L'approche Smoothed Particle Hydrodynamics (SPH) est une méthode de calcul pour simuler des écoulements fluides avec une méthode Lagrangienne de type suivi de particules. A l'inverse des méthodes Euleriennes, ce type d'approche ne nécessite pas de maillage. C'est là l'un des atouts majeurs de l'approche SPH puisqu'elle permet de s'affranchir des méthodes de suivi d'interfaces couramment utilisées dans les approches Euléeriennes (par exemple Volume-of-Fluid, Level-Set ou Front-Tracking). L'approche SPH est donc de plus en plus utilisée dans les domaines de l'hydro-ingénierie et de la géophysique notamment de part le traitement naturel des écoulements à surface libre dans la méthode SPH. Cependant, l'approche SPH n'est utilisée que depuis peu pour simuler des écoulements multiphasiques complexes et de nombreux problèmes restent en suspens, notamment concernant une formulation adéquate ou le micro-mélange aux interfaces. L'un des principaux enjeux de ces travaux de thèse est d'analyser de façon objective les différentes approches de type SPH existantes et d'évaluer leur capacité à simuler des écoulements multiphasiques complexes. Ainsi, la modélisation des phénomènes liés à la tension de surface a été réalisée et validée via l'utilisation de techniques de type Continuum Surface Force. Les phénomènes de convection naturelle ont quant à eux été modélisés grâce à une nouvelle formulation plus générale (non-Boussinesq). Une partie de ces travaux est dédiée à l'étude des problèmes de micro-mélange aux interfaces: les effets indésirables (notamment la fragmentation de l'interface) sont analysés et des solutions sont proposées. Une autre part de travail porte sur la modélisation des mouvements ascendants de bulles dans des liquides, avec l'inclusion des interactions entre bulles. Des simulations SPH ont été réalisées pour différents régimes d'écoulement, chacun d'eux correspondant à un ratio spécifique entre la tension de surface, la viscosité et la flottabilité. Les prédictions numériques de la topographie des bulles, de leur vitesse ainsi que de leur coefficient de trainée ont été validées. Pour ce faire, les résultats numériques ont été comparés non seulement aux données expérimentales de référence mais également à d'autres simulations numériques de bulles ascendantes. Dans ces travaux de thèse, une étude détaillée des concepts liés aux contraintes d'incompressibilité a été réalisée. Dans cet objectif, deux traitements différents ont été comparés: l'approche faiblement compressible (où une équation d'état adéquate est choisie) et l'approche incompressible (où une projection des champs de vitesse sur un espace sans divergence est réalisée de deux facons différentes). La pertinence de ces modèles pour des simulations d'écoulements multiphasiques est également évaluée. Les problèmes associés aux paramètres numériques sont discutés et un choix approprié de ces paramètres est proposé. Pour ce faire, de nombreux calculs de validation en deux et trois dimensions ont été réalisés. Enfin, une extension est proposée pour simuler les phénomènes liés à l'ébullition via une approche SPH. Ce sujet étant encore en friche, de nouvelles idées et schémas sont proposés pour le changement de phase liquide-vapeur dans l'approche SPH / Smoothed Particle Hydrodynamics (SPH) is a fully Lagrangian, particle based approach for fluid-flow simulations. One of its advantages over Eulerian techniques is no need of a numerical grid. Therefore, there is no necessity to handle the interface shape as it is done in Volume-of-Fluid, Lavel-Set or Front-Tracking methods. Thus, the SPH approach is increasingly used for hydro-engineering and geophysical applications involving free-surface flows where the natural treatment of evolving interfaces makes it an attractive method. However, for real-life multi-phase simulations this method has only started to be considered and many problems like a proper formulation or a spurious fragmentation of the interface remain to be solved. One of the aims of this work is to critically analyse the existing SPH variants and assess their suitability for complex multi-phase problems. For modelling the surface-tension phenomena the Continuum Surface Force (CSF) methods are validated and used. The natural convection phenomena are modeled using a new, more general formulation, beyond the Boussinesq approximation. A substantial part of the work is devoted to the problem of a spurious fragmentation of the interface (the micro-mixing of SPH particles). Its negative effects and possible remedies are extensively discussed and a new variant is proposed. Contrary to general opinion, it is proven that the micro-mixing is not only the