In this thesis, large-eddy simulation (LES) is used to simulate both Newtonian and non-Newtonian physiological pulsatile flows in constricted channels to gain insights into the physical phenomenon of laminar-turbulent flow transition due to the presence of an artificial arterial stenosis. The advanced dynamic nonlinear subgrid-scale stress (SGS) model of Wang and Bergstrom (DNM) was utilized to conduct numerical simulations and its predictive performance was examined in comparison with that of the conventional dynamic model (DM) of Lilly. An in-house LES code has been modified to conduct the unsteady numerical simulations, and the results obtained have been validated against available experimental and direct numerical simulation (DNS) results. The physical characteristics of the flow field have been thoroughly studied in terms of the resolved mean velocity, turbulence kinetic energy, viscous wall shear stress, and turbulence energy spectra along the central streamline of the domain.

Large Eddy Simulation is a method of obtaining high accuracy computational results for modelling fluid flow. Unfortunately it is computationally expensive limiting it to users of large parallel machines. However, it may be that the use of LES leads to an over-resolution of the problem because the bulk of the computational domain could be adequately modelled using the Reynolds averaged approach. A study has been undertaken to assess the feasibility, both in accuracy and computational efficiency of using a parallel computer to solve both LES and RANS type turbulence models on the same domain for the problem flow over a circular cylinder at Reynolds number 3 900 To do this the domain has been created and then divided into two sub-domains, one for the LES model and one for the kappa-epsilon turbulence model. The hybrid model has been developed specifically for a parallel computing environment and the user is able to allocate modelling techniques to processors in a way which enables expansion of the model to any number of processors. Computational experimentation has shown that the combination of the Smagorinsky model can be used to capture the vortex shedding from the cylinder and the information successfully passed to the kappa - epsilon model for the dissipation of the vortices further downstream. The results have been compared to high accuracy LES results and with both kappa - epsilon and Smagorinsky LES computations on the same domain. The hybrid models developed compare well with the Smagorinsky model capturing the vortex shedding with the correct periodicity. Suggestions for future work have been made to develop this idea further, and to investigate the possibility of using the technology for the modelling of mixing and fast chemical reactions based on the more accurate prediction of the turbulence levels in the LES sub-domain.

Cette thèse est dédiée à l'amélioration des modèles d'atomisation pour les injecteurs automobiles. Le but est de développer et d'évaluer des modèles numériques permettant de capturer le passage de structures liquides en cours d'atomisation depuis les grandes échelles vers les petites échelles de sous-maille dans des configurations complexes. Dans un premier temps, nous mettons en place une procédure de calcul permettant le passage d'une description Eulérienne d'un spray à une procédure Lagrangienne. Afin de ne pas perdre les plus petites structures liquides, celles-ci seront transformées en particules Lagrangienne. Une analyse sur différentes grandeurs physiques, telles que la masse, la quantité de mouvement ou l'énergie cinétique turbulente, lors de cette transformation a été réalisée. L'autre partie de ce travail est consacrée au développement d'un modèle de simulation aux grandes échelles des écoulements diphasiques. La simulation de l'atomisation requiert un traitement spécifique de l'interface. Deux cas limites sont traités dans la littérature : • L'interface peut bien être capturée par le maillage. A ces endroits, une méthode classique de type DNS (Direct Numerical Simulation), comme les méthodes VOF (Volume of Fluid), doit être utilisée. • Lors de la création de plissements inférieurs à la taille de la maille, le maillage ne permet plus de suivre fidèlement l'interface. Il faut alors que le calcul reproduise des résultats d'une méthode LES (Large Eddy Simulation) considérant des structures et des gouttes inférieures à la taille de la maille. Ainsi, la problématique principale consiste à déterminer la configuration dans laquelle se trouve l'interface. La mise en œuvre de ce modèle a permis d'obtenir des résultats dans une configuration proche de l'injection Diesel, qui sont alors comparés à une DNS de référence. / This thesis is dedicated to improve atomization models for automobile injectors. The aim is to develop and evaluate numerical models to capture the liquid structure while they are being atomized from large scales to small sub grid scales in complex configurations. Initially, a calculation procedure is introduced for the transition to an Eulerian description of a spray into a Lagrangian description. In order not to lose the smallest fluid structures, they will be transformed into Lagrangian particles. During this process, an analysis is been performed with various physical parameters such as mass, momentum, or turbulent kinetic energy. The other part of this work is dedicated to the development of a LES (Large Eddy Simulation) for multiphase flow. The simulation of the spray requires a specific treatment of the interface. Two limiting cases are treated in the literature: • The interface may be captured by the mesh. At these locations, a conventional method of DNS (Direct Numerical Simulation) should be used, like the VOF method (Volume of Fluid). • When creating pleating smaller than the size of the mesh, the mesh can no longer match the interface. Then, the calculation must reproduce results from a LES method that take into account structures and drops smaller than the mesh size. Thus, the main problem is to define the configuration of the interface. The development of this model allows to obtain results in a configuration close to the Diesel injection's, which are then compared to a reference DNS.

