Hydrodynamic and thermal analyses performed during the conceptual design of an unmanned underwater vehicle are presented in this study. The hull shape is determined by considering alternative shapes and the dimensions are determined from the internal arrangement of components. Preliminary thermal analyses of the watertight section are performed with a commercial software called FLUENT to check the risk of over-heating due to the heat dissipation of devices. Performance of the proposed hull design is analyzed by FLUENT. Before simulations of the vehicle, validation studies are performed. Models 4159, 4158 and 4154 of Series 58 are chosen as the experimental reference. Their total resistance coefficients are compared with the results of the validations analyses. Mesh densities, turbulence models, near wall modeling approaches and inlet turbulence intensities are varied to understand their effects on the accuracy of predictions. A suitable turbulence modeling approach is chosen to analyze forward and vertical motions of the vehicle to check whether speed requirements are fulfilled. Hull configurations with and without appendages are used to observe their effects on total drag. It is observed that the proposed design satisfies speed requirements of the vehicle and no overheating is expected in the watertight section.

Inter-turbine burners are useful devices for increasing engine power. To reduce the size of these combustion devices, ultra-compact combustor (UCC) concepts are necessary. One such UCC concept is the centrifugal-force based high-g combustor design. Here, a model ultra-compact combustor (UCC) with fuel spray injection is simulated using large eddy simulation (LES) and Reynolds-Averaged Navier-Stokes (RANS) methodologies to understand mixing and spray dispersion inside centrifugal-based combustion systems. Both non-evaporating and evaporating droplet simulations were carried, as well as the tracking of a passive scalar, to explore this multiphase system. Simulation results show that mixing of fuel and oxidizer is based on a jet-in-crossflow system, with the fuel jet issuing into a circulating oxidizer flow stream. It is seen that a a high velocity vortex-like ring develops in the inner core of the combustor, which has enough momentum to obstruct the path of combustion products. There is minimal fuel droplet and vapor segregation inside the combustor and enhanced turbulent mixing is seen at mid-radius. / text

In a novel application of the temporal numerical simulation, an investigation ofturbulence modeling techniques is carried for the turbulent wall jet over aconvex surface (Coanda wall jet.) The simultaneous presence of multipleinstability mechanisms and the interaction with the turbulence dynamics at thesolid boundary produces a unique combination of different large turbulentcoherent structures that constitutes both a consistent challenge for numericalsimulations and an ideal test bed for turbulence models.The Temporal Direct Numerical Simulation (TDNS) of the Coanda wall jetrestricts the focus from the global turbulent Coanda wall jet to a smaller, localportion of the flow and offers a wide array of advantages to the present work. Inparticular, the size of the computational domain can be arbitrarily chosen inboth the spanwise and the streamwise directions. This allows to either suppressor enhance individual physical mechanisms and, consequently, to selectivelyreproduce different large coherent structures within the local flow. In the firstpart, temporal numerical simulations are employed to reproduce four differentflow scenarios of the local Coanda wall jet with a level of numerical resolutionthat, because of the reduced size of the computational domain, cannot be matchedby standard DNS of the entire physical flow (spatial DNS, or SDNS.)The TDNS of these four flow scenarios are then used in the second part for ana--posteriori analysis of different turbulence models in order to addresscommon shortcomings shown by Hybrid Turbulence Models (HTM). For each flowscenario, the turbulent flow field is deliberately decomposed in resolved andunresolved flows by the application of different filters in space correspondingto different grid resolution. The behavior of turbulence models can be reproducedfrom the resolved flow and compared to the turbulent stress tensor directlycalculated from the unresolved part of the flow field. Starting from the RANSlimit, turbulence models with different levels of complexity are studied.Successively, the performance of these models is analyzed at intermediatenumerical resolutions between RANS, LES, and DNS. Finally, an improvedformulation of the Flow Simulation Methodology (FSM) is proposed.

