Over the past decade, the offshore wind industry has experienced rapid growth due to significant enhancements in computational capacity. The technology provides larger wind turbines and efficient power production to meet global energy demand. As the size of the floating wind turbines increases, their response to wind and wave excitation changes. In addition to external loading, aerodynamic loading and multibody interactions contribute to the nonlinear dynamics that affect the global response of the wind turbine. Analyzing these structural vibrations is crucial for the environment and social acceptance because they are the primary source of underwater noise. Most of the energy of operational noise of floating wind turbines primarily originates from the rotation of the rotor, propagates through the tower, and is transmitted to the ocean via the platform vibration. Compared to gear-boxed technology, direct-drive radiates noise in lower frequencies due to the slower rotation of the rotor. However, due to the interactions between structure, wind, waves, and currents, the energy may be distributed in a wider frequency range, which can interfere with the hearing frequency of some underwater species by affecting their communication and navigation. In this study, we investigate the underwater noise originating from the structural vibration and vortex shedding using the open-source software OpenFOAM. We employ OpenFOAM’s libAcoustics to analyze noise radiation from a simplified circular cylinder model in uniform flow. For validation, we utilize a stationary cylinder and define the spatial and time resolution with consideration of literature guidelines. Once the resolution of the model is validated, we conduct the numerical simulation by applying a forced vibration to the cylinder in one direction. We sample pressure and velocity in the wake region to characterize the frequency components of vibration and vortex shedding by spectral analysis. Then, we calculate the sound pressure levels (SPLs) in the far-field using the hydrophone, which we defined using OpenFOAM’s acoustic library. Consequently, we compare the power spectral densities (PSDs) of the pressure sampling with the SPLs of the acoustic sampling in the frequency domain to show that the source of the hydroacoustic noise is vibration and vortex shedding. The stationary case successfully captures the first four tones of fundamental vortex shedding frequency. In the vibrating case, we observe individual tones of fundamental vortex shedding frequency and the vibration frequency, as well as their combinations. Overall, the results show that the vibration increases not only the sound levels but also redistributes the frequency components by interacting with flow-induced vortex shedding.

As offshore wind energy scales up to meet global climate and energy goals, the structural integrity and long-term reliability of support structures become increasingly critical. Laser-based welding offers distinct advantages for fabricating steel components in offshore wind applications, including high precision, low distortion, and compatibility with automated manufacturing. However, the fatigue performance of laser-welded joints under cyclic marine loading remains a key challenge, especially in the context of evolving design standards and harsh environmental conditions. Offshore wind components, including monopiles and towers, experience millions of load cycles from wind, waves, and turbine operation, exposing laser-welded joints to severe high-cycle fatigue conditions. Rapid heating and cooling incurred under laser welding often produce hard martensitic microstructures and high tensile residual stresses near the fusion zone, both of which can promote fatigue crack initiation and accelerate fatigue crack growth. In thick steel sections typical of offshore applications, these residual stresses are difficult to relieve, and small imperfections such as undercuts or lack of fusion can act as stress concentrators leading to fatigue crack initiation.

This review synthesizes current knowledge on fatigue design and material behaviour of laser-welded steels used in offshore wind structures. It explores the influence of weld geometry, microstructure evolution, residual stress profiles, and surface quality on fatigue life. The paper also discusses the limitations of existing fatigue design codes when applied to laser welding and highlights the need for improved predictive models and validation through representative testing.

To enable safe and cost-effective deployment of laser welding in offshore wind, future work should focus on: (1) developing fatigue-optimized weld designs specific to laser processes, (2) integrating digital tools for fatigue life prediction, (3) conducting full-scale fatigue testing under realistic offshore conditions, and (4) updating design guidelines to reflect the unique characteristics of laser welds.

By identifying research gaps and technological needs, this contribution aims to support the broader adoption of laser welding in offshore wind, contributing to more resilient and sustainable energy infrastructure.

IEA Wind TCP's Task 58 addresses the critical challenge of accelerating massive offshore wind development through offshore energy hubs, where production, conversion, storage, and consumption of multiple energy carriers can occur with local control capabilities. The EU targets of 60 GW offshore wind by 2030 and 300 GW by 2050 require moving from traditional point-to-point connections into hub-based transmission systems that provide increased interconnectivity between countries and power system synchronous areas.

This international research collaboration, initiated in January 2025, coordinates R&D efforts across three key areas: technology optimisation, system integration, and social/regulatory/economic aspects. The task addresses the significant gap between cost-effective solutions needed and current technological capabilities, which industry experts describe as a "Mars mission" for the energy sector.
The work focuses on three central questions: how technologies should evolve to maximise synergies in offshore energy hub design and operation; how to integrate offshore energy hubs optimally into power and energy systems; and how to build and operate them sustainably and economically.
The interdisciplinary approach combines three main work packages (WPs) covering technology (WPI), system integration & role in future energy systems (WP2), and sustainability / social acceptance / regulation / economics (WP3). Deliverables will include state-of-the-art technology reports, system integration methodologies, life cycle assessment frameworks, ownership structure analyses, regulatory roadmaps, and reference plant designs. The task will establish consensus on definitions, terminology, and best practices while identifying research gaps and deployment roadmaps for offshore energy hubs.
HVDC), storage, and conversion technologies, focusing on multi-vendor converter-based power system design and co-control optimisation. WP2 develops improved energy system analysis methods, assesses system value, determines optimal placement designs, and addresses resilient power system operation. WP3 applies life cycle assessment methodologies, ownership structure analysis, market design evaluation, and social acceptability frameworks.

TSOs, OEMs, developers, and energy agencies will be addressed in several workshops. The Task will leverage existing contacts in other IEA TCP tasks (e.g. 25, 35, 46, 50, 51, 53, 60, 62) and international initiatives including SET Plan for HVDC technologies and the ISGAN collaboration.
As a starting point, the Task has identified critical research priorities: scalable electrical design concepts for multi-GW systems, functional specifications for multivendor interfaces, optimisation strategies for wind turbine-HVDC-storage integration, and resilient power system operation approaches for large energy hub deployments.

This poster will present the research framework planned for IEA Wind TCP's Task 58, including its structure addressing technology, system integration, sustainability, and socio-economic aspects of offshore energy hubs. Key takeaways include awareness of the planned deliverables and research questions to be addressed, regarding scalable electrical design concepts, regulatory framework development needs, and ownership structure investigations.

EU-INTERCHANGE aims to advance regional wave and ocean climate predictions through the development of a next-generation Digital Twin of the Ocean. The project will integrate optimized wave, ocean circulation, biogeochemistry, and nearshore models spanning all European basins—from the Nordic Seas to the Mediterranean, Black Sea, and Arctic. By bridging the gap between coarse-resolution climate models and stakeholder-oriented products, EU-INTERCHANGE will deliver multi-resolution downscaled projections (up to 500 m) that explicitly resolve fine-scale processes and coastal dynamics. A basin-specific, open-access database will underpin applications in climate risk mitigation and Blue Economy sectors, with a particular focus on supporting the growth of offshore wind energy.

Analytical models offer computationally inexpensive and reasonably accurate estimates of wind turbine wakes. As a result, these models are highly popular in the wind energy community for predicting wake-induced power losses and fatigue loads during wind farm planning and layout optimization. These models can be subdivided into two categories: models for wake velocity deficit and those for wake-added turbulence. Most traditional wake models are derived under the assumption of neutrally stratified flow, whereas recent efforts have focused on accounting for the effects of wind veer and thermal stability. In this work, we compare some existing analytical models for wake velocity deficit and wake-added turbulence under different thermal stability conditions with veer effects. We selected five wake velocity deficit models: Bastankhah & Porté-Agel model, Ishihara & Qian model, Cheng model, Abkar model, and Narasimhan model, which can capture the effects of either thermal stability, wind veer, or both. In addition, we tested four one-dimensional wake-added turbulence models (Crespo model, Frandsen model, Abkar model, and IEC model) and two three-dimensional wake-added turbulence models (Ishihara & Qian model, and Khanjari model). The wake of an IEA 15 MW wind turbine is simulated using large-eddy simulation (LES) under three different atmospheric stability conditions. In addition, LES data of a 1.5 MW turbine under three different atmospheric stability conditions from an existing study is used to broaden the scope of model applicability to different scales of turbines, atmospheric conditions, and LES codes. The results provide a comparison of different models against the LES data and focus on identifying the strengths and weaknesses of different models. In addition, we aim to give a physical explanation of why certain models fail under certain scenarios. Finally, the study provides some recommendations for future development of these models.

