La recherche de la genèse des méandres en général se fait selon deux approches qui doivent se valider mutuellement: la dynamique géomorphologique et la dynamique des fluides, où la modélisation des flux 3D permet de simuler le mouvement hélicoïdal, avec des difficultés liés au temps et procédure de calcul et en simplifiant le problème par des géométries simples (Camporeal, et al, 2007). Smoothed particle hydrodynamics (SPH) est une méthode à maille libre, devenu très populaire, en particulier pour simuler les flux de surface libre. C’est une méthode robuste et puissante pour décrire les milieux soumis à des déformations (Gomez-Gesteira, et al, 2010). L’objectif de cette recherche est de modéliser les écoulements hélicoïdaux en 3D. Le modèle à éléments finis utilisé dans cette étude, RMA, a montré sa capacité à simuler les caractéristiques clés des méandres et sont en accord avec les expérimentations de Hasegawa (1983), et Xu et Bai (2013). Les procédures SPH sont élaborées à partir du modèle 3D d'écoulement du fluide, en tenant compte des collisions entre les particules et des conditions aux limites de canal courbe. Avec le code SPH, l’écoulement hélicoïdal est initié par l'addition de flux de tourbillon visqueux aux conditions initiales. Le modèle d'écoulement hélicoïdal est compatible avec les modèles issus des expérimentations de Wang et Liu (2015), ainsi que celles de Wu (2008) qui tient compte de flux secondaires dans un canal courbe. Ainsi, SPH est capable de simuler l’écoulement hélicoïdal lié à la courbure du canal, en accord avec Camporeal et al. (2007), et da Silva (2006) et Yalin (1993). / Meandering channels research in general is separated, but still correlated, into two approaches: geomorphologic and fluid dynamics, where 3D flow modeling receive more attention for its ability to simulate helicoidal motion even though it is high in computational efforts and limited to simple geometry (Camporeal, et al., 2007). Smoothed particle hydrodynamics (SPH) is one most noticeable meshfree method and now become very popular, particularly for free surface flows, it is a robust and powerful method for describing deforming media (Gomez-Gesteira, et al., 2010). SPH is a very promising method to answer 3D flow modeling in meander dynamics. Objective of this research that helical flow patterns from flow simulation with 3D nearly incompressible flow SPH method is comparable to flow simulation with 3D stratified flow finite element method with RMA. The finite element model using in this study, RMA has shown its capability to simulate the meander key characteristics and are agreed with experiments of Hasegawa (1983), and Xu and Bai (2013). SPH procedures are developed from 3D fluid flow model, collision handling between water particles, and curved channel boundary conditions. With SPH simulation, helical flow is initiated by adding up viscous flow and vorticity at initial conditions. The helical flow pattern is consistent with the patterns from very recently experiment investigation by Wang and Liu (2015), and theoretical sketch of secondary flows in a curved channel by Wu (2008). Thus, SPH method is able to develop helical flow as a result of curvature, agreed with Camporeal et al. (2007), and even without sediment transport, agreed with da Silva (2006) and Yalin (1993).

This document deals with the development of drilling tools by means of CAD and CAE technologies. At first, a brief overview of various design procedures of 3D drill models is presented, possibilities of measurement of force and moment loading during drilling are mentioned, a chip formation mechanism is briefly described and then a list of commonly used explicit (mesh) finite element methods used for cutting simulations is presented. A meshless SPH method have been selected for this work. Although it is able to handle the large deformations easily, it has been used for cutting simulations very rarely and only an orthogonal cutting simulations related information can be found in scientific databases. It has been demonstrated on the orthogonal cutting simulation of A2024-T351 alloy that was also the starting point for SPH simulation of drilling. The following is a decription of the design, simulation and prototyping of new drilling tools - drills with three and two cutting edges and an internal chip channel. This document is focused in detail on the variant with two cutting edges for which SPH drilling simulation has also been performed. Some drawbacks related to more precise chip simulation demands have been revealed, especially a rapid increase in number of SPH elements followed with prolongation of a computational time. Information related to the design of the drilling head with two cutting edges were then used to create the patent application.

