Turbulent plumes play a crucial role in geophysical and industrial processes, yet their behavior in rotating environments remains insufficiently understood. This study experimentally investigates the influence of background rotation on negatively buoyant turbulent plumes using Particle Image Velocimetry (PIV) and Laser-Induced Fluorescence (LIF). By varying the source Richardson and Rossby numbers, the effect of rotation on plume self-similarity, width and centerline velocity is examined.
Results reveal that rotation significantly alters plume dynamics, reducing entrainment and modifying self-similarity profiles at the edges of plumes. The evolution of plume half-width is inhibited, and centerline velocity decays faster in rotating environments. The curved area of both buoyancy and velocity self-similarity profiles do however overlap with a Gaussian curve.
These findings contribute to a better understanding of rotating turbulent plumes and provide insights relevant to oceanic convection, atmospheric dynamics, and industrial mixing processes.

Buoyancy-driven plumes are natural phenomena that occur in widespread applications, such as geological flows or pollutant dispersion from a chimney. One characteristic of plumes is the entrainment process, where the plume stream drags in ambient fluid and mixes with the ambient fluid, and classic plume theory has used an entrainment coefficient to describe the entrainment process. Literature research reveals that there is a general lack of experiment data on the entrainment process of an axisymmetric plume, and there is little consensus on the entrainment coefficient. The aims of the thesis are to address the problem of insufficient experimental evidence and to investigate the entrainment coefficient using two theories (classic entrainment theory and new energy-consistent theory).

The thesis’s method is experimental. Buoyancy-driven plumes with different initial conditions were created in a water tank, and the velocity field and buoyancy field were measured using particle image velocimetry (PIV) and laser induced fluorescence (LIF). In addition, a new combination of urea and sodium sulfate was proposed to perform the refractive index matching (RIM), which is a crucial step for accurate velocity and buoyancy measurements.

The thesis’s results highlighted that the entrainment coefficient is approximately 0.11, despite large variations when using the classic theory to determine the entrainment coefficient, which may help explain different values in previous literature. In addition, the existence of a refractive index field was observed to affect both the velocity and buoyancy measurements. Specifically, the refractive index field caused PIV and LIF to overestimate the velocity field and underestimate the buoyancy field, respectively.

Pollution in urban environments is increasingly becoming a major concern for our societies which, together with the increasingly powerful capabilities provided by computational fluid dynamics (CFD), has made pollution dispersion a common topic of research during the last two decades. Several factors are relevant to the dispersion of pollutants in an urban environment, such as the variable roughness of the urban canopy and the thermal stratification conditions. However, the effects of these two factors are only partially understood, especially when it comes to their combined effects.

The present research work has been motivated by an unintuitive strengthening of the streamwise-elongated coherent structures of the atmospheric boundary layer (ABL), observed by Jayaraman & Brasseur (2021), for a certain range of weakly unstable thermal stratification conditions. Such observation evinced the existence of a sweet-spot where the large-scale streamwise-oriented counter-rotating vortical structures, characteristic of any turbulent boundary layer, were able to strengthen themselves by collecting small-scale thermal plumes, eventually creating the so-called large-scale atmospheric rolls. The importance of these findings lies in the fact that large-scale atmospheric rolls are known to be one of the most important structures when it comes to vertical transport of momentum and scalars (i.e. pollutants) in the ABL.

Following from this observation, we hypothesize that the counter-rotating roll structures behind a high-rise building could potentially exhibit a similar behaviour and strengthen themselves under weakly unstable thermal stratification conditions by collecting small-scale thermal plumes. If this was the case, it would enhance large-scale vertical mixing which could be beneficial for street canyon ventilation, reducing urban pollution levels. Therefore, the importance of studying the evolution of the roll structures behind a high-rise building under different thermal stratification conditions becomes evident.

In this regard, the present work has focused on the study of the roll structures behind a high-rise building under different thermal stratification conditions, by means of wall-resolved Large-Eddy Simulations (LES) using the Nek5000 Spectral Element Method (SEM) code. The present work has succeeded in providing a qualitative proof of concept of the presented hypothesis, proving that the rolls behind a surface-mounted cube (model for a high-rise building) under weakly unstable thermal stratification conditions do strengthen themselves in a similar manner as observed by Jayaraman & Brasseur (2021) for the ABL. Moreover, a reliable LES numerical model for simulating flow past a surface-mounted cube (high-rise building) has been built and thoroughly validated against DNS data, tuning the optimal mesh, and simulation and filtering settings.

