Plastic injection moulding machines are a positive example of the possibilities in terms of performance and energy efficiency of modern hydraulic drives technology. In addition to the performance and energy efficiency of the machines, the quality of the plastic mouldings and an easy to use machines control is the focus. To ensure a constant plastics part quality the set process parameters of the injection moulding machines are kept constant by appropriate closed loop control strategies today. Assuming a constant quality of the processed plastic raw material, this strategy is effective. If it comes to a qualitative variation in the processed plastics, which often leads to a change in viscosity of the plastics melt, keeping processing parameters constant will not lead to a constant quality of the moulded parts. The deviations in the plastics viscosity have such a great influence on the moulding process that the relevant process parameters have to be adjusted manually in many cases. Often the stroke of the reciprocating screw system has to be adapted to reach a constant filling volume of the cavity and therefore avoid burr formation or short shots. In this paper an approach for adaptive process control is introduced. This control loop is able to correct the set points of specific machines parameters online within the production cycle and therefore is able to avoid changes in the produced parts quality.

La spintronique dans les semiconducteurs vise à utiliser le spin de l’électron comme degré de liberté supplémentaire (en plus de la charge électrique) afin de véhiculer l’information, ce qui permettrait la mise au point de composants intégrant de nouvelles fonctionnalités. Ce travail de thèse porte sur deux étapes importantes qui doivent être maîtrisées : l’injection électrique de porteurs polarisés en spin dans les semiconducteurs III-V, et la manipulation du spin de l’électron (par champ magnétique) dans ces matériaux optimisés. Dans un premier temps, la grande efficacité des injecteurs de spin à base de CoFeB/MgO/GaAs est démontrée dans des dispositifs de type Diodes Electroluminescentes polarisées en spin (SpinLEDs). La comparaison entre des injecteurs comprenant une barrière tunnel fabriquée soit par pulvérisation cathodique, soit par épitaxie par jets moléculaires (MBE), permet de montrer que ces deux techniques donnent des résultats comparables. Dans les deux cas, l’efficacité de l’injection est améliorée par un recuit de l’échantillon autour de 300−350◦C. Le recuit induit une amélioration de la qualité de l’interface CoFeB/MgO. De plus, l’efficacité de l’injection de spin n’est stable en fonction du courant injecté que lorsque la barrière tunnel est fabriquée par pulvérisation cathodique. Ceci est dˆu aux caractéristiques de l’interface MgO/GaAs qui diffèrent selon la technique de croissance de la barrière. Dans un deuxième temps, l’injection de spin en l’absence de champ magnétique externe appliqué est réaliséegrâce à un nouveau type d’injecteur constitué d’une électrode de CoFeB ultrafine présentant une aimantation rémanente de la couche le long de l’axe de croissance de l’échantillon. Pour la première fois des taux de polarisation circulaire de l’électroluminescence de l’ordre de 20% sont mesurés à 25 K à champ magnétique nul. Ensuite, la problématique de la relaxation de spin des porteurs injectés dans les vallées L de haute énergie dans GaAs (phénomène non négligeable sous injection électrique) est également traitée. Nous observons qu’une fraction de la mémoire du spin photogénéré en L est conservée lorsque les électrons sont diffusés vers la vallée Γ, malgré une relaxation d’énergie de plusieurs centaines de meV. Le temps de relaxation de spin dans les vallées L est estimé autour de 200 fs. Enfin, nous avons exploré le matériau GaAsBi dilué (x ∼ 2.2%) dont la perturbation de la matrice par l’élément Bi permet d’attendre des propriétés électroniques et de spin fortement modifiées. Des mesures de photoluminescence ont mis en évidence une diminution de l’énergie de bande interdite de l’ordre de 85meV/%Bi. De plus, par la mesure directe des battements quantiques de la polarisation de photoluminescence nous avons déterminé un facteur de Landé des électrons de conduction de l’ordre de deux fois supérieur à celui de GaAs. Ces résultats témoignent de la forte perturbation des états de valences et de l’augmentation de l’interaction spin-orbite / Spintronics of semiconductors aims at using carrier spins as supplementary means of information transport. Thiswould lead to components showing extended functionalities. This thesis work is dedicated to the study of injectionand manipulation of electron spin in semiconductors, which are the basis of any spintronic application. In a first stepwe demonstrate the high efficiency of CoFeB/MgO/GaAs - based spin injectors. Circular polarization degrees of electroluminescence over 20% are measured on spin polarized LEDs (SpinLEDs) at 0.8 T and 25 K. Comparison betweensputtering- and MBE- grown spin injectors has shown similar results. In both case, spin injection efficiency is increasedby thermal annealing of the sample, in the range 300 − 350◦C. Indeed, annealing improves the quality