[1]金阳,张潇,王垒,等.AI 驱动的锂离子电池状态估计与故障诊断[J].郑州大学学报(工学版),2026,47(6):1-14.[doi:10. 13705 / j. issn. 1671-6833. 2026. 06. 019]
　JIN Yang,ZHANG Xiao,WANG Lei,et al.AI-Driven Lithium-Ion Battery State Estimation and Fault Diagnosis[J].Journal of Zhengzhou University (Engineering Science),2026,47(6):1-14.[doi:10. 13705 / j. issn. 1671-6833. 2026. 06. 019]
点击复制

AI 驱动的锂离子电池状态估计与故障诊断()
分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
47
期数:
2026年6期
页码:
1-14
栏目:
出版日期:
2026-11-30

文章信息/Info

Title:
AI-Driven Lithium-Ion Battery State Estimation and Fault Diagnosis
文章编号:
1671-6833(2026)06-0001-14
作者:
金阳1, 张潇1, 王垒2, 何志超2, 李秀勇1, 刘豪杰1,2
1. 郑州大学 电气与信息工程学院,河南 郑州 450001;2. 北京海博思创科技股份有限公司,北京 100084
Author(s):
JIN Yang1, ZHANG Xiao1, WANG Lei2, HE Zhichao2, LI Xiuyong1, LIU Haojie1,2
1. School of Electrical and Information Engineering, Zhengzhou University, Zhengzhou 450001, China; 2. Beijing Hyper Strong Science and Technology Co. , Ltd. , Beijing 100084, China
关键词:
锂离子电池;  电池状态估计;  故障诊断;  热失控预警;  人工智能
Keywords:
lithium-ion battery;  state estimation;  fault diagnosis;  thermal runaway warning;  artificial intelligence
分类号:
TM911TP18
DOI:
10. 13705 / j. issn. 1671-6833. 2026. 06. 019
文献标志码:
A
摘要:
大规模电网储能技术的发展,使得锂离子电池状态估计与故障诊断已成为保障储能系统安全高效运行的核心技术。 针对传统建模方法在复杂电化学老化机制与动态工况干扰下参数辨识难度大、跨工况泛化能力弱等显著瓶颈,以人工智能( artificial intelligence, AI)算法演进为主线进行系统概述。 首先,系统综述了 AI 在锂离子电池荷电状态( state of charge, SOC) 、健康状态( state of health, SOH) 及剩余使用寿命( remaining useful life, RUL) 估计中的应用进展,并深入剖析了各类方法的性能差异与局限;然后,进一步分析了 AI 在内短路、锂枝晶及热失控等典型故障诊断辨识与早期预警中的应用潜力;最后,总结了当前方法在可解释性、小样本适应性、跨工况泛化及算力约束等方面面临的核心挑战,并探讨了物理信息神经网络( physics-informed neural network, PINN) 、迁移学习、数字孪生及大语言模型( large language model, LLM)等新兴技术的发展趋势,为下一代高可信、强泛化电池管理系统算法研发提供了理论参考。
Abstract:
With the development of large-scale grid energy storage technology, state estimation and fault diagnosisof lithium-ion batteries were regarded as the core to guarantee the safe and efficient operation of energy storage systems. To address the prominent bottlenecks of traditional modeling methods, such as significant difficulties in parameter identification and weak cross-condition generalization ability under complex electrochemical aging mechanisms and dynamic operating conditions, taking the evolution of AI algorithms as the main thread, the applicationprogress of AI in the estimation of state of charge ( SOC) , state of health ( SOH) and remaining useful life (RUL)of lithium-ion batteries was systematically reviewed. The performance differences and limitations of various methodswere critically analyzed. Furthermore, the application potential of AI in the diagnosis, identification and earlywarning of typical faults such as internal short circuit, lithium dendrite growth and thermal runaway was analyzed.Finally, the core challenges confronted by current methods in interpretability, few-shot adaptability, cross-conditiongeneralization and computational power constraints were systematically summarized, and the development trends ofemerging technologies including physics-informed neural network ( PINN) , transfer learning, digital twin and largelanguage model ( LLM) were discussed pertinently, thereby providing a theoretical reference for the research anddevelopment of algorithms for the next-generation highly reliable and strongly generalizable battery management systems.

参考文献/References:

[1] Dunn B, Kamath H, Tarascon J M. Electrical energystorage for the grid: a battery of choices [ J] . Science,2011, 334(6058) : 928-935.

