Laminar–turbulent transition has long been a central topic in flow-control research due to the significant differences in the aerodynamic characteristics of laminar and turbulent boundary layers. In particular, maintaining laminar flow over a larger portion of a surface reduces skin-friction drag and can improve overall aerodynamic efficiency. Achieving this requires suppressing or delaying the growth of flow instabilities that ultimately trigger transition.

This thesis presents the development of an optimisation framework to identify an optimal passive wall-modulation configuration for attenuating Tollmien–Schlichting wave amplification in a Blasius boundary layer. The framework couples steady-state base-flow computations in OpenFOAM with linear stability analysis using the Harmonic Navier–Stokes Solver (DeHNSSo) within a Bayesian optimisation algorithm. The wall geometry is parameterised using a spline representation, enabling an efficient exploration of the design space.

The framework was demonstrated to produce robust and consistent optimisation results across different optimisation settings and design constraints. The resulting optimal wall configurations successfully reduced the amplitude of Tollmien–Schlichting waves at a monitored downstream location. Furthermore, the underlying physical mechanisms responsible for the observed stabilisation are discussed by examining the influence of the wall modulation on the base flow and the subsequent evolution of disturbance growth.

The developed framework provides a systematic and computationally efficient methodology for the design of optimal passive flow-control strategies and demonstrates their potential for the suppression of Tollmien–Schlichting waves in the Blasius boundary layer. ... Read more

Laminar-flow control offers a route to reducing aerodynamic drag by delaying transition in boundary-layer flows. On swept wings, transition is often initiated by stationary crossflow disturbances whose amplification is set by the laminar base flow. This thesis investigates wall-temperature control to reduce the linear amplification of these disturbances in a compressible swept boundary layer. The BROADCAST framework is adapted to a swept base flow with non-zero spanwise velocity and a prescribed wall-temperature distribution. For each controlled base flow, the dominant resolvent gain is computed and its wall-temperature sensitivity is evaluated using an adjoint approach verified against finite differences. Two optimisation paths are explored: a normalised gradient-descent update based on the sensitivity, and an optimisation constrained by a modelled heat-flux budget, which determines the wall-temperature distribution for different prescribed heat-flux limits. Results show that localised upstream cooling reduces the dominant amplification, while the constrained optimisation achieves suppression at lower modelled thermal effort. ... 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

In high-performance motorsport, thermal management systems are critical for maintaining optimal vehicle operating conditions. Protective grilles installed ahead of radiators and heat exchangers shield these components from track debris but introduce static pressure losses due to aerodynamic blockage. Accurate and cost-effective modeling of grille effects is essential to ensure adequate cooling while maximizing aerodynamic performance.

This research project aimed to identify the main pressure drop drivers and develop an accurate, cost-effective numerical model for simulating arbitrary grille geometries. The impact of grille location and geometrical construction on radiator performance was analyzed to optimize cooling efficiency and aerodynamic performance. The investigation employed Computational Fluid Dynamics (CFD) RANS simulations using Ansys Fluent. Although multiple validation attempts were conducted using a radiator test bench, the experimental results showed unsatisfactory reproducibility and coherence.

The investigation identified the hexagonal mesh pattern with circular wire as the optimal grille configuration, delivering the lowest static pressure losses while maintaining equivalent debris protection. Pressure loss proved slightly more sensitive to wire thickness changes than to opening size variations. In both cases, the optimal configuration exists at the structural and functional limits: wire should be as thin as structurally viable, and openings as large as possible while maintaining protective capability.

Although pressure loss across the grille and radiator was independent of their separation distance, the radiator demonstrated increased aerodynamic efficiency when distance was minimized, making this the preferable setup. The research also explored using protective grilles to deflect and align airflow with the radiator inlet face. However, the complexity of achieving optimized grille design, combined with additional custom manufacturing costs, rendered this concept impractical.

