Wind turbines experience a wide variety of load cases, both when in operation and when parked. While a turbine in operating conditions often experiences larger loads than in parked conditions, uniquely large inflow angles may occur in parked conditions resulting in less well-understood aeroelastic conditions. Vibrations under these conditions are referred to as stall and vortex-induced vibrations (SVIV). While there has been some success in simulating these conditions, there has been a lack of experimental studies in the literature.

To address this gap, this study analyses the experimental data obtained from a 3.8 MW research wind turbine. This turbine was equipped with strain gauges at the root of 2 blades and in the tower, as well as pressure sensors on one blade and accelerometers in the nacelle and tower. Several experiments were performed with a focus on simulating various conditions where SVIV may occur. This included pitch traverses of a single blade, yaw traverses, and a traverse of the azimuth angle at a yaw angle of 90 degrees to determine the effect of the inclination angle.

Severe stall-induced vibrations were identified at the yaw angle of 110 degrees, when a single blade pointing up in the sky was pitched 180 degrees, and the other two blades were in vane positions at 85 degrees pitch. The severity of these vibrations strongly varied with the wind speed, at an average wind speed of 19.5 m/s the test needed to be stopped early for the safety of the wind turbine, while at 16.6 m/s, the vibrations appeared to reach a limit that was still considered to be safe. The first tower mode experienced the most severe increase in magnitude during these conditions of SIV. In other conditions, no severe case of SIV was identified. However, under several conditions, smaller increases in vibration magnitude were identified. These are likely the result of slightly reduced aerodynamic damping.

No vortex shedding was identified in the data obtained from these experiments. This means that the tested turbine/blade design either did not experience vortex shedding or that it simply could not be measured by the installed instruments. Having pressure sensors installed in multiple locations along the blade would likely help in identifying vortex shedding for future experiments.

Simulations from the experimental conditions were performed using the aeroelastic software PHATAS together with the aero module from ECN (now part of TNO). This simulation used beam-based structural modeling with either a BEM or free vortex wake aerodynamic model. These simulations confirmed the need for realistic turbulent inflow conditions and the need for a good dynamic stall model. However, the structural model failed to simulate all the natural frequencies with the expected accuracy. This should be improved upon for future use of this structural model in parked conditions.

Several structural models that could serve as building blocks for a next-generation FSI-based kite design tool have been developed. The models represent the leading-edge inflatable V3 soft wing membrane kite of Kitepower B.V., by using a wireframe multi-plate representation. Where the shape deformations are calculated using bridle line system models. Both by using a trilateration algorithm and a particle system model were accurate results found. Accurate in comparison to the experimentally obtained data from a photogrammetry analysis. The existence of slack and its non-dependence on empirical relations made the force-based particle system model best.

The rapid development of the wind industry over the past few years has pushed turbine manufacturers to meet the growing energy demands by designing and producing large scale wind turbines.This also means development of larger monopile foundations for the foundation designers in the case of offshore wind turbines.\ Generally, the turbine tower and monopile are modeled together and the loads from the rotor-nacelle assembly are provided by turbine manufacturers.\ The offshore industry is now showing more interest in extracting the loads from the top mass by developing their own tools in order to reduce the dependency on the manufacturers. In order to aid in this process, the present master thesis aims to develop a linear model based on the concept of Dynamic Substructuring which employs a set of equations to compute the interface forces using the kinematics.\ Furthermore, the developed prediction model is used to analyze the loads occurring at the interface between the rotor-nacelle assembly and the tower for different wind speeds and wind conditions.Consequentially, the model was found to produce acceptable loads at higher wind speeds for selected degrees of freedom at the interface while failing to do the same for other degrees of freedom.The results in time domain were converted to the frequency domain to analyse the resonance.The influence of resonance on the interface degrees of freedom was found to be higher at wind speed below the rated condtion.\ These findings can be used as a basis to conduct further investigations into the application of numerical integration concepts to aeroelastic structures.