Background. Fluid simulation is an area of ongoing research. In recent years, simulators have become more realistic and stable, partly by employing the condition of having divergence-free velocity fields. A divergence-free velocity field is a strict constraint that requires a high level of correctness in a simulation. Another recent development is in the subject of performance optimization, where asynchronous time integration is used. Asynchronous time integration means integrating different parts of a fluid with varying time step sizes. Doing so leads to overall larger time step sizes, which improves performance. This thesis combines the divergence-free velocity field condition with asynchronous time stepping in a particle-based simulator. Objectives. This thesis aims to achieve a performance speedup by implementing asynchronous time integration into an existing particle-based simulator that assures the velocity field is divergence-free. Methods. With an open source simulator employing a divergence-free velocity field as a starting point, asynchronous time integration is implemented. This is achieved by dividing the fluid into three regions, each with their own time step sizes. Introducing asynchronous time integration means significantly lowering the stability of a simulation. This is countered by implementing additional steps to increase stability. Results. Roughly a 40\% speedup is achieved in two out of three scenes, with similar visual results as the original synchronous simulation. In the third scene, there is no performance speedup as the performance is similar to that of the original simulation. The two first scenes could be sped up further with more aggressive settings for asynchronous time integration. This is however not possible due to stability issues, which are also the cause for the third scene not resulting in any speedup. Conclusions. Asynchronous simulation is shown to be a valid option even alongside a divergence solver. However, occasional unrealistic behavior resembling explosions among the particles do occur. Besides from being undesirable behavior, these explosions also decrease performance and prevent more aggressive performance settings from being used. Analysis of their cause, attempted solutions and potential future solutions are provided in the discussion chapter.

Ce mémoire décrit le modèle numérique Smoothed Particle Ice Dynamics Equations for Rivers (SPIDER) développé par Nolin, Roubtsova et Morse afin de modéliser le transport et l'accumulation de glace à la surface d'une rivière. Dans SPIDER, la méthode Smoothed Particle Hydrodynamics (SPH) est utilisée pour résoudre les équations de dynamique de la glace en deux dimensions. Le modèle SPIDER a été utilisé afin de simuler la formation d'embâcle fluvial. Les profils d'embâcle obtenus sans considérer le frottement des berges sont identiques à ceux présentés dans les études antérieures. Les profils obtenus en considérant le frottement des berges diffèrent de ceux présentés dans les études antérieures mais sont conformes à la méthode utilisée pour modéliser le frottement. Deux études paramétriques ont été réalisées afin de caractériser l'influence de l'angle de frottement interne de la glace ([phi]) et du coefficient de traînée eau-glace (cw) sur les profils d'embâcle. Ces études montrent que plus l'angle ʹ est grand, plus l'embâcle est long et moins la glace au pied de l'embâcle est épaisse. De même, plus cw est grand, plus l'embâcle est court et plus l'épaisseur de glace au pied de l'embâcle est grande. Le relâchement d'un embâcle fluvial a été simulé en couplant SPIDER au modèle hydrodynamique KIMBULA développé par Roubtsova V. et Kahawita R. à l'École Polytechnique de Montréal. Les résultats préliminaires simulés par ce modèle hybride sont similaires à ceux présentés dans les études antérieures.

