High-aspect-ratio wings offer a direct method to improved aerodynamic efficiency through reduced induced drag, but their increased span also increases structural loads, aeroelastic sensitivity and creates airport-gate compatibility
challenges. The Flared Folding Wingtip (FFWT) concept addresses these competing requirements by combining an outboard folding panel with a flared hinge axis. When released during a gust encounter, the folding motion can reduce the local wingtip angle of attack and unload the outboard wing, thereby reducing the Wing-Root Bending Moment (WRBM). Previous numerical and experimental studies have demonstrated the potential of the concept, but also show that its performance depends strongly on hinge dynamics, release timing and post-gust oscillatory behaviour. This paper presents the development and application of a non-linear, time-domain aeroelastic framework for analysing FFWT release dynamics. The framework couples a Simscape Multibody representation of a flexible main wing and rigid folding tip to an aerodynamic solver based on an Unsteady Vortex Lattice Method (UVLM). The model is used to assess prescribed release strategies as well as based on hinge moment, under discrete vertical gust excitation. The results show that the FFWT response is governed primarily by release phase: early release reduces the critical WRBM peak, whereas release at the locked peak-load instant consistently increases the critical WRBM response. A subsequent hinge-parameter study shows that low hinge stiffness improves load alleviation but increases demands on the hinge angle, while damping mainly affects post-gust dynamic quality. For the simulated configuration, the best compromise is obtained with a low-to-moderate post-release stiffness, sufficient damping, and a low hinge-moment threshold, retaining most of the peak-load reduction of the most compliant setting while substantially reducing hinge-angle demand. ... Read more

This thesis presents a trim optimization methodology for aircraft featuring distributed electric propulsion (DEP) systems and horizontal thrust units (HTU). The Unifier C7A-HARW, a 19-passenger hybrid-electric commuter aircraft with 12 wing-mounted propellers and a tail-mounted HTU, serves as the reference configuration. Multiple trim solutions with different types of propulsion systems usage and performance indicators, such as maximum range, endurance or lift-to-drag ratio, are explored through optimization, revealing complex relationships between angle of attack, airspeed, flap deflection, ruddervator deflection, required aerodynamic power and electric power consumed in steady level flight. Some unexpected results demonstrate that maintaining a constant low angle of attack and gradually reducing flap deflection as airspeed increases is desirable for lowering aerodynamic power requirements in trim conditions and that the wing tip propeller has an important role under certain circumstances. Empirical correlations between power requirements, angle of attack, airspeed, flap and ruddervator deflections are established, providing insights into the performance characteristics of DEP aircraft configurations and enabling efficient trim performance predictions useful for conceptual design. ... Read more

Blended Wing Body aircraft offer the potential for major improvements in aerodynamic efficiency, fuel burn, and internal volume compared to conventional designs. However, current parametric modelling tools are often configuration-specific, which limits design exploration across different levels of wing–fuselage integration.This thesis presents a novel three-dimensional parametric framework that enables a unified geometric representation of aircraft ranging from conventional layouts to fully blended wing-body configurations. The method introduces a hybridization factor to control the degree of blending, combined with a 3D Class-Shape Transformation (CST) approach to define the outer mold line with a compact set of parameters. The framework is validated against multiple reference aircraft and concepts, and then applied to a hydrogen-powered A320-class BWB design study to evaluate how hybridization and wing parameters influence aerodynamic efficiency, volume, and stability. Results show that the hybridization factor strongly drives performance trends, enabling the identification of promising configurations that balance efficiency and volumetric requirements. ... Read more

Advancements in deep reinforcement learning (RL) open the door to the development of robust flight control systems (FCS) that have the potential to improve both safety and performance during off-nominal flight conditions. Simulation-based work on offline-RL FCS has already demonstrated robustness to adverse weather conditions, mechanical failures, and a wide range of operational conditions. However, it has neglected important dynamical phenomena that limit its applicability to reality. In anticipation of a future flight testing campaign of similar RL-based FCS, this research emulates the transition from simulation to reality by modelling prevalent sensor and actuator dynamics, and introduces a method to incorporate a long short-term memory (LSTM) artificial neural network (ANN) into the policy of a Soft Actor-Critic (SAC) agent. The approach is found to largely diminish the sensitivity of the controller to sensor noise and actuator dynamics, while increasing its robustness to delays in comparison with the ubiquitous feed forward deep neural network (DNN) and a traditional linear controller. ... Read more

