Table of contents

This Special Issue presents selected papers from EERA DeepWind'2020, 15-17 January, Trondheim, Norway. This was the 17th Deep Sea Offshore Wind R&D Conference. The conference is hosted by SINTEF and NTNU and organized in cooperation with the European Energy Research Alliance (EERA) joint programme on wind energy. A total of 32 papers are included addressing the research topics of the conference

▪ New turbine and generator technology

▪ Grid connection and power system integration

▪ Met-ocean condition

▪ Operation and maintenant

▪ Installation and sub-structures

▪ Wind farm optimization

▪ Experimental testing and validation

▪ Wind farm control systems

All papers presented in this special issue have gone through a careful peer-review organized by the international Scientific Committee of EERA DeepWind'2020:

Anaya-Lara, Olimpo, Strathclyde University

Badger, Jake, DTU

Berkhout, Volker, Fraunhofer IWES

Berge, Erik, MET

Bredmose, Henrik, DTU

Cutululis, Nicolaos, DTU

Eecen, Peter, ECN

Heggelund, Yngve, CMR

Kitzing, Lena, DTU

Kvamsdal, Trond, NTNU

Madsen, Peter Hauge, DTU

McKeever, Paul, ORE Catapult

Merz, Karl, SINTEF Energi

Munduate, Xabier, CENER

Muskulus, Michael, NTNU

Nielsen, Finn Gunnar, UiB

Nygaard, Tor Anders, IFE

Reuder, Joachim, UiB

Robertson, Amy, NREL

Sperstad, Iver Bakken, SINTEF Energi

Tande, John Olav, SINTEF Energi

Uhlen, Kjetil, NTNU

Van Bussel, Gerard, TU Delft

Van Wingerde, Arno, Fraunhofer rWES

Økland, Ole David, SINTEF

Acknowledgements

The financial support from the Research Council of Norway covering a share of the cost of organizing the conference is greatly appreciated. The conference chairs and guest editors of this Special Issue like to express our sincere gratitude to the Scientific Committee of EERA DeepWind'2020, all authors that have contributed to this Special Issue, and to colleagues that have so kindly contributed to the review process. We hope that this Special Issue provides results that contribute to the fruitful development of offshore wind energy. Details of the DeepWind2020 conference and all conference presentations and posters are available at the conference web page.

Conference chairs and guest editors:

John Olav Giæver Tande, Chief Scientist, SINTEF Energi

Trond Kvamsdal, Professor, NTNU

Michael Muskulus, Professor, NTNU

All papers published in this volume of Journal of Physics: Conference Series have been peer reviewed through processes administered by the Editors. Reviews were conducted by expert referees to the professional and scientific standards expected of a proceedings journal published by IOP Publishing.

□ Type of peer review: Single-blind / Double-blind / Triple-blind / Open / Other (please describe) Single-blind

□ Conference submission management system: by email: deepwind@sintef.no and Sharepoint

□ Number of submissions received: 36

□ Number of submissions sent for review: 36

□ Number of submissions accepted:32

□ Acceptance Rate (Number of Submissions Accepted / Number of Submissions Received X 100): 88,88

□ Average number of reviews per paper: 2

□ Total number of reviewers involved: 16

□ Any additional info on review process:

Contact person for queries: Overall conference chairs:

John Olav Giaever Tande, SINTEF Energi AS, john.o.tande@sintef.no,

Trond Kvamsdal, NTNU, trond.kvamsdal@ntnu.no

Michael Muskulus, NTNU, michael.muskulus@ntnu.no

Questions concerning adm:

Randi Aukan, SINTEF Energi AS, randi.aukan@sintef.no

Harsh environmental conditions, time pressure due to tightly calculated maintenance windows, complex technical systems and high quality standards determine the working environment for the maintenance and repair of offshore wind farms. Despite good qualifications and diverse training courses, maintenance technicians are reliant on information from manuals and checklists when performing their tasks and must document their work results in protocols. Both the supply of information and the data acquisition show a multitude of media disruptions. Hence, the processes on site and in the postprocessing are carried out slowly and uncertainly. The assistance system presented is intended to contribute to a consistently digital information flow and to support the technicians in a task-specific manner. Augmented Reality (AR) offers a way of presenting information and interacting with the assistance system during work execution. Using the example of the maintenance of crane systems used in a variety of ways in offshore wind farms, this article shows which potential can be tapped through digital assistance systems and AR applications.