Reducing the uncertainties in the extreme meteocean statistics is key to optimizing the offshore wind technologies and making this sector even more competitive. Indeed, the design of offshore wind turbines requires the knowledge of the 50-year wind speed at hub height and 50-year significant wave height to correctly assess the aerodynamics and hydrodynamic loads on, respectively, the turbines and foundations (which can be bottom fixed or floating). This is particularly critical in areas exposed to Tropical cyclones that results extreme wind and waves whose magnitude and statistics are tedious to model.
This abstract deals with the statistics of extreme winds and waves under tropical cyclones (TC) and relates the work carried out in the frame of the collaborative OROWSHI project, applied to offshore wind.
For sites located in areas exposed to TC risk of the main difficulties in estimating these statistics lies in the low probability of occurrence of tropical cyclones, which makes the traditional approaches based on the exploitation of (typically) 25-year of wind and waves hindcasts unapplicable. To circumvent this difficulty, another solution is to generate synthetic wind fields representative of 10 000 years and to derive the 50-year wind speed, as described in Ishihara and Yamaguchi (2015).
The OROWSHI project proposes to set new standards in terms of extreme wind and waves statistics assessments, for the design of OWT exposed to TC-risk. It is bringing three major breakthroughs:
1) The improvements of the Ishihara & Yamaguchi method for wind extreme statistics assessment, referred to in the IEC standards
2) The extension the IEC method to the estimation of extreme waves statistics using a Monte Carlo framework
3) The assessment of the contribution of extra-tropical weather systems to the extreme wind and wave statistics
The poster will present the overall OROWSHI extreme wind-wave assessment method, and will focus on the underlying challenges of developing wave modeling tools able to capture the complexity of the wave fields generated by rotating and translating wind fields induced by TC, within a computational time compatible with a Monte Carlo approach (i.e. the generation of the equivalent of 10 000 years of TC-generated wave field).

The expansion of offshore wind energy in Norway, places new demands on coastal infrastructure to support long-term Operations and Maintenance (O&M) activities. Selecting suitable O&M ports is a complex, multi-dimensional decision problem involving technical, logistical, environmental, and socio-economic factors. This work proposes a structured multi-criteria decision analysis (MCDA) framework for assessing port suitability, with a focus on the definition, justification, and weighting of key evaluation criteria. The Norwegian context serves as a case study to illustrate the framework’s application and relevance.
The study first establishes a comprehensive set of O&M port requirements derived from literature, international standards, and interviews with industry stakeholders. These requirements are grouped into five primary criteria domains:
• Infrastructure readiness (quay dimensions, water depth, crane capacity, storage, and office facilities),
• Logistics and accessibility (port connectivity, vessel maneuverability, air and road links, and operational flexibility),
• Workforce and industrial ecosystem (availability of skilled labor, proximity to offshore clusters, training institutions, and synergies with the oil and gas sector),
• Sustainability and energy infrastructure (shore power, alternative fuel access, emission mitigation, and environmental management), and
• Scalability and development potential (space for expansion, policy incentives, and long-term strategic fit).
Each criterion’s importance is discussed in detail, highlighting how port selection influences O&M cost, reliability, and environmental footprint throughout a project’s life cycle. The MCDA methodology integrates both quantitative and qualitative indicators within a transparent, weighted scoring system. Weight factors were derived through a combined literature-review and expert-judgement approach, ensuring that stakeholder perspectives and operational relevance were adequately reflected. The aggregation method follows a hierarchical structure, allowing sensitivity testing of criterion weights and demonstrating how variations affect final rankings. This framework ensures traceability and adaptability for future offshore wind developments, where data uncertainty and incomplete standardization often hinder early-stage decision-making.
To demonstrate the methodology, the paper presents a case study comparing three Norwegian ports (Tananger, Egersund, and Mandal) as potential O&M bases for SNII. The analysis applies the defined MCDA structure to real port data and stakeholder inputs. Results indicate that infrastructure readiness, logistical accessibility, and integration to an existing offshore industry have a high influence on overall suitability, while distance to site, traditionally considered critical, plays a limited role when employing Service Operation Vessels with extended offshore autonomy. Tananger emerges as the most capable short-term base due to its mature facilities and industrial integration, Egersund demonstrates strong potential with moderate upgrades, and Mandal represents a promising long-term expansion site.
Beyond the specific case results, the study’s main contribution lies in providing a replicable decision-support framework for national port planning and offshore wind logistics. By explicitly linking quantitative and qualitative parameters through MCDA, the approach enables balanced evaluation across economic, operational, and environmental dimensions.

As offshore wind turbine rotor designs continue to increase in size, spatial and temporal wind variations across larger scales have increasing importance for dynamic responses. For floating wind, the platform weight is also being reduced as designers optimize their substructures to save costs. Successful design, installation and maintenance of future mooring systems require understanding how the combination of lighter floaters and larger rotors impacts mooring lines loads. Previous work on smaller rotors suggests that the spatial coherence in the wind field is particularly important for mooring line responses.
A preliminary sensitivity analysis answering the above challenge is thus proposed, based on the INO Optiflex 22MW semi-submersible with the IEA 22MW rotor. The model can consider the platform (hull) as either flexible or rigid. A semi-taut mooring system is used with polyester rope between top chain and bottom chain sections. Non-linear time domain simulation is performed with two IEC standard turbulent wind loads - Kaimal spectrum with exponential coherence, and Mann uniform shear turbulence. Results show how turbulent wind input affects the fatigue life along the mooring lines as well as its extreme responses, for both the flexible and rigid platform models. Frequency content of the mooring loads is investigated to understand what key factors are driving mooring system response, both at low frequencies where differences between these models are well-documented for smaller rotors, and also at higher frequencies related to 3p excitation, tower vibration, and platform vibration.

Station-keeping technologies are significant drivers for the safe operation of offshore floating wind and key components for reducing the cost of energy generation. The sustainable potential of synthetic mooring line technology compared to traditional systems (chains) are related to their mechanical properties, easier installation and the reduced use of scarce primary raw materials.

These technical and economic aspects need to be weighed against any ecological impacts associated with this type of infrastructure. Mooring lines can act as artificial reefs, providing new habitats for benthic species and impacting the structure and biodiversity of local communities. In this work, the environmental impact of new fibre ropes for use in station keeping technologies for floating offshore wind farms will be investigated. An experimental field set-up is used to quantify and compare the marine communities colonising multilayered polymer mooring ropes (new mooring technology) and stainless-steel chains (traditional mooring system). The short-term colonisation patterns and diversity of micro- and macro-organisms, in addition to biomass accumulation on each material type is assessed by a combination of traditional visual assessment and molecular analyses.

Information provided by these studies will increase our knowledge on the environmental impact of synthetic mooring lines and further advance the development of new station keeping technologies for floating offshore wind structures.

Offshore wind structures are exposed to harsh marine environments that accelerate material degradation, necessitating effective corrosion protection strategies. Common methods include protective coatings, sacrificial anodes, and impressed current cathodic protection (ICCP), often used in hybrid configurations to balance cost and performance. Anodic corrosion protection, particularly via sacrificial anode systems, is a widely implemented method for mitigating corrosion in offshore wind turbine substructures. This technique involves attaching anodic materials, typically aluminium (Al), zinc (Zn), or magnesium (Mg) based alloys, to steel components. The anode undergoes oxidation, which protects the steel surface. While this method offers reliable protection in chloride-rich marine environments, it also introduces potential environmental challenges due to the continuous emission of metallic ions into the surrounding seawater, with sacrificial anodes identified as the second most significant source of chemical emissions from OWFs. Natural background concentrations of Al and Zn in seawater vary by location, ranging from 0.0005 to 0.017 mg/L for Al and from 0.0006 to 0.005 mg/L for Zn. However, these concentrations can be significantly higher at anthropogenically contaminated sites.

