Wind direction uncertainty is often overlooked in wind resource assessment, despite its potential to cause significant deviations in annual energy production (AEP). This study quantifies directional uncertainty arising solely from wind data, decomposed into three components: measurement uncertainty, long-term wind rose difference (LTWRD), and interannual variability (IAV). Two offshore sites with contrasting layouts and inflow regimes were analysed using statistical and resampling-based approaches.

The results show total directional uncertainties of 0.71% and 4.39% of AEP, with LTWRD emerging as the dominant contributor at the more directionally sensitive site. No consistent pattern in the composition of directional uncertainty was observed across sites, highlighting the need for site-specific assessments.

Given that typical total AEP uncertainties for wind projects average 8–10%, the findings indicate that directional uncertainty can represent a substantial fraction of the total, especially at sites with strong wake sensitivity and unidirectional inflow. Therefore, wind direction should be explicitly included in uncertainty analyses, and long-term correction procedures should address wind direction as well as wind speed.

This thesis investigates the potential of data-driven modelling to enhance wind farm monitoring, control, and operations and maintenance (O&M) strategies by leveraging the extensive SCADA data generated by offshore wind farms. Focusing on the Lillgrund offshore wind farm, the study develops regression-based surrogate models using XGBoost, artificial neural networks, and Gaussian process regression to predict two critical targets: the blade root flapwise and tower bottom fore-aft damage equivalent loads. All models are found to effectively capture the underlying patterns from the input features. Among them, XGBoost consistently outperforms the others in terms of prediction accuracy, computational efficiency, and robustness. Its superior performance is further validated across multiple preprocessing settings and operational scenarios, including its capacity to generalise across turbines and exploit spatial information. Finally, the applicability of the models is demonstrated through a simplified use case that estimates fatigue damage over a specific period. The findings underline the value of integrating machine learning-based surrogate models into operational workflows to reduce O&M costs and support decision-making in wind farm management.

Accurate modelling of wind flow in complex terrain remains a significant challenge in wind resource assessment. Traditional linear models, such as those used in Wind Atlas Analysis and Application Program (WAsP), often fail to capture non-linear effects like recirculation, separation, and stability-driven phenomena typical of steep or mountainous sites. Computational Fluid Dynamics (CFD) methods based on Reynolds-Averaged Navier–Stokes (RANS) equations offer improved accuracy but must be carefully verified and validated for reliability. This thesis evaluates the predictive accuracy of steady-state RANS simulations using PyWakeEllipSys against field measurements from the Perdigão campaign, characterised by complex double-ridge terrain. Three atmospheric stability regimes (stable, neutral, unstable) were simulated, employing various turbulence closures, including standard k–ε and Monin–Obukhov-based models (k–ε–MO). Grid convergence studies ensured robust simulation accuracy at turbine-relevant heights. Results indicate that unstable conditions are modelled most effectively, particularly in predicting terrain-induced speed-up and turbulence intensity profiles. Stable conditions were reasonably well captured in turbulence intensity and flow patterns but showed consistent underprediction of speedup due to overly persistent recirculation zones. Neutral conditions exhibited inconsistent accuracy across all metrics. Wind direction variability, especially bimodal flow patterns observed in the valley, was not captured by steady-state RANS, highlighting limitations in representing time-dependent, thermally driven flow mechanisms. The outcomes reinforce that steady-state RANS simulations, particularly when stability-adjusted turbulence models are employed, provide strong predictive capabilities for wind resource assessments in complex terrain, although inherent limitations related to transient phenomena must be acknowledged.

