Injection molding is a widely used process for manufacturing a large range of polymer parts, including fiber-reinforced polymer composites. The complex nature of this process, combined with the non-Newtonian behavior of molten polymers, implies significant challenges in terms of control, prediction, and optimization of the processing parameters that influence manufacturing quality. While various computational and numerical models have been proposed to simulate the injection molding process and ensure the desired part quality (see, for instance, industrial software such as Moldex3D or Moldflow, they bring their own challenges, including high computational costs, the need for solver tuning, frequent parametric studies, and the economic burden associated with licensing.

In this study, machine learning surrogates are explored as an alternative to conventional modeling software. They are based on Graph Neural Networks (GNNs) that have recently demonstrated remarkable potential for simulating physical systems. In a graph, physical systems are encoded as nodes end edges, capturing their organization. Then, state-of-the-art techniques such as the message passing framework are tested for their potential to learn physics. In the long term, this project aims to establish a novel graph-based framework capable of efficiently simulating complex physics across various material models and processing conditions.

A numerical model is developed for 2D mold filling and quantitively validated against a benchmark. The prediction results demonstrate that the learned simulator is able to capture the physics of melt front behavior. Further, the prediction runtime is significantly faster than the ground truth, completing in under 1 second compared to \textasciitilde hours, representing a substantial improvement in efficiency. However, the accuracy of the predictions can vary, with errors accumulating especially during testing. While the scalar loss metric provides an estimate of the model's performance, it lacks expressiveness and can be unreliable on its own.

The increasing demand for sustainable energy, results in more wind turbines being built offshore. The blades of wind turbines consist of composites which makes them difficult to design. Because composites consist of a micro- and macro-structure of which their interplay determines the global behavior under loads. Therefore, traditional methods of analysis are often infeasible: full-scale experiments take too long and are too costly, small-scale experiments can not capture the interaction between the micro- and macro-structure, and analytical theories are often only tractable for simple cases.

$FE^2$ is one such method that can analyze the behavior of multi-scale materials, such as composites. It consists of a macro finite element model where the constitutive behavior of each integration point is obtained by homogenizing representative volume elements (RVE) that represent the microstructure. Although this method is capable of accurately predicting composites, the computational cost is high, as for each integration point another finite element model must be computed.

This process can be sped up by replacing the RVE with a surrogate that is more efficient to compute. Recently, Gaussian Process Regression (GPR), a probabilistic machine learning model, is used as a surrogate for the RVEs. It uses prior knowledge and observations of the RVE to make its predictions. It is based on Gaussian Processes, meaning that the predictions exhibit a multivariate Gaussian distribution which provides an uncertainty measure besides its prediction. This is useful in determining where new observations must be collected from the RVE.

To fully capture the RVE a lot of observations are needed from it. This is still a computational bottleneck for the surrogate as they are expensive to compute. The GPR model can be enhanced with observations from a low-fidelity model, for example, linear elastic, which is called GPR with multi-fidelity. The low-fidelity observations, inexpensive and inaccurate, enhance the prediction of the high-fidelity model, which uses high-fidelity observations that are expensive but accurate. This is shown to decrease the number of observations needed for the high-fidelity model. Thus, fewer observations need to be collected from the RVE. However, the correlation between the low- and high-fidelity is often assumed to be constant in these models. This presents a problem as the correlation in the case of surrogate modeling is often non-linear. The literature investigates several extensions of the GPR with multi-fidelity that assume a non-linear correlation, but these models become more complex and lose their simplicity as their predictions are not Gaussian anymore.

Instead, this thesis keeps the constant correlation assumption but improves the correlation inference by splitting the model. Splitting means that multiple independent GPR with multi-fidelity models are used in different regions. As splitting results in discontinuous predictions, the thesis also investigates two stitching methods to remove these discontinuities. The first, the constrained boundary (CB) method, constrains the predictive mean and predictive variance of to neighboring models to be equal at their respective boundaries. These constraints to the optimization procedure of the hyperparameters of the local GPR with multi-fidelity models. The second stitching method, the random variable mixture (RVM) uses a weighing procedure based on a mixture of experts approach. However, this method mixes the random variables instead of the probability density distributions. This creates a much simpler and analytically tractable method. For both of these stitching methods, only the high-fidelity uses the stitching procedure while the low-fidelity is split...