This study aimed to evaluate the acute toxicity of Al and Zn ions on early life stages of Atlantic cod (Gadus morhua), a species of ecological and commercial importance in the Northern Atlantic Ocean. Newly fertilized (3 days post fertilization, dpf) cod eggs were incubated in 100 mL beakers with semi-static renewal of exposure solutions containing 6 concentrations of AlCl₃ (nominal 0.0019 - 6.05 mg/L) or ZnCl₂ (0.0115 – 36 mg/L). Exposure lasted until approximately 3 days post hatch. Water samples were analyzed to validate exposure concentrations. Toxicological endpoints included mortality rates, hatching success and timing, larvae morphology and developmental deformations. Statistical analyses were performed to determine toxicity thresholds, including Lethal Concentration (LC50) and No Effect Concentrations (NOEC). To compare the sensitivity of Atlantic cod to other marine species, literature data collected from the US-EPA ECOTOX database were used to generate Species Sensitivity Distributions (SSD).

Elemental analyses showed that measured concentrations closely matched nominal values, although Al exhibited precipitation challenges at the highest exposure concentration. Al was significantly more toxic than Zn, with LC50 values of 0.212 mg/L for Al and 5.63 mg/L for Zn. Larval deformations were observed in both metal exposures, with increased incidence at higher concentrations. Morphological indices revealed impaired growth and development at Al and Zn concentrations exceeding 0.05 mg/L. Comparison to literature NOECs for Zn using SSDs revealed that Atlantic cod are sensitive to Zn exposure. Very little toxicity data exists for Al for marine species thus restricting sensitivity comparison of Atlantic cod to other marine species for this metal. Our study underscores the sensitivity of Atlantic cod embryos and larvae to Al and Zn exposure, with potential long-term growth effects down to 0.05 mg/L. The study also highlighted a lack of comprehensive ecotoxicological data for Al, pointing to the need for further research on its chronic impacts on marine organisms.

This paper presents a co-simulation methodology for analyzing the dynamic interactions between modern wind turbines and the electrical power grid. By coupling a high-fidelity structural and aerodynamic simulation tool with average-value electrical representations, it is possible to capture complex wind turbine–grid behaviors that cannot be represented using conventional models. This is enabled and made more accessible through the utilization of the Functional Mock-up Interface (FMI). This standardized interface provides access to state-of-the-art open-source libraries and software developed by the community, incorporating advanced control logic without additional implementation effort. In particular, the wind turbine controller includes the full complexity required for realistic operation, moving beyond overly simplified MPPT-based approaches.

The key contribution of this paper is a demonstration of the use of a Functional Mock-up Unit (FMU) that enables coupling between high-fidelity structural and aerodynamic modeling from OpenFAST and average-value models of the generator, DC-link, converter systems, and the grid. The grid is represented using the open-source RMS-based power system simulator TOPS. At the core of the electrical model is a direct-drive permanent magnet synchronous machine (PMSM), implemented with electric drive control schemes and fault-handling mechanisms. The framework supports two-way coupling between mechanical and electrical domains, enabling realistic simulation of electro-mechanical interactions across multiple time scales.

Key features of the combined methodology include DC-link voltage control, chopper-based overvoltage protection, and torque-limiting strategies for fault response. A grid-side converter model with PQ or PV control enables interactions with the dynamic power system. Preliminary case studies, including a curtailment event and a three-phase fault, illustrate the framework’s ability to capture coordinated transient behavior across subsystems. The proposed methodology provides a flexible and scalable approach for studying electro-mechanical interactions in converter-based wind turbines under diverse operating conditions, supporting detailed analysis of control strategies and fault responses within integrated simulation environments.

Life Cycle Assessment (LCA) is a key tool for evaluating the environmental impacts of electricity generation technologies, including wind power. However, current LCA practices across academia and the wind industry are fragmented, with results varying due to differences in modelling choices, data availability, system boundaries, and inventory databases. This inconsistency hampers comparability across technologies and geographies, and limits the utility of LCA in policy-making, tendering processes, and sustainability reporting.

The IEA WIND Task 60 – CYCLEWIND addresses this challenge by developing a harmonised framework for LCA of wind power. Building on the outcomes of the IEA Wind Topical Expert Meeting TEM#105, which gathered 30 experts from 11 countries, CYCLEWIND emphasises the importance of stakeholder collaboration. By engaging actors from academia, industry, and government, the task builds an intersectoral platform for information exchange and consensus-building. This collaborative approach is essential for developing guidelines that are both scientifically sound and practically applicable. Therefore, CYCLEWIND aims to improve transparency, consistency, and comparability of LCA results across the wind energy sector. The task is co-chaired by SINTEF Ocean and Zurich University of Applied Sciences and runs from 2025 to 2028.

Key objectives of CYCLEWIND include: (1) producing methodological guidance for harmonised LCA at the plant, turbine, and technology levels; (2) aligning with existing LCA standards; (3) supporting the integration of LCA aspects into energy policy and auction design, including non-price criteria; (4) enabling digitalised data exchange platforms for primary data collection across the wind energy supply chain; and (5) providing state-of-the-art data and case studies that demonstrate the methodological guidelines produced by CYCLEWIND.

In its initial stage, the project has identified several key challenges in current LCA practices, such as lack of consistent approaches, limited access to primary data, confidentiality concerns, technological diversity, geographical diversity from one site to another and the absence of standardised reference cases. To address these, CYCLEWIND proposes a set of harmonisation guidance that will be applied to different LCA use cases—site-specific, product-specific, and technology-level assessments.

Ultimately, CYCLEWIND will deliver a harmonised LCA framework that supports long-term decision-making in wind energy development. It will facilitate transparent environmental profiling of wind technologies, enhance the credibility of LCA results, and enable their effective use in policy instruments and market mechanisms. The task will also explore the implications of different end-of-life approaches and discuss potential opportunities to capture aspects of material efficiency within the context of LCA.

Traditional operation and maintenance (O&M) for offshore wind turbines are fraught with high costs, safety risks, and logistical challenges. This presentation explores a transformative approach that integrates artificial intelligence (AI), robotics, and digital twin technology to enable predictive and autonomous O&M throughout a turbine's lifecycle. We introduce a comprehensive framework featuring AI-enabled autonomous vehicles capable of performing precise inspections and repairs in harsh ocean environments. Concurrently, a digital twin platform—powered by internet of things, paralleling computing, 3D virtual reality and advanced hybrid models including Physics-Informed Neural Networks (PINNs), Multi-Task Gaussian Processes, and Enhanced Auto-associative Kernel Regression—facilitates predictive maintenance of wind turbines. This system proactively detects component anomalies, such as bearing, blade and tower failures, to optimize intervention timing. By leveraging IoT and remote monitoring, this integrated solution minimizes human intervention in hazardous tasks. The result is a significant enhancement in turbine reliability, a reduction in operational costs, and improved safety, thereby positioning autonomous systems as pivotal for the sustainable expansion of offshore wind energy.

In the present work, the wind and current climate in an offshore wind farm in the North Sea is investigated. Together, the wind energy resource available in the area is assessed and analyzed. The analysis is based on 10 years of data (2015-2024) generated by means of the well-known atmospheric model WRF and the inhouse oceanographic model SINMOD. For the assessment, several statistical features have been used, such as seasonal variability, annual and inter-annual variability, quantiles of the probability distribution, directional distribution, to mention some of them.

This is a work in connection with EU project WindTwin, which aims to develop and validate an offshore wind farm digital twin (DT) for highly accurate prediction of power production and energy demand of the end user. The DT will give users tailored access to high-quality information, services, models, scenarios, forecasts, and visualizations, as a central hub for offshore wind decision-makers. And will also serve as platform, offering users access to a comprehensive array of high-quality resources, services, models, scenarios, forecasts, and visualizations. WindTwin seeks to revolutionize the way industry professionals make informed choices. The WinDTwin consortium consists of 13 organizations from 7 different Member States.

