Orientador: Antonio Carlos Bannwart / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica e Instituto de Geociências / Made available in DSpace on 2018-08-23T09:29:35Z (GMT). No. of bitstreams: 1 Daun_LeandroGalvao_M.pdf: 5084421 bytes, checksum: 56e4661dc939fdef8c0383b09f63721f (MD5) Previous issue date: 2013 / Resumo: Os desafios tecnológicos na produção de petróleo e gás têm aumentado consideravelmente nos últimos anos devido à busca por segurança, redução de custos e otimização da produção, como indicado pelo grande número de estudos sobre a garantia de escoamento. Uma grande preocupação está relacionada ao escoamento multifásico em equipamentos submarinos, onde podem ocorrer diferentes padrões de escoamento e fenômenos termo-fluido-dinâmicos. O presente trabalho tem por objetivo avaliar experimentalmente possíveis fenômenos de garantia de escoamento na entrada de um Sistema de Bombeio Centrifugo Submerso em Skid (S-BCSS), como o instalado no campo de Espadarte na costa brasileira. A atenção foi dada a possíveis sinais de surging da BCS, devido à presença de gás na fase líquida e mudança na orientação do escoamento a partir da horizontal para a vertical descendente, a montante da entrada da bomba. As condições de transição entre os padrões de escoamento "bolhas dispersas" e "intermitente" foram investigados em uma linha experimental com configuração do tipo U horizontal com braços superior, inferior e curvas de 90 graus. A tubulação da linha possui diâmetro interno de 60 milímetros e comprimento total de 32 metros, sendo testada com ar e água em uma ampla gama de composições de mistura permitindo a comparação com mapas de fluxo horizontal e vertical descendente encontrados na literatura. O comportamento dos limites de padrões de escoamento foi experimentalmente avaliado, indicando possibilidades de acumulo de gás na curva superior devido à mudança de padrão "bolhas dispersas" para "intermitente" quando há alterações de orientação do escoamento horizontal para vertical descendente. A existência de uma velocidade mínima da mistura capaz de assegurar o arrasto das bolhas pelo líquido foi também investigada. Diferentes correlações para a velocidade de deslizamento e parâmetro de distribuição foram testadas para o cálculo da fração de vazio. Isto foi realizado através da obtenção da velocidade média do gás local, por meio da obtenção de imagens em alta velocidade seguido de um processamento digital das mesmas. A fração de vazio foi calculada a partir da relação entre as velocidades superficiais e locais do gás. Através da comparação entre 2124 pontos experimentais, foi possível obter as melhores correlações de fração de vazio para diferentes intervalos, resultando na obtenção de diferentes valores de velocidades mínimas de mistura / Abstract: The technological challenges in the oil and gas production have increased considerably in recent years due to the search for safety, cost reduction and production optimization, as indicated by the large number of studies on flow assurance. One major concern is the multiphase flow in subsea equipment where different flow patterns and thermo-fluid-dynamic phenomena may occur. The present work is aimed at evaluating experimentally possible flow assurance phenomena at the inlet of a Subsea Skided ESP (mudline ESP) such as one installed at Espadarte field offshore Brazil. Attention was paid to possible surging of the pump due to the presence of gas bubbles within the liquid and change in flow orientation from horizontal to vertical downward upstream the ESP intake. First, the conditions at the transition between dispersed bubbles and intermittent flow were investigated in a horizontal U-type tube with top and down arms and 90-degree bends. A flow circuit of 60mm i.d. size and 32m was built and tested using air and water in a wide range of mixture compositions, allowing comparisons with horizontal and downward flow maps found in the literature. The behavior of flow pattern boundaries was experimentally evaluated and indicated no possibility of gas accumulation on the upper bend due when the flow pattern changes from dispersed bubbles to intermittent as the flow orientation changes from horizontal to downward. The requirement of a minimum mixture velocity capable of ensuring enough drag of the bubbles by the liquid was also investigated. Several different correlations for drift velocity and distribution parameter were tested for calculation of the void fraction. This was achieved by measuring the local average gas velocity using a high speed imaging and digital image processing. The void fraction was calculated from the ratio between superficial and local gas velocities. From comparison with 2,124 data points, the best void fraction correlation for each range was determined, from which reliable values for the minimum mixture velocity were determined / Mestrado / Explotação / Mestre em Ciências e Engenharia de Petróleo

Orientador: Milton Mori / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Química / Made available in DSpace on 2018-08-21T06:39:27Z (GMT). No. of bitstreams: 1 GonzalezSilva_German_D.pdf: 4427269 bytes, checksum: bb45c122c3faad0613f060ad4cbd3b61 (MD5) Previous issue date: 2012 / Resumo: O objetivo principal desta tese é propor uma metodologia para aplicar simulação de grandes escalas (LES) em uma unidade de craqueamento catalítico industrial. Para atingir este objetivo, iniciou-se propondo uma metodologia para a construção da malha, sendo uma malha quase-uniforme. As malhas propostas foram implementadas para minimizar o esforço computacional e procedimento para a refinação uniforme no domínio do sistema. Inicialmente, foi estudada a fluidodinâmica de um leito fluidizado gás-sólido na escala de laboratório sem reação química, utilizando Simulação de Grandes Escalas. Com base nos resultados numéricos obtidos a partir de