly coupled, thin dust layers cannot cause axisymmetric instabilities; (2) radial dust edges are unstable if the dust is vertically well-mixed; (3) the streaming instability necessarily involves a gas pressure response that lags behind dust density; and (4) dust-loading introduces buoyancy forces that generally stabilize the vertical shear instability associated with global radial temperature gradients. We also discuss dusty analogs of other hydrodynamic processes (e.g., Rossby wave instability, convective overstability, and zombie vortices) and how to simulate dusty protoplanetary disks with minor tweaks to existing codes for pure gas dynamics.

Massive circumstellar disks are prone to gravitational instabilities, which trigger the formation of spiral arms that can fragment into bound clumps under the right conditions. Two-dimensional simulations of self-gravitating disks are useful starting points for studying fragmentation because they allow high-resolution simulations of thin disks. However, convergence issues can arise in 2D from various sources. One of these sources is the 2D approximation of self-gravity, which exaggerates the effect of self-gravity on small scales when the potential is not smoothed to account for the assumed vertical extent of the disk. This effect is enhanced by increased resolution, resulting in fragmentation at longer cooling timescales beta. If true, it suggests that the 3D simulations of disk fragmentation may not have the same convergence problem and could be used to examine the nature of fragmentation without smoothing self-gravity on scales similar to the disk scale height. To that end, we have carried out local 3D self-gravitating disk simulations with simple beta cooling with fixed background irradiation to determine if 3D is necessary to properly describe disk fragmentation. Above a resolution of similar to 40 grid cells per scale height, we find that our simulations converge with respect to the cooling timescale. This result converges in agreement with analytic expectations which place a fragmentation boundary at beta(crit) = 3.