Posters

Abstract

This paper conducts a comprehensive comparison between virtual and actual data obtained from an operational wind farm, with a specific focus on validating and correcting wind flow within the designated area. The objective is to enhance the accuracy of wind resource assessments by addressing deviations between predicted and actual wind patterns, leading to more precise results in terms of energy production estimations for potential repowering or overpowering scenarios. The paper particularly emphasizes an overpowering scenario as a real world case study. Given the absence of local wind measurements, the wind climate was characterized using a Vortex MAST dataset. Although Vortex MAST provides reasonable results, it may sometimes under or over-predict wind behaviour in certain regions, leading to associated uncertainties greater than those with local wind measurements. Additionally, practical constraints such as economic, political, or strategic considerations may render a measurement campaign infeasible. In this context, a well-established operational wind farm not only generates electricity but also serves as a vast measurement campaign, recording detailed wind speed and direction information, albeit influenced by wakes of neighbouring wind turbines and of the rotor itself. Utilizing this extensive information to validate a virtual met mast can yield more accurate results for overpowering or repowering scenarios. In this case study, the Vortex MAST was modelled for the site of an existing wind turbine near the overpowering wind turbines, thus allowing a direct comparison between wind speed and direction of Vortex MAST and the Wind Farm's SCADA information over a 5-year period. The energy production figures in the Wind Farm SCADA were corrected for unavailability. Through criteria such as wind speed deviations and Annual Energy Production (AEP) comparisons, corrections were implemented in the wind flow simulation model used (WAsP) to refine the representation of the local wind environment. Before applying the wind flow correction methodology, discrepancies between the average wind speed derived from the virtual mast and the SCADA data were approximately 18%, with a 38% error in gross after-wake productions. The adoption of the methodology outlined in this paper resulted in a significant reduction of these errors to 1.8% and 5.5%, respectively. The validated wind flow data serves as a crucial foundation for conducting overpowering and repowering studies, facilitating a more precise estimation of the wind farm's energy production potential. By aligning theoretical projections with empirical observations, this research contributes to optimizing wind farm performance and aids in the effective early-stage prediction of wind farm projects.