The black-box nature of deep generative models like Variational Autoencoders (VAEs) limits their practical application in engineering design, particularly in aerospace, where interpretability is crucial for reliability and safety. This work investigates the use of Symbolic Regression (SR) to improve the interpretability of VAE latent spaces for airfoil shape optimization. Two approaches are developed: a latent analysis of a trained β-VAE and a novel SR-VAE model that integrates SR directly into VAE training.

The first approach shows that SR can approximate the decoder via analytical equations linking latent variables to geometric airfoil features. This even enabled parametric reconstruction independent of the decoder, though accuracy was limited for airfoils with extreme thickness or uncommon trailing edges.

The second approach investigates several SR integration strategies, with a per-batch method followed by retraining with fixed equations achieving the best balance between generalizability and reconstruction accuracy. The parametric equations of this final SR-VAE show an improvement over those from the latent analysis while preserving, and in some aspects even improving, the generative capability of the decoder. The latent space itself showed limited change due to the use of warm starts, suggesting that interpretability through SR is improved primarily at the output level.

The practical applicability of the decoders and equations obtained in this work are tested and compared to CST parameterization, using inverse design tests, as well as constrained and unconstrained optimization cases. The SR-VAE decoder consistently showed the highest reconstruction fidelity for inverse design, but limited use in practical optimizations due to its poor generalizability far from the training set mean. Although SR-based parameterizations show limited reconstruction fidelity in inverse design, they demonstrate competitive performance and the fastest convergence in optimization tasks.

Overall, this work demonstrates that SR can bridge the gap between black-box generative models and interpretable, equation-based design, opening new pathways for explainable AI in engineering contexts and beyond.
... Read more

This work explores a paradigm in which isolated aircraft components are analyzed separately and combined together in a multi-body system representative of the whole geometry. The FLAPERON (FLight mechanics Analysis and PERformance OptimizatioN) toolbox is developed and employed to assess the impact of vortex-filament-based aerodynamic interactions on the simulated aircraft's response when considering aerodynamic data only at the component level. Physical networks created in Simscape are used to simulate symmetrical flight and assess key stability characteristics of a configuration including a main wing, a horizontal stabilizer, and a vertical stabilizer. Vortex filaments are modeled for the main wing and the horizontal stabilizer. The introduced aerodynamic interactions change the simulated response of the aircraft in accordance with predictions. This work suggests that multi-body analysis of flying qualities is possible based on the data prescribed at the component level. ... Read more

