AI Moves Wind Farms Collectively to Improve Performance

Researchers from the University of Illinois devised a method for making wind farms move collectively in order to maximize their performance, a press statement reveals.

Besides feeling like an apt metaphor for the need for widespread collaboration in the global fight against climate change, the University of Illinois team's work has the potential to greatly improve wind farm efficiency, further incentivizing renewable energy adoption.

As wind passes through a turbine, it slows down as it transfers its energy, and effectively creates a wake that reduces the average downwind velocity.

When every turbine is optimized to achieve the best possible results for itself, it can create a problem for the wind farm as a whole as it is not controlled in a way that accounts for the reduced downwind velocity.

In a paper published in the Journal of Renewable and Sustainable Energy, the researchers from the University of Illinois detail their method for controlling upstream turbines in a manner that prevents downstream turbines from being adversely affected.

The researchers designed controllers that view the wind farm system as a coupled network, allowing them to generate renewable energy more efficiently. 

"If you think of a wind farm as a group of turbines each vying for the incoming wind, if every turbine is greedy and tries to maximize its own power, the system as a whole is suboptimal," said author Lucas Buccafusca. "Our work seeks to design controls for turbines to work collectively, thereby improving performance."

Wake steering implementation for higher-performance wind farms

For their research, the University of Illinois team applied a model predictive control (MPC) framework that analyzed shifting wind speeds and applies wake steering techniques.

These techniques are the key to improving the overall performance of wind farms, as they aim the upstream turbine's wakes away from turbines further downstream.

Using simulations, the researchers found that by using artificial intelligence algorithms that account for the downstream effect of turbines, they could noticeably increase performance.

"When observing the power extractions, it is surprising just how much the gains can be for even small wind turbine arrays simply by implementing wake steering techniques," said Buccafusca.

Other innovations in wind turbines, that stand to increase renewable energy uptake, include GE's floating wind turbines that are set to enable widespread expansion of wind farms in the oceans. Another concept from startup Alpha 311 aims to harness gusts of winds from traffic on highways.

The University of Illinois team believes their method is a worthy addition to the latest wind farm innovations. If applied to future wind turbine control algorithms, it stands to improve humanity's capacity to reverse the adverse effects of climate change by providing more energy than previously possible via renewables.