Dynamic conditions-categorical boosting-long short-term memory network via penalized huber loss for accurate state of health estimation of proton exchange membrane fuel cells

Jiang, Zhe, Yang, Bo, Liang, Boxiao, Zhang, Haoyan, Shi, Xinqi, Wang, Xiaosa, Li, Hongbiao, Gao, Dengke and Jiang, Lin ORCID: 0000-0001-6531-2791 (2026) Dynamic conditions-categorical boosting-long short-term memory network via penalized huber loss for accurate state of health estimation of proton exchange membrane fuel cells SUSTAINABLE ENERGY TECHNOLOGIES AND ASSESSMENTS, 88. 104941-. ISSN 2213-1388, 2213-1396

Text Cleaned Manuscript z jiang.pdf - Author Accepted Manuscript Available under License Creative Commons Attribution. Download (1MB) | Preview

Abstract

Proton exchange membrane fuel cells (PEMFCs) have a relatively limited operational lifespan, thus necessitating accurate and reliable monitoring of their state of health (SoH). Existing SoH estimation methods generally suffer from insufficient rigor in algorithm selection and neglect of loss function design. To address these issues, this study proposes a hybrid estimation framework integrating the Penalized Huber Loss (PHL), Categorical Boosting (CatBoost) and Long Short-Term Memory (LSTM) network, denoted as PHL-CatBoost-LSTM. This framework fully leverages CatBoost’s superiority in handling categorical features, LSTM’s capability in capturing temporal dependencies, as well as the strong adaptability and anti-interference performance of the PHL loss function. Experimental results demonstrate that the proposed method achieves a root mean square error (RMSE) of 0.0011 on the FC2 dataset, which indicates an extremely small estimation deviation and meets the requirements of practical engineering applications. These findings confirm that the proposed method provides a practical and high-precision solution for SoH monitoring of PEMFCs, which is of great significance for extending the service life of fuel cells and improving the reliability of their engineering applications.

Item Type:	Article
Uncontrolled Keywords:	Proton exchange membrane fuel cells, State of health, Dynamic conditions, Loss function, SimuNPS
Divisions:	Faculty of Science & Engineering Faculty of Science & Engineering > School of Engineering Faculty of Science & Engineering > School of Engineering > Electrical Engineering and Electronics
Depositing User:	Symplectic Admin
Date Deposited:	29 Apr 2026 09:55
Last Modified:	16 Jun 2026 05:59
DOI:	10.1016/j.seta.2026.104941
Related Websites:	Author Publisher
URI:	https://livrepository.liverpool.ac.uk/id/eprint/3198214
Disclaimer:	The University of Liverpool is not responsible for content contained on other websites from links within repository metadata. Please contact us if you notice anything that appears incorrect or inappropriate.