| Title: |
Machine learning to rapidly predict turbine yaw angles for wake steering |
| Authors: |
Stanley, Andrew P. J.; Mulder, Tim; Doekemeijer, Bart; Kreeft, Jasper |
| Source: |
Journal of Physics: Conference Series ; volume 2767, issue 8, page 082011 ; ISSN 1742-6588 1742-6596 |
| Publisher Information: |
IOP Publishing |
| Publication Year: |
2024 |
| Description: |
Wake steering is an important control strategy to boost power production of a wind farm. Because of computational expense and problem complexity, wind farm layouts are typically optimized assuming they will operate without wake steering. However, performance gains are possible by simultaneously optimizing wind farm layout and the yaw angles for wake steering. In this paper, we present a method to train a machine learning model to predict turbine yaw angles as a function of their position relative to other turbines in the wind farm and the inflow wind speed. This model is able to predict turbine yaw angles with an R 2 value of 0.98. The model also produces turbine yaw angles with wind farm power production that is similar to yaw angles that have been directly optimized. This method to rapidly compute optimal turbine yaw angles for wake steering enables control co-design of wind farms and the associated performance increase. |
| Document Type: |
article in journal/newspaper |
| Language: |
unknown |
| DOI: |
10.1088/1742-6596/2767/8/082011 |
| DOI: |
10.1088/1742-6596/2767/8/082011/pdf |
| Availability: |
https://doi.org/10.1088/1742-6596/2767/8/082011; https://iopscience.iop.org/article/10.1088/1742-6596/2767/8/082011; https://iopscience.iop.org/article/10.1088/1742-6596/2767/8/082011/pdf |
| Rights: |
https://creativecommons.org/licenses/by/4.0/ ; https://iopscience.iop.org/info/page/text-and-data-mining |
| Accession Number: |
edsbas.F079C285 |
| Database: |
BASE |