Periodic Reporting for period 1 - WILLOW (Wholistic and Integrated digitaL tools for extended Lifetime and profitability of Offshore Wind farms)
Reporting period: 2023-10-01 to 2025-03-31
Current approaches applied to wind farm (WF) control operate in a static, open-loop manner, avoiding the need for knowledge about available power but lacking the benefits of recent wind farm control and optimal operation planning research. Studies on these issues are limited and have not been fully integrated into farm-level decision-making. In offshore, the importance of structural health monitoring (SHM) for addressing issues like corrosion and fatigue degradation must be highlighted, and the role of smarter models for assessing remaining useful life (RUL) and condition levels (CL) is key.
The WILLOW project aims to tackle the complexity of integrating component degradation and grid integration by focusing on power dispatch control and leveraging recent research for multiscale, multidisciplinary modelling solutions.
WILLOW integrated system will provide an open-source, data-driven smart curtailment solution to the Wind Farm Operators with the basis of an integrated Wind Farm Control system looking for a trade-off between the power production and the lifetime consumption. With this aim, WILLOW pretends to design a novel Structural Health Monitoring System able to provide high quality data to perform a reliable fleet life assessment using physical models and AI methods which will be used for decision-making and maintenance scheduling. This will contribute to the reduction of the LCOE and to the increase of the AEP towards the current trend of design and operating life of offshore wind farms with up to 20 MW turbines beyond 50 years.
The WILLOW project aims to tackle the complexity of integrating component degradation and grid integration by focusing on power dispatch control and leveraging recent research for multiscale, multidisciplinary modelling solutions.
WILLOW integrated system will provide an open-source, data-driven smart curtailment solution to the Wind Farm Operators with the basis of an integrated Wind Farm Control system looking for a trade-off between the power production and the lifetime consumption. With this aim, WILLOW pretends to design a novel Structural Health Monitoring System able to provide high quality data to perform a reliable fleet life assessment using physical models and AI methods which will be used for decision-making and maintenance scheduling. This will contribute to the reduction of the LCOE and to the increase of the AEP towards the current trend of design and operating life of offshore wind farms with up to 20 MW turbines beyond 50 years.
The most significant achievements reached during the first 18 months of the project are:
WP1: Specifications, use case definition, cybersecurity guidelines & environmental/economical assessment
• To populate a use case definition document addressing the key objectives of each work package in the WILLOW project and thus not solely the use case definition by the end-user.
• The definition of the design requirements and data for WFC strategies and O&M scheduling.
• A comprehensive cybersecurity strategy was developed focusing on three key pillars: Industrial Control Systems security, Identity and Access Management for IoT devices, and Security in the Software Development Life Cycle. This work provides a robust basis for enhancing Wind Farm cybersecurity, the path forward involved practical implementation of these strategies, evaluation of their effectiveness in real-world scenarios, and the subsequent development of a set of best practices.
WP2: Novel Structural Health Monitoring system combined with drone-based inspections
• Virtual sensing of wind turbine support structure loads, most importantly bending moments to consider the effect of both aerodynamic and hydrodynamic loads: creation of SubDyn/OpenFAST model of Norther NRTC3 wind turbine support structure, implementation of Ritz vector-based modal expansion done by Wölfel and publication of an operational dataset for public use https://zenodo.org/records/11093262(opens in new window).
• The construction of a 1:75 scale physical model of an offshore wind turbine based on geometric and material data provided by VUB to investigate the sensitivity of higher-order vibration modes to localized structural defects and the deployment of accelerometers for this analysis done by TSI.
• The deployment of C-CUBE sensors and coated coupons at the Blue Accelerator, with the first data coming in and waiting for interpretation. CEIT worked on the development of two different models based on ultrasound data to estimate the coating degradation: modelling the reflection approach of the medium and a CNN model.
• SIRRIS has made notable advancements in the use of electrochemical techniques to study sample behavior under various test sequences. Significant correlations were made between OCP, LPR, ECN, and EFM. Pitting corrosion was identified by common peaks across the techniques. CEIT has made significant strides in ultrasound testing, evaluating various transducers on different coupon materials. A major achievement here is the clear differentiation observed in the echo’s temporal shape between pitted and non-pitted areas.
• CEIT and ALERION have proposed a novel methodology to correlate/combine both technologies thermography as drone inspections and ultrasounds as continuous monitoring with the aim of detecting and quantifying the uniform corrosion of structures.
WP3: Data driven farm-wide corrosion and load prognosis, lifetime assessment & novelty detection
• A metadatabase platform which facilitates digitization of structural design information has been developed by VUB and 24SEA.
• For the evaluation of the remaining useful lifetime based on monitoring data different extrapolation methods have been compared by VUB.
• Towards the fleetwide fatigue load prognosis a physics informed neural network (PINN) has been trained and validated on the Norther data.
• FMAKE investigated the use of machine learning modelling for predictive coating degradation and corrosion activity on the instrumented assets and fleetwide extrapolation.
• Preliminary research towards the identification of novelties, anomalies, in monitoring signals by machine learning has been conducted by VUB to directly identify possible undesirable structural responses by evaluating acceleration data only.
WP4: Fleet Life Assessment and Integrated Wind Farm Control
• The farmwide insights with updated loads in WP3 (damage equivalent load predictions at interface level) serves as the basis for health aware turbine controller actions in WP4.
• A plausible scenario for the 2040 power system in the North Sea region, with a characterisation of wind power curtailment and other aspects of the system.