Mechanical metamaterials are materials that appear to have unusual elastic mechanical properties, with the most prominent being a negative Poisson’s ratio. The key to their unique properties lies in the microarchitecture of the material rather than the material itself. For advanced applications in the realm of prosthetics, such as form matching, aperiodic metamaterials have been recommended above repetitive ones. In a restricted lattice, metamaterials with disordered microarchitecture have displayed a wide variety of anisotropic conventional and auxetic behaviour. Because of the non-unique and non-intuitive link between the random network (RN) structures and their mechanical responses, their inverse design, which tries to extract the ideal structure according to specified requirements, remains a complex process. The current study looks at the forward prediction of mechanical response and the inverse design of 2D random network metamaterial unit cells for use in aperiodic combinatorial designs. The associated mechanical properties of stiffness and Poisson's ratio in two directions were numerically simulated for the unit cells. The configuration of random network designs that provide the least complicated unit cells with the greatest range of mechanical properties were selected. A dataset of such unit cells and their mechanical responses was then created for training machine learning models. A forward predictor of their mechanical response through a regression model was trained, to accelerate the production of the datasets and assist in the inverse design modelling. It was suggested to model the inverse design issue in a probabilistically generative manner, allowing for an interpretable investigation of the complicated structure–performance connection. A variational autoencoder (VAE) was trained to map complex microstructures into a low-dimensional, continuous, and organized latent space based on their mechanical response. A systematic approach to train machine learning models was investigated based on hyperparameter optimization methods, with cross-validation and a search space that expands over the model hyperparameters and data analysis alternatives. The trained forward predictor and generative model were evaluated, and the reported R^2 values were 0.99 and 0.96, respectively. Compared to numerical simulations, the estimation of mechanical properties through the regressor resulted in a hundredfold acceleration of the process. The trained deep generative model can be an effective tool for expediting the creation of RN unit cells. On-demand inverse creation of metamaterials for both observable and intentionally designed mechanical responses is achievable. To illustrate a practical use of such unit cells, a technique for aperiodic combinatorial design of unit cells for 2D form matching on two directions and all four sides was also provided. The mechanical properties of the unit cells, as well as the combinatorial design capabilities, were evaluated using experimental tests of 3D printed specimens. The experimental results of the unit cells were in good agreement with the numerical simulations. Similar forms to those programmed were achieved by tensioning the combinatorial patterns, albeit with a less degree of distortion. Overall, the findings show the RN unit cells' potential for combinatorial design.

High-dimensional optimization problems with expensive and non-convex cost functions pose a significant challenge, as the non-convexity limits the viability of local optimization, where the results are sensitive to initial guesses and often only represent local minima. But as the number of expensive cost function evaluations required for a full exploration of the search space grows exponentially with the increasing number of dimensions, the use of standard global optimization algorithms is also not practical. To overcome this obstacle and to lower the dimensionality of the problem, the use of an autoencoder for model order reduction is proposed. In the resulting lower dimensional space, standard global optimization methods can then be utilized, as fewer cost functions evaluations are necessary. For problems with comparatively more expensive cost functions, this optimization includes the employment of a surrogate model, which reduces the necessary number of these computationally expensive evaluations further. This proposed method is then tested firstly on a number of benchmark functions, where it shows the ability to find global optima under certain conditions. Secondly, the proposed method is used to solve a compliance minimization problem, where it shows the ability to improve upon a large number of designs generated by local optimization.

Previous research has shown automated robotic mechanism design to be both deceptive (prone to local minima) and rife with linkage problems (having highly interdependent parameters). This results in a barrier to optimization that is unable to be breached by simply applying more iterations and computational power. The research also indicates that a graph structure model of the robot in combination with an evolutionary algorithm yields useful robotic mechanisms for a limited set of simple problems. This thesis expands on this pre-existing representation by introducing an indirect model that can be used to include both controllers, motors and other new elements in the representation. Besides this extension of the mechanism model, a framework for the automated design optimization task itself is introduced. This thesis shows an equivalence between an operator based representation of the mechanisms and the graph based representation. These operators represent modifications on the mechanism structure and/or parameters. By recognizing the operators as paths in this model a graph of the search space itself can be constructed. In this graph the vertices are mechanisms and the edges are operators. Using the the operator-mechanism equivalence it is shown that designing an optimization algorithm is equivalent to (1) choosing how vertices in the space are grouped together. (2) choosing how the vertices of this search space are connected beforehand by either implicitly or explicitly picking operators and projecting onto their corresponding domain. (3) picking which of the connected paths to traverse based on accumulated information at runtime. This represents a framework that allows the accumulation of knowledge about optimization algorithms acting within it by defining a set of meta-heuristics. With these it is possible to make informed choices to build better optimization algorithms. To show the effectiveness of the framework a novel quality diversity algorithm is developed, Reduced Operator-Space Evolution (ROSE), which uses the insights mentioned above to generate a large diversity of well performing mechanisms simultaneously on a representative pick-and-place task. This confirms the theoretical results about the effect of the operator-mechanism equivalence and locality properties. A new step forward to breaching the barrier to optimization. Alongside this thesis a performant simulation and analysis Python library for mechanisms was developed called PyMechs. The library visualizes the framework and handles mechanism simulation and evaluation, as well as implements ROSE. It is available at https://github.com/kooswestra/pymechs.

