基于状态监测的风电机组变桨系统故障诊断

Pitch system fault diagnosis for wind turbine based on condition monitoring

以风电机组中故障率较高的变桨系统作为研究对象,从数据采集与监视控制系统的数据库中选择与变桨系统运行相关的特征参数,基于相似性原理,利用非线性状态评估方法,建立能够涵盖变桨系统全部正常运行状态的健康模型。当变桨系统发生故障时,会出现模型预测值与正常状态的偏差,根据每一个特征参数对偏差的影响来确定故障的原因。应用实例验证表明,该模型能够准确地识别故障类型,可以解决在排除故障及设备维修时,因缺少相关信息而造成停机时间过长、维修难度大等问题。

The wind turbine pitch system, which has high failure rate, is researched in this paper. First the pitch system feature parameters are selected from the SCADA system database; Then, based on the principle of similarity, the method of nonlinear state assessment is used to establish a health model,which covers all normal operating states of pitch system. When the pitch system fails, it will produce a deviation between the predicted value of model and the normal value, the deviation contribution rate of each characteristic parameter will indicate the type of fault. Finally, an example shows that the model can accurately identify the type of failure so as to reduce the downtime and the difficulty of maintenance caused by lack of relevant information about the accident.