数据驱动风电机组偏航误差在线智能识别研究

Research on on-line intelligent identification of yaw error of data-driven wind turbine

针对不同风速下风向仪动态特性、风轮尾流、风向仪安装误差等因素导致的风电机组偏航误差问题,文章采用基于运行数据驱动的风电机组偏航误差方法进行在线智能识别。该方法通过改进DBSCAN聚类方法剔除过度离群数据,采用移动最小二乘法拟合“风速-功率-偏航误差”特性曲面,识别出不同风速下的偏航误差曲线,结合在线运行数据采集,可以实现不同风速下偏航误差的动态识别和持续矫正。算例分析表明,与偏航误差设定值相比,在有限数据下识别的偏航误差的识别结果较为准确,且识别误差在合理范围内。该方法的应用能够更为精确识别不同风速下风电机组偏航误差,进一步提高风电机组发电效率。

An online intelligent identification method based on operational data is used to identify the yaw error of wind turbines caused by the dynamic characteristics of anemometer,wind turbine wake,and anemometer installation error at different wind speeds.The method uses a modified DBSCAN clustering method to remove excessive outliers and a moving least squares method to fit the"wind speed-power-yaw error"characteristic surface to identify the yaw error curve at different wind speeds.Combined with the online operation data collection,the dynamic identification and continuous correction of yaw error at different wind speeds are achieved.The analysis of the example shows that the identified yaw error under limited data is close to the set value of yaw error,and the identification error is within a reasonable range.The application of this technology and method can more accurately identify the yaw error of wind turbine under different wind speeds,and further improve the generation efficiency of wind turbine.