A data-driven method based on SCADA and large-eddy simulation for nacelle north offset diagnosis

Nacelle north offset is a frequently overlooked issue in wind energy. It degrades wind farm performance assessment, causes potential AEP (Annual Energy Production) losses and mechanical fatigue, and hinders the implementation of active wake control strategies. Diagnosing the offset is challenging due to the lack of direct measurement data. In this work, we propose a diagnostic method for nacelle north offset that integrates SCADA (Supervisory Control and Data Acquisition) measurements and large-eddy simulations. The proposed method is based on the assumption of feature similarity among wind turbines in a wind farm (or nearby wind turbines), and achieves the identification and computation of north offsets via farm-level spatiotemporal correlations. The test is conducted in a wind farm located in complex terrain. The results demonstrate that this method can effectively detect nacelle north offset: all 33 wind turbines exhibit such offsets, among which 18 turbines (54.5%) have offsets falling within the range of 90∘≤θN≤270∘. Cross-validation across two wind directions and two months verifies the stability of the method. Further comparison of calibrated farm-averaged SCADA data with ERA5 (European Reanalysis Atmosphere 5) data reveals a high degree of consistency, with the correlation coefficients of wind components exceeding 0.9 for the two months under consideration.