Aerodynamic resistance is a major factor in track-cycling performance, particularly at the high speeds reached in pursuit and sprint events. Most cyclist-aerodynamics studies consider straight-line motion or a uniform yaw angle, since these conditions can be reproduced relatively easily in a wind tunnel or conventional computational domain. True cornering is more difficult to investigate experimentally because the cyclist follows a curved trajectory while leaning into the bend, causing the local relative velocity to vary over the cyclist bicycle system. A standard wind-tunnel setup cannot reproduce this spatially varying flow exactly with one uniform freestream direction. As a result, the aerodynamic influence of cornering remains less well understood than straight-line and fixed-yaw conditions. This thesis therefore performs CFD simulations to study the aerodynamic wake of a track cyclist under straight-line and yawed-flow conditions, while also developing the numerical framework required to simulate cornering.

The simulations are performed with the finite-volume solver INCA using an implicit large-eddy simulation approach, a Cartesian adaptive mesh and a cut-cell immersed-boundary method. The Generic Cyclist Model is used throughout the study, allowing the straight-line simulations to be compared with available wind-tunnel Particle Image Velocimetry measurements. The numerical sensitivity to mesh resolution, wall treatment and immersed-boundary mixing is first assessed. The numerical approach is then used to investigate two fixed leg positions, a complete cyclist–bicycle configuration, and yaw angles of +6° and −6°. Finally, a constant rotating reference frame in a rectangular domain is introduced to represent cornering motion.

The straight-line simulations reproduce the main large-scale features of the measured wake. The 12 mm mesh with Werner–Wengle wall-function treatment provides the best compromise between wake resolution and computational cost for the cyclist-only cases, while changing the immersed-boundary mixing threshold has little influence on the time-averaged velocity and vorticity fields. The absolute drag forces remain strongly underestimated, showing that the present near-wall and immersed-boundary treatment is not sufficiently accurate for reliable force prediction.

Changing the fixed leg position alters both the lower-body wake and the larger wake structures behind the hips and torso. The results show that the effect of crank position extends several hundred millimetres downstream and that a more symmetric leg position does not necessarily produce a more symmetric wake. The fine full cyclist–bicycle simulation gives a more detailed representation of the lower wake and shows the influence of the bicycle frame, crankset and other components on the flow between and behind the legs. The corresponding surface results confirm that the simulation remains wall modelled, with relatively high local y+ values on exposed and separated regions.

The yawed-flow simulations show that a fixed yaw angle of 6° already causes a clear lateral displacement and reorganisation of the wake. The upper-body, hip and leg-related structures all change with yaw direction, while the asymmetric crank position causes the positive- and negative-yaw cases to respond differently rather than as exact mirror images.

For the cornering case, the required geometrical configuration was established using a rotating reference frame and a partitioned rectangular domain. However, a stable production simulation was not obtained. The treatment of the outer boundary conditions remains the main issue to address before the cornering setup can be used for aerodynamic analysis.
... Read more

Turbulence modelling remains a key challenge in the field of computational fluid dynamics (CFD) for accurately resolving fluid flows. A new class of data-driven models have recently gained popularity that aim to quantitatively incorporate data from higher fidelity simulations to improve turbulent closure modelling in Reynolds Averaged Navier Stokes (RANS) models. These models have been particularly well researched in modelling the Reynold stress anisotropy tensor (RST).

This thesis re-formulates the data-driven framework in the context of resolving passive scalar fluxes. Corrections to the scalar flux vector are regressed for two different frameworks: framework 1 that uses the simple gradient diffusion hypothesis (SGDH) model and framework 2: that additionally introduces transport equations for the scalar variance and dissipation. These models are compared to each other for the Jets in-crossflow (JICF) case for resolving the scalar field where baseline RANS models struggle. Both data-driven frameworks show an improvement in modelling the scalar field and turbulent scalar flux vector than the baseline models. The study also indicates that inclusion of additional transport equations doesn’t necessarily improve regression of the corrective fields. ... Read more

Over the past decades, a plateau has been reached in the development of conventional civil aviation passenger aircraft. To achieve significant improvements in efficiency, the aviation industry must investigate unconventional aircraft designs, such as flying wings, which are characterized by a blunt-nosed, highly swept cranked wing. These unconventional configurations can offer more than 10% improvement in cruise performance. At low speeds and high angles of attack, however, these wings exhibit complex vortical flow phenomena and may experience a pitch break, in which the aircraft becomes statically unstable about the pitch axis. Because previous studies mostly focused on wind-tunnel-scale models, this study investigates the effect of Reynolds number on vortex flow features at high angles of attack. Steady-state Reynolds Averaged Navier Stokes (RANS) simulations were performed at both sub- and full-scale Reynolds numbers using the generalized k-ω (GEKO) turbulence model. The results indicate two different flow mechanisms underlying the pitch break phenomenon: vortex lift-off at the sub-scale Reynolds number, and outboard wing flow separation combined with vortex breakdown at the full-scale condition. The findings show that higher Reynolds numbers postpone the onset of pitch break to higher angles of attack, extending the usable flight envelope compared to the sub-scale case. ... Read more

