Wind farm interactions with the surrounding airflow leads to a reduction in velocity greater than the linear sum of single turbine inductions and is known as global or upstream blockage. The mechanisms and magnitude of global blockage effect are not yet fully understood. Models to simulate upstream blockage to improve efficiency estimates and better understand global blockage have thus far not been refined. The aim of this research is to investigate the sensitivity of upstream blockage to the numerical configuration of CFD simulations to improve model accuracy and understanding of global blockage. RANS simulations are executed under steady state conditions with a k − ϵ turbulence model. A neutral stability, pressure-driven atmospheric boundary layer is modelled with fully developed uni-directional flow. Wind farms are modelled as actuator discs with 5 rows of turbines on flat terrain. Streamwise and spanwise spacing is set to 7 turbine diameters (D) and 5D respectively. Global blockage is measured against the induction of a single turbine at 2.5D upstream. Variables investigated include the domain height, lateral extent, inlet and outlet distances from the wind farm. Domain heights ranging from 5D to 25D are investigated for change in magnitude and scale of upstream blockage for a laterally infinite wind farm. A clear trend of increasing blockage with domain height is observed. At domain heights of less than 15D, upstream velocity is increased by a maximum of 0.18% (5D). Larger domain heights produce a maximum velocity reduction of 0.23% (25D). The shape of upstream blockage is independent of domain height. A finite wind farm of 5 columns and lateral extents ranging from 2.5D to 20D on each side are utilized to investigate the impact on blockage. Wider domains of 5D to 20D display increasing blockage with width, while a domain of 2.5D exhibits behaviour similar to a laterally infinite wind farm. Blockage ranging from 0.13% (10D) to 0.31% (20D) reduction in velocity is shown to be highest at the center column of turbines and decreases toward the outer columns. Inlet and outlet distances ranging from 15D to 100D are modelled. Upstream blockage for inlet distances of 50D to 100D produce consistent upstream blockage magnitude and extent of 0.22% and 30D respectively. Shorter inlet distances result in decreased upstream blockage with a minimum of 0.12% (15D). The shape of blockage remains consistent through all inlet ranges. Outlet distance have no identifiable impact on upstream blockage magnitude and extent. Changes to the numerical configuration show a clear correlation of increased blockage with cross sectional area of the domain. Constraining the domain in the vertical and lateral directions constricts flow resulting in reduced blockage. Blockage becomes independent of inlet distances at values of 50D and higher. Outlet distance has no identifiable impact on upstream blockage. Choosing a numerical configuration with adequately sized domain boundaries is pertinent in producing realistic upstream blockage.

Studies revolving around data-driven methods have been on a rise in recent years to improve highly modelled methods such as the two-equation turbulence models of Reynolds-averaged Navier-Stokes (RANS). Similarly, such data-driven methods are implemented into partially-averaged Navier-Stokes (PANS). PANS is a young bridging method that fulfils the requirements of bridging methods set by Speziale [1]. It can be adjusted according to the fraction of a flow field that is desired by the user to be resolved and modelled by changing the value of fk, a parameter that takes a value between 0 and 1. In this thesis, PANS is extended via a combination of two data-driven methods: k−corrective frozen RANS [2] and data-driven stochastic closure simulation (DSCS) [3]. The k−corrective frozen RANS method aims to correct for the model errors in the k−equation and the anisotropy of the Reynolds stress tensor derived from the Boussinesq approximation. While DSCS also aims to correct for the anisotropy of the Reynolds stress tensor, it considers that PANS solves for the unresolved part of the flow field and thus corrects for the unresolved anisotropy. While PANS and the two data-driven methods have independently been proved to work as they were theoretically desired, combining these ideas has not yet been attempted. As any turbulence kinetic energy solving RANS turbulence model can be developed into PANS form, the k − ω SST model was chosen for the best initial prediction. This SST model for PANS is extensively derived and then reformulated to produce the target correction terms. The correction terms are analysed at various values of fk and they show good agreements with fk.

