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2

动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	42.24	20.07	0.975
	Transformer	35.77	14.93	0.982
	LSTM-TS	15.17	7.73	0.997
	CNN-LSTM	29.8	25.2	0.987
	CNN	14.8	7.79	0.997
60	LSTM	40.931	21.46	0.974
	Transformer	32.06	14.08	0.984
	LSTM-TS	15.43	8.39	0.996
	CNN-LSTM	65.8	45.3	0.932
	CNN	26.2	12.1	0.990
70	LSTM	51.27	27.15	0.955
	Transformer	34.6	14.99	0.980
	LSTM-TS	15.89	8.96	0.996
	CNN-LSTM	55.1	37	0.949
	CNN	48.3	28.4	0.961
80	LSTM	64.17	35.25	0.946
	Transformer	41	16.01	0.978
	LSTM-TS	20.68	10.89	0.994
	CNN-LSTM	33.6	28.3	0.985
	CNN	69.2	42.2	0.939

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.