Through a series of laboratory friction tests on small scale specimens an understanding of the physical behaviour is gained. The test program of the friction tests contained 12 test specimen variants, with a combination of dry air and seawater environment and various surface conditions and contact pressures.

Offshore projects, like the installation of offshore wind farms, consist of a number of different and often weather dependent activities. These tasks and their relations are defined in project schedules, which have to be assessed for weather effects before the project realization. Here, often Weather Window Statistics are used to calculate probable weather delay times using relative frequencies of weather windows. Another possible approach is the Weather Time Series Scheduling (WaTSS) method which combines the given project schedule, weather restrictions and historical weather time series data for several decades. The aim of this paper is a comparison of the two approaches especially when it comes to risk analysis for project schedules using percentile values or related risk measures. The same ERA5 model data is used both as basis for the Weather Window Statistics and the WaTSS method. We calculate weather downtimes for each task, as well as for the total project, and apply the risk measures Value at Risk (VaR), Conditional Value at Risk (cVaR) and Upper Partial Moments (UPM). This is followed by an analysis of the results from Weather Window Statistics and the WaTSS method considering their applicability and consistence. We obtain that risk measurement for project schedules under use of Weather Window Statistics is not always practicable because certain risk values tend to overestimation. WaTSS method provides project adjusted modeling and leads to more realistic risk measurement for complex project schedules.

This article provides a techno-economic study on coupled offshore wind farm and green hydrogen production via sea water electrolysis (OWF-H₂). Offshore wind energy, wind farms (OWF) and water electrolysis (WE) technologies are described. MHyWind (the tool used to perform simulations and optimisations of such plants) is presented, as well as the models of the main components in the study. Three case studies focus on offshore wind farms, either standalone or connected to the grid via export cables, coupled with a battery and electrolysis systems either offshore or onshore. Exhaustive searches and optimisations performed allowed for rules of thumb to be derived on the sizing of coupled OWF-H₂ plants, that minimize costs of hydrogen production (LCoH₂, in €/kgH₂): Non-connected OWF-H₂, coupled to a battery, offers the lowest LCoH₂, without the costs of H₂ transportation, when compared to cases where the WE is installed onshore and connected to the OWF. Using a simple power distribution heuristic, increasing the number of installed WE allows the system to take advantage of more OWF energy but doesn't improve plant efficiency, whereas a battery always does. Finally, within the scope of this study, it is observed that power ratios of optimized plant architectures (leading to the lowest LCoH₂) are between 0.8-0.9 for P_WE/P_OWF and 0.3-0.35 for P_Battery/P_OWF.

An accurate numerical simulation of the structural lifetime of offshore wind turbines is a challenging task due to several reasons. One of them is the uncertainty of met-ocean conditions acting on a turbine, e.g. wind and waves. This uncertainty can be divided into two kinds of uncertainty: aleatory and epistemic uncertainty. If both types of uncertainty occur, this is called polymorphic uncertainty. According to the state of the art, for met-ocean conditions, mainly aleatory uncertainty is considered or both types of uncertainty are modelled using a single probability density function. This leads to a simplification of the actual uncertainty, whose effect on the lifetime estimation has not been analysed so far. In that sense, in this work, the influence of various uncertainty models for met-ocean conditions on long-term damage equivalent loads (DELs) - representing the wind turbine fatigue lifetime - is investigated. For this purpose, different uncertainty models for met-ocean conditions are derived using real measurement data. Not only purely probabilistic models are applied, but imprecise probabilities - here interval random variables - as well. It is shown that the uncertainty models have a considerable influence on the fatigue life of offshore wind turbines. Especially, the large fatigue load intervals, which are determined, clarify the importance of a well-founded decisions regarding uncertainty modelling of met-ocean conditions.