As the need for abundant and reliable renewable energy increases, there is a growing interest in floating wind turbines, which would allow to harness the wind resource in areas where bottom-founded wind turbines cannot be used. However, the movements of the platform are expected to cause unsteady aerodynamics effects, including different wake dynamics, a variation of the induction field at the rotor and blade-vortex interactions. Consequently, there is no general consensus on whether Blade-Element Momentum (BEM) codes could be employed to model the aerodynamics of floating wind turbines. This poses a serious issue as BEM models are widely used in industry practice.
This project proposes to analyze the impact of surge motion on the induction field of a horizontal-axis wind turbine. A suitable actuator disc model is developed, followed by an actuator line model that allows to study the effect of the finite number of blades. Both models are implemented in OpenFOAM, an open-source CFD software. The simulations are run for a range of case studies with imposed baseline thrust, amplitude of the thrust variations, surge frequency and surge amplitude and the resulting induction factors are compared to those obtained with a dynamic inflow model, to assess whether a momentum method could lead to accurate predictions. Particular attention is given to the identification of the wake states of the streamtube, since both turbulent wake state and vortex ring state imply a breakdown of momentum theory.
The results of the actuator disc simulations show that in all cases there is a good agreement with the induction factors obtained with the examined dynamic inflow model. Turbulent wake state is only entered when a high thrust coefficient is reached at low frequencies, while propeller state is only entered when a negative thrust coefficient is reached at low frequencies. Furthermore, no signs of vortex ring state were detected.
Some of the cases considered during the first part of the project were also run with the actuator line model, to examine the finite blade effect. A rotor with three blades was modelled. The resulting disc average induction factors are in excellent agreement with those obtained with the actuator disc model, while the induction at the disc center is lower. The contour plots show that the conclusions on the wake states entered by the streamtube remain valid. It is advised to test this model at different tip speed ratios and for rotors with different numbers of blades.
Overall, this project contributes to a better understanding of the aerodynamics of floating wind turbines and gives confidence in the possibility of using momentum methods during their design phase. Furthermore, the CFD models developed are a flexible tool that may be used for future research on related topics.

Under real sea conditions, floating vertical-axis tidal turbines experience motions in six degrees of freedom which influence the relative velocity perceived at the turbine's blades, with a direct effect on the loading and performance of the rotor. Understanding the fluid-structure interaction of a vertical-axis tidal turbine under the floating carrier’s motions provides insights about critical aspects for the design of floating hydrokinetic systems.
In this project, the TU Delft in-house U2DiVA code is indicated as a conservative and time-effective tool to assess the hydrodynamic response of a vertical-axis tidal turbine under surge, sway and yaw motions compared to the computational fluid dynamics approach commonly used in the scientific literature. The effect of surge motion on the fluid-structure interaction of two existing vertical-axis tidal turbines is investigated using U2DiVA. The analysis focuses on the details of the time evolution of the flow field perceived at the blades and on the modified cyclic loading and power extraction when given surging conditions are imposed.

Wind farms experience significant efficiency losses due to the aerodynamic interaction between turbines. A possible control technique to reduce these losses to a minimum is yaw-based wake steering. This thesis investigates the feasibility of this technique by calibrating a surrogate model called the FLOw Redirection and Induction in Steady-state (FLORIS) model on a data set from the Lillgrund wind farm and using it to estimate the potential energy gain. The data set available is processed methodically to remove outliers and erroneous data points, resulting in a reliable and useful data set. It is used to obtain free stream wind conditions per time step and relate those to power measurements. The data set is consequently used to calibrate the tuning parameters of the FLORIS model. The calibration is done using a newly proposed method that determines the tuning parameters per combination of wind speed and turbine spacing. A difference with commonly applied calibration methods is that power measurements are used instead of predicted powers or flow field data from high-fidelity models. The performance of the calibrated model is tested through multiple uncertainty analyses. It is found that the model has a significant bias but low uncertainty by comparing the predicted wake losses with measured wake losses. This bias can potentially be reduced if atmospheric stability is taken into account. With the bias and uncertainty quantified, the FLORIS model is used to optimize the annual energy production of the Lillgrund wind farm by finding the ideal yaw angles for specific inflow conditions. A significant energy gain can be achieved when the optimal yaw angles are determined deterministically. However, the energy gain decreases drastically when uncertainty in input conditions is considered, showing that these yaw angles are not robust in terms of performance under uncertainty. More robust yaw angles can be obtained when the input uncertainty is taken into account during the yaw optimization. The energy gain achievable with these more robust yaw angles is approximately 3.4%. Therefore, it can be concluded that achieving an energy gain using yaw-based wake steering is feasible for the Lillgrund wind farm.

