Project

The project aims to develop wind farm models based on data and artificial intelligence algorithms. The model and data will support the design of intelligent control algorithms for wind farms. This modeling method is used to solve the problem that existing models cannot be used for actual wind farm control. The model uses machine learning models to learn high-fidelity model data to improve the performance of low-fidelity models. So as to achieve the balance between the fidelity required by the control algorithms and the computational cost. The wind farm control algorithm based on this model aims to improve the power production and turbine life of the total farm by intelligent wake redirection. The wind power industry will also benefit from the development of artificial intelligence algorithms. Reinforcement learning is used to design intelligently optimized controllers for wind farms.

This project enables Danish participation in IEA Wind Task 44: Farm Flow Control. The focus is on control strategies to mitigate wake effects in wind farms. The purpose of IEA Wind Task 44 is to coordinate international research in the field of wind field control inside wind farms. The technology used for this task covers a wide range, but focuses primarily on control algorithms and strategies and how they are transferred to real-world operational improvements.

The intention is to bring together ongoing research results as well as best industry practice, create an overview of control strategies and algorithms and investigate how uncertainties affect the performance and potential for implementation of wind farm control.
The result is guidance for the wind industry and researchers on the current control algorithms, requirements, barriers to adoption, future directions and expected benefits of wind farm control.