No description available.

No description available.

Unsteady large-scale vortices, formed by the roll-up of free shear layers separating along sharp edges, are the dominant flow characteristics of the turbulent flow over buildings. These vortical structures interact with each other and with the building surface resulting in secondary separation and severe pressure fluctuations. Moreover, the interaction of the large-scale vortices with the multiplicity of turbulence scales in the incoming wind exacerbates their unsteady motion and hence significantly affects the pressure fluctuations spectra experienced by the building. Large-eddy simulations are conducted to study the interaction of homogeneous turbulence in the incident flow with a surface-mounted prism. A compact fifth-order upwind difference scheme is used to effectively and accurately perform the simulations. Three cases of incident flow are considered. In one case, the prism is placed in a smooth uniform flow. In the second case, homogeneous isotropic turbulence with von Karman spectrum is superimposed on the uniform flow at the inflow boundary. The integral length scale is one-half the prism height. In the third case, the integral length scale is equal to the prism height. The numerical results are compared with experimental measurements reported by Tieleman et al. (2002). The results show that the highest negative mean value of the pressure coefficient on the roof and the sides is about 30% larger in case two of turbulent inflow and takes place closer to the windward edge of the prism. Moreover, the pressure coefficients on the roof and sides of the prism in the case of turbulent inflow show a higher level of variations in comparison with the case of smooth inflow conditions. The predicted mean characteristics of the pressure coefficients in the turbulent case match the experimental values in terms of both magnitude and location on the roof of the prism reported in Tieleman et al. (1998) and Tieleman et al. (2002). As for the peak value, the peak value of -2 obtained in the turbulent inflow case two is about 20% smaller than the values measured experimentally by Tieleman et al. (2002). On the other hand, it is stressed that the peak value in the simulations would increase as the duration of the simulation is increased to match that of the experimental measurement. The results also show that the turbulent case yields a non-exceedence probability for the peak pressure coefficient that is closer to the one obtained from the measured data than the smooth case data. Also, spectral and cross-spectral analysis are carried out using complex Morlet wavelet transform to investigate pressure-velocity relation. The study shows that the nonlinearity in the relationship of velocity-pressure is detected using wavelet bicoherence. / Ph. D.