This thesis aims to automatically and reliably detect large-scale structures in turbulent flows. To achieve this, a U-net (a type of neural network) is trained using artificially generated data. From the network output, continuous structures are derived and general statistics, including, volume fraction, relative kinetic energy and length scales (using PCA) are computed.

Detections were done on two homogenous isotropic turbulence (HIT) direct numerical simulation (DNS) datasets with Taylor Reynolds numbers of 175 and 1131, respectively. For both cases, the detected structures contained most of the volume and kinetic energy in the domain and were of the integral length scale. However, for the high Reynolds number, there were relatively half as many structures and the structures were roughly 4 times larger compared to those found in the low Reynolds number case.

With ongoing research towards clean combustion, hydrogen has been identified as a potential alternative to natural gas fuel, for example in power generation sectors utilizing gas turbines because of their inherent nature of being a carbon-free energy carrier. However, it is crucial to clarify that the ultimate goal is not just carbon-free combustion but clean combustion, which entails addressing other post-combustion emissions, such as NOx (nitric oxide) emissions. To meet stringent NOx emissions regulations, gas turbine fuels are combusted under premixed conditions. But in these premixed conditions, it has been observed that flames produced as a result of combustion have a tendency to flashback, which means that the flame travels back into the premixing chamber causing severe structural damages. This is particularly concerning when using hydrogen, which due to its high reactivity and flame speed is more prone to flashback than natural gas. To understand this high propensity of hydrogen to flashback, there is a strong requirement to examine the topology or structure of flame which can be obtained using optical combustion diagnostics technique.

Considering a flame has a three-dimensional structure, in this thesis, an optical combustion diagnostics technique of Computed Tomography of chemiluminescence (CTC) was applied to Bunsen burner flames using six CCD cameras. The cameras were arranged around the flame and a tomographic algorithm was used to reconstruct the three-dimensional structure of the flame. The technique was applied to turbulent 100% by volume fraction Dutch Natural Gas (DNG) flames at various Reynolds numbers representing stable, close to flashback, and flashback case. The reconstructed DNG flame results highlighted the capability of the CTC technique to offer valuable insights into the intricate features of the flame. Furthermore, these results not only indicated the possible location of the origin of the flashback within the structure of flame but also revealed specific features associated with events prior to the flashback. recognizing the potential of this technique, it was subsequently applied to a turbulent flame consisting of 50% hydrogen blended with 50% DNG. The reconstruction results offered insights into the fundamental structural differences between a flame consisting of 50% hydrogen and 50% DNG and a pure 100% DNG flame. The conclusions drawn from the visual assessments of the reconstruction results were further supported by the subsequent statistical analysis and the resulting cone angle values.

