CT has completed the IADGENOL project on automating AWE System trajectories using deep learning.

29 May 2024

CT has successfully completed the IADGENOL R&D project aimed at researching deep learning models for the automatic control and characterization of AWES (Airborne Wind Energy Systems) for wind energy generation. Over the last two years, the CT team has investigated the use of state-of-the-art Deep Learning technologies to enhance control and gain a deeper understanding of wind energy generation systems known as AWES.

“AWES represents an innovative approach to harnessing wind at high atmospheric layers, but they face significant challenges, particularly in operational autonomy. Currently, achieving robust autonomous control is one of the main hurdles; the systems must operate completely autonomously and withstand variable weather conditions over long periods. To enhance the adaptability of traditional control systems, we studied the use of machine learning and reinforcement learning,” explains Pablo Egea Hervás, project manager at CT.

The primary objective of the IADGENOL initiative was the creation of a Deep Learning-based control model for the automatic trajectory control of AWE systems, as well as employing these models to understand and characterize the dynamic challenges faced by such systems. This project has been entirely carried out by CT using its own resources and the shared AWES test machine available through a collaborative agreement with Carlos III University of Madrid for the development of AWES technologies. The tasks performed range from preliminary research on the state of the art, data mining and its processing for later use in the development of models, and the development itself of data-based models and the controller, split into four phases.

During last month, the AWES control system developed by CT and Carlos III University of Madrid underwent real-operation validation testing after being successfully trained in a simulation environment using reinforcement learning algorithms and an incremental learning methodology. This system, which teaches the aircraft to maintain its position in the air under various conditions and to maximize the tension of the cables, demonstrated its ability to align the kite with the wind and perform figure-eight trajectories, thus optimizing energy generation.

The preliminary results have been presented at the European AWES congress, AWEC2024, in Madrid, highlighting the potential of these technologies to develop robust and adaptive controllers, although further research in this field is acknowledged as necessary.