This study separated two parts: PART ¢¹ Meteorologically adjusted trends of ozone Since meteorological changes strongly affect ambient ozone concentrations, trends in concentrations of ozone upon the adjustment of meteorological variations are important of evaluating emission reduction efforts. This work is to study meteorological effects on the long-term trends of ozone concentration using a multi-variable additive model in Kaohsiung. The long-term trends of ozone concentration were analyzed using the Holland model (without meteorological-adjusted) and the robust MM Regression model (with meteorological-adjusted) based on the data of eight EPA air quality stations from 1997 to 2006 in Kaohsiung area. According to the result of the simulation, the simulated value of the robust MM-Regression model present more valid than the Holland model.The simulated results show that the long-term ozone concentration increases at 13.84% (or 13.06%) monthly (or annually) after meteorological adjustments, less than at 26.10% (or 23.80%) without meteorological adjustments in Kaohsiung county. The simulated results show that the long-term ozone concentration increases at 9.01% (or 6.88%) monthly (or annually) after meteorological adjustments, less than at 22.01% (or 19.67%) without meteorological adjustments in Kaohsiung city. Wind speed, duration of sunshine and pressure are the three dominant factors that influence the ground-level ozone levels in Kaohsiung area. PART ¢º Dispersion of air pollutants in the Hsuehshan Tunnel Concentrations of carbon monoxide (CO) and nitrogen oxides (NOx) were measured from November 14 ¡V 17 2008 in a cross-mountain Hsuehshan traffic tunnel stretching 12.9 km and containing eastward and westward channels. Air pollutants of CO (carbon monoxide) and NOx (nitrogen oxides) will be monitored at the inlet, outlet and vertical shafts of the tunnel. Meanwhile, numerical simulation of three-dimensional turbulent flow will be performed using STAR-CD software. Traffic and pollutant concentrations during the weekends exceeded those during the weekdays. Measured concentrations of CO at the two tunnel outlets (14.5 ¡V 22.8 ppm) were approximately three times higher than those at the two tunnel inlets (3.2 ¡V 7.3 ppm), while concentrations of NOx at the two tunnel outlets (1.9 ¡V 2.9 ppm) were approximately four to five times higher than those at the two tunnel inlets (0.3 ¡V 0.8 ppm). The outlet of vertical draft 2 had the highest pollutant concentrations (CO = 12.3 ppm; NOx = 1.9 ppm), followed by vertical drafts 1 and 3. Three-dimensional turbulence modeling results indicate that airflow in the tunnel was primarily driven by the combined effects of axial fans and vehicles. Results of this study demonstrate that simulated pollutant concentrations increase downstream and are vertically stratified, due to tailpipe exhausts close to tunnel floor. Simulations agreed fairly well with measurements.

In this thesis study, implementation, numerical stability and convergence rate issues of turbulence modeling are explored. For this purpose, a one equation turbulence model, Spalart-Allmaras, and a two-equation turbulence model, SST k-w, are adapted to an explicit, cell centered, finite volume method based, structured / hybrid multi grid flow solver, SENSE2D, developed at TUBITAK-SAGE. Governing equations for both the flow and the turbulence are solved in a loosely coupled manner, however, each set of equations are solved using a coupled, semi-implicit solution algorithm. In multigrid solutions, the semi-implicit solution algorithm and the turbulence model equations are employed only in the finest level grid. As a result, stable and convergent numerical solutions are obtained. In order to validate the turbulence models and the semi-implicit solution algorithm implemented, turbulent flow solutions over a flat plate, RAE2822 airfoil and NLR7301 multi element airfoil are performed. The results are compared with the experimental data and the numerical results of the commercial CFD package FLUENT. It is shown that the numerical results obtained by SENSE2D are in good agreement with the experimental data and the FLUENT results. In addition to the turbulence modeling studies, convergence rate studies are also performed by multigrid and semi-implicit solution methods. It is shown that, the convergence rates of the semi-implicit solutions are increased about 5 times for single grid and 35% for multigrid solutions in comparison to the explicit solutions.

