Understanding atmospheric motion is crucial for analyzing wind-structure interactions, particularly within the Atmospheric Boundary Layer (ABL). Traditionally, wind tunnels have simulated unsteady ABL conditions using active and passive devices. Recently, multi-fan wind generators have emerged as a flexible, cost-effective alternative, allowing independent fan control to replicate wind conditions the wind turbines, airborne devices or civil structures are subjected to.

This thesis investigates the capability of a newly manufactured 3 x 3 multi-fan wind generator to produce idealized uniform flow, linear shear flow profiles, and streamwise sinusoidal gusts. The system consists of off-the-shelf computer fans, controlled via open-source software, offering a low-cost and adaptable approach. Particle Image Velocimetry (PIV) measurements show a 96% uniform core region with 11.5% turbulence intensity 1 m downstream from the system. There, the area of the core is reduced to 1/9 of the system’s area due to the outer shear layer. The system struggles to replicate linear wind shear, but can generate oscillatory gusts at 0.2 Hz, 0.4 Hz and 0.8 Hz, corresponding to the operating frequency of the fans. Further improvements should focus on extending the configuration to maintain a larger core region, and on reducing swirl dynamics and recirculation to enhance flow quality for wind engineering applications.

The thesis presents an analysis of unsteady vortex shedding and vortex induced vibrations (VIV) on wind turbine blades using a combination of Stereoscopic Particle Image Velocimetry (SPIV) and Computational Fluid Dynamics (CFD). A spanwise uniform blade with a symmetric NACA0021 airfoil section, 0.075 m chord, and 0.4 m span was tested experimentally for a static, plunging, and surging blade oscillating at 90 deg angle of attack, 5 Hz, and 2.5 Hz frequency, 1 chord oscillation amplitude. Equivalent CFD simulation cases were run to compare the results and validate the CFD setup, which was used to analyse additional cases of interest. The extent of the three-dimensional flow behaviour due to the tip vortex was described, along with determining the lock-in region for the two motions using additional simulations. The results indicated that the spanwise convection of the tip vortex and the transition to a ‘2D’ flow were only dependent on the blade aspect ratio and independent of the freestream wind speed, the type of motion, and the oscillation frequency. The lock-in region for the plunging and surging blades was presented, noting that all the surging cases were locked in, most likely due to the large oscillation amplitude. However, the surging cases had a positive aerodynamic damping and hence, VIV would not be excited for a freely vibrating blade. It is suggested as future work to test a flexible blade and study the onset and development of VIV.

Wind farm control is an area of research that looks to maximise the performance of individual but most importantly the collective performance of wind farms. The performance of turbines located in a wind farm is heavily dependent on the effects of turbine wakes, and improving the efficiency of wind farms by mitigating the wake effects is an ongoing and promising field of research. Simulating the wake effects of a turbine is generally done with computationally expensive physics based modelling techniques like CFD simulations. The computational complexity of these models limits the scope of optimisation that can be performed on wake mitigating controls strategies such as dynamic induction control. A substitute engineering model like the free vortex wake model could provide better insight into dynamic induction control by virtue of its computational efficiency allowing a wider optimisation parameter space. In this thesis an optimisation framework based on a 2D implementation of the free vortex wake model and genetic algorithm optimisation is used to investigate dynamic induction control for a simple two turbine wind farm beyond the research that has been done up to this point. While yet to be confirmed with higher fidelity simulations, initial results corroborate earlier findings for simple dynamic induction signals and additionally indicate that dynamic induction control could benefit from multiple harmonics. The optimisation framework achieves its goal in allowing a wider parameter space to be searched for optimisation, even on consumer grade desktop hardware, and shows potential as a tool for further investigation of dynamic induction control.

Wind turbines experience a variety of different operating conditions during their operation. As a result, sections of their blades experience a range of angles of attack. Additionally, several commonly used modelling methods use the sectional loading of the blade directly. To adapt this sectional loading for higher angles of attack dynamic stall models are often used. However, these models are generally based in empirical results from pitching airfoils at moderate angles of attack. In this context moderate angles of attack are considered to be below 30 degrees. As a result, their applicability to stationary airfoils at much larger angles of attack is not known. Placing an airfoil at such angles causes it to enter deep stall conditions. During deep stall, vortex shedding becomes a dominant effect on the loading of the airfoil. This behaviour can greatly affect the loading on the turbine and thus should be modelled accurately.