Reduced-order models (ROMs) based on Proper Orthogonal Decomposition (POD) are widely employed to reduce the computational cost of high-fidelity fluid simulations. However, their direct application to multiphase flow fields poses challenges due to the presence of discontinuities at phase interfaces. Abrupt
variations in the input data are poorly represented by truncated linear modal bases, leading to non physical oscillations and a loss of physical fidelity in reconstructed fields. This limitation is particularly critical when reconstructing flow fields in multiphase flow simulations, where accurate predictions of
interface dynamics are vital.

This work presents a physics-informed reduced-order model for multiphase flows across geometries parameterised by a scalar quantity d. High-fidelity simulations corresponding to distinct geometrical configurations are mapped onto a common reference domain, enabling the extraction of a shared modal basis.
These modes capture flow features that are common across the dataset and form the foundation for reconstructing flow fields corresponding to previously unseen parameter values. The role of training data distribution is also investigated, revealing that the ROM operates predominantly as a local interpolator
in the parameter space. In particular, predictive accuracy is found to depend strongly on the proximity of training configurations to the target case, suggesting a trade-off between local accuracy and global robustness.

To address the limitations of conventional POD, the solution fields are decomposed into smooth and discontinuous components. For the volume fraction, the interface is explicitly identified using the α = 0.5 contour and regressed across geometries using a Gaussian Process Regression (GPR) model. The full field is subsequently reconstructed from the predicted interface geometry. For the pressure field, the discontinuous capillary jump is analytically modelled using the Young-Laplace relation, ∆p = σκ, where κ represents the curvature of the interface profile. POD is then applied exclusively to the smooth pressure field, after which the discontinuity is reintroduced using the predicted interface curvature.

The performance of this adapted approach is evaluated against a baseline approach involving the direct application of POD to the raw, discontinuous data. Performance is assessed in terms of standard L²-norm based measures such as the root mean squared error, as well as newly defined discontinuity aware error metrics, including interface sharpness, volume fraction phase purity, and localised gradient and laplacian metrics for the pressure. While the direct POD yields lower global RMSE values for the volume fraction, it produces oscillatory and smeared interface representations. In contrast, the adapted approach delivers physically consistent reconstructions, yielding more accurate predictions of meniscus characteristics. For the pressure field, the adapted approach demonstrates superior performance across all metrics within the practical operating range of reduced-order models.

A comparative analysis further shows that removing discontinuities prior to applying POD leads to a more compact and efficient modal basis, with improved energy capture. The results demonstrate that incorporating physical insight into the reduced-order modelling of discontinuous flows significantly enhances both accuracy and computational efficiency. ... Read more

Machine-learning surrogates are increasingly used to accelerate computational fluid dynamics, yet progress is limited by the lack of benchmarks capturing realistic, time-dependent turbulent flows. This thesis introduces a 13 TB dataset of high-fidelity implicit large-eddy simulations of three-dimensional turbulent wakes behind super-elliptical cylinders. Unlike existing datasets, it captures three-dimensional turbulence with an active energy cascade driven by vortex stretching, combining (i) 380 long-horizon trajectories of 400 time steps each with 3–9 million points per frame on irregular meshes, (ii) systematic variation across geometry, Reynolds number, and angle of attack, and (iii) a temporal resolution that preserves the full inertial subrange of the turbulent energy spectrum. Building on this dataset, state-of-the-art neural operators are evaluated across three prediction tasks of increasing complexity: mean-field prediction from governing parameters, the inverse pressure problem, and long-horizon autoregressive spatio-temporal forecasting. Neural operators accurately recover mean flow fields, yet fail progressively as the target fields gain high-frequency content: fine-scale spatial structure is systematically suppressed in instantaneous flow fields, and all evaluated architectures collapse for temporal predictions. These failure modes are attributed to several concurrent mechanisms: memory constraints forcing sparse point-cloud subsampling, latent-space compression discarding high-frequency spatial content, and a spectral bias of the mean-squared-error objective that de-prioritises the high-frequency residuals dominant in turbulent flows. Together, these results provide a new standard benchmark for turbulent CFD surrogates and show that progress requires methodological advances beyond architecture scaling, including physically motivated loss functions and latent representations capable of retaining fine-grained spatial content. ... Read more

Propellers are often the sole form of ship propulsion, making their design uniquely important in the vessel's operation. Despite many advantages, they suffer from issues like cavitation, vibrations and underwater radiated noise. Flexible composite materials have been proposed as a way to address these issues and increase the overall efficiency envelope, achieved by a passive pitch distribution reduction in the vessel's wake. Due to an increase in variables, optimization is increasingly dependent on the accuracy and validity of numeric simulations.

