基于深度学习的锂电池热失控预警技术研究进展

钟智栋; 倪明; 彭继宇; 王嘉晶; 赵章风; 蒋建东; 应光耀; 马恒; 徐静蕾

doi:10.19799/j.cnki.2095-4239.2026.0145

您当前的位置：

首页 >

文章列表页 >

基于深度学习的锂电池热失控预警技术研究进展

储能测试与评价 | 更新时间：2026-07-28

- 基于深度学习的锂电池热失控预警技术研究进展
- Research progress on deep learning-based thermal runaway early warning technology for lithium-ion batteries
- 储能科学与技术 2026年15卷第7期页码：2882-2894
- 作者机构：
  
  1.浙江华电器材检测研究院有限公司，浙江杭州 310000
  2.浙江工业大学机械工程学院，浙江杭州 310023
- 作者简介：
  
  钟智栋（1992—），男，工程师，研究方向为新能源研究与检测，E-mail：597102440@qq.com。
  彭继宇，副教授，研究方向为光学感知技术和智能控制技术，E-mail：jypeng@zjut.edu.cn。
- 基金信息：
  
  浙江华电器材检测研究院有限公司科技项目(5211HD250009);浙江省“尖兵领雁”研发攻关计划(2025C02001)
- DOI：10.19799/j.cnki.2095-4239.2026.0145
  中图分类号： TM 912
- 收稿：2026-02-10，
  
  修回：2026-05-02，
  
  网络首发：2026-07-28，
  
  纸质出版：2026-07-28
- 稿件说明：
移动端阅览
钟智栋, 倪明, 彭继宇, 等. 基于深度学习的锂电池热失控预警技术研究进展[J]. 储能科学与技术, 2026, 15(7): 2882-2894.

ZHONG Zhidong, NI Ming, PENG Jiyu, et al. Research progress on deep learning-based thermal runaway early warning technology for lithium-ion batteries[J]. Energy Storage Science and Technology, 2026, 15(7): 2882-2894.
钟智栋, 倪明, 彭继宇, 等. 基于深度学习的锂电池热失控预警技术研究进展[J]. 储能科学与技术, 2026, 15(7): 2882-2894. DOI： 10.19799/j.cnki.2095-4239.2026.0145.

ZHONG Zhidong, NI Ming, PENG Jiyu, et al. Research progress on deep learning-based thermal runaway early warning technology for lithium-ion batteries[J]. Energy Storage Science and Technology, 2026, 15(7): 2882-2894. DOI： 10.19799/j.cnki.2095-4239.2026.0145.

摘要

锂电池型号多样、老化程度差异显著，且运行工况复杂，使热失控过程具有较强的不确定性，导致传统预警手段存在误报率高、时效性低等固有局限。深度学习方法具备挖掘复杂热失控机理和特征演化规律的能力，可为锂电池热失控早期预警提供有效技术手段。本文围绕“基于深度学习的锂电池热失控预警技术”展开系统性综述，首先梳理了锂电池热失控关键预警特征、常用数据预处理方法及典型深度学习方法；其次，介绍了深度学习方法在锂电池热失控关键参数预测中的应用进展，并分析了其与物理约束、多物理场耦合模型融合的研究现状；随后，总结了基于热失控参数的多级预警策略与多源特征融合技术的实践路径；最后，针对锂电池热失控预警面临的故障数据稀缺、模型可解释性不足和模型轻量化部署等挑战开展系统讨论，并对未来改进方向进行展望。研究表明，深度学习技术能够有效缓解传统预警手段误报率高、时效性低等问题，对提升锂电池热失控预警的准确性和及时性具有积极作用。本文可为理解深度学习技术在锂电池热失控预警中的作用机制提供理论参考，并为推动其在锂电池安全管理系统中的工程应用、完善锂电池安全管理体系提供支撑。

Abstract

Lithium-ion batteries exhibit diverse configurations suited to operation under complex and dynamic working conditions-including dynamic charge-discharge rates

ambient temperature fluctuations

and mechanical vibrations-causing varying degrees of aging throughout their service life. These factors collectively introduce significant uncertainty into the thermal runaway process. Traditional early warning methods

which rely primarily on fixed thresholds and empirical rules

struggle to adapt to such complexity. Consequently

they suffer from high false alarm rates

missed detections

and insufficient timeliness

failing to capture the subtle precursor signals of thermal runaway. Contrastingly

deep learning methods offer strong nonlinear modeling capabilities and automatic feature extraction

