基于电化学性能数据驱动的神经网络预测海水电解液金属空气电池放电终止时间

doi:10.61558/2993-074X.3616

电化学（中英文） ›› 2026, Vol. 32 ›› Issue (8): 2614002. doi: 10.61558/2993-074X.3616

基于电化学性能数据驱动的神经网络预测海水电解液金属空气电池放电终止时间

申鹏鹏^#, 潘奕池^,^#, 刘育荣^,^#, 张露丹, 牛宁, 王冠军^*()(), 杨德坤^*()(), 田新龙^*()(), 饶鹏^*()()

中国海南大学热带海洋工程材料及评价全国重点实验室，信息与通信工程学院，电子科学与技术学院，海南，海口 570228

收稿日期:2026-03-23 修回日期:2026-05-02 接受日期:2026-05-14 发布日期:2026-05-14 出版日期:2026-08-28

Neural Network Driven by Electrochemical Performance Data for Predicting the Discharge Termination Time of Seawater Electrolyte-Based Metal-Air Batteries

Peng-Peng Shen^#, Yi-Chi Pan^,^#, Yu-Rong Liu^,^#, Lu-Dan Zhang, Ning Niu, Guan-Jun Wang^*()(), De-Kun Yang^*()(), Xin-Long Tian^*()(), Peng Rao^*()()

State Key Laboratory of Tropic Ocean Engineering Materials and Materials Evaluation, School of Information and Communication Engineering, School of Electronic Science and Technology Hainan University, Haikou 570228, Hainan, China

Received:2026-03-23 Revised:2026-05-02 Accepted:2026-05-14 Online:2026-05-14 Published:2026-08-28
Contact: *Guan-Jun Wang, E-mail: wangguanjun@hainanu.edu.cn (G. Wang) De-Kun Yang, E-mail: dekun.yang@hainanu.edu.cn (D. Yang) Xin-Long Tian, E-mail: tianxl@hainanu.edu.cn (X. Tian) Peng Rao, E-mail: raopeng@hainanu.edu.cn (P. Rao) E-mail:wangguanjun@hainanu.edu.cn;dekun.yang@hainanu.edu.cn;tianxl@hainanu.edu.cn;raopeng@hainanu.edu.cn
About author:Author Contributions
Peng-Peng Shen, Peng Rao and Xin-Long Tian designed the research. Yu-Rong Liu carried out the experiments and data analysis. Peng-Peng Shen and Yi-Chi Pan designed the neural network algorithm and analyzed the data. Peng-Peng Shen, Yi-Chi Pan, Yu-Rong Liu, Lu-Dan Zhang, Ning Niu, Guan-Jun Wang, Peng Rao, De-Kun Yang and Xin-Long Tian participated in data analysis. Peng-Peng Shen, Peng Rao and Xin-Long Tian wrote the manuscript. Guan-Jun Wang and Xin-Long Tian supervised the project.
^#These authors contributed equally to this work.

1. JOE2614002SI.pdf(442KB)

摘要/Abstract

摘要：

海水电解液基金属空气电池在海洋能源供给系统中展现出巨大应用潜力。然而，传统的统计分析手段虽可应用于海水金属空气电池的寿命预测，却存在预测精度不足、误差偏大的固有局限。本文提出一种基于InceptionTime并融合先验偏置注意力池化的深度时序回归框架，用以构建电化学性能序列与催化剂放电终止时间之间的非线性映射关系。具体而言，采用计时电流曲线提取长期稳定性特征，同时引入源自线性扫描伏安法的先验知识，以强化模型对关键电位区间的注意力权重。在嵌套留一催化剂交叉验证框架下，结合单标准误差准则进行迭代轮次选取，该模型在小样本测试集上，经多种聚合策略验证均表现出高度一致性，决定系数可达0.90以上。研究结果表明，通过注意力驱动的回归框架将多尺度时序特征与电化学先验知识相结合，可显著提升放电终止时间预测的准确性，从而为海水金属空气电池的催化剂评价及未来工程应用提供一种数据驱动范式。