Axial compressors are used in a wide variety of aerodynamic applications and are one of the most important components in aero-engines. The operability of compressors is however limited at low-mass flow rates by fluid dynamic instabilities such as stall and surge. These instabilities can lead to engine failure and loss of engine power which can compromise the aircraft safety and reliability. Therefore, a better understanding of how stall occurs and the causes behind its inception is extremely important. In the vicinity of the stall line, the flow field is inherently unsteady due to the interactions between adjacent rows of blades, formation of separation cells, and the viscous effects including shock-boundary layer interaction. Accurate modeling of these phenomena requires a proper set of stable and accurate boundary conditions at the rotorstator interface that conserve mass, momentum and energy, while eliminating false reflections. As a part of this effort, an existing 3D Navier-Stokes analysis for modeling single stage compressors has been modified to model multi-stage axial compressors and turbines. Several rotor-stator interface boundary conditions have been implemented. These have been evaluated for the first stage (a stator and a rotor) of the two stage fuel turbine on the space shuttle main engine (SSME). Their effectiveness in conserving global properties such as mass, momentum, and energy across the interface, while yielding good performance predictions has been evaluated. While all the methods gave satisfactory results, a characteristic based approach and an unsteady sliding mesh approach are found to work best. Accurate modeling of the formation of stall cells requires the use of advanced turbulence models. As a part of this effort, a new advanced turbulence model called Hybrid RANS/KES (HRKES) has been developed and implemented. This model solves Menter's k--SST model near walls and switches to a Kinetic Eddy Simulation (KES) model away from walls. The KES model solves directly for local turbulent kinetic energy and local turbulent length scales, alleviating the grid spacing dependency of the length scales found in other Detached Eddy Simulation (DES) and Hybrid RANS/LES (HRLES) models. Within the HRKES model, combinations of two different blending functions have been evaluated for blending the near wall model to the KES model. The use of realizability constraints to bound the KES model parameters has also been studied for several internal and external flows. The current methodology is used in the prediction of the performance map for the NASA Stage 35 compressor configuration as a representative of a modern compressor stage. The present approach is found to satisfactory predict the onset of stall. It is found that the rotor blade tip leakage vortex and its interaction with the shock wave is mainly the reason behind the stall inception in this compressor stage.

Most flows in nature and engineering are turbulent, and many are wall-bounded. Further, in turbulent flows, the turbulence generally has a large impact on the behavior of the flow. It is therefore important to be able to predict the effects of turbulence in such flows. The Navier-Stokes equations are known to be an excellent model of the turbulence phenomenon. In simple geometries and low Reynolds numbers, very accurate numerical solutions of the Navier-Stokes equations (direct numerical simulation, or DNS) have been used to study the details of turbulent flows. However, DNS of high Reynolds number turbulent flows in complex geometries is impractical because of the escalation of computational cost with Reynolds number, due to the increasing range of spatial and temporal scales. In Large Eddy Simulation (LES), only the large-scale turbulence is simulated, while the effects of the small scales are modeled (subgrid models). LES therefore reduces computational expense, allowing flows of higher Reynolds number and more complexity to be simulated. However, this is at the cost of the subgrid modeling problem. The goal of the current research is then to develop new subgrid models consistent with the statistical properties of turbulence. The modeling approach pursued here is that of "Optimal LES". Optimal LES is a framework for constructing models with minimum error relative to an ideal LES model. The multi-point statistics used as input to the optimal LES procedure can be gathered from DNS of the same flow. However, for an optimal LES to be truly predictive, we must free ourselves from dependence on existing DNS data. We have done this by obtaining the required statistics from theoretical models which we have developed. We derived a theoretical model for the three-point third-order velocity correlation for homogeneous, isotropic turbulence in the inertial range. This model is shown be a good representation of DNS data, and it is used to construct optimal quadratic subgrid models for LES of forced isotropic turbulence with results which agree well with theory and DNS. The model can also be filtered to determine the filtered two-point third-order correlation, which describes energy transfer among filtered (large) scales in LES. LES of wall-bounded flows with unresolved wall layers commonly exhibit good prediction of mean velocities and significant over-prediction of streamwise component energies in the near-wall region. We developed improved models for the nonlinear term in the filtered Navier-Stokes equation which result in better predicted streamwise component energies. These models involve (1) Reynolds decomposition of the nonlinear term and (2) evaluation of the pressure term, which removes the divergent part of the nonlinear models. These considerations significantly improved the performance of our optimal models, and we expect them to apply to other subgrid models as well. / text

The aerospace industry has been concerned with propeller noise levels for years. This interest is two-fold: government regulation and comfort in cabin. This report attempts to create a simulation mechanism needed to evaluate the far-field noise generation levels. However, in order to do that, the tandem cylinder case was evaluated first as a validation step before the SR-7A propeller case was performed. Both cases use STAR-CCM+, a commercial software, to perform the simulations.