enabling deeper exploration of the complex internal reaction mechanisms of thermal runaway and effective tracking of the dynamic evolution of key parameters during its initiation and propagation. These advantages overcome the limitations of conventional approaches and provide a promising solution for accurate advance warning of lithium-ion battery thermal runaway. This paper systematically reviews recent advances in deep learning-based early warning technologies for lithium-ion battery thermal runaway. First

it outlines key early warning features

common data preprocessing techniques (including cleaning

denoising

normalization

and dataset partitioning)

and the architectures and application scenarios of representative deep learning models. Second

it considers the application of deep learning algorithms in predicting key thermal runaway parameters such as voltage

temperature

gas concentration

state of charge

and state of health. It also discusses emerging frameworks that integrate deep learning with physical constraints and multiphysics coupling models to address the limited interpretability of purely data-driven approaches. Furthermore

this review summarizes practical strategies for multi-level early warning aligned with the staged evolution of thermal runaway

as well as multi-source data fusion methods that combine electrical

thermal

gas

and mechanical signals. Finally

it highlights key challenges hindering practical deployment

including the scarcity of real thermal runaway fault data

limited interpretability of deep learning models

and difficulties in lightweight implementation on resource-constrained embedded systems. Future research directions are proposed accordingly. Overall

it is demonstrated that deep learning has the potential to reduce the high false alarm rates and improves the timeliness and accuracy of thermal runaway early warning. This review provides a theoretical foundation for understanding deep learning applications in this field and supports their integration into battery safety management systems

contributing to enhanced lifecycle safety of lithium-ion batteries.

关键词

Keywords

references

MACHLEV R. EV battery fault diagnostics and prognostics using deep learning: Review, challenges & opportunities[J]. Journal of Energy Storage, 2024, 83: 110614. DOI:10.1016/j.est.2024. 110614.

王海波, 张春强, 林虹, 等. 2024年中国电池行业运行情况[J]. 电池, 2025, 55(1): 11-18. DOI:10.19535/j.1001-1579.2025.01.002.

WANG H B, ZHANG C Q, LIN H, et al. Operation condition of Chinese battery industry in 2024[J]. Battery Bimonthly, 2025, 55(1): 11-18. DOI:10.19535/j.1001-1579.2025.01.002.

王海波, 钱飞宏, 林虹, 等. 2025年中国电池行业运行情况[J]. 电池, 2026, 56(1): 6-15.

WANG H B, QIAN F H, LIN H, et al. Operation status of China's battery industry in 2025[J]. Battery Bimonthly, 2026, 56(1): 6-15.

陈素华, 白莹. 锂离子动力电池热失控机理及热管理技术研究进展[J]. 中国科学基金, 2023, 37(2): 187-198.

CHEN S H, BAI Y. Thermal runway mechanism and research progress on thermal management of lithium-ion power batteries[J]. Bulletin of National Natural Science Foundation of China, 2023, 37(2): 187-198.

李远茂. 动力锂离子电池性能关键估计参数建模与计算方法研究[D]. 广州: 华南理工大学, 2024.LI Y M. Research on modeling and calculation method of key estimation parameters for power lithium-ion battery performance[D]. Guangzhou: South China University of Technology, 2024.

WANG Z L, SADEGHI H, HUANG X Y, et al. Thermal runaway and flame propagation in battery packs: Numerical simulation and deep learning prediction[J]. Engineering Applications of Computational Fluid Mechanics, 2025, 19(1): 2445160. DOI:10.1080/19942060.2024.2445160.

张小龙, 孙承章, 王旭, 等. 基于虚拟防火墙的电化学储能系统热蔓延抑制策略[J]. 太阳能学报, 2025, 46(2): 41-51.

ZHANG X L, SUN C Z, WANG X, et al. Virtual firewall-based thermal spread suppression strategy for electrochemical energy storage system[J]. Acta Energiae Solaris Sinica, 2025, 46(2): 41-51.

朱学慢. 锂电池的热失控研究[D]. 上海: 上海海洋大学, 2021.ZHU X M. Study on thermal runaway of the lithium battery[D]. Shanghai: Shanghai Ocean University, 2021.

HINTON G E, SALAKHUTDINOV R R. Reducing the dimensionality of data with neural networks[J]. Science, 2006, 313(5786): 504-507. DOI:10.1126/science.1127647.

MCCULLOCH W S, PITTS W. A logical calculus of the ideas immanent in nervous activity[J]. 1988: 15-27. DOI:10.5555/65669.104377.

王天城. 电动汽车动力电池热失控状态评估方法及应用研究[D]. 哈尔滨: 哈尔滨工业大学, 2021.WANG T C. Research on evaluation method and application of thermal runaway state of electric vehicle power battery[D]. Harbin: Harbin Institute of Technology, 2021.