Data from numerical tests on various grille configurations formed the foundation for developing a mathematical model capable of predicting pressure losses for arbitrary grilles at incoming flow speeds ranging from 6 to 22 m/s. This formulation was subsequently used to represent the grille as a porous medium. Numerical validation demonstrated that the modeled grille in series with a radiator exhibited total pressure loss only 0.5% higher than the geometrically modeled configuration, confirming the approach's robustness and accuracy.

This research advanced understanding of the aerodynamic impact of motorsport protective grilles, clarifying how key geometrical parameters and positioning relative to the radiator affect pressure losses. The developed grille modeling provides a robust and accurate method for simulating aerodynamic impact, enabling reduced safety margins in cooling system design. This work paves the way for further iterations and validation tests, ultimately supporting implementation in full-scale car simulations and contributing to more efficient thermal management solutions in high-performance motorsport applications.
... Read more

Turbulent skin friction drag accounts for approximately 50% of the total drag in large aircraft. Helmholtz resonators (HRs) tuned to interact with near-wall turbulent structures have shown promise as a passive flow control technique for reducing this drag. Previous studies have shown that HRs attenuate velocity fluctuations at sub-resonance frequencies while amplifying them at resonance and super-resonance.
This study experimentally investigates whether inclining the HR orifice against the flow can induce a phase shift between the orifice pressure and velocity. If achieved, this could lead to attenuation at the resonance frequency as well. In addition, since ambient noise is significant in practical applications and its effect on resonator-turbulence interaction remains unexplored, the influence of noise on a single HR in grazing flow is also investigated. The impact of noise and orifice orientation on sweep, ejection, and turbulence production is also studied. Experiments were conducted in the Delft University Boundary Layer Facility at Reτ ≈ 2500. Time-resolved particle image velocimetry, hot-wire anemometry, and microphone measurements were employed. Three HR orientations (Vertical, Flow-Opposed, and Flow Aligned) were tested under no-noise conditions, tonal excitation at the resonance frequency of the HR, and white noise at two levels. The hot-wire results show that the Flow-Opposed HR attenuates more at sub-resonance and amplifies less at resonance than the other configurations. This suggests that a phase shift is induced, but it is not sufficient to achieve attenuation at resonance. The effect of tonal noise is strongly orientation-dependent: Vertical HR generally shows attenuation of Reynolds stresses, while Flow-Opposed HR shows amplification. Quadrant analysis reveals that sweep and ejection events follow the same trend across all configurations and frequencies. ... Read more

Acoustic liners have been extensively studied from a purely acoustic perspective. Recently, the interest in characterising their aerodynamic performance has increased, as they are known to provide an increase in drag, compared to their respective smooth counterpart, up to 70%. The goal of this work is to experimentally test various acoustic liner samples to study this drag contribution. The hypothesis is that it is possible to decrease the added drag without having to compromise the acoustic performance.
A variety of orifice cross-section configurations are investigated. Samples are tested in different experimental campaigns, combining the use of two different wind tunnels (the Delft University- Boundary
Layer Facility in Delft and the Flow Duct Facility at NLR, Marknesse) and two different experimental measurement setups, both to investigate the setup effects and to provide a deep understanding of liner performances in different operating conditions. All sample porous plates are 3D-printed by means of Stereolitography (SLA). Direct drag measurements are performed to obtain the aerodynamic response. Insertion Loss measurements are carried out to compare the acoustic attenuation of the different samples.
The direct drag measurements show an increase in added drag for all samples in all tested conditions,
with 0% < ∆CD < 37%. Compared to the conventional cylindrical orifice shape, two configurations show a lower added drag: the tapered and the inverse chamfered. Due to shape simplicity, which makes it easier to manufacture with different techniques, the tapered is taken as reference ’improved’
configuration. The relative drag reduction caused by the tapered configuration ranges from 20% to 40%, with consistent results between different campaigns. The insertion loss measurements, on the other hand, show no significant difference between different samples. All samples have a low acoustic
attenuation in the no-flow case, reaching a maximum of 2.5dB. With increasing M, the attenuation increases in the whole frequency range, probably due to the testing configuration and wind tunnel effects.
The measurements show important novel insights on the drag increase caused by acoustic liners. Data from different campaigns, facilities and measurement systems compare positively and a possible model is proposed, which relates the drag increase to the product of the Open Area Ratio σ and the inner-scaled pore diameter at the top of the liner porous facesheet, d+. It allows to estimate the drag increase,
given the geometric parameters of the liner and the flow conditions (in terms of a viscous parameter, e.g. uτ). This model does not take into account the effect of the material surface roughness, which has
a considerable impact, due to the contribution of the ’pure skin-friction drag’ on the total drag. From the acoustic measurements, no significant difference is found between the acoustic attenuation of all
liner samples, showing that a lower drag increase can be achieved without sacrificing the acoustic performance. All in all, further development in acoustic liners is considered crucial to achieve optimal noise attenuation as well as aerodynamic performance.
... Read more

