La problématique des contaminations croisées (transfert inter-lots de micro-ingrédients : additifs et/ou produits médicamenteux) dans le secteur de l'alimentation animale est entrée, depuis quelques décennies, au cœur des préoccupations de la profession. Ce phénomène se décompose en 2 phases : le dépôt de particules par un lot de fabrication et sa récupération par le/les lot(s) suivant(s). Plusieurs études expérimentales, ont incriminé l’élévateur à godets post-mélangeur dans l’augmentation du niveau de contamination d’une ligne de fabrication. Cet appareil de manutention achemine verticalement des mélanges pulvérulents de diverses matières premières, pouvant contenir, en particulier, des additifs ou des produits médicamenteux. L’objectif de ces travaux est de comprendre dans quelles mesures cet appareil est source de transfert inter-lots, et quels paramètres liés au procédé entrent en jeu. Plusieurs outils ont été mis en place et la réalisation d'un plan d’expériences a permis de mettre en lumière l’impact de certains facteurs sur la contamination croisée :d’une part, le mode de vidange (lié à la vitesse de la sangle) en tête de l’élévateur et l’angle de la bavette de jetée agissent sur la quantité de micro-ingrédient déposée lors du passage d’un lot, et d’autre part, l’espace autour des godets, modifie la capacité du lot suivant à collecter les reliquats.Ces informations définissent une position optimale des paramètres de l’élévateur qui, sur pilote, réduit le niveau de contaminations croisées de 9 à 7 %. Enfin, des observations des champs de vitesses pendant la jetée du produit ont apporté de éléments de compréhension sur mouvements d'air et d'aérosols à l'origine des contaminations. / Carry-over of additives and/or medicated products is a common issue in feed industry and, by extension in most of powder handling industries. Currently carry-over rate of a production line can be accurately defined but the causes are not identified yet. It can be broken into 2 phases : firstly, particle deposit during one batch processing and then, their collecting during the following batches. Experimental studies, carried out on industrial sites or on test benches, charged the bucket elevator situated just after the mixer to be responsible for a significant increase of cross contamination rate of industrial feed production lines. Therefore this work focuses on this handling device. It transfers mixing of several raw materials in powdery forms, which may contain micro-ingredients, especially additives or medicated products. The aim of this study is to understand how process operations affect cross contamination rates during bucket elevator handling. A test bench of this handling device, a reference product and laboratory methods have been setup. Moreover, an experimental fractional factorial design highlights the effects of several process parameters : on one hand the discharge phase on elevator head (linked to belt velocity) and the discharge spout angle act on microingredients deposit mass. On the other hand, spacing between buckets and the leg’s inner surface influences micro-ingredients collected mass. Furthermore, ideal position of process parameters has been defined. By this way, cross contamination rate on the test bench has been decreased from 9 to 7 percent. Finally, velocity fields observations during the discharge phase leads to better understanding of how these process parameters influence cross contamination rate.

Le but de cette étude expérimentale est de caractériser la dynamique de particules solides, à flottabilité nulle, incluse dans un écoulement turbulent isotrope en décroissance libre. Les particules utilisées sont de forme sphérique et ont un diamètre de 4 à 5 fois plus grand que l'échelle spatial de Kolmogorov de l'écoulement. De part leur taille, les particules ont également un nombre de Stokes proche de l'unité. On s'attend alors à ce que ces particules aient une dynamique différente de celle du fluide environnant. Dans cette étude, ont se propose de quantifier les différences de vitesses entre les deux phases à l'aide d'une technique de vélocimétrie simultanée / In this experimental study, the focus is made on the characterization of the dynamics of solid neutrally buoyant particles embedded in a freely decaying, nearly isotropic turbulence, with a weak mean flow. The particles are spherical with diameters several times larger than the Kolmogorov scale. The study of this flow configuration is still challenging both theoretically and numerically. Due to large particle sizes, the local flow around particles can not be considered as uniform and due to fluid-particle density ratio of around unity, the history and Basset forces cannot be neglected in comparison with the viscous drag force. Particle equation of motion is then fully non-linear, in contrast to the equation for heavy particles with diameters smaller then the Kolmogorov scale, for which only the Stokes drag is considered. In several experimental and numerical studies, the effect of particle size on velocity and acceleration statistics has been investigated (Homann and Bec 2010 ; Qureshi et al. 2008 ; Ouellette et al. 2008 ; Xu and Bodenschatz 2008). In the case of isotropic turbulence, Homann and Bec (2010) show that while the PDF of the particle velocity normalized by the square root of its variance does not vary with particle size, the variance itself is size dependent. A scaling relation for particle velocity variance has been proposed by using the Faxen correction (Gatignol 1983) which takes into account the non uniformity of the fluid flow at the scale of the particle. The aim of our research is to further study the dependence of particle dynamics on particle size. To that purpose, a turbulence generator has been set-up and the resulting turbulence is characterized. Then the flow was seeded with millimeter sized, neutrally-buoyant particles and the velocity of the two phases have been measured simultaneously. Simultaneous measurements of particle and surrounding fluid velocities show that although the global velocity statistics of the two phases have comparable values, the particles may have different local velocity from the velocity of the neighboring fluid

