Directive-based programming of graphics processing units (GPUs) has recently appeared as a viable alternative to using specialized low-level languages such as CUDA C and OpenCL for general-purpose GPU programming. This technique, which uses directive or pragma statements to annotate source codes written in traditional high-level languages, is designed to permit a unified code base to serve multiple computational platforms and to simplify the transition of legacy codes to new architectures. This work analyzes the popular OpenACC programming standard, as implemented by the PGI compiler suite, in order to evaluate its utility and performance potential in computational fluid dynamics (CFD) applications. Of particular interest is the handling of stencil algorithms, which are an important component of finite-difference and finite-volume numerical methods. To this end, the process of applying the OpenACC Fortran API to a preexisting finite-difference CFD code is examined in detail, and all modifications that must be made to the original source in order to run efficiently on the GPU are noted. Optimization techniques for OpenACC are also explored, and it is demonstrated that tuning the code for a particular accelerator architecture can result in performance increases of over 30%. There are also some limitations and programming restrictions imposed by the API: it is observed that certain useful features of modern Fortran (2003/8) are effectively disabled within OpenACC regions. Finally, a combination of OpenACC and OpenMP directives is used to create a truly cross-platform Fortran code that can be compiled for either CPU or GPU hardware. The performance of the OpenACC code is measured on several contemporary NVIDIA GPU architectures, and a comparison is made between double and single precision arithmetic showing that if reduced precision can be tolerated, it can lead to significant speedups. To assess the performance gains relative to a typical CPU implementation, the execution time for a standard benchmark case (lid-driven cavity) is used as a reference. The OpenACC version is compared against the identical Fortran code recompiled to use OpenMP on multicore CPUs, as well as a highly-optimized C++ version of the code that utilizes hardware aware programming techniques to attain higher performance on the Intel Xeon platforms being tested. Low-level optimizations specific to these architectures are analyzed and it is observed that the stencil access pattern required by the structured-grid CFD code sometimes leads to performance degrading conflict misses in the hardware managed CPU caches. The GPU code, which primarily uses software managed caching, is found to be free from these issues. Overall, it is observed that the OpenACC GPU code compares favorably against even the best optimized CPU version: using a single NVIDIA K20x GPU, the Fortran+OpenACC code is seen to outperform the optimized C++ version by 20% and the Fortran+OpenMP version by more than 100% with both CPU codes running on a 16-core Xeon workstation. / Master of Science

In this Thesis we explore, both theoretically and numerically, splitting strategies for a hyperbolic system of one-dimensional (1D) blood flow equations with a passive scalar transport equation. Our analysis involves a two-step framework that includes splitting at the level of partial differential equations (PDEs) and numerical methods for discretizing the ensuing problems. This study is inspired by the original flux splitting approach of Toro and Vázquez-Cendón (2012) originally developed for the conservative Euler equations of compressible gas dynamics. In this approach the flux vector in the conservative case, and the system matrix in the non-conservative one, are split into advection and pressure terms: in this way, two systems of partial differential equations are obtained, the advection system and the pressure system. From the mathematical as well as numerical point of view, a basic problem to be solved is the special Cauchy problem called the Riemann problem. This latter provides an analytical solution to evaluate the performance of the numerical methods and, in our approach, it is of primary importance to build the presented numerical schemes. In the first part of the Thesis a detailed theoretical analysis is presented, involving the exact solution of the Riemann problem for the 1D blood flow equations, depicted for a general constant momentum correction coefficient and a tube law that allows to describe both arteries and veins with continuous or discontinuous mechanical and geometrical properties and an advection equation for a passive scalar transport. In literature, this topic has been already studied only for a momentum correction coefficient equal to one, that is related to the prescribed velocity profile and in this case corresponds to a flat one, i.e. an inviscid fluid. In the case of discontinuous properties, only the subsonic regime is considered. In addition we propose a procedure to compute the obtained exact solution and finally we validate it numerically, by comparing exact solutions to those obtained with well-known, numerical schemes on a carefully designed set of test problems. Furthermore, an analogous theoretical analysis and resolution algorithm are presented for the advection system and the pressure system arising from the splitting at the level of PDEs of the complete system of 1D blood flow equations. It is worth noting that the pressure system, in case of veins, presents a loss of genuine non-linearity resulting in the formation of rarefactions, shocks and compound waves, these latter being a composition of rarefactions and shocks. In the second part of the Thesis we present novel finite volume-type, flux splitting-based, numerical schemes for the conservative 1D blood flow equations and splitting-based numerical schemes for the non-conservative 1D blood flow equations that incorporate an advection equation for a passive scalar transport, considering tube laws that allow to model blood flow in arteries and veins and take into account a general constant momentum correction coefficient. A detailed efficiency analysis is performed in order to showcase the advantages of the proposed methodologies in comparison to standard approaches.

