A FUEL CELL DEGRADATION TREND PREDICTION METHOD BASED ON PATCHTST MODEL

Shi Yong; Hu Zhilong; Xie Di; Wang Liangliang; Yao Jigang; Su Jianhui

doi:10.19912/j.0254-0096.tynxb.2024-1762

PDF(1544 KB)

Acta Energiae Solaris Sinica ›› 2026, Vol. 47 ›› Issue (2) : 761-767. DOI: 10.19912/j.0254-0096.tynxb.2024-1762

A FUEL CELL DEGRADATION TREND PREDICTION METHOD BASED ON PATCHTST MODEL

Shi Yong¹, Hu Zhilong¹, Xie Di², Wang Liangliang², Yao Jigang², Su Jianhui¹

Author information +

History +

Abstract

This paper proposes a fuel cell aging trend prediction method based on the PatchTST model. By segmenting time series data into multiple local time windows and combining the Transformer architecture to capture long-and short-term dependencies, the method achieves precise prediction of fuel cell aging trends. In the experiments, the model was trained under steady-state and quasi-dynamic operating conditions using data with training set proportions of 50%, 60%, and 70%, respectively, to predict aging trends for future horizons of 50 h, 100 h, and 150 h. Analysis based on error evaluation metrics, such as URMSE and UMAE, indicates that the model achieves the lowest prediction error when the training set proportion is 60% for the FC1 dataset and 50% for the FC2 dataset. Although the error increases slightly as the prediction horizon extends, the overall performance remains relatively stable. Under the conditions where the FC1 dataset is divided by 50% and 60% ratios, the prediction errors of the PatchTST model are lower than those of the Informer, Transformer, GRU, and LSTM models.

Key words

fuel cell / time series analysis / degradation / prediction / self-attention mechanism / PatchTST

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

Shi Yong, Hu Zhilong, Xie Di, Wang Liangliang, Yao Jigang, Su Jianhui. A FUEL CELL DEGRADATION TREND PREDICTION METHOD BASED ON PATCHTST MODEL[J]. Acta Energiae Solaris Sinica. 2026, 47(2): 761-767 https://doi.org/10.19912/j.0254-0096.tynxb.2024-1762

