Research on Life Prediction of Hydrogen Fuel Cell Based on Bi-LSTM-SSA

Aiming at the problem of life prediction of hydrogen fuel cells, this paper takes proton exchange membrane fuel cells(PEMFC) as the research object, a Life prediction method combining singular spectrum analysis algorithm(SSA) and Bi-directional Long Short-Term Memory(Bi-LSTM) is proposed. Based on t...

Full description

Saved in:
Bibliographic Details
Published in2024 IEEE 5th International Conference on Advanced Electrical and Energy Systems (AEES) pp. 147 - 152
Main Authors Huang, Amin, Dong, Haiying, Sun, Na
Format Conference Proceeding
LanguageEnglish
Published IEEE 29.11.2024
Subjects
Bi-LSTM
Data models
Fuel cells
Indexes
life prediction
Long short term memory
Noise
PEMFC
Prediction algorithms
Predictive models
Protons
Spectral analysis
SSA
Voltage
Online AccessGet full text

Cover

More Information
Summary:Aiming at the problem of life prediction of hydrogen fuel cells, this paper takes proton exchange membrane fuel cells(PEMFC) as the research object, a Life prediction method combining singular spectrum analysis algorithm(SSA) and Bi-directional Long Short-Term Memory(Bi-LSTM) is proposed. Based on the the publicly available long-running data set of PEMFC, the stack voltage of PEMFC is used as the life prediction index. Firstly, SSA is utilized to eliminate the noise signals and retain the main degradation features of the original data. on this basis, a PEMFC voltage prediction model is constructed based on Bi-LSTM. finally, the lifetime of the fuel cell is predicted based on Bi-LSTM-SSA, and the results show that the error of the Bi-LSTM-SSA model is smaller compared to that of the LSTM and the model's coefficient of determination \left(R^{2}\right) is larger, thus verifying that the Bi-LSTM-SSA method proposed in this paper is capable of life prediction of PEMFC with higher accuracy.
DOI:10.1109/AEES63781.2024.10872582