关键词: 海水金属空气电池, 电化学性能, 催化剂失效时间预测, 时间序列回归

Abstract:

Seawater electrolyte-based metal-air batteries exhibit great promise for marine energy supply systems. However, conventional statistical analysis methods, though applicable to seawater metal-air battery lifetime prediction, have inherent limitations of insufficient prediction accuracy and large error. Herein, a deep time-series regression framework based on InceptionTime and incorporating prior-biased attention pooling is proposed to construct a nonlinear mapping between electrochemical performance sequences and the discharge termination time of catalysts. Specifically, chronoamperometric profiles are employed to extract long-term stability features, while prior knowledge derived from linear sweep voltammetry is introduced to strengthen the attention weighting over critical potential regions. Under a nested leave-one-catalyst-out cross-validation framework with the one-standard-error rule for epoch selection, the model demonstrates high consistency across various aggregation strategies on a small-sample test set containing different air cathode catalysts, yielding a coefficient of determination over 0.90. These findings suggest that integrating multiscale temporal features with electrochemical prior knowledge through an attention-driven regression framework can improve the prediction accuracy and reduce the prediction error of discharge termination time in the current dataset, thereby providing a preliminary data-driven approach for catalyst evaluation in seawater metal-air batteries.

Key words: Seawater metal-air batteries, Electrochemical performance, Catalyst failure time prediction, Time series regression

申鹏鹏, 潘奕池, 刘育荣, 张露丹, 牛宁, 王冠军, 杨德坤, 田新龙, 饶鹏. 基于电化学性能数据驱动的神经网络预测海水电解液金属空气电池放电终止时间[J]. 电化学（中英文）, 2026, 32(8): 2614002.

Peng-Peng Shen, Yi-Chi Pan, Yu-Rong Liu, Lu-Dan Zhang, Ning Niu, Guan-Jun Wang, De-Kun Yang, Xin-Long Tian, Peng Rao. Neural Network Driven by Electrochemical Performance Data for Predicting the Discharge Termination Time of Seawater Electrolyte-Based Metal-Air Batteries[J]. Journal of Electrochemistry, 2026, 32(8): 2614002.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://electrochem.xmu.edu.cn/CN/10.61558/2993-074X.3616