[pt] Alguns projetos pilotos de injeção de polímeros em campos de produção de petróleo mostraram um incremento na recuperação de óleo, embora os mecanismos que governam a dinâmica do escoamento não são bem compreendidos. Recentes investigações experimentais mostraram que as propriedades viscoelásticas de soluções poliméricas podem alterar o comportamento do escoamento na escala de poros e reduzir a saturação residual de óleo. Para entender esses fenômenos em meios porosos, é importante estudar o escoamento de soluções viscoelásticas através das gargantas dos poros. Este trabalho apresenta um estudo experimental do escoamento de uma solução viscoelástica de PEO (0,1 porcento em peso de óxido de polietileno) de alto peso molecular escoando através de uma constrição, utilizado como modelo de uma geometria de garganta de poro de um meio poroso. Medições de queda de pressão e campos de velocidade do escoamento são obtidos utilizando a técnica de velocimetria por imagem de partículas (Micro-PIV). Experimentos com uma solução viscosa de glicerina (45 porcento em peso de glicerina em água) de viscosidade similar à solução de PEO foram também realizados com a finalidade de estimar os efeitos elásticos da solução de PEO. O escoamento da solução de PEO exibiu uma queda de pressão extra (comportamento não linear) acima de uma condição crítica do escoamento, acima da qual os efeitos elásticos passam a ser preponderantes. Para toda a faixa de vazão explorada, os campos de velocidade da solução de glicerina mostraram um regime de escoamento Newtoniano, enquanto a solução de PEO apresenta instabilidade no escoamento a partir de um número de Weissenberg crítico, coincidindo com o aumento da queda de pressão. Esta instabilidade pode ser relacionada ao aumento da viscosidade extensional na entrada da garganta acima de uma determinada taxa de extensão. Os resultados obtidos indicam a variação do padrão do escoamento da solução polimérica de PEO devido à presença dos efeitos elásticos do polímero, e fornecem informações importantes sobre o comportamento das soluções poliméricas viscoelásticas em um meio poroso e que podem impactar sua utilização na recuperação melhorada de óleo. / [en] Some pilot projects of polymer injection in oil fiel ds have shown an increase in oil recovery, although the mechanisms that govern the flow dynamics are still not well understood. Recent experimental investigations have shown that the viscoelastic properties of polymer solutions may change the pore-scale flow behavior and reduce the residual oil saturation. To understand these phenomena in porous media, it is important to understand viscoelastic flow behavior through the pores-throats. This work presents experimental study of the flow of a high molecular viscoelastic PEO solution (0.1 wt percent Polyethylene Oxide) flowing through a constricted capillary, used as model for a pore-throat geometry of a porous media. Pressure drop measurements are performed and velocity fields are obtained using the micro-particle image velocimetry (Micro-PIV) technique. Experiments with a viscous solution of glycerin (45 wt percent glycerin in water), of similar shear viscosity to the PEO solution were also performed in order to isolate the elastic effects of the PEO solution. The flow of the PEO solution exhibited an extra pressure drop (nonlinear behavior) above a critical flow condition beyond which the elastic forces become relevant. For the entire flow rate range explored, the velocity field of the glycerin solution showed a Newtonian flow regime, while the PEO solution shows instability in the flow above a critical Weissenberg number, coinciding with the onset of the extra pressure drop. This instability in the flow is associated with the high extensional viscosity near the constriction at high enough extension rates. The results show the changes in the flow pattern of the PEO polymer solution due to the presence of the elastic effects of polymer, and provide important information on how viscoelastic polymer solutions behave in a porous media and can impact their use in Enhanced Oil Recovery operations.