resultados da escala de laboratório foi aplicado o modelo cinético de 4 lump em uma simulação CFD tridimensional para um reator industrial FCC, utilizando LES para a fase gasosa, e considerando o catalisador como uma fase contínua (Euleriana). Os pacotes de simulação utilizados foram Ansys ICEM versão 13 para a construção da malha e Ansys CFX versão 13 para o pós-processamento dos resultados. No pós-processamento dos resultados foi proposta uma metodologia para determinar as médias azimutais das variáveis em planos perpendiculares ao escoamento e finalmente foram validadas as simulações com dados reportados na literatura. As principais conclusões do trabalho foram que foi possível diminuir o tamanho da malha e o tempo de processamento. Notou-se também que, ao usar malhas com discretização quase-uniforme não foi necessário fazer um refinamento de malha elevado, nem refinar perto da parede para o sistema gás-sólido / Abstract: The main objective of this thesis is to propose a methodology in how to apply Large Eddy Simulation (LES) on a unit of catalytic cracking. In order to achieve this, it was proposed an alternative way of constructing the computational mesh, by using quasi-uniform meshes. The proposed meshes were implemented to minimize the computational effort and procedure for refining them in the entire domain of the system. Initially it was studied the fluid dynamics of a lab scale gas-solid system without chemical reaction, using Large Eddy Simulation. Based on the numerical results obtained from lab scale results it was implemented the 4 lump kinetic model in a three dimensional CFD simulation of an FCC industrial reactor, using LES for the gas phase and considering the catalyst as a continuous phase (Eulerian). The simulation packages used were Ansys ICEM, version 13 for mesh construction and Ansys CFX 13 for computation and post-processing of the results. In the data post-processing it was proposed a methodology for calculating average values of fluctuating variables between two circular sections in the azimuthal direction. The results were compared with data reported in literature. The main conclusions of the results showed that it was possible to decrease the mesh size and the computational time. It was also noticed that by using quasi-uniform discretization it was not necessary to make a high mesh refinement near the wall for a gas-solid system / Doutorado / Desenvolvimento de Processos Químicos / Doutor em Engenharia Química

Cette thèse est consacrée au mécanisme conduisant à des taux de collisions importants dans les suspensions turbulentes de particules inertielles. Le travail a été effectué en suivant numériquement des particules, par simulations directes des équations de Navier–Stokes, et également par étude de modèles simplifiés. Les applications de ce domaine sont nombreuses aussi bien dans un contexte industriel que naturel (astrophysique, géophysique). L’approximation des collisions fantômes (ACF), souvent utilisée pour déterminer les taux de collision numériquement, consiste à compter dans une simulation, le nombre de fois que la distance entre les centres de deux particules devient plus faible qu’une distance seuil. Plusieurs arguments théoriques suggéreraient que cette approximation conduit à une surestimation du taux de collision. Cette thèse fournit non seulement une estimation quantitative de cette surestimation, mais également une compréhension détaillée des mécanismes des erreurs faites par l’ACF. Nous trouvons qu’une paire de particules peut subir des collisions répétées avec une grande probabilité. Ceci est relié à l’observation que, dans un écoulement turbulent, certaines paires de particules peuvent rester proches pendant très longtemps. Une deuxième classe de résultats obtenus dans cette thèse a permis une compréhension quantitative des très forts taux de collisions souvent observés. Nous montrons que lorsque l’inertie des particules n’est pas très petite, l’effet « fronde/caustiques », à savoir, l’éjection de particules par des tourbillons intenses, est responsable du taux de collision élevé. En comparaison, la concentration préférentielle de particules dans certaines régions de l’espace joue un rôle mineur. / This thesis is devoted to the mechanisms leading to strong collision rates of inertial particles in turbulent suspensions. Our work is based on simulating the motion of particles, using both direct numerical simulations of the Navier–Stokes equations, and a simpler model (kinematic simulations). This subject is important for many applications, in industrial as well as natural (astrophysical, geophysical) contexts. We revisit the ghost collision approximation (GCA), widely used to determine the rate of collisions in numerical simulations, which consists in counting how many times the centers of two particles come within a given distance. Theoretical arguments suggested that this approximation leads to an overestimate of the real collision rate. This work provides not only a quantitative description of this overestimate, but also a detailed understanding of the error made using the GCA. We find that a given particle pair may undergo multiple collisions with a relatively high probability. This is related to the observation that in turbulent flows, particle pairs may stay close for a very long time. We have provided a full quantitative characterization of the time spent together by pairs of particles. A second class of results obtained in this thesis concerns a quantitative understanding of the very strong collision rates often observed. We demonstrate that when the particle inertia is not very small, the “sling/caustics ” effect, i.e., the ejection of particles from energetic vortices in the flow, is responsible for the high collision rates. The preferential concentration of particles in some regions of space plays in comparison a weaker role.