Achieving performance optimisation and cost savings by advanced data analysis techniques and improved digital communication is a significant focus of wind farm operators and research organisations often framed under the terms IoT or Industry 4.0. Within the research project 'ModernWindABS', Fraunhofer IEE has conducted a twofold approach to identify new applications using modern methods that will innovate operation and maintenance processes. Based on a systematic structure of O&M processes and mathematic methods, innovative applications were identified and evaluated in expert workshops. Separately a survey among wind industry professionals with the focus on innovative applications through digital platforms has been conducted in partnership with German industry associations. From both approaches, applications, that enable failure risk monitoring, turn out as the highest priorities from the professionals consulted. Additionally, the survey results yield insights on participants' expectations on benefits and financial contribution per organisational role and preferences regarding the platform setup and points out gaps between these expectations and current platform service designs. It further identifies the main barriers to the broader use of platforms, of which organisational and legal obstacles seem to outweigh technical problems.

This paper presents an original contribution to the coupled hydroelastic simulation of offshore wind turbines foundations. A non-linear coupled simulation tool has been developed to calculate the response of these systems. The hydrodynamic loads are computed by the non-linear potential flow solver WSCN, developed at Centrale Nantes and based on the weak-scatterer theory. The response of the structure is calculated with a finite element analysis tool using an Euler-Bernoulli beam model. The two solvers are tightly coupled. A first case study is presented here. It consists of a uniform monopile foundation in regular waves. Numerical simulations are compared to results obtained with the SIMA software, which uses Morison equation and non-linear wave kinematics. The two simulation tools show a very good agreement in a series of regular waves up to the third-order load harmonic. Higher-order harmonics appear with WSCN in steep waves, which are not computed by SIMA.

The share of wind energy in the public electricity supply in Europe is constantly growing, so that reliability and availability of wind turbines are becoming increasingly important. High availability is ensured by continuous condition monitoring, because it allows long downtimes to be avoided by reacting immediately to (imminent) failures. This paper presents a portable and real-time capable simulation model for determining internal loads on components in the mechanical drivetrain. The loads are apt for being utilized for a subsequent generation of a system reliability index in the scope of a decision support tool for demand- and degradation-oriented adjustments of the operational management. Thus, the work contributes to ensuring reliability. By determining the state of degradation of similarly loaded plants, under- or overloading of individual turbines can be proactively prevented. The analysis of the influence of individual model parameters is of particular importance here and provides evidence that even with a restricted parameter set the outputs of the load calculation model are accurate enough as inputs for a reliability calculation.

Spar-buoy floating wind turbines (FWTs) have been deployed at full-scale off the coast of Scotland. However, their deep draught restricts their wider-suitability for assembly and installation at shallow water ports. Here a barge-type installation vessel is investigated experimentally for supporting a FWT such that a draught reduction of up to 30% can be achieved. Tests in irregular waves are conducted to understand the design loads required of a mechanical linkage between the vessel and FWT. Overturning moments increase by over double for a doubling in significant wave height, H_s, and peak overturning moment occurs at the pitch eigenfrequency of the combined vessel and FWT. For H_s=1.5 m, these loads can be accommodated for with a steel truss-frame structure. Collisions between vessel and FWT are also tested for regular head waves and show that, whilst a number of collision forces exceed those of the relevant design standards, these could be reduced to within the existing design limits by use of a fender.

