Wind farm set point optimization with surrogate models for load and power output targets

Abstract

This study presents a methodology for wind farm set point optimization, which allows including both loads and power output as optimization criteria. The primary control strategy investigated involves de-rating of individual turbines in the wind farm. The optimal de-rating level of each individual turbine is determined by using accurate and computationally efficient surrogate models, mapping the dependency between the choice of de-rating strategy for a given turbine and the load and power outputs of downwind turbines. A case study based on the Lillgrund offshore wind farm shows that for specific wind directions with strong wake interactions, it may be possible to achieve a net gain in power output in the order of 5% at specific mean wind speeds. Alternatively, maintaining nominal power output but optimizing the load distribution could lead to substantial fatigue load reductions.