The recent growth of the size of wind farms highlights the need for a deeper understanding of the mesoscale phenomena in a stably stratified atmosphere, such as atmospheric gravity waves, as the effect on the power generation can be significant. This thesis is a numerical study of the effect of wind farm layout on atmospheric gravity wave excitation and the resulting feedback on wind farm performance. Wind farms in this study have varying power density, streamwise or spanwise turbine spacing, aspect ratio, hub height, rotor diameter, shape, or orientation with respect to the freestream, and are situated in a conventionally neutral boundary layer in offshore conditions. Moreover, the wind farms can be horizontally or vertically staggered. To investigate the atmospheric gravity wave excitation, the effect of wind farm layout on the Froude number and inversion Froude number governing the internal and interfacial waves respectively is studied using several high-fidelity Large Eddy Simulations. Then, AGW wavelength and wind farm efficiency are parametrically studied using a fast reduced-order model. Specifically, the non-local efficiency is considered, which is a measure of the global blockage effect induced by the atmospheric gravity waves. It is found that the length scale used in the Froude number must be adjusted to the wind farm, and is dependent on the turbine spacing and farm shape. The inversion Froude number is based on the phase speed of the interfacial waves. It is suggested that the phase speed must be based on shallow-water theory and deep-water theory for small and large turbine spacings respectively. In other words, for small turbine spacings the wind farm acts as an entity, while for large turbine spacings, the farm acts as a collection of individual turbines. Finally, the streamwise and spanwise turbine spacing (and consequently the power density), and the aspect ratio primarily govern the non-local efficiency.

Offshore wind is central to a decarbonized future, but its high costs remain a challenge. This research investigates the drivers of turbine sizing and their impact on wind farm performance using a Multi-Disciplinary Design Analysis and Optimization (MDAO) framework. Rated power and rotor diameter are explored as key design variables, with implications for both costs and energy production. Case studies for the North Sea show that the optimal size for minimizing the Levelized Cost of Electricity (LCoE) is around 15.5 MW with a 230 m rotor. Sensitivity analyses reveal wind speed and farm density as dominant drivers, while many modeling assumptions shift results along directions where LCoE is relatively flat. Extending the framework to revenue-based objectives shows that market-price correlations influence optimal specific power, though gains are limited. For hydrogen production, the high cost of electrolysis favors low-specific-power turbines. Overall, the findings indicate that continued turbine upscaling offers marginal benefits, highlighting the value of design standardization.

Dunkelflaute, meaning "dark doldrums" in German, denotes prolonged periods characterized by low wind and solar energy production, attributed to overcast skies and calm weather conditions. With wind and solar power assuming increasingly crucial roles within the European energy landscape, these extreme weather conditions (Dunkelflaute events) pose a significant challenge to grid stability. Addressing this challenge in power production, the objective of this dissertation is to comprehensively analyze Dunkelflaute events and devise both physical and machine learning-based methodologies for their prediction. The research goal is approached through three key aspects: 1) conducting a statistical analysis of the frequency, duration, seasonal variations, and associated weather patterns of Dunkelflaute events to gain insights into their impact and underlying characteristics; 2) developing diverse strategies for the identification and prediction of these events from both a modelling and data availability standpoint to enhance predictability; and 3) projecting future weather patterns under climate change scenarios and investigate the impact of climate changes on this extreme weather....

Renewable energy, particularly wind power, plays a vital role in combating global warming and air pollution. However, the inherent variability of wind complicates turbine performance predictions, especially for offshore wind farms. Standardised power curve models often fail to account for site-specific conditions, leading to performance deviations.

This thesis addresses these limitations by developing a new method to correct power curve predictions using detailed, site-specific data. Nacelle-mounted LiDAR and SCADA data from RWE’s offshore wind farms are utilised to enhance performance assessments during the development phase. Existing methods only rely on turbulence intensity (TI) to explain performance deviations, but these methods are based on smaller, onshore turbines and do not adequately account for the complexities of larger offshore turbines. The research introduces power performance matrices that incorporate additional wind characteristics such as shear, veer, and yaw misalignment, rather than relying solely on TI. The study finds that while LiDAR is less effective in measuring TI, it excels at capturing shear and veer, both of which show consistent patterns in relation to turbine performance. The findings reveal that extreme shear values negatively affect power production, while neutral or low shear has positive effects. Similarly, high veer correlates with decreased performance. 

Based on these insights, new power performance matrices were developed, using shear, veer, and TI, along with normalised wind speed, to adjust the power curve. These matrices can be dynamically created using site-specific data, making them flexible and adaptable for different turbines and wind conditions. The method enables selecting “inner range” parameters, providing a more accurate reference for power deviations. Additionally, this approach is particularly advantageous because shear and veer are well-captured by both floating and nacelle LiDAR, making it applicable across various stages of wind farm development. The primary limitation of this method is its reduced effectiveness in regions with significant mechanical atmospheric mixing, such as near coastal or mountainous terrain. In such areas, local topography can create complex wind flows, requiring more parameters than shear and veer to fully capture the wind dynamics. Another limitation is wake analysis; despite directional filtering to remove wake effects, some residual wake impact persists, suggesting further refinement is needed. 

