This is an experimental and theoretical study of a laminar separation bubble and the associated transition dynamics in its early stages. The separation of a laminar boundary layer from a solid surface is prevalent in very many flow situations such as over gas turbine blades (especially in the low-pressure turbine stage) and the wings of micro-aero-vehicles (MAVs) that operate at fairly low Reynolds numbers. Flow separation occurs in such cases due to the presence of an adverse pressure gradient. The separated shear layer becomes unstable due to the presence of an inflection point and presumably transitions to turbulence rapidly. Eventually, there is reattachment back to the solid surface further downstream, if conditions are right. The region enclosed by the shear layer is called a laminar separation bubble and has been a subject of many studies in the past. The present experiments have been conducted in a closed-circuit wind tunnel. A separation bubble was obtained on the upper surface of a flat plate by appropriately contouring the top wall of the tunnel. Four different techniques were used for qualitative and quantitative study viz. surface flow visualisation, smoke flow visualisation, surface pressure measurements and hotwire anemometry. Response of the bubble to both natural as well as artificial (impulsive excitation) disturbance environment has been studied. Linear stability analyses (both Orr-Sommerfeld and Rayleigh calculations), in the spatial framework, have been performed for the mean velocity profiles starting from an attached adverse pressure gradient boundary layer all the way up to the front portion of the separation bubble region (i.e. up to the end of the dead-air region where linear evolution of disturbances could be expected). The measured velocity profiles (both attached and separated) were fitted with analytical model profiles for doing stability calculations. A separation bubble consists of aspects of both wall-bounded and wall-free shear layers and therefore both viscous and inviscid mechanisms are expected to be at play. Most of the studies in the literature point to the inviscid instability associated with the shear layer to be the main mechanism. The main aim of the present work is to understand the exact origin of the primary instability mechanism responsible for the amplification of disturbances. We argue that at least up to the front portion of the bubble, the instability mechanism is due to the inflectional mode associated with the mean velocity profile. However, the seeds of this inviscid inflectional instability could be traced back to the attached boundary layer upstream of separation. In other words, the inviscid inflectional instability of the separated shear layer should be logically seen as an extension of the instability of the upstream attached adverse-pressure-gradient boundary layer. This modifies the traditional view that pegs the origin of the instability in a separation bubble to the free shear layer outside the bubble with its associated Kelvin-Helmholtz mechanism. Our contention is that only when the separated shear layer has moved considerably away from the wall (and this happens near the maximum height of the mean bubble) that a description by Kelvin-Helmholtz instability paradigm with its associated scaling principles could become relevant. We also propose a new scaling for the most amplified frequency for a wall-bounded shear layer in terms of the inflection point height and the vorticity thickness, and show its universality. Next, we theoretically investigate the role played by the re-circulating region of the separation bubble in the linear instability regime. In the re-circulating region near the wall, associated with the so-called wall mode, the production of disturbance kinetic energy is found to be negative. This is a very interesting observation which has been cursorily noted in earlier studies. Here we show that the near-wall negative production region exerts a stabilising influence on the downstream travelling disturbances. A theoretical support for such a mechanism to exist close to the wall is presented. It is shown that the stabilising wall-proximity effect is not a peripheral aspect but has a significant effect on the overall stability especially for the waves close to the upper neutral branch. We demonstrate the appropriateness of inviscid analysis for the stability of the separated flow velocity profile away from the wall, by comparing the numerical solutions of Rayleigh and Orr-Sommerfeld equations. Following this, the analytical consequences of the Rayleigh equation such as the inflection point criterion and the Fjortoft criterion are derived for the wall-bounded inflectional velocity profiles. Furthermore, we also discuss the relevance of the negative production region towards flow control and management for the wall-bounded flows. It appears fruitful to divide the separation bubble region into two parts with respect to the nature of disturbance dynamics: one outside the mean dividing streamline (which behaves as an amplifier) and the other inside the bubble corresponding to the re-circulating region (having oscillator type characteristics). To explore the oscillator-like behaviour of the bubble further, we have carried out spatio-temporal stability analysis of the reversed flow velocity profiles and determined the conditions for the onset of absolute instability. We contend that the presence of the negative production region for the upstream travelling waves has a restraining effect arresting the tendency of the flow (both wall-free and wall-bounded) to become absolutely unstable and thereby requiring a particular threshold of the backflow velocity to be crossed for its realisation. Moreover, the delay in the onset of absolute instability for a wall-bounded profile as compared to a free shear layer is attributed to a certain ‘negative-drag’ effect of the wall on the overall flow which increases the group velocities for the wall-bounded flows. A related theme in the literature regarding the dynamics of laminar separation