Sediment erosion is one of the key challenges in hydraulic turbines from a design and maintenanceperspective in Himalayas and Andes. Past research works have shown that the optimization of theFrancis turbine runner blade shapes can decrease erosion by a signicant amount. This study conductedas a Master's Thesis has taken the proposed designs from past works and conducted a CFDanalysis on a single passage of a Francis runner blade to choose an optimized design in terms of erosionand eciency. Structural analyses have been performed on the selected design through one-way andtwo-way FSI to compare the structural integrity of the designs.Two types of cases have been considered in this thesis work to dene the boundary condition of thestructural model. In the rst case, a runner blade is considered to have no in uence of the joint andother stier components. In the second case, a sector of the whole runner has been modeled withnecessary boundary conditions. Both one-way and two-way FSI have been performed on the casesfor the designs. Mesh independent studies have been performed for the designs, but only for the rstcase, whereas in the second case, a ne mesh has been used to make the analysis appropriate.The loads have been imported into the structural domain from the uid on the interfaces for one-wayFSI. In the case of two-way FSI, the Multi-Field Solver (MFX) supported by ANSYS has been usedto solve the coupled eld analysis. A fully coupled FSI in ANSYS works by writing an input le inthe structural solver containing the information about the interfaces in the structural domain, whichis imported in the uid solver. The interaction between the two domains is dened in ANSYS-CFX,including the mesh deformation and solver setups. The results have been post-processed in CFX-Post,where the results from both the elds are included. It has been found that the structural integrity ofthe optimized design is better than the reference design in terms of the maximum stress induced inthe runner. The two-way FSI analysis has been found as an inevitable part of the numerical analysis.However, with the advancement of the computational capability in the future, there could be a greatscope in the research eld to carry out a fully-coupled transient simulation for the whole runner toget a more accurate solution.

Behaviour prediction of hydraulic steel structures with the view to surrounding influences in various design dispositions is a fundamental condition for operational reliability assessment of the analyzed construction. Reliable characteristics of construction behaviour defined by the specification of its movement within changes caused by time and environmental influences is of great importance. In currently used engineering mechanics formulation it concerns setting the response of the defined construction or its part to the given time variable mechanic load. Required response values, which are necessary for evaluation terminal dispositions of capacity and usability of the construction, are trans-location and tension, or values thence derived. Calculation is basic means for response prediction of construction. The thesis presented deals with complex multi-physical behaviour problems of water supply constructions in fluid structure interaction. There are presented various approaches to calculations of static and dynamic qualities of constructions. These approaches are divided into so called “direct method”, which is based on direct connection between two physical fields and the calculation is performed by the method of final elements, and so called “indirect method” , which is based on connection of two physical fields by means of various interfaces, which are described in this thesis. In case of indirect method, the calculation of running liquid is performed by the method of final volumes and the construction calculation is performed by the method of final elements. Within the scope of this thesis, static and dynamic responses of water supply constructions have been solved with the use of the above mentioned approaches. The results of the calculations in the scope of this thesis have been compared with the findings of performed experiments. The final part of the thesis describes the results and generalized findings gathered from the tasks by various approaches.

This work extends a previously developed finite-volume overset-grid fluid flow solver to enable the characterisation of rigid-body-fluid interaction problems. To this end, several essential components have been developed and blended together. The inherent time-dependent nature of fluid-solid interaction problems is captured through the laminar transient incompressible Navier-Stokes equations for the fluid, and the Euler-Newton equations for rigid-body motion. First and second order accurate time discretisation schemes have been implemented for the former, whereas second and third order accurate time discretisation schemes have been made available for the latter. Without doubt the main advantage the overset-grid method offers regarding moving entities is the avoidance of the time consuming grid regeneration step, and the resulting grid distortion that can often cause numerical stability problems in the solution of the flow equations. Instead, body movement is achieved by the relative motion of a body fitted grid over a suitable background mesh. In this case, the governing equations of fluid flow are formulated using a Lagrangian, Eulerian, or hybrid flow description via the Arbitrary Lagrangian-Eulerian method. This entails the need to guarantee that mesh motion shall not disturb the flow field. With this in mind, the space conservation law has been hard-coded. The compliance of the space conservation law has the added benefit of preventing spurious mass sources from appearing due to mesh deformation. In this work, two-way fluid-solid interaction problems are solved via a partitioned approach. Coupling is achieved by implementing a Picard iteration algorithm. This allows for flexible degree of coupling specificationby the user. Furthermore, if strong coupling is desired, three variants of interface under-relaxation can be chosen to mitigate stability issues and to accelerate convergence. These include fixed, or two variants of Aitken’s adaptive under-relaxation factors. The software also allows to solve for one-way fluid-solid interaction problems in which the motion of the solid is prescribed. Verification of the core individual components of the software is carried out through the powerful method of manufactured solutions (MMS). This purely mathematically based exercise provides a picture of the order of accuracy of the implementation, and serves as a filter for coding errors which can be virtually impossible to detect by other means. Three instances of one-way fluid-solid interaction cases are compared with simulation results either from the literature, or from the OpenFOAM package. These include: flow within a piston cylinder assembly, flow induced by two oscillating cylinders, and flow induced by two rectangular plates exhibiting general planar motion. Three cases pertaining to the class of two-way fluid-interaction problems are presented. The flow generated by the free fall of a cylinder under the action of gravity is computed with the aid of an intermediate ‘motion tracking’ grid. The solution is compared with the one obtained using a vorticity based particle solver for validation purposes. Transverse vortex induced vibrations (VIV) of a circular cylinder immersed in a fluid, and subject to a stream are compared with experimental data. Finally, the fluttering motion of a rectangular plate under different scenarios is analysed.

