Insights into wind fields from Lidar measurements, powered by TNO Wind Energy and UbiOps

  • We are happy to announce that TNO chooses to work with UbiOps for their machine learning algorithms to reconstruct wind velocity fields from Lidar measurements.
Photo over EWTW wind test site, Wieringermeer, The Netherlands. Credit: Gerben Bergmann

Photo over EWTW wind test site, Wieringermeer, The Netherlands. Credit: Gerben Bergmann

Why wind fields?

Understanding wind conditions is central to many design activities, from large buildings to wind farms. TNO performs measurements both on- and offshore, with a particular focus on assisting with the design of large offshore wind farms.

Traditionally, measuring the wind velocity required building a tall mast to reach the height of interest, and then installing a small cup anemometer. However, Light Detection and Ranging (“Lidar”) wind sensing devices are able to measure wind at a distance, and in many places, depending on how the laser beam is pointed. This offers a step-change in information. The only problem is that raw Lidar data does not directly measure the full wind velocity field, so TNO has since 2017 been working on applying machine learning to this challenge.

Wind field measurement at Maasvlakte II, Rotterdam, The Netherlands. Read the report here.

 Wind field measurement at Maasvlakte II, Rotterdam, The Netherlands. Read the report here.

Move your machine learning algorithms into production

TNO’s machine learning algorithms have proved to be highly effective at this task; however, they now need to ‘move into production’: to be able to quickly process the very large amounts of data generated, but at a low cost. Moreover, the data processing pipelines must be modular, reproducible, scalable and quick to get started with. UbiOps enables them to do so, without requiring DevOps skills.

“By collaborating with UbiOps we can efficiently process large amounts of Lidar data using our state-of-the-art machine learning algorithms. Their platform allows us to develop and deploy our concepts in a professional way, improving both accessibility and reliability. We have found them to be a very supportive partner, very keen to adapt and extend their platform’s capabilities to support our rather specific needs.”
Jan Willem Wagenaar – Portfolio Manager, TNO Wind Energy

We are really happy to support TNO Wind Energy running their sophisticated machine learning algorithms to help grasp wind fields from large amounts of Lidar data. We expect more great analytics applications from TNO in the near future and we are happy to play a supportive part in this journey.

Yannick MalthaCo-founder and CEO at UbiOps