Le travail présenté concerne le développement d’un cadre d’applications pour le traitement d’écoulements dispersés, tenant compte de l’effet de la coagulation sur la distribution des tailles des particules. Nous explorons également quelques techniques de validation expérimentale. Les modèles sont valables pour un écoulement incompressible et isotherme, avec des particules qui ont un temps de relaxation faible en comparaison à celui du fluide. Pour la phase dispersée, une méthode eulérienne est utilisée, ce qui permet d’extrapoler la vitesse des particules de celle du fluide. La distribution des tailles est modélisée à l’aide du « Direct Quadrature Method of Moments ». Cette approche permet de résoudre des équations de transport pour les poids et les abscisses d’une approximation de la distribution à l’aide des fonctions Dirac delta. L’effet de la coagulation est pris en compte à l’aide d’un noyau de collisions qui utilise la vitesse instantanée du fluide. Toutes les équations de transport sont résolues à l’aide de la méthode des éléments finis. Pour le fluide, les stabilisations « Streamline Upwind » et « Pressure Stabilized Petrov-Galerkin » sont utilisées ensemble avec une stabilisation grad-div. Afin de limiter le temps de calcul pour une simulation directe, une formulation utilisant un traitement explicite des termes d’advection est proposée. Avec l’apparition de gradients élevés, les équations de transport pour les particules nécessitent une stabilisation supplémentaire. Tout le travail est disponible dans le projet de logiciel libre Coolfluid 3, en utilisant un langage spécifique permettant une implémentation directe pour des modèles en éléments finis. Le code qui en résulte ressemble à la forme variationnelle des équations utilisées. Le programme est générique en termes de dimensions spatiales et de type d’éléments. Une première validation utilise des résultats trouvés dans la littérature comme référence. La précision des méthodes est vérifiée à l’aide des vortices Taylor-Green. Pour l’écoulement et les concentrations des particules, une simulation directe d’un canal turbulent est effectuée. Le noyau de coagulation est vérifié à l’aide de particules de différentes tailles qui tombent à travers un vortex Burgers. Finalement, quelques techniques de validation expérimentale sont utilisées dans une cellule d’essai. La technique « Particle Image Velocimetry » est utilisée pour les vitesses du fluide, tandis que la distribution des tailles est mesurée à l’aide du « Phase Doppler Anemometry » et « Multiple Wavelength Light Extinction ». La technique d’extinction de lumière est capable de produire des distributions des tailles qui peuvent être comparées facilement avec les résultats numériques. / The present work deals with the development of a framework for the modeling of dispersed flows, including the effect of coagulation on the particle size distribution. We also explore some techniques for experimental validation. Models are developed for incompressible, isothermal flow containing particles that have a small relaxation time compared to the fluid time scale. For the dispersed phase, an equilibrium Eulerian approach is used, extrapolating the particle velocity from the fluid velocity. The size distribution is modeled using the Direct Quadrature Method of Moments. In practice, this results in solving transport equations for the weights and abscissa of a Dirac delta approximation of the size distribution. To model the effect of coagulation, a collision kernel that makes use of the resolved instantaneous velocity is developed. All transport equations are solved using the Finite Element Method. For the fluid, the Streamline Upwind and Pressure Stabilized Petrov-Galerkin method are used, with additional grad-div stabilization. To decrease the solution time for DNS, a segregated formulation with an explicit advection term is proposed. The particle transport equations require cross-wind diffusion in addition to the streamline upwind stabilization when large gradients occur. All work is available in the open source Coolfluid 3 framework, using an Embedded Domain Specific Language we developed for the implementation of finite element models. The resulting code closely resembles the variational form of the equations and is generic in terms of element type and the number of spatial dimensions. A first validation uses literature results as reference. Correctness and accuracy of the methods are verified using the Taylor-Green vortex flow. For the fluid and particle concentration, direct numerical simulation of a turbulent channel flow is performed. The particle coagulation kernel is tested using particles of different sizes falling through a Burgers vortex. Finally, some experimental validation techniques are used on a small test chamber. Particle image velocimetry is used for the fluid motion, while the size distributions are measured using Phase Doppler Anemometry and Multiple Wavelength Light Extinction. The light extinction technique was found to produce size distributions that could provide valuable reference data for our particle model.