The need to reduce greenhouse gas emissions is emergent throughout every industrial sector worldwide. For aviation, this has opened a market gap for small scale electric aircraft configurations (e-V/STOL). These configurations are very diverse and they are classified based on how lift is obtained. However, their flight envelope consists of the same operational regimes which in turn, are very diverse compared with the ones of a traditional aircraft. This challenges the way wind tunnel testing was developed and urges the need for new testing techniques and wall correction methods. It is known for over a century that the flow inside a wind tunnel is not exactly the same as a free-air flow. This has to do with the existence of the wind tunnel test section boundaries which affect the flow over the model. Thus, the measured wind tunnel data deviate from the desired free-air performance measurements, except if the model is extremely small relative to the test section size. This is almost never satisfied due to physical scaling and structural considerations. Wall or boundary interference, is the quantification of those effects with the aim of correcting the wind tunnel performance measurements to their actual free-flight values. This is achieved by wall correction methods. These methods were developed in absence of high computing power and their purpose is to perform real-time adjustments to wind tunnel data. Consequently, they depend on simplified modeling assumptions, such as the linearized potential flow. As a result, these corrections might prove ineffective in generating reliable outcomes when dealing with non-linear boundary interference fields characterized by significant variations. For a given model, it is generally known that there is a limit at which wind tunnel test data are able to represent free-air conditions. This limit is called wind tunnel flow breakdown and it occurs when the model wake impinges on the boundaries of the wind tunnel test section, causing flow phenomena which would never occur in a free-flight situation. This phenomenon succeeds the non-linearities of boundary interference described above and is the worst case scenario for wind tunnel tests. The wall corrections completely fail to produce any valid outcome and the wind tunnel data are considered useless. This happens at low-speed/high-thrust conditions which are typical for a V/STOL aircraft during transition from hover to forward flight. Thus, powered wind tunnel testing at such conditions poses challenges in terms of data representation to an interference-free, free-air situation. Such conditions may also be encountered when the aircraft is operating in-ground vicinity and thus, the interference field for those cases is of interest. In the latter case, it is expected that the magnitude of the corrections and the wind tunnel flow breakdown limits would be significantly lower. This Thesis aims to investigate wall interference effects between the rotor and the wind tunnel test section boundaries at low-speed/high-thrust conditions, establish the flow breakdown limits for a given propeller and test section and to gain insight into the effect that the bottom wall of the test section exhibits to the boundary interference field. This is realized by conducting wind tunnel tests on two different fixed-pitch propellers with varying incidence angles in both a closed-wall and a 3/4 open-jet test section of the NLR Aeroacoustic Wind Tunnel. Static wall pressure measurements are acquired and the flow-field in the vicinity of the propeller is quantified using large-scale, Tomographic Particle Tracking Velocimetry (PTV) with the use of helium filled soap bubbles (HFSB) as tracers. Finally, the interference-free performance data for these propellers are obtained from the second test campaign performed in the industrial Low-Speed-Wind Tunnel (LST) of the DNW. The global flow topology of the wall-bounded tests in the AWT-closed test section is presented by means of wall pressure measurements and PTV. It was found that a distinct rise of the pressure coefficient leading to a maximum peak, corresponds to wake impingement. This firstly occurs on the advancing side of the rotor due to the larger vortex which is deflected further downward than the one on retreating side. A flow breakdown criterion which makes use of the static pressure distribution on the bottom wall of the test section is introduced. It aligns well with existing benchmarks available in literature for closedwall test sections. This criterion can be applied to any test section consisting of a solid lower wall, as long as the distribution of static pressure is monitored. On the other hand, the comparison between the LST and AWT-closed test section data do not show any signs of flow breakdown. This could prompt additional questions regarding the necessity of an even larger test section for obtaining the interference free propeller performance data. Various wall correction methods are utilized in order to evaluate their applicability for powered wind tunnel testing and to assess the impact of wall interference. It is shown that the developed Two-VariableMethod (TVM) does not provide satisfying results with respect to lift interference, when compared with the Heyson method. Nevertheless, blockage predictions follow the expected trends even for lifting cases, at least for the higher advance ratios where no flow impingement is present. For a simple non-lifting case, this method is deemed more reliable since its underlying assumptions are satisfied to a greater extend and its blockage predictions are also in good agreement with the Glauert’s method. For the purely lifting angle of attack, the comparison between the two test campaigns is not satisfactory both for in-ground and out-of-ground effect. This discrepancy arises from the fact that the predictions of all wall correction methods applied do not abide by the trend that is discernible based on the deviations of the bounded (AWT) to the unbounded (LST) data. That is mainly attributed to the load measurement system and to the possibly stalling conditions at the higher wind tunnel speeds. For the 0◦ incidence angle, there was a very good agreement between the AWT-closed test section and the LST data for both flight conditions (OGE, IGE). This was validated by both applied correction procedures (Glauert’s method for off-center propellers & Two-Variable-Method). Even in that case, the AWT-3/4 open-jet results were not in good agreement implying effects which may not only be attributed to boundary interference.

Aerodynamics has played a significant role in the industry of motorsports in improving the performance and handling of the race car. Rob Smedley, the former head of vehicle performance at Williams Racing stated that - "Where teams have problems is when their development or simulation environment – so CFD [Computational Fluid Dynamics] or wind tunnel – doesn’t describe well what happens in reality (although in truth, no-one’s wind tunnel correlates absolutely 100%)". One of the reasons for this poor correlation could be arising from the fact that, in real life on track scenarios, the race cars undergo accelerating, decelerating, or cornering motion which can have a different influence on the aerodynamics of a race car which is not accounted for in the wind tunnel and simulation environment where a steady constant flow is employed. This research aims to numerically investigate the flow past the front wing of a Formula One car in ground effect subjected to accelerating and decelerating flows to understand the trends in the aerodynamic performance.

