基于多通道振动主元特征的风电机组叶片自监督异常识别方法

王博特,王卿,刘强,金波

Self-supervised anomaly recognition method for wind turbine blade based on multi-channel vibration principal features

Bote WANG,Qing WANG,Qiang LIU,Bo JIN

表 2 不同模型振动主元特征回归性能对比

Tab.2 Comparison of regression performance of principal vibration characteristics of different models

模型名称 MSE MAE RMSE SVR 0.0574 0.1784 0.2396 MLP 0.0262 0.1162 0.1619 GRU 0.0334 0.1226 0.1829 Transformer 0.0103 0.0920 0.1147 ABT-AK 0.0141 0.1047 0.1433 ABT-BD 0.0089 0.0881 0.0974 ABT 0.0054 0.0523 0.0728