This thesis addresses thermoacoustic instabilities and flashback in hydrogen combustion. A simplified model of Ansaldo Energia's GT36 reheat combustor was simulated at 20 bar using Large Eddy Simulation (LES), revealing unsteady flame dynamics driven by strong pressure oscillations. To detect flashback precursors, LES-derived time series were sampled at the combustor wall after an analysis identified suitable monitoring locations, representing a step toward practical sensor placement.

Fourteen thermodynamic, velocity, and species mass fraction signals were reduced via autoencoders with 2–4 latent variables; the three-latent representation emerged as optimal, isolating transition sharpness and mid-frequency modes. Clustering of this space with a modularity-based algorithm consistently identified precursors with maximum lead times of ~42 μs and virtually no false positives. In one case, a flashback was successfully predicted and suppressed. Robustness analyses confirmed generalization across locations and noise levels, demonstrating that wall-based latent clustering advances predictive flashback control toward real-world deployment. ... Read more

Flow control plays a vital role in improving efficiency in aerospace, maritime, and energy systems by delaying transition to turbulence and suppressing instabilities such as Tollmien–Schlichting waves. This work uses the Kuramoto–Sivashinsky (KS) equation as a model to study convective instabilities in boundary layers. A finite-difference discretization yields a state-space formulation, enabling systematic application of control strategies. Linear Quadratic Regulator (LQR) and adjoint-based optimization methods are developed to minimize flow perturbations. Results show LQR effectively suppresses disturbances, while adjoint-based control scales efficiently to nonlinear cases, highlighting promising avenues for future turbulence management. ... Read more

Identifying large-scale coherent structures in homogeneous isotropic turbulence is crucial for advancing the understanding of turbulent phenomena, including intermittency and energy transfer. However, the current knowledge and statistical characterization of these structures remain limited due to the absence of efficient and consistent identification techniques. This thesis presents an automated methodology based on the HDBSCAN clustering algorithm to identify large-scale coherent structures.
The method successfully detects coherent large-scale structures, characterized by a quasi-uniform velocity direction. It has been tested across a wide range of Reynolds numbers, providing insights into their spatial organization and interactions. High dissipation regions were observed to occur between neighbouring structures, indicating a direct link between large-scale motions and small-scale intermittency through shearing mechanisms. Additionally, the methodology was extended to time-resolved datasets, enabling temporal tracking and analysis of the evolution of these structures in physical space.
This thesis provides a robust and efficient framework for coherent large-scale structure identification in homogeneous isotropic turbulence and provides new insights into their statistical properties, dynamics, and role in turbulent energy transfer.