栗家傲. 基于深度学习的动力电池异常检测算法研究[D]. 北京: 北京邮电大学, 2023.LI J A. Research on abnormal detection algorithm of power battery based on deeplearning[D]. Beijing: Beijing University of Posts and Telecommunications, 2023.

YAO L, FANG Z P, XIAO Y Q, et al. An intelligent fault diagnosis method for lithium battery systems based on grid search support vector machine[J]. Energy, 2021, 214: 118866. DOI:10.1016/j.energy.2020.118866.

KONG X D, LU L G, YUAN Y B, et al. Foreign matter defect battery and sudden spontaneous combustion[J]. eTransportation, 2022, 12: 100170. DOI:10.1016/j.etran.2022.100170.

YANG H S, LI X Y, FU K, et al. Behavioral description of lithium-ion batteries by multiphysics modeling[J]. DeCarbon, 2024, 6: 100076. DOI:10.1016/j.decarb.2024.100076.

LI F, ZUO W, ZHOU K, et al. State of charge estimation of lithium-ion batteries based on PSO-TCN-Attention neural network[J]. Journal of Energy Storage, 2024, 84: 110806. DOI:10.1016/j.est.2024.110806.

张伟, 杨旭, 黄鑫, 等. 基于PLSR-CNN的锂电池热失控特征气体拉曼检测方法[J]. 消防科学与技术, 2025, 44(4): 525-531. DOI:10.20168/j.1009-0029.2025.04.0525.07.

ZHANG W, YANG X, HUANG X, et al. Analysis of thermal runaway gases in lithium batteries based on Raman PLSR-CNN[J]. Fire Science and Technology, 2025, 44(4): 525-531. DOI:10.20168/j.1009-0029.2025.04.0525.07.

MA M N, WANG Y, DUAN Q L, et al. Fault detection of the connection of lithium-ion power batteries in series for electric vehicles based on statistical analysis[J]. Energy, 2018, 164: 745-756. DOI:10.1016/j.energy.2018.09.047.

LI Z W, LIU F, YANG W J, et al. A survey of convolutional neural networks: Analysis, applications, and prospects[J]. IEEE Transactions on Neural Networks and Learning Systems, 2022, 33(12): 6999-7019. DOI:10.1109/TNNLS.2021.3084827.

BERTONI F, MONTOBBIO N, SARTI A, et al. Emergence of lie symmetries in functional architectures learned by CNNs[J]. Frontiers in Computational Neuroscience, 2021, 15: 694505. DOI:10.3389/fncom.2021.694505.

NAJJAR R. Redefining radiology: A review of artificial intelligence integration in medical imaging[J]. Diagnostics, 2023, 13(17): 2760. DOI:10.3390/diagnostics13172760.

JEON J, CHEON H, JUNG B, et al. ProADD: Proactive battery anomaly dual detection leveraging denoising convolutional autoencoder and incremental voltage analysis[J]. Applied Energy, 2024, 373: 123757. DOI:10.1016/j.apenergy.2024.123757.

CUI B H, WANG H, LI R L, et al. Long-sequence voltage series forecasting for internal short circuit early detection of lithium-ion batteries[J]. Patterns, 2023, 4(6): 100732. DOI:10.1016/j.patter.2023.100732.

TIAN L Y, DONG C Y, WANG R, et al. Anti-interference lithium-ion battery intelligent perception for thermal fault detection and localization[J]. IET Energy Systems Integration, 2024, 6(4): 593-605. DOI:10.1049/esi2.12158.

PEARLMUTTER B A. Gradient calculations for dynamic recurrent neural networks: A survey[J]. IEEE Transactions on Neural Networks, 1995, 6(5): 1212-1228. DOI:10.1109/72.410363.

PASCANU R, MIKOLOV T, BENGIO Y. On the difficulty of training recurrent neural networks[C]//International Conference on Machine Learning, 2012. DOI:10.1007/s12088-011-0245-8.

VAN HOUDT G, MOSQUERA C, NÁPOLES G. A review on the long short-term memory model[J]. Artificial Intelligence Review, 2020, 53(8): 5929-5955. DOI:10.1007/s10462-020-09838-1.

VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Advances in Neural Information Processing Systems 30(NIPS 2017). Long Beach: Curran Associates Inc., 2017: 6000-6010.

CAI L, LI J X, XU X F, et al. Automatically constructing a health indicator for lithium-ion battery state-of-health estimation via adversarial and compound Staked autoencoder[J ] . Journal of Energy Storage, 2024, 84: 110711. DOI:10.1016/j.est.2024. 110711.

