Most fluid flows of practical interest involve and are affected by turbulence. One of the most promising computational methods for the prediction of turbulent flows is the so-called large eddy simulation (LES) methodology. Experience over the past decades have shown the capability of LES to provide accurate predictions for several types of flow at a reasonable computational cost. It has also become clear, however, that the LES methodology fails when applied to boundary layer flows at high Reynolds numbers. Since many engineering applications fall in exactly that category, this failure is often considered the most severe bottleneck of LES. <br /><br /> The present thesis is an attempt to move towards a solution of this problem. Inspired by the idealized picture of a turbulent boundary layer, a statistical model is used for the approximately universal turbulence in the inner boundary layer, whereas the more flow dependent outer boundary layer is solved by LES. Ideally, this results in a computational method that provides accurate predictions of rather general turbulent flows, while maintaining a tractable computational cost. In practice, the results are a vast improvement compared to LES without any inner layer modeling, but a transition layer appears where the state of the turbulence changes from being modeled statistically to resolved by LES. This so-called 'artificial buffer layer' results in the skin friction being consistently underpredicted by 10-15%. <br /><br /> The physics and dynamics of this artificial buffer layer are investigated and characterized, and it is argued that there exist several similarities with true buffer layer turbulence. Additional forcing of the momentum equations is used as a means to trigger resolved turbulence motions more quickly, and it is demonstrated that the results are better: the artificial buffer layer is smaller, the skin friction is accurately predicted, and the dynamics in the inner layer have more correct length scales. The results with the additional forcing are very sensitive to the forcing amplitude, and a simple control algorithm for this parameter is proposed and tested with favourable results.

This dissertation consists of two parts, i.e. dynamic approaches for subgrid-scale (SGS) stress modelling for large eddy simulation and advanced assessment of the resolved scale motions related to turbulence geometrical statistics and topologies. The numerical simulations are based on turbulent Couette flow. The first part of the dissertation presents four contributions to the development of dynamic SGS models. The conventional integral type dynamic localization SGS model is in the form of a Fredholm integral equation of the second kind. This model is mathematically consistent, but demanding in computational cost. An efficient solution scheme has been developed to solve the integral system for turbulence with homogeneous dimensions. Current approaches to the dynamic two-parameter mixed model (DMM2) are mathematically inconsistent. As a second contribution, the DMM2 has been optimized and a modelling system of two integral equations has been rigorously obtained. The third contribution relates to the development of a novel dynamic localization procedure for the Smagorinsky model using the functional variational method. A sufficient and necessary condition for localization is obtained and a Picard's integral equation for the model coefficient is deduced. Finally, a new dynamic nonlinear SGS stress model (DNM) based on Speziale's quadratic constitutive relation [J. Fluid Mech., 178, p.459, 1987] is proposed. The DNM allows for a nonlinear anisotropic representation of the SGS stress, and exhibits a significant local stability and flexibility in self-calibration. In the second part, the invariant properties of the resolved velocity gradient tensor are studied using recently developed methodologies, i.e. turbulence geometrical statistics and topology. The study is a posteriori based on the proposed DNM, which is different than most of the current a priori approaches based on experimental or DNS databases. The performance of the DNM is further validated in terms of its capability of simulating advanced geometrical and topological features of resolved scale motions. Phenomenological results include, e.g. the positively skewed resolved enstrophy generation, the alignment between the vorticity and vortex stretching vectors, and the pear-shape joint probability function contour in the tensorial invariant phase plane. The wall anisotropic effect on these results is also examined.

