Real-time modeling technique optimizes power output from wind turbines

The wake flow behind a wind turbine is an additional source of wind energy that can be tapped by properly tuning parameters of a wind turbine array, like the yaw angle or blade pitch of the turbines. To help optimize power output, Naseem Ali and Raúl Bayoán Cal developed a data-driven approach to a dynamical, real-time method for wind turbine control strategies.

Using training data unique to individual wind farms’ operating conditions, the method forecasts fluctuations in wind velocity with time. By continuously updating model parameters using real-time measurements, this data-driven approach allows farms to determine the future state of the system and modify the orientations of turbines and turbine blades accordingly.

“Usually, you take weather forecasting to prepare yourself to go out with an umbrella or a jacket,” said Ali. “This is the exact same thing. We predict what will happen next to control the wind farm to extract more power.”

In testing the method, Ali and Cal took velocity measurements behind a 3-by-3 turbine array inside a wind tunnel. They found they were able to reconstruct the underlying dynamics of the system to simulate its behavior by correcting each prediction with the current measurements.

Though a wind tunnel is much simpler than an outdoor farm, the process is generalizable, and the main mechanism remains the same. Data from different operating conditions can be used to train the model and customize its predictions for a specific location.

“Weather is constantly changing. The direction is constantly changing,” said Cal. “One must be ready to be able to account for those.”

Source: “Data-driven modeling of the wake behind a wind turbine array,” by Naseem Ali and Raúl Bayoán Cal, Journal of Renewable and Sustainable Energy (2020). The article can be accessed at https://doi.org/10.1063/5.0004393 .