A scaling analysis is performed to determine the relevant non-dimensional numbers that influence the flow for translationally accelerating airfoils and two dimensionless numbers are arrived at, namely, the Reynolds number and the Froude number. Numerical investigations are carried out for translationally accelerating wings in ground effect to determine the influence of these dimensionless numbers on the aerodynamic forces.

Transient simulations were performed on a two-dimensional airfoil and a three-dimensional wing in ground effect subjected to translational acceleration and deceleration. The Shear Stress Transport (SST) based on k-ω was employed to model the turbulent flow. The results from the numerical investigations revealed a temporary change in the downforce and the drag force coefficients, as the airfoil (or wing) in ground effect is subjected to translational acceleration (or deceleration). In this study, the mechanisms that contribute to this temporary change in the aerodynamic force coefficients are discussed.

Reduction of skin-friction drag over a fully developed canonical zero-pressure gradient turbulent boundary layer (ZPGTBL) subjected to spanwise oscillation is measured using planar particle image velocimetry (PIV). The experiments are conducted at Re_θ of 1000 and 1800, the chosen range of spanwise oscillations amplitude and frequency are around the optimum reported in literature studies (T⁺_osc = [100-700], A⁺_osc= [50, 150]). A high-resolution planar PIV apparatus is employed to measure the drag reduction directly by wall shear stress estimates on the oscillating wall. The measurement uncertainty of the drag estimates from PIV measurements is examined. The results show drag reduction in the order of 15% after 6 boundary layer thicknesses from the beginning of the oscillating section. Variations of the drag reduction follow the trends reported in literature. The PIV measurements enable the analysis in terms of Reynolds shear stresses, turbulence production and allow visualising vortex packets by vorticity. A pronounced drop of turbulence production in the range y⁺ = [5, 30] is observed. The vorticity analysis indicates a distortion from the well-known hairpin-packet arrangement, suggesting that the mechanism of hairpin auto-generation may be inhibited by spanwise wall oscillations

A tip-vortex in the wake of a finite length lift generating surface has a low pressure region near the axis of the tip-vortex. This low pressure region can trigger tip-vortex cavitation especially in ship propellers which has adverse effects by causing vibration and underwater noise production. One way of minimising the tip-vortex cavitation is by reducing the propeller blade loading but at a cost of lower operating efficiency of the propeller. An alternative passive approach, to delay the onset of cavitation, by increasing the minimum pressure in the core of the tip-vortex through modification of the tangential velocity profile of the tip-vortex is investigated in this thesis. In this thesis, the effect of upstream turbulence on the peak tangential velocity of the tip-vortex from a NACA 662-415 hydrofoil is investigated through stereo-Particle Image Velocimetry measurements. The turbulence is produced by a passive grid placed in the upstream of the hydrofoil. The tangential velocity profile at three different turbulent intensities were compared with measurements from a baseline no grid case that had nearly uniform and steady inflow to the hydrofoil. The tangential velocity profile of the vortex exhibits a reduction in the average peak tangential velocity under the influence of the grid generated turbulence compared to the no grid case. The measurements taken at two downstream locations, 0.5 and 1.3 chord length downstream from tip of the hydrofoil, showed that a significant reduction in the magnitude of the average peak tangential velocity of the tip-vortex is observed only at the latter position. Also, on examining the axial velocity profiles, a significant reduction in the magnitude of the average peak axial velocity in the core of the tip-vortex was observed under the influence of grid generated turbulence. An apparent effect of the upstream turbulence on the tangential velocity of the tipvortex is seen to appear only when sufficient time is given for a secondary vortex structure, from the upstream turbulence, to form around and be intensified by the primary tip-vortex. At 1.3 chord length downstream from the tip of the hydrofoil, the reduction in the magnitude of the peak tangential velocity is accompanied by increase core radius of the tip-vortex.

