LiDAR-assisted closed-loop control of a wind farm

Abstract A common approach to perform wake steering control in wind farms, involves using pre-calibrated, low-fidelity wake models. Due to their low computation time, these models can be efficiently used to precalculate the yaw angles that maximize the total wind farm power across a wide range of wind conditions. This “open-loop” control strategy, however, is highly dependent on the accuracy of the model. Any errors in the model can lead to suboptimal power output. To overcome this issue, a “closed-loop” control method can be used, which leverages available measurements to adjust certain model parameters, ensuring a closer alignment with the actual operation of the wind farm. To implement this closed-loop approach, a new control-oriented wake model is presented, designed to be adapted in real time using measurements from the wind farm. These measurements are assumed to come from nacelle-mounted multi-range LiDAR sensors, which provide partial wind field data. The entire closed-loop control scheme is validated using the medium-fidelity dynamic simulator FAST.Farm. Several wind condition scenarios are tested, all showing a significant increase in the total power output of the wind farm.