Posters

Abstract

Traditionally, wind energy research has predominantly focused on modeling wind speed, often overlooking the intricate nuances of wind direction, which play a crucial role in understanding losses attributed to curtailments. This study addresses this gap by adopting a comprehensive approach that explores various methodologies, including Circular Statistics, to model wind direction in long-term assessments. The research identifies circular regression as the most suitable model for long-term predictions, with a significant enhancement using angle dispersion to improve the accuracy of long-term wind rose assessments, resulting in a remarkable 1.12% Root Mean Square Error (RMSE) in frequency predictions per bin. Furthermore, the study delves into vertical extrapolation and interpolation of wind direction, examining the application of veer for each time step and comparing it with the use of mean veer values only. The findings reveal a substantial improvement of 84.73% when utilizing veer per timestep, with a reduction of almost 2.04° in RMSE in the results. The implications of these modeling advancements are assessed by evaluating the difference in Wind Sector Management (WSM) strategy estimates in an operational wind farm, yielding a significant reduction in the RMSE of WSM percentage loss estimates from 2.38% to 0.55%. This highlights the importance of the improvement and utilization of directional modeling in enhancing estimates of WSM-related losses.