The objective of this report is to study experimental deformations of flexible propellersin terms of their underlying parametric properties. These are compared to simulations using the unsteady Reynolds averaged Navier-Stokes equations for fluid calculations, coupled with a finite element model for the structural calculations.

A custom preprocessor was developed and used to reconstruct the full parametric blade geometry from digital image correlation measurements. The first 10 mode shapes of the undeformed mesh were found using modal decomposition, which was fed as input to a flexible point cloud registration. The participation factors for each mode shape were optimized to minimize an error function, the symmetric chamfer distance. The deformed mesh was processed with a software library, PropArt, to recover the underlying geometry parameters.

It was found that the hydrodynamic pitch β was the best predictor of rake, skew, and camber deformations, with increased inertial forces increasing the magnitude of the deformation. Pitch deformations are largely predicted by the blades' inertial loading, and is mostly independent of β. The blade thickness and chord length deformations were negligible. Simulations tended to over predict the deformations of the propeller, especially at low β, being exacerbated by increased inertial loading. At high β, the deformations were small and well predicted.

The relation between pitch deformation and inertial loading suggests an optimal material stiffness per operating condition exists, to be tuned for an appropriate amount of pitch reduction. The lack of correlation to β also poses a challenge in the design as the wake peak is where the change in pitch is desired. The increase in camber also leads to an increase in thrust at low β, which is contrary to the desired wake peak thrust reduction. These relations should be investigated further, for a broader range of material stiffnesses to see if this holds. Finally, the difference between experimental results and simulations suggests a systematic error due to the error scaling with the inertial forces, which should be further investigated.
... Read more

The increasing size of wind turbines, the expansion of wind farm areas, and the growing offshore space occupied by wind warms highlight the need for accurate modelling of wake interactions between wind farms. These interactions can significantly influence power production and the optimal layout of wind farms in densely developed offshore regions. This thesis modifies and expands existing forest canopy models by introducing a wind-farm induced momentum sink, a turbulent kinetic energy source and a turbulent kinetic energy dissipation source in the transport equations that represent the effect of the turbines within the wind farm. Wind farms are parametrised and modelled using a 𝑘 − 𝜀 RANS approach in which a neutrally stable atmospheric boundary layer is simulated under offshore conditions. Different modelling approaches and wind turbine spacings within the wind farm are investigated, considering a range of spacings representative of modern offshore wind farm layouts to evaluate their influence on wake recovery and turbulence generation. The results are first verified with higher-fidelity LES results and subsequently validated against wind speed and power production measurements from operational offshore wind farms. The study shows that a modelling approach inspired by forest canopies agrees well with high-fidelity simulations. Moreover, a relationship between wind turbine spacing and the Wind Farm Canopy model coefficients is found. The results indicate that as turbine spacing increases, relatively more turbulent kinetic energy dissipation than turbulent kinetic energy needs to be added to the 𝑘 − 𝜀 transport equations. As a result, increased turbulent kinetic energy dissipation is required to reproduce the observed velocity deficit. This indicates that larger turbine spacings correspond to lower background turbulence levels, leading to a weaker and slower recovery of the velocity deficit. These findings provide guidance for selecting Wind Farm Canopy model coefficients based on turbine spacing, thereby improving the predictive capability of RANS-based simulations for large offshore wind farm clusters and their wake interactions.
... Read more