Although more data is necessary to fully validate the approach, the current methodology shows promise in enhancing energy yield estimations and improving long-term predictions. The use of site-specific power matrices, particularly for offshore turbines, could significantly increase the accuracy of performance predictions, benefiting wind farm developers and manufacturers. As more data from LiDAR measurements and power performance tests become available, the matrices can be adapted to different turbine types and conditions by normalising wind speeds and creating dynamic “inner ranges” for site-specific use. Future research should focus on refining wake filtering methods and understanding the full correlation between shear, veer, and TI to further enhance prediction accuracy. In conclusion, this thesis proposed a novel method for adjusting turbine power predictions using site-specific wind characteristics, with potential to greatly improve the reliability of performance estimates. By integrating shear and veer into the power curve adjustments, this approach allows more reliable performance estimations. With further data collection, this methodology holds the potential to become a key tool for optimising wind energy production in the future.

The ever-increasing need for improving the energy yield in wind farms while minimising fatigue loads has created the need for exploring new concepts in the wind sector. One is the notion of wind farms with turbines of varying hub heights, often called vertically- staggered (VS) configurations. Multiple studies have demonstrated the potential benefits of such arrangements, yet the effect of erecting larger-scale wind turbines into an already existing control wind farm has not been explored. This concept has gained remarkable industrial attention, particularly in the German onshore market, as it simplifies the grid connection and land acquisition, among other benefits. Understanding the implications of vertical staggering on turbine performance is therefore essential. In addition, fast and accurate modelling of the inflow conditions modulated by the atmospheric boundary layer (ABL), is crucial in predicting the flow properties and energy yield in wind farms. A novel steady-state, RANS-based inflow model has recently been proposed, yet it lacks detailed validation. The present research aims to assess the reliability of that model in large wind farm flow simulations and subsequently apply it to identify flow patterns and power production characteristics in VS configurations representative of the German onshore wind standards through numerical simulations. RANS modelling combined with the actuator disk method in the PyWakeEllipSys software was employed to demonstrate that despite the inability of that inflow model to yield accurate prediction of the turbulence intensity, the computed velocity field is in fair agreement with LES results in conventionally neutral boundary layer conditions. This is evident, especially for deeper ABL cases. This study argues for the complexity of the flow caused by the collocation of different turbine scales in the investigated VS configurations. It is also highlighted that rotor overlap negatively impacts the power production of both turbine types in VS wind farms. Hence, it is suggested that optimal performance in VS setups can be achieved by minimising the rotor vertical overlap through appropriate adjustments of the hub height.

In recent years, the offshore wind energy industry has been facing significant barriers, ranging from supply chain disruptions to financial challenges due to rising commodity prices and increasing inflation. To keep offshore wind energy competitive, more efforts towards reducing costs and accelerating project timelines are needed. The optimization of load calculation processes, which are central services in the development of any wind energy project, would significantly contribute to the achievement of these goals.

Conventionally, load calculations are performed using high-fidelity numerical models that simulate the dynamic response of the turbine or wind farm. Although this approach is highly reliable, it is computationally intensive. For applications requiring a large number of load evaluations, this translates into high costs and long project timelines. Load response models have been developed as a data-driven alternative to the high-fidelity numerical models, offering sufficiently reliable load estimates at a cheaper computational cost. However, currently-existing load response models suffer from key limitations pertaining to their applicability to modern offshore wind farms.

This thesis aims to develop a load catalogue, consisting of a database of fatigue loads that can serve as a foundation for a load response model, that better represents the loading conditions in modern offshore wind farms. The database is constructed by running dynamic simulations for a variety of operating conditions tailored to represent those observed in offshore sites. In addition, the thesis seeks to use the catalogue to study the impact of employing the different wind and wake models recommended by the IEC 61400-1 standard on calculated loads. Specifically, it does so by comparing the 90th percentile and Weibull distribution turbulence intensity models, the Mann and Kaimal turbulence models, and the Frandsen and DWM wake models with respect to annual equivalent fatigue loads.