In this thesis project, it is investigated to what extent the energy consumption of an electric aircraft can be improved by using a variable pitch and RPM propeller, including energy recuperation during the descent, with respect to a constant pitch propeller. The largest percentage of total mission energy saved using the variable pitch propeller with respect to the constant pitch propeller is about 4.1% for a cruise distance of 5 km when using two blades and about 3.0% for three blades. This percentage reduces to approximately 0.7%, both for two and three blades, when the cruise distance is 200 km. A two-bladed propeller therefore turns out to be the best option. Regarding the effect of the propeller airfoil camber and thickness with respect to the NACA 4415 airfoil, the NACA 0015 airfoil performs best on a short range mission and the NACA 2415 performs best on a long range mission.

This thesis investigates the potential of a semi-passive trailing-edge flap on a large conceptual wind turbine. The mechanism passively reacts to blade and tower accelerations by changing the airfoil camber, opposing the dynamic loads on the turbine. An active element is present, which influences the mean of the flap oscillation. First, a low-fidelity, parameter study was done in MATLAB. Next, the flap model was implemented in the aeroelastic code HAWC2, to capture dynamic and structural effects due to blade accelerations. Results show that the semi-passive design reduces ultimate and fatigue loads, during normal power production. Effects on AEP are minimized by the active element and are an improvement to the passive model. The present study motivates simulation of more design load cases, e.g. parked or grid failure. Also, the benefits of the mechanism should be investigated in combination with a new, enlarged, rotor at similar key loads.

To increase the total wind farm power output, the wind farm layout needs to be optimized. The power output of a wind turbine depends on the incoming velocity, while the velocity is influenced by the wake of the upstream wind turbines. Wind Farm Layout Optimization (WFLO) problems makes use of the so called low fidelity wake models, which predicts the velocity downstream of a turbine. The analytical Jensen wake model with a top hat velocity wake profile is commonly used to perform the WFLO during the preliminary design phase of a wind farm. However a top hat velocity profile is not an accurate depiction
of the actual velocity profile downstream of the turbine wake. To get a more accurate wake profile the model needs to be extended. To improve the wake model, the role of the stability of the Atmospheric Boundary Layer (ABL) on the development of the turbine wake is analyzed using the software openFOAM and SOWFA. It is noticed that the analytical Jensen-Gaussian wake model is in better agreement with measurement data than the Jensen top hat wake model. It is verified that it is necessary to include the added Turbulence Intensity (T.I) induced by the wind turbine. For the Jensen- Gaussian wake model, the Gao turbulence model gives results that are in good agreement with the experimental data. The Jensen-Gaussian wake model is extended to be used inside a wind farm with multiple wakes. The power output for a row of 10 Vestas V-80 turbines in the Horns rev Wind farm is computed. Using the equivalent velocity by weighted area averaging over a discretized wake turbine-cross section in combination with the power curve, the power output of a turbine can be computed. Using the energy superposition method the equivalent velocity for alligned turbines can be computed. Comparison with measurement data shows that there still is a difference between the results from the wake model and the measurement data. To further improve the Jensen-Gaussian wake model it is important to take into account the effect of the stability of the ABL on the wake. The different stabilities for an offshore ABL are simulated with SOWFA and the turbine wakes are computed. The different wake recovery rates and elliptical shapes due to the stability of the ABL are included in the extended model.
Using the offshore Horns rev wind farm data, the extended Jensen-Gaussian model in combination with the mixed-discrete Particle Swarm Optimization (MDPSO), the WFLO is carried out. The WFLO predicts that it is important to take the stability of the different models into account. However it is concluded that the improvement cannot be quantified, due to the uncertainties in the computation in the power output of each wind turbine.