La présente étude est dédiée à la compréhension des mécanismes de malaxage intervenant lors du soudage de matériaux dissimilaires par des sources de haute énergie et en particulier sur deux couples de matériaux présentant des problèmes métallurgiques différents : • cuivre - inox (lacune de miscibilité, différence de propriétés thermophysiques),• TA6V- inox (oxydation, formation de phases intermétalliques fragilisant la soudure).Pour le premier couple de matériaux, le soudage par laser Nd:YAG continu et par faisceau d'électrons a été utilisé. L'étude des évolutions de la morphologie des soudures, de la composition et de la microstructure des zones fondues ainsi que des propriétés mécaniques a permis de proposer des hypothèses sur les mécanismes de formation du mélange hétérogène à solubilité limitée. Afin de quantifier les phénomènes physiques intervenant en soudage continu de matériaux dissimilaires, la modélisation numérique a été mise en œuvre en utilisant le logiciel FEM "Comsol Multiphysics". Une série des modèles simulant les champs de températures, les mouvements convectifs et le malaxage (diffusion, méthode level set, méthode des champs de phases) a été créée. Dans le cas du laser, la formulation pseudo-stationnaire du transfert de chaleur basée sur la géométrie du capillaire simplifiée et la convection a été couplée avec les problèmes 2D de diffusion et de malaxage des matériaux dans différents plans horizontaux. En soudage par faisceau d'électrons, la morphologie de la microstructure a nécessité une formulation temporelle. Le modèle multiphysique final en couplage complet (solution multiphysique simultanée) reproduit le processus de formation d'une structure périodique de solidification lors du soudage par faisceau d'électrons et permet d'expliquer l'aspect des structures alternées entre matériaux immiscibles ou présentant de grandes différences de propriétés thermophysiques.Le deuxième couple de matériaux présente des problèmes métallurgiques majeurs liés à la formation des phases intermétalliques rendant l'assemblage direct par fusion impossible. La composition locale devient donc l'aspect-clef de la formation d’une soudure correcte : l'introduction d’un troisième matériau (cuivre) ayant une meilleure compatibilité avec le titane est nécessaire. Pour pouvoir déterminer les fenêtres optimales des conditions opératoires, les modèles numériques, créés précédemment, ont été adaptés pour quatre procédés de l’assemblage : faisceau d'électrons, soudage lasers Nd:YAG continu et pulsé, brasage par laser avec apport de fil. L'analyse élémentaire des microstructures dans les soudures résistantes mécaniquement a permis de développer le scénario de la solidification d'une zone fondue et de comprendre l'influence de la composition aux interfaces sur la résistance mécanique des assemblages.Les modèles numériques multiphysiques créés au cours de cette étude permettent l'accès rapide à la grande quantité d'information sur le comportement de la zone fondue en fonction des paramètres de soudage en se basant sur le nombre des données de départ relativement limité et sur quelques hypothèses simplificatrices. L'approche multiphysique à la modélisation de soudage permet de reproduire la forme de la zone fondue, visualiser les écoulements du liquide et cartographier la distribution de certains éléments avec une bonne corrélation avec les résultats expérimentaux. L'ensemble des modèles permet de déterminer les conditions opératoires répondant aux critères fixes en fonction de la métallurgie d'un couple hétérogène. / The present study is dedicated to the comprehension of the mechanism of materials mixing during dissimilar welding by high power beam sources. We have been interested in joining of two couples of metallic materials which present different metallurgical problems: • copper- stainless steel (miscibility gap, important difference in physical properties);• TA6V- stainless steel (oxidation on air, formation of intermetallic phases which made the joint brittle).For the first couple of materials, continuous laser Nd:YAG welding and electron beam welding have been applied. The experimental study of morphology evolution, composition, microstructure and mechanical properties has allowed establishing the hypotheses on formation of heterogeneous mixture between the materials having limited solubility. To quantify the physical phenomena of continuous dissimilar welding, the numerical modeling has been carried out by means of FEM software package "Comsol Multiphysics". A number of models reproducing temperature field, convection movements and mixing (diffusion, level set method, phase field method) between the materials has been created. In case of continuous laser welding, the pseudo-stationary formulation of heat transfer based on simplified key-hole geometry and convection has been coupled with two-dimensional problems of diffusion and mixing in horizontal planes. The electron beam welding presenting the nonlinear development of the weld has needed employing of temporary formulation. Final model including complete