https://electrochem.xmu.edu.cn/CN/Y2026/V32/I8/2614002

图/表 6

参考文献 33

[1]	Tian H J, Li Z, Feng G X, Yang Z Z, Fox D, Wang M Y, Zhou H, Zhai L, Kushima A, Du Y G, Feng Z X, Shan X N, Yang Y. Stable, high-performance, dendrite-free, seawater-based aqueous batteries[J]. Nat. Commun., 2021, 12(1): 237. https://doi.org/10.1038/s41467-020-20334-6. doi: 10.1038/s41467-020-20334-6 URL pmid: 33431888
[2]	Zhuang L, Cai W B, Ji H X, Li Q, Wang G W, Xin S, Zhao Q, Cheng F Y, Guo Y G, Mao L Q, Tian Y, Wu F, Zhang L M, Xiang Y, Hu J S, Cao R, Xiao L, Tao H B, Xing W, Zhan D P, Liao H G, Xiao M L, Ren B, Peng Z Q, Wen R, Wang X, Song Y F, Lv H F, Xia B Y, Wang G X, Cheng J, Liu Z P, Zhou M, Huang B, Li C P, Zou Y Q, Wang S Y, Lin H B, Wei Z D. Development trends and priority research fields of electrochemical discipline in the 15th five year plan period[J]. J. Electrochem., 2025, 31(10): 2510081. https://doi.org/10.61558/2993-074X.3590. doi: 10.61558/2993-074X.3590 URL
[3]	Li Y F, Huang A J, Zhou L X, Li B H, Zheng M Y, Zhuang Z W, Chen C, Chen C, Kang F Y, Lv R T. Main-group element-boosted oxygen electrocatalysis of Cu-N-C sites for zinc-air battery with cycling over 5000 h[J]. Nat. Commun., 2024, 15(1): 8365. https://doi.org/10.1038/s41467-024-52494-0. doi: 10.1038/s41467-024-52494-0 URL
[4]	Li Y F, Wang H L, Chen C, Xie X S, Yang Y, Tan X H, Jiang K R, Chen N, Zhang H, Li Z. Ten thousand hour stable zinc air batteries via Fe and W dual atom sites[J]. Nat. Commun., 2025, 16(1): 8085. https://doi.org/10.1038/s41467-025-63540-w. doi: 10.1038/s41467-025-63540-w URL
[5]	Zhao H F, Li L, Zhang T, Yao J Q, Peng X, Peng J, Zhu M, Xu B B, Liu X W, Yu H B. Discovering high-entropy electrocatalysts through a batch-alloy targeting approach[J]. Sci. Adv., 2025, 11(28): eadx6121. https://doi.org/10.1126/sciadv.adx6121.
[6]	Kang D K, Zhang C H, Wang X K, Wang F Q, Gai H Y, Yao H X, Liu X, He Z Z, Huang M H, Jiang H Q. Efficient atomically dispersed Fe catalysts with robust three-phase interface for stable seawater-based zinc-air batteries[J]. Green Carbon, 2025, 3(1): 1-10. https://doi.org/10.1016/j.greenca.2024.09.002. doi: 10.1016/j.greenca.2024.09.002 URL
[7]	Kang W D, Meng S Y, Zhao Y C, Xu J Y, Wu S, Zhao K, Chen S, Niu J F, Yu H T, Quan X. Scaling-free cathodes: enabling electrochemical extraction of high-purity nano-CaCO₃ and -Mg(OH)₂ in seawater[J]. Environ. Sci. Technol., 2024, 58(31): 14034-14041. https://doi.org/10.1021/acs.est.4c04700. doi: 10.1021/acs.est.4c04700 URL
[8]	Fang W H, Xu K L, Wang X L, Zhu Y H, Li X T, Liu H, Li D L, Wu J. Heteroatom-coordinated Fe-N₄ catalysts for enhanced oxygen reduction in alkaline seawater zinc-air batteries[J]. Nano-Micro Lett., 2026, 18(1): 96. https://doi.org/10.1007/s40820-025-01943-6. doi: 10.1007/s40820-025-01943-6 URL
[9]	Xia Y F, Guo P, Li J Z, Zhao L, Sui X L, Wang Y, Wang Z B. How to appropriately assess the oxygen reduction reaction activity of platinum group metal catalysts with rotating disk electrode[J]. iScience, 2021, 24(9): 103024. https://doi.org/10.1016/j.isci.2021.103024. doi: 10.1016/j.isci.2021.103024 URL
[10]	Ruan M B, Liu J, Song P, Xu W L. Meta-analysis of commercial Pt/C measurements for oxygen reduction reactions via data mining[J]. Chin. J. Catal., 2022, 43(1): 116-121. https://doi.org/10.1016/S1872-2067(21)63854-8. doi: 10.1016/S1872-2067(21)63854-8 URL
