REWIND | Operational maintenance for the wind energy domain
The aim of the REWIND project is an increased productivity and smarter wind energy generation, by optimising via monitoring and modelling techniques enabled by artificial intelligence (AI) and digital twins. To achieve this, the project partners will work on data enrichment, in order to produce high-quality annotated ground truth data and will couple AI solutions for operational monitoring data from wind farms with a multiscale wind resource model chain for advanced fault detection and diagnosis with a view to improving predictive maintenance capabilities. This will make it possible to define smart KPIs that support wind farm operators during their decision-making processes.
The monitoring and diagnostic needs of wind farms are growing significantly more complex and demanding with every passing year. Today, underperformance in wind turbines is typically detected and assessed via a semi-manual top-down approach, i.e. fleet, site, turbine. As a result, monitoring-based fault analysis and diagnosis are time-consuming, expert-dependent and often of insufficient accuracy. This means that multiple underperformance issues and failure modes may either remain undetected, be falsely diagnosed or may not even have their root cause identified.
These limitations are largely due to the low level of automation, an area where AI could help. AI can assist in deducing smart and understandable insights from complex wind turbine machinery involving many nonlinear mechanical, electrical and thermal interactions. However, the lack of high-quality ground truth data for training and evaluating AI models is problematic. In particular, this lack of annotated trustworthy data means that AI solutions cannot currently provide precisely quantified and realistic rates in terms of fault diagnosis accuracy and confidence level.
The outcomes of the project will provide tools and intelligence for the optimisation of onshore and offshore wind farm operations through smart automatic fault detection and advanced diagnosis enabling risk mitigation and empowering decision making. The envisaged solutions include a virtual met mast service, sensor anomaly detection, yaw and blade pitch misalignment analysis, drivetrain fault detection, performance degradation quantification and remaining lifetime estimation of components. These solutions will be validated in an industrially relevant environment (TRL5) via this project.
We will address the issue relying on direct access to over 15 GW of operational wind farm monitoring data collected through SynaptiQ, 3E’s monitoring and asset management platform to which over 10,000 wind and solar farms are connected. In addition, the involvement of a user committee, comprising selected 3E customers, will ensure additional insights on technical issues as well as access to O&M records and field data, and will further enrich the value of the monitoring data.
Joint R&D (The Industry of Tomorrow: Green, Human & Smart): O&M optimization for wind energy generation