One of the challenges related to the design of floating wind turbines (FWTs) is the strong interactions between the controller and the support structure, which may result in an unstable system. Several control strategies have been proposed to improve the dynamic behaviour, all of which result in trade-offs between structural loads, rotor speed variation, and blade pitch actuator use, which makes controller design a challenging task. Due to the interactions, simultaneous design of the controller and support structure should be performed to properly identify and compare different solutions. In the present work, integrated design optimization of the blade-pitch controller and support structure is performed for a 10 MW spar FWT, considering four different control strategies, to evaluate the effect of the controller on the structural design and associated costs. The introduction of velocity feedback control reduces the platform pitch response and consequently the fatigue loads in the tower, which leads to a decrease in the tower costs compared to a simple PI controller. Low-pass filtering of the nacelle velocity signal to remove the wave-frequency components results in reduced rotor speed variation, but offers only small improvements in costs, likely due to the limited wave-frequency response for the considered designs. Comparisons with nonlinear time-domain simulations show that the linearized model is able to capture trends with acceptable accuracy, but that significant overpredictions may occur for the platform pitch response.

In the EU H2020 project LIFES50+, a 1:36 scaled model test campaign was carried out for the NAUTILUS-DTU10 semi-submersible floating offshore wind turbine with active ballast. This paper concentrates on the modelling capabilities of a state-of-the-art time domain simulation model FAST8 for floating offshore wind turbines and specific challenges associated with the validation process. For the modelling, the platform is considered as a rigid body, and the frequency dependant radiation damping, added mass, and wave excitation are evaluated with a panel code using potential theory. The results from the scaled model are compared to the simulations, looking into the effects on the platform when the first order radiation-diffraction hydrodynamics through the Cummins equation, the mean drift coefficients from the nearfield solution for Newman's approximation of the second-order difference-frequency wave forces, and full quadratic transfer function (QTF) are taken into account. Moreover, viscous forces on the floating platform are modelled through Morison elements with coefficients of drag. Frequency domain analysis of the motions showed good agreement after modifications of the coefficients of drag of the Morison elements for tests performed with a pink wave spectrum. On the other hand, the extreme wave test showed large discrepancies between results and simulations, which could not be overcome by the inclusion of the QTFs or tuning of the coefficients of drag which model the viscous forces. This is followed by a discussion of the challenges in the modelling approach, and other validation techniques are proposed for future research. The main goal is to define the Morison elements for the floater, and tune the drag coefficients to enable the numerical tool to capture the floater's motions. The effect of the change of the coefficients on the simulation outputs are shown.

Prestressed concrete is gaining growing interests as an alternative to steels for the construction material for floating platforms of wind turbines due to its low material costs. Investigation of the characteristics of the concrete structure under the fluctuating loadings are necessary to ensure the structural integrity, where a reference model of both the floating concrete platform and accompanying structural design are highly useful. In this study, a reference model for prestressed-concrete spar-type floater for 10 MW wind turbines is designed together with its concrete structure. Frequency-domain analysis showed that an optimum floater diameter that gives minimum pitch angle could be found. Although the maximum pitch angles were similar for floaters with diameters of 15 m to 17 m, structural design showed that the case with the wall thickness of 0.35 m will subject to prestressed-concrete steel yielding, concrete bending cracks, and shell buckling due to water pressure. Finally the reference model is designed for a floater diameter of 16 m and wall thickness of 0.45 m.

As the rotor diameter of offshore wind turbines increases, an improved understanding of the structure of the turbulent wind field approaching the rotor is important. Present standards for computing wind loads are resting upon assumptions of neutral atmospheric conditions and a simplistic formulation of the coherence of the turbulence. In the present work, various formulations of the wind field are applied and the dynamic responses of a bottom fixed and a floating wind turbine are computed to investigate the sensitivity to the formulation of the wind field. Focus is on wind situations with above average turbulence intensity as these are expected to have a significant contribution to fatigue damage of the structure. It is observed that choice of wind spectrum, coherence formulation as well as assumptions related to atmospheric stability conditions significantly influences the dynamic loads in tower bending moment, yaw moment and blade flap moment. The differences are significant in particular in the low frequency range. This implies that in particular floaters are sensitive to the formulation of the wind field.