<p>A droplet subjected to external aerodynamic disturbances disintegrates into smaller droplets and is known as secondary atomization. Droplet breakup has been studied for low Ohnesorge (<b><i>Oh < </i></b>0.1) numbers and good agreement has been seen amongst researchers. However, when it comes to cases with high the <b><i>Oh</i></b> number, i.e. atomization where the influence of viscosity is significant, very little data is available in the literature and poor agreement is seen amongst researchers. </p> <p> </p> <p>This thesis presents a complete analysis of the modes of deformation and breakup exhibited by a droplet subjected to continuous air flow. New modes of breakup have been introduced and an intermediate case with no droplet fragmentation has been discovered. Further, results are presented for droplet size-velocity distributions. In addition, Digital in-line holography (DIH) was utilized to quantify the size-velocity pdfs using a hybrid algorithm. Finally, particle image velocimetry (PIV) was employed to characterize the air flow in the unique cases where drops exhibited no breakup and cases with multiple bag formation. </p> <p> </p> <p>A droplet subjected to external aerodynamic disturbances disintegrates into smaller droplets and is known as secondary atomization. Secondary breakup finds relevance is almost every industry that utilizes sprays for their application. </p> <p> </p>

La demande croissante de poudres d‟alliages métalliques aux propriétés spécifiques utilisées en particulier en projection thermique et fabrication rapide pousse les chercheurs à améliorer et à optimiser sans cesse les procédés de production de ces poudres. L‟objectif affiché sur ces procédés est de maîtriser à la fois morphologie/ distribution de taille des particules produites et coût de fabrication. Actuellement, la majorité de ces poudres est produite par des procédés d‟atomisation par fluide et essentiellement par le procédé d‟atomisation gazeuse. Parmi ces procédés, le procédé Nanoval utilisant une buse «De Laval» est l‟un des plus performants en termes de distribution granulométrique et de rendement.L‟objectif principal de ce travail de thèse vise à améliorer la compréhension des phénomènes physiques mis en jeu par le procédé Nanoval afin d‟en optimiser le fonctionnement. Deux approches composent ce travail :- une partie numérique de modélisation sous Fluent. Deux modèles ont été étudiés, un modèle monophasique relatif à l‟écoulement gazeux dans l‟unité d‟atomisation (passage de l‟autoclave à la chambre d‟atomisation) et un modèle diphasique relatif à la constriction du filament de métal liquide en sortie de buse de coulée. Cette étude numérique a permis de mettre en évidence l‟effet des paramètres opératoires tels que la pression d‟atomisation et le diamètre de la buse de coulée sur la dynamique du jet de gaz, sur la striction du filament de métal liquide ainsi que les zones de forte pression et de haute vitesse avant, pendant et après la désintégration du filament métallique.- une partie expérimentale pour laquelle la mise en place d‟outils de diagnostic in–situ a été nécessaire pour la caractérisation du procédé en cours de fonctionnement. Trois analyses ont été conduites. La première renseigne de la dynamique du jet d‟atomisation évaluée à partir de mesures de Vélocimétrie par Images de Particules (PIV) à proximité de la sortie de la buse De Laval. La deuxième concerne les caractéristiques à l‟écrasement des particules sur un substrat placé dans la chambre d‟atomisation. La troisième et dernière analyse porte sur les propriétés des particules produites et la comparaison avec la matière récupérée après refroidissement dans l‟autoclave. Différents paramètres opératoires ont été explorés (pression d‟atomisation, diamètre de la buse de coulée, pression dans la chambre d‟atomisation, nature du métal) et reliés à leur influence sur la vitesse et le diamètre des particules. Des relations directes entre les résultats de ces trois analyses ont pu être démontrées ainsi qu‟une bonne adéquation entre résultats expérimentaux et résultats issus de la modélisation. / The growing demand for metal alloy powders with specific properties used in thermal spray application and rapid manufacturing encourages researchers to improve and optimize their manufacturing processes. The aim of these processes is to master both morphology/particle size distribution and manufacturing cost. Today the vast majority of powders are produced by fluid atomization and mainly gas atomization process. Among them, the Nanoval process, consisting of a De Laval nozzle is one of the most outstanding process in terms of granulometric distribution and output.The main objective of this thesis is to improve the understanding of the physical phenomena occurring in the Nanoval process to optimize the way it operates. Two approaches will be developed:- A numerical study using Fluent. The two following models were studied, one monophasic concerning gas flow in the atomization unit form autoclave to atomization chamber and the other, a diphasic model concerning the finest part of the filament in the exit of the melt nozzle. This numerical study has highlighted the effect of parameters such as atomization pressure, nozzle diameter on the gas dynamics, fine filament, high-pressure and high speed areas before, during the process and after the disintegration of the metallic filament.- And an experimental study which required the implementation of the in-situ diagnosticTools to characterize the process under working conditions. Three analyses were carried out. The first concerns the dynamics of the atomization jet from Velocimetry measures by Particle Image Velocimetry (PIV) close to De Laval nozzle exit. The second deals with the characteristics obtained when particles impact the substrate in the atomization chamber. The third describes the particle properties and deals with the comparison with the matter in the autoclave after cooling process. Different operating parameters were explored (atomization pressure, melt nozzle diameter, pressure in the atomization chamber, nature of metal) and linked to their influence on the particle velocity and diameter. Narrow links between the analysis results were demonstrated as well as a good adequacy between experimental and modeling results.