In this work, the mathematical implications of merging two different turbulence modeling approaches are addressed by deriving the exact hybrid RANS/LES Navier-Stokes equations. These equations are derived by introducing an additive-filter, which linearly combines the RANS and LES operators with a blending function. The equations derived predict additional hybrid terms, which represent the interactions between RANS and LES formulations. Theoretically, the prediction of the hybrid terms demonstrates that the hybridization of the two approaches cannot be accomplished only by the turbulence model equations, as it is claimed in current hybrid RANS/LES models. The importance of the exact hybrid RANS/LES equations is demonstrated by conducting numerical calculations on a turbulent flat-plate boundary layer. Results indicate that the hybrid terms help to maintain an equilibrated model transition when the hybrid formulation switches from RANS to LES. Results also indicate, that when the hybrid terms are not included, the accuracy of the calculations strongly relies on the blending function implemented in the additive-filter. On the other hand, if the exact equations are resolved, results are only weakly affected by the characteristics of the blending function. Unfortunately, for practical applications the hybrid terms cannot be exactly computed. Consequently, a reconstruction procedure is proposed to approximate these terms. Results show, that the model proposed is able to mimic the exact hybrid terms, enhancing the accuracy of current hybrid RANS/LES approaches. In a second effort, the Two Level Simulation (TLS) approach is proposed as a near-wall model for LES. Here, TLS is first extended to compressible flows by deriving the small-scale equations required by the model. The full compressible TLS formulation and the hybrid TLS/LES approach is validated simulating the flow over a flat-plate turbulent boundary layer. Overall, results are found in reasonable agreement with experimental data and LES calculations.

Des méthodes numériques sont présentées qui permettent la simulation de jets chauds débouchants dans un écoulement transverse aux grands nombres de Reynolds et aux rapports des vitesses faibles. Différentes approches pour la modélisation de turbulence, c'est-à-dire URANS, SAS, DDES et ELES, sont validées par comparaison à des données expérimentales pour une configuration générique, soulignant la nécessité de résoudre les différentes échelles turbulentes pour une prévision correcte du mélange thermique. L'analyse de la solution instationnaire permet l'identification de processus dynamiques intrinsèques ainsi que des phénomènes de mélange et l'application de l'analyse en composantes principales révèle l'ondulation latérale du sillage de jet. Du fait du caractère multi-échelles qui se manifeste dans la simulation d'un jet débouchant sur une configuration avion, l'approche séquentielle basée sur le modèle SAS est mise en place. Comme les résultats pour la sortie d'un système de dégivrage de nacelle sont en bon accord avec les données d'essai en vol, cette approche est finalement appliquée à la sortie complexe d'un système de pre-cooler, mettant en valeur sa capacité à être appliquée dans un processus industriel. / Numerical methods for the simulation of hot jets in cross flow at high Reynolds numbers and small momentum ratios are presented. Different turbulence modeling strategies, i.e. URANS, SAS, DDES and ELES, are validated against experimental data on a generic configuration, highlighting the necessity of scale-resolution for a correct prediction ofthermal mixing. The analysis of transient flow simulations allows the identification of inherent flow dynamics as well as mixing phenomena and the application of the Proper Orthogonal Decomposition revealed the lateral wake meandering as being one of them. Due to the multi-scale problem which arises when simulating jets in cross flow on real aircraft configurations, the sequential approach based on the SAS turbulence model is introduced. As results for the exhaust of a nacelle anti-icing system comprising multiple jets in cross flow agree well with flight test data, the approach is applied in a last step to the complex exhaust of a pre-cooling system, emphasizing the capabilities of this methodology in an industrial environment.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The turbulent flow simulation through the Boussinesq s hypothesis is represented, currently, by two distinct methodologies, the Large-Eddy Simulation (LES) and the Reynolds Averaged Navier-Stokes Equations (RANS). New Hybrid RANS/LES methods are in development, taking off advantage of LES and RANS potentialities through a one only model. The present work deals with the evaluation of these three methodologies, LES, RANS and Hybrid RANS/LES through the turbulent backward-facing step flow simulation. This classical flow is a benchmark for new turbulence models due to the fact that, despite its simple geometry, it presents a very complex generation of three-dimensional structures, influencing the transition phenomenon and properties such as characteristics frequencies of vortex emission and reattachment length. Parallel to this, an inlet turbulent boundary condition influence study showed that the statistical and topological content of the inlet boundary layer profile can modify substantially results like reattachment length and pressure coefficient. A recycling method for generating three-dimensional, time-dependent turbulent boundary layer inflow data for Large-Eddy and Hybrid RANS/LES simulation is employed. Results for the three methodologies disclose that Large-Eddy Simulation and Hybrid RANS/LES methods present very similar descriptions for the turbulent backward-facing step flow, differing from the RANS s results, where the second order statistical moments are totally suppressed, with absence of three-dimensional and transient structures. / A simulação numérica de escoamentos turbulentos através da hipótese de Boussinesq é representada, atualmente, por duas grandes metodologias distintas, a Simulação de Grandes Escalas (LES Large-Eddy Simulation) e as Equações Médias de Reynolds (RANS Reynolds Averaged Navier-Stokes). Uma nova metodologia, chamada de Híbrida RANS/LES, está em desenvolvimento, tirando proveito das potencialidades das metodologias tradicionais LES e RANS através de um único modelo. O presente trabalho trata da avaliação das três metodologias, LES, RANS e Híbrida RANS/LES de modelagem da turbulência através da simulação numérica do escoamento turbulento sobre um degrau. Os modelos são avaliados através deste escoamento, que apesar de simples geometricamente, é capaz de gerar um escoamento complexo, com regiões de escoamento parietal e cisalhante livre. Juntamente com a modelagem da turbulência, um estudo de imposição de condições de contorno turbulentas na entrada do domínio utilizado revelou que tão importante quanto o modelo de turbulência, as condições de contorno empregadas modificam substancialmente os resultados obtidos. Foi implementado um modelo de geração de contorno baseado no escalonamento de informações internas do escoamento de forma a satisfazer estatística e topologicamente o caráter turbulento da condição de contorno na entrada. Resultados para as três metodologias revelam que a Simulação de Grandes Escalas e métodos Híbridos RANS/LES apresentam descrições muito semelhantes para o escoamento turbulento sobre o degrau, diferindo dos resultados da metodologia RANS, onde momentos estatísticos de segunda ordem são suprimidos, com ausência de estruturas tridimensionais e transientes. / Mestre em Engenharia Mecânica

