Health Status Assessment for Wind Turbine with Recurrent Neural Networks

Table 7

Performance criterions of the traditional LSTM neural network for wind turbine 2.

SCADA variables	RMSE	MAE	BIAS	SDE
Generator speed	0.296	0.173	-0.147	0.256
Generator temperature	0.056	0.037	0.015	0.056
Generator cooling air temperature	0.076	0.045	0.13	0.075
Generator bearing temperature	0.057	0.043	0.019	0.054
Gearbox temperature	0.035	0.028	0.017	0.031
Gearbox bearing temperature	0.066	0.052	-0.02	0.063
Rotor speed	0.3	0.178	-0.151	0.242
Battery box temperature	0.105	0.1	0.1	0.033
Hydraulic pressure	0.087	0.046	-0.015	0.085
Shaft bearing temperature	0.11	0.11	-0.11	0.026
Top box temperature	0.03	0.025	-0.005	0.03