[11]	Zheng W R, Chen S J. Decoding electrocatalyst degradation using time-resolved electrochemical impedance analysis[J]. JACS Au, 2025, 5(12): 6396-6409. https://doi.org/10.1021/jacsau.5c01196. doi: 10.1021/jacsau.5c01196 URL
[12]	Zhan Y, Ding Z B, He F, Lv X, Wu W F, Lei B, Liu Y, Yan X B. Active site switching of Fe-N-C as a chloride-poisoning resistant catalyst for efficient oxygen reduction in seawater-based electrolyte[J]. Chem. Eng. J., 2022, 443: 136456. https://doi.org/10.1016/j.cej.2022.136456. doi: 10.1016/j.cej.2022.136456 URL
[13]	Chen X Q, Zheng X R, Yin Z X, Lu J D, Wang Y, Guo Y Y, Zhang J F, Wang H Z, Zhao Z W, Wu Y Q, Deng Y D. Pre-adsorption of chlorine enhances the oxyphilic property and oxygen reduction activity of Fe/Se-NC electrocatalyst in seawater electrolyte[J]. Chem. Eng. J., 2024, 482: 148856. https://doi.org/10.1016/j.cej.2024.148856. doi: 10.1016/j.cej.2024.148856 URL
[14]	Vo T G, Gao J J, Liu Y. Recent development and future frontiers of oxygen reduction reaction in neutral media and seawater[J]. Adv. Funct. Mater., 2024, 34(23): 2314282. https://doi.org/10.1002/adfm.202314282. doi: 10.1002/adfm.v34.23 URL
[15]	Houssein E H, Mohamed M, Younis E M G, Mohamed W M. Artificial intelligence and classical statistical models for time series forecasting: a comprehensive review[J]. J. Big Data, 2025, 12(1): 271. https://doi.org/10.1186/s40537-025-01318-z. doi: 10.1186/s40537-025-01318-z URL
[16]	Hu X S, Xu L, Lin X K, Pecht M. Battery lifetime prognostics[J]. Joule, 2020, 4(2): 310-346. https://doi.org/10.1016/j.joule.2019.11.018. doi: 10.1016/j.joule.2019.11.018 URL
[17]	Zhang Y W, Tang Q C, Zhang Y, Wang J B, Stimming U, Lee A A. Identifying degradation patterns of lithium ion batteries from impedance spectroscopy using machine learning[J]. Nat. Commun., 2020, 11(1): 1706. https://doi.org/10.1038/s41467-020-15235-7. doi: 10.1038/s41467-020-15235-7 URL pmid: 32249782
[18]	Hsieh W W. Nonlinear multivariate and time series analysis by neural network methods[J]. Rev. Geophys., 2004, 42(1): RG1003. https://doi.org/10.1029/2002RG000112.
[19]	Kong X J, Chen Z H, Liu W Y, Ning K L, Zhang L C, Muhammad Marier S, Liu Y C, Chen Y H, Xia F. Deep learning for time series forecasting: a survey[J]. Int. J. Mach. Learn. Cybern., 2025, 16(7): 5079-5112. https://doi.org/10.1007/s13042-025-02560-w. doi: 10.1007/s13042-025-02560-w URL
[20]	Zhou J M, Rong J F, Zhang J M, Liu C R, Yi F Y, Jiao Z P, Zhang C Z. Deep learning estimation of state of health for lithium-ion batteries using multi-level fusion features of discharge curves[J]. J. Power Sources, 2025, 653: 237781. https://doi.org/10.1016/j.jpowsour.2025.237781. doi: 10.1016/j.jpowsour.2025.237781 URL
[21]	Hu J C, Fu P Y, Wei Z B, Huang Y J, Early J, Fly A, Zhang Y J. Early prediction of lithium-ion battery degradation with a generative pre-trained transformer[J]. Nat. Commun., 2025, 17(1): 126. https://doi.org/10.1038/s41467-025-66819-0. doi: 10.1038/s41467-025-66819-0 URL
[22]	Rosser D A, Leonard K C. High-speed cyclic voltammetry regressions using machine learning[J]. ACS Electrochem., 2025, 1(7): 1038-1043. https://doi.org/10.1021/acselectrochem.5c00012. doi: 10.1021/acselectrochem.5c00012 URL
