Le travail engagé dans cette thèse porte sur l'étude numérique des Interactions Fluide-structure en hydrodynamique. Dans une première partie, une analyse détaillée des méthodes de couplage (schémas décalés) a été effectuée sur un cas académique. Il s'agit de la résolution de l'équation non-linéaire de Burgers dans un domaine mobile, dont I'interface mobile est représentée par un système de type masse ressort. Selon la discrétisation en temps et la linéarisation du problème couplé, on distingue quatre schémas de couplages différents : explicite, semi-implicite, implicite-externe et implicite-interne. Une étude comparative des performances en vitesse de convergence et en temps de calcul de ces schémas a été effectuée. Les performances varient suivant le schéma de couplage utilisé. Le schéma explicite permet un calcul rapide en comparaison des autres schémas. En revanche il n'assure pas la conservation de l'énergie mécanique à I'interface fluide-structure. D'où le problème de stabilité du schéma numérique. Ce problème ne se pose pas pour les algorithmes de couplage implicites, car dans ce cas la conservation de l'énergie à I'interface est assurée. Il s'agit en effet d'une condition de convergence du schéma implicite. Ce schéma requière plus de temps de calcul, mais il est nécessaire pour avoir plus de précision dans les résultats. Par ailleurs, I'analyse des déplacements de I'interface fluide-structure montre que l'écart entre la position de I'interface comme étant le bord mobile du fluide et la position de la structure, dépend principalement du schéma d'actualisation du maillage choisi.Dans une deuxième partie une extension de l'étude des algorithmes de couplage à un problème plus concret d'IFS est effectuée. Un hydrofoil en pilonnement et tangage est ainsi étudié. L'équation de la dynamique de I'hydrofoil est écrite en considérant un centre de rotation situé à une distance non nulle du centre de gravité.Ce qui rend l'équation non-linéaire et introduit un couplage des deux modes pilonnement et tangage) ainsi qu'un amortissement du tangage. La dynamique de I'hydrofoil est étudiée pour différentes configurations : en mouvement libre ou forcé, dans un fluide au repos ou en écoulement. On observe que le mouvement de I'hydrofoil est pseudo périodique amorti. L'évolution des charges hydrodynamiques suit également cette tendance et tend vers un point d'équilibre. L'étude vibratoire montre bien une modification des fréquences propres du système, qui varient suivant que le fluide est au repos ou en écoulement. Le problème est également couplé à l'équation de la position du centre de pression, qui dépend de la position de I'hydrofoil et de l'écoulement. Celle-ci présente une singularité lorsque la portance et la traînée s'annulent simultanément.Enfin Les équations prenant en compte la présence d'un fluide non-homogène à I'interface fluide-structure, du type des écoulements cavitants par poche stationnaire ou auto-oscillante, ont été développés. La méthode consiste à séparer les variables du fluide en écoulement autour d'un hydrofoil immobile d'une part et celles de l'écoulement généré par la vibration de I'hydrofoil d'autre part. Il en résulte un opérateur de masse ajoutée non symétrique en milieu non homogène et un opérateur d'amortissement ajouté dû au taux de variations de masse volumique à l’interface dans le cas auto-oscillant. L'ensemble se traduit par une modulation au cours du temps des fréquences propres et des amplitudes du système. / A numerical study of Fluid Structure Interaction (FSI) in hydrodynamic case is adressed in this thesis. Thirstly, the analysis of coupling methods (staggered schemes) was established to an academic case. It corresponds to the resolution of non linear Burgers equation in a moving domain where the moving interface is assimilated to a mass spring system. According to the time discretisation and linearization of the coupled problem, four coupling scheme can be defined : explicit, semi-implicit, implicit-outer and implicit-inner. A comparative performance study in convergence and computing time were performed. The performance depends on the coupling scheme used. The explicit scheme requires less time compared to the others schemes. However it does not allow the mechanical energy conservation at the interface, inducing the stability issue of the numerical scheme. This instabilities does not arise for the implicit coupling algorithms because the energy conservation at the interface is fulfilled. lndeed, a convergence condition is added for implicit schemes. Even though these schemes require more computing time, they are necessary to get better precision. Inter alia, the fluid-structure interface analysis shows that the gap between the interface taken as the moving boundary and the structure position mostly depends on the actualization scheme of the chosen mesh.In the second part, the coupling algorithm study is extended to physical problem of FSI. A hydrofoil in heave and pitch immersed in a fluid flow is then studied. The equation of hydrofoil movement takes account the distance between the rotation center and the center of gravity. This causes the equation to be nonlinear and introduces a coupling of the two movements (heave and pitch) and a damping of the heave movement. The hydrofoil dynamic is studied for different configurations : forced movements or not, immersed in a fluid at rest or a flowing one. It shows that the hydrofoil movement is pseudo-periodic followed by a damping movement. The hydrodynamic forces tend to follow the same evolution and converge to an equilibrium point. The vibration study clearly shows a frequency modification of the system that depends on the fluid flow (at rest or with an inflow). The problem is also coupled to center of pressure position's equation which depends on the hydrofoil position and the fluid flow. The trend of the position presents a singularity when the lift and drag coefficients vanishes at the same time.Last part, the equation that take into account the inhomogeneous characteristic of the fluid at the fluid-structure interface as well as sheet cavitation in steady or unsteady case, was developed. The method allows the separation of the fluid variables when flowing around the fixed hydrofoil on one hand and the flow generated by the hydrofoil vibration one the other. This introduces an asymmetric added mass operator and an added damping operation due to the variation of the density of the fluid at the interface in unsteady case.The whole system results in a natural frequencies and amplitudes modulation over time.

