Symposium 2021
Highlighting the HPB projects 2021
We are very proud to present the brief abstracts of the research projects that were performed by students during their Honours Program for Bachelor students during the academic years 2019-2021. In the same folder you can also find short videos about those researches. Each of these students selected a topic or supervisor and worked on it during their 2nd and 3rd year of their Bachelor Study. This online “Symposium” is the Covid-proof alternative for a regular “Symposium” at a venue in or near Delft. We hope, many people will visit this site and view their achievements.
Marlou van Nus, Student Counsellor
Jos Sinke, coordinator Honours Programs Aerospace
Research Projects
-
Abstract
Distributed swarm systems pose many control challenges. The decentralised control architecture of these systems means that trajectory planning is done on-board with the information the agents are able to collect on their immediate surroundings only. With current technology the amount of information the agents are able to sense about their environment and the other agents in the neighbourhood is quite limited. In this project, an algorithm developed by PhD Mario Coppola is presented. Following this algorithm, the agents in a swarm shuffle themselves into the desired pattern only requiring their local state as input. The agents keep shuffling until a state in the desired states set is achieved, at which point they remain static. This way, the swarm ends up shuffling itself into the desired pattern. My work in this project consisted of working a bit further on the adaptation of the algorithm to continuous time and space and finally testing it on real drones.
-
Abstract
The conventional fixed wings that are currently used in the aviation industry are designed to maximise the performance of the aircraft for the main phase of the flight, which typically is represented by the cruise. Hence, the wing's configuration is not optimal for other phases, such as take-off, loiter or descent. A morphing wing is a compliant, continuous, jointless structure that can change its shape by obeying an objective function, such as drag minimization or load alleviation. Thus, the morphing technology allows the wing optimization at any point of the mission profile, enabling the maximisation of the L/D ratio for each flight phase, reducing the overall fuel consumption. This entails increments in payload capacities, range or endurance. The aim of this project is to prove experimentally the aerodynamic benefits of a trailing edge morphing wing, by wind tunnel tests conducted on SmartX-Alpha, developed at TU Delft. Furthermore, its drag reduction capabilities are investigated for a small UAV mission profile.
-
Abstract
In order to reduce the greenhouse emissions of the aviation industry in the short term, the development of alternative fuels and improvement in combustion efficiency could be essential. Computed Tomography of Chemilluminescence (CTC) can help in this endeavour by producing visualizations of the topology of the flames through the use of the radiation generated by the excited radicals produced during the combustion process. The goal of this project is to present the design, and test of a CTC set up for the Swan Band radiation. Furthermore, the test also aims to investigate the effect of changing the number of views and the number of voxels in the reconstruction of a V-shape flame.
-
Abstract
Autonomous drones is an emerging topic in research and technology. The field of autonomous drone racing has the recent years been heavily incentivised with price money up to one million USD. During fast flight, commands are issued quickly, and it would be beneficial to know where you will end up before you are there. How can one make a drone predict the consequence of current motor commands? The study investigates whether deep learning is a feasible method for drones to predict future states. Training on flight data, different models were tested in order to find a suitable combination of network structure, inputs and outputs. The results can further be used in trajectory planning and optimisation.
-
Abstract
This paper presents an implementation of a novel mesh deformation technique using radial basis function interpolation (RBFI) with sliding boundary nodes into a C++ framework. For fluid-structure interaction problems, the deformation of the structure requires an adaptation of the fluid mesh to account for the deformation of the fluid-structure interface. The resulting deformed fluid meshes are susceptible to low-quality when subjected to large deformations in regions of small wall clearances, resulting in a reduced accuracy of the solution or even divergence of the solver. Previous work by M. T. Mathew found that allowing the boundary nodes to slide reduced the skewness of the cells and maintained cell orthogonality. The increased robustness of the sliding boundary RBFI method comes at the expense of increased computational cost relative to the fixed boundary node method. For this reason, a greedy algorithm was used to decrease the computational runtime. The implementation into a C++ framework will allow for future integration of the RBFI mesh deformation with sliding boundary nodes into current open source CFD suites such as SU2.