Closer propulsion-airframe integration through the use of boundary layer ingesting promises significant improvements in aircraft fuel consumption. However, little research has been performed on how these intakes are to be designed for swept-wing aircraft. In this study, a trailing-edge-mounted, parametrically designed buried engine intake is designed for a highly swept transonic wing. To connect the wing to the nacelle, two diagonal ramps are created. These allow for a smooth transition between the two, while minimizing the wetted area and allowing more control over the bottom surface of the connection. The intake is initially designed as an axisymmetric body, after which the intake is reshaped to allow for elliptical highlight and throat profiles. The intake duct can also be angled left and right, up and down, to align it with the incoming flow. This is a key feature, as a highly swept wing causes a significant level of cross-flow. Since the engine is placed near the trailing edge, there is also a large downwashing component to the flow. By angling the intake, it can be aligned with the flow. The intake is designed and parameterized in 3DX’s CATIA to allow quantifiable changes to be made to the design. The designs are tested using a compressible RANS simulation with a 𝑘 − 𝜔 SST turbulence model and third-order MUSCL spatial discretization at 𝑀 = 0.85, 𝐶𝐿 = 0.1719, and 𝑅𝑒 = 104.8 million. The initial baseline design has a TPR of 0.973, despite a section of flow reversal at the crown of the intake, equal to 3.8% of the fan area. The ingested boundary layer is significantly thickened ahead of the intake due to excessive curvature. This causes a DC60 of 0.349, unacceptable for modern aero-engines. The lift coefficient has increased by 12.2%, mainly caused by a sectional lift increase outboard of the engine. The drag coefficient has increased by 4.8%, although the wetted area has increased by 8.8%, indicating an increase in aerodynamic efficiency. The highest pressure recovery found is 0.985 for version 1 and is acceptable compared to conventional podded engines. However, it still featured a large portion of separated and reversed flow at the crown. Therefore, 5 further iterations were made, finally arriving at an intake with a similar TPR of 0.983, but with almost no reversed flow. The remaining total pressure losses are predominantly caused by the ingestion of the wing’s boundary layer. The DC60 has reduced to 0.171, while the SC60 is 0.0874. The lift coefficient has remained the same w.r.t the baseline design, but the drag coefficient has significantly reduced, being 3.0% less than that of the clean wing. During the iteration process, the angling of the intake into the flow was found to be most critical when trying to improve performance. Specifically, the centerline toe angle, angle of incidence, and throat offset were found to have the greatest effect on the flow quality at the fan face. ... Read more

Composite materials can be used to fabricate flexible marine propellers, which can improve efficiency and reduce underwater radiated noise. Since the hydrodynamic performance of flexible propellers is determined by their deformation under fluid loading, similarity laws for flexible propeller scaling should account for the deformation in model experiments. This study introduces a non-dimensional parameter to characterise the deformation of flexible propellers and evaluates it through time-domain fluid–structure interaction simulations. A coupled solver combines unsteady Reynolds-Averaged Navier–Stokes equations with a finite-element structural solver. The study focuses on the Wageningen C4-40 propeller geometry under uniform inflow and is limited to isotropic materials.

A set of non-dimensional relations is derived through dimensional analysis, with a form of the Cauchy number expressing deformation amplitude. Validation using a Reynolds–Cauchy similarity approach on two geometrically similar propellers of different diameters confirms consistent deformation and less than 5\% difference in thrust and torque coefficients between model and full-scale propeller. The disparities in performance results are attributed to numerical artefacts in the fluid solver, as the results indicate that the $k-\omega$ SST turbulence model is sensitive to near-wall resolution.

Achieving full-scale Reynolds numbers in propeller test facilities is not feasible, yet simulations demonstrate that flexible propellers are sensitive to viscous forces. The study observes disparities in deformation extent across Reynolds numbers. The deformation of flexible propellers improves flow attachment over the blades. The overall Reynolds-number trends remain similar to the rigid results: thrust coefficients increase and torque coefficients decrease as Reynolds numbers increase. The open-water efficiency depends on both coefficients, and larger Reynolds numbers result in higher efficiencies.

The Froude–Cauchy scaling approach proves suitable for model experiments; however, material availability limits practical implementation. This study indicates that the steady-state deformation is primarily governed by stiffness, with negligible impact from the structural-to-fluid density ratio. In contrast, in unsteady conditions, the structural-to-fluid density ratio affects the modal frequencies, which describe the dynamic behaviour of propeller blades. Particularly, propellers with high skew, rake, or with anisotropic material properties are affected by structural-to-fluid density. Furthermore, this analysis observes that the first blade mode of a zero-skew angle propeller is pure bending, and a 30\% variation in structural density does not alter the natural frequency of this blade. However, coupled bend-twist modes are sensitive to the structural density, which therefore affects blade deformation in unsteady conditions.