of CoFeB/MgOinterface, and induces the crystallization of CoFeB above 300◦C. A higher stability of spin injection with current injectionis found when the tunnel barrier is grown by sputtering. This is due to the MgO/GaAs interface characteristicswhich is related to the growth technique. In a second step, we demonstrate spin injection without external appliedmagnetic field, through an ultra-thin (a few atomic layers) CoFeB electrode, taking advantage of the perpendicular magnetic anisotropy of the layer which leads to a remanant magnetization along the growth axis. For the first time in this configuration, circular polarization degrees of electroluminescence of about 20% are measured at 25 K at zero magnetic field. In a third step, due to the crucial role it may play in electrical injection, electron spin dynamics in high energy L-valleys is investigated. Using polarization resolved excitation photoluminescence in the range 2.8-3.4 eV, we observe that a fraction of photogenerated spin polarization is preserved when electrons are scattered hundreds of meV down to Γ valley. Spin relaxation time in L valleys is estimated to 200 fs. Finally we investigate electron and spin properties of GaAsBi dilute bismide alloy. We observe that the bandgap energy is reduced by 85meV/%Bi when Bi element is introduced into GaAs matrix. Moreover, the electron Land´e factor is about twice the one in GaAs for a 2.2% Bi composition. These features are evidence of the strong perturbation of host states and spin-orbit interaction enhancement

Cette étude se focalise sur des matrices polyamide 66 vierges, mises en œuvre sous la forme de plaques injectées. Son objectif est la mise en évidence des corrélations qui existent entre l’organisation microstructurale de ces matrices et leur comportement mécanique.L’étude se limite dans un premier temps à l’analyse d’une seule plaque. Une cartographie des états microstructuraux a été réalisée, et dévoile une organisation typique du procédé d’injection moulage, avec des évolutions dans l’épaisseur ainsi que dans la direction de l’écoulement. Ainsi, on peut différencier les régions superficielles, en contact avec le moule, dans lesquelles la cristallisation a été impactée par des gradients thermiques et des contraintes de cisaillement relativement élevés, a l’inverse des régions profondes qui ont pu cristalliser dans des conditions plus stables. Les évolutions dans la direction de l’écoulement sont plus subtiles et semblent relever du transport de matière ou des différences de pression durant le procédé d’injection.Ces différences sur la microstructure se traduisent sur le plan du comportement mécanique et des évolutions dans les mêmes directions ont pu être observées. Il apparait que la dimension des entités cristallines, leur degré de perfectionnement, le niveau de contrainte de l’amorphe ou encore le taux de cristallinité impactent tous les propriétés mécaniques, même s’il les interrelations sont complexes et difficiles à discriminer.L’influence de l’humidité sur le matériau polyamide 66 a été observée et discutée. En conditionnant le positionnement du matériau par rapport à sa transition vitreuse, la teneur en eau pilote la façon dont la microstructure intervient dans les mécanismes de déformation du polymère. Sa prise en compte est donc essentielle.Enfin une extension de l’étude à des plaques de conditions de mise en œuvre différentes a été menée. Bien que leur impact sur les propriétés étudiées se sont avérées minimes, les évolutions observées ont permis d’affiner l’analyse et de confirmer certains points. / This study focuses on polyamide 66 matrices, formed as injection molded plates. The aim of the study is to highlight existing interrelations between microstructural organization and mechanic behavior.At first, the analysis is limited to one plate. A microstructural mapping has been done, and shows a typical injection molded organization, with evolutions in thickness and injection directions. Therefore, some areas can be differentiated, the superficial ones, those in contact with the mold, where the crystallization is induced by thermal gradient and shear stresses quite important, contrary to bulk areas which have known more stable conditions. Evolutions along the injection direction are more subtle and seem to be induced by the transportation of matter or different pressure conditions during the injection process.These microstructural differences are reflected on the mechanical plan and evolutions in the same directions have been observed. It appears that crystallites dimensions, perfection levels, constraint of amorphous phase or crystallinity ratio all impact on mechanical properties, even if interrelations are complex and not easy to discriminate.The influence of humidity on polyamide 66 material has been observed and discussed. This parameter controls the position of the material relatively to its glass transition and determines the way microstructure is implicated in deformation mechanisms.Finally, the study has been extended to different processing conditions. Although the evolutions that have been noticed were very small, they have helped to refine the analysis and confirm some points.