La méthode de mesures par PIV (Particle Image Velocimetry) est une technique pour mesurer un champ de vitesse de manière non intrusive et multipoints. Cette technique utilise l'algorithme de corrélation entre deux images consécutives pour déterminer les vecteurs vitesse. La quantité de calcul requis par cette méthode limite son usage à des traitements en temps différé sur ordinateur. Les performances des ordinateurs demeurent insuffisantes pour ce type d'applications sous contrainte temps réel sur des cadences de données élevés. Au vu de ces besoins, la définition et la conception d'architectures dédiées semblent être une solution adéquate pour atteindre le temps réel. L'évolution des niveaux d'intégration permet le développement des structures dédiées au traitement d'images en temps réel à bas prix. Dans ce travail de thèse, nous nous sommes intéressés à la conception d'une architecture de type SoC (System on-Chip) dédiée aux mesures de paramètres physiques par traitement d'images en temps réel. C'est une architecture hiérarchique et modulaire dédiée à des applications de type flot de données d'entrée dominant. Cette description hiérarchique permet la modification du nombre et/ou de la nature de ces éléments sans modifier profondément l'architecture. Pour le calcul d'une mesure, il faut 267 s avec un FPGA à 50 MHz. Pour estimer les performances du système, un imageur CMOS a été connecté directement au FPGA. Les avantages de ce prototype sont de réduire au minimum le mouvement de grands ensembles de données ainsi que la latence en commençant à traiter des données avant leur complète acquisition.

A series of laboratory experiments were conducted at three fetches of 4.8, 8.8 and 12.4 m, and at six wind speeds ranging from 4.1 to 9.6 m/s at each fetch in a wind-wave-current research facility. In addition, five surfactant-influenced experiments were conducted at concentrations ranging from 0.1 to 5.0 ppm at a wind speed of 7.9 m/s and a fetch of 4.8 m. The goals were to examine the momentum transfer and to characterize the turbulent flow structure beneath wind waves, and to investigate the relationship between wind waves and the gas transfer rate at the air-water interface. Digital particle image velocimetry (DPIV) was used to measure two-dimensional instantaneous velocity fields beneath the wind waves. The friction velocities and roughness lengths of the coupled boundary layers were used to characterize the flow regime and momentum transfer. The air-side flows were found to be aerodynamically rough and the water-side flows were found to be in transition and then become hydrodynamically smooth as wind speed increased. Airflow separation from the crests of breaking waves may be responsible for making the air-side boundary layer rougher and water-side boundary layer smoother. Momentum transfer was studied by examining the partitioning of the wind stress into the viscous tangential stress and wave-induced stress. It was found that the wave steepness was the most important wind-wave property that controls the momentum transfer in the coupled boundary layers. Two distinct layers were observed in the near-surface turbulence in the presence of a surfactant and three layers in clean water. In the surfactant-influenced experiments, the energy dissipation rate decayed as zeta^(-0.3) in the upper layer and in the lower layer energy dissipation rate decayed as zeta^(-1.0) similar to a wall-layer. For clean experiments, the energy dissipation rate could be scaled using the depth, friction velocity, wave height and phase speed as proposed by Terray et al. (1996) provided that layer based friction velocities were used. In the upper layer, the near-surface turbulence was dominated by wave-induced motions and the dissipation rates decayed as zeta^(-0.2) at all fetches. Below this in the transition layer turbulence was generated by both wave-induced motions and shear currents and the dissipation rate decayed as zeta^(-2.0) at a fetch of 4.8 m. However, at fetches of 8.8 and 12.4 m, the dissipation rate decayed at two different rates; as zeta^(-2.0) in the upper region and as zeta^(-4.0) in the lower region. In the third layer, the dissipation rate decayed as zeta^(-1.0) similar to a wall-layer at a fetch of 4.8 m. Four empirical relationships commonly used to predict the gas transfer rate were evaluated using laboratory measurements. The gas transfer rate was found to correlate most closely with the total mean square wave slope and varied linearly with this parameter. The three other parameterizations using wind speed, wind friction velocity and energy dissipation did not correlate as well. / Water Resources Engineering