bubbles is the so-called ‘bursting’ of the bubble wherein there is a sudden increase in the length and height of the bubble as some critical conditions are reached. Bubbles before bursting are termed as ‘short’ bubbles and those after bursting as ‘long’ bubbles. In this work, we provide a criterion to predict bursting which is a refinement over the existing criteria. The proposed criterion takes into account not just the length of the bubble but also the maximum height and it is shown to be more universal in differentiating short bubbles from the long ones, as compared to the other criteria. We also present a hypothesis regarding the sequence of events leading to bubble bursting by relating its onset to the instability of the re-circulating region. For this we observe that as the amount of backflow velocity is increased for a reversed flow velocity profile, the inflection point moves inside the mean dividing streamline and this happens before the onset of absolute instability. This causes a vorticity maximum to develop inside the re-circulating region which could lead to the instability of the closed streamlines with respect to two-dimensional cylindrical disturbances. The actual bursting process may be expected to involve non-linear interactions of the disturbances and the long bubble could be a nonlinearly saturated state of the instability of the re-circulating region. In order to explore the three-dimensionality associated with the bubble, extensive surface flow visualisation experiments have been performed. The surface streamline pattern is obtained for the entire span of the plate for three different freestream velocities. The patterns have been interpreted using topological ideas and various critical points have been identified. It is shown that the arrangement of critical points satisfies the ‘index theorem’ which is a topological necessity and the streamline patterns are ‘structurally stable’. An interesting observation from these patterns is the presence of three-dimensionality upstream of the separation line close to the wall even though the oncoming flow is nominally two-dimensional. Using the critical point theory, we propose a hypothesis which could be used to construct a semi-empirical model wherein the critical points are assigned with a quantity called ‘strength’ for determining the extent of upstream influence of a given separation line. Finally, we derive a necessary condition for the existence of inviscid spatial instability in plane parallel flows. It states that for spatial instability the curvature of the velocity profile should be positive in some region of the profile. This includes Rayleigh’s inflection point theorem (which was proposed and proved by Rayleigh for temporal instability) as a special case. It thus provides a rigorous basis for applying the inflection point criterion to the flows in the framework of spatial stability theory (which we have used extensively in the present thesis). Moreover, the condition derived here is more general as it also includes velocity profiles with the curvature positive everywhere which are excluded by Rayleigh’s theorem in the temporal framework. An example of such a profile is presented (Couette-Poiseuille flow with adverse pressure gradient) and it is shown that this flow is an exceptional case which is temporally stable but spatially unstable. Eigenvalue calculations as well as energy considerations suggest that the mechanism governing instability of this flow is inviscid and non-inflectional in character. This is a new result which could have important implications in understanding the instability dynamics of parallel flows.

Gradient-based aerodynamic shape optimization, based on Computational Fluid Dynamics analysis of the flow, is a method that can automatically improve designs of aircraft components. The prospect is to reduce a cost function that reflects aerodynamic performances. When the shape is described by a large number of parameters, the calculation of one gradient of the cost function is only feasible by recourse to techniques that are derived from the theory of optimal control. In order to obtain the best computational efficiency, the so called adjoint method is applied here on the complete mapping, from the parameters of design to the values of the cost function. The mapping considered here includes the Euler equations for compressible flow discretized on unstructured meshes by a median-dual finite-volume scheme, the primal-to-dual mesh transformation, the mesh deformation, and the parameterization. The results of the present research concern the detailed derivations of expressions, equations, and algorithms that are necessary to calculate the gradient of the cost function. The discrete adjoint of the Euler equations and the exact dual-to-primal transformation of the gradient have been implemented for 2D and 3D applications in the code Edge, a program of Computational Fluid Dynamics used by Swedish industries. Moreover, techniques are proposed here in the aim to further reduce the computational cost of aerodynamic shape optimization. For instance, an interpolation scheme is derived based on Radial Basis Functions that can execute the deformation of unstructured meshes faster than methods based on an elliptic equation. In order to improve the accuracy of the shape, obtained by numerical optimization, a moving mesh adaptation scheme is realized based on a variable diffusivity equation of Winslow type. This adaptation has been successfully applied on a simple case of shape optimization involving a supersonic flow. An interpolation technique has been derived based on a mollifier in order to improve the convergence of the coupled mesh-flow equations entering the adaptive scheme. The method of adjoint derived here has also been applied successfully when coupling the Euler equations with the boundary-layer and parabolized stability equations, with the aim to delay the laminar-to-turbulent transition of the flow. The delay of transition is an efficient way to reduce the drag due to viscosity at high Reynolds numbers.

