Scaled testing of maximum-reserve active power control

Abstract

We present the scaled experimental validation of an active power control (APC) algorithm designed to optimize the tracking accuracy of a wind farm in the presence of turbulent wind lulls. This is obtained by maximizing the minimum local power availability (called reserve) across the farm. This method combines an offline-computed open-loop set-point scheduler aimed at the power reserves, with a fast closed-loop corrector tasked with enhancing tracking accuracy. The open-loop component is synthesized using an augmented engineering wake model, which captures the combined effects of yaw misalignment and power curtailment, wind-tunnel-specific inflow characteristics, and the influence of the chord-based Reynolds number (caused by the small size of the models) on the performance of the rotors. A preliminary simulation-based steady-state analysis indicates that the approach effectively enhances local power availability across the wind farm by leveraging both induction and wake steering control. The methodology is then experimentally demonstrated in a large boundary-layer wind tunnel with minimal blockage, in different scenarios, including dynamically varying wind direction conditions. Its performance is benchmarked against three alternative APC strategies sourced from the literature. During the experiments, the APC algorithms run in real time on a dedicated cabinet and communicate with the turbine controllers via a dedicated network. Dynamic wind direction changes are generated using a large turntable driven by a field-recorded wind direction time series, scaled to match the temporal dynamics of the wind tunnel. Results show that in both static and dynamic scenarios, the proposed control strategy outperforms the reference methods in power tracking accuracy, particularly under high power-demand conditions, while maintaining a limited impact on structural fatigue. Notably, the algorithm effectively manages wake interactions and redistributes local power demands to increase local power reserves, thereby mitigating saturation effects and improving overall tracking performance.