This study investigates the optimization of power cable configurations for floating axis wind turbines (FAWT) in offshore demonstration projects. We simulated the unique precession motion characteristic of FAWTs as a forced vibration using OrcaFlex. Through a comprehensive exploration of the parameters necessary for configuration and the variations in water depth within the designated marine environment, we aimed to identify the optimal cable configuration. The natural frequency of the cables was estimated using free vibration analysis. Our findings reveal that while the maximum tension, minimum bending radius, and minimum distance between floating body cables show no dependence on the period of precession motion, a clear amplitude dependence was observed. Furthermore, we demonstrated that near the cable's natural frequency, synchronization occurs with the wave period, which results in the minimum bending radius and minimum distance between floating body cables operating within safe limits.

This investigation specifically focuses on the power cables for offshore demonstrations of FAWTs, which are expected to have a service life of several years. However, for commercially deployed wind turbines, it is crucial to conduct fatigue analysis due to cyclic loading and to evaluate the impact of marine growth. Given that the period estimated from the precession motion is sufficiently long, we hypothesize that the wave period predominantly influences the number of cycles experienced by the cables. To mitigate the increase in deformation associated with the amplitude of the precession motion, we propose ensuring a sufficiently long flexible region extending to the seabed fixation point. Additionally, the horizontal displacement of the power cable connection points is anticipated to increase, which may lead to variations in marine growth characteristics compared to existing floating offshore structures. However, due to the numerous environmental factors involved, further empirical investigations through offshore demonstrations will be necessary.

The OBLEX–F1 campaign held in the southern North Sea provided more than one year of scanning lidar data with some scans showing recurring streak patterns in addition to random noise. The streaks appeared as local value shifts along random azimuths. The value shifts were most pronounced in the carrier-to-noise (CNR) data and often affected the corresponding radial velocity data, visibly altering the measured field. Most of the streaks did not have a fixed position and could appear at any azimuth of the scanned sector. Widely used filtering methods, e.g., median filter or clustering-based filters, are calibrated to remove random noise, but do not consistently filter streaks due to a combination of their regular structure and randomness of occurrence. Although the filtering parameters can be adjusted to improve the result, the adjustment requires prior knowledge of streaks' presence. Otherwise, the adjusted filter becomes too aggressive and may classify valid data as noise.

Instead of building a decision tree on how and when to adjust parameters of a specific filter, we suggest a supplementary method as a second-pass filter that targets streak structures specifically. The method is run on CNR data, where the streaks are most prominent. An additional check is performed on radial velocity data to determine that the found CNR streaks correspond to strong shifts in radial velocity. The suggested algorithm is based on standardizing the CNR data with an interquartile range, so that the streaks are amplified compared to a background and can be filtered via thresholding. We also provide a verification method that is independent of the filtering algorithm. The verification is based on comparing data in consecutive scans and, thus, can be used to verify any improved method for filtering streaks or other non-persistent structures.

Nowadays, the offshore floating wind sector in Italy is entering its commercial phase. As of the end of September 2025, over 110 offshore wind projects—totalling more than 77 GW—are at various stages of the authorization process. Over 90% of these projects are based on floating technology. Among them two have been already obtained the final consent. Due to the complexity and unique characteristics of floating wind technology, a significant share of the manufacturing phase is expected to involve the local economy. In Italy, for example, existing ports, steel mills, and shipyards have the potential to be revitalized and strengthened, and a significant number of jobs is expected to be created both during the construction and operational phases. Given the large scale of the announced floating offshore wind farms, a substantial number of households can be supplied with clean energy, and a significant reduction in CO₂ emissions can be achieved. These metrics are crucial not only for decision-makers at the local, regional, and national levels, but also for shaping an accurate and effective narrative about the overall impact of this technology—both in the media and among citizens.
However, providing a quantitative and reliable assessment of these parameters remains a significant challenge. The main reasons are analysed in the work and briefly outlined below. Regarding job creation, the key issues include the lack of long-term operational data and installation experience in specific contexts (such as country, region, or area), as well as the balance between local and foreign content. As for the number of households supplied, this depends on both current and projected average annual electricity demand. For the estimation of avoided emissions, the present and future energy mix should be considered. Although some metrics for evaluating the above-mentioned parameters are already known, data provided by operators still show a wide range of variability. The authors’ methodology for assessing these parameters, along with a critical discussion on the results obtained from selected case studies carried out in Italy are presented.

As offshore wind farms scale to multi-GW capacity, effective control strategies are vital to maintain energy yield and cost efficiency. Wake interactions can lower downstream inflow and cut annual energy production by 10–20 %. Farm-level approaches such as yaw-based wake steering offer a way to redirect wakes and recover these losses. Yet, large-scale application remains challenging due to site-specific winds, complex flow coupling, and the trade-off between energy gain and turbine loading. This study quantifies the aerodynamic and economic impacts of yaw-based wake steering under realistic offshore conditions.

A custom numerical framework was developed to evaluate and optimize yaw control at the wind-farm scale. Its aerodynamic core couples a Blade Element Momentum (BEM) solver with a Gaussian wake model to capture wake deficits and flow recovery across the array. Yaw angles are optimized through a gradient-based algorithm that maximizes total power output within defined yaw limits.

To capture a wide range of atmospheric flow characteristics, two complementary datasets were employed. (i) Idealized wind fields with uniform and Weibull-distributed velocity profiles were generated to perform controlled sensitivity analyses and isolate the influence of statistical wind variability on the results.
(ii) Site-specific wind data from the Alpha Ventus offshore wind farm were incorporated to represent more realistic environmental conditions. These data include wind direction frequencies and speed distributions obtained from FINO1 meteorological mast observations (2009–2011). The simulated power outputs were integrated over the annual wind distributions to estimate annual energy production (AEP), followed by an economic analysis based on predefined cost and revenue models.

Results show that yaw optimization increases total AEP by about 6–7 % compared to baseline operation, reducing wake losses by over 95 % and improving power balance across the array. This corresponds to a reduction in the levelized cost of energy (LCOE) of roughly 7–9 %, which represents the average cost per unit of electricity produced over the system’s lifetime, accounting for capital, operational, and maintenance expenses. The performance gain remains stable across wind directions and speeds, and sensitivity analyses with uniform and Weibull inflows confirm the robustness of the approach. The Alpha Ventus case study demonstrates consistent performance under realistic offshore conditions, validating the framework’s practical applicability. The economic evaluation indicates higher annual revenue, while load redistribution across turbines may lower maintenance frequency. Overall, the results emphasize the value of integrating aerodynamic optimization with cost modeling to achieve a balanced, lifecycle-oriented control strategy.

This study demonstrates that integrating yaw-based wake steering with site-specific wind data and economic evaluation can significantly enhance the efficiency and profitability of offshore wind farms. The results advance understanding of control-oriented design for long-term sustainability in wind energy. Future work will focus on extending the framework toward real-time, closed-loop operation and dynamic wake steering in floating turbine arrays, paving the way for more adaptive and intelligent offshore wind control strategies.

Abstract
The rapid expansion of floating offshore wind technology (FOWT) demands standardized and holistic design methodologies that integrate structural, hydrodynamic, and control aspects across multiple simulation environments. Current design processes are often fragmented, relying on isolated tools and manual iterations, which can lead to suboptimal performance and limited reproducibility. To address these challenges, Fraunhofer IWES is developing a unified optimization framework, FloOpt, that connects a suite of in-house and open-source modules for modeling, analysis, and optimization of FOWT systems.

The proposed framework (FloOpt) is established around a reference FOWT model representative of a large-scale offshore wind turbine. It couples Fraunhofer IWES in-house aero-servo-hydro-elastic simulation software, test automation python framework and post-processing tool — MoWiT, PyWiT, and Island — with the open-source optimization library iwopy, forming an integrated workflow for parametric design and optimization studies. FloOpt acts as the central orchestrator, managing model generation, simulation execution, post-processing, and optimization loops. The workflow enables systematic exploration of key design parameters, including column inclination and geometry, ballast distribution, mooring line characteristics, and global performance indicators such as acceleration response and reliability index.