... Read more

Hypersonic flows regard a special class of flows where dire conditions for spacecraft may be found. Simulation of these flows requires taking chemistry and thermal nonequilibrium into account as to acquire the correct wall quantities for the vehicles. Flux reconstruction -- a higher order method -- enables fluid simulations with a coarser grid than finite volume, as well as with easier shock discretisation for the practitioner. Computational Object-Oriented Libraries for Fluid Dynamics (COOLFluiD) allows the simulation of hypersonic flows with flux reconstruction, but not yet with thermochemical nonequilibrium due to positivity issues encountered. A survey of the literature for positivity methods in higher order methods, together with a presentation of flux reconstruction, the chemical model used and COOLFluiD were given. An entropy based exponential filter and a Fejér filter were implemented in COOLFluiD based on a pre-existing framework of filter based positivity and tested against the previous positivity method in place. The exponential filter was found to be equivalent for a 2D simulation with the Euler equations to the previous method while the Féjer filter behaved in a more destabilising manner. The entropy based exponential filter was implemented for cases with multiple species and multiple temperatures, with modularity for species sets. The derivation of entropy for this case was given. A new positivity method based on least squares optimisation was developed to target density positivity issues where the filter fails. The case was tested against two cases from literature in P0, ensuring positivity and validated against literature. The cases were found not to be yet fully spatially resolved, as expected in P0 simulations. P1 simulations were tested for the same cases and found to have issues with the time integration method and the artificial viscosity. An inspection of the code base was executed, revealing that the artificial viscosity was not fully implemented for the cases tested and that the current base implementation of backwards Euler cannot be used for these cases, requiring being solved through the already implemented Newton iterations method. The artificial viscosity was fully implemented for the tested cases. The P1 simulations were retested. The artificial viscosity was found to be active and to diffuse the shock wave. No positivity failures of the method were found. A stringent CFL condition was found to exist, and possibly linked to the stiffness of the equations. The exponential filter was found to lead to a growth of the temperatures just upstream of the shock wave, where the filter is active. This behaviour was linked to a possible interaction with the highly nonlinear model or with the artificial viscosity. While positivity was found to work, further issues in testing it are still found and expected to be part of future work. Possible extensions or alterations to the code were found to possibly be required in order to allow further testing. Thus, positivity was found not to be the only issue with running thermochemical nonequilibrium simulations, being only part of the problem, requiring further research in order to fully test it in the context of COOLFluiD. Future work was defined to focus on studying the exponential filter interactions, on enabling future tests, on expanding on these testes based on presented literature, and on expanding on the model used to include more robust but complicated methods. ... Read more

This thesis presents the first systematic experimental study of boattail radius effects on Hammerhead Fairings (HHFs) in transonic flow, using the Coe and Nute Model 11 as a representative platform. Boattail radii were selected based on ratios of radius to boundary-layer thickness, and tested in two campaigns: high-speed schlieren imaging, oil-flow visualisation, and PIV at selected cases. Results show that increased radius generates a third expansion region, often with a lambda-shock, reducing separation length by 10–25% but increasing reattachment-point oscillations. Strong radius effects promoted shock-induced separation and lifted the shear layer at the shock foot. Two dominant unsteady modes were identified: a flapping mode linked to reattachment motion and shock features, and an undulation mode related to momentum injection and ejection from the separation bubble. Findings hint at trade-offs between separation reduction and increased unsteadiness, with implications for launch vehicle stability and structural loading. ... Read more

Floating Offshore Wind Turbines (FOWTs) have low natural frequencies of motion, typically around 0.01−0.001 Hz. If external wind fluctuations match these frequencies, they can excite the motion, leading to increased loads and higher fatigue at specific locations of the wind turbine. Until now, wind turbine design has only considered high-frequency, microscale turbulence, modelled using the Mann (1994) model, as recommended by the IEC. However, this model does not capture the full energy content of the wind fluctuations around the FOWT motions natural frequencies. Syed and Mann (2024) proposed a model for these low-frequency, mesoscale fluctuations. In this thesis, the two turbulence models were implemented, combined, and input in the Siemens Gamesa Renewable Energy (SGRE) in-house servo-hydro-aeroelastic wind turbine solver for a SGRE FOWT loads simulation. The goal was to assess whether adding low-frequency turbulence increases damage on the tower bottom and mooring lines compared to using high-frequency turbulence alone. It was found that indeed for the simplified load simulations done here, the lifetime of a FOWT could be impacted by mesoscale turbulence and thus could be considered in the FOWT design. ... Read more