Using a hydrodynamical smoothed particle hydrodynamics (SPH) zoom-in simu lation of a galaxy group, we present a set of tail identification methods, and study the statistical properties of galaxy tails and their correlations with their expected formation mechanisms. We have a sample of 4548 M > 108 M⊙ galaxies across 58 snapshots from z = 0.67 to z = 0. For each galaxy, we apply a series of velocity and density cuts to identify the tail. We observed no significant correlations between galaxy tail mass and ram pressure, though we note some issues with our sampling. Tracking four visually identified jellyfish galaxies over time showed some evidence of increased ram pressure driving ISM mass loss, as well as spikes in tail mass pre ceding spikes in ram pressure with temporal offsets ranging from 500 Myr to 2 Gyr. No correlation was found between ISM mass and tail mass. We track the tail gas of a particularly well defined jellyfish galaxy 3.2 Gyrs back in time. We find that a lower bound of 30% of the tail gas was never in the ISM. Distinguishing between former ISM tail material and never ISM-accreted tail material, we see evidence of temperature mixing with the IGM in the former. Velocity and radial trajectory maps show a sharp impulse of ∆v ≈ 50 km s−1 over 4 snapshots, affecting both the never ISM-accreted tail material and CGM material, with the former showing evidence of momentum mixing onto the former ISM material. Combined with ob servations of CGM stripping, we propose that a significant portion of galaxy tails consists of stripped CGM that got swept up into the stripped ISM / Thesis / Master of Science (MSc)

No / An incompressible smoothed particle hydrodynamics (SPH) method is presented to simulate Newtonian and non-Newtonian flows with free surfaces. The basic equations solved are the incompressible mass conservation and Navier¿Stokes equations. The method uses prediction¿correction fractional steps with the temporal velocity field integrated forward in time without enforcing incompressibility in the prediction step. The resulting deviation of particle density is then implicitly projected onto a divergence-free space to satisfy incompressibility through a pressure Poisson equation derived from an approximate pressure projection. Various SPH formulations are employed in the discretization of the relevant gradient, divergence and Laplacian terms. Free surfaces are identified by the particles whose density is below a set point. Wall boundaries are represented by particles whose positions are fixed. The SPH formulation is also extended to non-Newtonian flows and demonstrated using the Cross rheological model. The incompressible SPH method is tested by typical 2-D dam-break problems in which both water and fluid mud are considered. The computations are in good agreement with available experimental data. The different flow features between Newtonian and non-Newtonian flows after the dam-break are discussed.

Cette thèse a pour objet le développement d'un modèle numérique de simulation des fluides fondé sur la méthode Smoothed Particle Hydrodynamics (SPH). SPH est une méthode de simulation numérique sans maillage présentant un certain nombre d'avantages par rapport aux méthodes Eulériennes. Elle permet notamment de modéliser des écoulements à surface libre ou interfaces fortement déformées. Ce travail s'adresse principalement à quatre problématiques liées aux fondements de la méthode SPH : l'imposition des conditions aux limites, la prédiction précise des champs de pression, l'implémentation d'un modèle thermique et la réduction des temps de calcul. L'objectif est de modéliser des écoulements industriels complexes par la méthode SPH, en complément de ce qui peut se faire avec des méthodes à maillage. Typiquement, les problèmes visés sont des écoulements 3-D à surface libre ou confinés, pouvant interagir avec des structures mobiles et/ou transporter des scalaires, notamment des scalaires actifs (e.g. température). Dans ce but, on propose ici un modèle SPH incompressible (ISPH) basé sur une représentation semi-analytique des conditions aux limites. La technique des conditions aux limites semi-analytiques permet d'imposer des conditions sur la pression de manière précise et physique, contrairement à ce qui se fait avec des conditions aux limites classiques en SPH. Un modèle k-epsilon a été incorporé à ce nouveau modèle ISPH, à partir des travaux de Ferrand et al. (2013). Un modèle de flottabilité a également été ajouté, reposant sur l'approximation de Boussinesq. Les interactions entre flottabilité et