No / Owing to unpredictable bed topography conditions in natural shallow flows, various numerical methods have been developed to improve the treatment of source terms in the shallow water equations. The surface gradient method is an attractive approach as it includes a numerically simple approach to model flows over topographically-varied channels. To further improve the performance of this method, this study deals with the numerical improvement of the shallow-flow source terms. The so-called surface gradient upwind method (SGUM) integrates the source term treatment in the inviscid discretization scheme. A finite volume model (FVM) with the monotonic upwind scheme for conservative laws is used. The Harten–Lax–van Leer-contact approximate Riemann solver is used to reconstruct the Riemann problem in the FVM. The proposed method is validated against published analytical, numerical, and experimental data, indicating that the SGUM is robust and treats the source terms in different flow conditions well.

Yes / Based on the common approach, the adaptation length in sediment transport is normally estimated in the temporal independence. However, this approach might not be theoretically justified as the process of reaching of the sediment transport equilibrium stage is affected by the flow conditions in time, especially for those fast sediment moving flows, such as scour-hole developing flow. In this study, the 2D shallow water formulation together with a sediment continuity-concentration (SCC) model were applied to flow with mobile sediment boundary. A time-varying approach was proposed to determine the sediment transport adaptation length to treat the flow sediment erosion-deposition rate. The proposed computational model was based on the Finite Volume (FV) method. The Monotone Upwind Scheme of Conservative Laws (MUSCL)-Hancock scheme was used with the Harten Lax van Leer-contact (HLLC) approximate Riemann solver to discretize the FV model. In the flow applications of this paper, a highly discontinuous dam-break fast sediment transport flow was used to calibrate the proposed time-varying sediment adaptation length model. Then the calibrated model was further applied to two separate experimental sediment transport flow applications documented in literature, i.e. a highly concentrated sediment transport flow in a wide alluvial channel and a sediment aggradation flow. Good agreements with the experimental data were presented by the proposed model simulations. The tests prove that the proposed model, which was calibrated by the discontinuous dam-break bed scouring flow, also performed well to represent the rapid bed change and the steady sediment mobility conditions. / The National Natural Science Foundation of China NSFC (Grant Number 20101311246), Major State Basic Research Development Program (973 program) of China (Grant Number 2013CB036402) and Open Fund of the State Key Laboratory of Hydraulics and Mountain River Engineering, Sichuan University of China (Grant Number SKLH-OF-1103).