A rotary-wing aircraft, or helicopter, is crucial for maritime operations due to its Vertical Take-Off and Landing (VTOL) capabilities, making it essential for small-deck Naval ships. Helicopters serve multiple roles, enhancing the Royal Netherlands Navy’s (RNLN) missions. However, the rise of Unmanned Aerial Vehicles (UAVs) in the past decade has introduced a cheaper, expendable, and more efficient alternative for certain operations, as UAVs require no pilot and can operate continuously. UAVs are ideal for tasks related to improving situational awareness, thereby increasing mission effectiveness, reducing pilot risk, and lowering fuel consumption. The RNLN aims to utilize UAVs alongside crewed helicopters on new ships to maximize operational efficiency. Both helicopters and UAVs follow three mission phases: launch, mission, and recovery, with the recovery phase being the most challenging. During the recovery phase, a crewed helicopter approaches the sailing vessel in a standardized way for which operational limits are established. These limits are essential for rough weather situations and are determined through many flight tests, which is a time-consuming, expensive and hazardous task. In addition, these operational limits only apply to a standardized approach manoeuvre. No approaches for UAVs are prescribed yet, making it difficult to determine such limits, but opens the possibility to standardize newly optimal UAV-ship approaches for which operational limits can be established. Moreover, the ship needs to manoeuvre to obtain optimal wind conditions for an approaching helicopter, which is not desired for UAVs. This means that UAVs should be able to approach and land on the ship in any wind condition. This study investigated the effect of wind on optimized UAV-ship approaches, which can be used to aid and standardize maritime UAV-ship approaches. For this objective, a trajectory optimization framework was created that combined helicopter dynamics and performance, a ship, three wind models of increasing levels of accuracy and a solver to find optimal wind-affected approaches. Trajectories were optimized to be smooth and fast because this is most important for operational purposes. Experiments involved several wind directions, wind speeds and helicopter starting positions, affected by three wind models. This work shows that the created framework functions properly and can be used in a variety of situations. The optimized trajectories were evaluated by the Longest Common SubSequence (LCSS) similarity measure to investigate the effect of both wind direction, speed and model fidelity. With an output between 0 and 1, the LCSS algorithm provided intuitive results and clear trends. A conclusion is that the helicopter wants to exploit the wind as much as possible. It diverts its path to acquire a higher ground speed at the start and a stronger headwind to decelerate at the end of the approach. When introducing a wind gradient, it was observed that the helicopter adjusted its vertical manoeuvre with the same objective. In some wind cases, constraint limits were reached or violated, marking preliminary operational limits. Results were also interpreted with a different perspective to find the best starting positions for the helicopter to approach the vessel in a certain wind condition. These results showed a variety of optimal starting locations, based on the total objective function value, total energy consumption and by distance normalized quantities. Wind model fidelity did not influence this outcome significantly, but it showed that it is important to estimate performance parameters with higher-fidelity models for increased wind speeds. This could also be seen from the downdraft behind the hangar; while the optimized approaches did not deviate geometrically, performance indicators such as the power required were affected significantly. ... Read more

This study explores the integration of machine learning with space mapping techniques to enhance the mapping of optimal control sequences between low- and high-fidelity flight mechanic models. Space mapping is a methodology that simplifies the control optimisation process by approximating a high-fidelity model using less computationally demanding low-fidelity models, which are then iteratively corrected to converge towards high-fidelity outputs. The main research question investigates how the integration of space mapping with sequence-to-sequence neural networks can improve control sequence mapping compared to traditional model predictive control (MPC) methods, particularly in managing the trajectory differences in non-linear flight regimes.

In the pursuit of sustainable aviation, with a sharp focus on reducing emissions through innovative designs and enhanced flight mechanics, the computational cost of high-fidelity models becomes a significant limitation. These models, crucial for capturing complex interactions in advanced aircraft designs, often require simplification to reduce computational demands. This research proposes a novel approach by combining the strengths of machine learning, particularly sequence-to-sequence neural networks like Gated Recurrent Units (GRUs) and transformers, with space mapping techniques to bridge the gap between low- and high-fidelity models effectively.

The study delves into two main machine learning architectures: GRUs and transformers. GRUs excel in managing sequences with fewer changes, maintaining stable predictions with minimal error. Transformers on the other hand are well suited at handling complex sequences with frequent changes, thanks to their ability to process entire sequences simultaneously through self-attention mechanisms. This capability makes transformers particularly suitable for dynamic scenarios where anticipating future states is crucial.

A significant contribution of this study is the implementation of the Prior Knowledge Input-Difference (PKI-D) architecture, which uses the low-fidelity model output as a baseline that the neural network corrects, providing a robust framework for the machine learning models to accurately predict trajectory adjustments. This architecture not only enhances the predictive accuracy but also optimises computational efficiency by reducing the dependency on extensive high-fidelity simulations.

Comparative analyses reveal that MPC methods typically provides superior mapping performance for trajectories requiring no anticipation, while the hybrid machine learning-space mapping approach offers improved performance comparably or better in complex scenarios requiring advanced anticipation. This study highlights the critical role of active learning in adapting the machine learning models to new data dynamically, a feature that proves essential in maintaining accuracy over prolonged operational periods.