Swirling flows abound in nature and numerous engineering applications. Under conditions which are not completely understood, the swirling cores could undergo a sudden enlargement of their vortex core, leading to a ’vortex breakdown’. The physics of vortex breakdown and strategies to control it have been active areas of research for nearly half a century. There are many competing theories of vortex breakdown in the literature; broadly, these are surmised on similarities to flow separation, hydrodynamic instability or transition from a supercritical to a subcritical state. However, a rational criterion for vortex breakdown continues to be elusive. One of the most well known criteria in the literature is the one due to Brown and Lopez (1990) based on an inviscid vortex dynamics model which suggests that the helix angle of the velocity vector should enclose the helix angle of the vorticity vector. However it appears that this only suggests that the stream surface would diverge and not necessarily constitute a condition for breakdown. In this work, we propose a new criterion based on helicity (scalar product of velocity and vorticity vectors) for characterizing breakdown since it has fundamental topological interpretations relating to change in linkages of vortex lines. In particular, it is suggested that the breakdown location corresponds to the location where helicity becomes zero. We study the problem of vortex breakdown in a cylindrical container with a rotating top lid in order to clarify and elucidate our hypothesis. We present results from Direct Numerical Simulation of this problem for three different Reynolds numbers and evaluate the utility of our proposed helicity criterion. Our studies indicate that helicity is indeed a better choice for characterizing vortex breakdown.

The interaction of a shock wave with a boundary layer is a classic viscous/inviscid interaction problem that occurs over a wide range of high speed aerodynamic flows. For example, on transonic wings, in supersonic air intakes, in propelling nozzles at offdesign conditions and on deflected controls at supersonic/transonic speeds, to name a few. The transonic interaction takes place at Mach numbers typically between 1.1 and 1.5. On an aerofoil, its existence can cause problems that range from a mild increase in section drag to flow separation and buffeting. In the absence of separation the drag increase is predominantly due to wave drag, caused by a rise in entropy through the interaction. The control of the turbulent interaction as applied to a transonic aerofoil is addressed in this thesis. However, the work can equally be applied to the control of interaction for numerous other occurrences where a shock meets a turbulent boundary layer. It is assumed that, for both swept normal shock and unswept normal shock interactions, as long as the Mach number normal to the shock is the same, then the interaction, and therefore its control, should be the same. Numerous schemes have been suggested to control such interaction. However, they have generally been marred by the drag reduction obtained being negated by the additional drag due to the power requirements, for example the pumping power in the case of mass transfer and the drag of the devices in the case of vortex generators. A system of piezoelectrically controlled flaps is presented for the control of the interaction. The flaps would aeroelastically deflect due to the pressure difference created by the pressure rise across the shock and by piezoelectrically induced strains. The amount of deflection, and hence the mass flow through the plenum chamber, would control the interaction. It is proposed that the flaps will delay separation of the boundary layer whilst reducing wave drag and overcome the disadvantages of previous control methods. Active control can be utilised to optimise the effects of the boundary layer shock wave interaction as it would allow the ability to control the position of the control region around the original shock position, mass transfer rate and distribution. A number of design options were considered for the integration of the piezoelectric ceramic into the flap structure. These included the use of unimorphs, bimorphs and polymorphs, with the latter capable of being directly employed as the flap. Unimorphs, with an aluminium substrate, produce less deflection than bimorphs and multimorphs. However, they can withstand and overcome the pressure loads associated with SBLI control. For the current experiments, it was found that near optimal control of the swept and unswept shock wave boundary layer interactions was attained with flap deflections between 1mm and 3mm. However, to obtain the deflection required for optimal performance in a full scale situation, a more powerful piezoelectric actuator material is required than currently available. A theoretical model is developed to predict the effect of unimorph flap deflection on the displacement thickness growth angles, the leading shock angle and the triple point height. It is shown that optimal deflection for SBLI control is a trade-off between reducing the total pressure losses, which is implied with increasing the triple point height, and minimising the frictional losses.