De nos jours, le LASER reste l'outil le plus efficace et le plus utilisé pour injecter des fautes au sein des micro-contrôleurs sécurisés modernes. Parmi ses principaux avantages nous pouvons citer ses fortes résolutions spatiale et temporelle. Ces avantages ne sont toutefois accessibles qu'au prix d'investissements conséquents en temps et argent avec un coût oscillant entre deux et quatre centaines de milliers d'euros selon la qualité du LASER.Outre ces aspects financiers, la publication par les scientifiques, ainsi que l'intégration par les fabricants de cartes à puce, de contre-mesures efficaces, comme les détecteurs de lumière, ont incité aux développements de techniques d'injection de fautes alternatives et à coûts plus modérés. Parmi ces techniques alternatives, nous trouvons les techniques d'injection de faute(s) par médium électromagnétique qui permettent de perturber le comportement des circuits. C'est dans ce contexte que cette thèse présente les principaux effets de ce type d'injection de fautes en procédant à une première décomposition face avant face arrière, puis à une seconde lié à la forme du signal perturbateur utilisé (harmonique ou pulsé). Nous avons ainsi pu mettre en exergue des effets de coulage avec les lignes de métallisations mais aussi des effets de Forward sur le circuit cible. / Nowadays, LASER remains the tool the most effective and most used to inject faults within the modern secure microcontrollers. Among its main advantages we can quote its strong spatial and temporal resolutions. These advantages are however accessible only to the price of consequent investments in time and money with a cost oscillating between two and four hundreds of thousand euro according to the quality of the LASER. Besides these financial aspects, the publication by scientists, as well as the integration by the manufacturers of smart cards, effective countermeasures, as light detectors, incited to the development of alternative faults injection techniques with moderate costs. Among these alternative techniques, we find the electromagnetic fault injection techniques which allow perturbing the behavior of circuits. It is in this context that this thesis presents the main effects of this kind of fault injection by proceeding to a first decomposition in front side and back side, then in one second bound to the shape of the disturbing signal (harmonic or pulsed) used. So, we highlight coupling effects with metals lines but also a Forward effect on the target circuit.

La consommation des circuits intégrés n'a cessé d'augmenter cette dernière décennie. Avec l'augmentation du prix de l'énergie et la démocratisation des systèmes embarqués, des méthodes permettant de gérer le compromis consommation performance, comme la gestion dynamique de la fréquence et de la tension d'alimentation ou encore du potentiel de substrat, ont été élaborées. Ces méthodes, qui sont de plus en plus couramment mises en œuvre dans les systèmes intégrés, permettent de diminuer la consommation de ceux-ci, et mieux de gérer le compromis consommation performance. Certains de ces circuits, embarquant ces méthodes peuvent avoir à effectuer des opérations traitant des informations confidentielles. Il est donc nécessaire de s'interroger sur l'éventuel impact de ces sur la sécurité des systèmes intégrés. Dans ce contexte, les travaux de thèse reportés dans le présent document, ont eu pour objectif d'analyser la compatibilité de ces méthodes de gestion de la consommation avec la conception de circuits robustes aux attaques matérielles. Plus particulièrement, l'objectif a été de déterminer si ces techniques de conception faible consommation, constituent des obstacles réels ou bien facilitent les attaques matérielles par observation et perturbation exploitant le canal électromagnétique. Dans un premier temps, une étude sur l'efficacité des attaques par observation en présence de gestion aléatoire de la tension, de la fréquence et de la polarisation de substrat a été conduite. Dans un deuxième temps, l'impact de la gestion dynamique des tensions d'alimentation et de substrat sur la capacité à injecter des fautes par médium électromagnétique a été étudié. Ce document présente l'ensemble des résultats de ces analyses.Mots-clés : Attaques Matérielles, Attaques par Canaux Auxiliaires, Attaques par fautes, Canal électromagnétique, DVFS, Body-Biasing. / The consumption of integrated