Several approaches to avoid control-induced resonances of floating wind turbines have been proposed. The main focus has been on reductions in global motions and loads in the tower base. In the present work, we examine the consequences of three such controllers on the loads on drivetrain components. One common advantage that comes along implementing all the alternative controller designs is an improved motion response in surge and pitch directions. However, these reductions come at other costs. Evaluating the drivetrain performance through multi-body simulations identifies new considerations for controller design. For example, tower top shear stress may not have been perceived as an important design criteria from a structural load perspective, but contributes to the radial load of bearings and gears and should be taken into consideration when comparing controller performance.

In this study, numerical analysis of a tension leg platform wind turbine is conducted and the responses with focus on surge motions and tendon tension are compared with available experimental test data. The main scope of the study is to establish the numerical model for which the damping coefficients for rigid-body motions are tuned based on the comparison of the sway free decay test results (natural periods and damping ratios) between the numerical and the experimental studies. The differences between the test model properties and the numerical model information have been discussed. Numerical model tuning with available test data resulted with relatively good accordance but also slight to moderate differences in the responses. These differences are credited for the uncertainties in the model testing and the solution methodology of the numerical model. Numerical study is under development with regular and irregular wave analyses and analyses including wind excitation.

The calculation of the velocity deficit in the wake of individual wind turbines is a fundamental part of the wind farm analysis. A good approximation of the wake deficit behind a single wind turbine will improve the power estimation for downwind turbines. Large-eddy simulation (LES) is a research tool widely used in studying the velocity deficit and turbulence intensity in the wake. However, the computational cost of the LES prevents its application in wind farm performance analysis and control. Existing analytical wake models provide a fast estimation of the velocity deficit and the wake expansion rate downstream from the rotor. The Gaussian wake models use a Gaussian distribution to improve the prediction of the wake velocity deficit. With the number of analytical models available, an extensive evaluation of their performance under different flow parameters is needed. In this work, we simulate a wake of a single wind turbine using the LES code PALM (Parallelized LES Model) combined with an actuator disc model with rotation. We compare the computed flow field with the predictions made by Gaussian models and fit their parameters to obtain the best possible fit for the wake field data as computed by LES.

High variability of wind in the farm areas causes a drastic instability in the energy markets. Therefore, precise forecast of wind speed plays a key role in the optimal prediction of offshore wind power. In this study, we apply two deep learning models, i.e. Long Short-Term Memory (LSTM) and Nonlinear Autoregressive EXogenous input (NARX), for predicting wind speed over long-range of dependencies. We use a four-month-long wind speed/direction, air temperature, and atmospheric pressure time series (all recorded at 10 m height) from a meteorological mast (Vigra station) in the close vicinity of the Havsul-I offshore area near Ålesund, Norway. While both predictive methods could efficiently predict the wind speed, the LSTM with update generally outperforms the NARX. The NARX suffers from vanishing gradient issue and its performance declines by abrupt variability inherited in the input data during training phase. It is observed that this sensitivity will significantly decrease by integrating, for example, the wind direction at low frequencies in the learning process. Generally, the results showed that the predictive models are robust and accurate in short-term and somewhat long-term forecasting of wind.

In this paper, the primary objective is to investigate flow structures in the wake of wind turbines based on applying a truncated Proper Orthogonal Decomposition (POD) approach. This scheme decomposes the three-dimensional velocity fields produced by the high-fidelity PArallelized LES Model (PALM) into a number of orthogonal spatial modes and time-dependent weighting coefficients. PALM has been combined with an actuator disk model with rotation to incorporate the effects of a turbine array. The time-dependent deterministic weights from applying the POD scheme are replaced by stochastic weights, estimated from two independent stochastic techniques that aim to account for unresolved small-scale features for a number of POD modes. We then reconstruct the flow field by a small number of stochastic modes to investigate how well the applied stochastic methodologies can reproduce the flow field compared to the original LES results.

Floating offshore wind farms represent the next frontier in wind power industry. However, the development of this technology is strongly dependent on its economic feasibility. There follows that the development of economic analyses is crucial to highlight the possible greater potential of floating offshore wind farms and to support their sustainability and technical value.