problem of flows with neglegible surface tension. A significant part of this work is devoted to the modelling of bubbles rising through liquids, including bubble-bubble interactions. The SPH simulations were performed for several flow regimes corresponding to different relative importance of surface tension, viscosity and buoyancy effects. The predicted topological changes, bubble terminal velocity and drag coefficients were validated with respect to reference experimental data and compared to other numerical methods. In the work, fundamental concepts of assuring the incompressibility constraint in SPH are also recalled. An important part of work is a thorough comparison of two different incompressibility treatments: the weakly compressible approach, where a suitably chosen equation of state is used, and truly incompressible method (in two basic variants), where the velocity field is projected onto a divergence-free space. Their usefulness for multi-phase modelling is discussed. Problems associated with the numerical setup are investigated, and an optimal choice of the computational parameters is proposed and verified. For these purposes the study is supported by many two- and three-dimensional validation cases. In addition, the present work opens new perspectives to future simulations of boiling phenomena using the SPH method. First ideas and sketches for the implementation of the liquid-vapour phase change are presented

Submitted by Flaviani Cristina da Silva Cruz (flavianicruz@gmail.com) on 2018-05-28T16:39:36Z No. of bitstreams: 1 dissertacao finallllll.pdf: 1641212 bytes, checksum: a98d0d98cacf7206b9dca507069233a8 (MD5) / Rejected by Ana Paula Santulo Custódio de Medeiros null (asantulo@rc.unesp.br), reason: - Folha de aprovação: falta o ano na data da defesa: 07 de maio .... - Ficha catalográfica incorreta: a ficha catalográfica só pode ser elaborada por um bibliotecário e não é a que você recebeu pela biblioteca em 21/05/2018: NOTA: De acordo com a Resolução CFB nº 184/2017 de 29/09/2017 – na Ficha catalográfica deve constar o nome do Bibliotecário/CRB, e ser elaborada de acordo com as normas vigentes segundo à AACR2. OBSERVAÇÃO: Caso se faça necessário acrescentar ou modificar alguma informação na Ficha catalográfica do seu trabalho, informá-la no campo recomendado do Sistema 3S ou responder este e-mail. É proibido perante a lei (Art. 297 – Código Penal) qualquer alteração documental, sem autorização do Bibliotecário responsável. DA FALSIDADE DOCUMENTAL: (I) FALSIDADE DE DOCUMENTO PUBLICO ART. 297: Falsificar, no todo ou em parte, documento publico, ou alterar documento publico verdadeiro: Pena – reclusão, de dois a seis anos, e multa. DOCUMENTO PUBLICO: é aquele elaborado por funcionário publico, de acordo com as formalidades, e desempenho de suas funções. Art. 232, CPP - Consideram-se documentos quaisquer escritos, instrumentos ou papéis, públicos ou particulares. on 2018-05-28T19:00:26Z (GMT) / Submitted by Flaviani Cristina da Silva Cruz (flavianicruz@gmail.com) on 2018-05-29T13:52:43Z No. of bitstreams: 1 dissertacao-final-final.pdf: 1600591 bytes, checksum: d5829fa27d80336f1e24b663680531f8 (MD5) / Approved for entry into archive by Adriana Aparecida Puerta null (dripuerta@rc.unesp.br) on 2018-05-29T14:23:36Z (GMT) No. of bitstreams: 1 cruz_fcs_me_rcla.pdf: 1625066 bytes, checksum: c010643fc4461e0b244b4ddd4414af66 (MD5) / Made available in DSpace on 2018-05-29T14:23:36Z (GMT). No. of bitstreams: 1 cruz_fcs_me_rcla.pdf: 1625066 bytes, checksum: c010643fc4461e0b244b4ddd4414af66 (MD5) Previous issue date: 2018-05-07 / Esta dissertação objetiva-se ao estudo analítico de técnicas de interpolação integral – fundamentadas na teoria de distribuição, também conhecida como teoria de funções generalizadas – e suas aplicações a reconstrução de imagens formadas por partículas (noise image) e à forma anisotrópica das equações da hidrodinâmica com partículas suavizadas, no inglês smoothed particle hydrodynamics, bem conhecida como SPH. A aplicação a processamento de imagens é um caso particular da técnica de interpolação utilizada em SPH. A imagem original é fragmentada na forma de partículas (em duas dimensões), aleatoriamente colhidas dos píxeis da imagem original, cuja densidade de probabilidade é proporcional à escala de cinza de cada píxel vizinho. Então é feita uma reconstrução da imagem através de interpolação anisotrópica, evidenciando que detalhes estruturais do contexto original são mais bem recuperados do que a correspondente técnica isotrópica. Apesar de não realizarmos diretamente simulações de mecânica dos fluidos neste trabalho, o que é proposto aqui é uma revisão das equações fundamentais da técnica