The efficiency of a wind turbine depends largely on the wake of the upstream turbine. Seeking to contribute towards the development of a wind farm solver using a Lagrangian scheme to analyze the wake, this thesis analyses and validates (using MATLAB) a Vortex Particle Method (VPM) algorithm by simulating the behavior of vortex rings. Due to the computationally expensive nature of VPM schemes to solve n-body problems, such as the simulation of vortex rings and wind turbine wake, a Fast Multipole Method (FMM) library called BBFMM3D has been studied and validated (using C++) for the acceleration of computation of matrix-vector products which are essential in solving n-body problems. The execution of the VPM algorithm has been shown to be possible to be accelerated using the validated FMM library by implementing the library in the VPM scheme to simulate the case of a single vortex ring. Using the VPM solver developed to simulate vortex rings, the simulation of a wake of a wind turbine, that was modeled using actuator lines, was performed (without FMM) and the results have been analyzed and attempted to be validated. It has been found that the VPM scheme generates less than accurate results of the velocity profile of the wake with respect to other CFD simulations and the model could not be validated due to the results not being accurate up to the mark. However, the accuracy of the results has been found to rely significantly on the formulation of the strengths of the vortices shed into the wake. Two formulations for this purpose have been presented with the results showing signs of improvement from one formulation to another.

Modern wind turbines frequently operate at off-design conditions during their life cycle. They undergo dynamic loads characterized by unsteady aerodynamics. Predicting these unsteady aerodynamic loads has been very difficult due to the non-linear nature of unsteady aerodynamics. Especially when operating near the stall region, these turbine are prone to increased loads because of dynamic stall. Dynamic stall is typically observed when there is a turbulent inflow, yaw misalignment, or severe wind shear causing periodic variations in angle of attack. Nonetheless, the nature of dynamic stall phenomenon is still a topic under investigation. The aim of this research was to investigate the performance of dynamic stall models in yawed and standstill conditions by using current state-of-the-art engineering models and validating the results with the New MEXICO (Model Rotor Experiments under Controlled Conditions) measurement campaigns. The first part of the research dealt with a detailed analysis of the New MEXICO experiments in standstill and yawed flow conditions. This part also encompassed extracting 3D polars from pressure measurements and a spectral analysis to characterize any vortex shedding phenomenon in standstill conditions. The second part of the research was concerned with validating dynamic stall models implemented in ECN’s in-house aeroelastic tool Aero-Module. Three different dynamic stall models namely: Snel, ONERA, and Beddoes Leishman model, were extensively validated and improved using New MEXICO measurements in standstill and yawed flow conditions. Finally, a case study was performed on the AVATAR rotor, using afore-mentioned dynamic stall models, to access their effect on aerodynamic damping and, consequently, in predicting the onset of aeroelastic instabilities. The research was able to shed light on our current understanding of dynamic stall phenomenon and the way we model it, hoping to improve the dynamic stall modeling capabilities in the future.

The wind field and, more importantly, the power production change when downstream wind turbines are located in the wake of an upstream wind turbine. Wind farms become larger and therefore the field becomes more complex. Turbines will experience the influence of the wake of multiple wind turbines. The interaction of these wakes can be modelled using different approaches. Using numerical solvers is very computationally costly and accordingly, there is a need for simple engineering wake models which represent the wind field in a good way.
The focus of this MSc Thesis project is to find a superposition method in combination with the Jensen/Park model, which is in good agreement with a representative reference for mixed wakes in reality.
This reference could be either Large Eddy Simulations (LES) or large-scale measurements data from the BEACon campaign, carried out by Ørsted. Comparisons between some datasets showed that more research is needed to identify the discrepancies between the wake fields of both datasets.
Studies have been carried out in literature, but mostly focussing on wind speeds below rated wind speed. Therefore, cases with an inflow wind speed just above and below rated wind speed are considered. The superposition methods looked at are linear superposition, quadratic superposition and the maximum deficit method. The modelling of the wake boundary, rotor averaged wind speed and power are discussed. Some superposition methods are in good agreement with the LES results, but because a uniform profile is modelled, more research is needed to assess if these conclusions also hold for sheared inflow wind profiles.
Apart from examining the combination of superposition methods and the Park wake, a study is also carried out to examine if single LES wakes can be superposed to mimic LES wake fields with multiple wind turbines. Based on the available single LES wakes, there are still essential differences in the results, but these might be overcome if more single LES wakes can be used.
As the proposed superposition methods are not necessarily "true", a preliminary study is carried out in which the superposition method is optimized. This gives an insight in the number of upstream wakes that need to be included and the possible scaling or improvement of the superposition methods.