Computational Fluid Dynamics has been widely used to model flows for engineering applications. The challenge has been to model or resolve turbulent flows accurately and there exist different approaches such as Direct Numerical Simulations (DNS), Large Eddy Simulations (LES) and Reynolds Averaged Navier Stokes (RANS). With limitations on the available computational power and time, RANS is the favorable choice to model such flows despite significant modeling errors.
With recent advancements in machine learning (ML) algorithms and availability of high-fidelity (accurate) data, the turbulence community has been actively working to improve RANS modeling by informing it from the reference data and developing corrections. This is also known as data driven turbulence modeling. There exist several algorithms that have improved RANS modeling but their computational cost and lack of physical interpretability remains an issue to gain an understanding of flow physics and the relation between different flow quantities.

The present study implements two data driven methods to provide corrections ad improve RANS modeling, focusing on Turbulent Kinetic Energy and the Reynolds Stress Tensor. However, the goal is
to also obtain symbolic models and physically interpretable corrections, while ensuring computational cost remains minimum.

A priori analysis, where only the reference data is used to develop corrections, shows significant re- ductions in the computational cost (function calls to the CFD solver) while converging to acceptable
correlations as well. The superior of the two methods is selected for a posteriori study where it is
coupled with a CFD solver. The correction is now formulated by differentiating the CFD code. This
information is given to the ML algorithm which adjusts its parameters accordingly to optimize the objective function and improve turbulence modeling. Results show that, with regularization, a generalized symbolic formula can be obtained which when tested on different geometries improves the prediction of flow quantities and turbulence modeling, as a whole. ... Read more

Accurate modelling of wind flow in complex terrain remains a significant challenge in wind resource assessment. Traditional linear models, such as those used in Wind Atlas Analysis and Application Program (WAsP), often fail to capture non-linear effects like recirculation, separation, and stability-driven phenomena typical of steep or mountainous sites. Computational Fluid Dynamics (CFD) methods based on Reynolds-Averaged Navier–Stokes (RANS) equations offer improved accuracy but must be carefully verified and validated for reliability. This thesis evaluates the predictive accuracy of steady-state RANS simulations using PyWakeEllipSys against field measurements from the Perdigão campaign, characterised by complex double-ridge terrain. Three atmospheric stability regimes (stable, neutral, unstable) were simulated, employing various turbulence closures, including standard k–ε and Monin–Obukhov-based models (k–ε–MO). Grid convergence studies ensured robust simulation accuracy at turbine-relevant heights. Results indicate that unstable conditions are modelled most effectively, particularly in predicting terrain-induced speed-up and turbulence intensity profiles. Stable conditions were reasonably well captured in turbulence intensity and flow patterns but showed consistent underprediction of speedup due to overly persistent recirculation zones. Neutral conditions exhibited inconsistent accuracy across all metrics. Wind direction variability, especially bimodal flow patterns observed in the valley, was not captured by steady-state RANS, highlighting limitations in representing time-dependent, thermally driven flow mechanisms. The outcomes reinforce that steady-state RANS simulations, particularly when stability-adjusted turbulence models are employed, provide strong predictive capabilities for wind resource assessments in complex terrain, although inherent limitations related to transient phenomena must be acknowledged. ... Read more

The present Master's thesis investigated the low-frequency unsteadiness characteristic of highly separated transitional oblique shock wave-boundary layer interactions (hereinafter "OSBLIs"). Such phenomena, encountered notably in engine components operating at transitional Reynolds numbers, are relevant due to their impact on aerodynamic efficiency, structural integrity, and system reliability. The primary research question examined how variations in different parameters such as Mach number, Reynolds number and inviscid pressure jump influence a low-frequency shock oscillation mechanism which was previously identified in literature. To this end, experimental studies were conducted in the TST-27 transonic-supersonic wind tunnel at TU Delft, using high-speed and spark-light Schlieren visualizations capture the relevant flow phenomena. These recordings were processed using digital and spectral analysis.