The way in which mean strain affects the turbulent structures is imperative to understand various natural flows such as flow over a hill, the flow of a river in the delta, jet streams in the upper atmosphere etc. Further, it also has industrial implications viz; flow over bodies such as airfoil, turbomachinery, gas pipelines.
The strained pipe flows, in particular, have huge engineering interest due to its prevalence in industrial fittings wherein a larger pipe diameter is connected to a smaller one and vice versa. This subject also has a fundamental interest as strain highlights the interaction of various scales of turbulence. This scale interaction
essentially dictates transfer of energy in turbulence and hence is fundamental in understanding turbulence dynamics itself. The present work deals with the experimental study of the response of pipe turbulence to axisymmetric, irrotational strain using high-resolution planar Particle Image Velocimetry (PIV). A mildly rapid strain (s^*s ~ 3.2) is imposed on turbulence via a spatial contraction. It is seen that turbulence is suppressed upon straining. As a response to mean strain, transverse Reynold stress increases at the expense of streamwise Reynolds stress and anisotropy is induced in the turbulence. Despite strain being only mildly rapid, Rapid Distortion Theory (RDT) is found to predict the correct trend of normal Reynolds stress although traverse Reynolds stress is over-predicted. The effect of strain on different scales of turbulence is discerned. The large scales of turbulence are seen to get compressed in the radial direction although they do not get affected significantly in the streamwise direction. Near-wall coherent structures which were initially inclined w.r.t. the wall are seen to get aligned with the flow as they also get severely compressed in the radial direction. On the other hand, the small scales of turbulence are found to be spatially organised in the form of sheets or layers. Upon straining, these sheets are found to get aligned with the mean flow. Further, they get elongated in streamwise and compressed in the radial direction. It is observed that the small scales are more severely distorted than the large scales upon staining inside the contraction. At the Reynolds number (Re) range employed in this thesis, there is no substantial difference in which turbulence is strained inside the contraction at disparate Re. Downstream of the contraction, at the axis of the pipe, the anisotropy of Reynolds stress is found to recover slowly. Further, this relaxation is seen to be Re dependent with higher Re turbulence relaxing slightly faster.

Turbulence is a commonly encountered state of fluid dynamics. Unsteady turbulent flows in pipes are present in many engineering applications and also in biological flows. However, the various processes active in such flows are not well understood. The present work employs stereo-PIV to investigate the effects of a sinusoidal pressure gradient on the various turbulence parameters, including the terms of the turbulent kinetic energy budget equation. The bulk flow rate was oscillated with a frequency of 0.5 Hz with a mean Re 26,000 and an amplitude of modulation, 0.23 times the mean value. It is seen that there is a delay in the response of turbulence to the oscillations of the bulk flow and the delay increases with increasing distance from the wall. The axial and the in-plane turbulence parameters show a difference in the delay of their responses. This delay extends into the small scales responsible for the dissipation of turbulent kinetic energy. Changes are also observed in azimuthal length scales when the flow oscillates.The effects of oscillation on the streaks of low momentum are also discussed and the structural organization in unsteady pipe flows are found to be different from that in steady pipe flows

Aerodynamics has become increasingly important these days in the design of automobiles due to the increased constraints placed on the emissions. In the past the focus has been primarily on the shaping of the upper body. However, due to practical requirements, further scope of improvements to the upper body has been greatly reduced. The focus has thus shifted towards the optimization of the underbody which includes the wheel and the wheelhouse. The contribution of the wheel and the wheelhouse to the total drag and lift coefficient of a body has been estimated at 30\% and 40\% respectively. Therefore, understanding the behavior of the flow in the wheelhouse would potentially result in reducing the drag coefficient of an automobile.

In the present study, the wheelhouse flow has been investigated experimentally using Particle Image Velocimetry (PIV). A generic simplified car body with a wheel and a wheelhouse has been designed and used for this purpose. The study has been done for a rotating wheel and stationary ground configuration. Planar PIV and Stereo PIV have been used to study the flow inside the wheelhouse and in the wake of the wheel respectively. The study was undertaken for the Reynolds number, based on wheel diameter, of 9000 and 72000. Measurements were taken at planes inside and outside the wheelhouse as well as upstream and downstream of the wheel. In addition to providing data for the Large Eddy Simulation (LES), the results from the PIV measurements showed that the behavior of the flow depended on the Reynolds number. Velocity profiles in the wake of the wheel highlighted a three dimensional flow in the wheelhouse.

Finally, the effect of a Vortex Generator (VG) on the flow was investigated. The VG was found to have an effect at the low Reynolds number wherein the flow inside the wheelhouse and in the underbody was found to be quite different compared to the configuration without a VG.