In this context, the purpose of this paper is to present a sensitivity analysis of the main cost parameters for floating offshore wind farms, namely the distance from the coast, the distance from the closest port and the sea depth. It can give specific information on which parameters are more important, and how much they affect the total cost.

To this aim, a comprehensive life cycle cost assessment of floating offshore wind farms has been developed. In this study the cost model has been applied to the Italian waters.

The results shown should provide guidance on how to preliminary assess the quality of a given site for floating offshore wind farm installation, and should be helpful for future development of decision-making procedures in the offshore wind sector.

Ship-based profiling Lidar systems experience a strong influence of rotational and translational motion on beam direction and hence the line-of-sight velocity. This motion error is inherited by the retrieved 3-dimensional wind vector and is especially visible in the velocity spectra and cross-spectra of velocities at different measurement heights (coherence). Applying motion compensation on the line-of-sight velocity observations was found to have a strong impact on the statistical properties of the retrieved wind vector and successfully improved the corresponding velocity distributions. The impact of motion correction on the spectra of the horizontal wind speed components was found to be neglectable. The Lidar measurement principle, in particular the effect of cross-contamination at higher frequencies, was found to have a larger impact in shaping the horizontal spectra than motion correction. Vertical velocity spectra were strongly affected by ship motion and the motion correction was only partly successful. Precisely, this effect was present at frequencies larger than the resonance frequency of the ship.

The best practise for structural damage detection currently relies on the installation of structural health monitoring systems for the collection of dedicated high frequency measurements. Switching to the employment of the wind turbine's SCADA (Supervisory Control and Data Acquisition) signals and their commonly recorded low frequency statistics can lead to a reduction in the number of ad-hoc monitoring sensors and quantity of data required. In this paper, aero-hydro-servo-elastic simulations for a model of a turbine are used to assess its loads and any changes in the dynamics under healthy state and a damaged configuration case study. To prove the feasibility of the damage detection through low-resolution data, the statistics of the typically recorded signals from the SCADA and the structural monitoring systems are fed into a database for training and testing of classification algorithms. The ability of the machine learning models to generalise the classification for both stochasticity and uncertainties in the environmental conditions are tested. Decision tree-based classifiers showed the capability to capture the damage for the majority of the operating conditions considered. Though the setup of the traditional SCADA sensors had to be supplemented with an additional structural health monitoring sensor, the detection of the damage has been shown feasible by referring to low-frequency statistics only.

The development within the offshore wind sector towards more powerful turbines combined with increasing water depth for new wind parks is challenging both the designer as well as the manufacturer of bottom fixed support structures. Besides XL-monopiles, the market developed an innovative and economic jacket support structure which is based on automatically manufactured tubular joints combined with standardized pipes. Besides the improvements for a serial manufacturing process the automatically welded tubular joints show a great potential in terms of fatigue resistance e.g. due to a smooth weld geometry without sharp notches. However, these benefits are not considered yet within the fatigue design process of automatically manufactured jacket substructures according to current standards due to the lack of suitable S-N curves. Therefore, 32 axial fatigue tests on single and double-sided automatically welded tubular X-joints have been performed to determine a new hot spot stress related S-N curve. Based on these constant amplitude fatigue tests a new S-N curve equal to a FAT 126 curve was computed which implicitly includes the benefits of the automatically welding procedure.

For the simulation of the coupled dynamic response of floating offshore wind turbines, it is crucial to calibrate the hydrodynamic damping with experimental data. The aim of this work is to find a set of hydrodynamic drag coefficients for the semisubmersible platform of the Offshore Code Comparison Collaboration, Continuation, with Correlation and unCertainity (OC6) project which provides suitable results for an irregular sea state. Due to the complex interaction of several degrees of freedom (DOF), it is common to calibrate drag coefficients with the time series of decay tests. However, applying these drag coefficients for the simulation of an irregular sea state results in misprediction of the motions. By using numerical optimization, it is possible to calibrate multiple drag coefficients simultaneously and effectively, while also considering several DOF. This work considers time series of structural displacements from wave tank tests of the OC6 project and from simulations of the same load cases in OpenFAST. Results are transferred into the frequency domain and the deviation between power spectral densities of surge, pitch and heave from experiment and numerical simulation is used as an objective function to obtain the best fitting drag coefficients. This novel numerical optimization approach enables finding one set of drag coefficients for different load cases, which is a major improvement compared to decay-test-tuned drag coefficients.