This work presents direct and large eddy simulations of a wide range of heat augmentation surfaces roughened by modified dimples/protrusions. The dissertation is composed of two main parts: Part I (Chapters 2-4) for compact heat exchangers and Part II (Chapter 5) for internal cooling of rotating turbine blades. Part I consists of three phases: Phase I (Chapter 2) investigates flow structure and heat transfer distribution in a channel with dimples/protrusions; Phase II (Chapter 3) studies the application of dimples as surface roughness on plain fins; and Phase III (Chapter 4) considers a new fin shape, the split-dimple fin, that is based on modifying the conventional dimple shape. Chapter 2 presents direct and large eddy simulations conducted of a fin bank over a wide range of Reynolds numbers, ReH=200-15,000, covering the laminar to fully turbulent flow regimes and using two channel height geometries. While the smaller fin pitch channel has better performance in the low to medium Reynolds number range, both channel heights show similar trends in the fully turbulent regime. Moreover, analysis of the results shows that vortices generated in the dimple cavity and at the dimple rim contribute substantially to heat transfer from the dimpled surface, whereas flow impingement and acceleration between protrusions contribute substantially on the protrusion side. Chapter 3 considers applying dimples as surface roughness on plain fin surfaces to further enhance heat transfer from the fin. Three fin geometries that consider dimple imprint diameter effect and perforation effect are considered. The dimple imprint diameter has a minimal effect on the flow and heat transfer of the fin. However, the introduction of perforation in the dimple significantly changes the flow structure and heat transfer on the dimple side of the fin by eliminating recirculation regions in the dimple and generating higher intensity vortical structures. Chapter 4 presents a novel fin shape, the split-dimple fin, which consists of half a dimple and half a protrusion with an opening between them. The split dimple provides an additional mechanism for augmenting heat transfer by perturbing continuous boundary layer formation on the fin surface and generating energetic shear layers. While the protruding geometry of the split dimple augments heat transfer profoundly, it also increase pressure drop. The split dimple fin results in heat conductance that is 60–175% higher than a plain fin, but at a cost of 4–8 times the frictional losses. Chapter 5 studies the employment of dimples/protrusions on opposite sides for internal cooling of rotating turbine blades. Two geometries with two dimple/protrusion depths are investigated over a wide range of rotation numbers, Rob=-0.77 to 1.10. Results show that the dimple side is more sensitive to the destabilizing forces on the trailing surface, while both react similarly to the stabilizing effect on the leading side. It is concluded that placing the protrusion on the trailing side for low rotation number, |Rob|<0.2, provides better performance, while it is more beneficial to place the dimple side on the trailing side for higher rotation numbers. / Ph. D.

The research was motivated by the desire to use Large Eddy Simulations (LES) to calculate liner heat transfer in industrial scale gas turbine combustors, which operate at high Reynolds numbers and high Swirl numbers. LES has several challenges which need to be surmounted for general application to complex high Reynolds number turbulent flows. The primary challenge in wall bounded flows is the need for very fine grids in the vicinity of walls, which makes LES impractical at high Reynolds numbers. An additional challenge is the accurate representation of inlet turbulent conditions for developing flows such that the computational domain size is limited to the immediate region of interest. The generalization of solutions to surmount these issues in complex geometries and grids is yet another challenge. To meet these challenges, a novel formulation, implementation, and validation of a two layer velocity and temperature zonal wall model along with the implementation of the synthetic eddy method in a generalized coordinate system LES framework is presented in this thesis. The wall model greatly alleviates the grid requirements, whereas the synthetic eddy method provides accurate turbulent inlet boundary conditions. The methods are validated in turbulent channel flow up to a Reynolds number of 2x106, a backward facing step at Re=40,000, before application to a model swirl combustor at Re=20,000 with a Swirl number of 0.43 and flow and heat transfer in an industrial scale can combustor at Re=80,000 and Swirl number of 0.7. The integrated zonal near wall approach for velocity and temperature is then successfully used to investigate flow and heat transfer in a statistically three-dimensional flow of a ribbed duct passage used for the internal cooling of turbine blades. The zonal wall model is further modified to take in to account the effects of surface roughness and successfully used to investigate flow in a rod roughened channel at high Reynolds numbers up to 60,000. In all cases it is shown that the zonal wall model used with the synthetic eddy method for inlet turbulence generation can result in large savings in computational cost without any significant loss in accuracy when compared to wall resolved LES and experiments. In a turbulent channel flow at Re=45,000, computational complexity was reduced by a factor of 285 using wall modeled LES, whereas in a statistically three-dimensional flow and heat transfer in a ribbed duct, at Re=20,000, the computational complexity was reduced by a factor between 60 and 140. In a swirl dominated can combustor at Re=20,000, the reduction was more modest at a factor of 9. / Ph. D.