-
Abstract
Self-healing polymers have received increasing attention over the past decades, driven by their many engineering applications and concerns related to plastic overconsumption. The intrinsic self-healing processes leading to chemical barrier restoration in phase-separated polycyclooctene and polyurethane blends are reasonably well understood. However, little is known about the impact of the hard, disperse micro domains on the self-healing capability of these blends. In particular, there is a lack of proven in-situ mechanical characterisation techniques to tackle this knowledge gap in the first place. This research aims to investigate the potential of micro- and Nano-indentation for resolving the mechanical property gradients in phase-separated polymers. In order to locate the µmsized domains across the sample surface, µ-ATR-FTIR microscopy is employed. Nanoindentation is shown to achieve a sufficient resolution to measure the differences in elastic modulus and hardness. Interestingly, the phases are found to be highly sensitive to such variables as time, temperature and pressure.
-
Abstract
Everyone often wonders whether there is life outside of the Earth. We then look around us and may ask whether life is present on Venus or Mars. However, a better question might be whether there has ever been life on these planets. Several studies have indicated that Venus, Earth’s sister planet, was very much Earth-like in the past. Therefore, it is not unreasonable to believe that life might have evolved on the planet, during that period. The aim of this research is to estimate the depth at which potential traces of this life might have survived the many years of constant radiation exposure. In this work I have applied Monte-Carlo technique, which is able to simulate the complex particle transport through matter and estimate radiation doses in different scenarios of radiation environment and shielding conditions. By analysing the results, it can be found whether and how deep future missions should look for biological samples. Hopefully, this could be confirmed by the 2 new missions DAVINCI+ and VERITAS, just announced by NASA, which will investigate this fascinating topic and provide valuable measurements.
-
Abstract
One of the most important today's world challenges is reducing people's impact on the environment. Especially this process can be seen in the transportation area, where electrically powered vehicles are on the rise and design ideas for electric aircraft are arising. One of which is having closely coupled propellers distributed along the leading edge of the wing to augment the lift performance and allow a more efficient flight or much better performance at take-off and landing. This propulsion system is currently being heavily researched, however, no research focused on its performance at varying angles of attack and stall angle was found. Thus, the aim of the research, presented in this presentation, is to analyse the performance of closely coupled propellers at varying angles of attack compared to a clean wing and a wing with one propeller operating.
-
Abstract
The Lunar Meteoroid Impact Observer (LUMIO) is a Lunar CubeSat mission, which main goal will be to study the flux of meteoroid impacts on the Moon's far side. These are fragments of Near Earth Objects (NEOs) and have sizes ranging from micrometres to meters, presenting a hazard to satellite missions. This project is focused on LUMIO's propulsion system. When analysing the requirements for this subsystem, it is noticed that the weight, the volume, and the thrust level have low values as a requisite. This is why micro-propulsion is needed for this mission. However, as downscaling a propulsion system is a difficult task, many types are still under development. From the currently available groups, only two of them provide satisfactory levels of thrust for the mission: mono- and bi-propellant systems. Hence, several off-the-shelf propulsion systems were selected from these two categories, and compared through a trade-off. This resulted in the system MPS130 being the selected one. However, it needs to be noted that some characteristics of this commercially available system were modified, such as the propellant mass or the volume, in order to achieve the desired performance for the mission. Further steps of this project will be to perform modifications on the system itself, such as the specific impulse or the type of propellant used, and analyse in a qualitative manner the impact these would have on the chosen subsystem.
-
Abstract
In space propulsion, ongoing research aims to find a green alternative for hydrazine, because of its expensive handling procedures and toxicity. Hydrogen peroxide (HP) is one promising option, due to its low toxicity and competitive performance. Its performance as a monopropellant increases by adding alcohol as a blending fuel. The stability and ignition of HP blended with ethanol, propanol and butanol were investigated. For the O/F ratios of 60/40, 70/30 and 80/20, the blended monopropellant samples were stable. However, reactivity was generally low, and further investigation is required. In the four sample ignitions of the heating plate drop test, it was found that combustion energy increases with an increasing carbon chain. Furthermore, it was found that ignition delay increases with the carbon chain and decreases with a rise in heating plate temperature. The highest flame temperature was 1328°C and the shortest ignition delay was 35 ms.
HPB presentation Pepijn Deroo
-
Abstract
Numerical simulation of complex natural phenomena is nowadays an essential tool across science and engineering research. However, high complexity can make it more difficult to judge whether the numerical solution is physical or not. This issue is addressed by mimetic discretisation, which uses insights about the geometric structure of physical laws to conserve physically meaningful properties at any resolution. In this project, we derive a mimetic boundary element method for potential flow around aerofoils based on the mimetic spectral element method. The derivation of this boundary element method is based on finding a linear combination of the harmonic solution with a particular solution such that the Kutta condition is satisfied. We investigate the convergence behaviour of the method for both the potential-flux formulation and the stream function-circulation formulation of the Laplace equation. Using mimetic spectral basis functions, we achieve exponential convergence and an exact representation of discrete divergence and curl operators.