Based on these observations, the presented work aims to investigate the applicability of current dynamic stall models to deep stall conditions. Additionally, an adapted dynamic stall models is created. This model was designed to improve the prediction of the vortex shedding frequencies for stationary airfoils in deep stall. In this process, three different experiments were used for comparison to experimental results.

Three previously developed dynamic stall models were considered, which all showed different responses in terms of vortex shedding. These responses ranged from being completely aperiodic to undamped oscillations. However, none of the models was able to accurately predict the vortex shedding frequencies of the considered experiments. Therefore, the current dynamic stall models were found not to be applicable to the deep stall regime.

Therefore, an adapted model was constructed and calibrated based on the experimental results. This model greatly improves the prediction of vortex shedding frequencies in all considered experiments. Additionally, the accuracy of this model when applied to pitching airfoils at moderate angles of attack was examined. When applied to those conditions the model performed only slightly worse compared to a reference dynamic stall model. Hence, the presented models shows promising results for extending current dynamic stall models to the deep stall conditions.

Finally, it should be noted that the general validity of the adapted model remains to be investigated further. For this purpose, more experiments are to be considered. However, the approach outlined in this research has been shown to provide good results for creating a dynamic stall model which includes vortex shedding in deep stall conditions. Therefore, the presented method can be used to develop a validated dynamic stall model for such conditions provided more experimental results are available.

The hunt for new methods to harness wind energy is at an all-time high as wind energy quickly overtakes other renewable energy sources. Deep waters are promising alternatives with higher average wind speed (thus, higher yields) and also due to lesser availability of land on shore. Even though the design of the blades and rotor of a FOWT is largely similar to that of onshore counterparts, they are mounted on floating platforms, unlike the fixed platforms on onshore wind turbines. Due to the excitation from wind and waves, these foundations are made to oscillate in all six degrees of freedom. Pitch and surge motions have the most significant effects out of all the possible types of motions. Since surging motion comes closest to the pitching motion, this study attempts to compare the aerodynamics of these two motions by simulating a pitching wind turbine using an actuator disk CFD model. This study can also be considered a continuation of the actuator disk study of the surging motion.

The primary objective of this thesis is to understand the differences between the aerodynamics of a surging and a pitching wind turbine using an actuator disk model. The actuator disk is made to pitch sinusoidally with the prescribed pitching frequency and amplitude. In order to identify the effects of motion and thrust prescription separately, three different types of actuator disk models were created. Using the first model, which simulates a pitching actuator disk with constant thrust, the effects of pitch motion are captured. The second model which simulates a still actuator disk with dynamic loading gives the effects of change in thrust on the rotor. The last model, which is a combination of the first two with a pitching actuator disk under dynamic loading, attempts to simulate a more realistic pitching actuator.

On comparing the surge and pitch motions of the actuator disk, the stark difference observed is the meandering in wake due to flow shear and asymmetric tip vortex shedding. At high thrust and low-frequency cases, turbulent wake
states are also visible in the wake. At higher frequencies, they are nearly as strong as the tip vortices and introduce non-linear effects on the induction field. They also tend to introduce varied dynamic responses based on the radial location. It was also observed that the motion introduces a phase delay that varies with the radial position on the rotor.

Thus, this project provides a comprehensive view of the effects of platform pitching motion on a FOWT and it differences with respect to the surging dynamics. This project also confirms that an actuator disc model is capable of reproducing the effects of pitching in a FOWT. Using the insights provided in this thesis, a better dynamic inflow can be developed for use in industries

As the world grapples with the challenges of global warming, there has been a significant push for the development of new methods of harnessing renewable energy. Energy harvesters have been a new avenue to generate electrical power locally for remote applications. An energy harvester is a device that taps into aeroelastic phenomena to convert vibrational energy into electrical energy. Additional benefits of harvesters include being lower in complexity, ease of maintenance, and the resulting cost-effectiveness.

The splitter plate represents a wake control device that alters the wake to make the energy capture more efficient and practical when attached to a cylinder. However, the flow mechanisms under different conditions have not been fully understood. The motivation of the present work is to bridge this gap by characterizing the mechanisms and wake dynamics of a cylinder with a flexible splitter plate. This will help advance the understanding of the wake behavior of these devices to make safe and efficient devices. To characterize the behavior of the wake, measurements are made at different flow speeds. A preliminary analysis of the splitter plate motion showed that increasing the Reynolds number produced four distinct regimes to be studied further: crossover, baseline, reduced amplitude, and chaotic regimes. The initial analysis also revealed a condition named burst, defined as a period of resonant, two-dimensional, high amplitude oscillations of the splitter plate. All regimes are compared to a bare cylinder with no splitter plate to ascertain differences in wake behavior.