The findings from this study revealed that the investigated transitional OSBLIs exhibited low-frequency shock oscillations strongly correlated with the periodic formation and disappearance of a Mach stem, which was denoted as the "dual domain" phenomenon. Through carefully chosen variations in Mach number and Reynolds number, it was shown that slight adjustments significantly impacted both the presence of the dual domain and the characteristics of the shock oscillations. Moreover, the Reynolds number regime identified as transitional for the natural flat plate boundary layer in previous research was validated.

Another aspect of this thesis involved the implementation of passive flow control techniques, specifically the introduction of thin two-dimensional steps (in height increments of 60 microns) designed to artificially trip the boundary layer. The experimental results demonstrated that even minimal boundary layer tripping significantly dampened the shock oscillations and modified the interaction dynamics of cases where the oscillation mechanism had previously clearly been identified. The frequency analysis confirmed this, as none of the oscillation peaks which had previously been identified were observed when tripping the boundary layer.

A non-dimensional analysis of the dominant oscillation frequencies indicated a consistent Strouhal number convergence around St = 0.33, particularly in cases where the "dual-domain" behavior and high oscillation amplitudes were observed. An increase in the Reynolds number consistently resulted in reduced laminar separation amplitudes and increased oscillation frequencies, which aligned with the theoretical expectations of accelerated boundary-layer transition dynamics.

In conclusion, this study was successful in identifying the main parameters that cause unsteadiness in transitional OSBLIs. It confirmed the existing transitional Reynolds number ranges which had previously been analyzed in the context of weak OSBLIs and the natural boundary layer of the flat plate which was used, and showed that even simple flow control methods with low 2D step heights can effectively reduce shock oscillations. Additionally, a meaningful non-dimensional scaling was done, which can aid in further research and comparison of the phenomena which were investigated in the present thesis. These findings provide a solid basis for future studies aiming to apply this knowledge to more general cases and improve the prediction and design of aerospace components affected by transitional shock-induced boundary-layer interactions. ... Read more

Accurately predicting flow-induced vibrations (FIV) in high-precision lithography equipment is challenging, as standard Reynolds-Averaged Navier-Stokes (RANS) simulations fail to capture the driving turbulent fluctuations.
This work evaluates the accuracy and uncertainty of a Stochastic Noise Generation and Radiation (SNGR) pipeline that reconstructs time-resolved, divergence-free velocity fluctuation fields from RANS statistics to improve in FIV and aero-acoustic assessments. Therefore, helping to ensure nanometer-scale manufacturing precision. The MATLAB based implementation ingests CGNS/HDF5 solver output, assembles modal Fourier fields from a prescribed energy spectrum, enforces incompressibility, applies anisotropic tensor mapping to match Reynolds stresses, and offers both a time-marching (single time-loop) and an ensemble snapshot mode. Validation is performed against canonical
Direct Numerical Simulation (DNS) reference data for turbulent channel flow (Lee & Moser, up to Reτ ≈ 5200)
and a benchmark backward-facing step, and comparisons are made with RANS (StarCCM+) results. A modular MATLAB pipeline automates diagnostics and produces slice-wise statistics (RMSE, absolute/relative errors), spatial heatmaps, PSDs, and GIF visualizations. Results show that SNGR successfully reproduces spectral content and the
spatial distribution of turbulence, yielding instantaneous fields that correlate with RANS TKE maps. However, the analysis reveals that discrepancies - particularly concentrated in near-wall amplitudes, outer-layer overshoots and the smallest resolved scales - are primarily inherited from biases in the initial RANS model. The report quantifies how these RANS biases, domain and boundary conditions choices and numeric resolution propagate into SNGR
reconstructions, and it recommends practical diagnostics and sensitivity checks for the robust industrial application of the SNGR method. ... Read more

... Read more

Hydrogen represents a promising pathway for decarbonizing heavy-duty transport; however, accurately modeling its injection and combustion behavior in dual-fuel engines remains challenging. This work addresses two critical aspects of hydrogen high-pressure direct injection (HPDI) systems: the characterization of the injection process and the physical mechanisms governing combustion initiation.