3D environment reconstruction has received great interest in recent years in areas such as city planning, virtual tourism and flood hazard warning. With the rapid development of computer technologies, it has become possible and necessary to develop new methodologies and techniques for real time simulation for virtual environments applications. This thesis proposes a novel dynamic simulation scheme for flood hazard warning. The work consists of three main parts: digital terrain modelling; 3D environmental reconstruction and system development; flood simulation models. The digital terrain model is constructed using real world measurement data of GIS, in terms of digital elevation data and satellite image data. An NTSP algorithm is proposed for very large data assessing, terrain modelling and visualisation. A pyramidal data arrangement structure is used for dealing with the requirements of terrain details with different resolutions. The 3D environmental reconstruction system is made up of environmental image segmentation for object identification, a new shape match method and an intelligent reconstruction system. The active contours-based multi-resolution vector-valued framework and the multi-seed region growing method are both used for extracting necessary objects from images. The shape match method is used with a template in the spatial domain for a 3D detailed small scale urban environment reconstruction. The intelligent reconstruction system is designed to recreate the whole model based on specific features of objects for large scale environment reconstruction. This study then proposes a new flood simulation scheme which is an important application of the 3D environmental reconstruction system. Two new flooding models have been developed. The first one is flood spreading model which is useful for large scale flood simulation. It consists of flooding image spatial segmentation, a water level calculation process, a standard gradient descent method for energy minimization, a flood region search and a merge process. The finite volume hydrodynamic model is built from shallow water equations which is useful for urban area flood simulation. The proposed 3D urban environment reconstruction system was tested on our simulation platform. The experiment results indicate that this method is capable of dealing with complicated and high resolution region reconstruction which is useful for many applications. When testing the 3D flood simulation system, the simulation results are very close to the real flood situation, and this method has faster speed and greater accuracy of simulating the inundation area in comparison to the conventional flood simulation models

<p align="justify">This work deals with the development of numerical methods for compressible flow simulation with application to the interaction of fluid flows and structural bodies.</p> <p align="justify">First, we develop numerical methods based on multidimensional upwind residual distribution (RD) schemes. Theoretical results for the stability and accuracy of the methods are given. Then, the RD schemes for unsteady problems are extended for computations on moving meshes. As a second approach, cell centered and vertex centered finite volume (FV) schemes are considered. The RD schemes are compared to FV schemes by means of the 1D modified equation and by the comparison of the numerical results for scalar problems and system of Euler equations. We present a number of two and three dimensional steady and unsteady test cases, illustrating properties of the numerical methods. The results are compared with the theoretical solution and experimental data.</p> <p align="justify">In the second part, a numerical method for fluid-structure interaction problems is developed. The problem is divided into three distinct sub-problems: Computational Fluid Dynamics, Computational Solid Mechanics and the problem of fluid mesh movement. The problem of Computational Solid Mechanics is formulated as a system of partial differential equations for an anisotropic elastic continuum and solved by the finite element method. The mesh movement is determined using the pseudo-elastic continuum approach and solved again by the finite element method. The coupling of the problems is achieved by a simple sub-iterative approach. Capabilities of the methods are demonstrated on computations of 2D supersonic panel flutter and 3D transonic flutter of the AGARD 445.6 wing. In the first case, the results are compared with the theoretical solution and the numerical computations given in the references. In the second case the comparison with experimental data is presented.</p>

Ce travail de thèse porte sur l'extension et l'analyse d'un solveur volumes finis eulérien, co-localisé avec capture d'interfaces pour la simulation des écoulements multi-matériaux non miscibles. Les extensions proposées s'inscrivent dans la volonté d'élaborer un outil de simulation multi-physiques. Dans le cadre de ce mémoire, le caractère multi-physiques recouvre les champs que nous allons détailler. Nous traitons le cas des écoulements radiatifs modélisés par un système à deux températures qui couple les phénomènes purement hydrodynamiques aux phénomènes radiatifs. Nous proposons un solveur permettant la prise en compte des effets de tension superficielle à l'interface entre deux fluides. Nous développons un solveur implicite permettant la simulation précise d'écoulements faisant intervenir de faibles nombres de Mach par le biais d'une méthode de renormalisation de la diffusion numérique. Enfin, les effets tri-dimensionnels sont considérés ainsi que la possibilité d'étendre le schéma de base aux écoulements à un nombre quelconque de matériaux. A chaque étape, les solveurs développés sont validés sur des cas-tests. / This work is devoted to the extension of a eulerian cell-centered finite volume scheme with interfaces capturing for the simulation of multimaterial fluid flows. Our purpose is to develop a simulation tool which could be able to handle multi-physics problems in the following sense. We address the case of radiating flows, modeled by a two temperature system of equations where the hydrodynamics are coupled to radiation transport. We address a numerical scheme for taking surface tension forces into account. An implicit scheme is proposed to handle low Mach number fluid flows by means of a renormalization of the numerical diffusion. Eventually, the scheme is extended to three-dimensional flows and to multimaterial flows, that is with an arbitrary number of materials. At each step, numerical simulations validate our schemes.