References

[1] 陈家城, 周苏. 大功率质子交换膜燃料电池建模及仿真[J]. 太阳能学报, 2024, 45(3): 290-297.
CHEN J C, ZHOU S.Modeling and simulation of high-power proton exchange membrane fuel cells[J]. Acta energiae solaris sinica, 2024, 45(3): 290-297.
[2] CHEN H C, PEI P C, SONG M C.Lifetime prediction and the economic lifetime of proton exchange membrane fuel cells[J]. Applied energy, 2015, 142: 154-163.
[3] 李从心, 岳美玲, 李昕彤, 等. 基于条件神经网络的质子交换膜燃料电池的老化性能预测[J]. 储能科学与技术, 2024, 13(9): 3094-3102.
LI C X, YUE M L, LI X T, et al.Prediction of aging performance of proton exchange membrane fuel cells based on conditional neural networks[J]. Energy storage science and technology, 2024, 13(9): 3094-3102.
[4] LIU J, GAO Y, SU X, et al.Disturbance-observer-based control for air management of PEM fuel cell systems via sliding mode technique[J]. IEEE transactions on control systems technology, 2018, 27(3): 1129-1138.
[5] BURLATSKY S F, GUMMALLA M, O'NEILL J, et al. A mathematical model for predicting the life of polymer electrolyte fuel cell membranes subjected to hydration cycling[J]. Journal of power sources, 2012, 215: 135-144.
[6] ZHOU D, WU Y, GAO F, et al.Degradation prediction of PEM fuel cell stack based on multiphysical aging model with particle filter approach[J]. IEEE transactions on industry applications, 2017, 53(4): 4041-4052.
[7] LIU H, CHEN J, HISSEL D, et al.Remaining useful life estimation for proton exchange membrane fuel cells using a hybrid method[J]. Applied energy, 2019, 237: 910-919.
[8] MA R, YANG T, BREAZ E, et al.Data-driven proton exchange membrane fuel cell degradation predication through deep learning method[J]. Applied energy, 2018, 231: 102-115.
[9] MA R, BREAZ E, LIU C, et al.Data-driven prognostics for PEM fuel cell degradation by long short-term memory network[C]//2018 IEEE Transportation Electrification Conference and Expo. Long Beach, CA, USA, 2018: 102-107.
[10] 薛阳, 燕宇铖, 贾巍, 等. 基于改进灰狼算法优化长短期记忆网络的光伏功率预测[J]. 太阳能学报, 2023, 44(7): 207-213.
XUE Y, YAN Y C, JIA W, et al.Photovoltaic power prediction model based on IGWO-LSTM[J]. Acta energiae solaris sinica, 2023, 44(7): 207-213.
[11] 莫易敏, 余自豪, 叶鹏, 等. 基于迁移学习与GRU神经网络结合的锂电池SOH估计[J]. 太阳能学报, 2024, 45(3): 233-239.
MO Y M, YU Z H, YE P, et al.Lithium battery SOH estimation method based on combination of transfer learning and GRU neural network[J]. Acta energiae solaris sinica, 2024, 45(3): 233-239.
[12] 柏帆, 王路达, 左红群, 等. 储能用质子交换膜燃料电池长期老化预测[J]. 电池, 2024, 54(2): 160-164.
BAI F, WANG L D, ZUO H Q, et al.Long-term aging prediction of proton exchange membrane fuel cells for energy storage[J]. Battery, 2024, 54(2): 160-164.
[13] 吴航宇, 王玮, 朱文超, 等. 基于ARIMA-BiGRU双数据驱动的燃料电池性能退化预测方法[J]. 中国电机工程报, 2025, 45(7): 2690-2699.
WU H Y, WANG W, ZHU W C, et al.Performance degradation prediction of fuel cells based on ARIMA-BiGRU dual data-driven method[J]. Proceedings of the CSEE Engineering, 2025, 45(7): 2690-2699.
[14] 李浩, 李浩, 杨扬, 等. 基于改进鲸鱼算法优化GRU的PEMFC老化预测[J]. 中国电机工程学报, 2024, 44(20): 8166-8178.
LI H, LI H, YANG Y, et al.Aging prediction of PEMFC optimized by improved whale algorithm and GRU[J]. Proceedings of the CSEE, 2024, 44(20): 8166-8178.
[15] NIE Y, NGUYEN N H, SINTHONG P, et al. A time series is worth 64 words: long-term forecasting with transformers[J]. arXiv, 2022, preprint arXiv: 2211-14730.
[16] ZENG A, CHEN M, ZHANG L, et al.Are transformers effective for time series forecasting?[C]//Proceedings of the AAAI Conference on Artificial Intelligence. Washington DC, USA, 2023, 37(9): 11121-11128.
[17] VASWANI A, SHAZEER N, PARMAR N, et al.Attention is all you need[J]. Advances in neural information processing systems, 2017, 30: 6000-6010.
[18] FCLAB Research. IEEE PHM2014 data challenge[EB/OL]. http://eng.fclab.fr/ieee-phm-2014-data-challenge/.
[19] 潘诗媛, 华志广, 王光伟, 等. 基于级联回声状态网络的氢燃料电池剩余使用寿命预测[J]. 中国电机工程学报,2025, 45(12): 4718-4728.
PAN S Y, HUA Z G, WANG G W, et al.Remaining useful life prediction of hydrogen fuel cells based on cascade echo state network[J]. Proceedings of the CSEE, 2025, 45(12): 4718-4728.
[20] WILLMOTT C J, MATSUURA K.Advantages of the mean absolute error (MAE) over the root mean square error (RMSE) in assessing average model performance[J]. Climate research, 2005, 30(1): 79-82.