The analysis reveals that using a 90th percentile for the turbulence intensity overestimates the annual fatigue loads obtained when sampling the turbulence intensity from a Weibull distribution by an average of 20% for tower base loads and 6% for blade root loads. Moreover, it shows that the Kaimal and Mann turbulence models lead to significantly different fatigue load estimates. Furthermore, it finds that Frandsen's effective turbulence model consistently predicts higher loads than the DWM model for inter-turbine spacings smaller than 5 rotor diameters, but it can predict lower loads for inter-turbine spacings exceeding 5 rotor diameters.

The study concludes that using a 90th percentile turbulence intensity might lead to overly conservative designs, which can be avoided by using a Weibull distribution that accurately characterizes site turbulence. Also, it suggests that a combined modeling approach to long-term fatigue calculations that selectively toggles between the Kaimal and Mann turbulence models based on the short-term environmental conditions is expected to yield more accurate load estimates compared to employing a single model exclusively. Lastly, it implies that the effective turbulence model might not always be conservative, especially for wind farms with large inter-turbine spacings.

Ultimately, this research culminates in the delivery of a valuable instrument, the load catalogue, that enhances decision-making in the preliminary stages of project development and facilitates the development of better load response models. Additionally, it enriches the theoretical understanding of load calculation methodologies, paving the way for the development of better load calculation techniques.

This thesis addresses the critical issue of underestimated wake effects between neighboring windparks by developing efficient long-distance wind farm flow models using Convolutional NeuralNetworks (CNNs). The study compares three wake deficit models (Jensen, Bastankhah, andTurbOPark) and four neural network architectures (Convolutional Autoencoder (CAE), U-Net,CAE/MLP, and U-Net/MLP) to improve long-distance wake predictions.A novel method for random wind park layout generation was developed, simulating diversescenarios up to 768 rotor diameters downstream. Each wake model dataset comprised 1000simulations, split 80/20 for training and testing. Results demonstrate that all neural net-works effectively simulate wake datasets, with U-Net and U-Net/MLP consistently outperform-ing CAE approaches. Mean Absolute Errors (MAE) range from 4.75 × 10−4 m/s (Jensen) to1.44 × 10−2 m/s (TurbOPark). The U-Net/MLP model also successfully predicted turbulenceintensity, achieving MAEs between 1.30 × 10−4 (Frandsen model) and 2.21 × 10−4 (Crespo-Hernández model).Crucially, neural networks significantly outperform traditional engineering models in compu-tational efficiency. While engineering models’ computational time scales linearly with turbinecount, neural networks maintain a constant execution time of approximately 3 ms, regardlessof wind park size. This breakthrough enables rapid assessment of large-scale wind farm layoutsand performance optimization.

Climate change and wind energy are closely intertwined, with wind energy playing a vital role in reducing greenhouse gas emissions that drive climate change. However, climate change also poses a potential threat to the wind energy industry due to possible reductions in wind resources in key regions, though it may enhance resources in other areas. This study evaluates the impact of climate change on wind energy by using climate projections from Global Climate Models (GCMs). To improve resolution and capture finer details, downscaling methods—both statistical and dynamical—are employed.
This research investigates mean wind speed variations at nine global sites using non-downscaled CMIP6 (Coupled model intercomparison project phase 6) GCMs, statistically downscaled CMIP5/6 GCMs, and dynamically downscaled CMIP5 GCMs from CORDEX, under the climate change scenarios RCP4.5/SSP2-4.5 and RCP8.5/SSP5-8.5. It assesses impacts on wind resource availability, annual energy yield, sensitivity factor, and capacity factor, comparing these to historical data. The study also evaluates revenue from Annual Energy Production (AEP) and compares historical projections from GCMs with reanalysis data.
Findings indicate that non-downscaled CMIP6 GCMs and statistically downscaled GCMs exhibit similar trends in predicting wind speed decline at most sites, suggesting the effectiveness of non-downscaled CMIP6 GCMs for this purpose. CORDEX models reveal a significant influence of GCMs on mean wind speed projections, highlighting the need to include multiple GCMs for reliable analysis. The study observes larger decreases in AEP and capacity factors at onshore sites compared to offshore sites, attributed to complex terrain and higher sensitivity factors at the former.
Comparisons with reanalysis data show higher percent bias, mean absolute percentage error (MAPE), and lower correlation for onshore sites, due to their complex terrain. Additionally, no significant correlation was found between the distance of GCM grid points from reanalysis sites and MAPE. These discrepancies arise from differences in dataset resolution, topographical representation, and data assimilation processes, emphasizing the need for careful consideration when comparing GCM projections with reanalysis data.