[23]	Fawaz H I, Lucas B, Forestier G, Pelletier C, Schmidt D F, Weber J, Webb G I, Idoumghar L, Muller P A, Petitjean F. InceptionTime: Finding AlexNet for time series classification[J]. Data Min. Knowl. Discov., 2020, 34(6): 1936-1962. https://doi.org/10.1007/s10618-020-00710-y. doi: 10.1007/s10618-020-00710-y URL
[24]	Liu Y R, Feng S Y, Shan L T, Zhu Y S, Zhou C C, Li J, Shi X D, Kang Z Y, Tian X L, Rao P. Localized negatively charged interfaces for seawater electrolyte-based zinc-air batteries[J]. Adv. Funct. Mater., 2025, 35(26): 2422874. https://doi.org/10.1002/adfm.202422874. doi: 10.1002/adfm.v35.26 URL
[25]	Liu Y R, Zhu Y S, Zhang Y X, Rao P, Li J, Shan L T, Tang B Y, Shi X D, Kang Z Y, Tian X L. Pyridine nitrogen decorated carbon support for high-performance seawater electrolyte-based zinc-air battery[J]. Angew. Chem., 2025, 137(37): e202509911. https://doi.org/10.1002/ange.202509911. doi: 10.1002/ange.v137.37 URL
[26]	Liu Y R, Zhang M, Yu Y H, Liu Y L, Li J, Shi X D, Kang Z Y, Wu D X, Rao P, Liang Y, Tian X L. Local electric fields coupled with Cl^- fixation strategy for improving seawater oxygen reduction reaction performance[J]. J. Electrochem., 2025, 31(9): 2504132. https://doi.org/10.61558/2993-074X.3566. doi: 10.61558/2993-074X.3566 URL
[27]	Wang Y X, Zhang C H, Wang X K, Duan J R, Tong K C, Dai S X, Chu L, Huang M H. Engineering carbon-chainmail-shell coated Co₉Se₈ nanoparticles as efficient and durable catalysts in seawater-based Zn-air batteries[J]. Acta Phys. Chim. Sin., 2024, 40(6): 2305004. https://doi.org/10.3866/PKU.WHXB202305004. doi: 10.3866/PKU.WHXB202305004 URL
[28]	Luo M C, Koper M T M. A kinetic descriptor for the electrolyte effect on the oxygen reduction kinetics on Pt(111)[J]. Nat. Catal., 2022, 5(7): 615-623. https://doi.org/10.1038/s41929-022-00810-6. doi: 10.1038/s41929-022-00810-6 URL
[29]	Beltrán D E, Litster S. Half-wave potential or mass activity? Characterizing platinum group metal-free fuel cell catalysts by rotating disk electrodes[J]. ACS Energy Lett., 2019, 4(5): 1158-1161. https://doi.org/10.1021/acsenergylett.9b00790. doi: 10.1021/acsenergylett.9b00790 URL
[30]	Wang Q, Li L Y, Yang Y. Knockoff-ML: a knockoff machine learning framework for controlled variable selection and risk stratification in electronic health record data[J]. npj Digit. Med., 2025, 8(1): 723. https://doi.org/10.1038/s41746-025-02102-2. doi: 10.1038/s41746-025-02102-2 URL
[31]	Tan Y Z, Tan W L, Liang Y C, Long Y Z, Chen S H, Hu Q H, Ou Y J, Fu J L, Chen H, Ren F Y, Ye J, Zhou Q, Li S, He X J, Wang Q Q, Shen Y M, Lu H Y, Wu D C, Gao A B, Chen X, Li Y K. Machine learning-enabled spatial multi-omics uncovers lactate-driven targets and tumor microenvironmental reprogramming in cancer[J]. npj Digit. Med., 2025, 9(1): 109. https://doi.org/10.1038/s41746-025-02286-7. doi: 10.1038/s41746-025-02286-7 URL
[32]	Ouyang T C, Gong Y B, Ye J L, Deng Q Y, Su Y Y. State co-estimation for lithium-ion batteries based on multi-innovations online identification[J]. Renew. Sustain. Energy Rev., 2025, 210: 115204. https://doi.org/10.1016/j.rser.2024.115204. doi: 10.1016/j.rser.2024.115204 URL
[33]	Liu Z, Zhao Z H, Qiu Y, Jing B Q, Yang C S, Wu H F. Enhanced state of charge estimation for Li-ion batteries through adaptive maximum correntropy Kalman filter with open circuit voltage correction[J]. Energy, 2023, 283: 128738. https://doi.org/10.1016/j.energy.2023.128738. doi: 10.1016/j.energy.2023.128738 URL