Um método lagrangeano-euleriano arbitrário para a resolução de escoamentos dominados por tensão superficial é apresentado neste trabalho. Tais escoamentos são importantes em muitas aplicações, especialmente em canais capilares que frequentemente aparecem em escoamentos em microescala. A resolução deste tipo de escoamento apresenta vários desafios que são abordados neste trabalho. O escoamento é resolvido somente para a fase líquida, com condições de contorno apropriadas para a superfície livre que delimita o líquido e o gás, que é representada por arestas e vértices da malha computacional. Esta se move e se deforma, sendo que sua qualidade é mantida sob controle para não degradar a solução numérica. As equações de Navier-Stokes são discretizadas pelo método de elementos finitos em um referencial arbitrário. O método de incorporação dos efeitos de tensão superficial e linha de contato é explicado em detalhes. Validações comprovam a precisão do método proposto, com comparações através de soluções pseudo-analíticas para casos simples. Finalmente alguns resultados sobre escoamentos em capilares são apresentados / An arbitrary lagrangian-eulerian finite element method to solve surface tension dominated flows is presented. Such flows are important in many applications, particularly in capillary channels, that appear in microscale flows. The resolution of such flows presents several challenges that are addressed in this work. The flow is solved only in the liquid phase, and proper boundary conditions are applied on the free-surface, bounding the liquid and gas, which is explicitly represented by vertices and edges of the computational mesh. The mesh is moved and deformed, but its quality is kept under control in order to control errors in the numerical solution. The Navier-Stokes equations are discretized by standard Galerkin finite element method in an arbitrary reference. Details of the computation of surface tension and contact line effects are presented. The methodology is validated for a number of simple test cases against known pseudo-analytical solutions, and numerical results are presented, showing the robustness and accuracy of the methodology. Finally, some results about surface-tension-driven flows in capillaries are presented

Cette thèse traite de la modélisation et de l'approximation numérique des écoulements liquide-gaz compressibles. La difficulté centrale est la modélisation et l'approximation de l'interface liquide-gaz. Le modèle bifluide est constitué d'un système de lois de conservation fermé par une loi d'état du mélange. La loi d'état conditionne les bonnes propriétés (hyperbolicité, existence d'une entropie de Lax) du système. Les schémas classiques de type Godunov conduisent à des imprécisions les rendant inutilisables en pratique. L'existence de solutions discontinues rend difficile la construction de schémas d'ordre élevé et nécessite des maillages très fins pour une précision acceptable. Il est indispensable de proposer des algorithmes performants pour les calculateurs parallèles les plus récents. Nous aborderons chacune de ces problématiques: construction d'une "bonne" loi de pression, construction de schémas numériques adaptés, programmation sur calculateur massivement multicoeur. / This thesis deals with the modeling and numerical approximation of compressible gas-liquid flows. The main difficulty lies in modeling and approximation of the liquid-gas interface. The two-fluid model is a system of conservation laws closed with a mixture pressure law. The law has to be chosen carefully, it conditions good properties of the system as hyperbolicity or existence of a Lax entropy. Classic conservative Godunov-type schemes lead to inaccuracies that make them unusable inpractice. The existence of discontinuous solutions makes it difficult to build high order schemes and requires very fine meshes to an acceptable accuracy. It is therefore essential to provide efficient algorithms for the High Performance Computing. In this thesis, we will partially treat each of these issues : construction of a "good" pressure law, building adapted numerical schemes, programming on GPU or GPU cluster.