In conclusion, the extent of propeller deformation can be controlled by a non-dimensional parameter expressing the ratio of elastic to hydrodynamic forces. For steady open-water conditions, deformation is additionally a function of Reynolds number. In unsteady conditions, the structural-to-fluid density ratio becomes relevant, as it alters coupled bend-twist modal frequencies. Thus, the deformation extent in this regime is, in addition to the Reynolds number, a function of fluid damping and the ratio of natural frequency to revolution rate. This study offers a basis for accurate scaling of flexible propellers in both experimental and computational studies. ... Read more

The development of reusable hypersonic vehicles presents significant challenges due to the complex and computationally intensive nature of high-fidelity simulations required for aerodynamic performance prediction. This thesis explores the use of Gaussian Process Regression (GPR) as a surrogate modelling technique to efficiently and accurately predict the aerodynamic coefficients—namely drag, lift, and moment—of re-entry vehicles such as capsules and gliders. A multi-output GPR architecture is implemented to capture interdependencies between outputs and reduce the number of required simulations. High-fidelity CFD simulations using the DLR TAU code serve as the training dataset for the surrogate models. The study evaluates various kernel functions, sampling strategies, and model configurations to optimize predictive performance, achieving high accuracy with significantly reduced data requirements. Results show that GPR models can reliably predict aerodynamic coefficients across a wide range of flow conditions, with a mean relative error below 1% for drag in realistic re-entry trajectories. This approach enables the rapid generation of aerodynamic databases, offering a valuable tool for early-stage design and trajectory planning of hypersonic vehicles. ... Read more

Floating offshore wind turbines enable renewable energy expansion into deep-water regions with stronger, more consistent winds. However, they are subject to continuous platform motion from wind and waves, which complicates blade loading and wake aerodynamics. While surge and pitch effects have been extensively studied, the aerodynamic influence of roll motion remains underexplored. Roll motion uniquely induces a non-uniform tangential velocity field across the rotor, which would have implications on the blade loading, structure fatigue and the wake aerodynamics.

This thesis presents a baseline analysis of blade loading and near-wake aerodynamics for a FOWT under prescribed roll motions. A coupled high-fidelity GPU-based Large-Eddy Simulation code GRASP resolved turbulent wake dynamics, while OpenFAST calculated blade loads. Coupling was achieved via the Filtered Actuator Line Method and the AspFAST application programming interface. The IEA 15MW reference turbine was modeled under steady, uniform inflow, with variations in roll amplitude (5° and 10°), roll frequency (0.03Hz and 0.05Hz), and tip-speed ratio (TSR 7 and TSR 9) in a full factorial test set.

Results show that roll motion leads to asymmetric fluctuations across the rotor plane. Normal forces display strong vertical asymmetry, with greater variations in the upper side of the rotor due to its larger distance from the roll center. A kinematic analysis deriving the variation of tangential velocity both azimuthally and over time supported this finding. Tangential forces exhibit lateral variation, with a left-right asymmetry across the rotor plane, though the magnitude of variation is similar on both sides. Both roll amplitude and frequency increase the magnitude of loading fluctuations, while TSR influences their blade spanwise distribution.

Power Spectral Density analysis reveals that roll introduces spectral content at the roll frequency and at 1P sideband frequencies, the latter arising from modulation of the 1P frequency by roll. Higher roll amplitudes, frequencies, and TSRs amplify these load fluctuations. While the prominence of the 1P frequency is negligible at the blade, it is strong in the near wake along with the roll frequency, 3P frequency, and 3P sidebands.

The unsteady blade loads directly shape the near wake. The stable helical vortex system of a bottom-fixed turbine is replaced by an oscillatory corkscrew structure that breaks down earlier and more chaotically. Increased roll amplitude causes stronger lateral oscillations and fragmented vortices, while higher roll frequency produces shorter wavelength perturbations, accelerating vortex pairing. Higher TSR strengthens the initial vortices and accelerates breakdown. Standard deviation of velocity fields in the wake confirms larger fluctuations at the top of the rotor due to its greater distance from the roll axis.

This study establishes a direct link between roll-induced blade loads and near-wake dynamics, providing a foundation for improved FOWT design, control, and wake modeling. Future work should assess far-wake impacts, the persistence of lateral velocities induced by rotational motion, and inertial effects under highly unsteady roll conditions. ... Read more

With the resurgence of interest in propeller-powered aircraft for short-haul and regional missions, understanding the aerodynamic interaction between propellers and nearby surfaces has become increasingly crucial. While propellers offer superior propulsive efficiency and sustainability benefits, their integration introduces complex unsteady flow phenomena that remain insufficiently explored. Additionally, concepts related to regenerative braking and the negative thrust regime aim to harness the full potential of propellers in the pursuit of sustainable aviation. However, the interactions are more complex in the negative thrust regime and need to be accounted for.

This study investigates the unsteady aerodynamic effects of a pusher-propeller operating upstream of a downstream airfoil, with a focus on both positive and negative thrust regimes. A flexible PCB embedded with microphones and pressure sensors was used to capture unsteady surface pressure fluctuations across different operating conditions. The motivation stems from the limited understanding of unsteady surface pressure fluctuations in negative thrust conditions.