Although spectral methods have been in use for decades, there is still room for innovation, refinement and improvement of the methods in terms of efficiency and accuracy, for generalized homogeneous turbulent flows, and especially for specialized applications like the computation of atmospheric flows and numerical weather prediction. In this thesis, two such innovations are presented. First, inspired by the adaptive mesh refinement (AMR) technique, which was developed for the computation of fluid flows in physical space, an algorithm is presented for accelerating direct numerical simulation (DNS) of isotropic homogeneous turbulence in spectral space. In the adaptive spectral resolution (ASR) technique developed here the spectral resolution in spectral space is dynamically refined based on refinement criteria suited to the special features of isotropic homogeneous turbulence in two, and three dimensions. Applying ASR to computations of two- and three-dimensional turbulence allows significant savings in the computational time with little to no compromise in the accuracy of the solutions. In the second part of this thesis the effect of explicit filtering on large eddy simulation (LES) of atmospheric flows in spectral space is studied. Apply an explicit filter in addition to the implicit filter due to the computational grid and discretization schemes in LES of turbulent flows allows for better control of the numerical error and improvement in the accuracy of the results. Explicit filtering has been extensively applied in LES of turbulent flows in physical space while few studies have been done on explicitly filtered LES of turbulent flows in spectral space because of perceived limitations of the approach, which are shown here to be incorrect. Here, explicit filtering in LES of the turbulent barotropic vorticity equation (BVE) as a first model of the Earth's atmosphere in spectral space is studied. It is shown that explicit filtering increases the accuracy of the results over implicit filtering, particularly where the location of coherent structures is concerned. / Ph. D.

Large Eddy Simulation (LES) is a relatively more accurate and reliable alternative to solution of Reynolds Averaged Navier Stokes (RANS) equations in simulating complex turbulent flows at a lesser computational cost than a direct numerical simulation (DNS). However, LES of wall-bounded flows still requires a very high grid resolution in the inner wall layer making its widespread use difficult. Different attempts have been made in the past time to overcome this problem by modeling the near wall turbulence instead of resolving it. One such approach is a two-layer wall model that solves for a reduced one-dimensional equation in the inner wall layer, while solving for the filtered Navier-Stokes equations in the outer layer. The use of such a model allows for a coarser grid resolution than a wall resolved LES. This work validates the performance of a two-layer wall model developed for an arbitrary body fitted non-orthogonal grid in the flow over a wall mounted hump at Reynolds number 9.36x105. The wall modeled large eddy simulation (WMLES) relaxes the grid requirement compared to a wall resolved LES (WRLES) by allowing the first off-wall grid point to be placed at a y+ of approximately 20-40. It is found that the WMLES results are general good agreement with WRLES and experiments. Surface pressure coefficient, skin friction, mean velocity profiles, and the reattachment location compare very well with experiment. The WMLES and WRLES exhibit some under prediction of the peak values in the turbulent quantities close to the reattachment location, with better agreement with the experiment in the separated region. In contrast, a simulation that did not employ the wall model on the grid used for WMLES failed to predict flow separation and showed large discrepancies with the experimental data. In addition to the relaxation of the grid requirement in the wall normal direction, it was also observed that the wall model allowed a reduction in the number of computational cells in the span-wise direction by half. However an LES calculation on a grid with reduced number of cells in span-wise direction turned unstable almost immediately, thereby highlighting the effectiveness of the wall model. Besides reducing the number of grid points in the spatial domain, the relaxed grid resolution for the WMLES also permitted the use of a larger time step. This resulted in an order of magnitude reduction in the total CPU time relative to WRLES. / Master of Science