Remaining useful life prediction method of PEMFC based on joint feature extraction of LSTM and improved Transformer | Xibei Gongye Daxue Xuebao/Journal of Northwestern Polytechnical University

Open Access

表1

静态工况不同训练集多步预测量化结果

训练集占比/%	预测方法	E_RMS	E_MA	R²
50	LSTM	160.4	124.8	-0.02
	Transformer	37.52	19.44	0.998
	LSTM-TS	151.2	120.5	0.099
	CNN-LSTM	194.2	172.4	-0.49
	CNN	251.9	221	0.15
60	LSTM	145.9	107.1	-0.21
	Transformer	36.69	17.62	0.92
	LSTM-TS	76.42	54.52	0.67
	CNN-LSTM	116.1	104.3	0.24
	CNN	155.6	124.7	0.27
70	LSTM	48.14	28.7	0.83
	Transformer	41.5	18.7	0.88
	LSTM-TS	67.9	47.4	0.67
	CNN-LSTM	70.4	54.9	0.64
	CNN	101	77	0.48
80	LSTM	34.62	16.7	0.90
	Transformer	34.8	16.8	0.90
	LSTM-TS	15.1	8.91	0.98
	CNN-LSTM	18.1	13.5	0.98
	CNN	16	18.6	0.98

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.