A gradient-based optimization strategy is implemented to improve convergence efficiency compared to generic algorithms, reducing computational cost and enabling faster iteration cycles. Boundary conditions derived from environmental and operational inputs are applied to each design iteration, allowing automated evaluation of dynamic responses and fatigue life under representative design load cases (DLCs). Results from PyWiT simulations, including free-decay and DLC scenarios, are post-processed in Island, with key performance metrics returned to FloOpt for assessment. The iterative process continues until all optimization criteria are satisfied, yielding an optimized and verifiable floating wind turbine configuration.

Preliminary results demonstrate the capability of FloOpt to streamline FOWT concept development and significantly reduce manual design effort. The modular architecture ensures compatibility with future extensions, such as novel control strategies, enhanced hydrodynamic modeling, or alternative floating platform topologies. This work contributes to the broader goal of establishing standardized and reproducible procedures for the holistic design and optimization of floating offshore wind systems, supporting comparability and transparency across the research community.

Future developments will focus on extending the methodology toward multi-objective optimization, incorporating cost and manufacturability constraints, and validating results against experimental and field data. The presented framework represents a significant step toward a unified digital environment for FOWT design, accelerating the transition from conceptual studies to industrial-scale floating wind deployment.

**Keywords:** Floating Offshore Wind Turbine (FOWT), Optimization, Holistic Design, Numerical Modeling, FloOpt, Fraunhofer IWES, Reference FOWT Model

Floating offshore wind turbines (FOWT’s) are more expensive than bottom-fixed or on-shore wind turbines, however they allow for more wind resources to be utilized, making use of the large offshore space with ideal wind resources available. While FOWT’s are likely to play a significant role in the energy transition, further work is required to improve their cost effectiveness to make them a more viable option. Mooring designs are one possible area of optimization, potentially reducing the material cost for a system and for site-specific design, while ensuring limits on safety and efficiency are still met in an array context.
In order to reduce the FOWT costs, this research has aimed to develop an optimization toolset to optimize mooring designs directly, while also designing the mooring system to maintain limits on platform and cable motions to indirectly impact the systems efficiency. The toolset uses non-linear time-domain simulations through OrcaFlex to assess the constraints of the process, ensuring high accuracy, leading to a low cost, feasible design. These dynamic simulations of FOWT’s have proven to be computationally expensive, leading to long wait times and larger computational resources required than less accurate static or linearised frequency domain simulations. This has led to the process being made computationally efficient, reducing the number of expensive OrcaFlex simulations required using surrogate models trained using a reduced number of simulations. The accuracy is increased around the optimum region by retraining of the surrogate models using data collected in the evolved optimal region of the design space, to ensure the final designs are accurate.
This toolset can be used within industry to allow for moorings to be cost-optimized, while also improving the computational efficiency of design iterations for site-specific mooring system designs. The challenges and solutions outlined may also provide a better understanding of how to further improve the optimization process in terms of turn-around time and ensuring accuracy of optimal solutions with respect to design constraints.

This work presents a simulation-based approach to automated weld groove localization for crawler robots, which are mobile robotic platforms designed to traverse steel pipes in the jacket foundations for offshore wind turbines. Manual welding of these large tubular structures is labor-intensive and time-consuming, and it exposes operators to significant safety hazards. This study investigates the potential of simulation and sensor fusion to facilitate autonomous or semi-autonomous welding operations, thereby improving operational efficiency, process repeatability, and weld quality.
The crawler robot traverses a tensioned rail that encircles the pipe and employs a line laser scanner to localize and measure welding grooves. Fiducial markers affixed to the pipe surface provide reference points for localization and alignment. Sensor fusion combines three-dimensional measurements, inertial measurement unit (IMU) data, and motor encoder feedback to estimate the robot’s position and orientation relative to the pipe. This research demonstrates how real-world sensor data can be combined to support the generation of weld trajectories and robot guidance. A custom simulation environment has been developed that enables the modeling of pipe geometry, crawler kinematics, and sensor behavior, including realistic representations of line laser scanning, IMU drift, and encoder noise.
The environment supports virtual testing of seam detection algorithms and assessment of localization accuracy under diverse operating conditions. Simulation-driven development of robotic welding systems addresses the demand for scalable and high-quality manufacturing processes in offshore wind substructures. The developed approach also supports broader objectives, such as reducing production costs, enhancing quality assurance, and advancing the transition to automated fabrication within the offshore wind sector. The system is developed as part of the WindRise project and contributes to the digitalization of welding operations in offshore wind manufacturing.

Offshore wind turbines have rapidly scaled up over recent decades, a trend expected to continue, with the next generation of wind turbines (NEXTgenT) exceeding 25 MW and reaching tip heights of over 350 metres.
These turbines will operate far beyond the surface layer (SL; roughly the lowest 10% of the atmosphere, i.e., <200 m) within parts of the atmosphere that are currently under-resolved, and our understanding of turbulence remains limited.
The turbulence wind load on large offshore wind turbines (LOWT) is calculated using power spectral density functions, normalized by the variance of the longitudinal wind component (σu).
This variance changes with height, yet present (engineering) standards often treat it as uniform, assuming homogeneous SL-parametrizations developed for onshore conditions.
Above approximately 100 metres, direct measurements of the longitudinal variance profile are sparse, and SL assumptions tend to break down.
As a result, we cannot verify how turbulence behaves across the turbine, leading to large uncertainties in turbine load and performance predictions.

In other words, the the longitudinal turbulence variance is currently not measured at turbine-relevant heights, including the hub height of NEXTgenTs and beyond.
Nor is there an established method to infer this vertical profile from available LIDAR remote-sensing data.
Our aim is to infer this missing profile using quantities that can be measured reliably, such as the vertical variance and mean wind speed.
Obtaining a realistic, height-dependent profile of σu will enable a more accurate representation of non-uniform turbulence intensity across the rotor plane.

We address this gap by asking: Can a data-driven model, trained on physical simulations, infer the unmeasured structure of turbulence?
To answer this, we explore a simple yet powerful data-driven framework: A shallow neural network (S-NN), trained on large-eddy simulation (LES) data covering various stability regimes and surface conditions.
The S-NN learns to reconstruct the along-wind turbulence component (σu) from quantities typically available in real-world measurements, such as mean wind components and vertical turbulence (σw).
This ’virtual instrument’ mimics real-world conditions, where turbulence can only be observed at discrete altitudes.
Linking this with physically grounded simulations, we aim to reconstruct the missing parts of the profile.

First results show that the S-NN accurately reproduces LES turbulence profiles and captures stability-dependent vertical structures, without requiring explicit knowledge about atmospheric stability.
The next step is to apply the virtual instrument to mast measurements, to test its ability to retrieve physically meaningful turbulence relationships.
This constitutes a necessary first step towards more accurate, non-uniform turbulence models for large offshore wind turbine design.

Our approach aims to build a bridge between measurements and modelling.
By enabling the use of UAV–LIDAR systems to infer missing turbulence characteristics, it offers a new methodological framework for reconstructing turbulence characteristics where direct observations are impossible.
With this, we are hopefully contributing to the overall goal of supporting the design and operation of safer, more efficient, next-generation offshore wind turbines, and thus contributing to the renewable energy transition.

The design and analysis of mooring systems for floating offshore wind turbines traditionally rely on constitutive models that are approximations, calibrated from experimental data, and oftenly cannot capture and predict complex behaviors under highly nonlinear and chaotic load conditions. In this work, we introduce and investigate a novel disruptive alternative: a data-driven computational mechanics framework for modeling mooring lines, enhancing design reliability and operational safety for floating offshore wind systems. In particular, we develop a data-driven greedy optimization algorithm based on the Alternating Direction Method, called GO-ADM, enabling direct utilization of experimental data and bypassing empirical constitutive assumptions. Numerical validation demonstrates that our approach achieves global optimal solutions and convergence, offering insights into computational efficiency and robustness. Furthermore, we apply the framework to the structural analysis of a nylon mooring line using experimental data from cyclic loading tests conducted at industrial facilities. Our results underscore the potential of data-centric mechanics as a disruptive technology for next-generation offshore wind infrastructure.