This master thesis deals with stress-strain analysis of simplified model of the thin wall transformer case subjected to impact load of electrical blast. Electrical blast is replaced by chemical blast (detonation of high explosive). The problem is solved using computational modeling utilizing the Finite Element Method (FEM) and LS-DYNA solver. After the introduction where detonation and shock wave propagation is explained the analytical approach is presented. This approach serves to results verification. In the next chapter is conducted research of applicable methods from which ALE method is chosen. In preliminary study is performed the mesh size analysis that is focused on finding the size of element which is both computational effective and gives accurate results. Next the infulence of input conditions (shape, location and parametres of high explosive, location of detonation point, boundary conditions) on distribution and time progress of pressure is investigated. Then influence of the opening on upper side of the case on overall pressure redistribution and strain and stress of the case is analysed. The stress-strain analysis of the case´s door which are connected to case by various types of contact models is performed as well as stiffness analysis of these types of contact.

Title: Numerical solution of flows in time dependent domains with elastic walls Author: Martin Hadrava Department: Department of Numerical Mathematics Supervisor: prof. RNDr. Miloslav Feistauer, DrSc., dr. h. c. Supervisor's e-mail address: feist@karlin.mff.cuni.cz Abstract: This work is devoted to the numerical solution of flow in time dependent domains with elastic walls. This problem has several applications in engineering and medicine. The flow is described by the system of Navier-Stokes equations supplemented with suitable initial and boundary conditions. A part of the boundary of the region oc- cupied by the fluid is represented by an elastic wall, whose deformation is described by a hyperbolic partial differential equation with initial and boundary conditions. Its right- hand side represents the force by which the fluid flow acts on the elastic wall. The goal of this work is to elaborate a numerical method for solving this coupled problem based on the finite element method and the ALE formulation of the equations describing flow. A for- mulation and analysis of the problem together with discretization, algorithmisation and programming of modules, which were added to an existing software package, is presented. The method that was worked out is applied to solve test problems. Keywords: interaction of flow and an...

Gegenstand dieser Arbeit ist die Bereitstellung eines geeigneten Simulationswerkzeuges für die numerische Untersuchung der beim Rollkontakt zwischen Rad und Schiene auftretenden Phänomene. Hierbei liegt das Hauptaugenmerk auf der kontinuumsmechanischen Formulierung des mechanischen Feldproblems kontaktierender Körper sowie dessen numerischer Lösung mittels der Finite-Element-Methode. Zur Reduzierung des bei der Simulation von Rollkontakt aus der notwendigen sehr feinen Diskretisierung der Kontaktgebiete resultierenden numerischen Aufwandes wird eine relativkinematische Beschreibung herangezogen. Diese gemischte LAGRANGE-EULER-Betrachtungsweise beruht auf der Zerlegung der Bewegung in einen Starrkörperanteil und eine dazu relative Deformation. Die Herleitung der Bewegungsgleichung für das Kontaktproblem erfordert die relativkinematische Formulierung der kontinuumsmechanischen Grundgleichungen, d.h. der Bilanzgleichungen sowie der konstitutiven Beziehungen. Eine geeignete Kontaktmechanik einschließlich der Berücksichtigung des Kontakts rauer Oberflächen und veränderlicher Kontaktrandbedingungen ist ebenfalls notwendig. Die physikalische Einbindung der Körper in die Umgebung erfolgt über NEUMANNsche und DIRICHLETsche Randbedingungen. Auf dieser Basis können die Bewegungsgleichungen der Elastomechanik hergeleitet werden, welche sich jedoch einer analytischen Lösung verschließen. Somit werden sie in ihrer schwachen Form im integralen Mittel formuliert, was der Anwendung des Prinzips der virtuellen Verschiebungen als Ausgangspunkt für die numerische Lösung entspricht. Die rechentechnische Umsetzung erfordert die inkrementelle und diskrete Formulierung der Bewegungsgleichungen unter besonderer Beachtung der Trägheits-und Kontaktterme, wobei auf die Unterscheidung zwischen Haften und Gleiten beim Tangentialkontakt besonderes Augenmerk gelegt wird. Die numerische Lösung des Finite-Element-Gleichungssystems liefert den aktuellen Beanspruchungszustand zweier Körper im Rollkontakt. Die Funktionsfähigkeit der entwickelten Algorithmen wird abschließend anhand aussagekräftiger Beispielrechnungen zum statischen Kontakt und zum stationären Rollkontakt demonstriert, deren Ergebnisse gute Übereinstimmung mit analytischen Vergleichslösungen, soweit verfügbar, aufweisen. / Scope of this work is the preparation of a suitable simulation tool for the numerical investigation of rolling contact phenomena. The main focus lies on the continuum–mechanical formulation of the mechanical field problem of contacting bodies and its numerical solution within the framework of the Finite Element Method. For reducing the numerical effort in rolling contact simulation, induced by the necessity of a very fine discretization within the expected contact area, a relative–kinematical description is utilized. This arbitrary LAGRANGian–EULERian approach is based upon the decomposition of the total motion into a rigid body motion and a superimposed deformation. The derivation of the equation of motion for the contact problem requires the relative–kinematical formulation of the continuum–mechanical fundamental equations, i. e. the balance equations and the constitutive relations. A suitable contact model including the contact of rough surfaces and varying contact boundary conditions is also necessary. The physical embedding into the environment is accomplished by NEUMANN and DIRICHLET boundary conditions. Based upon that foundation the elastomechanics’ equations of motion are derived, which however can not be solved analytically in general. Hence, the equations of motion are transferred into their weak form by the application of the principle of virtual displacements serving for the numerical solution. The implementation of the problem demands for an incremental and discrete formulation of the equations, especially regarding the terms of inertia and the contact terms. Thereby, special attention has to be paid to the distinction between sticking and sliding within the framework of the tangential contact analysis. The numerical solution of the finite elements’ system of equations provides the state of stress, displacement and contact of two bodies in rolling contact. The reliability of the developed algorithms is finally verified by means of meaningful numerical examples for both static contact and for stationary rolling contact, whereby the numerical results coincide well with available analytical reference solutions.