基于电化学性能数据驱动的神经网络预测海水电解液金属空气电池放电终止时间

Neural Network Driven by Electrochemical Performance Data for Predicting the Discharge Termination Time of Seawater Electrolyte-Based Metal-Air Batteries

RichHTML

PDF (PC)

补充材料

可视化

摘要/Abstract

引用本文

使用本文

图/表 6

参考文献 33

相关文章 15

Metrics

[1]	杨云锐, 董欢欢, 郝志强, 何祥喜, 杨卓, 李林, 侴术雷. 高性能锂硫电池用钴/碳复合材料硫宿主[J]. 电化学（中英文）, 2023, 29(4): 2217003-.
[2]	贠潇如, 陈宇方, 肖培涛, 郑春满. 关于水系锌离子电池中无氧钒基正极材料的综述[J]. 电化学（中英文）, 2022, 28(11): 2219004-.
[3]	梁振浪, 杨耀, 李豪, 刘丽英, 施志聪. 基于不同前驱体制备的硬碳负极材料的储锂性能[J]. 电化学（中英文）, 2021, 27(2): 177-184.
[4]	吴凯. Na₃V₂(PO₄)₂O₂F的合成及其在钠离子电池中的应用[J]. 电化学（中英文）, 2021, 27(1): 56-62.
[5]	孙梦雷, 张达奇, 冯金奎, 倪江锋. 钒基电极材料研究进展[J]. 电化学（中英文）, 2019, 25(1): 45-54.
[6]	何大平, 木士春. 质子交换膜燃料电池铂电催化剂稳定策略[J]. 电化学（中英文）, 2018, 24(6): 655-663.
[7]	王鸿辉，马明洁，冯婕，康黄雅，黄文杰. 钕掺杂二氧化铅复合阳极的电化学性能研究[J]. 电化学（中英文）, 2018, 24(4): 367-374.
[8]	张伶潇，赵惠慧，张丽娟，付予. Ag-TiO₂-MnO₂复合材料的制备与电化学性能研究[J]. 电化学（中英文）, 2018, 24(3): 292-299.
[9]	李全一，杨琪，赵艳红. 二氧化钼-碳复合涂层的电化学性能研究[J]. 电化学（中英文）, 2018, 24(2): 160-165.
[10]	王友，曾一文，钟星，刘星，汤泉. 锂离子电池负极材料Li₃V₂(BO₃)₃/C 复合材料的合成及电化学性能研究[J]. 电化学（中英文）, 2018, 24(2): 174-181.
[11]	刘兴亮，杨茂萍，汪伟伟, 曹勇. 分级过程对LiFePO₄/C电池性能的影响[J]. 电化学（中英文）, 2017, 23(6): 661-666.
[12]	罗化峰，乔元栋. 纤维素微孔锂电隔膜的制备及性能研究[J]. 电化学（中英文）, 2017, 23(5): 610-616.
[13]	程琥, 聂晓燕, 申叶丹, . 哌啶型离子液体混合电解液在Li/LiCoO₂电池中的性能研究[J]. 电化学（中英文）, 2017, 23(1): 59-63.
[14]	杨亚雄，马瑞军，高明霞，潘洪革，刘永锋. 晶态Li₁₂Si₇锂离子电池负极材料的电化学性能研究[J]. 电化学（中英文）, 2016, 22(5): 521-527.
[15]	朱立伟*, 阎云海, 贾慧, 梁润芬. 利用均相沉淀剂和形貌导向剂水热一步合成层状镍钴氢氧化物[J]. 电化学（中英文）, 2016, 22(4): 412-416.