Nanomechanical resonators with low dissipation rates are ideal tools in fundamental science applications. They have been used in the field of cavity optomechanics for example in ground-state cooling, and in sensing applications, such as atomic resolution mass sensors. Their great sensitivity is due to their high Q-factor, which is a metric that shows how fast a system loses its energy. In ultra-high Q nanomechanical resonators, energy loss is limited to intrinsic and radiation losses, the latter is due to energy dissipation from the resonator into the substrate. Experiments have shown that the Q-factor of resonators with thin substrates are limited by radiation loss. However, the precise role of the substrate remains a topic that has not received much attention, but has significant implications for how we design nanomechanical microchips. Here we show that the resonator mode can couple to nearby substrate modes, which reduces the Q-factor. We found that the strength of this mode-coupling depends on the mode-shape of the substrate, with stronger coupling at anti-nodes of the mode-shape and hardly any coupling at the nodes. Furthermore, we show that clamping down the substrate with double-sided tape reduces the Q-factor of the resonators, this is explained by a reduction in Q of substrate modes due to the tape. Lastly, we found that in thin substrates, which have a higher density of modes, the Q-factor can be limited due to mode-coupling with the substrate. Our results demonstrate that the substrate choice, as it can strongly affect the Q-factor of resonators, should become an integral part of the resonator design phase. These results can likely be used by all types of nanomechanical resonators limited by radiation loss. We can use this knowledge to design chips with resonators that have an even higher Q-factor.

Carbon fibre reinforced polymer (CFRP) composites are well-known for their specific strength and stiffness properties. For the past two decades, various industries have increasingly used these materials to improve performance and reduce emissions. This increased use has caused a problematic growth in both production and end-of-life CFRP waste. Recent developments in recycling processes allow recycled carbon fibres (rCF) to have undamaged and clean surfaces with intact sizing, retaining the mechanical properties of non-recycled virgin carbon fibres (vCF) at a considerable cost reduction.
Unfortunately, CF recyclate from these recycling procedures is highly diffuse making them useless for structural application. For making rCF material viable to be used in structural applications, further processing of the recyclate is necessary. To manufacture high-performance rCF materials which feature homogenously distributed, orientated and aligned fibres, considerable effort has to be spent in controlling the microstructure during processing.
The presented thesis has focused on introducing an initial characterisation method that allows for quantitative analysis and comparison between rCF microstructures. Characterisation is performed by analysing elliptical CF footprints in perpendicular cross-section micrographs. Fibres in the micrograph are identified and segmented through a semi-supervised machine learning tool. The elliptical fibre shape is measured by the method of ellipses (MoE) in FIJI ImageJ. This MoE extracts the position and elliptical shape of every fibre footprint in the micrograph. Analysis of the microstructure is done in python to quantify the microstructural morphology by fibre diameter, fibre orientation distribution (FOD), global-local fibre volumes (Vf), fibre density distribution (FDD) and Fibre-to-Fibre (FtF) spacing.
It was found that the method is accurate and reliable in identifying and measuring fibres present in the cross-section micrograph. Similarly, calculated values for Vf, FDD and FtF spacing can be used for characterising the quality of studied materials.
Unfortunately, despite being frequently used in literature to characterise highly aligned and orientated CF materials, all fibre orientation measurements were flawed by the inherent susceptibility of the MoE to calculate fibre orientation with large biases. This orientation bias was caused by the error sensitivity of the arccosines function used for calculating the out-of-plane (OoP) angle. Minute elliptical shape differences in near perpendicular fibres with almost circular footprints resulted in large non-linear errors.
Recommended is to study FODs in samples that are sectioned at a 45 or 60° angle, resulting in large elliptical fibre footprints which are less error sensitive when analysing with the MoE. The induced section angle can easily be transformed back to yield an unbiased FOD measurement.