Buoyant flows are characterized with unsteady large-scale structures and thus time-dependent large eddy simulation (LES) is generally favored. In this dissertation, to further explore LES for buoyant flow, an LES code based on a collocated grid system is first developed. A multigrid solver using a control strategy is developed for the pressure Poisson equations. The control strategy significantly accelerated the convergence rate. A temperature solver using a fourth-order Runge-Kutta approach is also developed. The LES code is extensively tested before it is applied. Although the collocated grid system will introduce conservation errors, in tests of a steady lid-driven cavity flow and transient start-up flow, the effect of the non-conservation of the collocated grid system was not significant. <p>In LES, the effect of SGS scales is represented by SGS models. A novel dynamic nonlinear model (DNM) for SGS stress is tested using isothermal channel flow at Reynolds number 395. The kinetic energy dissipation and geometrical characteristics of the resolved scale and SGS scale with respect to the DNM are investigated. In general, the DNM is reliable and has relatively realistic geometrical properties in comparison with the conventional dynamic model in the present study. In contrast to a pure advecting velocity field, a scalar (temperature) field displays very different characteristics. The modelling of SGS heat flux has not been as extensively studied as that of SGS stress partly due to the complexity of the scalar transport. In this dissertation, LES for a turbulent combined forced and natural convection is studied. The DNM model and a nonlinear dynamic tensor diffusivity model (DTDM-HF) are applied for the SGS stress and heat flux, respectively. The combined effect of the nonlinear models is compared to that of linear models. Notable differences between the nonlinear and linear SGS models are observed at the subgrid-scale level. At the resolved scale, the difference is smaller but relatively more distinguishable in terms of quantities related to the temperature field. <p>Finally, the geometrical properties of the resolved velocity and temperature fields of the thermal flow are investigated based on the LES prediction. Some universal geometrical patterns have been reproduced, e.g. the positively skewed resolved enstrophy generation and the alignment between the vorticity and vortex stretching vectors. The present research demonstrates that LES is an effective tool for the study of the geometrical properties of a turbulent flow at the resolved-scales. The wall imposed anisotropy on the flow structures and orientation of the SGS heat flux vector are also specifically examined. In contrast to the dynamic eddy diffusivity model, the DTDM-HF successfully predicts the near-wall physics and demonstrates a non-alignment pattern between the SGS heat flux and temperature gradient vector.

Study of turbulent flow over a rough bed is highly important due to its numerous applications in the areas of sediment transport and pollutant discharge in streams, rivers and channels. Over the past few decades, many experimental studies have been conducted in this respect to understand the underlying phenomenon. However, there is a scarcity in the number of computational studies conducted on this topic. Therefore, a Large Eddy Simulation (LES) of turbulent flow over a rough channel bed was conducted to contribute further understanding of the influence of bed roughness on turbulent flow properties. For this purpose, an efficient, second order accurate 'immersed boundary method' was implemented into the LES code Hydro3d-GT, and validated for flow past bluff bodies. LES results from the present study showed excellent agreement with previous experimental studies on flow over rough beds. An in-depth analysis of time varying turbulent quantities (like the velocity fluctuations) revealed the presence of coherent structures in the flow. Also, a three dimensional visualization of the turbulent structures provided a good picture of the flow, especially in the near bed region, which is quite difficult to accomplish using experimental studies.

Flame based synthesis is a major manufacturing process of commercially valuable nanoparticles for large-scale production. However, this important industrial process has been advanced mostly by trial-and-error based evolutionary studies owing to the fact that it involves tightly coupled multiphysics flow phenomena. For large scale synthesis of nanoparticles, different physical and chemical processes exist, including turbulence, fuel combustion, precursor oxidation, and nanoparticle dynamics exist. A reliable and predictive computational model based on fundamental physics and chemistry can provide tremendous insight. Development of such comprehensive computational models faces challenges as they must provide accurate descriptions not only of the individual physical processes but also of the strongly coupled, nonlinear interactions among them. In this work, a multiscale computational model for flame synthesis of TiO2 nanoparticles in a turbulent flame reactor is presented. The model is based on the large-eddy simulation (LES) methodology and incorporates detailed gas phase combustion and precursor oxidation chemistry as well as a comprehensive nanoparticle evolution model. A flamelet-based model is used to model turbulence-chemistry interactions. In particular, the transformation of TiCl4 to the solid primary nucleating TiO2 nanoparticles is represented us- ing an unsteady kinetic model considering 30 species and 70 reactions in order to accurately describe the critical nanoparticle nucleation process. The evolution of the TiO2 number density function is tracked using the quadrature method of moments (QMOM) for univariate particle number density function and conditional quadrature method of moments (CQMOM) for bivariate density distribution function. For validation purposes, the detailed computational model is compared against experimental data obtained from a canonical flame- based titania synthesis configuration, and reasonable agreement is obtained. / text