This thesis deals with determination of effect of wing flexibility on load distribution. FSI analysis using modal superposition was used for determination of effect of wing flexibility. Analysis was verified by analytic calculation. Differences of load distribution, between rigid and flexible wing, was determined for the selected flight regimes. Change of the bending moment was up to 3,9 %. Thesis shows importance of including effect of wing flexibility for sailplane design.

Předložená diplomová práce se zabývá interakcí mezi proudící krví a krční tepnou, obsahující aterosklerotický plat, za pomoci konečnoprvkové fluid-structure interaction analýzy. První část práce obsahuje souhrn teoretických poznatků, sestávající z kardiovaskulárního systému, cév, souvisejících konstitutivních modelů, reologie krve a úvodu do teorie proudění. Dále je v práci obsažen stručný souhrn současného poznání výpočtového modelování v této oblasti, s důrazem na strukturní a fluid-structure interaction analýzy v oblasti krční tepny, a na použité konstitutivní modely. Experimentální část se soustředí na tvorbu zjednodušeného modelu krční tepny, obsahující aterosklerotický plat, a na tvorbu odpovídajícího modelu krve. Oba modely poté společně vstupují do fluid-structure interaction analýzy, která si klade za cíl pochopit důsledky pulzujícího toku krve na stěnu tepny a na růst aterosklerotického plátu; primárními zkoumanými veličinami jsou první hlavní napětí na stěně tepny, celková deformace stěny tepny, časově zprůměrovaná hodnota smykového napětí na stěně tepny a oscilační smykový index. Všechny výsledky jsou porovnány napříč několika typy analýz, tak aby bylo možné zhodnotit rozdíly a důsledky zvoleného přístupu. Součástí práce je také zjednodušená parametrická studie, která porovnává vliv rostoucího procenta stenózy na vyhodnocované veličiny. V poslední částí práce jsou zhodnoceny výsledky, její limitace a další možnosti výzkumu v této oblasti.

This master’s thesis is made to increase the understanding of the dynamic characteristics of a typical large swing check valve used in a system that transports pressurized water to a reactor tank.3D FSI-simulations are performed for a number of transients in order to study the dynamic characteristics their dependence of the deceleration rate. The purpose is to find information about the dynamics that could be used in a future improvement of a 1D-model.Steady state simulations are performed for angles in the whole spectrum. Seven transient FSI-simulations with different constantly decelerating flows from 630 kg/s2 (6.7 m/s2) to 40 320 kg/s2 (430 m/s2) have been performed. The pressure on the disc caused by the hydraulic torque is integrated and the corresponding torque contribution, together with the weight torque, is used in the second law of motion to calculate the movement of the disc throughout the transients.Steady state simulations yield the pressure drop over the valve, which could be compared with field measurements in order to validate the CFD-simulations. Comparison of the pressure distribution on the disc for the steady state and transient simulations shows the importance of taking the disc angular velocity into account when modelling in 1D. Correlations between the angle, angular velocity, torque and mass flow are obtained from the transient FSI-simulations. Torque coefficients according to (Li & Liou, Vol. 125) are also brought out from the simulated transients, but in order to create a model in line with this approach further simulations have to be performed. A prediction of the pressure rise that occurs when a swing check valve closes in backward flow according to the Joukowsky equation is brought out and gives an idea of the loadings that the system has to be able to handle.