This research considers the offline training stage of the Reduced Order Models (ROM), that has been getting attention recently on the endeavor to come up with efficient solutions for the highly complex numerical models. In this work, a simply supported beam problem has been considered, for which a reduced basis creation has been investigated. Reduced basis creation is in utmost importance for the accuracy and reliability of the ROM. Main focus is on the efficient parameter sampling strategies to enrich the reduced basis, which brings forth computational burden. To decrease this burden, a statistical tool Gaussian Processes Regression (GPR) based Bayesian Optimization (BO) is utilized. These tools are used to create a surrogate function of error indicator that is used to select additional training points for ROM. Results of this work show that randomness in the proposed procedure influences parameter sampling but does not have an impact on the overall accuracy. Finally, this work suggests further work on creation of a stopping criteria and finding a method of storing previous information and combining it with current information regarding training points without losing information. With the help of proposed further research topics, this work intends to be used as a foundation for efficient reduced basis construction.

The advent of machine learning and the availability of big data brought a novel approach for researchers to discover fundamental laws of motion. Computers allow to quickly find underlying physical laws from experimental data, without having in-depth knowledge of the system. Applications are widespread among numerous fields such as physics, chemistry, engineering, biology, climate science and finance. However, the problem is that often experimental data are polluted with noise. This causes symbolic regressors to capture the noise but not the underlying physics. In this work, we leverage the noise reduction properties of non-parametric machine learning, to improve symbolic regression methods. This combination allows for numerically robust differentiation and significantly increases the noise tolerance of common symbolic regressors. A new strategy is exemplified by combining Gaussian processes and the symbolic regression toolbox SINDy for finding governing equations from data. The method balances the quality of the fit and parsimony to avoid overfitting. The method is tested on well-known dynamical systems with varying properties. These will include the Duffing oscillator, as an example of a forced system, the Van der Pol oscillator, as an example of an autonomous system and the Lorenz attractor as an example of a chaotic system.

The society of Delft Aerospace Rocket Engineering (DARE) aims at launching the Stratos IV sounding rocket to an apogee of 100km; thus cementing their place as one of the few student teams that have had the unique opportunity to design and build a rocket that is capable of reaching the frontiers of space. In doing so, DARE has entered flight supersonic regimes during atmospheric entry. With the development of cryogenic engines in the pipeline, future DARE missions look at reaching the hypersonic velocity regimes resulting in the current cork and carbon fibre composite heat shield design becoming redundant. Thus, hypersonic flight brings with it the challenge of designing new a heat shield; capable of withstanding the elevated temperatures.

Alginate and montmorillonite nanoclay based biopolymer nanocomposites have become a well-researched topic in recent years. This stems from the ability of the nanoclay fillers to align themselves within the suspension to mimic natural materials such as nacre (mother of pearl). The alignment of the nanoclay not only reinforces the biopolymer to provide a high storage modulus but also provides a tortuous path for transport phenomenon such as diffusion to take place; thereby making them flame retardant.

In their native state, the nanocomposites are film-like and offer limited viability. But, simple production practices may be borrowed from culinary disciplines to transform the nanocomposites into foams with different relative densities. The introduction of the gaseous voids should further heighten the thermal characteristics; allowing them to function as heat shields. However, in transforming the nanocomposite to a foam, little is understood about the impact it will have on the mechanics of the native nanocomposite. As multiple analytical and numerical models have been provided over the years to estimate the mechanical properties of foams, a careful assessment of the various literature sources becomes necessary. Once a suitable model is identified, a detailed experimental campaign can be carried out to characterise the biopolymer nanocomposite foams mechanically.