This thesis presents the implementation of several two-equation turbulence models into a finite difference, two- and three-dimensional Navier-Stokes Solver. Theories of turbulence modeling and the historical development of these theories are briefly investigated. Turbulence models that are defined by two partial differential equations, based on k-&amp / #969 / and k-&amp / #949 / models, having different correlations, constants and boundary conditions are selected to be adapted into the base solver. The basic equations regarding the base Navier-Stokes solver to which the turbulence models are implemented presented by briefly explaining the outputs obtained from the solver. Numerical work regarding the implementation of turbulence models into the base solver is given in steps of non-dimensionalization, transformation of equations into generalized coordinate system, numerical scheme, discretization, boundary and initial conditions and limitations. These sections of implementation are investigated and presented in detail with providing every steps of work accomplished. Certain trial problems are solved and outputs are compared with experimental data. Solutions for fluid flow over flat plate, in free shear, over cylinder and airfoil are demonstrated. Airfoil validation test cases are analyzed in detail. For three dimensional applications, computation of flow over a wing is accomplished and pressure distributions from certain sections are compared with experimental data.

Les premières simulations numériques d'écoulements réactifs sur des configurationscomplexes ont été réalisées à l'aide d'approches RANS (Reynolds Averaged Navier Stokes). Ces dernières, bien adaptées aux écoulements de type couches limites attachées et relativement peu coûteuses en temps de calcul, ne donnent accès qu'à des résultats stationnaires qui s'éloignent parfois de la réalité. Pour réaliser des simulations instationnaires d'écoulements, les méthodes de type LES (Large Eddy Simulation) -- plus précises mais plus coûteuses -- sont de plus en plus utilisées. Cependant, ces méthodes sont mal adaptées à la simulation de la dynamique pariétale, car elles nécessitent un effort de maillage souvent prohibitif près de la paroi. Cette thèse est consacrée au développement dans le code CEDRE (code de simulation d'écoulements réactifs complexes de l'Onera) d'une méthode hybride RANS/LES, appelée Delayed Detached Eddy Simulation (DDES), et à son application à des écoulements réactifs massivement décollés. Après une étape de validation sur des couches limites attachées, la DDES a été appliquée à la simulation des écoulements inerte et réactif dans une chambre de combustion en forme de marche descendante (A3C) et comparée aux résultats des approches RANS et LES classiques, ainsi qu'aux résultats expérimentaux. Cette méthode a ensuite permis de réaliser l'étude de la dynamique de l'écoulement réactif décollé dans la tuyère ATAC montée sur le banc cryotechnique MASCOTTE de l'Onera. / The first numerical simulations of reactive flows on complex configurations were performed using RANS (Reynolds Averaged Navier Stokes) approaches. These methods, which are well adapted to attached boundary layer flows and relatively not expensive in computation time, provide only steady results, which may not correctly reproduce reality. For unsteady flow simulations, LES (Large Eddy Simulation) method -- more accurate but more expensive – are increasingly employed. However, these approaches are poorly suited to simulate wall turbulence since they often require a prohibitive meshing effort close to the wall. This PhD thesis is devoted to the development in the CEDRE code (Onera code for the simulation of complex reactive flows) of a hybrid RANS/LES approach, called Delayed Detached Eddy Simulation (DDES) and to its application to massively separated reactive flows. After a step of validation on attached boundary layers, the DDES is applied to the simulation of the inert and reactive flows in a backward facing step combustor (A3C) and compared to the results obtained by RANS and LES approaches, as well as to experimental results. The DDES has then been used to study the dynamics of the separated reactive flow in the ATAC nozzle mounted on the cryotechnic MASCOTTE facility of Onera.