de simulação SPH para o caso anisotrópico. As interpolações apresentadas neste trabalho foram feitas a partir de um núcleo normalizado de suavização, definido na forma de uma função regulada por partes, conhecida como cubic-B-spline, que é de classe analítica C2. / This mastered dissertation aims at the analytical study of integral interpolation techniques – based on the theory of distribution, also known as generalized function theory – and its applications the reconstruction of images formed by particles (noise image) and the anisotropic form of the fundamental equations of smoothed particle hydrodynamics, well known as SPH. The application to image processing is a particular case of the interpolation technique used in SPH. The original image is fragmented in the form of particles (in two dimensions), randomly drawn from the píxels of the original image, whose density is proportional to the gray scale of each neighboring píxel. Then a reconstruction of the image is made through anisotropic interpolation, evidencing that structural details of the original context are better recovered than the corresponding isotropic technique. Although we do not directly perform fluid mechanics simulations in this work, what is proposed here is a review of the fundamental equations of the SPH simulation technique for the anisotropic case. The interpolations presented in this work were made from a normalized smoothing kernel, defined as cubic-B-spline, which is of analytic class C2.

Cette thèse porte sur l'étude rhéocinétique de deux systèmes époxy/amine, DGEBA/DETDA et DGEBA/IPD pour le rotomoulage réactif. Les paramètres cinétiques du mécanisme auto-catalytique et non catalytique de la réticulation isotherme ont été identifiés par une méthode inverse à l'aide des résultats du suivi cinétique par spectroscopie proche infra rouge (PIR). La différence du comportement rhéocinétique entre les deux systèmes est attribuée à la différence de réactivité intrinsèque des groupes amines mis en jeu et l'importance de l'effet de substitution (réactivité des amines primaires par rapport aux amines secondaire) des deux durcisseurs. Une nouvelle méthode de détermination de l'effet de substitution a été établie, en utilisant les taux de conversion de l'époxy et de l'amine primaire mesurées par PIR. L'analyse des phénomènes de vitrification et gélification des deux systèmes a permis d'établir le diagramme Temps-Température-Transformation (TTT). La comparaison de la rotomoulabilité des deux systèmes a été effectuée par l'ajout des courbes d'iso-viscosité sur le diagramme TTT. L'écoulement au cours du rotomoulage a été modélisé par un écoulement à surface libre sous gravité d'un fluide réactif à l'intérieure d'un cylindre horizontal en rotation autour de son axe principal. Dans le cas des fluides de viscosité faible, la méthode lagrangienne de Smoothed Particle Hydrodynamics SPH, a été appliquée en utilisant une expression adaptée de la viscosité.

Le rotomoulage réactif est un procédé de fabrication de pièces creuses en polymère où la synthèse du matériau intervient pendant la mise en œuvre. Cette méthode présente plusieurs avantages comparée à la méthode traditionnelle utilisant des poudres thermoplastiques : réduction du temps de cycle, utilisation possible de matériaux techniques, et baisse de la consommation d'énergie et du coût des matières premières. Cependant le rotomoulage réactif est plus complexe à mettre en œuvre car la polymérisation provoque un changement important et rapide de la viscosité. Une des solutions pour optimiser ce procédé est de simuler l'écoulement du système réactif pendant la mise en œuvre.Pour ce travail nous avons utilisé un polyuréthane thermodurcissable. Des analyses thermiques et rhéologiques ont permis d'étudier les phénomènes de gélification et de vitrification du matériau et le diagramme Temps-Température-Transformation a été établi. Le comportement rhéocinétique du système a également été modélisé.Le procédé a été simulé en utilisant un code de calcul basé sur la méthode " Smoothed Particle Hydrodynamics " (SPH). Ce code a été développé par notre équipe et plusieurs améliorations ont été apportées au cours de cette étude. Pour effectuer des simulations plus réalistes en utilisant un plus grand nombre de particules, la première amélioration a consisté à accélérer la résolution des calculs. Ensuite l'évolution de la viscosité a été prise en compte grâce à l'utilisation d'un modèle rhéocinétique et une nouvelle condition limite a été développée pour simuler l'adhésion du polymère sur la paroi du moule. Les modifications nécessaires à la simulation d'écoulements 3D ont également été apportées au code SPH.