Dans le cadre d’une collaboration avec la société PHYTOREM, nous avons élaboré un prototype de distribution autorégulé afin d’épandre des Eaux Usées après un simple dégrillage et via la Phytorémédiation (dépollution par les plantes).La première approche du projet de thèse a été de comprendre les comportements rhéologiques des effluents, mis à disposition par Phytorem, et mécaniques du matériau élastomère de type EPDM. Nous avons exposé les différentes façons de retrouver leurs propriétés rhéologiques et mécaniques par le biais de divers tests de rhéométrie, concernant les effluents, et de traction uni-, bi- et équibi-axiale, pour la partie matériau. Ceci nous a permis d’obtenir d’une part, la viscosité de nos effluents, et d’autre part, la loi de comportement la mieux adaptée à notre matériau.La deuxième et dernière approche porte sur les interactions entre un fluide et une membrane hyperélastique ayant pour fonction de réguler un écoulement. Le comportement de la membrane contrainte par la pression a été simulé sous Abaqus. Ces résultats ont permis de modéliser l’écoulement (code CFD commercial) lorsque la membrane est déformée et de déterminer numériquement la loi débit/pression du dispositif. Ces développements numériques s’appuient sur la méthode des éléments finis et un couplage partitionné simple en une étape pour une première approche entre le fluide, la membrane et la structure. Les modèles numériques sont validés expérimentalement. Ces travaux participent à l’élaboration d’un prototype de distributeur auto-régulé. / In collaboration with PHYTOREM, we have developed a prototype of self-regulated drip emitter to spread the Wastewater after a simple screening using phytoremediation (remediation by plants).The first approach of the thesis project was to understand the rheological behaviour of waste provided by PHYTOREM, and mechanical properties behaviour of EPDM elastomer type. We have explained the different ways to find their rheological and mechanical properties through various rheometry tests on waste, and tension uni-, biand equibi-axiale, for the material part. This allowed us to obtain first, the viscosity of our waste, and secondly, the behaviour law of best suited to our material.The second and final approach focuses on the interactions between a fluid and a hyperelastic membrane whose function is to regulate flow. The membrane behaviour under pressure stress was simulated using Abaqus. These results were used to model the flow (commercial CFD) when the membrane is distorted and to determine numerically its flow versus pressure law. These developments are relying on numerical finite element method and partitionned into a single coupling step for a first approach between fluid, membrane and structure. The numerical models are validated experimentally. This work contributes to the development of a prototype of self-regulated drip emitter.

La thèse s'intéresse aux deux écoulements présents en aviron, autour du bateau et de la palette, et aux interactions avec le système bateau-avirons-rameur. Le premier est inhabituel en hydrodynamique, à cause du cavalement important et des mouvements secondaires. La complexité du second provient de l'instationnarité et de la déformation de la surface libre. L'objectif consiste à mettre en oeuvre des méthodes numériques performantes et précises puis à les valider pour, à plus long terme, les réutiliser à des fins d'analyse et d’optimisation de la performance en aviron.Ces simulations instationnaires à surface libre sont coûteuses en ressources pour les codes RANS. Un algorithme de sub-cycling a été développé et validé sur plusieurs cas test, diminuant les temps CPU d'un facteur 3 à 4, sans perte de précision. Il est compatible avec la déformation et le raffinement automatique de maillage. Deux bases de données expérimentales sont exploitées pour chaque écoulement afin de valider le cadre de simulation. Pour celui autour de la palette, une campagne in situ et une autre en laboratoire sont utilisées. Dans les deux cas, les profils d'efforts sont bien capturés, compte tenu des incertitudes cumulées liées à la mesure indirecte de la cinématique de la palette par rapport à l'eau. Pour le skiff en configuration instationnaire, les efforts fluctuants sont bien capturés, en amplitude et en phase, pour des fréquences typiques. Des écarts inattendus (de l'ordre de 10%) sont constatés sur la valeur moyenne et restent pour le moment sans réponse probante. La structure d'une co-simulation entre les résolutions des écoulements et celle de la dynamique du système multicorps est initiée. / The thesis focuses on the two flows occurring in rowing,around the boat and the blade, and on interactions with theboat-oars-rower system. The first flow is unusual in hydrodynamics because of the large surge and secondary motions. The complexity of the second one comes from the unsteadiness and the free surface deformation. The goal is to set up efficient and accurate numerical methods to reproduce these flows and then to validate them for the purpose of analysis and optimisation of the performance in rowing.Those unsteady computations with free surface are cost lyin resources for RANS codes. A sub-cycling algorithm was developed and validated on several test cases, allowing to decrease the CPU time by a factor of 3 to 4, without loss of accuracy. It is compatible with mesh deformation and automatic grid refinement. Two experimental databases are exploited for each flow in order to validate the frame of simulation. For the flow around the blade, an in-situ campaign and a more controlled one conducted in laboratory, are used. In both cases, the profiles of the efforts are well captured, considering the cumulative uncertainties linked to the indirect measurement of the blade kinematics relative to the water. For the skiff in unsteady state, the fluctuating forces are well captured, in terms of amplitudes and phases, for typical frequencies. Unexpected errors (around10%) are observed for the mean value and remain unexplained for now. The structure of a co-simulation between the resolutions of the flows and the resolution of the dynamics of the multibody system is initiated.