The interaction of a shock wave with a boundary layer is a classic viscous/inviscid interaction problem that occurs over a wide range of high speed aerodynamic flows. For example, on transonic wings, in supersonic air intakes, in propelling nozzles at offdesign conditions and on deflected controls at supersonic/transonic speeds, to name a few. The transonic interaction takes place at Mach numbers typically between 1.1 and 1.5. On an aerofoil, its existence can cause problems that range from a mild increase in section drag to flow separation and buffeting. In the absence of separation the drag increase is predominantly due to wave drag, caused by a rise in entropy through the interaction. The control of the turbulent interaction as applied to a transonic aerofoil is addressed in this thesis. However, the work can equally be applied to the control of interaction for numerous other occurrences where a shock meets a turbulent boundary layer. It is assumed that, for both swept normal shock and unswept normal shock interactions, as long as the Mach number normal to the shock is the same, then the interaction, and therefore its control, should be the same. Numerous schemes have been suggested to control such interaction. However, they have generally been marred by the drag reduction obtained being negated by the additional drag due to the power requirements, for example the pumping power in the case of mass transfer and the drag of the devices in the case of vortex generators. A system of piezoelectrically controlled flaps is presented for the control of the interaction. The flaps would aeroelastically deflect due to the pressure difference created by the pressure rise across the shock and by piezoelectrically induced strains. The amount of deflection, and hence the mass flow through the plenum chamber, would control the interaction. It is proposed that the flaps will delay separation of the boundary layer whilst reducing wave drag and overcome the disadvantages of previous control methods. Active control can be utilised to optimise the effects of the boundary layer shock wave interaction as it would allow the ability to control the position of the control region around the original shock position, mass transfer rate and distribution. A number of design options were considered for the integration of the piezoelectric ceramic into the flap structure. These included the use of unimorphs, bimorphs and polymorphs, with the latter capable of being directly employed as the flap. Unimorphs, with an aluminium substrate, produce less deflection than bimorphs and multimorphs. However, they can withstand and overcome the pressure loads associated with SBLI control. For the current experiments, it was found that near optimal control of the swept and unswept shock wave boundary layer interactions was attained with flap deflections between 1mm and 3mm. However, to obtain the deflection required for optimal performance in a full scale situation, a more powerful piezoelectric actuator material is required than currently available. A theoretical model is developed to predict the effect of unimorph flap deflection on the displacement thickness growth angles, the leading shock angle and the triple point height. It is shown that optimal deflection for SBLI control is a trade-off between reducing the total pressure losses, which is implied with increasing the triple point height, and minimising the frictional losses.

Submitted by Maike Costa (maiksebas@gmail.com) on 2016-03-28T14:37:42Z No. of bitstreams: 1 arquivototal.pdf: 2200397 bytes, checksum: fa2b77afd6348b68a616a33acb7c7cb2 (MD5) / Made available in DSpace on 2016-03-28T14:37:42Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 2200397 bytes, checksum: fa2b77afd6348b68a616a33acb7c7cb2 (MD5) Previous issue date: 2014-03-20 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / The aim of this thesis is to present some controllability results for some fluid mechanic models. More precisely, we will prove the existence of controls that steer the solution of our system from a prescribed initial state to a desired final state at a given positive time. The two first Chapters deal with the controllability of the Burgers-α and Leray-α models. The Leray-α model is a regularized variant of the Navier-Stokes system (α is a small positive parameter), that can also be viewed as a model for turbulent flows; the Burgers-α model can be viewed as a related toy model of Leray-α. We prove that the Leray-α and Burgers-α models are locally null controllable, with controls uniformly bounded in α. We also prove that, if the initial data are sufficiently small, the pair state-control (that steers the solution to zero) for the Leray-α system (resp. the Burgers-α system) converges as α → 0+ to a pair state-control(that steers the solution to zero) for the Navier-Stokes equations (resp. the Burgers equation). The third Chapter is devoted to the boundary controllability of inviscid incompressible fluids for which thermal effects are important. They will be modeled through the so called Boussinesq approximation. In the zero heat diffusion case, by adapting and extending some ideas from J.