Tsunamis are among the most terrifying and complex physical phenomena potentially affecting almost all coastal regions of the Earth. Tsunami waves propagate in the ocean over thousands of kilometres away from their generating source at considerable speeds. Among several other tsunamis that occurred during the past decade, the 2004 Indian Ocean Tsunami and the 2011 Tohoku Tsunami in Japan, considered to be the deadliest and costliest natural disasters in the history of mankind, respectively, have hit wide stretches of densely populated coastal areas. During these major events, severe destruction of inland structures resulted from the action of extreme hydrodynamic forces induced by tsunami flooding. Subsequent field surveys in which researchers from the University of Ottawa participated ultimately revealed that, in contrast to seismic forces, such hydrodynamic forces are not taken into proper consideration when designing buildings for tsunami prone areas. In view of these limitations, a novel interdisciplinary hydraulic-structural engineering research program was initiated at the University of Ottawa, in cooperation with the Canadian Hydraulic Centre of the National Research Council, to help develop guidelines for the sound design of nearshore structures located in such areas. The present study aims to simulate the physical laboratory experiments performed within the aforementioned research program using a single-phase three-dimensional weakly compressible Smoothed Particle Hydrodynamics (SPH) numerical model. These experiments consist in the violent impact of rapidly advancing tsunami-like hydraulic bores with individual slender structural elements. Such bores are emulated based on the classic dam-break problem. The quantitatively compared measurements include the time-history of the net base horizontal force and of the pressure distribution acting on columns of square and circular cross-sections, as well as flow characteristics such as bore-front velocity and water surface elevation. Good agreement was obtained. Results show that the magnitude and duration of the impulsive force at initial bore impact depend on the degree of entrapped air in the bore-front. The latter was found to increase considerably if the bed of the experimental flume is covered with a thin water layer of even just a few millimetres. In order to avoid large fluctuations in the pressure field and to obtain accurate simulations of the hydrodynamic forces, a Riemann solver-based formulation of the SPH method is utilized. However, this formulation induces excessive numerical diffusion, as sudden and large water surface deformations, such as splashing at initial bore impact, are less accurately reproduced. To investigate this particular issue, the small-scale physical experiment of Kleefsman et al. (2005) is also considered and modeled. Lastly, taking full advantage of the validated numerical model to better understand the underlying flow dynamics, the influence of the experimental test geometry and of the bed condition (i.e. dry vs. wet) is investigated. Numerical results show that when a bore propagates over a wet bed, its front is both deeper and steeper and it also has a lower velocity compared to when it propagates over a dry bed. These differences significantly affect the pressure distributions and resulting hydrodynamic forces acting on impacted structures.