European wind turbine manufacturers face mounting challenges in maintaining competitiveness while ensuring high-quality and sustainable production. Blade manufacturing, in particular, remains labour-intensive and blades at their end-of-life remain difficult to recycle. This poses significant barriers to industrial scale-up and wind turbine blade circularity. The MADE4WIND project addresses these challenges by advancing three key technologies: preform manufacturing for efficient and defect-free blade production, toughening interleaves to enhance structural integrity and reduce blade mass, and novel recycling processes to enable closed-loop material recovery.
Preforms, in this context, are large stacks of dry fabric reinforcement materials shaped, cut, and stabilized outside the main blade casting mould and subsequently placed in the blade mould. This decoupling allows for automated, high-quality manufacturing. In the MADE4WIND project a virtual process chain of blade manufacturing to simulate preform production is developed. The virtual process chain combines a high-fidelity model capable of predicting defect formation during forming with an efficient kinematic model for process optimization. This approach supports scalable, repeatable, and cost-effective blade manufacturing.
Toughening interleaves are nonwoven polymeric layers inserted between fabric plies to significantly enhance the mechanical properties of the laminate. Within the MADE4WIND project, these materials and their toughening mechanisms are investigated to improve blade integrity at critical design points. This will enable lighter blades without compromising their performances. Preliminary results show a substantial increase in laminate toughness, supporting the potential for reduced material usage and improved blade longevity.
Lastly, the project explores innovative recycling processes based on the Recyclamine® resin technology. The project investigates structural-level dissolution of the recyclamine epoxy resin, allowing recovery of the reinforcement materials with preserved architecture and fiber sizing. Composites cast using recycled reinforcement materials are mechanically tested and expected to demonstrate comparable mechanical performance to virgin composites, offering a viable pathway to blade circularity with reduced waste and raw material demand. The MADE4WIND project is also developing new handling and processing methods for the recycled materials to support this recycling approach, contributing to a more sustainable end-of-life strategy for wind turbine blades.
These three critical blade technologies will be presented at the conference. Together, they represent a significant step towards meeting the European Union’s climate and sustainability goals, while strengthening the competitiveness of the European wind energy sector.

The design and optimisation of floating offshore wind farms require accurate modelling of wake dynamics, particularly under the influence of platform motions such as yaw and pitch. These motions alter the rotor inflow conditions, which in turn affect the wake propagation and direction. This has a direct impact on the performance and loading of downstream turbines. However, current engineering wake models lack the capability to represent these dynamic effects, limiting their accuracy and applicability for floating wind turbine scenarios.

The primary objective of this work is to develop analytical wake steering models that can be integrated into existing engineering tools, such as DIWA (Dynamic Wake Meandering Tools developed by SINTEF) and PyWake, to enhance their ability to simulate floating wind turbines. These analytical models are derived from high-fidelity Computational Fluid Dynamics (CFD) simulations, which are essential to capture the complex flow physics associated with yaw and pitch-induced wake deflection. By bridging the gap between high-fidelity modelling and low-order engineering tools, this work contributes to the development of next-generation modelling capabilities for floating wind applications.

To achieve this, we perform detailed CFD simulations of the IEA 15 MW reference wind turbine using the Actuator Line Method (ALM) implemented in OpenFOAM. The ALM, based on blade element theory, allows for accurate representation of blade-induced forces and near-wake dynamics. These forces are introduced as momentum sink terms in the Navier-Stokes equations, and the unsteady PISO-based solver is extended to handle the turbine-induced flow modifications.

The CFD simulations are conducted under both steady and unsteady inflow conditions. For steady-state validation, Reynolds-Averaged Navier-Stokes (RANS) models are used, while Large Eddy Simulations (LES) are employed to resolve unsteady wake meandering and turbulence structures. Inflow turbulence is generated using the Mann turbulence model, which is coupled with the CFD solver to replicate realistic atmospheric conditions. The high-fidelity simulations are validated against available measurement data to ensure the accuracy of the predicted wake behavior.

The results demonstrate how yaw and pitch motions influence the rotor plane orientation and steer the wake, leading to significant changes in wake trajectory and intensity. Multiple wake detection methods, including Gaussian fitting, momentum deficit, and power-based techniques, are applied to quantify wake deflection and structure. These insights are then used to formulate analytical models that describe the wake steering behaviour, including meandering as a function of yaw and pitch angles.

The developed analytical models are designed for seamless integration into engineering wake models, enabling fast yet accurate prediction of wake behaviour in floating wind farms. This work represents a significant advancement in the modelling of floating wind turbines, providing tools that are currently lacking in standard engineering practice. By incorporating the effects of platform motions, the proposed models enhance the predictive capability of existing tools and support more informed decisions in wind farm design, control, and layout optimisation.

Traditional methods for simulating wind fields in load case design typically focus on short-duration, 10-minute high-frequency simulations, capturing only the microscale portion of the wind spectrum directly linked to atmospheric turbulence. Meanwhile, large-scale low-frequency mesoscale models are primarily used for wind resource assessment and energy production forecasting. Long-duration turbulent (high-frequency) wind time series are becoming increasingly important when moving from turbine to farm level. A typical application is related to wind farm control, not only for power production but also in grid integration applications and structural damage. Moreover, bridging the mesoscale and microscale wind fluctuations to generate long, high-frequency time series has received limited attention until recently.

The developed methodology allows for the generation of long, turbulent, high-frequency, statistically realistic wind time series with a low computational cost.

These long time series are based on the NORA3 Hindcast model for the mesoscale fluctuations and the model from Cheynet et al. based on measurement data from FINO-1 for the microscale fluctuation. Each unique case of wind speed, wind direction, atmospheric stability and friction velocity from NORA3 is used to generate 2-hours of turbulent wind data using SINTEF’s synthetic turbulence generation code FLAggTurb. Our method is based on generating in advance a collection of those 2-hour stochastic turbulent time series with multiple seeds for each unique set of atmospheric conditions, before creating the long time series in a second step. The main novelty is the splicing method developed to transition between two consecutive wind files during the 1-hour overlapping period to ensure the linear evolution of the power spectral density and the variance. This is made possible by using independent time series for two consecutive time series issues from two different randomly picked seeds.

The methodology has been validated against observations at the weather mast FINO1 in the North Sea, both using the 10-minute wind statistics (standard deviation) as the generated time series reproduce the wind fluctuation only statistically. The spectrum of the generated time series was also compared with the observations for periods up to 24 hours.

The results show an excellent agreement of the 10-minute mean and standard deviation for the along-wind component, which gives an accurate estimation of the average turbulence intensity, as well as the 10-minute standard deviation for the side wind and vertical wind components. Nevertheless, some time periods with specific weather phenomena underestimate the fluctuations.

From a spectral point of view, the generated spectrum (NORA3) represents precisely the observation in the mesoscale domain, which confirms the suitability of NORA3 in modelling large mesoscale fluctuations, while the microscale spectrum is properly represented as well. The modelling of the 10-minute to 1-hour periods is correctly represented in stable and neutral conditions, but the power density of the generated time series is significantly lower than the observation for this range in unstable conditions.

Further applications of this method, using an operational weather forecasting system as mesoscale forcing, would allow for the forecasting of wind fluctuations and turbulence intensity at a low computational cost.

