EP EP4796782A1
The present invention concerns a control method (100) for controlling a wind farm comprising a plurality of wind turbines.The method comprises a learning phase (1100) and an exploitation phase (1200).During the learning phase receiving (1110) a model architecture is received. The model architecture comprises an attention comprising at least on multi-head self-attention neural network. The learning phase further comprises training (1120) the model architecture according to a reinforcement learning technic implying tuning the tunable parameters of the model architecture to obtain a trained model optimizing a performance of the wind farm. The performance of the wind farm depends on an estimated power output by the wind farm, and losses due to wake effect,The exploitation phase comprises receiving (1210) estimated wind velocities and directions. The exploitation phase further comprises applying (1230) the trained model to derive rotations commands, and sending (1240) the rotation commands to wind turbines.
A document turning up in a search is not the same as a document that anticipates your claims. Our registered Patent Agents read the claims, not just the abstract, and tell you where you still have room.
Protect Your InventionBibliographic data via the European Patent Office's Open Patent Services (DOCDB) . Republished here for research and prior-art review. This page is not legal advice, and its presence in our index says nothing about the validity or enforceability of the document. For the Indian legal position (status, oppositions, renewals), see InPASS.