A two-dimensional axisymmetric implementation of the smoothed particle hydrodynamics (SPH) technique, called for short AxisSPH, has been described in this thesis, along with a number of basic tests and realistic applications. The main goal of this work was to fill a gap on a topic which has been scarcely addressed in the published literature concerning SPH. Although the application of AxisSPH to the simulation of real problems is restricted to those systems which display the appropriate symmetry there are, however, many interesting examples of physical systems which evolve following the axisymmetric premise. These examples belong to a variety of scientific and technological areas such as, for example, astrophysics, laboratory astrophysics or inertial confinement fusion. Additionally AxisSPH can be also useful in convergence studies of the standard 3D-SPH technique because the higher resolution achieved in 2D can be used to benchmark the three-dimensional codes. The main improvements implemented in AxisSPH with respect existing axisymmetric SPH formulations are summarized as follows: 1) We have derived simple analytical expressions for correction factors which largely improves the calculation of density and velocity in the vicinity of the z-axis. These expressions and their derivatives were given as a function of an adimensional parameter and do not increase the computational load of the scheme. 2) We have obtained the appropriate expression of the fluid Euler equations containing the new correction functions and their derivatives. Far enough from the singular axis, the scheme reduces to the standard formulation discussed by Brookshaw (2003). 3) A novel expression for the heat conduction term, which has to be added to the energy equation was devised and checked. This new term improves the description of the heat flux for those particles located at the axis neighborhoods. 4) Until now axisymmetric SPH hydrocodes handle artificial viscosity using a crude approach because it was treated as a simple restriction of the standard 3D Cartesian viscosity to 2D. Here we propose to calculate the viscous pressure as a combination of two terms, the first one is the (standard) Cartesian part and the second is the axis-converging part of the viscosity respectively. As expected this last term is of special relevance to simulate implosions. 5) We have developed an original method to incorporate gravity into AxisSPH. First the direct ring to ring force was found as a function of the Euclidean distance between the 2D particles. In second place the gravitational force on a given particle was obtained by summing the contributions of all N particles. We have also developed a more efficient scheme to obtain the gravitational force calculating the potential of the ring, instead the force because it involves lesser algebraic operations. The scheme has been checked using a large number of tests cases. These tests range from very specific oriented to check a particular algorithm or a piece of physics, to rather complex ones intended to analyze the behavior of the scheme in potential real applications (ICF, jets, astrophysics). At least in one case, the head on collision of a pair of white dwarfs, the result of the simulations carried out using AxisSPH brings new, unpublished, scientific material. / En esta tesis se ha desarrollado un código, que hemos llamado AxisSPH, en dos dimensiones axisimétrico a partir de la técnica conocida como SPH (“smooothed particle hydrodynamics”). AxisSPH ha sido validado después de realizar una serie de tests básicos y algunas simulaciones de situaciones reales. El objetivo principal de este trabajo ha sido llenar, en parte, el vacío existente al respecto en la literatura sobre SPH. Aunque sólo se puede aplicar AxisSPH en problemas reales que presenten la apropiada simetría, existen muchos ejemplos interesantes de sistemas físicos que presentan la simetría axial demandada. Existen ejemplos en campos de aplicación tanto científica como tecnológica, por ejemplo en astrofísica, en el llamado laboratorio de astrofísica o en fusión por confinamiento inercial (ICF). Otra interesante aplicación de AxisSPH puede ser su utilización en estudios de convergencia con otros códigos 3D-SPH debido a su mayor resolución, al tratarse de un código 2D. Las mejoras implementadas en el código AxisSPH en comparación con otros códigos axisimétricos SPH existentes se pueden resumir en los siguientes puntos: 1) Hemos deducido expresiones analíticas simples para unos factores de corrección que mejoran el cálculo de la densidad y la velocidad en las proximidades del eje z. Dichas expresiones y sus derivadas dependen de un parámetro adimensional que no incrementa mucho el peso computacional del esquema propuesto. 2) Hemos obtenido las expresiones adecuadas de las ecuaciones de Euler que contienen estas nuevas funciones correctoras y sus derivadas. Lejos del eje de singularidad estas ecuaciones se transforman en las de la formulación estándar propuesta por Brookshaw (2003). 3) Una expresión novedosa del término de conducción, que debe de añadirse a la ecuación de la energía, se ha propuesto y validado. Este nuevo término mejora la evolución del flujo de calor de las partículas situadas en las proximidades del eje z. 4) Hasta el momento los códigos hidrodinámicos SPH axisimétricos existentes trabajaban con una aproximación poco elaborada de la viscosidad artificial ya que consistían en una restricción a dos dimensiones de la viscosidad estándar 3D. En este trabajo proponemos el cálculo de la presión debida a la viscosidad como combinación de dos términos, el primero reflejo de la parte cartesiana y la segunda da cuenta de la parte relacionada con la convergencia en el eje. Como era de esperar este último término es de relevante importancia en la simulación de implosiones. 5) Hemos desarrollado un método original para incorporar el cálculo de la gravedad en el código AxisSPH. En primer lugar la fuerza directa de anillo a anillo y en segundo lugar la fuerza de la gravedad que sufre una determinada partícula a partir de la contribución del resto de las N partículas existentes. También hemos desarrollado un esquema más eficiente para calcular la gravedad a partir del cálculo del potencial del anillo en lugar del cálculo directo de la fuerza ya que implica un menor número de operaciones algebraicas. El método ha sido verificado con un gran número de test numéricos. Desde los más específicos orientados a comprobar la validez de un algoritmo particular o la capacidad para simular un fenómeno físico en particular, hasta simulaciones bastante más complejas, con la intención de validar la capacidad de simular aplicaciones potencialmente más reales (ICF, jets, astrofísica). Así, en al menos un caso, en la colisión frontal de dos enanas blancas, los resultados de la simulación utilizando AxisSPH pueden aportar material científico publicable.