Wind energy is one of the most important renewable energy sources, effectively addressing climate change issues and promoting sustainable development on a global scale. Blade failures may cause long shut-down times and may present a safety hazard. Continuous and real-time monitoring of the blade conditions is helpful for finding blade damage at an early stage and for predicting its development. Non-contact damage detection methods have the advantage of easy and flexible installation and deployment, especially for current in-service wind turbines. This thesis aims to investigate and develop a new non-contact method for wind turbine blade damage detection based on measurements of aerodynamic noise. The principle of the proposed method relies on the fact that damage to the blade may modify the boundary layer over the blade surface and the flow field around the blade, and, as a consequence, alter the noise generated aerodynamically. This noise propagates to the far-field and be measured by microphones, which could provide a remote way to detect blade damage. In this thesis, the detection of two types of damage, trailing edge crack and leading edge erosion, is experimentally investigated in the wind tunnel. The results show that the proposed aeroacoustics-based approach can effectively detect the damage mentioned above under some circumstances, which might be a promising solution complementing traditional damage detection methods in wind farms in the future.

The potential of utility-scale airborne wind energy (AWE) systems to contribute significantly to the energy transition hinges on their large-scale deployment, which depends on the cost-competitiveness and complementarity with conventional wind turbines. Central to the assessment of these metrics is understanding long-term energy production, which is influenced by the variability of wind profiles. This thesis investigates the significance of wind profile variability on annual energy production estimation for AWE systems. The study establishes the climatology of vertical wind profiles and expands flight operation models of AWE systems. By synthesising these aspects, a new energy production estimation framework is developed to incorporate variations in the wind profile shape. This framework is utilised to assess the impact of different wind profile shapes on the energy production estimation. The research underlines the need to move away from conventional wind energy calculation methods and offers a more suitable alternative for AWE systems. The framework offers a valuable tool for increasing the understanding of the viability of large-scale deployment of AWE systems.

In urban areas, vertical-axis wind turbines (VAWTs) show promise due to their omnidirectional design, addressing challenges faced by traditional horizontal-axis turbines (HAWTs). Despite significant progress in urban VAWTs, extensive multidisciplinary research is needed to optimise their efficiency and use in such environments.

This dissertation addresses this gap in four aspects. First, a low-fidelity noise model based on state-of-the-art literature is developed, allowing fast, acceptable, and accurate predictions for preliminary design stages of the primary noise sources on an urban VAWT. Then, a wind speed estimator and tip-speed ratio (WSE-TSR) tracking controller is designed to maximise the power production of an urban VAWT in turbulent wind conditions. This WSE-TSR tracking controller turned out to be an ill-posed problem, impacting the turbine and controller performance in the presence of model uncertainty. Follows the presentation of an approach that combines frequency-domain analysis and multi-objective optimisation, demonstrating its effectiveness in assessing and calibrating torque control strategies, thereby contradicting earlier assumptions and establishing new perspectives on performance optimisation for real-world wind turbines. Based on these collective findings, a decision-making framework is derived, capable of striking a balance between VAWT performance and noise acceptance, allowing for the first time to consider psychoacoustic annoyance as a metric.

In summary, this thesis contributes significantly to advancing the understanding of the complex dynamics of VAWTs, specifically focusing on human acoustic perception nearby, laying the groundwork for the successful integration of VAWTs into urban landscapes.

Routine maintenance is an essential requirement for the optimal functioning and longevity of any technical system that has been constructed. The issue occurs when the maintenance planning for such a structure becomes necessary. The adverse weather conditions prevalent in the Netherlands contribute to the heightened danger associated with the duty of a maintenance worker. Additionally, it is vital to comprehend the optimal time frame for minimising revenue losses when allocating time towards maintenance activities rather than operational tasks.