This work is concerned with the theoretical analysis of the space-time discontinuous Galerkin method applied to the numerical solution of nonstationary nonlinear convection-diffusion problem in a time- dependent domain. At first, the problem is reformulated by the use of the arbitrary Lagrangian-Eulerian (ALE) method, which replaces the classical partial time derivative by the so-called ALE derivative and an additional convection term. Then the problem is discretized with the use of the ALE space-time discontinuous Galerkin method. On the basis of a technical analysis we obtain an unconditional stability of this method. An important step in the analysis is the generalization of a discrete characteristic function associated with the approximate solutionin a time-dependentdomainand the derivationof its properties. Further we derive an a priori error estimate of the method in terms of the interpolation error, as well as in terms of h and tau. Finally, some practical applications of the ALE space-time discontinuos Galerkin method in a time-dependent domain are given. We are concerned with the numerical solution of a nonlinear elasticity benchmark problem and moreover with the interaction of compressible viscous flow with elastic structures. The main attention is paid to the modeling of flow induced vocal fold...

Modelling the full event of an underwater explosion (UNDEX) is complex and requires advanced modelling methods in order to achieve accurate responses. The process of an UNDEX includes a series of events that has to be considered. When a detonation is initiated, a shock-wave propagates and the rest products from the explosive material creates a gaseous bubble with high pressure which pulsates and impacts the surroundings. Reflections of the initial shock-wave can also appear if it hits the sea floor, water surface or other obstacles. There are different approaches how to numerically model the impact of an UNDEX on a structure, some with analytical approaches without a water domain and others where a water domain has to be modelled. This master’s thesis focuses on two modelling methods that are available in the finite element software LS-DYNA. The simpler method is called Sub-Sea Analysis (SSA) and does not require a water domain, thus it can be beneficial to use in an early design stage, or when only approximated responses are desired. To increase the accuracy, a more complex method called S-ALE can be used. By implementing this method, the full process of an UNDEX can be studied since both the fluid domain and explosive material are meshed. These methods are studied separately together with a combination of them. Another important aspect to be considered is that oscillations of a structure submerged in water differs from the behavior it has in air. Depending on the numerical method used, the impact of the water can be included. Natural frequencies of structures submerged in water are studied, how it changes and how the methods takes this into account. To verify the numerical models, experiments were executed with a cylindrical test object where the distance and weight of charge were altered through out the test series. It was found that multiple aspects affects the results from the experiments, that are not captured in the numerical models. These aspects have for instance to do with reflections, how accurate the test object is modelled and the damping effects of the water. It is concluded that the numerical models are sensitive when small charges and fragile structures are studied. High frequency oscillations were not triggered in the experiment but found for both methods. It should be further investigated if the methods are more accurate for larger charges and stronger structures. Experiments with larger water domain would also be beneficial to reduce effects from reflections, as well as a more accurate model of the cylinder in the simulations.