Competition within the energy generation industry provides an incentive for developers to build offshore wind farms with a low levelised cost of energy. Therefore, there is a need for design optimisation to reduce costs and increase energy capture. A sequential approach to optimise turbine placement and cable layout is presented, using a heuristic k-opt algorithm and mixed-integer linear programming respectively. Energy storage is considered as a means to further improve the cable selection process. A case study is carried out on the Lillgrund offshore wind farm and the resulting layout improves energy capture by 6%. Cable costs are increased but the electrical losses are reduced such that there is an overall saving over the project lifetime of 20%. Energy storage as a means to peak shave the power seen by a cable in order to reduce electrical losses or de-rate a cable section was found to be impractically large and not profitable. Future work will consider secondary revenue streams to remedy this.

Floating LiDAR is a new aspiring technology for replacing offshore meteorological masts for applications like offshore site assessment or power curve measurement. Since the beginning of the development of the first Floating LiDAR Systems (FLS) in the late 2000s, the systems are more and more maturing. The focus of current research is moving from the pure technology development to the application of the obtained data. As a result, procedures for assuring the measurement quality and the assessment of the data quality - or the quantification of estimated measurement uncertainty, respectively - is a key aspect for further increasing the confidence in the data. This is an important pillar for a further distribution of the technology, and for decreasing risks and costs for its application in offshore wind.

In this paper, we introduce the methodologies of FLS verification and classification, with the goal to verify the measurement of individual FLS and study the characteristics of a certain type of LiDAR systems. A classification case study is presented, showing low effects of the buoy motion and sea states on the FLS measurement for the Fraunhofer LiDAR buoy. A detailed study of the interdependencies between influencing parameters and the possible correlations is presented as well.

The robustness and issues of the classification methodology - originally defined for fixed LiDAR systems - is discussed. Especially the non-robustness of the bin-fitting is identified as an issue. Alternative methods for the sensitivity analysis within the classification process are discussed.

The purpose of the study is to show the importance of wake meandering effect with regard to power production, in terms of annual energy production (AEP), and velocity deficit for wind turbines affected by wake meandering. A simplified economical and environmental analysis is also presented. A replica of Equinor's layout for Hywind Tampen wind farm, referred to as the original layout, is made and used as a basis for comparison. An alternative configuration is produced through a pragmatical optimization process, applying the SIMA-DIWA software, which reduces the effect of wake meandering on the downstream wind turbines, when the wind approaches from the most frequent direction. To show the aerodynamic effects of applying greater spacing between turbines, an enhanced version of Hywind Tampen is also presented. The main findings of the project correlate greater wake effect meandering for wind turbine configurations spread over a smaller area. The estimated annual energy production for the three configurations is 488.5 GWh for the original, 497.6 GWh for the up-scaled and 486.2 GWh for the alternative configuration, respectively. Furthermore, the optimization process revealed that a reduction in turbine distance can contribute to reduced wake effect in particular cases.

The vortex interaction in the wake behind a two- and three-bladed model scale wind turbine is investigated. The two rotors have equal solidity, and produce similar power and thrust at the design tip speed ratio. Phase-averaged quantities of the wake flow from one to four rotor diameters behind the turbines are measured in a wind tunnel. It is found that the two-bladed turbine has slower wake recovery than the three-bladed turbine, and a larger velocity deficit is produced in the far wake. The tip vortices behind the two-bladed turbine is more stable than behind the three-bladed turbine, and the vortex structures exist further downwind. In a wind farm, this could reduce the power production and increase fatigue loads for the turbines operating in the wake flow, especially during stable atmospheric conditions.