In conclusion, this research demonstrates that integrating space mapping with machine learning can significantly enhance the mapping of control sequences in aerospace applications. It provides a starting point for future studies to explore tailor made machine learning solutions using extremely small data sets in situations where data availability is sparse. This research could further open up avenues where the advanced capabilities of machine learning can be applied to problems in aerospace engineering previously inaccessible. ... Read more

Trajectory optimization has proven to be a powerful tool in solving a wide variety of optimal control problems in the aerospace field. However, in many cases, numerical complexities prevent the analysis of optimal trajectories for high-fidelity models, particularly due to the inherent difficulty of transcribing high-order dynamic systems. This research project proposes a methodology incorporating reduced-order modeling that retains the most critical dynamic characteristics from a full-order model while allowing the resulting simplification to be manageable for a trajectory optimization solver. The study applies this methodology to evaluate optimal landing trajectories for the UNIFIER19 C7A, a hybrid-electric aircraft equipped with a distributed electric propulsion system that was previously developed under the UNIFIER19 project. Results show that the reduced-order models generated for the aircraft can be used to generate flyable trajectories, verified by tracking the resulting landing approach paths using the base high-fidelity model. It is envisioned that this methodology will also be applicable to other aircraft models and mission phases. ... Read more

In this paper, the Proximal Policy Optimization (PPO) algorithm is used to perform a constrained wing shape optimization. The PPO algorithm is a Machine Learning (ML) algorithm that improves itself by repeatedly performing the same optimization and learning from its results. The complete adaptation of the PPO framework to the design problem is detailed and evaluated. Not only was the PPO framework able to consistently optimize the wing 4% further than the Particle Swarm Optimization (PSO) algorithm, it was able to do so 35 times faster once the model is fully trained. The PPO framework was able to find more efficient wing shapes than the PSO framework. The trained PPO model was able to optimize the wing of other similar aircraft, even without direct retraining. These results illustrate that PPO could be a promising technique for automated aerospace design problems. Due to the significant training time of the ML approach, the PPO algorithm is not an effective replacement of traditional optimization algorithms for design problems where only a single optimization is required. ... Read more

Growing concerns about the environmental impact of aviation have sparked interest in hydrogen aircraft as a greener alternative. Hydrogen can be used to power existing turbofan engines or electrical motors via a fuel cell, eliminating carbon emissions not only during flight, but also during production, provided renewable energy sources are used. However, adopting hydrogen as fuel introduces technological challenges, particularly with regard to on-board storage. Integral tanks, which are part of the aircraft's main structure, seem promising but existing designs show limitations in their integration with the airframe and insulation capabilities.

To address these issues, this study proposes an integral tank concept featuring a double wall architecture with vacuum insulation. The main advantage of this design is the use of an external stiffened wall that can be directly connected to the remaining airframe. In addition, having stiffeners on the outside ensures the required space for systems routing and addresses concerns with the crash worthiness of the structure. A parametric method, coupled with finite element analysis is developed to size the external load bearing wall, enabling quick analysis and mass estimations of different tank configurations. The method consists of a sizing optimization with the objective of minimizing the structural mass under constraints on the strength, buckling stability and fatigue behaviour.

The feasibility of the concept is then evaluated on an aft tank for a short/medium range aircraft in configurations with and without a forward tank. Preliminary results under this realistic scenario point to fuel containment efficiencies of up to 0.71, which are consistent with existing designs. Moreover, buckling stability is identified as the critical design criterion, highlighting the importance of using a stiffened shell design. These findings show the viability of the proposed concept from a structural standpoint and provide the basis for further research. The optimum solution at an aircraft level can be obtained by integrating the developed framework into a multidisciplinary aircraft design tool. ... Read more

This paper presents a System of Systems Engineering approach to aircraft design. For this purpose, conventional design disciplines are coupled with Agent-Based Modeling and Simulation (ABMS) defining a unique optimization problem. The proposed methodology is applied to design seaplanes for an on-demand transportation system connecting the Greek islands. Within this network, diverse scenarios are analyzed by varying parameters of the model such as fleet size and travel demands at each seaport. The objective is to show the impact of including ABMS in the design workflow on the optimized seaplane design parameters. The optimum designs are evaluated on the basis of a number of classic performance metrics, to assess to what extent they can represent a competitive alternative to existent maritime means of transportation. The results reveal optimal fleet performance for seaplanes characterized by lower cruise speeds and passenger capacities, as compared to those derived from conventional methodologies and to existing designs. ... Read more