蒋建杰, 楼平, 徐国华, 等. 基于预测误差的锂离子电池热失控预警方法研究[J]. 储能科学与技术, 2024, 13(11): 4187-4197.

JIANG J J, LOU P, XU G H, et al. Research on lithium-ion battery thermal runaway early warning method based on prediction error[J]. Energy Storage Science and Technology, 2024, 13(11): 4187-4197.

SUN J, REN S, SHANG Y L, et al. A novel fault prediction method based on convolutional neural network and long short-term memory with correlation coefficient for lithium-ion battery[J]. Journal of Energy Storage, 2023, 62: 106811. DOI:10.1016/j.est. 2023.106811.

CHENG Y M, GAO D X, ZHAO F M, et al. A data-driven early warning method for thermal runaway during charging of lithium-ion battery packs in electric vehicles[J]. Measurement Science and Technology, 2025, 36(1): 016233. DOI:10.1088/1361-6501/ad9d68.

TANG M, ZHANG X, ZHANG C, et al. An intelligent sensing array for thermal runaway characteristic gas concentration prediction based on SACNN-Mamba[J]. Sensors and Actuators B: Chemical, 2025, 431: 137368. DOI:10.1016/j.snb.2025.137368.

HAN C Y, YUAN H M. Capacity prediction of lithium-ion batteries based on adaptive sliding window pooling extreme learning machine algorithm[J]. Physica Scripta, 2024, 99(1): 016003. DOI:10.1088/1402-4896/ad1704.

DAS GOSWAMI B R, MASTROGIORGIO M, RAGONE M, et al. A combined multiphysics modeling and deep learning framework to predict thermal runaway in cylindrical Li-ion batteries[J]. Journal of Power Sources, 2024, 595: 234065. DOI:10.1016/j.jpowsour.2024.234065.

DAS GOSWAMI B R, ABDISOBBOUHI Y, DU H, et al. Advancing battery safety: Integrating multiphysics and machine learning for thermal runaway prediction in lithium-ion battery module[J]. Journal of Power Sources, 2024, 614: 235015. DOI:10.1016/j.jpowsour.2024.235015.

LIN L M, EZEKOYE O A. Time-resolved characterization of toxic and flammable gases during venting of Li-ion cylindrical cells with current interrupt devices[J]. Journal of Loss Prevention in the Process Industries, 2025, 94: 105488. DOI:10.1016/j.jlp.2024. 105488.

黄峥, 秦鹏, 石晗, 等. 过热条件下86Ah磷酸铁锂电池热失控行为研究[J]. 高电压技术, 2022, 48(3): 1185-1191. DOI:10.13336/j.1003-6520.hve.20210126.

HUANG Z, QIN P, SHI H, et al. Study on thermal runaway behavior of 86 Ah lithium iron phosphate battery under overheat condition[J]. High Voltage Engineering, 2022, 48(3): 1185-1191. DOI:10.13336/j.1003-6520.hve.20210126.

ZHANG Q S, LIU T T, HAO C L, et al. In situ Raman investigation on gas components and explosion risk of thermal runaway emission from lithium-ion battery[J ] . Journal of Energy Storage, 2022, 56: 105905. DOI:10.1016/j.est.2022.105905.

WANG H B, XU H, ZHANG Z L, et al. Fire and explosion characteristics of vent gas from lithium-ion batteries after thermal runaway: A comparative study[J]. eTransportation, 2022, 13: 100190. DOI:10.1016/j.etran.2022.100190.

暴祥. 基于多任务学习和对抗域适应的车载锂电池状态估计研究[D]. 包头: 内蒙古科技大学, 2024.BAO X. Research on state estimation of on-board lithium battery based on multi-task learning and adversarial domain adaptation[D]. Baotou: Inner Mongolia University of Science & Technology, 2024.

吕媛媛, 洪颖, 丁斌, 等. 新能源锂离子电池热失控性能研究[J]. 储能科学与技术, 2026, 15(2): 391-397. DOI:10.19799/j.cnki.2095-4239.2025.1162.

LYU Y Y, HONG Y, DING B, et al. Investigation on the thermal runaway performance of new energy lithium-ion cells[J]. Energy Storage Science and Technology, 2026, 15(2): 391-397. DOI:10.19799/j.cnki.2095-4239.2025.1162.

汪红辉, 李嘉鑫, 储德韧, 等. 磷酸铁锂电池存储失效机理及热安全性研究[J]. 储能科学与技术, 2025, 14(5): 1797-1805.

WANG H H, LI J X, CHU D R, et al. Study on the electrochemical performance failure mechanisms and thermal safety of lithium iron phosphate battery during storage conditions[J]. Energy Storage Science and Technology, 2025, 14(5): 1797-1805.