When designing an airborne wind energy system, it is necessary to be able to estimate the traction force that the kite produces as a function of its flight trajectory. Being a flexible structure, the geometry of a soft kite depends on its aerodynamic loading and vice versa, which forms a complex fluid-structure interaction (FSI) problem. Currently, kite design is usually done on an experimental basis since no model meets the requirements of being both accurate and fast. In this project, an FSI methodology is developed to study the steady-state aerodynamic performance of leading-edge inflatable (LEI) kites by coupling two fast and simple models. On the structural part, the deformations are calculated with a particle system model, based on the assumption that the shape of the kite can be modelled using a wireframe wing model represented by the bridle line attachment points, whose coordinate changes are modelled using a bridle line system model and canopy billowing relations. On the aerodynamic side, the load distribution is calculated with a 3D nonlinear vortex step method, coupled with 2D polars obtained with a correlation model derived from Reynolds averaged Navier-Stokes (RANS) analysis, to account for viscous effects and flow separation. Furthermore, with the 2D correlation model it is possible to consider changes in the thickness and the camber of each section. Based on 2D thin airfoil theory, the three-quarter chord point is used to determine the magnitude of the forces, and the one-quarter chord point is used to determine the direction of these forces. Moreover, the model developed for LEI kites can consider canopy billowing and variations in kite and airfoil geometry while proving robust and inexpensive. This model has been validated with several geometries and a RANS analysis of the LEI kite, showing great accuracy for pre-stall angles of attack. The coupling of these two models results in a fast aeroelastic model of LEI kites capable of predicting the steady-state deformations and aerodynamic forces on the kite for the range of actuation settings and inflow conditions expected during a normal pumping cycle. Furthermore, the results show that the deformations follow the same trends as the results from the photogrammetry analysis and that, by taking into account the deformations that the kite undergoes, the aerodynamic forces more closely resemble experimental data.

The recurred idea of developing multi-rotor wind turbines has led to the need of more accurate surrogate wake models which allow for a fast annual energy production (AEP) calculation and further understanding of the aerodynamic power losses of multi-rotor wind turbines. The present thesis develops a surrogate wake model of a multi-rotor-two turbine validated against computational fluid dynamics (CFD) simulations of type RANS-AD. The outcome is a superposition model of an analytical representation of the wake which base function coefficients are stored in look-up tables as a function of the wind inflow conditions affecting the turbine. The derived surrogate model is able to predict the overall wind farm efficiency with more than 90% accuracy while compared to RANS-AD models. Towards the end of the thesis, a comparison between a single-rotor wind farm of 18 V29 turbines and a multi-rotor wind farm composed by nine 2R-V29 turbines (hypothetical turbine) is evaluated through RANS-AD simulations within the same wind-farm area. The energy ouput showed to be highly dependent on the wind-farm geometry, and the wind direction average suggest that 5% more energy yield is obtained from the multi-rotor-farm for velocities below rated speed.

The power production of downstream wind turbines in a wind farm is significantly impacted by wake effects of upstream turbines. Improving the layout optimization process could reduce these wake losses and therefore result in more efficient wind farms. The wakes are also affected by atmospheric stability conditions, as stable conditions lead to a reduced wake recovery, and unstable conditions lead to an improved wake recovery. The topic of atmospheric stability has quickly gained more attention in wind energy research over the recent years. However, there is still little research done that focuses specifically on the effects of atmospheric stability on the wind farm layout optimization process. This thesis aims to determine the effects of atmospheric stability on the optimal layout of a wind farm and to quantify the potential benefits of considering stability conditions in the layout optimization process. A stability-dependent Jensen wake model is developed, using a stability-dependent wake decay coefficient based on the non-dimensional Obukhov length. The developed model is implemented in FLORIS, a wake modeling utility for Python, to calculate the wind field and the annual energy production (AEP). A simplistic layout optimization method is used, in which the positioning of wind turbines relative to each other is fixed and only the orientation of the entire wind farm is varied. To quantify the potential benefits of considering stability conditions, the layout optimization is done twice for each analyzed case: once using the determined stability conditions and once using the assumption of neutral stability conditions. The resulting benefit of considering stability effects is expressed as the potential AEP gain. The results of the first cases looked promising, showing potential AEP gains of 7.4%, 5.6%, and 9.2%. However, these cases consist of unrealistic wind conditions and were mostly intended to study the effects of different stability conditions on the resulting optimal layout. For example, it is found that it is more beneficial to reduce wake overlap for stable conditions than for unstable conditions, which results in stable wind directions playing a dominant role in the optimization process. Cases with semi-realistic wind conditions showed significantly lower potential AEP gains of 0.1% and 0.7%. Finally, a real case based on meteorological data from an offshore site in the Netherlands resulted in a potential AEP gain of 0.0%. It is concluded that the benefits of considering stability effects in the layout optimization process are likely to be insignificant. In most cases, the layout optimization under neutral stability conditions already optimizes for wind directions with stable conditions, as stable conditions tend to be more frequent in wind directions with high wind speeds. It is expected that there can be a small potential AEP gain in cases with unusual stability distributions that differ significantly from the described trend. However, even in such cases it is likely that the benefits are still very small.