The experimental campaign was conducted in two wind tunnel labs: the M-Tunnel and the Small Low-Turbulence Tunnel in the Low-Speed Wind Tunnel Laboratory of Delft University of Technology, Netherlands. The research was carried out in two experimental phases. In the first phase, the device was validated to verify the response of the microphones and the pressure sensors. A known tonal excitation case, along with a case involving an upstream cylinder, was conducted to assess whether the results aligned with theory and expected trends from the literature. These validation experiments confirmed the device’s capability to capture unsteady flow behaviour. However, certain limitations, such as overshoots in measured pressure data and restricted chordwise and spanwise resolution, were encountered. The second phase was conducted in the Small Low-Turbulence Tunnel featuring a fixed airfoil section downstream of a rotating propeller. The device, comprising the microphones and BMP390 pressure sensors, was wrapped around the leading edge of the airfoil to measure pressure fluctuations across the airfoil surface. Initial validation confirmed the reliability of the device in measuring the flow in the propeller slipstream, with good agreement with results from the literature. Comparative analysis was performed across multiple cases, including nacelle-only baselines and propeller-on conditions at two different advance ratios. In the positive thrust regime, the propeller generated a strong tip vortex trace, which significantly influenced the laminar separation bubble and led to elevated pressure fluctuations and peaks at the tonal harmonic of the blade passage frequency. In contrast, the negative thrust regime featured a weaker tip vortex trace and a broadband-dominated spectrum, with reduced suction observed on the upper surface due to lower dynamic pressure in the slipstream.

The study also highlights that the influence of the propeller slipstream extends well beyond its boundary across the span of the airfoil model. Key limitations included discrepancies in the data measured by two rows of microphones due to surface mounting issues.

Overall, the device proved to be a valuable measurement tool for investigating the unsteady surface pressure fluctuations associated with propeller–wing interaction. The insights gained contribute to a better understanding of surface pressure fluctuations on a body immersed in a propeller slipstream, particularly in energy-harvesting operating regimes. Recommendations for future work include improving sensor mounting fidelity, increasing chordwise resolution, and incorporating time-resolved flow visualisation techniques to complement the surface pressure measurements and provide additional insight into the spatial and temporal evolution of the flow field. ... Read more

With increasing computational power, Large-Eddy Simulations (LES) are becoming more prevalent for high-fidelity wind farm analysis. Accurately representing wind turbines in these simulations requires knowledge of blade load distributions, often obtained from Blade Element Momentum (BEM) theory. However, BEM requires numerous geometric and aerodynamic parameters, which are frequently unavailable in industrial settings due to confidentiality. Consequently, industry-scale LES are often limited to uniform Actuator Disk (AD) models, which lack accuracy and cannot capture load distributions.

This thesis implements an Analytical Body Force Model (ABFM) that estimates blade load distributions from LES data using limited turbine information. Both one-way and two-way coupling strategies are explored. In the one-way approach, the ABFM supplements the AD model to compute blade forces without influencing the LES. In the two-way approach, the ABFM forces are actively fed back into the LES in a feedback loop, replacing the uniform AD model and improving simulation fidelity. ... Read more

Aerodynamics significantly influences race car performance, with wheels contributing 35–50% of total drag and affecting underbody downforce. This study investigates the flow around a rotating motorsport slick wheel using CFD to evaluate various turbulence models, validated against wind tunnel measurements. Simulations are conducted in a wind tunnel configuration using RANS, URANS, and DDES models in OpenFOAM. Results show that k-w SST outperforms EB Lag k-e in predicting near-wake features. In addition, DDES offers the highest accuracy overall compared to RANS and URANS, especially in capturing flow separation. Indeed, a trade-off must be made between solution accuracy and computational cost. Three main effects are analyzed in RANS. First, the wind tunnel rig induces wake asymmetry, increasing both lift and drag. Second, increasing Reynolds number delays flow separation from the wheel top, raising lift and reducing drag. Third, yaw angle alters wake symmetry and vortex strength, increasing drag and causing non-monotonic changes in lift. ... Read more