The goal of this thesis is to develop a constraint based viscoplastic fluid model suitable for time-efficient dynamics simulation in 3D of granular matter. The model should be applicable to both the static and dense flow regime and at large pressures. The thesis is performed for UMIT Research Lab at Umeå University. It is a part of the research at UMIT connected to LKAB and Volvo CE and its applications can be in simulating industrial processes or training simulators. My work is based on previous work done by Claude Lacoursière, Martin Servin and Kenneth Bodin. They have created a constraint fluid model based on {\sph} and Claude's PhD. thesis. This model is extended with additional constraints to handle shear forces, which is necessary to model granular material. Some test cases are specified and compared visually to each other and to the results of other work. The model seems to work visually but more analysis and larger systems are needed to be certain. The model should scale well and is well suited for parallellization.

Nous étudions les mécanismes d'endommagement dus à l'érosion pluie sur des matériaux massifs et sur des matériaux revêtus. Pour cela, un banc d'essais spécifique a permis de mener des recherches en laboratoire. Le principe de l'essai repose sur l'émission de jets à haute vitesse et à haute fréquence.L'étude des matériaux massifs met en évidence un mécanisme de propagation de fissures par fatigue qui aboutit à des pertes de matière. Ces mêmes endommagements ont été observés sur des échantillons usés en service. Il a été confirmé que la dureté augmente la résistance à l'érosion pluie des matériaux métalliques.L'étude des revêtements s'est focalisée sur 2 technologies, à savoir la pulvérisation cathodique magnétron, qui est un procédé de déposition phase vapeur, et la projection thermique sous flamme supersonique. Les revêtements obtenus par projection thermique (dont l'épaisseur était supérieure à 200 μm), se sont révélés moins résistants à cause d'un manque d'adhérence ou de la présence de défauts au sein du matériau. Les revêtements obtenus par PVD (dont l'épaisseur était inférieure à 30 μm) ont permis d'obtenir des gains de résistance significatifs. Dans tous les cas, quel que soit la technologie utilisée, l'adhérence du revêtement s'est révélé être un paramètre critique en ce qui concerne la résistance de la surface à l'érosion pluie. Enfin, une simulation numérique en dynamique a permis d'étudier les champs de contraintes dans des feuillets métalliques, et ce, en faisant varier leurs épaisseurs, les matériaux qui les composent, et la vitesse d'impact. Les résultats de la simulation tendent à prouver que la propagation des ondes de contraintes dans le matériau peut entrainer des phénomènes de sur-contraintes dans les feuillets les plus fins à cause de la réflexion des ondes sur la face antérieure de la plaque.