turbulence sont prises en compte. Enfin, une formulation pour les frontières ouvertes dans le nouveau modèle est établie. La validation du modèle en 2-D a été réalisée sur un ensemble de cas-tests permettant d'estimer les capacités de prédiction du nouveau modèle en ce qui concerne les écoulements isothermes et non-isothermes, laminaires ou turbulents. Des cas confinés sont présentés, ainsi que des écoulements à surface libre (l'un d'eux incluant un corps solide mobile dans l'écoulement). La formulation pour les frontières ouvertes a été testée sur un canal de Poiseuille plan laminaire et sur deux cas de propagation d'une onde solitaire. Des comparaisons sont présentées avec des méthodes à maillage, ainsi qu'avec un modèle SPH quasi-incompressible (WCSPH) avec le même type de conditions aux limites. Les résultats montrent que le modèle permet de représenter des écoulements dans des domaines à géométrie complexe, tout en améliorant la prédiction des champs de pression par rapport à la méthode WCSPH. L'extension du modèle en trois dimensions a été réalisée dans un code massivement parallèle fonctionnant sur carte graphique (GPU). Deux cas de validation en 3-D sont proposés, ainsi que des résultats sur un cas simple d'application en 3-D / In this work a numerical model for fluid flow simulation was developed, based on the Smoothed Particle Hydrodynamics (SPH) method. SPH is a meshless Lagrangian Computational Fluid Dynamics (CFD) method that offers some advantages compared to mesh-based Eulerian methods. In particular, it is able to model flows presenting highly distorted free-surfaces or interfaces. This work tackles four issues concerning the SPH method : the imposition of boundary conditions, the accuracy of the pressure prediction, the modelling of buoyancy effects and the reduction of computational time. The aim is to model complex industrial flows with the SPH method, as a complement of what can be done with mesh-based methods. Typically, the targetted problems are 3-D free-surface or confined flows that may interact with moving solids and/or transport scalars, in particular active scalars (e.g. the temperature). To achieve this goal, a new incompressible SPH (ISPH) model is proposed, based on semi-analytical boundary conditions. This technique for the representation of boundary conditions in SPH makes it possible to accurately prescribe consistent pressure boundary conditions, contrary to what is done with classical boundary conditions in SPH. A k-epsilon turbulence closure is included in the new ISPH model. A buoyancy model was also added, based on the Boussinesq approximation. The interactions between buoyancy and turbulence are modelled. Finally, a formulation for open boundary conditions is proposed in this framework. The 2-D validation was performed on a set of test-cases that made it possible to assess the prediction capabilities of the new model regarding isothermal and non-isothermal flows, in laminar or turbulent regime. Confined cases are presented, as well as free-surface flows (one of them including a moving body in the flow). The open boundary formulation was tested on a laminar plane Poiseuille flow and on two cases of propagation of a solitary wave. Comparisons with mesh-based methods are provided with, as well as comparisons with a weakly-compressible SPH (WCSPH) model using the same kind of boundary conditions. The results show that the model is able to represent flows in complex boundary geometries, while improving the pressure prediction compared to the WCSPH method. The extension of the model to 3-D was done in a massively parallel code running on a Graphic Processing Unit (GPU). Two validation cases in 3-D are presented, as well as preliminary results on a simple 3-D application case

Ce travail traite des méthodes de calcul numérique pour les simulations hydrodynamiques appliquées principalement sur des produits développés par ANDRITZ HYDRO. Il s’agit ici de mettre en place des schémas d’ordre élevé pour des simulations CFD en utilisant le code de calcul ASPHODEL développé et utilisé par ANDRITZ HYDRO. Les principales motivations sont l’augmentation de la fiabilité des résultats de calculs numériques avec un coût de calcul raisonnable. Cette fiabilité s’exprime à travers l’augmentation de la précision et de la robustesse des schémas numériques. Le code de calcul ASPHODEL est basé sur la méthode sans maillage SPH-ALE. Mélange entre les volumes finis et la méthode SPH (Smoothed Particle Hydrodynamics), la méthode SPH-ALE emploie un ensemble de points appelés particules servant à la