A CFD-based methodology was developed to reconstruct injection profiles from apparent heat release rate (aHRR) data and to apply nozzle flow theory for approximating injector behavior, validated through experimental comparison. The findings indicate that classical convergent-nozzle theory fails to capture the observed injection trends. While convergent nozzle theory predicts variable, pressure-dependent mass flow rates, the reconstructed mass flow profiles exhibit nearly constant injection rates when the needle is open, consistent with convergent–divergent nozzle theory predictions. However, the observed variation in maximum injection rates across different cases suggests these deviations may arise not only from geometric constraints but also from aerodynamic phenomena such as boundary layer separation or recirculation within the injector. Such flow features can induce pressure-dependent effects that limit injector performance beyond what nozzle geometry and convergent-nozzle theory alone would predict.

Regarding combustion, the study reveals that hydrogen ignition, triggered by a small diesel pilot, is dominated by localized high-temperature regions produced by the diesel flame. This accelerated autoignition contrasts with alternative hypotheses involving radical transport or direct flame interaction.

Overall, these results advance the understanding of injection and ignition phenomena in hydrogen HPDI engines, providing valuable insights for refining CFD models and supporting the development of efficient hydrogen-powered heavy-duty engines. ... Read more

This master's thesis explores the application of a matrix-free GMRES solver for implicit time integration of unsteady turbulent flows using a high-order DGSEM spatial discretisation. Efficient time-integration methods remain one of the primary challenges towards enabling industrial use of high-order methods for scale-resolving simulations. The focus of this study is on the development of such a prototype implicit solver for the DGSEM solver in the CFD framework TRACE, developed by DLR, which specialises in internal aerodynamics of turbomachinery components. To gain a fundamental understanding of the method and have a lightweight experimentation tool, a one-dimensional DGSEM solver that solves the Burgers equation was first developed from scratch. With this gained knowledge, the implementation of the solver in TRACE could start. The prototype solver developed in TRACE is capable of using both finite-difference and forward automatic differentiation techniques to enable the matrix-free GMRES method. Performance tests show that the automatic differentiation method is approximately twice as slow as the finite-difference differentiation method when comparing them for an equal number of linear iterations. However, the finite-difference method produces inaccurate gradient estimations that could influence the linear solve. Lastly, the solver was applied to perform an LES of a turbulent channel flow. The implicit solver yielded similar results compared to the time-explicit reference simulations at time step sizes greater than the explicit limit. Using the finite-difference differentiation method, the implicit solver was also found to require fewer CPU hours than the explicit reference simulations. ... Read more

Loss of Control In-flight (LOC-I) remains a leading cause of fatal accidents in commercial aviation, with aerodynamic stall identified as a frequent precursor. Effective Upset Prevention and Recovery Training (UPRT) requires representative Flight Simulation Training Devices (FSTDs), which in turn depend on accurate stall models. Current models are typically derived through in-flight system identification, but rely solely on global aircraft states and thus depend on assumptions about the onset and progression of flow separation. This study introduces a methodology to derive quantitative flow-separation maps from in-flight recordings of tufts and flow cones on the PH-LAB research aircraft during stall maneuvers.
High-speed video was analyzed through a dedicated image-processing pipeline to classify tuft states and reconstruct separation patterns. The resulting maps provide direct, spatially resolved evidence of stall onset and progression. This allows for a direct validation of existing Kirchhoff-inspired formulations, while highlighting their shortcomings in capturing spanwise variations and local effects of separation. Furthermore, synchronization with flight-test data allows for correlation of tuft measurements with global aircraft states and supports the development of a logistic model to describe local separation behavior. Beyond validating existing models, the proposed approach establishes a robust experimental framework for integrating flow visualization into stall model identification, with direct implications for the improvement of stall models and, ultimately, flight safety. ... Read more