AIMWIND

Description: About 65 GW of onshore wind turbine installations in Europe will reach end-of-design-life by 2028. It is time for the operators to decide on one of the three end-of-life scenarios, namely, decommissioning, lifetime extension, or repowering. The last two options will increase the operating life and thus reduce lifecycle costs. These end-of-life decisions require careful consideration of the accumulated fatigue life of each turbine in a wind farm to minimize monetary risk for the wind farm operators. Today, this decision is primarily based on a single point assessment by the certification authority. AIMWind proposes a continuous evaluation of wind farm health based on big data analytics using multimodal data such as wind, operational data, weather, condition monitoring, and inspection logs across a wind farm. Conventional approaches to fatigue estimation are slow and inadequate to achieve these goals, especially in large wind farms. Such a continuous health assessment will facilitate not only accurate life predictions but also continuous improvement of wind turbine operations to ensure long life and high availability.

The project AIMWind will take a three-pronged approach.

1. We will also extend the condition monitoring systems as existing systems today focus only on a selected set of components providing incomplete health information.
2. We will develop big data analytics using a fusion of physics-based models and novel deep-learning techniques to adequately estimate the accumulated fatigue in real-time, which does not exist today. We will use NORCOWE wind measurements, reference wind farm data, and other open data sources to achieve this (more details in the proposal document).
3. We develop health-aware control technologies to achieve the dual objectives of efficiency and long life.

Thus, AIMWind plans to build the essential knowledge to reliable and efficient wind farm operation and improved chances for lifetime extension and repowering.

For more information check the project web site and the project page at the Research Council of Norway.