As the offshore wind industry moves toward larger wind turbines and deeper water, wave-induced loads on large-diameter monopiles are of increasing importance for ultimate limit state design checks. The combination of a relatively large diameter with steep waves in intermediate water depth presents challenges for numerical methods, and small-scale hydrodynamic testing of monopiles is therefore a necessary step in reducing the uncertainties in numerical analyses. Here, we aim to summarize the experimental observations in a new set of tests carried out with a flexible monopile wind turbine, and to understand the similarities and differences between these results and previous studies. Compared to previous studies, the present tests consider a larger monopile diameter and hub height, and include a larger number of realizations and repetitions. The distribution of extreme values and the contributions from different structural modes are studied. These experimental results provide insight into the physical effects which must be accurately captured by numerical tools that are used in design.

Atmospheric flows are governed by a broad variety of spatio-temporal scales, thus making real-time numerical modeling of such turbulent flows in complex terrain at high resolution computationally unmanageable. In this paper, we demonstrate a novel approach to address this issue through a combination of fast coarse scale physics based simulator and a family of advanced machine learning algorithm called the Generative Adversarial Networks. The physics-based simulator generates a coarse wind field in a real wind farm and then ESRGANs enhance the result to a much finer resolution. The method outperforms state of the art bicubic interpolation methods commonly utilized for this purpose.

Since operational managers often monitor large numbers of wind turbines (WTs), they depend on a toolset to provide them with highly condensed information to identify and prioritize low performing WTs or schedule preventive maintenance measures. Power curves are a frequently used tool to assess the performance of WTs. The power curve health value (HV) used in this work is supposed to detect power curve anomalies since small deviations in the power curve are not easy to identify. It evaluates deviations in the linear region of power curves by performing a principal component analysis. To calculate the HV, the standard deviation in direction of the second principal component of a reference data set is compared to the standard deviation of a combined data set consisting of the reference data and data of the evaluated period. This article examines the applicability of this HV for different purposes as well as its sensitivities and provides a modified HV approach to make it more robust and suitable for heterogeneous data sets. The modified HV was tested based on ENGIE's open data wind farm and data of on- and offshore WTs from the WInD-Pool. It proved to detect anomalies in the linear region of the power curve in a reliable and sensitive manner and was also eligible to detect long term power curve degradation. Also, about 7 % of all corrective maintenance measures were preceded by high HVs with a median alarm horizon of three days. Overall, the HV proved to be a promising tool for various applications.

Reduced basis methods (RB methods or RBMs) form one of the most promising techniques to deliver numerical solutions of parametrized PDEs in real-time with reasonable accuracy [1]. For the Navier-Stokes equation, RBMs based on stable velocity-pressure spaces do not generally inherit the stability of the high-fdelity method. Common techniques for working around this problem (e.g. [2]) have the effect of deteriorating the performance of the RBM in the performance-critical online stage.

We show how divergence-free reduced formulations eliminates this problem, producing RBMs that are faster by an order of magnitude or more in the online stage. This is most easily achieved using divergence-conforming compatible B-spline bases, using a transformation that can maintain the divergence-free property under variable geometries. See [3] for more details.

We also demonstrate the flexibility of RBMs for non-stationary flow problems using a problem with two stages: an initial, finite transient stage where the flow pattern settles from the initial data, followed by a terminal and infinite oscillatory stage characterized by vortex shedding. We show how an RBM whose data is only sourced from the terminal stage nevertheless can produce solutions that pass through the initial stage without critical problems (e.g. crashing, diverging or blowing up).

Splipy is a pure Python library for the creation, evaluation and manipulation of B-spline and NURBS geometries. It supports n-variate splines of any dimension, but emphasis is placed on the use of curves, surfaces and volumes. The library is designed primarily for analysis use, and therefore allows fine-grained control over many aspects which cannot be achieved with conventional CAD tools. It is packaged and distributed through Python Package Index: PyPi which gives easy installation.