An improved data-driven methodology and field-test verification of yaw misalignment calibration on wind turbines

Abstract

Yaw misalignment is being gradually recognized as a critical aspect for repowering aged wind turbines. For large-scale wind farms, the detection and calibration of yaw misalignment need to balance the relationship among accuracy, efficiency, and cost. This paper presents a data-driven yaw misalignment calibration method, and implements the field test with slight hardware modification. The specific procedures, including data preprocessing, regionalization and inference are listed step by step, which sustain the calibration direction and value of yaw misalignment calibration. Then, the historical data are extracted from system database of six adjacent commercial 2 MW wind turbines, and the proposed data-driven calibration method is applied, where the nacelle-mounted LiDAR is also introduced for auxiliary calibration and comparison. Field verification is implemented in these six wind turbines, through implementing pre-defined yaw offsets. The performance evaluations are conducted from different aspects, and compared with the same period of previous year. Consistent results show that both power production and yaw control accuracy are improved significantly, while the data-driven calibration method is with the same direction and acceptable deviation as the LiDAR. Hence, the proposed yaw misalignment calibration could be considered a propagable repowering technique with advantages of low cost, high efficiency, and acceptable accuracy.