-M. Coron [14] and O. Glass [45], we establish the simultaneous global exact controllability of the velocity field and the temperature for 2D and 3D flows. When the heat diffusion coefficient is positive, we present some additional results concerning exact controllability for the velocity field and local null controllability of the temperature. In the last Chapter, we prove the local exact controllability to the trajectories for a coupled system of the Boussinesq kind, with a reduced number of controls. In the state system, the unknowns are: the velocity field and pressure of the fluid (y, p), the temperature θ and an additional variable c that can be viewed as the concentration of a contaminant solute. We prove several results, that essentially show that it is sufficient to act locally in space on the equations satisfied by θ and c. / O objetivo desta tese é apresentar alguns resultados controlabilidade para alguns modelos da mecânica dos fluidos. Mais precisamente, provaremos a existência de controles que conduzem a solução do nosso sistema de um estado inicial prescrito à um estado final desejado em um tempo positivo dado. Os dois primeiros Capítulos preocupam-se com a controlabilidade dos modelos de Burgers-α e Leray-α. O modelo de Leray-α é uma variante regularizada do sistema de Navier-Stokes (α é umparâmetro positivo pequeno), que pode também ser visto como um modelo de fluxos turbulentos; já o modelo Burgers-α pode ser visto como um modelo simplificado de Leray-α. Provamos que os modelos de Leray-α e Burgers-α são localmente controláveis a zero, com controles limitados uniformemente em α. Também provamos que, se os dados iniciais são suficientemente pequenos, o par estado-controle (que conduz a solução a zero) para o sistema de Leray-α (resp. para o sistema de Burgers-α) converge quando α → 0+ a um par estado-controle (que conduz a solução a zero) para as equações de Navier-Stokes (resp. para a equação de Burgers). O terceiro Capítulo é dedicado à controlabilidade de fluidos incompressíveis invíscidos nos quais os efeitos térmicos são importantes. Estes fluidos são modelados através da então chamada Aproximação de Boussinesq. No caso emque não há difusão de calor, adaptando e estendendo algumas idéias de J.-M. Coron [14] e O. Glass [45], estabelecemos a controlabilidade exata global simultaneamente do campo velocidade e da temperatura para fluxos em 2D e 3D. Quando o coeficiente de difusão do calor é positivo, apresentamos alguns resultados sobre a controlabilidade exata global para o campo velocidade e controlabilidade nula local para a temperatura. No último Capítulo, provamos a controlabilidade exata local à trajetórias de um sistema acoplado do tipo Boussinesq, com um número reduzido de controles. Nesse sistema, as incógnitas são: o campo velocidade e a pressão do fluido (y, p), a temperatura θ e uma variável adicional c que pode ser vista como a concentração de um soluto contaminante. Provamos vários resultados, que essencialmente mostram que é suficiente atuar localmente no espaço sobre as equações satisfeitas por θ e c.

Cette thèse vise à une meilleure compréhension de la dynamique du voilier et à la validation des outils numériques de prédiction de performances et d’optimisation par l'étude expérimentale in situ du problème aéro-élastique d'un gréement. Une instrumentation est développée sur un voilier de 8m de type J80 pour la mesure dynamique des efforts dans le gréement, de la forme des voiles en navigation, du vent et des attitudes du bateau. Un effort particulier est apporté à la mesure des caractéristiques géométriques et mécaniques des éléments du gréement, la calibration des capteurs et au système d'acquisition des données. Les principaux résultats montrent que le voilier instrumenté est un outil adapté pour les mesures instationnaires et soulignent l'amplitude de variation d'effort rencontrée en mer (20 à 50% de l'effort moyen dans une houle modérée). En outre, les variations du signal d'effort sont déphasées avec l'angle d'assiette, créant un phénomène d'hystérésis. Le comportement dynamique d'un voilier en mouvement diffère ainsi de l'approche quasi-statique. Les simulations numériques proviennent du code ARAVANTI, couplage implicite d’un code structure éléments finis ARA et d’un code fluide parfait, limitant son domaine de validité aux allures de près Les résultats de simulation sont très proches des cas stationnaires et concordent bien avec les mesures en instationnaire dans une houle de face. L'expérimentation numérique d'un gréement soumis à des oscillations harmoniques en tangage souligne l'importance de l'approche Interaction Fluide Structure (IFS) et montre que l’énergie échangée par le système avec la houle est reliée à la fréquence réduite et l'amplitude du mouvement. Certaines informations n'étant pas disponibles sur le voilier instrumenté, une expérience contrôlée en laboratoire est développée. Elle consiste en un