基于特征选择和BP神经网络的风电机组故障分类监测研究

Wind Turbine Fault Classification Based on BP Neural Network and Feature Selection Algorithm

为了减少风电场的经济损失,采用ReliefF特征选择与BP神经网络相结合的方法,对风电机组进行状态监测研究.基于风电场实际运行数据,重点分析了桨距角不对称故障.结果表明:ReliefF特征选择与BP神经网络相结合的方法可以有效地分辨出是否发生了桨距角不对称故障,且准确率较高.

To reduce the maintenance cost of wind farms,a condition monitoring method was proposed for wind turbines based on BP neural network combined with ReliefF feature selection algorithm,with which blade angle asymmetrical faults were analyzed using wind farm operation data.Results show that the proposed method can effectively distinguish the asymmetrical faults with high accuracy.