In the recent past, LES methodology has emerged as a viable tool for modeling turbulent combustion. LES computes the large scale mixing process accurately, thereby providing a better starting point for small-scale models that describe the combustion process. Significant effort has been made over past decades to improve accuracy and applicability of the LES approach to a wide range of flows, though the current conventions often lack consistency to the problems at hand. To this end, the two main objectives of this dissertation are to develop a dynamic transport equation-based combustion model for large- eddy simulation (LES) of turbulent spray combustion and to investigate grid- independent LES modeling for scalar mixing. Long-standing combustion modeling approaches have shown to be suc- cessful for a wide range of gas-phase flames, however, the assumptions required to derive these formulations are invalidated in the presence of liquid fuels and non-negligible evaporation rates. In the first part of this work, a novel ap- proach is developed to account for these evaporation effects and the resulting multi-regime combustion process. First, the mathematical formulation is de- rived and the numerical implementation in a low-Mach number computational solver is verified against one-dimensional and lab scale, both non-reacting and reacting spray-laden flows. In order to clarify the modeling requirements in LES for spray combustion applications, results from a suite of fully-resolved direct numerical simulations (DNS) of a spray laden planar jet flame are fil- tered at a range of length scales. LES results are then validated against two sets of experimental jet flames, one having a pilot and allowing for reduced chemistry modeling and the second requiring the use of detail chemistry with in situ tabulation to reduce the computational cost of the direct integration of a chemical mechanism. The conventional LES governing equations are derived from a low-pass filtering of the Navier-Stokes equations. In practice, the filter used to derive the LES governing equations is not formally defined and instead, it is assumed that the discretization of LES equations will implicitly act as a low-pass filter. The second part of this study investigates an alternative derivation of the LES governing equations that requires the formal definition of the filtering operator, known as explicitly filtered LES. It has been shown that decoupling the filter- ing operation from the underlying grid allows for the isolation of subfilter-scale modeling errors from numerical discretization errors. Specific to combustion modeling are the aggregate errors associated with modeling sub-filter distribu- tions of scalars that are transported by numerical impacted turbulent fields. Quantities of interest to commonly-used combustion models, including sub- filter scalar variance and filtered scalar dissipation rate, are investigated for both homogeneous and shear-driven turbulent mixing. / text

Extreme hydrological events associated with global warming are likely to produce an increasing number of flooding scenarios resulting in significant bridge inundation and associated damages. During large floods, the presence of a bridge in an open channel triggers a highly turbulent flow field including 3D complex coherent structures around bridge structures. These turbulence structures are highly energetic and possess high sediment entrainment capacity which increases scouring around the bridge foundation and consequently lead to structural stability problems or even failure of the structure. Hence, understanding the complex turbulent flow field for these extreme flow conditions is crucial to estimate the failure risks for existing bridges and better design of future bridges. This research employs the method Large Eddy Simulation (LES) to predict accurately the 3D turbulent flow around bridge structures. The LES code is refined with a novel free surface algorithm based on the Level Set Method (LSM) to determine the complex water surface profiles. The code is used to analyze the hydrodynamics of compound channel flow with deep and shallow overbanks, free flow around a bridge abutment, pressure flow with a partially submerged bridge deck and bridge overtopping flow. All simulations are validated with data from complementary physical model tests under analogous geometrical and flow conditions. Primary velocity, bed shear stress, turbulence characteristics and 3D coherent flow structures are examined thoroughly for each of the flow cases to explain the hydrodynamics of these complex turbulent flows.

Die vorliegende Arbeit beschreibt die Entwicklung des hochauflösenden Atmosphärenmodells ASAMgpu. Dabei handelt es sich um ein sogenanntes Grobstrukturmodell bei dem gröbere Strukturen mit typischen Skalen von Deka- bis Kilometern in der atmosphärischen Grenzschicht explizit aufgelöst werden. Hochfrequentere Anteile und deren Dissipation müssen dabei entweder explizit mit einem Turbulenzmodell oder, wie im Falle des beschriebenen Modells, implizit behandelt werden. Dazu wurde der Advektionsoperator mit einem dissipativen Upwind-Verfahren dritter Ordnung diskretisiert. Das Modell beinhaltet ein Zwei-Momenten-Schema zur Beschreibung mikrophysikalischer Prozesse. Ein weiterer wichtiger Aspekt ist die verwendete thermodynamische Variable, die einige Vorteile herkömmlicher Ansätze vereint. Im Falle adiabatischer Prozesse stellt sie eine Erhaltungsgröße dar und die Quellen und Senken im Falle von Phasenumwandlungen sind leicht ableitbar. Außerdem können die benötigten Größen Temperatur und Druck explizit berechnet werden. Das gesamte Modell wurde in C++ implementiert und verwendet OpenGL und die OpenGL Shader Language (GLSL) um die nötigen Berechnungen auf Grafikkarten durchzuführen. Durch diesen Ansatz können genannte Simulationen, für die bisher Supercomputer nötig waren, sehr preisgünstig und energieeffizient durchgeführt werden. Neben der Modellbeschreibung werden die Ergebnisse einiger erfolgreicher Test-Simulationen, darunter drei Fälle mit mariner bewölkter Grenzschicht mit flacher Cumulusbewölkung, vorgestellt.