carré de tissu tenu par deux lattes en oscillation forcée. Les mesures sur cette « voile oscillante » permettent d'étudier les phénomènes IFS avec décollement et sont utilisées pour la validation du couplage ARA-ISIS entre un code fluide Navier-Stokes (RANS) et le même code structure. / This work presents a full scale experimental study on the aero-elastic wind/sails/rig interaction in real navigation condition with the aim to give a reliable database of unsteady measurement. This database is used for the investigation of the dynamic behavior and loads in the rigging and for an experimental validation of an unsteady Fluid Structure Interaction (FSI) model. An inboard instrumentation system has been developed on a 8 meter yacht (J80 class) to simultaneously and dynamically measure the navigation parameters, yacht's motion, sails flying shape, wind and loads in the rigging. A special effort is made on mechanical and geometrical characteristics measurement, sensors calibration and data acquisition system synchronization. Results show that the instrumented boat is a reliable tool to measure the unsteady phenomena in navigation. Dynamic measurements at sea underline the load variation encountered, which represent 20 to 50% of the mean value in a moderate sea state. Oscillations of loads exhibit phase shift with the trim angle, reason for an hysteresis phenomenon, which shows that the dynamic behavior of a sail plan subject to yacht motion clearly deviates from the quasi-steady theory. Simulations are made with ARAVANTI, an implicit coupling of a Finite Element Method structural model ARA and an inviscid fluid model which restricts the simulation domain to upwind conditions. The simulation results compare very well with the experimental data for steady sailing conditions and show a good agreement in unsteady conditions (head swell). Numerical investigation of a sail plan submitted to harmonic pitching motion underlines the importance of FSI modeling and shows that the energy exchanged by the system with the swell increases with the motion reduced frequency and amplitude. Some information is not accessible on the instrumented boat and requires developing a controlled test case in laboratory. The experiment consists of a spinnaker fabric square mounted on two carbon battens moved in forced oscillation. This test case is used to study FSI phenomena with a separated flow and gives experimental results for the validation of the coupling ARA-ISIS of a RANS fluid model with the same structure model.

These thesis review some recent results on the construction of very high order multidimensional upwind schemes for the solution of steady and unsteady conservation laws on unstructured triangular grids.<p>We also consider the extension to the approximation of solutions to conservation laws containing second order dissipative terms. To build this high order schemes we use a subtriangulation of the triangular Pk elements where we apply the distribution used for a P1 element.<p>This manuscript is divided in two parts. The first part is dedicated to the design of the high order schemes for scalar equations and focus more on the theoretical design of the schemes. The second part deals with the extension to system of equations, in particular we will compare the performances of 2nd, 3rd and 4th order schemes.<p><p>The first part is subdivided in four chapters:<p>The aim of the second chapter is to present the multidimensional upwind residual distributive schemes and to explain what was the status of their development at the beginning of this work.<p>The third chapter is dedicated to the first contribution: the design of 3rd and 4th order quasi non-oscillatory schemes.<p>The fourth chapter is composed of two parts: we start by understanding the non-uniformity of the accuracy of the 2nd order schemes for advection-diffusion problem. To solve this issue we use a Finite Element hybridisation.<p>This deep study of the 2nd order scheme is used as a basis to design a 3rd order scheme for advection-diffusion.<p>Finally, in the fifth chapter we extend the high order quasi non-oscillatory schemes to unsteady problems.<p>In the second part, we extend the schemes of the first part to systems of equations as follows:<p>The sixth chapter deals with the extension to steady systems of hyperbolic equations. In particular, we discuss how to solve some issues such as boundary conditions and the discretisation of curved geometries.<p>Then, we look at the performance of 2nd and 3rd order schemes on viscous flow.<p>Finally, we test the space-time schemes on several test cases. In particular, we will test the monotonicity of the space-time non-oscillatory schemes and we apply residual distributive schemes to acoustic problems. / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