Steady Reynolds-Averaged Navier-Stokes (RANS) simulations utilizing the k-ω SST turbulence model are conducted to investigate the aerodynamic performance of a wing-integrated ram-air duct housing a heat exchanger for propeller-driven aircraft, including its impact on wing-body junction flow. The research is conducted in two stages: first, a 2D aerodynamic analysis employing a Design of Experiment (DoE) methodology to assess the sensitivity of key geometrical parameters–including stagger angle, leading-edge droop, duct gap, and heat exchanger characteristics–on lift, drag, and duct mass flow rate; and second, a 3D investigation of the junction flow behavior in the nacelle/ducted-wing configuration. The heat exchanger pressure drop is modeled as a porous media zone using the Darcy-Forchheimer quadratic drag law. Heat transfer is incorporated through a variable energy source term applied via a userdefined function (UDF) based on the ε-NTU correlation. Findings from the 2D aerodynamic analysis indicate that heat exchanger characteristics, particularly porosity and thickness, have a more pronounced impact on aerodynamic performance than external duct geometry. However, intake stagger angle and leading-edge droop play critical roles in mitigating flow separation and optimizing the wing pressure distribution. In addition, the redistribution of pressure due to flow restriction alters the stagnation point location, inlet-velocity ratio, and static pressure distributions, all of which influence the aerodynamic loading of the ducted wing. The optimal ducted airfoil configuration, featuring a lower-surface outlet aft of the maximum thickness and a thin heat exchanger, minimizes aerodynamic penalties while maximizing duct mass flow rate. However, thermal feasibility assessments reveal that meeting the cooling demands of fuelcell systems necessitates a thicker heat exchanger to accommodate sufficient heat transfer area within the constrained wing volume. This increase in thickness impairs aerodynamic performance through increased pressure drop and resultant drag. Although higher porosity mitigates flow resistance, the required thickness offsets this advantage, reinforcing the inherent trade-off between aero-thermal performance. In 3D, the presence of the heat exchanger inside the duct fundamentally alters the local aerodynamics by modifying boundary layer interactions at the wing-body junction. The flow resistance imposed by the heat exchanger directly affects the strength and topology of secondary flow structures, particularly the horseshoe vortex (HSV), which governs junction flow behavior and whose strength scales with the Reynolds number based on the momentum thickness of the incoming boundary layer. At low porosity levels, the stronger HSV , with an increased vertical extent above the wing, entrains high-momentum freestream flow into the chordwise and spanwise boundary layers, mitigating corner flow separation. Conversely, at high porosity levels, lower flow resistance alters HSV topology, reducing its vertical extent and allowing part of the vortex to enter the duct, inducing a secondary vortex at the lower lip. This weakens the HSV’s ability to stabilize the boundary layer, leading to earlier separation, increased pressure losses, and higher drag. A moderate porosity level provides an optimal balance between HSV strength, vertical positioning, and junction flow stability, reducing corner flow separation and associated pressure losses. Collectively, these findings yield critical insights into integrating ram-air cooling ducts within the wings of propeller-driven aircraft, offering a compelling approach to achieving efficient thermal management systems with minimal aerodynamic penalty. This investigation provides unprecedented detail in visualizing and understanding the intricate coupling between ducted wing aerodynamics and heat exchanger-induced flow interactions, while emphasizing the need for further research to validate and expand upon these findings ... Read more

The transition to renewable energy sources is essential to mitigate climate change, and floating wind turbines (FOWTs) present a promising solution to harness offshore wind resources. Light Detection and Ranging (LiDAR) systems mounted on nacelles provide a cost-effective and efficient means to measure wind fields, critical for turbine performance, control and load simulation. However, FOWT's motion introduces complexities in LiDAR measurements due to velocity and positional changes. This thesis focusses on developing a correction method for LiDAR measurements on FOWTs, addressing the influence of motion on wind velocity, position, and direction. The accuracy and uncertainty of these corrected measurements are quantified.

Simulated six degrees of freedom (6DOF) motion and a power law wind field are inserted in a numerical LiDAR model, in which corrected and uncorrected measurement position, direction and line of sight velocity are constructed. The corrected outputs are validated through reconstructed wind fields and the uncertainty of the correction is quantified. In this study, significant motion-induced bias is identified in the reconstructed wind fields. The dominant motion affecting measurement accuracy was identified as pitch motion, especially when it exhibits a non-zero mean. The relative error of the reconstructed power law wind field parameters is reduced by 3 orders of magnitude. Despite an increase in uncertainties associated with the correction method applied, the correction remains effective in reducing the error in LiDAR measurements induced by FOWT motions. The findings highlight the necessity and feasibility of motion correction for LiDAR measurements, offering substantial improvements in the accuracy and reliability of reconstructed wind fields for floating wind turbine applications. ... Read more

Recent advancements in the UAM (urban air mobility) sector pave the way for making urban eVTOL (electric vertical take-off and landing) flight a reality. Operating in urban environments, noise generated by aerial vehicles plays an imperative role in their certification. Urban flight is expected to subject aerial vehicles to a broad spectrum of large-scale turbulent structures generated by various sources such as flow around high-rise buildings, terrain-level obstacles, incoming ABL (atmospheric boundary layer)
characteristics, local windspeed variations, etc.. Unlike the well-documented effects of aerial vehicles flying under steady freestream or small-scale turbulence impingement, research on their performance when subjected to large-scale turbulence representative of an urban airspace is relatively scanty. This experimental work aims to generate large scale turbulence in the order of the rotor diameter and analyze its influence on propeller loading and noise emission.