This thesis research is set up to increase the knowledge on the influence of atmospheric stability on the Global Blockage Effect (GBE) of offshore wind farm flows, thereby aiming to improve the accuracy of energy yield calculations and predictions of offshore wind farms.

With an ever increasing demand for sustainable energy, limitations of current sustainable technologies are studied widely. In wind farms, the so-called wake effect provides the biggest limitation on wind farm total power output. Using wind from the unaffected boundary layer to re-energize the wind flow in the wake provides a method of limiting this wake effect. In this study, kites are introduced to steer the wind flow of the unaffected boundary layer into the wake through a downwash velocity. RANS (Reynolds-averaged Navier-Stokes) simulations are performed in Computational Fluid Dynamics (CFD) software OpenFOAM of the atmospheric boundary layer (1), a small four-turbine wind farm (2) and a wind farm with static kites between the turbines (3). The turbines are modelled through the actuator disc approach, and kites are introduced through the more complex actuator line method. Results of the athmospheric boundary layer (ABL) and wind farm simulations correspond well with literature. Through extensive kite parameter studies, an optimal layout of kites in the wind farm is presented yielding a wind farm efficiency increase of 2.3 %, which increases over 5% for even larger kites. Kite size and the kite’s downstream location show to impact the re-energising levels of the wake flow the most. The kites generate a downwash wake instead of a single downwash velocity, a finding that should further be studied in future research.

Operating in real-world conditions, modern large capacity wind turbines often experience off-design situations, enduring dynamic loads characterized by complex unsteady aerodynamics. Key among the challenges in predicting these dynamic loads is understanding the effects of wind shear and turbulence, both individually and in their complex interplay. This research aims to shed light on these phenomena, with an emphasis on their impacts on wind turbine fatigue loads and power production. The research first provides an in-depth analysis of the influence of atmospheric stability on wind shear profile, aiming to extend the wind shear profile beyond the range of LiDAR measurements. Recognizing the limitations of existing power law and logarithmic law extrapolation methods, the study validates the use of multiple stability correction functions for accurate wind speed extrapolation. Subsequently, the research delves into the intricate effects of wind shear and turbulence on fatigue loads at the blade root of wind turbines, leveraging aeroelastic simulations. This research addresses the challenge of assessing wind turbine suitability for sites where one or several wind climate parameters surpass their design class values. It investigates the potential of the Response Surface Methodology (RSM) to estimate site-specific fatigue loads, a process that conventionally requires extensive aeroelastic simulations. This research also extends the scope to include the assessment of site-specific wind turbine power curves, validating the use of the Rotor Equivalent Wind Speed (REWS) and turbulence renormalization methods. Both methods show promise in estimating site-specific wind turbine power curves using a power curve measured under varying wind conditions. In essence, this study emphasizes the significant impact of wind shear and turbulence on the performance and longevity of wind turbines. By shedding the light on potential improvements, this study hopes to contribute towards accurate power output and fatigue load assessments.