Ce travail s’inscrit dans le projet VOILENav qui vise à améliorer la compréhension des phénomènes d’Interaction Fluide-Structure appliqués aux voiles. Des comparaisons numériques expérimentales sont réalisées sur des mesures « in situ » au près à l’aide d’un code fluide parfait. Un critère, fondé sur l’équilibre du couple aérodynamique avec le couple de redressement, est proposé, permettant de vérifier l’hypothèse d’un écoulement attaché. Les précédentes études sur un voilier instrumenté ont montré les limites d’une approche « in situ » de par l’instationnarité naturelle liée aux évolutions du vent et de l’état de mer. Les autres limites résident dans la mesure de ces dernières – et tout particulièrement la mesure du vent réel – ainsi que dans le spectre des conditions rencontrées au réel. Des essais en soufflerie sont ainsi réalisés dans le cadre de ces travaux pour répondre, par une approche systématique et contrôlée, aux interrogations soulevées par les mesures « in situ ». Deux campagnes expérimentales successives, soutenues par le programme d’échange Sailing Fluids ont été menées dans la soufflerie du Yacht Research Unit de l’Université d’Auckland se focalisant sur les essais de voiles au près puis au portant. Les essais au près sont réalisés sur trois modèles réduits de grand-voiles d’IMOCA60 dans des conditions de réglages statiques et dynamiques. Le meilleur réglage statique est obtenu grâce à l’utilisation d’un algorithme d’optimisation original puis l’influence de l’amplitude et de la fréquence du « pumping » sont étudiés. Les performances aérodynamiques du système soumis à un réglage dynamique sont supérieures à celles du réglage optimum statique et un maximum est observé autour d’une fréquence réduite de 0.25 à 0.3. Au portant, les effets de l’instationnarité naturelle du spinnaker connue sous le terme « curling » (repliement du bord d’attaque) sont étudiés. Quatre modèles de spinnakers de J80 de forme identique sont testés pour différents matériaux et différentes coupes. Les mesures en soufflerie montrent que, pour des angles de vent apparent supérieurs à 100°, l’apparition du « curling » conduit à une augmentation de la force propulsive pouvant atteindre 10%. Les effets de la vitesse et de l’angle de vent apparent sont également étudiés et permettent d’extraire une fréquence réduite de curling indépendante de la vitesse de l’écoulement de 0.4 pour un vent apparent de 120°. L’étendue de la gamme de mesures explorées et le soin particulier apporté aux données expérimentales font de ces travaux une base de données remarquable pour des comparaisons avec des simulations de l’Interaction Fluide-Structure. / This work is part of the VOILENav project which aims to improve the understanding of Fluid-Structure Interaction applied to sails. Full-scale numerical experimental comparisons are achieved in upwind conditions with an inviscid flow code. A criterion using the equilibrium between the righting and heeling moment is suggested to check the attached flow hypothesis. Previous fullscale studies on instrumented boat are limited by the natural unsteadiness of wind and sea conditions and the measurement of these conditions. True wind computation and the wide range of encountered sailing conditions are still challenging. Complementary wind tunnel tests are carried out in this PhD project, using controlled conditions, to address some issues observed at full-scale. Thanks to the Sailing Fluids collaboration, two experimental campaigns in the Twisted Flow Wind Tunnel of the Yacht Research Unit of the University of Auckland have investigated upwind and downwind conditions. Upwind tests investigate static and dynamic trimming on three model IMOCA60 mainsails. The optimum static trim is determined thanks to an innovative optimization algorithm then the pumping amplitude and frequency are investigated. Aerodynamic performances under dynamic trimming are better than the optimum static trim with a maximum located for a reduced frequency about 0.25 to 0.3. For the downwind test, the natural unsteadiness known as curling (repeated foldingunfolding of leading edge) is studied. Four model J80 spinnakers with identical design shape are tested with different materials and cuts. Wind tunnel measurements show that for apparent wind angles higher than 100°, the curling apparition increases the drive force by up to 10%. Wind speed and wind angle effects are investigated and show a reduced curling frequency of 0.4 independent from the flow velocity for an apparent wind angle of 120°. The variety of the experimental conditions tested makes this work a precious database for Fluid Structure Interaction numerical-experimental comparison in the future.

The main objective of this Master´s thesis is the appropriate calculation proposal of pressure and discharge conditions in extensive ducts in unsteady flow. The calculation proposal was aimed at the conenction of two commercial programmes. Exacly the programme Ansis Fluent and Matlab, which deals with the connection of onedimensional (1D) calculation in Matlab and multidimensional (2D) calculation in Ansys Fluent programme. This Mastr’s thesis also deals with creation of the independent 1D model (Matlab, method of characteristic) and independent 2D model flow (Ansys Fluent, Inviscid model).