Bluff body shedding from a cylinder is utilized to generate rotor scale turbulence in the inflow, with a variation in upstream cylinder placement tested to quantify the influence of it’s proximity to the propeller. The presence of an upstream obstruction results in severe penalties to the aerodynamic performance and thrust generation compared to an isolated propeller. An increase in rotor scale fluctuations in the inflow leads to an increased intensity of low-frequency loading fluctuations. The highly turbulent inflow has a severe impact on the noise emissions, with an overall increase in the broadband content and retention of discrete tones emitted at BPF (blade passing frequency) harmonics across all tested advance ratios; contrasting drastically with the emission of the isolated propeller. Further, with the cylinder obstruction placed closer to the propeller, the effect of haystacking is observed in the noise emissions, which dominates the high-frequency range and results in a spectrum comprising of purely broadband content above the 2nd
harmonic of the BPF. The findings highlight a strong influence of large-scale turbulence on propeller performance and noise emissions, while emphasizing the requirement for further investigation which isolates and analyzes each underlying effect to obtain a complete understanding of the identified influence. ... Read more

This study investigates the aerodynamic behaviour of two-dimensional rigid leading-edge inflatable (LEI) airfoils through experiments in TU Delft’s Low Turbulence Tunnel and computational fluid dynamics (CFD) simulations. Lift and drag coefficients were measured for two steel scale models at Reynolds numbers of 5×10⁵ and 10⁶ and angles of attack between −10° and 25°. Corrections were applied for wall and wake-rake effects, and transition behaviour was analysed using infrared imaging and oil flow visualization. CFD simulations using OpenFOAM with the k–ω and 𝛾 − R̃e𝜃t models were compared to the experimental data. Fully turbulent CFD results showed the best agreement with measurements, while transition models performed inconsistently. Infrared and oil-flow data revealed fixed or moving transition zones, corner eddies, and a laminar separation bubble. This work provides the first dedicated experimental dataset for rigid LEI airfoils and validates CFD toolchains for future aerodynamic analysis of soft-wing kites. ... Read more

Auxiliary Power Units (APUs) are critical for the safe and reliable operation of modern aircraft, providing electrical power and compressed air during ground operations and in-flight emergencies. As APUs operate worldwide, they are exposed to harsh conditions such as sand ingestion leading to compressor deterioration and APU performance degradation. The objective of this thesis, conducted in collaboration with EPCOR, was to investigate whether Computational Fluid Dynamics (CFD) can be used to predict compressor performance degradation and its impact on overall APU performance to improve APU condition monitoring and predictive maintenance strategies.

In this study, the centrifugal compressor of a Pratt & Whitney APS5000, as used in the Boeing 787, was reverse engineered using 3D scans of the impeller and diffuser. These geometries were reconstructed and implemented in a CFD model that was validated against a pass-off test measurement. Through a literature study, it was concluded that the main compressor deterioration effects are increases in impeller tip clearance and impeller and diffuser surface roughness. The impact of these effects on compressor efficiency, pressure ratio, and flow capacity was simulated and incorporated into a Gas turbine Simulation Program (GSP) model to assess the resulting APU performance degradation by evaluating changes in Exhaust Gas Temperature (EGT), fuel flow, and compressor pressure ratio.

The results show that increased surface roughness and tip clearance both lead to reductions in compressor efficiency, pressure ratio and flow capacity, which translate into higher exhaust gas temperatures and increased fuel flow at the APU system level. Plotting the reduction in pressure ratio, increase in EGT and increase in fuel flow as a function of compressor efficiency deterioration and flow capacity deterioration, a compressor deterioration map was made. This map is overlaid with simulated points of varying surface roughness and/or tip clearance serving as a decision tool to aid in root cause determination during APU disassembly.

Although the absolute accuracy of the results is limited by assumptions in geometry reconstruction, turbulence modeling, and validation data, the study provides an indication of the relative reduction in APU system performance and demonstrates a working proof of concept in the form of a deterioration map. Therefore, it is concluded that compressor CFD with gas turbine simulation offers a viable approach to assess compressor deterioration effects and their impact on APU performance, thus enhancing APU condition monitoring and supporting root cause determination in a maintenance environment.
... Read more