Running robots at insect scale are an upcoming research field, because of their numerous potential applications like exploration of hazardous environments such as collapsed buildings, natural disaster sites and debris. To date no report exists on such a robot that either exploits its resonance, or that is monolithic. These properties, however, could improve the performance of these robotic insects in terms of efficiency, actuation complexity, manufacturing tolerance, suitability for rapid prototyping, large number reproduction and miniaturization. This article presents the monolithic design of FARbot; a Frequency Actuated Resonant robot that uses resonance to increase its stride length and consists of one single piece. To achieve the design of FARbot a novel design methodology is presented and utilized to systematically obtain the necessary compliant mechanism that resonates in a desired motion and at a desired frequency. The final design has been manufactured monolithically from the material HTM140-V2 using digital light processing 3D-printing technology. In terms of production time FARbot outperforms all other robots in this research field. An earlier produced, non-monolithic prototype shows resonance at the desired eigenfrequency at which its stride length is amplified by a maximum factor of 22.74 for constant energy demand. This proves the benefit of using resonance and also validates the proposed design methodology.

Additive manufacturing (AM) is currently making a transition from mainly being used for prototyping to an alternative for conventional production methods. AM provides more freedom in design geometry combined with the possibility to produce (complex) products in low quantities. As a result, the need arises for printed parts with consistent mechanical properties that rival those of conventionally produced parts. Currently, the layer-by-layer process used by AM introduces weaknesses at the interlayer (weld) bonds. Polymer science suggests that raising weld bonds to temperatures above the glass transition temperature (Tg) can improve their mechanical properties up to bulk strength. This research aims to verify if fused filament fabrication (FFF) printing at elevated print room temperatures (Tenv) raises interlayer temperatures above Tg and as a result enhances mechanical properties. In order to measure the increase in interlayer properties double cantilevered beam (DCB) testing is used. Additionally, tensile testing is used to evaluate the effect of elevating Tenv on general mechanical properties. DCB and tensile testing of samples printed at elevated Tenv values has shown an increase of 109 % in interlayer energy release rate (G1c), up to 50 % in ultimate tensile strength and 106 % in tensile toughness. This has been determined by comparing samples printed at elevated Tenv with samples printed at room temperature (23 ◦C). To further understand the mechanisms behind the enhanced properties, the temperature history during printing has been determined. This has been done by running simulations and using IR imaging. The temperature history of printed parts has been related to data obtained from mechanical testing. This showed that mechanical properties increased for samples in which the interlayer bonds resided above Tg for prolonged periods of time. In addition, optical microscopy and micro CT has been used to monitor the meso-structure of printed samples. This showed that voids in the sample caused by printing defects contributed to an increased spread in measured values. Mechanical properties corrected for effective surface areas showed a roughly 10 % increase for average values. While significant this is not in the same order of magnitude as the increase in mechanical values reported for printing at higher Tenv values. Combining these results it has been found that elevating Tenv does not significantly affect the meso-structure, but it does cause bonds to spend an increased amount of time above Tg. Therefore, it has been concluded that the enhanced mechanical properties are caused by weld bond healing due to the part residing above Tg.

The pipeline production rate and the pipeline’s superior mechanical properties are the crucial parameters for the offshore pipe laying company. Pulling on the bead, i.e. the action of moving the pipe on the firing line to the subsequent weld stall as the vessel moves forwards allows a fast and reliable pipe welding process. During this movement, the ends of the pipes are intermittently not supported and the bending moment in the weld is created as a consequence. Such operation imposes tensile loads in the welds, that are especially high for the 3 mm thick beadweld and the subsequent 3 mm thick hot pass (6mm thick in total). An important question arises whether imposing loads above the beadweld material yield capacity affects the overall fatigue performance of the girth weld. It is in the best interest of the offshore industry to assure that the pipeline welds properties during their 20-30 years operation lifetime are not adversely affected during its installation. This study investigates the effect of pulling on the bead on the weld’s fatigue behaviour.
The overstressing simulation of the test coupons, Finite Element Analysis (FEA) of the loaded beadweld and the small-scale specimens comparative fatigue study were conducted to investigate the effect of straining history on the full girth weld fatigue performance. Additionally, the holistic approach was implemented to investigate the consolidated effects of other factors: local weld root misalignment (Hi-Lo), weld porosity and microstructure quality. Four test coupons were welded together and overstressed to simulate the pulling on the bead action. To assess the stress-strain state in the beadweld and hot pass during the overstressing simulation the FEA analysis was done. The FEA results allowed to quantitatively determine the exact locations of the elastically and plastically strained beadweld material. Full welds were checked against the porosity and subsequently, 46 fatigue specimens were fabricated and tested in the 4-point-bending fatigue loading setup. Additionally, the fatigue crack observation techniques were implemented in order to separate fatigue crack initiation and growth phases out of the total fatigue life. The optical observations (with the optical-digital microscope) and the Potential Drop techniques were used. Tensile and hardness tests were done to compare the strength properties of specimens with different straining and porosity histories. Finally, the microstructure and fracture surfaces of the fatigue broken specimens were studied.
The research done in this study gives a good overview of the fatigue behaviour of the girth welds and provides insights into the straining history effects. Plastic straining was found not to affect the fatigue performance more than currently acceptable elastic straining does. Moreover, plasticity was found to reduce the negative effect of porosity. Additionally, the prevalent influence of Hi-Lo on the crack initiation phase was confirmed, making it the driving negative factor for the fatigue resistance.