Au sein d’un échangeur à air, les trompes à air permettent de créer l’écoulement d’air froid nécessaireà son bon fonctionnement. Ces dispositifs, qui peuvent ^etre assimilés à des jets subsoniques confinésen conduit, peuvent contribuer au bruit rayonné par les avions lors des phases au sol. Nous proposonsdans cette thèse de développer un outil numérique prédictif de l’acoustique rayonnée par ces dispositifsafin de pouvoir proposer des solutions de réduction de bruit appropriées. Cet outil est adapté au contexteindustriel de Liebherr Aerospace. Une méthode stochastique permet, à partir d’un calcul stationnaireRANS, de générer un champ de vitesse turbulente qui autorise la formation d’un terme de forçage dansles équations d’Euler linéarisées qui sont alors utilisées comme un opérateur de propagation. Un nouveaumodèle stochastique basé sur l’hypothèse de sweeping est développé. Ce dernier permet de produiredes champs instationnaires respectant certaines propriétés aérodynamiques statistiques dans le cadre dejets libres subsoniques. Cette méthode est couplée avec le solveur Euler de l’Onera sAbrinA_v0 et l’outilrésultant est appliqué sur le cas d’un jet libre subsonique à M=0.72. Moyennant une calibration duterme source, la méthodologie permet de reproduire les spectres acoustiques en champ lointain, exceptépour les angles faibles. L’outil numérique est ensuite couplé avec un solveur FW-H pour étudier le casconcret de la trompe à air. Les résultats aérodynamiques et acoustiques sont validés par comparaison àune base de données aérodynamique et acoustique constituée au préalable à partir d’une campagne d’essaiscomprenant des mesures par anémométrie laser Doppler à l’intérieur du conduit et des microphonesacoustiques en champ lointain. / In air exchangers, the cool air flow can be produced by jet pumps. These devices, which can be consideredas subsonic jets confined in ducts, could contribute directly to ramp noise. A predictive numerical toolof the acoustic radiated by jet pumps is therefore developped in order to be able to propose appropriatenoise reduction solutions. This tool is adapted to Liebherr Aerospace’s industrial context. A stochasticmethod allows, starting from a steady RANS computation, to synthetise a turbulent velocity fields andto enforce source terms in the linearized Euler equations therefore used as a wave propagator. A newstochastic model relying on the sweeping hypothesis is developped. Unsteady fields reproducing someaerodynamics features of a free subsonic jet flow can be generated. This method is then coupled withOnera’s Euler solver sAbrinA_v0 and the resulting tool is applied on a free subsonic jet configuration atMach 0.72. Assuming a cabration of the source terms, this methodology models properly the far fieldacoustic spectra except for small angles. The numerical tool is then coupled with a FW-H solver to studya realistic jet pump. Aerodynamic and acoustic results are validated by comparison with a data baseobtained from an experimental campaign including laser Doppler anemometry measures inside the ductand pressure recording in the far-field.

La turbulence axisymétrique est un écoulement bidimensionnel trois-composantes. L’étude de ce type de turbulence est motivée par le fait que celle-ci représente la limite asymptotique des écoulements anisotropes, et qu’elle a été le sujet des investigations théoriques dans le passé. Dans ce manuscrit, la turbulence axisymétrique a étudié en géométrie fermée en utilisant des simulations numériques spectrales et pseudo-spectrales.Études antérieures concernant la génération des structures cohérentes, obtenues dans les écoulements en déclin libre, sont considérées ici dans le contexte des écoulements statistiquement stationnaires, où l’énergie est injectée soit par un forçage spectralement localisé ou par une rotation des disques en haut et en bas du cylindre. On montre que les structures observées sont conformes aux prédictions théoriques.Lorsqu’un protocole de forçage anisotrope est utilisé, une bifurcation est observée entre un état non-tourbillonnant (bidimensionnel deux-composantes, 2D2C) et un écoulement tourbillonnant turbulent (bidimensionnel trois-composante, 2D3C). Cette transition est modélisée à travers un système de deux équations différentielles ordinaires (ODE), et on montre que ce modèle retient la physique essentielle de cette transition. La transition de l’écoulement axisymétrique à un écoulement tridimensionnel (3D3C) est ensuite étudiée à l’aide d’une dimension non-entière, en introduisant de façon continue la variation azimutale dans le système. On montre que la limite 2D2C est singulière et qu’une petite variation azimutale permet une redistribution d’énergie sur les différentes composante énergétiques. Le modèle ODE est adapté pour ce système et on montre que pour l’écoulement considéré la corrélation pression-déformation est responsable d'un niveau approximativement proportionnel à la dimension non-entière. Des Simulations des Grandes Echelles sont réalisées pour évaluer la robustesse des observations à grands nombres de Reynolds. / Axisymmetric turbulence is a two-dimensional three-component flow. The investigation of this type of turbulence is motivated by the fact that it represents the asymptotic limit of anisotropic flows and since it has been the subject of theoretical investigations in the past. In the present manuscript such a flow is investigated in wall-bounded cylindrical geometry using spectral and pseudo-spectral numerical simulations.Previous results on the generation of coherent structures, obtained for freely decaying flow, are here assessed in the context of statistically steady flow, where the energy is supplied by either a spectrally localized forcing, or by moving top and bottom plates of the cylinder. It is shown that the observed structures are consistent with theoretical predictions.When an anisotropic forcing protocol is used, a bifurcation is observed from a non-swirling (two-dimensional two-component, 2D2C) flow to a swirling (two-dimensional three-component 2D3C) turbulent flow. This transition is modelled by a system of two ordinary differential equations (ODE), and it is shown that this model retains the essential physics of the transition.The transition of the axisymmetric flow to three-dimensional (3D3C) flow is then studied using non-integer dimension, by smoothly introducing azimuthal variation into the system. It is shown that the 2D2C limit is singular and that small azimuthal variation allows a redistribution of energy over the different energy components. The ODE model is adapted for this system and it is shown that for the considered flow the pressure-strain correlation is responsible for a swirl-level approximately proportional to the non-integer dimension. Large-Eddy Simulations are carried out to assess the robustness of the observations at higher Reynolds number.