Fourier-based stochastic simulation of wind fields commonly used in wind turbine loads computations is unable to account for contrasting states of atmospheric stability. Flow fields in the stable boundary layer (SBL), for instance, have characteristics such as enhanced wind shear and veering wind direction profiles; the influence of such characteristics on utility-scale wind turbine loads has not been studied. To investigate these influences, we use large-eddy simulation (LES) to generate inflow wind fields and to estimate load statistics for a 5-MW wind turbine model. In the first part of this thesis, we describe a procedure employing LES to generate SBL wind fields for wind turbine load computations. In addition, we study how large-scale atmospheric conditions affect the characteristics of wind fields and turbine loads. Next, in the second part, we study the contrasting characteristics of LES-SBL and stochastic NBL flow fields and their influences on wind turbine load statistics by isolating effects of the mean wind (shear) profile and of variation in wind direction and turbulence levels over the rotor sept area. Among large-scale atmospheric conditions, the geostrophic wind speed and surface cooling rate have the greatest influence on flow field characteristics and, thus, on wind turbine loads. Higher geostrophic winds lead to increased mean and standard deviation values of the longitudinal wind speed at hub height. Increased surface cooling rates lead to steeper shear profiles and appear to also increase fatigue damage associated with out-of-plane blade root moments. In summary, our studies suggest that LES may be effectively used to model wind fields in the SBL, to study characteristics of turbine-scale wind fields, and to assess turbine loads for conditions that are not typically examined in design standards. / text

The present thesis deals with the development and performance analysis ofanisotropy-resolving models for the small, unresolved scales (”sub-grid scales”,SGS) in large eddy simulation (LES). The models are characterised by a descriptionof anisotropy by use of explicit algebraic models for both the subgridscale(SGS) stress tensor (EASSM) and SGS scalar flux vector (EASSFM). Extensiveanalysis of the performance of the explicit algebraic SGS stress model(EASSM) has been performed and comparisons made with the conventionalisotropic dynamic eddy viscosity model (DEVM). The studies include LES ofplane channel flow at relatively high Reynolds numbers and a wide range ofresolutions and LES of separated flow in a channel with streamwise periodichill-shaped constrictions (periodic hill flow) at coarse resolutions. The formersimulations were carried out with a pseudo-spectral Navier–Stokes solver, whilethe latter simulations were computed with a second-order, finite-volume basedsolver for unstructured grids. The LESs of channel flow demonstrate that theEASSM gives a good description of the SGS anisotropy, which in turn gives ahigh degree of resolution independence, contrary to the behaviour of LES predictionsusing the DEVM. LESs of periodic hill flow showed that the EASSMalso for this case gives significantly better flow predictions than the DEVM.In particular, the reattachment point was much better predicted with the EASSMand reasonably well predicted even at very coarse resolutions, where theDEVM is unable to predict a proper flow separation.The explicit algebraic SGS scalar flux model (EASSFM) is developed toimprove LES predictions of complex anisotropic flows with turbulent heat ormass transfer, and can be described as a nonlinear tensor eddy diffusivity model.It was tested in combination with the EASSM for the SGS stresses, and itsperformance was compared to the conventional dynamic eddy diffusivity model(DEDM) in channel flow with and without system rotation in the wall-normaldirection. EASSM and EASSFM gave predictions of high accuracy for meanvelocity and mean scalar fields, as well as stresses and scalar flux components.An extension of the EASSM and EASSFM, based on stochastic differentialequations of Langevin type, gave further improvements. In contrast to conventionalmodels, these extended models are able to describe intermittent transferof energy from the small, unresolved scales, to the resolved large ones.The present study shows that the EASSM/EASSFM gives a clear improvementof LES of wall-bounded flows in simple, as well as in complex geometriesin comparison with simpler SGS models. This is also shown to hold for a widerange of resolutions and is particularly accentuated for coarse resolution. The advantages are also demonstrated both for high-order numerical schemes andfor solvers using low-order finite volume methods. The models therefore havea clear potential for more applied computational fluid mechanics. / <p>QC 20140304</p> / Explicit algebraic sub-grid scale modelling for large-eddy simulations