Alkaline water electrolysis will become increasingly important for supplying the world with sufficient renewable energy, and with raw material for the chemical and pharmaceutical industry. Zero Emission Fuels B.V. (ZEF), a technology start-up based in the Netherlands, is developing a small scale alkaline electrolysis cell (AEC), which is integrated in a methanol producing micro-plant. The challenge of this project is to look into the use of polymers for the encapsulation of the ZEF AEC. The conditions of the ZEF AEC are not to be neglected. The system will run in a 30wt% KOH solution at a temperature of 90 °C and a pressure of 50 bar. Not many polymers will be able to withstand these conditions for a desired lifetime of 20 years. Using CES EduPack, a selection of 30 potentially suitable polymers has been made, from which high density polyethylene (HDPE), 40% glass-reinforced polyphenylene sulfide (PPS-40%gf) and polysulfone (PSU) are further investigated. Based on a literature review and a simple KOH ageing test, HDPE is found unsuitable for application in the ZEF AEC. PPS-40%gf and PSU have been subjected to durability testing for a range of different conditions involving a variety of KOH concentrations (15, 30 and 45wt%), and two oxygen partial pressures (O2 at 5 bar and air with pO2=20%) at different temperatures (90, 120 and 170 °C). The purpose of the ageing experiments is to give a better understanding of the effect of these parameters on the structure and integrity of the polymer; and to eventually be able to acquire a lifetime prediction. Extensive characterisation of the exposed samples has been carried out using different techniques, including weight measurement, tensile testing, DMA, creep-recovery testing, DSC, FTIR, XRD and SEM. After 12 weeks of ageing, it is found that glass-filled polymers are unsuitable for application in a strong alkaline solution at elevated temperatures, due to the dissolution of the glass fibres, which leads to a reduction in mechanical and barrier properties. However, the PPS matrix and PSU are found to be resistant to thermo-oxidative and chemical degradation in the tested ageing conditions. Only subtle changes in mechanical, visco-elastic and thermal behaviour are observed, which can be assigned to the effects of physical ageing. Due to the undesirable brittle nature of PPS, it can be concluded that PSU is the most promising candidate for the long-term application in alkaline electrolysis.

In order to save weight and reduce the impact on the environment, the aerospace industry keeps searching for strong and lightweight materials to reduce fuel consumption. A relatively new manufacturing technique made it possible to print nylon, reinforced with continuous unidirectional carbon fibres on a layer-by-layer basis and praised itself for making materials as strong as Aluminium 6061-T6. In this study, several coupons were 3D printed and tested in tensile, compression and three point bending mode to determine the material properties. When loaded in the direction of the fibres, the material did not live up to the upper limit of the rule of mixtures for fibre reinforced composites, which is mainly due to the amount of voids in the material. These voids were discovered when studying the micro-structure of the material and were both in between the layers, as well as in the layers themselves.
The mechanical tests were performed to validate constructed finite element models, simulated in Abaqus/CAE. The simulation results showed some mismatches with the mechanical test results, which were caused by the assumptions made to simplify the models. Finally, an air plane chair bracket, which has as task to transfer loads from the back seat panel to the chair, was modelled, simulated and designed. Major issues in the designing process were the limitations of the printer, such as not being able to choose which exact areas to reinforce in a specific layer. The final design of the chair bracket was manufactured and was tested to see if it could withstand a load of 375 N. The part did not fracture under the load, but a high displacement of the back seat was observed, as well as damage occurring at other locations than the simulations indicated.

The study showed that this 3D printing technique is able to print structural aerospace parts, given that the part is not too complex and that it is loaded in a static nature. More freedom in the design software of the 3D printer could allow the user to successfully design more complex structural aerospace parts. Since the constructed parts were only tested under static loading, further research has to be conduction into dynamic loading of these parts before it is possible to actually implement these in air planes.