It is expected that offshore wind capacity will increase significantly between now and 2030 with an expected capacity of 370GW of which 18.9GW will be floating platforms. A key challenge to this rapid increase in floating wind capacity is means by which they will be maintained. Tow to port presents high costs due to downtime and it may be more economically viable to perform some maintenance in-situ. One method for performing in-situ maintenance would be the use of a floating crane vessel.
This study builds upon prior research into the feasibility of using a floating crane for generator exchange on a semi-submersible FOWT, specifically the UMaine VolturnUS-S supporting a 15 MW reference turbine. The wind turbine model is validated by using a marine simulation environment at NDC with the FATHOM HYDRO physics engine by calibrating the model RAOs against published values from the NREL. The dynamic responses of the generator, nacelle, and crane barge were evaluated for regular and irregular waves during the removal of the wind turbine generator using a simple floating crane model. By considering a limiting acceleration on the generator of 5 m/s² it is possible to determine under what sea states it is safe to carry out the operation.
The results highlight that while generator accelerations are a significant operational factor, the primary constraint is the risk of collision between the generator and turbine structure during lifting operations. Parametric studies revealed critical wave periods that exacerbate generator motions and collisions, and while modifications to the lifting methodology proved ineffective, reorienting the crane barge parallel to incoming waves showed a modest reduction in collisions. These findings underline the importance of vessel selection, wave direction, and sea state limitations in ensuring the viability of in-situ maintenance using floating cranes for FOWTs.

Floating offshore wind represents a viable pathway toward sustainable energy production. As the floating wind farm moves to remote deep-sea areas with vast wind resources, hydrogen production provides an alternative promising solution for energy transmission. The semi-submersible floating wind turbines (FWTs), which inherently offer available onboard space, can accommodate hydrogen processing system equipment onboard. To meet the need, the design of decentralized hydrogen production on a semi-submersible FWT has been proposed [1-2].

The integration of hydrogen facilities on platform will modify the overall mass properties and, consequently, affect the performance of floating wind hydrogen system (FWHS). Moreover, the operational and survival constraints associated with hydrogen equipment can exceed those required for wind turbine operation alone. Therefore, a comprehensive understanding of the impact of integrating hydrogen facilities into FWT is essential for FWHS development. While both frequency-domain and time-domain coupled models have been employed to assess FWHS performance, long-term response analyses are still scarce.

The present study mainly investigates the long-term extreme responses of a decentralised floating wind hydrogen system on a 15 MW semi-submersible platform. A fully coupled aero-hydro-servo-elastic model is first developed using OpenFAST. Then, time-domain simulations are carried out considering combined wind wave conditions, and the extreme responses are estimated based on a direct inverse first order reliability method (D-IFORM) [3]. The analysed variables include the platform motion, mooring line tension, tower base loads, and the critical locations for the electrolyser and hydrogen pipe connection. These results offer valuable insights into the dynamic responses of FWHS and contribute to the design and development of offshore hydrogen production technologies.

[1] Castillo C A R, Collu M and Brennan F 2024 Int. J. Hydrogen Energy 89 496–506
[2] Pham T D, Dinh V N, Trinh L C, Judge F and Leahy P 2025 Int. J. Hydrogen Energy 150832
[3] Mackay Ed, de Hauteclocque G. Ocean eng 2023 273 113959

The reliable operation of wind turbine converters and transformers increasingly depends on the effectiveness of their thermal management systems. As turbine ratings rise and offshore environments pose new operational constraints, cooling concepts must be validated not only for nominal operation but also under disturbed grid conditions. However, physical testing of such systems remains costly and logistically challenging, particularly in offshore applications. This work presents a simulation framework for virtual testing of liquid-based cooling systems, designed to support the evaluation and development of advanced thermal management strategies for large wind turbine drivetrain components.

The model is implemented in MATLAB/Simulink and is structured around a set of interacting modules. Heat transfer is described using thermal masses and energy balance equations, enabling fast but representative dynamic simulations. Pumps and fans are parameterized through similarity rules, while the air-liquid heat exchangers are modelled using the ε-NTU method. The simplified turbine model links wind speed to heat dissipation in converter and transformer, establishing the thermal load input for the cooling system. Controllers are implemented in four layers, coordinating the response of pumps, fans, valves and derating strategies to both operational and environmental conditions. All modules allow extension and modifications with alternative components, control strategies, or cooling media.

The model is tested and evaluated according to specific goals and performance parameters defined in the Horizon 2020 research project RealCOE. The tests include normal turbine operation, reactive power production under no-wind conditions and grid faults.

The contribution of this work lies in enabling virtual testing of cooling systems for wind turbines, integrated with its operation, component design and control strategies. By coupling thermal models with a simplified turbine representation and control algorithms, the framework provides insights into system-level behavior that are difficult to capture with isolated component models, e.g., the exploration of alternative cooling media, and the development of improved control and operation strategies.

Mooring systems play a fundamental role in a wide range of offshore applications, including oil and gas platforms, floating offshore wind turbines, aquaculture installations such as salmon farms, and the deployment of scientific and industrial equipment for marine exploration. Traditionally, mooring systems have been constructed using steel chains or wire ropes, whose mechanical properties are well understood and can be accurately described assuming linear stress-strain models. However, in recent years, composite materials, particularly synthetic fiber ropes such as polyester and nylon, have been gaining ground due to their superior fatigue resistance and reduced weight, making them particularly suitable for deep-water applications, a key advantage for the expanding offshore wind sector.

This work addresses the numerical implementation and predictive capabilities of two distinct mooring models: a discrete Lumped-Parameter Model (LPM) and an advanced nonlinear torsion- and shear-free Kirchhoff rod model (called ARMoor). The models are benchmarked through a series of tests, including static analysis, dynamic horizontal loading, and a scenario with environmental forces and uneven seabed. Comparisons with reference solutions reveal a clear trade-off: the LPM provides computational efficiency for basic analyses, but the continuous rod model achieves higher fidelity in capturing critical aspects like elastic stretching, seabed contact location, and dynamic response under complex loading.

The study confirms that advanced, reliable mooring models are vital for the industry's growth and suggests future research paths, including materials with nonlinear stress-strain relationships and memory effects, improved contact algorithms, and hybrid modeling approaches where, for instance, wind and wave loads, control systems, and mooring lines are included to compute the structural response.

Substructure of floating wind turbine (WT) entails huge challenges in terms of design, as they must ensure structural stability and withstand extreme dynamic loads and harsh environmental conditions. This involves conflicting requirements on physical and mechanical properties of which advanced construction materials can provide tailored solutions. However, the integration of advanced materials and designs requires high-fidelity models and reliable simulation tools to evaluate their impacts on safety, efficiency and sustainability.
A flexible optimization framework integrating machine learning, with Structural Finite Element Analysis (FEA) and Computational Fluid Dynamics (CFD) simulations is developed by SINTEF in collaboration with the construction industry. The goal is to enable a design platform for reliable and lasting WT substructure in accordance with the R&I priority themes of EERA JP Wind.
Considering concrete as a critical component of WT substructure, the optimization framework addresses tailoring concrete mixes and evaluating its integration in the substructure design as a case study. The focus is on the evaluation of multi output ML algorithms and their capability to predict the performance of concrete depending on the quantity of data and the features classification of concrete mixes. Two multi output optimization approaches were evaluated, focusing on the density and the strength as relevant properties for floating substructure. The first involves simultaneous prediction with ML estimators that inherently support multi-output prediction such as decision tree, random forest and neural network. The second is based on ensemble model as warper to train the estimator for each output. The training and testing data were collected to represent varying concrete mixes, of which the constituents can be organized under hierarchical structure using class and subclass categories. While incorporating subclasses in the data features improve prediction performance, it limits the quantity of data for training and testing ML algorithms due to lack of material details. For limited amount of data, the decision tree regressor showed improved prediction compared to neural network, which is prone to overfitting leading to poor performance on validation data. The prediction is further improved using ensemble algorithms, demonstrating non corelation between density and strength of the investigated concrete mixes.
The ML model is used to tailor the concrete mixes to achieve the required properties on density and strength through the optimization of aggregate and binder compositions. FEA combined with CFD are applied to evaluate the integration of lightweight concrete in the substructure design to ensure design stability for defined sea conditions and evaluate the stress according to standard Ultimate limit state, with consideration of the forces in the mooring system. The platform is developed to enable the future integration of next generation circular aggregates and new types of binders (Such as geopolymers) in the substructure and improve WT circularity.