Le but de ce mémoire de recherche est d'étudier les convections naturelle et mixte d'écoulements fluides, et de développer et valider des méthodes numériques pour le suivi d'interfaces afin de traiter plus tard des écoulements incompressibles de fluides immiscibles. Dans une première étape, une méthode numérique originale, basée sur des discrétisations Volumes Finis, est développée pour modéliser les écoulements à faible nombre de Mach et grands écarts de température. Trois applications physiques, portant sur l'écoulement d'air à travers des plaques verticales parallèles chauffées, sont étudiées. Nous avons montré que l'espacement optimal, correspondant au pic de flux de chaleur transféré d'un réseau de plaques parallèles isothermes refroidies par convection mixte, est plus faible que ceux obtenus en convections naturelle ou forcée lorsque la chute de pression à la sortie est constante. Nous avons également prouvé que les écoulements de convection mixte à débit imposé peuvent présenter des solutions physiques inattendues ; un modèle alternatif basé sur une pression totale imposée à l'entrée et une pression fixée à la sortie donne de meilleurs résultats. Pour des canaux soumis un flux de chaleur sur une paroi seule, le rayonnement de surface tend à supprimer l'apparition des recirculations à la sortie et à uniformiser les températures des parois. Dans une seconde étape, le modèle mathématique couplant les équations de Navier-Stokes incompressibles et la méthode Level-Set pour le suivi d'interfaces est développé. Des améliorations de la conservation du volume fluide par l'utilisation de schémas de discrétisation d'ordres élevés (ENO-WENO) pour l'équation de transport et des variantes de l'équation de la distance signée sont discutées / The aim of this research dissertation is at studying natural and mixed convections of fluid flows, and to develop and validate numerical schemes for interface tracking in order to treat incompressible and immiscible fluid flows, later. In a first step, an original numerical method, based on Finite Volume discretizations, is developed for modeling low Mach number flows with large temperature gaps. Three physical applications on air flowing through vertical heated parallel plates were investigated. We showed that the optimum spacing corresponding to the peak heat flux transferred from an array of isothermal parallel plates cooled by mixed convection is smaller than those for natural or forced convections when the pressure drop at the outlet keeps constant. We also proved that mixed convection flows resulting from an imposed flow rate may exhibit unexpected physical solutions; alternative model based on prescribed total pressure at inlet and fixed pressure at outlet sections gives more realistic results. For channels heated by heat flux on one wall only, surface radiation tends to suppress the onset of recirculations at the outlet and to unify the walls temperature. In a second step, the mathematical model coupling the incompressible Navier-Stokes equations and the Level-Set method for interface tracking is derived. Improvements in fluid volume conservation by using high order discretization (ENO-WENO) schemes for the transport equation and variants of the signed distance equation are discussed