circuits has been increasing over the last decade. With the increase of energy prices and the democratization of embedded systems, methods to manage the consumption performance compromise, such as the dynamic management of the frequency and the supply voltage or the substrate potential, were developed. These methods, which are becoming more commonly implemented in integrated systems, allow to reduce the consumption of those latter, and to better manage the tradeoff between consumption and performance.Some of these circuits, embedding these methods, may have to perform some operations with confidential information. It is therefore necessary to consider the possible impact of these methods on the safety of the integrated systems. In this context, the work reported in this thesis aimed to analyze the compatibility of these methods of power management with the design of robust circuits to physical attacks.Specifically, the objective was to determine whether these low-power techniques constitute real obstacles or facilitate the attacks by observation or perturbation exploiting the electromagnetic channel. Initially, a study on the effectiveness of attacks by observation in the presence of random management of voltage, frequency and substrate polarization was done. Secondly, the impact of the dynamic management of supply voltages and substrate polarization on the ability to inject faults by electromagnetic medium was studied. This document presents the overall results of these analyzes. Keyword : Hardware Attacks, Side Channel Attacks, Faults Attacks, Electromagnetic canal, DVFS, Body-biasing

A injeção de carvão pulverizado (pulverized coal injection - PCI) é uma técnica largamente utilizada nos altos-fornos pelas siderúrgicas brasileiras, seguindo uma tendência mundial, que busca reduzir o consumo específico de coque por tonelada de gusa e, em consequência, do custo do ferro gusa produzido. A combustão do carvão pulverizado ocorre sob pressões médias (de até 4 atm), temperaturas de chama elevadas (em torno de 2 000 °C), altas taxas de aquecimento (104 a 105 °C/s), tempo de residência muito curtos (inferiores a 40 ms), e é seguida pela gaseificação na presença de CO2. Como não existem métodos padronizados para a avaliação das características de combustão dos carvões para sua utilização em PCI, utilizam-se geralmente equipamentos de injeção de carvão em escalas laboratorial e de bancada, além plantas-piloto com este propósito. O objetivo principal deste trabalho é apresentar as etapas do desenvolvimento de um equipamento em escala de bancada realizado no Laboratório de Siderurgia da Escola de Engenharia da UFRGS (LASID-UFRGS), desde o projeto conceitual até a fase pré-operacional, e os resultados iniciais obtidos, para estudar as características de combustão dos carvões em condições bastante similares às que ocorrem nos altos-fornos. Os testes iniciais incluem a avaliação da combustão de um carvão brasileiro objetivando seu uso potencial em PCI, em substituição parcial de carvões importados para este fim. Optou-se por um projeto moderno, altamente automatizado, que inclui aquisição rápida de dados, com escala e conceito operacional adequados para estudos acadêmicos, de configuração vertical e que possibilitasse a operação no modo de injeção de uma amostra de carvão em pulso único, bem como uma adequação futura ao modo de injeção contínua de carvão. Algumas características relevantes do equipamento são a medição e aquisição ultrarrápida de dados termodinâmicos de pressão e temperatura em diversos pontos do sistema, permitindo o registro dos fenômenos transientes que ocorrem durante a combustão, a possibilidade de programação e controle de tempos, pressões e temperaturas para testes de combustão e de pirólise através de rotinas especialmente desenvolvidas para isto, bem como a coleta representativa dos produtos sólidos e gasosos resultantes da combustão para posterior análise. É ainda possível a filmagem da xvi combustão em modo ultrarrápido, permitindo correlacionar as imagens aos dados termodinâmicos registrados durante a combustão ao longo de um segundo, em intervalos de poucos milissegundos. Os resultados iniciais obtidos na fase pré-operacional demonstram o bom funcionamento do sistema, permitindo distinguir claramente a influência da variação de parâmetros operacionais como tipo de carvão, massa e de amostra injetada, pressão e temperatura de