The present work studies the impact of self-induced Atmospheric Gravity Waves (AGWs) excited by an moderately sized offshore wind farm immersed in a Conventionally Neutral Boundary Layer (CNBL). A wind farm of finite span-wise length, consisting of 25 NREL 15MW wind turbines laid out in (5 X 5 Aligned manner) is considered. In order to achieve the same, two customized open source CFD RANS solvers are used, based on the OpenFOAM and SOWFA environments. The study primarily focuses on the impact of thermal stratification in the Free Atmosphere, the strength of the capping inversion and height of the same on the characteristics of the AGWs excited. Further, the work also evaluates the impact of the excited AGWs on the wind farm in the form of velocity deficits at each individual turbine column and also power down the analysis. The study finds that for a higher capping inversion strength, the AGWs excited by the wind farms are trapped and propagate horizontally along the capping inversion along with its vertically propagating counterpart. These cause velocity fluctuations at the hub height of the turbine, which results in AGW induced wind farm blockage that lowers the amount of incoming kinetic energy at the first turbine row and an increased velocity at the last two rows. Furthermore, the trapped AGWs cause the collective wake of the wind farm to recover faster. Moreover, the study proves that the height of the capping inversion determines the extent of AGW excitation by the wind farm. It was found lower capping inversion heights led to a higher AGW excitation and higher AGW induced flow blockage and wake recovery. Lastly, the study also finds that the thermal stratification in the Free Atmosphere determines the vertical wavelength of the AGWs propagating in the free atmosphere. Furthermore, the study also reports that for a higher Free Atmosphere stratification, a lower AGW induced blockage is observed.

Measurement campaigns and CFD simulations have recently identified a large-scale flow phenomenon called wind-farm flow blockage. This is found to bear a significant and far-reaching reduction in wind speed upstream of a wind farm. The wind farm blockage is attributed to the cumulative induction effects of multiple wind turbines placed in series. Wind-farm flow blockage has important consequences on energy production because it reduces the available kinetic energy in the incoming wind flow. In turn, this causes leading wind turbines in a wind farm to produce less energy than they each would in isolation. To date, the physics of this global blockage effect is not entirely understood, and they are therefore an active research topic. Due to the increasing demand for wind energy, reducing annual energy production (AEP) uncertainties and power production bias seems to be a challenge for wind energy researchers. Understanding wind farm blockage in complex terrain becomes crucial to account for uncertainties and power production bias. This thesis set out to perform Reynolds-Averaged Navier-Stokes (RANS) simulations to assess the impact of wind-farm flow blockage in complex terrain using the open-source software OpenFOAM. A laterally infinite row of turbines is simulated on top of a 2-D hill defined by the mathematical curve ’Witch of Agnesi’. The set of simulations is performed for varying atmospheric conditions: truly neutral and stable free atmospheric conditions. Thermal stratification imposed under stable conditions is of particular interest due to the excitation of atmospheric gravity waves (AGWs) by the turbine array and the topology. The velocity fields due to the presence of the turbine array on top of the hill are compared to the ones without. The resulting flow reduction is then compared to the cases without the hill in order to assess the impact of complex terrain on wind farm blockage. A series of sensitivity analyses are performed for varying inter-array spacing and hill size variations in order to further the understanding of wind farm blockage. The results obtained in this study show that the magnitude of wind farm blockage is amplified due to the presence of the hill. Additionally, the excitation of AGWs is seen to have a major impact on the wind farm blockage due to alterations caused to the pressure field. The impact of blockage is seen to be dominant up to at least 10-15 turbine diameters upstream of the turbine array under truly neutral conditions. While the effects are more pronounced and much more dominant under stable free atmosphere conditions. All the stable free atmosphere cases simulated show a reduction ranging from 1-4% at different upstream locations while neutral cases show slightly lower yet non-negligible reduction due to blockage. This study ultimately concludes that the existing ’wakes-only’ approach for estimating energy losses still has a significant power production bias. Therefore accounting for the blockage effects in the farm upstream is also equally important and must be analysed before commissioning a wind farm.