People living or working close to the highways are at risk to higher exposure levels of different air pollutants, one of which is particulate matter. The Antea group has recently proposed the addition of Electrostatic Precipitators (or ESPs) on the top of sound barriers adjacent to highways, in order to capture the particulate matter and safely discard them. This is an attractive, but expensive, supplement. While estimates of the collection efficiency (i.e. what percentage of particulate matter mass entering the ESP is collected by the device) are known, the aerodynamic efficiency of the ESP (i.e. what percentage of pollutants reach the ESP entrance as compared to the mass emitted) is yet to be determined. The net efficiency of the ESP will be the product of the above two efficiencies, making it necessary to study the latter, prior to large-scale installation.

Scaled down experiments were performed in a water channel, housed in the Laboratory for Aero and Hydrodynamics at the Delft University of technology. Simultaneous measurements of the flow velocities and dye intensities in the symmetry-plane of the channel were made through planar Particle Image Velocimetry (PIV) and Planar Laser Induced Fluorescence (PLIF). A few flow configurations were studied by varying certain parameters. It was found that: (a) ESP installation is more beneficial on top of shorter sound barriers. (b) Slightly raising the ESP from the top of the sound barrier is advantageous. (c) Installation of ESPs in isolated highway canyons should be done with care, as several flow regimes are observed.

First order estimates on the aerodynamic efficiencies of the ESP were determined. This was accompanied by the identification of ESP performance trends that should serve as a guideline for the initial testing of the ESPs at full scale. The guidelines include the physical location and entrance orientation of the ESP. It is known that the collection efficiency of the ESP is a function of the incoming flow velocity. It is recommended that the entrance flow velocities reported here be used as an input for calculating the associated particle collection efficiency. Together, the collection and aerodynamic efficiencies can be used to estimate the overall efficiency of the ESP. If the overall efficiency satisfies a minimum desirable threshold, the field experiments yield promising results, the device clears safety requirements, and the benefits outweigh the costs, the ‘Open Air Line ESP’ can be installed along the highways en masse in the future.

In this thesis a vortex generator is presented, that can generate controlled ring-vortex injections into a pipe flow. The performance of the generator and the coherence of its structures is tested by applying a dye visualisation method. Furthermore, iso-contour plots of the 3D velocity field, obtained from Stereoscopic Particle Image Velocimetry (SPIV) data, reveal that the injection develops into a hairpin- or ring-vortex with hairpin packets.
In future applications, this vortex generator will be used to study the interaction between the controlled ring-vortex structures and the turbulent structures in the pipe, which will act as a model for scale interactions in turbulence.

A combined PIV and Schlieren measurements have been carried out in the transonic-supersonic wind tunnel (TST-27) to investigate the effect of exhaust plume and the variation in nozzle length on the flow topology and mean pressure distribution on the wake of axisymmetric backward facing step model at freestream Mach numbers of 0.76 and 2.20, respectively. Four different nozzle length configurations with and without the presence of a supersonic exhaust plume have been tested. Testing with different nozzle length configuration resulted in flow cases where the shear layer reattachment occurred on the nozzle (solid reattachment), on the flow downstream of the nozzle (fluidic reattachment), and intermittently on the nozzle and on the flow (hybrid reattachment). A qualitative identification of the effect of exhaust plume and the variation in nozzle length on the flow topology on the wake of axisymmetric backward facing step model, at the above mentioned conditions, is successfully described by means of Schlieren visualization. The topological flow features for the subsonic and supersonic flow cases has been identified (i.e shock waves, Prandtl-Meyer expansion fan, boundary layer, separated shear layer, recompression and plume shocks). On the other hand, a quantitative identification of the effect of exhaust plume and the variation in nozzle length on the flow topology and mean pressure distribution on the wake of axisymmetric backward facing step model, at the above mentioned conditions, is successfully done by means of planar PIV. The planar PIV measurements in the wake region of the model provided detailed information of the mean flow field properties (i.e mean velocity, turbulent kinetic energy, Reynolds stresses and reverse flow). Using these mean flow field properties data, the mean pressure distribution is reconstructed based on the momentum equation.It has been shown that an increase in nozzle length and the presence of an exhaust plume caused an increase in mean reattachment length at freestream Mach number of 2.20, while no significant change in mean reattachment length was noticed at freestream Mach number of 0.76. Significantly higher turbulent kinetic energy levels have been observed for L/D = 1.8 cases where solid reattachment occurred, at freestream Mach number of 2.20. In contrast to supersonic flow cases, the flow cases at freestream Mach number of 0.76 showed a significantly lower turbulent kinetic energy levels. Comparisons of flow cases with a long nozzle without a plume and flow cases with a short nozzle but with a plume suggest that the presence of the plume cannot accurately be modeled by replacing the plume with a solid geometry. From the pressure results it is observed that the location of the low-pressure region downstream of the base remained unchanged for different flow cases with and without exhaust plume and for different nozzle lengths. Furthermore, it has been shown that an increase in nozzle length leads to higher local pressure at the nozzle exit and hence results in a less under-expanded for the supersonic flow cases or more over-expanded plume for the subsonic cases.