No description available.

<div>Shock-wave/turbulent boundary-layer interactions (SWTBLIs) are characterized by low-frequency unsteadiness, amplified aerothermal loads, and a complex three-dimensional flowfield. Presence of a broad range of length and time-scales associated with compressible turbulence generates additional gasdynamic features that interact with different parts of the flowfield via feedback mechanisms. Determining the physics of such flows is of practical importance as they occur frequently in different components of a supersonic/hypersonic aircraft such as inlets operating in both on- and off-design conditions, exhaust nozzles, and control surfaces. SWTBLIs can cause massive flow separation which may trigger unstart by choking the flow in an inlet. On control surfaces, fatigue loading caused by low-frequency shock unsteadiness, coupled with high skin-friction and heat transfer at the surface, can result in failure of the structure.</div><div><br></div><div>The objective of this study is twofold. The first aspect involves examining the causes of unsteadiness in SWTBLIs associated with two geometries – a backward facing step flow reattaching on to a ramp, and a highly confined duct flow. Signal processing and statistical techniques are performed on the results obtained from Delayed Detached-Eddy Simulations (DDES) and Implicit Large-Eddy Simulations (ILES). Dynamic Mode Decomposition (DMD) is used as a complement to this analysis, by obtaining a low-dimensional approximation of the flowfield and associating a discrete frequency value to individual modes. </div><div><br></div><div>In case of the backward facing step, Fourier analysis of wall-pressure data brought out several energy dominant frequency bands such as separation bubble breathing, oscillations of the reattachment shock, shear-layer flapping, and shedding of vortices from the recirculation zone. The spectra of reattachment shock motion suggested a broadband nature of the oscillations, wherein separation bubble breathing affected the low-frequency motion and shear-layer flapping, and vortex shedding correlated well at higher frequencies. A similar exercise was carried out on the highly confined duct flow which featured separation on the floor and sidewalls. In addition to the low-frequency shock motions, the entire interaction exhibited a cohesive back-and-forth in the streamwise direction as well as a left-right motion along the span. Mode reconstruction using DMD was used in this case to recover complex secondary flows induced by the presence of sidewalls.</div><div><br></div><div>For the final aspect of this study, a flow-control actuator was computationally modeled as a sinusoidally varying body-force function. Effects of high-frequency forcing at F⁺ =1.6 on the flowfield corresponding to a backward facing step flow reattaching on to a ramp were examined. Conditionally averaged profile of streamwise velocity fluctuations, based on reattachment shock position, was used for the formulation of spatial distribution of the actuator. The forcing did not change the mean and RMS profiles significantly, but affected the unsteadiness of the interaction significantly. The effects of forcing were localized to the recirculation zone and did not affect the evolution of the shear-layer. The acoustic disturbances propagating through the freestream and recirculation zone drove the motion of the reattachment shock, and did not alter the low-frequency dynamics of the interaction.</div>