coupling (simultaneous multiphysical solving) reproduces the process of development of periodic solidification structure during electron beam welding and allows explaining the mechanism of formation of altered structures between immiscible materials which have important difference in thermophysical properties.The second couple of materials presents weldability problems due to formation of brittle intermetallic phases making direct joining by fusion impossible. The local elementary composition becomes the key-aspect of successful joining: the introduction of the third material (pure copper) having better compatibility with titanium is necessary. To determine the ranges of optimal operational conditions, numerical models created previously have been adapted to the case of four joining techniques: electron beam and laser Nd:YAG (continuous and pulsed) welding and laser brazing with filler wire. Elementary analysis of microstructures of resistant welds has allowed developing the solidification scenario and understanding the influence of local composition of heterogeneous interfaces on tensile properties of the joints. The multiphysical models created during this study allow rapid access to high quantity of data on behavior of melted zone in function of welding parameters basing on relatively limited input data and several simplification hypotheses. The multiphysical approach to welding modeling allows recreating the shape of melted zone, to visualization the convection movements and providing the cartography of several elements in good correspondence with experimental results. A set of models allows determination of operational parameters respecting fixed criterions determined by metallurgy of dissimilar couple.

Die vorliegende Dissertationsschrift beinhaltet die Ergebnisse der Untersuchung von loka-len und globalen Prozessen der Wärmeübertragung beim unterkühlten Strömungssieden. Sie ist an der Schnittstelle zwischen Reaktorsicherheitsforschung und der experimentellen Thermofluiddynamik für Phasenübergänge einzuordnen. In technischen Anwendungen zur effizienten Übertragung großer Wärmemengen spielt der Prozess des Siedens eine wichtige Rolle. Dieser Vorgang bewirkt einen starken Anstieg des Wärmetransportes von der beizten Wand an das Fluid bei vergleichsweise geringem Anstieg der Wandtemperatur. Der maximal übertragbare Wärmestrom beim Sieden wird begrenzt durch die sogenannte kritische Wärmestromdichte, deren Überschreitung zum thermomechanischen Versagen der beheizten Komponente führen kann. Aufgrund der Komplexität dieser Prozesse ist es trotz intensiver Arbeiten in den letzten Jahrzehnten noch nicht gelungen, diese Vorgänge detailliert zu modellieren. Eine Weiter-entwicklung der Modelle zur realistischen Beschreibung des unterkühlten Strömungssie-dens erfordert neuartige Untersuchungen, welche eine genaue Klassifizierung der partiellen Wärmeübergänge des Blasensiedens ermöglichen. Die Analyse partieller Wärmetransportgrößen beim unterkühlten Strömungssieden sowie der Einfluss variierender thermohydraulischer Randbedingungen ist Schwerpunkt dieser Arbeit. In der entwickelten Versuchsanlage erfolgt die Erfassung der Siedevorgänge bei Strömungsgeschwindigkeiten von 0,1 – 2 m/s und Eintrittstemperaturen von 60 - 98 °C. Mit Hilfe empfindlicher Temperaturmessungen in einem elektrisch beheizten Kapillarrohr innerhalb des Strömungskanals werden die globalen Vorgänge beim Übergang von Kon-vektion zum Sieden erfasst. Durch eine modellbasierte Bestimmung der Oberflächentem-peratur lassen sich Phänomene nachweisen, welche bisher weitestgehend unbeachtet ge-blieben sind. Die transparente Versuchsstrecke ermöglicht eine Erfassung der lokalen Sie-devorgänge mit optisch und zeitlich hochauflösenden Messverfahren. Durch die Entwick-lung neuer Algorithmen der digitalen Bildverarbeitung wurde eine umfangreiche, kenngrö-ßenorientierte Auswertung der in großem Umfang entstandenen Datenmengen realisiert. Der Einsatz transparenter und elektrisch leitfähiger Beschichtungen ermöglicht die mikro-skopische Erfassung des Blasenwachstums in weiten thermohydraulischen Parameterberei-chen. Mit erweiterten Bildverarbeitungsalgorithmen erfolgt die detaillierte und dynamische Bewertung des Blasenwachstumsverhaltens. Die statistische Auswertung der Verläufe er-möglicht die Ableitung eines Blasenwachstumsmodells für unterkühltes Strömungssieden. In einer weiteren Versuchsanordnung werden die lokalen Wärmetransportvorgänge bei der Ablösung quasistatisch gewachsener Blasen mit Hilfe der Infrarot-Thermographie be-stimmt. Dadurch können erstmalig die aus der lokalen Abkühlung der beheizten Oberfläche durch Blasenablösung resultierenden Wärmeströme unter Vernachlässigung der Bla-senbildung experimentell quantifiziert werden. Weiterhin können die bisher theoretisch beschriebenen Driftströmungen beim Aufstieg der Blase experimentell nachgewiesen wer-den. Die ermittelten Größen und Zusammenhänge tragen zur Weiterentwicklung und zum Abbau von Unsicherheiten bei der Modellierung von Wärmetransportvorgängen beim unterkühlten Strömungssieden bei.