operação e composição dos gases oxidantes. / Pulverized coal injection (PCI) is a technique used in blast furnaces (BFs) by Brazilian steel industry, following a worldwide trend, to reduce coke consumption by ton of hot metal produced, and therefore reducing the overall cost. Burning of pulverized coal injected into tuyeres of BFs takes place under medium pressure (up to 4 atm), high flame temperatures (around 2 000 °C), very fast heating rates (104–105 °C/s) and very short residence times, less than 40 ms, followed by gasification in presence of CO2. Since there are no standard tests for evaluation of coal combustibility at PCI conditions, lab and bench scale coal injection rigs and pilot plants are usually employed for this purpose. This work shows the development steps of a bench-scale rig, built at the Iron and Steelmaking Laboratory of the School of Engineering - Universidade Federal do Rio Grande do Sul (LASID-UFRGS), from the conceptual design to the pre-operational step, as well as the initial results. This equipment permits to study the combustion characteristics of coals in conditions very close to those occurring in blast furnaces. Initial tests include the combustion evaluation of a Brazilian coal, aiming its potential usage for PCI, in partial substitution of imported coals for this purpose. The rig has a modern design and is highly automated. Its scale and operational concept is fitted for academic studies. It has a vertical configuration, to be operated with injection of a coal sample in a single pulse mode and is capable to be adjusted afterwards to continuous coal injection mode. Some relevant characteristics of the injection rig are: (1) the very fast measurement and acquisition of thermodynamic data of pressure and temperature in several points of the system, allowing the capture of transient phenomena occurring along the combustion process; (2) the possibility of programming and controlling time intervals, pressures and temperatures to perform combustion and pyrolysis tests, employing specially developed routines; and (3) the representative sample collection of solid and gaseous combustion products to be further analyzed. It is also possible to capture images of the combustion by a high-speed camera, allowing correlate the images, acquired during a time interval of one second, with the thermodynamic data collected in intervals of few milliseconds. xviii The good performance of the equipment was shown by the initial results obtained at the pre-operational phase. The experimental data clearly depicted the effect of operational parameters like coal type, injected sample mass, operational pressure and temperature, and oxidizing gas composition.

An extensive study has been made of the factors affecting sample dispersion in flow injection analysis. This has included the effect of different manifold parameters on peak height and shape, and the use of various flow models to describe the physical processes involved. The factors investigated included sample flow-rate, tube length and diameter, sample volume, and the diffusion coefficient of the solute. The role of the detector has also been investigated, and was found to exert a strong influence on the observed dispersion behaviour of a flow injection manifold. In the light of this, a low volume flow cell was constructed which minimised the contribution of the detector to dispersion. Different modes of sample injection ("time" and "slug" injection) have been studied, and an automatic valve constructed in connection with the work. Both new and existing flow models have been evaluated in terms of their ease-of-use and ability to predict peak shape. Conditions for which the different models were valid were derived both experimentally and from consideration of the assumptions of the models. The applications of such models to the characterisation and design of flow manifolds has also been considered, and this approach compared to the purely theoretical and empirical approaches that have been described in the literature. The models examined in this thesis are the convection, diffusion, tanks-in-series, well-stirred tank, two-tanks in- parallel and two-tanks-in-series models. Finally, the practical implications of this work for the design of flow injection manifolds and the derivation of so-called "rules for dispersion" have been discussed, and suggestions made for further areas of research.