Et si la santé était une chose trop sérieuse pour être confiée à des acteurs sanitaires ? Et si la fin des risques était possible ? Ces interrogations ont, entre autres, animé la planification en santé environnement(ale). Cette thèse sonde les genèse, conception et portée de cet instrument aussi sujet aux transferts et variations qu’aux aléas et controverses. Elle se penche sur l’OMS qui l’a initié, jusqu’au niveau national français où la planification est épousée de diverses manières. Une approche par les acteurs est privilégiée, un regard conjoint sur le déjà-là et le changement réclamé. Ainsi, l’analyse comprend la dimension politique de la planification comme dépendante d’un dépassement : celui d’acteurs du génie sanitaire, entrepreneurs historiques de santé environnement(ale), par des nouveaux groupes sanitaire et écologique disposant de ressources-clés dans l’appareil d’État. Plus particulièrement, la pérennité de la planification procède des acteurs écologiques. Renversant les projets de l’OMS, cette configuration suggère un glissement de propriété du problème santé environnement(ale). Depuis plus de vingt ans, la planification réduit l’ubiquité de la santé environnement(ale). Elle en esquisse aussi l’explosivité. À qui veut saisir la fabrique de ce problème, ses acteurs et débats, ses temps et grandeurs, ce travail propose quelques clés de lecture. Ces clés ouvrent sur l'intérieur, les sommets et les marges de l'État. S’y distingue une expertise systémique transcendant durablement sa pendante scientifique dans le gouvernement des relations humains/milieux. S’y dévoilent les sens composites des transversalités et leurs rapports vivaces aux sectorialités. / What if health was too serious to be left to health actors ? What if risks can disappear ? These interrogations have driven some of the national environmental health action plans. This specific planning is structurally a matter of transfers and changes, uncertainty and controversies. This thesis deals with its genesis, making, and effects. It integrates a transnational scene, the World Health Organization that initiates the planning. It also considers a national level, the French one precisely, that imports the planning. Two rationale are used : an actor-centered approach, a balanced examination of preexistence and innovation. This work thus foresees the political dimension of the national environmental health planning in France. It requires health and ecological state actors with adequate resources. These actors go beyond sanitary engineers that are historical advocates of environmental health. Besides, the ecological actors determine the continuity of the planning. This suggests that the owners of the environmental health problem have now changed. This also means a distance with the initial project of WHO. For more than twenty years, the national planning reduces the ubiquity of environmental health. It also suggests its exploding aspects. This thesis provides some insight to understand the making of this problem, its actors and debates, its times and quantities. It explores the inside, the highest level and the boundaries of the state. It enlightens the transcendence of systemic expertise with respect to scientific expertise for governing the human-environment relationships. More broadly, it leads to question the vigorous relation between the cross-cutting and sectoral aspects of public intervention.

A general finite element method is presented to solve the Euler equations in a Lagrangian reference frame. This FEM framework allows for separate arbitrarily high order representation of kinematic and thermodynamic variables. An accompanying hydrodynamics code written in Matlab is presented as a test-bed to experiment with various basis function choices. A wide range of basis function pairs are postulated and a few choices are developed further, including the bi-quadratic Q2-Q1d and Q2-Q2d elements. These are compared with a corresponding pair of low order bi-linear elements, traditional Q1-Q0 and sub-zonal pressure Q1-Q1d. Several test problems are considered including static convergence tests, the acoustic wave hourglass test, the Sod shocktube, the Noh implosion problem, the Saltzman piston, and the Sedov explosion problem. High order methods are found to offer faster convergence properties, the ability to represent curved zones, sharper shock capturing, and reduced shock-mesh interaction. They also allow for the straightforward calculation of thermodynamic gradients (for multi-physics calculations) and second derivatives of velocity (for monotonic slope limiters), and are more computationally efficient. The issue of shock ringing remains unresolved, but the method of hyperviscosity has been identified as a promising means of addressing this. Overall, the curvilinear finite elements presented in this thesis show promise for integration in a full hydrodynamics code and warrant further consideration.