Floating offshore wind turbines are expected to play a pivotal role in Norway’s transition to renewable energy, given its deep-water coastline. Large-scale deployment through wind farms requires port infrastructure capable of accommodating turbine components and floating substructures. This study evaluates the suitability of Norwegian ports for supporting floating offshore wind development. Key parameters analyzed include quay structural specification and environmental conditions around ports derived from the NORA3 dataset. Results indicate that Norwegian ports generally experience benign metocean conditions for assembly of turbine components. However, the load bearing capacity of existing quay structures in most of the ports does not satisfy expected load for larger floating wind farm deployment which requires 30 to 50 tons per square meter. Accommodating larger wind turbines may involve uncertainty reduction in structural assessments, optimized assembly methods, cost-effective retrofitting strategies, and advanced structural health monitoring systems. These findings highlight the need for port infrastructure upgrades to unlock Norway’s offshore wind potential.

A large-scale integration of renewable energy sources, such as wind farms, brings along an increased penetration of power electronics converters in the grid. These inverter-based resources (IBRs) create challenges for transmission and distribution system operators as their behavior during faults differs radically from conventional synchronous generators. This behavior is highly dependent on the specific IBRs and its control system, and available models in commercial software are not accurate for short circuit studies. Furthermore, with power grids becoming more dynamic, ensuring adequate protection systems becomes increasingly complex. Performing fast and reliable short circuit studies in an IBR-dominated grid is therefore emerging as a key challenge for network operators and solving this is strictly necessary for the successful large-scale integration of wind power.

This study employs a comprehensive literature review, and an examination of existing initiatives from system operators to manage IBRs in their power system models. The aim is to identify the steps operators must take to ensure accurate representation of IBRs in short circuit studies, and to identify areas where further research is needed to support the development of efficient modeling practices.

The study identifies several key aspects that must be considered when representing IBRs in short circuit studies. First, it outlines which components of an IBR must be included to capture relevant fault behavior. Second, it discusses whether electromagnetic transient (EMT) simulations are necessary, or if simpler approaches can be sufficient in some cases. Third, it examines the challenges operators face in accessing proprietary data from manufacturers, and how this affects their ability to model IBRs and enforce grid code requirements. Finally, examples from current practice are presented to illustrate how different operators are approaching these issues, and what can be learned from their experiences.

The findings point to a growing need for system operators to adapt their modeling practices to better reflect the behavior of IBRs during faults. While EMT simulations offer high accuracy, they are resource-intensive and not always practical for routine studies. A balance must be struck between model complexity and usability, especially given the limited access to proprietary control data.

Main author:
Bendik Hjertholm Voldseth

Co-Authors:
Etienne Cheynet
Joachim Reuder

Using NORA3 reanalysis tool to improve the accuracy of extreme wind estimates for design and insurance of offshore wind turbines.

Accurate estimation of offshore extreme winds is essential for both the design of offshore wind energy infrastructure and the financial risk assessment carried out by insurance companies. Traditional methods for modeling extreme winds in the North Sea and surrounding regions have relied on relatively low-resolution reanalysis data being extrapolated or site-specific measurements from offshore installations, both of which introduce significant uncertainties (Barthelmie, 2021). This study investigates the potential of the NORA3 reanalysis model, a high-resolution dataset with a spatial resolution of 3 × 3 km, to provide more precise estimates of offshore extreme wind conditions.

The study focuses on the validation of NORA3 against in-situ wind observations obtained from coastal lighthouses, within the last 55 years. Compared to data from oil platforms or other offshore structures, lighthouse measurements are less disturbed by surroundings and therefore more representative of the true wind field. Initial analysis indicates that NORA3 systematically underestimates wind speeds, particularly in the high-frequency range of the spectrum. To address this, a spectral correction is applied. This correction, developed by Malekmohammadi et al. (2025), and inspired by Bastine et al. (2018), is based on adjusting the slope of the power spectral density (PSD) within a defined subrange, where the PSD is expected to follow the canonical −5/3 power law. The correction effectively serves as a deconvolution procedure to restore the variability that is otherwise dampened in the reanalysis output.

Following validation, the corrected NORA3 data will be used to estimate extreme wind maximum with a 50-year return period for 106 offshore locations in the North Sea. The objective of the study is to demonstrate that NORA3 can provide reliable long-term extreme wind estimates to be used in calculation of insurance premiums and in support of more precise turbine designs. The goals of this study supports SDG 7 (Affordable and Clean Energy) and SDG 13 (Climate Action) by contributing to more cost-effective, accurate, and climate-resilient offshore wind energy systems.

References:

Pryor, S. C., & Barthelmie, R. J. (2021). A global assessment of extreme wind speeds for wind energy applications. Nature Energy, 6(3), 268–276. https://www.nature.com/articles/s41560-020-00773-7

Malekmohammadi, Shokoufeh, Etienne Cheynet, and Joachim Reuder. Observation of Kelvin–Helmholtz billows in the marine atmospheric boundary layer by a ship-borne Doppler wind lidar Scientific Reports 15.1 (2025): 5245.

Bastine, D., Larsén, X., Witha, B., Dörenkämper, M., & Gottschall, J. (2018). Extreme winds in the new European wind atlas. Journal of Physics: Conference Series, 1102, 012006. IOP Publishing.

The increasing penetration of wind power into modern electrical grids presents significant challenges for maintaining system stability, particularly in terms of frequency control. Traditionally, frequency support has been provided by synchronous generators with inherent inertia; however, as these are progressively displaced by converter-based renewable generation, wind farms must assume a more active role in delivering ancillary services. This transition necessitates advanced control strategies that go beyond conventional approaches, which often rely on simplified implementations of Wind Power Plant Control (WPPC) and Wind Farm Flow Control (WFFC). Existing models often overlook the intricate aerodynamic and electrical interactions within wind farms, leading to inaccurate estimates of available reserves and a lack of rigorous assessment of grid support capabilities under dynamic conditions.
The paper addresses these limitations by evaluating the impact of integrated wind farm control on the provision of frequency support services within a high-fidelity dynamic simulation environment. The research introduces two novel functionalities designed to enhance both energy capture and system stability. First, an optimised wake-steering algorithm is implemented, employing geometrically derived yaw misalignment setpoints to minimise wake losses and improve overall power output. This approach leverages the FLORIS-based wake model to capture aerodynamic interactions between turbines, enabling a more accurate representation of intra-farm flow dynamics. Second, two dedicated controllers for Fast Frequency Reserves (FFR) and Frequency Containment Reserves (FCR) are developed in MATLAB/Simulink, adhering to the technical specifications of the Nordic Synchronous Area. These controllers incorporate advanced control loops capable of rapid active power modulation, ensuring compliance with stringent response time and stability requirements.
The methodological framework encompasses a series of test scenarios that include low, medium, and high wind speeds, gust events, and varying yaw configurations. The simulated wind farm comprises four 15 MW IEA reference turbines in a 5D grid setting, representing a realistic, albeit minimal, offshore configuration. Simulations are employed to capture transient phenomena with high temporal resolution, while parametric sensitivity analyses quantify the influence of wind conditions and control parameters on reserve availability and frequency response performance.
Results indicate that the integration of optimised wake steering yields a measurable improvement in energy capture, with a 3.11% increase in power output compared to baseline configurations. Furthermore, the FFR and FCR controllers successfully meet activation and performance criteria under normal and disturbed operating conditions, including gust-induced transients. However, scenarios near cut-in wind speeds reveal inherent limitations in reserve provision due to constrained aerodynamic power, underscoring the need for adaptive reserve allocation strategies. Despite these constraints, the implemented frequency support mechanisms maintain operational integrity during critical events, demonstrating robustness against rapid system disturbances.
In conclusion, the findings substantiate that coordinated WPPC and WFFC significantly enhance the capability of wind farms to deliver reliable frequency support. By increasing available reserves and improving dynamic response, the proposed control architecture contributes to the development of converter-dominated power systems that meet future grid stability requirements.