Decades of research on compression ignition engines have led to a highly efficient combustion cycle in contemporary diesel engines. Nonetheless, the combustion process is being studied perpetually to meet both current and future emission regulations. One of the most influential parameters that impacts the combustion quality, is the fuel spray evolution during injection, and subsequent fuel-air mixture formation inside the engine cylinder. The spray evolution has been investigated to a high level of detail, and the highly complex processes of mixture formation and combustion are well-documented for diesel engines. Most of these investigations are limited to studying either research-grade injectors, or brand new production injectors. Injectors in real-world diesel engines, i.e. normal passenger cars and trucks that are used on a daily basis, are however subject to deposit formation at the tip of the injector nozzle. These deposits have the potential of altering the internal nozzle flow and fuel spray pattern, which in turn degrades combustion quality and increases engine emissions. In the work presented in this thesis the spray evolution of production injectors has been studied over a wide range of injector conditions. Common rail light-duty injectors with a usage history of up to 90 000 miles were acquired from the UK commuter car parc, and several brand new injectors were studied for comparison purposes. It is shown that the spray pattern of the injected fuel changes over the lifetime of the injector. For used injectors a reduced penetration rate was observed in the transient regime of fuel injection, during needle lift. The reduced penetration rate was often accompanied by anomalous radial expansions. Although the magnitude of the effects varied from injector to injector, the highest mileage injectors tended to produce the strongest spray deviations. For several high-mileage injectors the end of injection appeared retarded with respect to new injectors. Expulsions of liquid ligaments and droplets after the end of injection were observed from all injectors, irrespective of the mileage of the injector.

Afin d'améliorer la coulabilité de certaines nuances d'acier, du calcium est ajouté pour modifier et contrôler la nature des inclusions d'oxydes et des sulfures présentes dans l'acier. Grâce au traitement au calcium, les inclusions solides provoquant l'obstruction des buses de coulée sont converties en inclusions liquides qui ne bouchent pas les buses. L'injection du fil fourré est devenue l'une des technologies de pointe pour les ajouts de calcium, car elle offre une récupération de calcium plus élevée que d'autres méthodes et son fonctionnement est simple et peu coûteux. Ce travail est divisé en deux axes principaux. La première clarifie les mécanismes de dissolution du fil fourré lors de son injection dans le métal liquide. La deuxième traite du transport du calcium une fois qu'il est libéré dans l'acier, ainsi que de la modification chimique des inclusions par cet élément. Le premier axe s'articule autour de DissolFil, un modèle numérique décrivant les phénomènes thermiques présents au cours du processus d'injection d'un fil fourré dans un bain d'acier liquide. Celui-ci est complété par une étude expérimentale de l'immersion de fils fourrés dans des bains d'acier statiques utile à la validation du modèle. Les résultats de simulation ont mis en évidence l'importance de la vitesse d'injection du fil en tant que paramètre à adapter en fonction de la conception du fil et de la surchauffe du bain d'acier afin d'optimiser la profondeur de libération du calcium dans la poche. Le deuxième axe comprend un modèle de dissolution et du transport du calcium dans une poche d'acier et la modification des inclusions, couplant la mécanique des fluides et la thermodynamique à l'aide des codes de calcul Ansys Fluent (mécanique des fluides) et CEQCSI (thermodynamique). Le modèle a été validé à partir des prélèvements d’acier réalisés lors d’essais industriels spécifiques. Nous montrons que le transfert de masse en couche limite à partir des bulles de calcium vers l'acier liquide est le mécanisme dominant pour la dissolution du calcium. De plus, la modification globale des inclusions solides et liquides dépend fortement du mélange et de l'homogénéisation de la poche et de la composition chimique de l'acier / In order to improve the castability of some steel grades, calcium is added to the liquid steel to modify and control the nature of oxide and sulphide inclusions present in the steel. As a result of the treatment with calcium solid inclusions that provoke clogging of the casting nozzles are converted into liquid ones that do not clog the nozzles. Cored wire injection has matured as one of the leading technologies to perform calcium additions since it offers a higher calcium recovery than other methods and its operation is simple and inexpensive. This work, divided into two main axes, clarifies the mechanisms of dissolution of the cored wire during its injection, the dissolution and transport of calcium once it has been released into the steel ladle, and the chemical modification of inclusions by this element. The first axis revolves around DissolFil, a numerical model describing the thermal phenomena