Aircraft redesign and flight path optimisation offer promising methods of rethinking how we fly. As the effects of aviation on the planet are becoming better understood, focus has shifted from cost minimisation to climate impact mitigation. In this study, simultaneous aircraft design and trajectory optimisation is used to investigate the trade-off between direct operating costs (DOC) and the global average temperature response (ATR) associated with a typical medium range flight. It is shown for a representative mid-latitude atmosphere and narrowbody aircraft that the ATR can be reduced by 49% with approximately 0.5% increase in operating costs through 2-dimensional flight path optimisation. With simultaneous wing planform optimisation, the DOC-ATR trade-off is even more favourable, leading to a 56% ATR reduction with no increase in operating costs. Results indicate that contrail avoidance is a highly cost-effective method of minimising the climate effects of aviation. ... Read more

Flying characteristics of a multi-engined light general aviation (GA) aircraft during and after an engine failure is often a major safety consideration when both designing and operating the aircraft. In the meantime, the propellers, being large rotating masses, can exert considerable gyroscopic effect on the aircraft during flight, itself contributing to a coupling between the pitch and yaw axis, thus affecting flight dynamics. This study presents an investigation on the impact of propeller gyroscopic effects on the flying motion of a representative twin-engine GA aircraft. This is done using a modular flight mechanics toolbox that performs analyses in both frequency domain and time domain. A steady-state windtunnel aerodynamic and control surface model with empirically estimated unsteady aerodynamic coefficients, along with a propeller governor engine system simulation, complements the gyroscopic inertia model in the simulation setup.

Firstly, a modal analysis showed that all modes aside the spiral mode does not get discernibly affected by the rotating inertia typical to the reference aircraft’s propellers. Then, time-domain simulations of various rapid maneuvers show that gyroscopic effect does cause significant change in the angular response of the coupled axis, e.g. sideslip angle response during a pitch input only maneuver, whilst its impact on long term phugoid motion remained inconclusive due to undesired and uncontrolled roll motion. To compensate for this, maneuvers were performed again with a manually tuned simple wing leveler and results showed that pitch input maneuvers does not show much deviation in phugoid motion, whereas yaw input maneuvers such as sudden left engine failure shows discernible difference in airspeed and altitude responses, though the difference in magnitude is still small. Next, comparisons made with different powertrain responsiveness showed that in a power reducing case such as sudden one engine failure, the effect of the powertrain time delay is independent from the influence of gyroscopic effects, whereas for a power increase case, such as going around, the impact of the two is
simultaneous and intertwined. Finally, a sensitivity study on unsteady aerodynamic coefficients showed that their effects on flying motion are generally independent from the gyroscopic effect. ... Read more

The aerodynamic model of a combat aircraft is essential for its success and competitiveness compared to other combat aircraft. This thesis aims to research the most optimal machine learning model to create an aerodynamic model of a combat aircraft. The very large but still sparse, highly nonlinear dataset forms a challenge for using specific machine learning models. Tree-based models, artificial neural networks (ANNs), and Bayesian neural networks (BNNs) have been identified to be individually capable of modelling the aerodynamics of combat aircraft. For ANNs, additional research was performed into genetic algorithms, as a robust hyperparameter optimisation method. Transfer learning, which reduced the training time by around 14%. Finally, adding gradient information of the training data as an additional input, reduced the mean squared error (MSE) by 7%. Scalable BNNs, which used Bayes-by-backprop, were developed to handle the large dataset. The uncertainties coming from the BNN were overconfident in the results compared to the tolerances of the dataset. The uncertainties showed only a weak correlation with the MSE of a given prediction.
Finally, to leverage each of the model's advantages, a stacked model was created which improves the predictive performance on average by 27% compared to the best base model in terms of the MSE on the test dataset. With the stacked model implemented, each of the aerodynamic coefficients could be predicted with over 97% of the predictions of the test data within the tolerance. This is an average increase of 0.5% for each of the aerodynamic coefficients compared to the best base models. Due to the strict regulations related to this aerodynamic model, the machine learning model that is created cannot replace the aerodynamic model at this time. However, the implementation of this machine learning model allows engineers to design the aerodynamic model faster, and with greater precision.
... Read more