discrétisation du domaine fluide. Elle permet en particulier de par son caractère sans maillage, un suivi des surfaces libres sans effort de calcul supplémentaire. Cet aspect est véritablement attrayant pour bon nombre d’applications industrielles notamment la simulation des écoulements à surface libre se produisant dans une turbine Pelton, mais également le remplissage d’une turbine Francis. Cependant, le bémol à cette méthode est son manque de précision spatiale. En effet les points de calcul étant mobiles, les opérateurs spatiaux doivent être en mesure de conserver leur précision et leur robustesse au cours du temps. La qualité des résultats en est du coup impactée, en particulier le champ de pression souvent excessivement bruité. La montée en ordre et l’amélioration de la consistance des opérateurs pour un vaste panel de configurations géométriques sont donc les enjeux de ce travail. En utilisant des outils inspirés par les volumes finis non-structurés, il est possible d’améliorer les opérateurs spatiaux. En effet, la montée en ordre ou p-raffinement peut notamment se faire avec des reconstructions d’ordres élevés pour évaluer les états aux interfaces des problèmes de Riemann. La sommation des flux numériques résolus par un solveur de Riemann est ensuite retravaillée pour obtenir un schéma numérique d’ordre global cohérent. Le même soucis de cohérence avec les schémas en temps doit d’ailleurs être pensé. Le gain de précision apporté par les schémas numériques d’ordre élevé est comparé avec un raffinement spatial, c’est à dire une augmentation du nombre des particules de taille plus petite, aussi appelé h-raffinement. La méthode SPH-ALE améliorée est ensuite testée sur des cas représentatifs des applications visées. En conclusion, les développements effectués dans cette étude ont été guidés par l’application en turbine Pelton principalement mais il va de soi qu’ils sont applicables à des écoulements sans surface libre dans les turbines Francis par exemple. Ce travail montre les possibilités d’une méthode sans maillage pour des cas d’écoulements complexes autour de géométrie tournantes. / This work deals with numerical methods for hydrodynamic testing applied mainly on products developed by ANDRITZ HYDRO. This is to put in place high order schemes for CFD simulations using the ASPHODEL calculation code developed and used by ANDRITZ HYDRO. The main reasons are the increased reliability of the results of numerical calculations with a reasonable computational cost. This reliability is expressed through increasing the accuracy and robustness of numerical schemes. The ASPHODEL computer code is based on the meshfree method SPH-ALE. Mix between finite volume method and SPH (Smoothed Particle Hydrodynamics), the SPH-ALE method uses a set of points called particles serving as the fluid domain discretization. It allows track free surfaces without additional computational effort. This is truly attractive for many industrial applications including the simulation of free surface flows occurring in a Pelton turbine, but also filling a Francis turbine. However, the downside of this method is its lack of spatial accuracy. Indeed calculation points are mobile, space operators must be able to keep their accuracy and robustness over time. The quality of results is impacted especially the pressure field is often excessively noisy. The rise in order and improving the consistency of the operators for a wide range of geometric configurations are the challenges of this work. Using tools inspired by the unstructured finite volume, it is possible to improve the spatial operators. Indeed, the increasing order or p-refinement particular can be done with reconstructions of high order to assess the conditions at the interfaces of Riemann problems. The summation of discret fluxes solved by Riemann solver is then reworked to obtain a coherent global order scheme. The same concern for consistency with temporal schemes should also be considered. The precision gain provided by numerical schemes of higher orders is compared with a spatial refinement ie an increase in the number of smaller particles ; also called h -refinement . Improved SPH -ALE method is then tested on representative cases of intended applications. In conclusion, the developments made in this study were guided in accordance mainly with the Pelton turbine but it goes without saying that they are applicable to non- free surface flows in Francis turbines for example. This work shows the possibilities of a free mesh method for cases of complex flow around rotating geometry.