The use Helium Filled Soap Bubbles (HFSB) in flow diagnostic tools such as Particle Image Velocity (PIV) experiments has permitted large-scale measurements. A Lagrangian Particle Tracking (LPT) experiment was performed in which the HFSB is tracked individually with the help of an algorithm. The present study employs Shake-The-Box (STB), a novel LPT tracking algorithm that uses time-resolved experimental data to predict future tracer particle locations. The benefit of such an algorithm includes reduced ghost particle detection and computational efficiency. The details of this experimental method have been covered extensively in this thesis. The raw experimental data was processed to perform statistical analysis, proper orthogonal decomposition, and spectral analysis to provide insights to meet the thesis objective.

The placement of wind turbines in wind farms is economically beneficial to the industry. Consequently, the wake generated by the upstream turbines introduces interactions with downstream turbines that significantly affect power generation. However, it is common practice to control the individual turbines to operate at their optimal settings and neglect the wake effect. Turbines in a wake experience lower wind speeds and an increase in turbulence. This results in lower power generation and increased fatigue loads on the downstream turbine. Wind farm control aims to increase overall power generation and reduce fatigue loads by considering all turbines in the wind farm.
Recently, numerous studies have been conducted to control the wake itself. The helix approach is one such strategy and seeks to initiate better mixing of the wake. For this, it uses IPC to create sinusoidal blade bending moments acting on the rotor disc. This in turn creates a helical structure in the velocity of the wind, allowing it to better mix with the surrounding energy-rich wind. As a result, the waked downstream turbine experiences an increase in power due to the reduced velocity deficit in the wake. An extensive load analysis of the helix method has shown that the rotational frequency of this helical wake can be measured in the bending moments of the waked downstream turbine. To further develop the field of dynamic wake mixing, this work investigated the initialization of the helix method in a downstream turbine by amplifying these signals through synchronization with the incoming helical wake.
This method was tested in simulation environments with increasing levels of fidelity and number of turbines simulated. The results with low fidelity showed that synchronization could be achieved without loss of performance. In addition, less pitch actuation was required to generate large bending moments due to amplification. The final experiment was conducted for a three-turbine setup and was performed in a high-fidelity model using CFD methods. Here it was found that -90 degrees out-of-phase synchronization with the incoming helical wake resulted in a 9.8% increase in power generation in a turbine further downstream. These results demonstrate the clear potential of wake synchronization in a multi-turbine setting. This study, therefore, contributes to the field of dynamic wake mixing by introducing a novel field of wake synchronization.

A 3D panel method code, named AWSM3D, is developed as an extension to the AeroModule aerodynamic software suite of TNO which includes a Blade Element Momentum theory (BEM) and a Lifting Line (LL) code. The underlying panel code method is presented along with a gradual process of simulation validation culminating in validating the code against the New MEXICO experimental dataset and an established panel code. In view of exploring the avenue of sweeping wind turbine blades as a way of increasing wind turbine efficiency, the capabilities of the three proposed aerodynamic models in solving flows about swept geometries are studied. The unmodified BEM model is incapable of adequately modeling flows over swept geometries. The limits of the cross-flow theory which the LL simulation hinges on for the simulation of swept flows are put to the test. Both a fixed wing and a conventional Horizontal Axis Wind Turbine case show that the LL solution in proximity to body extremities where the cross-flow component is prominent fails to capture smaller-scale local effects caused by said cross-flow. Despite this, the larger-scale effect is reproduced well. Although AWSM3D requires no airfoil polars and produces an accurate inviscid flow solution for more complex body geometries, it currently excludes viscosity and has a much higher computation time in comparison to an LL. A number of further AWSM3D development and research into the effects of sweep suggestions are formulated.

The main objective of this master’s thesis is to experimentally investigate the effects of dynamic inflow conditions on the wake of an actuator disc. Dynamic inflow is the unsteady wake in momentum theory. Dynamic inflow can experimentally be generated by a gust generator. During the experiments, in which Particle Image Velocimetry (PIV) was used as a flow measurement technique, the gust generator produced different gust frequencies and gust amplitudes. The oscillating gust results in the oscillating wake of the actuator disc. The oscillations of the wake cause wake expansion and compression. The expansion and compression of the wake is mainly dependent on the gust frequency. The higher the gust frequency is, the more oscillations the wake will see. So, the actuator disc’s wake was influenced by the gust.

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.