大容量海上风电机组运行故障状态识别方法

Fault state identification method for large capacity offshore wind turbines

故障状态识别是海上风电机组运行监测中的重要项目。对于大容量海上风电机组,其故障状态识别难度较高,现行方法难以精准识别,提出大容量海上风电机组运行故障状态识别方法。利用无线传感器感知风电机组运行中风速、电网功率、机舱温度及齿轮箱油温等状态信息,采用小波包分解算法对机组运行数据分解处理,将信号分解成独立的频带,实现大容量海上风电机组运行故障状态识别。经实验证明,设计方法F1 Score值在0.95以上,ROC曲线更靠近左上方,能够实现对大容量海上风电机组运行故障状态的精准识别。

Fault status identification is an important project in the operation monitoring of offshore wind turbines.For large-capacity offshore wind turbines,it is difficult to identify the fault status,and the current method is difficult to accurately identify the fault status.Wireless sensors are used to perceive wind speed,power grid,engine room temperature and gearbox oil temperature during the operation of wind turbines,and wavelet packet decomposition algorithm is adopted to decompose the unit operation data and decompose the signals into independent frequency bands,so as to realize the fault status identification of large-capacity offshore wind turbines.Experiments show that the F1 Score of the design method is above 0.95,and the ROC curve is closer to the upper left,which can realize the accurate identification of the operating fault state of large-capacity offshore wind turbines.