风电机组高斯过程回归塔架振动监测研究

Vibration Monitoring by Gaussian Process Regression for Wind Turbine Towers

根据风电机组的运行原理,对运行数据中记录的塔架振动特征进行分析,发现塔架振动与风电机组数据采集与监视(SCADA)系统记录的多个其他变量存在密切关系,针对风电机组运行数据强随机性和高噪声的特点,采用高斯过程回归方法建立了描述塔架振动与相关变量关系的振动模型,并对该模型进行了验证.结果表明:通过分析塔架模型残差可以实现叶轮桨距角不对称故障的监测和诊断,证明塔架振动监测的有效性.

Based on wind turbine operating theories,an analysis was conducted to the tower vibration characteristics with operation data,and it was found that the tower vibration has close relationship with some variables recorded by the supervisory control and data acquisition(SCADA)system.A vibration model describing the relation between tower vibration and relevant variables was established using Gaussian process regression method according to the characteristics of wind turbine operation data,i.e.strong randomness and high noise,which was subsequently verified.Results show that the asymmetry of rotor blade angle can be monitored and diagnosed by analyzing the tower model residual,thus proving the vibration monitoring method to be effective.