Dans le secteur aéronautique, la performance des alliages d’aluminium connait une amélioration continue, grâce notamment à l’optimisation des procédés d’élaboration. Dans ce cadre, le travail de recherche vise à prédire le comportement des inclusions dans un bain d’aluminium brassé par induction afin d’améliorer la propreté inclusionnaire des alliages coulés. Un modèle numérique a été développé pour simuler le comportement magnétohydrodynamique du bain d’aluminium dans le creuset suivi par la modélisation du comportement d’une population d’inclusions non-métalliques. Le modèle 2D axisymétrique en régime transitoire s’appuie sur le code de CFD commercial ANSYS Fluent, bien que de nombreuses fonctions utilisateurs aient été introduites pour simuler les phénomènes spécifiques comme l’induction électromagnétique et la résolution des bilans de population. Le modèle MHD résout dans un unique maillage les phénomènes d’induction électromagnétiques, l’écoulement turbulent du bain d’Al, la déformation de la surface libre et les effets de la présence d’une couche de métal oxydée en surface du bain. Une méthode dite de vitesse de glissement (entre les particules et le fluide) a été choisie pour simuler à la fois le transport macroscopique des inclusions dans le bain d’Al et les interactions mésoscopique entre les inclusions (c.à.d. les mécanismes d’agrégation et de fragmentation). Des campagnes expérimentales à l’échelle d’un four de laboratoire et d’une installation industrielle accompagnent le travail numérique pour le valider. Les résultats de modélisation MHD exprimés sous la forme du profil de déformation du bain sont en accord raisonnable avec les mesures faites au laboratoire. Les résultats numériques démontrent également l'effet du frottement induit par la couche d'oxyde sur le profil du bain, ainsi que sur l'écoulement à proximité de la surface du dôme. Pour des conditions opératoires du four industriel en mode de maintien, l’évolution temporelle de la population au sein du bain est calculée. Il apparaît que la séparation magnétique est très intense, particulièrement dans la peau électromagnétique, et est ainsi responsable du transport et de la capture d’une grande fraction de la population d’inclusions à la paroi du four. / With an objective of improving processing and development of aerospace aluminum alloys, the current dissertation presents experimental and numerical tools which help comprehend the behavior of a non-metallic inclusion population in an Al bath stirred by induction. The mechanisms occurring in the metallurgical reactor were separated into two interlinked issues – (i) Magnetohydrodynamics (MHD) of the induction furnace, and (ii) Inclusion population dynamics in the Al bath, which were modeled using the ANSYS Fluent software and in-house User Defined Functions. For a 2D axisymmetric geometry, numerical simulations were performed in a single framework and calculated: (i) the electromagnetic forces using the A-V formulation, (ii) the free surface deformation using the Volume Of Fluid method, (iii) the turbulent stirring of the bath using a RANS-based k-omega model and (iv) the friction force due to the oxide layer by imposing a pseudo-wall condition on the bath free surface. The steady state MHD results and the physical properties of the inclusion population were used as input data for the transient inclusion behavior modeling. A combination of the Drift Concentration Method and the Population Balance Method was developed to respectively model the mean transport of inclusions within the bath at the macroscopic scale and the inclusion interactions (turbulent aggregation and fragmentation) at the mesoscopic scale. The performance of the MHD numerical tool was evaluated by comparing the model results with experimental results at laboratory and industrial scales. The simulation results in the form of the average bath surface profile were found to be consistent with the laboratory measurements. The results also illustrated the impact of the friction due to the oxide layer on the bath surface deformation as well as on the flow near the dome interface. The inclusion behavior simulations were performed for the holding mode operation of an industrial IMF. The deduced removal frequency compared the relative importance of each phenomenon. It was found that the electromagnetic migration, especially in the electromagnetic skin, dominates the inclusion dynamics and is responsible for the capture of a large fraction of the inclusion population.