The aerodynamic drag force is a crucial factor for passenger cars and trucks, where fuel economy is of utmost importance. While a lot of research has been performed to optimize the design of the upper-body of the vehicle, the flow over a rotating wheel that is enclosed by a wheelhouse is not understood to the same level. Estimates show that the addition of wheel and wheelhouse to a vehicle body, can increase its drag and lift coefficient by as much as 40 \%. With the major automobile manufacturers striving to improve the fuel economy of the vehicle, understanding the flow over a rotating wheel in a wheelhouse, offers a new dimension for the improvement of the drag coefficient of the vehicle. In the present research, the flow over a simplified car body with a wheel and a wheelhouse was studied using Large Eddy Simulation (LES). When performing an LES on a given numerical grid, it can be expected that the contributions from the subgrid viscosity model reduces on decreasing the Reynolds number of the flow. A study focusing on the dependence of the flow field on the Reynolds number was thus performed. Three Reynolds numbers based on the wheel diameter - 9 x $10^3$, 75 x $10^3$, 150 x $10^3$- were chosen for the dependence study. A comparison of the relative pressure distribution and the flow topology between the different Reynolds numbers showed that the flow over a rotating wheel in a wheelhouse was qualitatively independent of the Reynolds numbers considered in this work. Experimental results from Particle Image Velocimetry (PIV), performed on a similar vehicle model, were used to validate the numerical solution at a Reynolds number of 9 x $10^3$. The behaviour of the flow in the wheelhouse region was investigated and a new description of the flow field inside a wheelhouse, consisting of 11 vortex structures, was presented in this work. To determine the qualitative dependence between the flow field and the drag coefficient, a comparison of the flow between two geometric configurations, with different wheelhouse radius, was performed at a Reynolds number of 9 x $10^3$. It was found that the configuration with the larger wheelhouse radius had a higher wheelhouse drag coefficient and a lower wheel drag coefficient when compared to the configuration with lower wheelhouse radius. Based on a detailed flow field, four out of the eleven vortices were found to be crucial in determining differences in the drag coefficient of the wheel and wheelhouse. Based on this analysis, an alternate wheelhouse geometry, combining the advantages of the two geometric configurations considered in this work, was proposed, that could potentially lead to a reduction in the drag coefficient of the vehicle.

Shock Wave Boundary Layer Interactions (SWBLIs) occur frequently in supersonic and hypersonic flows and, depending on the strength of the shock wave and the state of the boundary layer, may lead to boundary layer separation. This can result in a multitude of problems, such as an increase in drag, a drop in air intake efficiency and high fatigue loads due to the inherent unsteadiness of the ow phenomenon. Since a turbulent boundary layer has a better capability to withstand the adverse pressure gradient that is inherent to the incident shock wave when compared to a laminar boundary layer, it is beneficial to have a turbulent boundary layer at the interaction. By forcing boundary layer transition only a short distance upstream of the interaction, the capability of the boundary layer to remain attached is increased while retaining the low skin friction coefficient of a laminar boundary layer for as long as possible.

Micro-ramps are currently being investigated as a means of reducing negative effects of shock wave-turbulent boundary layer interactions. A micro-ramp enhances the mixing of the ow by introducing vortices in the ow, thus energizing the boundary layer to overcome the adverse pressure gradient due to an impinging shock. The effect of changing micro-ramp height, Mach number and Reynolds number on the mean ow downstream of a micro-ramp's center-line has been studied extensively.

The coherent structures within the turbulent boundary layers have been investigated extensively over the past five decades. In particular, the long alternating low- and high speed streaks in the near-wall region. Along these streaks ejection and sweep events occur which play a major role in the production of Reynolds shear stress. Hence, studies on turbulent boundary layer control were focused on manipulating the low-speed streaks, to either reduce wall skin-friction by attenuating the streaks or by intensifying the streaks for separation control.