The objective of this project is to create a methodology for enhancing and improving the management of asset maintenance planning. This process is carried out by developing two statistical models. The primary objective of this study is to examine the feasibility of utilising API data from wind forecast sites to provide accurate production forecasts for individual wind turbines up to a 7-10 day period in advance. Furthermore, this study aims to forecast the electricity prices for the upcoming week by analysing historical data and recent day-ahead pricing. The outputs generated by these models are subsequently aggregated to yield a single outcome in terms of revenue.

The predictive model for electricity prices utilises data sourced from ENTSOE to generate an aggregate of electricity prices spanning the previous 7-8 years. This aggregate is subsequently adjusted by incorporating the electricity prices observed within the most recent three-week period. The model exhibits high accuracy in predicting day-ahead pricing during weekdays, but its performance is not consistently replicated on weekends.

The wind turbine output forecast model utilises operational archival data and high-resolution 10-day forecasts obtained from the European Centre for Medium-Range Weather Forecasts (ECMWF). A correlation has been established between the archival data and the historical turbine data provided by Green Trust Consultancy for the designated wind farm. The aforementioned correlation is subsequently employed to establish a connection between the output of a turbine and the real-time forecast data. The accuracy of the power generation forecast decreases from the initial day to the tenth day of the projection, resulting in inconsistent outcomes.

The combined outputs of these two models yield a solitary outcome that aids in predicting the potential revenue loss for the selected turbine during the period of maintenance-induced idleness. The limitations inherent in both models contribute to the generation of imprecise outcomes inside the revenue model pertaining to wind turbines. The model accurately predicts outcomes in two out of the four tested scenarios.

The objective of this thesis is to build an optimization algorithm with the aim of optimizing layouts for two objective functions - Annual Energy Production (AEP) and Levelized Cost of Energy (LCoE), for large offshore wind farms. The algorithm considers the four main factors that are taken into account when creating a preliminary system design for an offshore wind farm. They are - the geographical location of the turbines, the hub height of the turbines, the type of the turbine, and the total number of turbines in the design space.
The annual energy production (AEP) of the wind farm is calculated using PyWake which uses the simple NOJDeficit wake model combined with the required superposition and blockage models to resolve wind turbine wakes. This AEP is then fed into TOPFARM, an economic solver developed at DTU which uses scaling factors to derive the total cost of the wind farm. Constant factors such as the discount rate, distance from shore, foundation type, and drivetrain type are also considered when deriving the total cost of the wind farm.
The results of this process are used to determine whether a system design is better than another. Several constraints are applied when changing each optimization variable to keep each iteration as realistic as possible. The boundary is assumed to be a square. Both algorithms arrive at similar results, with random search providing a much better solution with approximately a 40\% reduction in LCoE.

The aim of this Design Synthesis Exercise was to design a floating large-scale wind farm of 1 GW in deep water using an Airborne Wind Energy System (AWES) that is cost-competitive, largely recyclable and uses less material than conventional wind turbines. By exploring the project foundation, carrying out an iterative single-system design process, and delving into the farm layout and management, this project assesses the feasibility of this novel concept. This report proposes an initial design with the resources currently available to the team while highlighting the next steps that need to be taken in order to pursue further design iterations, prototyping, and testing of this concept. This initial design, the tool developed to carry out the sizing, and a detailed reflection on the limitations of this design process and concept could be stated as the main contribution of this project to the field of airborne wind energy.

Measure Correlate Predict (MCP) is a method used to characterize the future wind resource at potential new wind farm locations. It finds a relationship between the wind data obtained at the target site and a nearby reference site over a short time period. This relationship is applied to long-term historical data from the reference site in order to predict the long-term wind resource at the potential wind farm location. Over the last few decades, new MCP techniques have been proposed, and modelled meteorological data in the form of reanalysis products have emerged. However, little research has been done regarding how the application of modelled data and different regression techniques in MCP affect the achievable wind resource prediction accuracy.