The main purpose of this thesis was to propose and test a new approach that captures the features of single and multiple bubbles dynamics using the Shan and Chen-type lattice Boltzmann method (LBM). Two dimensional bubbles motions were simulated considering the buoyancy effect for which the topology of the bubble is characterized by the Eötvös (Eo), and Morton (M) numbers. A qualitative and quantitative validation were performed using the Level set method. Bubble shape deformation was captured and analysis based on terminal Reynolds number and degree of circularity show very good agreement with the experimental results and with available simulation results. In sum, this study presents crucial preliminary information to further analyze multiphase fluid flows in various contexts.

Orientadores: Denis José Schiozer, Eliana Luci Ligero / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica e Instituto de Geociências / Made available in DSpace on 2018-08-26T15:17:18Z (GMT). No. of bitstreams: 1 Santana_GustavoMenezes_M.pdf: 19965408 bytes, checksum: 111a8315e511e6681fab2b5663345c52 (MD5) Previous issue date: 2014 / Resumo: Os reservatórios do pré-sal brasileiro apresentam grandes volumes de óleo leve com quantidade significativa de CO2 dissolvido. O CO2 produzido no pré-sal pode ser utilizado como gás de injeção no método especial de recuperação de injeção alternada de água e gás (WAG). Neste trabalho, a injeção WAG-CO2 miscível é empregada na recuperação de um óleo leve com teor de cerca de 8% molar de CO2 em condições análogas às do pré-sal de alta pressão e baixa temperatura, onde fenômenos físicos inerentes à injeção WAG-CO2 são incorporados aos modelos de simulação de reservatórios através dos modelos de histerese da permeabilidade relativa, tal como o modelo trifásico de Larsen e Skauge (1998). Este trabalho tem como foco o estudo dos efeitos da modelagem da histerese de permeabilidade relativa em simulações de injeção WAG-CO2 miscível. Dois modelos sintéticos de reservatório com óleo leve e com diferentes graus de heterogeneidade são estudados. Os modelos de simulações empregam a formulação composicional por ser mais apropriada em casos de injeção de gás miscível em óleo leve. A histerese causa redução da permeabilidade relativa aos fluidos, o que pode gerar dois efeitos: o aumento da eficiência local de varrido de óleo e a perda de injetividade. O primeiro efeito contribui para o aumento da recuperação de óleo, enquanto a perda de injetividade, dependendo das condições operacionais dos poços, resulta na redução da quantidade de fluidos injetada, prejudicando a recuperação. O predomínio de um desses dois efeitos faz com que a implementação dos efeitos da histerese de permeabilidade relativa nos modelos de simulação com injeção WAG resulte em recuperações finais de óleo diferentes das obtidas nos casos em que a histerese não é considerada, atingindo-se valores até 8% maior ou menor conforme efeito predominante e condições operacionais utilizadas / Abstract: The Brazilian pre-salt reservoirs present large volumes of light oil with a significant amount of dissolved CO2. It is intended to use CO2 produced in the pre-salt as an injection gas on the enhanced oil recovery method of water-alternating-gas (WAG) injection. In this work, the miscible WAG-CO2 injection is used in light oil recovery with an amount of 8% molar CO2 in similar conditions to pre-salt with high pressure e low temperature where physical phenomena involving WAG-CO2 injection applied to a reservoir of light oil with dissolved CO2 are built into the reservoir simulation models through the hysteresis of relative permeability models, as the Larsen and Skauge (1998) model. This work focuses on the study of the effects of modeling the hysteresis of relative permeability in simulation of miscible WAG-CO2 injection. Two synthetic reservoir models with light oil and different heterogeneities are studied. The simulation models employ the compositional formulation because it is more appropriate in cases of miscible gas injection in light oil. The hysteresis causes a reduction of the relative permeability of the fluid, which can cause two effects: increased local oil swept efficiency and the loss of injectivity. The first effect contributes to increase the oil recovery, while the loss of injectivity, depending on the operating conditions of the wells, results in a reduction of the amount of injected fluids, reducing the recovery. The predominance of one of these two effects makes the implementation of the effects of hysteresis of the relative permeability in simulation models with WAG injection exhibit different results compared to those models without hysteresis, reaching values up to 8% higher or lower according to the predominant effect and operational conditions used / Mestrado / Reservatórios e Gestão / Mestre em Ciências e Engenharia de Petróleo