Non-Newtonian fluids are of great scientific interest due to their range of physical properties, which arise from the characteristic shear stress-shear rate relation for each fluid. The applications of non-Newtonian fluids are widespread and occur in many industrial (e.g. lubricants, suspensions, paints, etc.) and environmental (e.g. mud, ice, blood, etc.) problems, often involving multiple fluids. In this study, the novel technique of Incompressible Smoothed Particle Hydrodynamics (ISPH) with shifting (Lind et al., J. Comput. Phys., 231(4):1499-1523, 2012), is extended beyond the state-of-the-art to model non-Newtonian and multi-phase flows. The method is used to investigate important problems of both environmental and industrial interest. The proposed methodology is based on a recent ISPH algorithm with shifting with the introduction of an appropriate stress formulation. The new method is validated both for Newtonian and non-Newtonian fluids, in closed-channel and free-surface flows. Applications in complex moulding flows are conducted and compared to previously published results. Validation includes comparison with other computational techniques such as weakly compressible SPH (WCSPH) and the Control Volume Finite Element method. Importantly, the proposed method offers improved pressure results over state-of-the-art WCSPH methods, while retaining accurate prediction of the flow patterns. Having validated the single-phase non-Newtonian ISPH algorithm, this develops a new extension to multi-phase flows. The method is applied to both Newtonian/Newtonian and Newtonian/non-Newtonian problems. Validations against a novel semi-analytical solution of a two-phase Poiseuille Newtonian/non-Newtonian flow, the Rayleigh-Taylor instability, and a submarine landslide are considered. It is shown that the proposed method can offer improvements in the description of interfaces and in the prediction of the flow fields of demanding multi-phase flows with both environmental and industrial application. Finally, the Lituya Bay landslide and tsunami is examined. The problem is approached initially on the real length-scales and compared with state-of-the-art computational techniques. Moreover, a detailed investigation is carried out aiming at the full reproduction of the experimental findings. With the introduction of a k-ε turbulence model, a simple saturation model and correct experimental initial conditions, significant improvements over the state-of-the-art are shown, managing an accurate representation of both the landslide as well as the wave run-up. The computational method proposed in this thesis is an entirely novel ISPH algorithm capable of modelling highly deforming non-Newtonian and multi-phase flows, and in many cases shows improved accuracy and experimental agreement compared with the current state-of-the-art WCSPH and ISPH methodologies. The variety of problems examined in this work show that the proposed method is robust and can be applied to a wide range of applications with potentially high societal and economical impact.

BARBOSA, Charles Welton Ferreira. Uma análise comparativa de métodos aproximativos baseados em smoothed particle hydrodynamics para animação de fluidos. 2013. 68 f. Dissertação (Mestrado em ciência da computação)- Universidade Federal do Ceará, Fortaleza-CE, 2013. / Submitted by Elineudson Ribeiro (elineudsonr@gmail.com) on 2016-07-11T12:31:20Z No. of bitstreams: 1 2013_dis_cwfbarbosa.pdf: 4026755 bytes, checksum: 3808122d6233e5855d6891854317c373 (MD5) / Approved for entry into archive by Rocilda Sales (rocilda@ufc.br) on 2016-07-14T15:52:26Z (GMT) No. of bitstreams: 1 2013_dis_cwfbarbosa.pdf: 4026755 bytes, checksum: 3808122d6233e5855d6891854317c373 (MD5) / Made available in DSpace on 2016-07-14T15:52:26Z (GMT). No. of bitstreams: 1 2013_dis_cwfbarbosa.pdf: 4026755 bytes, checksum: 3808122d6233e5855d6891854317c373 (MD5) Previous issue date: 2013 / Animations of fluids, such as water or smoke, are used to add details to virtual games or films. Various methods exist for simulating fluids using particle systems, including the Smoothed Particle Hydrodynamics (SPH) method. The SPH method is a good choice for simulating fluids for its easy description and implementation. This work describes some methods based on SPH for general fluid simulation, and for the interaction between fluids and solids. These methods are analysed and compared on their realism, and a few weak points of each method are presented. This analysis may be used for the creation of better methods based on SPH, and for the realistic animation of fluids. / Animações de fluidos, como a água ou fumaça, são utilizadas para introduzir detalhes em jogos virtuais ou filmes. Vários métodos existem para a simulação de fluidos utilizando sistemas de partículas e, em especial, utilizando o método Smoothed Particle Hydrodynamics, ou SPH. O método SPH é uma boa escolha para a simulação de fluidos por sua fácil descrição e implementação. Este trabalho apresenta alguns métodos baseados no SPH para a simulação de fluidos gerais, e para a interação entre fluidos e sólidos. Esses métodos são analisados e comparados relativos ao seu realismo, apresentando alguns pontos fracos de cada método. Essa análise pode ser utilizada para guiar a implementação de melhores métodos baseados no SPH, e para a animação realista de fluidos.