Dans cette étude, l'intérêt est porté sur l'effet des conditions opératoires sur l'enrobage en voie sèche de grosses particules " hôtes " par de fines particules " invitées " et aussi sur la modélisation de cet enrobage selon la méthode par éléments discrets (DEM) afin de mieux comprendre les phénomènes mis en jeu. Dans ce travail, les matériaux choisis comme particules hôtes sont les Suglets® et les particules invitées sont en stéarate de magnésium (StMg). Ces deux éléments sont introduits dans un mélangeur à fort taux de cisaillement, le Cyclomix. Les propriétés du produit final, comme la coulabilité, la mouillabilité et le degré d'avancement de l'enrobage, ont été caractérisés. La variation des propriétés est étudiée en fonction de la durée de traitement dans le Cyclomix pour diverses vitesses de rotation, taux de remplissage et rapport de taille de particules hôte et invitée. La coulabilité a été améliorée avec la durée du mélange ou la vitesse de rotationLe degré d'avancement présente une tendance semblable pour différentes conditions opératoires. Sa variation est représentée par une loi exponentielle empirique en fonction du temps de mélange, paramétrée par une constante ajustable. Cette constante permet d'estimer l'efficacité de l'enrobage. La simulation des mouvements de particules dans le mélangeur par DEM a permis d'obtenir des informations sur la position, la vitesse des particules, et d'autres paramètres énergétiques. Les champs de vitesse réelle ou numérique liés aux mouvements de particules, analysés par PIV (Particule Image Velocimetry), sont analogues. La constante d'enrobage dépend de la vitesse de rotation simulée et peut être ainsi prédite par DEM.

Bioreactors are devices used for the growth of tissues in a laboratory environment. They exist in many different forms, each designed to enable the production of high-quality tissues. The dynamic environment within bioreactors is known to significantly affect the growth and development of the tissue. Chondrocytes, the building blocks of articular cartilage, for example, are stimulated by mechanical stresses such as shear, as compared with those in tissues grown under static incubation conditions. On the other hand, high shear can damage cells. Consequently the shear-stress level has to be controlled in order to optimize the design and the operating conditions of bioreactors. Spinner flasks have been used for the production of articular cartilage in vitro. Assuming the existence of a relation between the cellular glycosaminoglycan (GAG) synthesis and the local shear stresses on the construct surfaces, this research focuses on the development of a model for cartilage growth in such devices. The flow produced in a model spinner flask is characterized experimentally using particle-image velocimetry (PIV). A computational fluid dynamic (CFD) model validated with respect to the laboratory measurements is constructed in order to predict the local shear stresses on the construct surfaces. Tissue growth experiments conducted in the prototype bioreactor permit construct histologies and GAG contents to be analyzed and then correlated with the shear-stress predictions. The integration of this relation into the CFD model enables the prediction of GAG synthesis through convective effects. Coupling this convective model to an existing diffusive model produces a complete cartilage-growth model for use in aiding the optimization of existing bioreactors, and in the design of new ones.

In this study the dynamical modeling of the unsteady flow over a flapping wing is considered. The technique is based on collecting instantaneous velocity field data of the flow using Particle Image Velocimetry (PIV), applying image processing to these snapshots to locate the airfoil, filling the airfoil and its surface with proper velocity data, applying Proper Orthogonal Decomposition (POD) to these post-processed images to compute the POD modes and time coefficients, and finally fitting a discrete time state space dynamical model to the trajectories of the time coefficients using subspace system identification (N4SID). The procedure is applied using MATLAB for the data obtained from NACA 0012, SD 7003, elliptic airfoil and flat plate, and the results show that the dynamical model obtained can represent the flow dynamics with acceptable accuracy.