Orientadores: Antonio Carlos Bannwart, Oscar Mauricio Hernandez Rodriguez / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica e Instituto de Geociências / Made available in DSpace on 2018-08-28T09:34:02Z (GMT). No. of bitstreams: 1 BonillaRiano_Adriana_D.pdf: 7155927 bytes, checksum: 63be57b0a5136f5e783cfb4f870b0189 (MD5) Previous issue date: 2015 / Resumo: Neste trabalho, é apresentado o desenvolvimento de uma nova técnica para a medição da espessura do filme de água com alta resolução espacial e temporal em escoamento óleo-água. É proposto o uso de um sistema de medição de capacitância elétrica para medir filmes finos de água na proximidade da parede do tubo. O sistema conta com um sensor planar e foi necessário determinar a melhor geometria via simulações baseadas no Método de Elementos Finitos (FEM) para o caso de escoamento óleo-água. As características comparadas foram a profundidade de penetração do campo elétrico no filme de água, a sensibilidade, a resolução espacial mínima e a resposta quase-linear. Padrões de escoamento óleo-água disperso e anular instável foram estudados numa tubulação vertical de 12 m de comprimento, feita de vidro, com 50,8 milímetros de diâmetro interno. Os fluidos usados foram óleo mineral (com densidade 828 kg/m3 e viscosidade 220 mPas) e água da torneira. O trabalho experimental foi realizado nas instalações de escoamento multifásico do Laboratório de Engenharia Térmica e Fluidos (NETeF) da EESC-USP. Foi medida a espessura média do filme de água usando o sistema capacitivo e uma câmera de vídeo de alta velocidade. Para obter a espessura do filme de água a partir das imagens, foi proposto um algoritmo de pré-processamento e um algoritmo de segmentação que combina vários métodos disponíveis na literatura. Os resultados experimentais do sensor capacitivo mostraram que o sistema pode medir espessuras entre 400 µm e 2200 µm. O escoamento anular instável é caracterizado por grandes flutuações na no sentido do escoamento e na direção do perímetro, e estruturas interfaciais grandes (gotas). Por sua vez, o escoamento disperso tem flutuações menores no sentido do escoamento e na direção do perímetro, e estruturas interfaciais menores (gotículas). Uma estrutura interfacial média no espaço e no tempo é proposta para modelar a interface entre a região próxima à parede do tubo e a região do núcleo, e sua análise foi feita no domínio do tempo e da frequência. Foram estudadas a amplitude, velocidade e o comprimento da estrutura interfacial em cada par transmissor-receptor do sensor. Foi possível estabelecer correlações para a velocidade das estruturas em escoamento de óleo em água / Abstract: The development of a new technique for high spatial and temporal resolution film thickness measurement in oil-water flow is presented. A capacitance measurement system is proposed to measure thin water films near to the wall pipe. A planar sensor was chosen for sensing and some geometries were compared using finite elements method (FEM). The penetration depth, the sensitivity, the minimum spatial resolution (high spatial resolution) and the quasi-linear curve were the analyzed characteristics. Dispersed and unstable-annular oil-water flows patterns were studied in a 12-m long vertical glass pipe, with 50.8 mm of internal diameter, using mineral oil (828 kg/m3 of density and 220 mPa s of viscosity) and tap water. The experimental work was carried out in the multiphase-flow facilities of The Thermal-Fluids Engineering Laboratory (NETeF) of EESC-USP. Experiments with a high-speed video camera and the proposed capacitance system were performed to obtain images of the oil-water flow near the pipe wall. A pre-processing enhancement algorithm and a combined segmentation algorithm are proposed and allowed the measurement of characteristic space and time averaged water film thickness. Experimental results of the capacitive technique showed that the system could measure thickness between 400 µm and 2200 µm. It was possible to recognize and characterize typical behaviors of the two different flow patterns studied. Unstable-annular flow can be described by huge fluctuations on the flow direction and perimeter direction, and big interfacial structures (drops). On the other hand, dispersed flow has tiny fluctuations on the flow direction and perimeter direction, and smaller interfacial structures (droplets). An interfacial structure is suggested in order to model the interface between wall and core regions and it was analyzed in time and frequency domains; amplitude, velocity and wavelength at each pair transmitter-receiver of the sensor were studied. Correlations for the interfacial structure velocity were found for dispersed oil-in-water flow and unstable-annular flow / Doutorado / Explotação / Doutora em Ciências e Engenharia de Petróleo / CAPES