• First public release of FLAggTurb, an open-source farm-level synthetic turbulence generator developed at SINTEF.
WP1: Specifications, use case definition, cybersecurity guidelines & environmental/economical assessment
• To populate a use case definition document addressing the key objectives of each work package in the WILLOW project and thus not solely the use case definition by the end-user.
• The definition of the design requirements and data for WFC strategies and O&M scheduling.
• A comprehensive cybersecurity strategy was developed focusing on three key pillars: Industrial Control Systems security, Identity and Access Management for IoT devices, and Security in the Software Development Life Cycle. This work provides a robust basis for enhancing Wind Farm cybersecurity, the path forward involved practical implementation of these strategies, evaluation of their effectiveness in real-world scenarios, and the subsequent development of a set of best practices.
WP2: Novel Structural Health Monitoring system combined with drone-based inspections
• Virtual sensing of wind turbine support structure loads, most importantly bending moments to consider the effect of both aerodynamic and hydrodynamic loads: creation of SubDyn/OpenFAST model of Norther NRTC3 wind turbine support structure, implementation of Ritz vector-based modal expansion done by Wölfel and publication of an operational dataset for public use https://zenodo.org/records/11093262(opens in new window).
• The construction of a 1:75 scale physical model of an offshore wind turbine based on geometric and material data provided by VUB to investigate the sensitivity of higher-order vibration modes to localized structural defects and the deployment of accelerometers for this analysis done by TSI.
• The deployment of C-CUBE sensors and coated coupons at the Blue Accelerator, with the first data coming in and waiting for interpretation. CEIT worked on the development of two different models based on ultrasound data to estimate the coating degradation: modelling the reflection approach of the medium and a CNN model.
• SIRRIS has made notable advancements in the use of electrochemical techniques to study sample behavior under various test sequences. Significant correlations were made between OCP, LPR, ECN, and EFM. Pitting corrosion was identified by common peaks across the techniques. CEIT has made significant strides in ultrasound testing, evaluating various transducers on different coupon materials. A major achievement here is the clear differentiation observed in the echo’s temporal shape between pitted and non-pitted areas.
• CEIT and ALERION have proposed a novel methodology to correlate/combine both technologies thermography as drone inspections and ultrasounds as continuous monitoring with the aim of detecting and quantifying the uniform corrosion of structures.
WP3: Data driven farm-wide corrosion and load prognosis, lifetime assessment & novelty detection
• A metadatabase platform which facilitates digitization of structural design information has been developed by VUB and 24SEA.
• For the evaluation of the remaining useful lifetime based on monitoring data different extrapolation methods have been compared by VUB.
• Towards the fleetwide fatigue load prognosis a physics informed neural network (PINN) has been trained and validated on the Norther data.
• FMAKE investigated the use of machine learning modelling for predictive coating degradation and corrosion activity on the instrumented assets and fleetwide extrapolation.
• Preliminary research towards the identification of novelties, anomalies, in monitoring signals by machine learning has been conducted by VUB to directly identify possible undesirable structural responses by evaluating acceleration data only.
WP4: Fleet Life Assessment and Integrated Wind Farm Control
• The farmwide insights with updated loads in WP3 (damage equivalent load predictions at interface level) serves as the basis for health aware turbine controller actions in WP4.
• A plausible scenario for the 2040 power system in the North Sea region, with a characterisation of wind power curtailment and other aspects of the system.
• First public release of FLAggTurb, an open-source farm-level synthetic turbulence generator developed at SINTEF.
The main results identified in WILLOW with potential impacts are:
- Closed-loop probabilistic decision-making tool for wind farm control based on calculated remaining useful lifetime and damage equivalent loads.
- Open-source smart curtailment tool.
- Multi-objective active power control algorithms complying grid requirements and preventing assets degradation in curtailed operations.
- Curtailed operation forecast scenarios to assess offshore wind ancillary power reserve service provision versus power maximisation.
- Population based novelty detection methods to identify anomalies within the windfarm.
- Framework and algorithms for the continuous consumed life (CL) + remaining useful life (RUL) monitoring of the monitored structures
- Methodology and algorithms for the continuous corrosion prognosis at wind farm level
- Virtual sensing and digital twin software tool to determine loads due to waves and current in offshore wind turbine structures.
- Tool (sensors & algorithms) to monitor and forecast coating degradation on the wind turbine foundation/substructure.
- Innovative solutions to monitor pit initiation and growth: method (physical and digital) based on ultrasounds and electrochemical sensors.
- AI/ML approaches with fused information for automatic classification of damages.
- Closed-loop probabilistic decision-making tool for wind farm control based on calculated remaining useful lifetime and damage equivalent loads.
- Open-source smart curtailment tool.
- Multi-objective active power control algorithms complying grid requirements and preventing assets degradation in curtailed operations.
- Curtailed operation forecast scenarios to assess offshore wind ancillary power reserve service provision versus power maximisation.
- Population based novelty detection methods to identify anomalies within the windfarm.
- Framework and algorithms for the continuous consumed life (CL) + remaining useful life (RUL) monitoring of the monitored structures
- Methodology and algorithms for the continuous corrosion prognosis at wind farm level
- Virtual sensing and digital twin software tool to determine loads due to waves and current in offshore wind turbine structures.
- Tool (sensors & algorithms) to monitor and forecast coating degradation on the wind turbine foundation/substructure.
- Innovative solutions to monitor pit initiation and growth: method (physical and digital) based on ultrasounds and electrochemical sensors.
- AI/ML approaches with fused information for automatic classification of damages.