La cavitation peut avoir lieu dans les turbines hydrauliques. Ce phénomène se produit lorsque les bulles de vapeur s’effondrent à proximité de la surface de la machine. Ceci entraîne des conséquences négatives, telles que l’érosion, affectant ainsi les performances de la machine. L’effondrement d’une bulle de gaz non-condensable dans l’eau est simulé en utilisant la méthode sans maillage SPH-ALE, qui intègre un modèle pour simuler les écoulements compressibles et multiphases. Le modèle résout les équations de conservation de masse, de quantité de mouvement et d’énergie du système d’Euler, en utilisant l’équation d’état de Stiffened Gas pour l’eau et l’équation d’état de gaz parfait pour le gaz non-condensable à l´ıntérieur de la bulle. Les deux phases sont modélisées comme compressibles et le changement de phase n’est pas considéré. La caractéristique sans maillage de la méthode SPH-ALE permet le calcul des écoulements diphasiques où l’interface est nettement définie. Pour les applications de cavitation, où le nombre de Mach atteint des valeurs de 0.5, la distribution de particules doit être corrigée. Cela est réalisé grâce à la fonctionnalité ALE. Le modèle compressible a été validé à l’aide de configurations monodimensionnelles, comme le cas du tube à choc pour des écoulements monophase et multiphases. L’effondrement de la bulle près d´une paroi a été abordé comme le mécanisme fondamental qui produit des dégâts. Son comportement général se caractérise par la formation d’un micro jet d’eau et par l’effondrement de la bulle sur elle-même. Le phénomène est analysé en tenant compte des principaux paramètres qui le régissent, comme la distance initiale entre le centre de la bulle et la paroi (H0), la taille de la bulle (R0) et le taux de pression qui entraîne l’effondrement (pw/pb). Il est démontré que l’intensité de l’effondrement dépend principalement du rapport de pression entre le liquide et la bulle (pw/pb). De plus, quatre indicateurs, comme la pression en paroi, l’impulsion, la pression du coup de bélier et la vitesse du micro jet d’eau, servent à déterminer le chargement. Cette analyse indique qu’une bulle initialement située à une distance inférieure à H0/R0 = 2 présente un haut potentiel d’endommagement. Afin de prédire cet endommagement, la mécanique du solide est analysée à l’aide de simulations d’interaction fluide-structure. On obtient que le matériau réagit aux charges hydrauliques en ayant des zones de compression et de traction. Ceci suggère qu’un mécanisme de fatigue entraîne le phénomène d’endommagement. En plus, on constate que les contraintes les plus importantes sont situées sous la surface du matériau, indiquant que cette zone peut être sujette à une déformation plastique. / Hydraulic turbines can experience cavitation, which is a phenomenon occurring when vapor bubbles collapse in the vicinity of the machine’s surface. This phenomenon can lead to negative consequences, such as erosion, that affect the machine’s performance. The compression of a non-condensable gas bubble in water is simulated with the Smoothed Particle Hydrodynamics method following the Arbitrary Lagrange Euler approach (SPHALE), where a compressible and multiphase model has been developed. The model solves the mass, momentum and energy conservation equations of the Euler system using the Stiffened Gas EOS for water and the ideal gas EOS for the non-condensable gas inside the bubble. Both phases are modeled as compressible and the phase change is not considered. The meshless feature of the SPH-ALE method allows the calculation of multiphase flows where the interface is sharply defined. For cavitation applications, where the Mach number reaches values of 0.5, the distribution of particles must be corrected, which is achieved by the ALE feature. The compressible model was validated through monodimensional configurations, such as shock tube test cases for monophase and multiphase flows. The bubble compression close to the wall has been addressed as the fundamental mechanism producing damage. Its general behavior is characterized by the formation of a water jet and by the collapse of the bubble by itself. The phenomenon is analyzed by considering the major parameters that govern the bubble collapse dynamics, such as the initial distance between the bubble center and the wall (H0), the bubble size (R0), and the collapse driven pressure ratio (pw/pb). It is shown that the intensity of the collapse depends mainly on the pressure ratio between the liquid and the bubble (pw/pb). As well, four indicators, such as the pressure at the wall, the impulse, the water-hammer pressure and the water jet velocity, are used to determine the loading. This analysis gives that the bubble initially located at a distance lower than H0/R0 = 2 presents high potential to cause damage. In order to predict the damage due to the bubble collapse, the solid mechanics is analyzed through fluid-structure interaction simulations. It is obtained that the material reacts to the hydraulic loads by having compression and traction zones, suggesting that a fatigue mechanism drives the damage phenomenon. Additionally, it is found that the highest stresses are located below the material surface, indicating that this zone may reach plastic deformation.