LI J, ZHANG Y L, YUAN B X, et al. A data-driven thermal runaway warning method for lithium-ion batteries under mechanical abuse[J]. IEEE Transactions on Transportation Electrification, 2025, 11(2): 5169-5179. DOI:10.1109/TTE.2024. 3459038.

RAISSI M, PERDIKARIS P, KARNIADAKIS G E. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J]. Journal of Computational Physics, 2019, 378: 686-707. DOI:10.1016/j.jcp.2018.10.045.

CAO R, ZHANG Z J, SHI R W, et al. Model-constrained deep learning for online fault diagnosis in Li-ion batteries over stochastic conditions[J]. Nature Communications, 2025, 16: 1651. DOI:10.1038/s41467-025-56832-8.

ZHU G Y, SUN T, XU Y W, et al. Identification of internal short-circuit faults in lithium-ion batteries based on a multi-machine learning fusion[J]. Batteries, 2023, 9(3): 154. DOI:10.3390/batteries9030154.

FENG X N, ZHANG F S, FENG J, et al. Propagation dynamics of the thermal runaway front in large-scale lithium-ion batteries: Theoretical and experiment validation[J]. International Journal of Heat and Mass Transfer, 2024, 225: 125393. DOI:10.1016/j.ijheatmasstransfer.2024.125393.

KIM S W, KWAK E, KIM J H, et al. Modeling and prediction of lithium-ion battery thermal runaway via multiphysics-informed neural network[J ] . Journal of Energy Storage, 2023, 60: 106654. DOI:10.1016/j.est.2023.106654.

张浩轩, 王寅达, 王一尧, 等. 高功率微波作用下三维结构中多物理场耦合过程仿真方法研究综述[J]. 电波科学学报, 2024, 39(5): 852-869. DOI:10.12265/j.cjors.2024055.

ZHANG H X, WANG Y D, WANG Y Y, et al. Review of simulation method of multiphysics coupling process in three-dimensional structures under high power microwave action[J]. Chinese Journal of Radio Science, 2024, 39(5): 852-869. DOI:10.12265/j.cjors.2024055.

DOMALANTA M R B, PARAGGUA J A D R. A multiphysics model simulating the electrochemical, thermal, and thermal runaway behaviors of lithium polymer battery[J]. Energies, 2023, 16(6): 2642. DOI:10.3390/en16062642.

GAO D X, DU Y R, CHENG Y M, et al. Research on dynamic multi-level warning method for thermal runaway charging of electric vehicles[J]. Engineering Applications of Artificial Intelligence, 2024, 132: 107919. DOI:10.1016/j.engappai.2024. 107919.

王弘. 大容量锂离子电池热失控多参数分级预警技术研究[D]. 武汉: 武汉理工大学, 2024.WANG H. Research on multi-parameter hierarchical early warning technology for thermal runaway of large-capacity lithium-ion batteries[D]. Wuhan: Wuhan University of Technology, 2024.

刘阳阳, 叶晓宇. 基于多参数融合圆柱形三元锂电池热失控预警研究[J]. 汽车与驾驶维修(维修版), 2024(8): 41-43.

LIU Y Y, YE X Y. Research on thermal runaway early warning of cylindrical ternary lithium battery based on multi-parameter fusion[J]. Auto Driving & Service, 2024(8): 41-43.

韩腾飞, 李然, 谢宇嘉, 等. 基于联邦对比学习航天器锂电池健康状态评估[J]. 航天控制, 2025, 43(3): 76-83. DOI:10.3969/j.issn.1006-3242.2025.03.010.

HAN T F, LI R, XIE Y J, et al. Federated contrastive learning based state-of-health estimation of spacecraft lithium batteries[J]. Aerospace Control, 2025, 43(3): 76-83. DOI:10.3969/j.issn.1006-3242.2025.03.010.

GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial netwoks[C]//GHAHRAMANI Z, WELLING M, CORTES C, et al. Advances in Neural Information Processing Systems 27 (NIPS 2014). Montreal: MIT Press, 2014: 2672-2680.

HINTON G E, OSINDERO S, TEH Y W. A fast learning algorithm for deep belief nets[J]. Neural Computation, 2006, 18(7): 1527-1554. DOI:10.1162/neco.2006.18.7.1527.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于人工智能的储能电池热失控早期预警与防控研究

深度学习框架在锂离子电池充电状态估计的应用与优化

基于深度学习的电力系统负荷预测与储能系统优化调度策略研究

融合特征多尺度池化的锂电池健康状态评估算法

溶剂锚定与界面强化协同构筑稳定聚合物锂金属电池