Les méthodes hybrides Moments/PDF sont bien adaptées pour la description des écoulements diphasiques turbulents, polydispersés, hors équilibre thermodynamique. Ces méthodes permettent d'avoir une description assez fine de la polydispersion, de la convection et des termes sources non-linéaires. Cependant, les approximations issues de telles simulations sont bruitées ce qui, dans certaines situations, occasionne un biais. L'approche alternative étudiée dans ce travail consiste à coupler une description Eulerienne des moments avec une description stochastique Lagrangienne à l'intérieur de la phase dispersée, permettant de réduire l'erreur statistique et d'éliminer le biais. La mise en oeuvre de cette méthode nécessite le développement de schémas numériques robustes. Les approches proposées sont basées sur l'utilisation simultanée des techniques de relaxation et de décentrement, et permettent d'obtenir des approximations stables des solutions instationnaires du système d'équations aux dérivées partielles, avec des données peu régulières provenant du couplage avec le modèle stochastique. Une comparaison des résultats de la méthode hybride Moments-Moments/PDF avec ceux issus de la méthode hybride "classique'' est présentée en termes d'analyse des erreurs numériques sur un cas de jet co-courant gaz-particules. / Hybrid Moments/PDF methods have shown to be well suitable for the description of polydispersed turbulent two-phase flows in non-equilibrium which are encountered in some industrial situations involving chemical reactions, combustion or sprays. hey allow to obtain a fine enough physical description of the polydispersity, non-linear source terms and convection phenomena. However, their approximations are noised with the statistical error, which in several situations may be a source of a bias. An alternative hybrid Moments-Moments/PDF approach examined in this work consists in coupling the Moments and the PDF descriptions, within the description of the dispersed phase itself. This hybrid method could reduce the statistical error and remove the bias. However, such a coupling is not straightforward in practice and requires the development of accurate and stable numerical schemes. The approaches introduced in this work rely on the combined use of the upwinding and relaxation-type techniques. They allow to obtain stable unsteady approximations for a system of partial differential equations containing non-smooth external data which are provided by the PDF part of the model. A comparison of the results obtained using the present method with those of the ``classical'' hybrid approach is presented in terms of the numerical errors for a case of a co-current gas-particle wall jet.

On étudie des phénomènes de séparation d'écoulement avec deux approches différentes. Dans la première partie, on considère des écoulements 2D, instationnaires, incompressibles et non visqueux. Un modèle analytique-numérique, basé sur la jonction d'une transformation conforme et d'une méthode aux tourbillons ponctuels, est construit pour définir l'écoulement potentiel dans un domaine doublement connecté où les corps sont caractérisés par une variation temporelle de leur circulation. En particulier, on s'intéresse à l'étude de l'écoulement autour d'un VAWT avec deux pales. Dans la seconde partie on considère des écoulements visqueux et compressibles. On construit un solveur qui résoud les équations de Navier-Stokes en y introduisant une technique de pénalisation: les corps sont modélisés comme des milieux poreux ayant une porosité très petite par rapport à la porosité du fluide extérieur. Cette technique permet d'utiliser des maillages cartésiens pour des géométries très complexes. / In the present work flow separation phenomena are investigated by means of two different approaches. In the first part, 2D unsteady incompressible inviscid flows are studied. An analytical-numerical model, based on the conjunction of a conformal mapping and a point vortex method, is built to define the potential flow field in a doubly connected domain where bodies are characterized by a variation in time of their circulation. In particular, the study of the unsteady flow past a 2-blade Darrieus VAWT is addressed. Until now the study of vortex motions has only been described in doubly-connected flow fields where the circulations have a constant null value. The flow field here analysed has a deep unsteadiness, which determines the circulations varying in time: so a technique is developed to uniquely define the circulations around the bodies. Three conditions result necessary to be imposed: in addition to the two Kutta conditions at the trailing edges, another one has to be imposed in order to respect the Kelvin theorem. With a classical configuration, this machine, experiencing angles of attack of opposite values, gives rise to complex vortex shedding phenomena that reduce its performances and stress its structure. In order to control the flow separation from the blades, an innovative solution is qualitatively investigated which consists of taking blade profiles provided with vortex trapping cavities. Interesting results are obtained, even if in the limit of inviscid flow. In the second part compressible viscous flows are taken into account. A fully Navier-Stokes equations solver is implemented introducing the penalization technique. The idea is to replace the bodies by the fluid, in a way that also into the bodies the penalized Navier-Stokes equations remain valid, respecting the boundary conditions on their contours. Starting from this purpose, the bodies are considered as porous media with a little porosity with respect to that of the external flow, which tends to infinity. This technique allows simple Cartesian meshes to be used, also for very complex geometries like those of industrial interest. The resulting code is tested on different flow fields, both steady and unsteady, both subsonic and supersonic, obtaining always a good agreement with other theoretical and numerical results described in literature.