The aim of this study is to investigate the fluid-structural response to pulsatile Newtonian and non-Newtonian blood flow through an axisymmetric stenosed vessel using FLOTRAN and ANSYS. This is to provide a basic understanding of atherosclerosis. The flow was set to be laminar and follows a sinusoidal waveform. The solid model was set to have isotropic elastic properties. The Fluid-Structural Interaction (FSI) coupling was two-way and iterative. Rigid and Newtonian cases were investigated to provide an understanding on the effects of incorporating FSI into the model. The wall expansion was found to decrease the axial velocity and increase the recirculation effects of the flow. To validate the models and methods used, the results were compared with the study by Lee and Xu [2002] and Ohja et al [1989]. Close comparisons were achieved, suggesting the models used were valid. Two non-Newtonian models were investigated with FSI: Carreau and Power Law models. The Carreau model fluid behaviour was very close to the Newtonian model. The Power Law model produced significant difference in viscosity, velocity and wall shear stress distributions. Pressure distribution for all models was similar. In order to quantify the changes, Importance Factor (IG) was introduced to determine the overall non-Newtonian effects at two regions: the entire flow model and about the vessel wall. The Carreau model showed reasonable values of IG whereas the Power Law model showed excessive values. Transient and geometrical effects were found to affect the Importance Factor. The stress distributions for all models were found to be similar. Highest stress occurred at the shoulders of the stenosis where a stress concentration occurred due to sharp corners of the geometry and large bending moments. The highest stresses were in the axial direction. Notable circumferential stress was found at the ends of the vessel. Carreau model produced slightly higher stresses than the other models. Wall stresses were found to be primarily influenced by internal pressure, rather than wall shear stresses.

El presente trabajo de título forma parte del proyecto FONDECYT Flow Dynamics and Arterial Wall Interaction in Realistic Cerebral Aneurysm Models. El principal objetivo de este trabajo es optimizar la metodología de simulación de la mecánica de aneurismas cerebrales desarrollada en trabajos anteriores [4],[5] en los siguientes sentidos: crear una metodología de reconstrucción para dotar a las geometrías de espesor variable y cuantificar la sensibilidad de los resultados al variar el modelo constitutivo del material de la pared arterial, la cantidad de sifones de la vasculatura precedente al aneurisma, la condición de salida del flujo y el espesor de la pared. El primer objetivo se logró en colaboración con Rojo [2]. Para cumplir el segundo objetivo, se seleccionó uno de los nueve casos reconstruidos en este trabajo de título y se realizaron tres simulaciones estructurales puras, tres simulaciones CFD puras y cinco simulaciones FSI con el programa ADINA 8.3®. Mediante las simulaciones se pudo hallar la proporción en que diferían los resultados al variar los parámetros mecánicos antes mencionados. Se concluyó que la utilización de un modelo elástico equivalente es suficientemente buena aproximación para simular la mecánica de la pared arterial. Por otro lado, se concluyó que los resultados fluidodinámicos están fuertemente influenciados por la creación de flujo secundario producto de la curvatura de la vasculatura precedente al aneurisma. Al realizar las simulaciones, se debe considerar al menos un sifón suficientemente curvo en la geometría. Los resultados de las simulaciones FSI indican que las diferencias más grandes que se producen al variar cualquier parámetro se hallan en el sólido, de hecho los resultados fluidodinámicos, en su gran mayoría, no superan el 10 %. Por otro lado, el parámetro que genera mayor sensibilidad en los resultados es el espesor de la pared. Esta variable produce diferencias cercanas al 60% en el sólido, mientras que la variación de la condición de presión o la elasticidad de la pared, tan sólo producen diferencias del orden del 30 %. Se concluyó que la simulación más precisa desarrollada es la FSI que considera modelo de pared elástica Seshaiyer equivalente, presión normal a la salida y espesor de pared variable. En general, bajo esta metodología los esfuerzos y deformaciones son sobreestimados, mientras que el esfuerzo de corte en la pared es subestimado, lo cual la convierte en una prueba conservadora del riesgo de ruptura.