Trinity researchers to use machine learning techniques to optimise wind farm performance

Researchers from Trinity, led by Dr Breiffni Fitzgerald, Ussher Assistant Professor in the School of Engineering, will use Science Foundation Ireland (SFI) funding of around €500,000 with machine learning techniques to optimise wind farm performance.

The funding, awarded to Professor Fitzgerald via SFI’s Frontiers for the Future programme, will support the MeLodiC project. The project will assess the aerodynamics of wind farms and develop new controllers to optimise wind farm power production. The new controllers will use operational wind farm data to adaptively “learn” the best way to control wind farms.

Speaking about MeLodiC, Professor Fitzgerald said:

“This award will enable me to build a team of researchers in the Department of Civil Engineering to investigate how we can model and control wind farms better using first principles-based aerodynamics coupled with data driven techniques.

Currently each turbine in a wind farm is controlled to maximise its own individual performance, but this ignores the effect each turbine has on the other turbines in the wind farm, which is a sub-optimal approach.

“MeLodiC will instead develop new wind farm models that take aerodynamic wake interaction effects into account and these new models will help us develop new wind farm controllers with the goal of holistically optimising power production.”

Advanced control algorithms will be integrated with powerful machine learning techniques (using Trinity’s high performance computing facilities) to adaptively learn the best way to control wind farms and, ultimately, make a greater contribution to sustainable energy production.

SFI’s Frontiers for the Future programme offers opportunities for independent investigators to conduct highly innovative, collaborative research with the potential to deliver impact, while also providing opportunities for high-risk, high-reward research projects.