No description available.

The goal of this work is to investigate the capabilities of two different commercial codes, OpenFOAM and ANSYS CFX, to predict forced convection heat transfer in low Prandtl number fluids and investigate the sensitivity of these predictions to the type of code and to several input parameters.The goal of the work is accomplished by predicting forced convection heat transfer in two different experimental setups with the codes OpenFOAM and ANSYS CFX using three different turbulence models and varying the input parameters in an extensive sensitivity analysis. The computational results are compared two the experimental data and analyzed for qualitative and quantitative parameters, such as shape of velocity and temperature profiles, thickness of the boundary layers and wall temperatures.The results show that predictions of the temperature and velocity field are generally sufficient to good, however, the sensitivity especially to the turbulent Prandtl number has to be taken into account when computing forced convection heat transfer in low Prandtl number fluids. The results also show that methods applied to OpenFOAM cannot directly be applied to ANSYS CFX. / <p>QC 20130531</p> / GENIUS

Development of a novel high-order flux correction method on strand grids is presented. The method uses a combination of flux correction in the unstructured plane and summation-by-parts operators in the strand direction to achieve high-fidelity solutions. Low-order truncation errors are cancelled with accurate flux and solution gradients in the flux correction method, thereby achieving a formal order of accuracy of 3, although higher orders are often obtained, especially for highly viscous flows. In this work, the scheme is extended to high-Reynolds number computations in both two and three dimensions. Turbulence closure is achieved with a robust version of the Spalart-Allmaras turbulence model that accommodates negative values of the turbulence working variable, and the Menter SST turbulence model, which blends the k-ε and k-ω turbulence models for better accuracy. A major advantage of this high-order formulation is the ability to implement traditional finite volume-like limiters to cleanly capture shocked and discontinuous flow. In this work, this approach is explored via a symmetric limited positive (SLIP) limiter. Extensive verification and validation is conducted in two and three dimensions to determine the accuracy and fidelity of the scheme for a number of different cases. Verification studies show that the scheme achieves better than third order accuracy for low and high-Reynolds number flow. Cost studies show that in three-dimensions, the third-order flux correction scheme requires only 30% more walltime than a traditional second-order scheme on strand grids to achieve the same level of convergence. In order to overcome meshing issues at sharp corners and other small-scale features, a unique approach to traditional geometry, coined "asymptotic geometry," is explored. Asymptotic geometry is achieved by filtering out small-scale features in a level set domain through min/max flow. This approach is combined with a curvature based strand shortening strategy in order to qualitatively improve strand grid mesh quality.