This project set out to investigate the accuracy of the MCP procedure under different configurations. Three comparative studies have been done in this project. Firstly, the attainable accuracy with either nearby MET-station data or ERA5 reanalysis data as a long-term reference source is assessed. Secondly, the use of different regression types for forming the relationship between target and reference data in the MCP procedure is assessed. Lastly, the accuracy achieved with standard MCP is compared to that achieved with a new wind resource estimation method, the method of analogs. The accuracy of the different configurations is assessed through the ability to accurately predict a period of wind speed values measured at 35 sites located in different terrain types. The predictions are evaluated using metrics such as the coefficient of determination, the root mean square error, the mean absolute error and the mean bias error.

This study found that ERA5 reanalysis data can serve as a reliable alternative to observed MET-station data. Generally, using ERA5 reanalysis data as a reference source always led to more accurate predictions then a MET-station reference source if the Pearson correlation between target and MET-station is lower than 0.8, and for offshore targets. If the Pearson correlation between target and MET-station reference is higher than 0.9, the achieved accuracy with either the MET-station or ERA5 data as a long-term reference is similar and depends on specific site conditions. In terms of regression methods it was found that the Matrix method, using the target site sectors for determining the regression parameters, generally outperforms other regression methods in terms of accuracy when determining the mean wind speed. Lastly, the method of analogs, a recently developed wind speed estimation method, yielded a similar prediction accuracy to standard MCP.

It should be noted that the different regression methods in employed in MCP all exhibited very similar prediction outcomes, with an average absolute difference in the predicted mean wind speed between the best and worst performing regression methods of only 0.046 m/s. Furthermore, the performance of the method of analogs in terms of accuracy improves with a longer concurrent period during which the relationships are formed. This project employed relatively short concurrent periods for certain targets, which may have contributed to sub-optimal performances of the method of analogs.

In the light of global need for renewable energy, wind energy plays a crucial role. Today, the majority of wind turbines are still being built on land. Given this critical role, accurate monitoring methods are needed to fully understand the performance within wind farms. This thesis aims to create a methodology to evaluate the power performance of wind turbines within an onshore wind farm using SCADA data. The primary objective is to identify deviations from the expected power generation patterns by using a multivariate machine learning approach. Due to the terrain complexity of onshore wind farms, a cluster-based approach is used. In total, 15 clusters consisting of 61 turbines have been evaluated. The training data is filtered by applying multiple filters with the objective of creating a normal behaviour model. The evaluation is based on a feedforward multilayer perceptron regressor to predict the power output for a test data set. A sequential methodology is explored to refine the model performance and is then applied to power performance analysis tasks to detect anomalies in the data sets. The results of the final analysis suggest that the approach taken is capable of detecting anomalies in the data. It is shown that in multiple clusters cases of under- or overperformance can be detected.

With the growing global utilisation of wind energy, lowering of its levelised cost of energy is pursued. This effort is hindered by wind turbine wakes and their detrimental effects on wind farm profitability. Most currently researched wake mitigation methods are based on wind farm system-level interventions, and there is a research gap on individual wind turbine design wake alleviation measures.

This thesis investigates the wake-diffusion rotor concept, a wind turbine design with an outboard shifted thrust distribution along the blade, and its effects on wake mitigation and power generation in wind farms.
Three wind turbine rotors with equivalent thrust coefficient are modelled as actuator disks in a RANS CFD software PyWakeEllipSys, corresponding to an inboard shifted thrust distribution, a conventional thrust distribution, and an outboard shifted thrust distribution representing the wake-diffusion rotor concept. Investigated scenarios include a single wind turbine and rows of three and ten wind turbines aligned to the inflow.
Compared to the baseline case, outboard thrust distribution produces lower wake velocity deficits, increased momentum transfer and increased turbulence generation. The wake-diffusion rotor concept increases the wind farm power generation by 3.8 % and 2.5 % in the three and ten wind turbine row scenarios respectively, with the largest relative gain in wind turbine power of 13.9 % for the second wind turbine in the row. Consecutive wind turbines in the row experience diminishing gains. The inboard shifted thrust distribution had opposite effects for all of these aspects.
The benefits in wake mitigation and power generation in wind farms from using the wake-diffusion rotor concept are concluded to come from higher turbulence mixing caused by the outboard shifted thrust distribution. Recommendations on its use for maximum effectiveness are given. Future research could focus on the design implementation of this concept, investigate combinations with other wake mitigation methods and perform parametric wind farm AEP studies.