Attention Mechanism-Based CNN-LSTM Model for Wind Turbine Fault Prediction Using SSN Ontology Annotation

Table 5

Characteristics of wind turbine icing fault after screening.

Features	Score
yaw_position	0.147249
environment_tmp	0.115538
power	0.097059
pitch3_angle	0.092650
int_tmp	0.090929
pitch3_moto_tmp	0.081600
pitch2_moto_tmp	0.076074
pitch1_moto_tmp	0.073881
pitch2_angle	0.067191
pitch1_angle	0.057764
wind_speed	0.037283
generator_speed	0.035642