Made available in DSpace on 2015-03-04T18:50:40Z (GMT). No. of bitstreams: 1 Apresentacao.pdf: 115092 bytes, checksum: 8c6ec19160c941d82efcb1e4536bf57c (MD5) Previous issue date: 2006-08-04 / Fundação Carlos Chagas Filho de Amparo a Pesquisa do Estado do Rio de Janeiro / The past two decades showed a rapid growing of physically-based modeling of fluids for computer graphics applications. Techniques in the field of Computational Fluid Dynamics (CFD) have been applied for realistic fluid animation for virtual surgery simulators, computer games and visual effects. In this approach, since the equation is solved numerically the next step is the rendering. A majority of fluid animation methods in computer graphics rely on a top down viewpoint that uses 2D/3D mesh based approaches motivated by the Eulerian methods of Finite Element (FE) and Finite Difference (FD), in conjunction with Navier-Stokes equations of fluids. Recently mesh-free methods like Smoothed Particle Hydrodynamics (SPH) have been applied. On the other hand, cellular automata (CA) are discrete models based on point particles that move on a lattice, according to suitable and simple rules in order to mimic a fully molecular dynamics. Such bottom-up framework needs low computational resources for both the memory allocation and the computation itself. In this work, we study the theoretical and practice aspects for computational animation of fluids in computer graphics, using cellular automata and SPH. We propose two models for animation of two-phase systems (e.g. gas-liquid), one based on SPH and CA and another only on CA. Finally, we describe a software developed in the context of this thesis for animation of fluids by CA. / Nas últimas décadas, observou-se um interesse crescente por aplicações de técnicas de dinâmica de fluidos na geração de efeitos visuais para a indústria cinematográfica e de jogos eletrônicos. Estas aplicações fazem parte da chamada Animação Computacional de Fluidos; a qual é uma área multidisciplinar, envolvendo também conceitos e métodos em computação gráfica e visualização científica. Nesta área, uma vez resolvidas numericamente as equações de fluidos, passa-se à fase de rendering, onde técnicas de visualização são aplicadas sobre os campos gerados, com o objetivo de criar efeitos visuais, tais como transparência, imagens refletidas na superfície de um líquido, ou mesmo, efeitos especiais que incluem deformação de paisagens, incêndios, etc. O métodos de Diferenças Finitas é o mais tradicional em trabalhos de animação de fluidos em computação gráfica. Nos últimos anos, porém, métodos baseados em sistemas de partículas, e livres de malhas, tais como o Smoothed Particle Hydrodinamics (SPH), foram utilizados na tentativa de resolver limitações inerentes aos métodos baseados em malhas. Por outro lado, métodos baseados em uma classe de autômatos celulares (AC), cuja evolução imita um sistema de partículas, vêm sendo também estudados como uma alternativa ao uso de equações diferenciais parciais e métodos numéricos para simulação de fluidos. Nesta tese, são estudados os aspectos teóricos e práticos da animação computacional de fluidos para computação gráfica, utilizando autômatos celulares e SPH. São propostos dois modelos para animação de sistemas bifásicos (gás-líquido, por exemplo), um deles baseado em SPH e AC, e um segundo totalmente baseado em AC. Finalmente, descrevemos um aplicativo, desenvolvido no âmbito desta tese, para animação de fluidos via AC.