occurring during the process of injection of a cored wire into a liquid steel bath. In parallel, an experimental study of immersion of cored wires into quiescent steel baths that was carried out with the aim of validating the model. The simulation results put into evidence the importance of the wire injection velocity as the operating parameter that needs to be tailored as a function of the wire design and steel bath superheat in order to optimize the release depth of calcium in the steel ladle. The second axis comprises a model of dissolution and transport of calcium in the steel ladle and inclusion modification, coupling fluid-dynamics and thermodynamics by using the Ansys Fluent (fluid-dynamics) and CEQCSI (thermodynamics) calculation codes. The model was validated with data from steel samples obtained in dedicated industrial trials. We highlight boundary layer mass transfer from calcium bubbles into the liquid steel as the dominant calcium dissolution mechanism. Furthermore, the global modification of solid inclusions into liquid ones depends strongly on the mixing and homogenization in the ladle and the steel chemical composition

La croissance et le détachement des bulles formées sur une paroi par injection ou ébullition sont présents dans de nombreuses situationsindustrielles (les contacteurs gaz/liquide, les évaporateurs ou des dispositifs d’impression à jet d’encre). Dans la plupart des applications, les bulles grossissent dans un écoulement cisaillé ou au repos et la plupart des études visent à prévoir le rayon de détachement par une analyse de l’équilibre des forces. Ainsi, une bonne modélisation des forces capillaires et hydrodynamiques est nécessaire pour prédire correctement le rayonau détachement. Dans un premier temps, la croissance quasi-statique d’une bulle d’air formée sur la paroi d’une cuve et d’un canal rectangulaire est étudiée. Trois substrats différents sont utilisés pour observer l’influence des conditions de mouillage. La dynamique de la croissance de la bulle et le détachement sont alors enregistrés différentes configurations : canal horizontal, canal incliné, avec et sans écoulement liquide. Ensuite,le comportement des bulles de vapeurs en croissance sur une paroi en ébullition en vase et convective est étudié. Les expériences sonteffectuées en gravité terrestre et en microgravité dans un liquide sous-refroidi (HFE 7000) à basse pression (1-2bars). Des bulles de vapeurssont nuclées sur la paroi par l’intermédiaire d’un générateur de bulles. Ce dernier chauffe le liquide environnant au-dessus de sa température de saturation. Les bulles croient sur le site de nucléation dans un écoulement cisaillé et se détachent soit directement soit en glissant le long de la paroi. Une copie de la cellule d’essai appelée RUBI développée par l’ESA est implémentée dans le dispositif expérimental pour effectuer lespremières expériences en microgravité avant son lancement dans la station spatiale ISS. Toutes les acquisitions sont filmées à l’aide d’une caméra rapide et l’extraction du contour de la bulle est réalisée par traitement d’images. Différentes forces agissant sur une bulle sont calculées à partir des paramètres géométriques de la bulle. Un modèle prédictif de détachement de bulles est proposé. Enfin, un bilan thermique diphasique est effectué. / The growth and departure of bubbles nucleated on a wall are of particular interest in industrial situations such as gas/liquid contactors in chemicalprocessing, vapor nucleation in evaporators or inkjet printing devices. In most of these industrial applications, the bubbles grow in a shear flow and most studies aim topredict the radius at detachment by a force balance analysis. Thus a good modeling of both the hydrodynamic and capillary forces is needed to predict correctly this radius. First, the quasi-static growth of an air bubble nucleated on the wall of a tank and a rectangular channel is investigated. Three different substrates are used to observe the wettability effect. The dynamics of the bubble growth and detachment is then recorded for different configurations : horizontal channel, inclined channel, with and without a shear flow. Then, behavior of vapor bubbles growing at a wall in pool boiling and flow boiling is investigated. Experiments are performed in normal and microgravity environments with a subcooled test liquid HFE-7000 at low pressure (1-2 bars). Vapor bubbles are nucleated on the wall by a bubble generator that heats the liquid above its saturation temperature. This bubble generator was designed to provide nucleation on an isolated site. These bubbles grow on this nucleation site under an imposed shear flow and depart downstream either along the wall or perpendicularly. A copy of the test cell called RUBI developed by ESA is included in our experiment to perform the first experiments in microgravity before its launch to the ISS. Geometric and kinematic features of the air bubbles and vapor bubblesare measured by processing images obtained by high-speed video recordings. Different forces acting on the bubbles are calculated from these measurements, using a point force approach. The validity of this mechanistic approach to the bubble dynamics is discussed. At last, heat transfer balance is estimated on the bubble.