It is the aim of this research to assess the mission performance of a boxwing aircraft by developing a configurationagnostic, multifidelity optimal control toolbox for performance and mission analysis. The boxwing aircraft, sometimes named a PrandtlPlane (PrP), is an unconventional aircraft. An instance with redundant controls is designed within the PARSIFAL project. This specific aircraft is designed for commercial transport in the shortrange segment (a 4000 km design range) and for a high passenger capacity (up to 308 passengers). Because of the beneficial induced drag characteristics inherent to the boxwing configuration, the PrP represents a possible solution towards the sustainable future of aviation. To investigate the potential of the PrP as an alternative to conventional commercial aircraft, the mission performance assessment of the aircraft has been split into two components. The first part of the assessment covers the comparison between the performance of the PARSIFALdesigned PrP and that of a competitor aircraft with a similar design range, the A320, while allowing only nonredundant controls. The second part of the assessment involves the quantification of the PrP’s performance when allowing redundant controls in the form of Direct Lift Control (DLC), enabling the aircraft to increase its net lift without a change in pitching moment. Analyses of the PrP and its competitor aircraft for various ranges have shown that the PrP outperforms its competitor in terms of relative fuel consumption. When flying its minimumfuel mission, the PrP’s competitor consumes less fuel in absolute terms. Nonetheless, because the PrP carries more than twice as many passengers, it consumes up to 14.5 % less fuel per passenger per kilometre. In other respects the PrP’s performance is inferior to that of its competitor. The 5400 km maximum range of the PrP is considerably lower than its competitor’s maximum range of 6200 km. Moreover, at a fueloptimal Mach number of approximately 0.7 the PrP cruises appreciably slower than the cruise Mach number for which it was designed, unlike its competitor. In general, the PrP flies its trajectories much slower than its competitor at an approximately 10 % lower average velocity in the minimumfuel missions. If both time and fuel are considered equally in the cruise altitude optimisation, the design altitude of 11 km is deemed appropriate. If only fuel consumption is considered, the PrP would benefit in fuel economy from lowering the initial cruise altitude at the cost of increased mission time. At an optimal altitude of 9.3 km, the PrP would consume 2.2 % less fuel than at its design altitude of 11 km at the cost of even slower flight. The sensitivities of the PrP’s mission time and fuel performance to changes in its design Zerofuel Mass (ZFM) have been investigated. Keeping the Maximum Takeoff Mass (MTOM) constant while varying the ZFM, design mission simulations were run for the PrP for several objective functions. It was found that when flying for minimum fuel, a 1 % increase in ZFM incurs a fuel consumption penalty of over 1 % through a nearlinear, direct proportionality. Likewise, the mission time varies nearly linearly with the ZFM; a 1 % increase results in an approximate mission time increase of nearly 0.5 %. The incremental aerodynamic lift and drag due to control surface deflections for DLC were modelled using a flatplate approximation. With this approach, the projected missionlevel benefits of using DLC are marginal. On the design mission, the results indicate an increase in fuel economy of 0.6 % on the minimumfuel mission and negligible temporal gains on the minimumtime mission. It is however emphasised that numerical uncertainties due to the discretisation of the problem pollute all obtained solutions to some degree, such that appropriate caution should be exercised when interpreting these results in an absolute sense. In future research, a grid refinement study would be a valuable addition to quantify and bound these uncertainties. It is deemed equally important to look into a more sophisticated way to model the control surface aerodynamics necessary for assessing the benefits of DLC. A broader recommendation pertains to future research on boxwing aircraft aerodynamic design. The current research has indicated that the optimal trajectories for the PrP result in very distinct flight profiles when optimising for different objectives. Therefore, it would be interesting to see how the aerodynamic design could evolve, such that flying for fuel economy wouldn’t require such a compromise in temporal performance and vice versa. ... Read more