-
Abstract
The deconsolidation of fibre reinforced thermoplastic prepreg tapes during rapid heating in the LAFP process is investigated. Prepreg tapes were hit by a laser during short periods of time, using variable nip point temperatures, heating times and heated spot lengths, simulating the rapid heating which occurs during LAFP. This led to the deconsolidation of the tapes, and hereafter microscopy images of the tapes' cross section were taken. A program was created which automatically quantifies the degree of deconsolidation in the cross section of the deconsolidated tape. Using the results from the image analysis program, it has been demonstrated how the degree of deconsolidation can be quantified. Furthermore, the results are used to explain how deconsolidation evolves during the heating phase (from ambient temperature to a 400oC nip point temperature), and to explain how heated spot length and heating time affect the deconsolidation state of the tapes.
Quantification of the Degree of Deconsolidation in LAFP - TU Delft Honours Research
-
Abstract
This paper presents an alternative weight saving method for the main landing gear (MLG) structure by reducing the Factor of Safety (FoS). Motivation to decrease the FoS is a common discussion. Firstly, there is inconsistency among engineers regarding its purpose, and its origins are unclear. Furthermore, based on progress in manufacturing, load determination and design it could be argued that the current 1.5 FoS is inappropriate for today's time. Lastly, reducing the FoS does not hamper safety, as the FoS is not the critical limitation of aircraft safety. Therefore, this validated the proposition of finding a new FoS using the probabilistic approach, which relates the FoS to the probability of failure by finding the probability that the applied load factor exceeds the ultimate load factor. A new FoS then corresponds to the FoS that coincides with the threshold probability of failure obtained from CS-25, in this case 3.86 × 10−5 per 60,000 landings, or to the FoS at which fatigue becomes more critical than static loading, whichever is highest. The model resulted in a considerable reduction in mass of the MLG, roughly 23% of the MLG structural mass. Nonetheless, although motivation to reduce the FoS exists and reconsidering it does result in a mass reduction, further investigation of the threshold probability of failure and the method's applicability to other aircraft subsystems is required to better access its validity.
TU Delft HP Research
-
Abstract
Current technologies for indoor drone positioning rely on devices, called anchors, installed throughout a building to guide drones around. My research project proposes a new method which may help this existing system. The idea is to use the communication between drones as a supplement to these anchors. In this scenario, the drones would act as flying anchors to each other, providing each drone with more reference points in order to improve their position estimation. The drones used for this research are CrazyFlies, which communicate to their anchors using Ultra-Wideband anchors. These anchors can provide high accuracy signals to the CrazyFlies, which then calculate their position using on-board technology. To research the potential improvement in positioning performance, I will be simulating their behaviour in an environment with drone-to-drone communication, and in one without.
-
Abstract
Since its first mention in 2015 in the Metropolis paper by E. Sunil et al., the layered urban airspace concept has received significant praise. Nevertheless, it is still far from optimal for all scenarios, due to a number of assumptions. In my research, I have tried to challenge these with an adaptive layered airspace, which assigns aircraft (or drones) to layers based on their heading, while trying to improve safety. This is done with the use of two supervised learning neural networks. In order to simulate the aircraft, BlueSky is used (a TU Delft open-source air traffic simulator, github.com/TUDelft-CNS-ATM/bluesky). The results are impressive: the neural network layer assignment appears to be even safer than state-of-the-art conflict resolution methods used to divert aircraft when they are on course for a crash.
Adaptive Layered Airspace - TU Delft Honours Programme Symposium
-
Abstract
Numerical simulation of complex natural phenomena is nowadays an essential tool across science and engineering research. However, high complexity can make it more difficult to judge whether the numerical solution is physical or not. This issue is addressed by mimetic discretisation, which uses insights about the geometric structure of physical laws to conserve physically meaningful properties at any resolution. In this project, we derive a mimetic boundary element method for potential flow around aerofoils based on the mimetic spectral element method. The derivation of this boundary element method is based on finding a linear combination of the harmonic solution with a particular solution such that the Kutta condition is satisfied. We investigate the convergence behaviour of the method for both the potential-flux formulation and the stream function-circulation formulation of the Laplace equation. Using mimetic spectral basis functions, we achieve exponential convergence and an exact representation of discrete divergence and curl operators.
Hybrid Mimetic Spectral Boundary Element Method Applied to Potential Flow Around Aerofoils