[2] Xia Xiangyang, Tan Xinxin, Shan Zhouping, et al. Keytechnology and development prospect of ontology safety forlithium-ion battery storage power stations [ J ]. ElectricPower, 2024, 57(11): 1-17. [夏向阳, 谭欣欣, 单周平, 等. 储能电站锂离子电池本体安全关键技术及新技术应用情况[J]. 中国电力, 2024, 57(11): 1-17. ]
[3] Hu Dunan, Huang Sheng, Wen Zhen, et al. A review onthermal runaway warning technology for lithium-ion batteries [ J ] . Renewable and Sustainable Energy Reviews,2024, 206: 114882.
[4] Xiong Rui, Yang Ruixin, Chen Zeyu, et al. Online faultdiagnosis of external short circuit for lithium-ion batterypack[ J] . IEEE Transactions on Industrial Electronics,2020, 67(2) : 1081-1091.
[5] Song Lei, Lu Chunguang, Liu Lin, et al. State of chargeestimation of LiFePO4battery based on modified amperhour integral method[ J] . Journal of Zhengzhou University(Engineering Science) , 2023, 44( 6) : 84- 90. [ 宋磊,陆春光, 刘琳, 等. 基于修正安时积分法的磷酸铁锂电池荷电 状 态 估 计 [ J] . 郑 州 大 学 学 报 ( 工 学 版) ,2023, 44(6) : 84-90. ]
[6] Severson K A, Attia P M, Jin N, et al. Data-driven prediction of battery cycle life before capacity degradation[ J] . Nature Energy, 2019, 4(5) : 383-391.
[7] Zhang Yongzhi, Xiong Rui, He Hongwen, et al. Longshort-term memory recurrent neural network for remaininguseful life prediction of lithium-ion batteries [ J] . IEEETransactions on Vehicular Technology, 2018, 67 ( 7 ) :5695-5705.
[8] Ren Lei, Zhao Li, Hong Sheng, et al. Remaining usefullife prediction for lithium-ion battery: a deep learning approach[ J] . IEEE Access, 2018, 6: 50587-50598.
[9] Thakuri S K, Li Huibo, Ruan Diwang, et al. The RULprediction of Li-ion batteries based on adaptive LSTM[ J] . Journal of Dynamics, Monitoring and Diagnostics,2025, 4(1) : 53-64.
[10] Yang Fangfang, Li Weihua, Li Chuan, et al. State-of-charge estimation of lithium-ion batteries based on gatedrecurrent neural network [ J ] . Energy, 2019, 175:66-75.
[11] Song Xiangbao, Yang Fangfang, Wang Dong, et al.Combined CNN-LSTM network for state-of-charge estimation of lithium-ion batteries[ J] . IEEE Access, 2019, 7:88894-88902.
[12] Fan Yaxiang, Xiao Fei, Li Chaoran, et al. A novel deeplearning framework for state of health estimation of lithium-ion battery[ J] . Journal of Energy Storage, 2020, 32:101741.
[13] Ren Lei, Dong Jiabao, Wang Xiaokang, et al. A data-driven auto-CNN-LSTM prediction model for lithium-ionbattery remaining useful life [ J] . IEEE Transactions onIndustrial Informatics, 2021, 17(5) : 3478-3487.
[14] Liang Haifeng, Yuan Peng, Gao Yajing. Remaining useful life prediction of lithium-ion battery based on CNN-Bi-LSTM network [ J ] . Electric Power Automation Equipment, 2021, 41(10) : 213-219. [梁海峰, 袁芃, 高亚静. 基于 CNN-Bi-LSTM 网络的锂离子电池剩余使用寿命预测[ J] . 电力自动化设备, 2021, 41(10) : 213-219. ]
[15] Zhu Bo, Liu Xueqin, Wang Y H, et al. SOC predictionfor lithium-ion batteries based on EMD-LSTM-attention[ J] . Journal of Power Supply, 2026, 24(3) : 242-248.[朱博, 刘雪芹, 王宇珲, 等. 基于 EMD-LSTM-Attention 的锂离子电池 SOC 预测[ J] . 电源学报, 2026, 24(3) : 242-248. ]
[16] Sherkatghanad Z, Ghazanfari A, Makarenkov V. A self-attention-based CNN-Bi-LSTM model for accurate state-of-charge estimation of lithium-ion batteries [ J] . Journal of Energy Storage, 2024, 88: 111524.
[17] Tang Aihua, Jiang Yihan, Yu Quanqing, et al. A hybridneural network model with attention mechanism for stateof health estimation of lithium-ion batteries [ J] . Journal of Energy Storage, 2023, 68: 107734.
[18] Wang Zhuqing, Liu Ning, Chen Chilian, et al. Adaptive self-attention LSTM for RUL prediction of lithium-ion batteries[ J] . Information Sciences, 2023, 635: 398-413.
[19] Li Lei, Li Yuanjiang, Mao Runze, et al. Remaining useful life prediction for lithium-ion batteries with a hybridmodel based on TCN-GRU-DNN and dual attention mechanism[ J] . IEEE Transactions on Transportation Electrification, 2023, 9(3) : 4726-4740.
[20] Li Jie, Kang Longyun, Yue Rui, et al. Lithium-ion battery SOC estimation method based on the Transformer[ J] . Power Electronics, 2023, 57(7) : 64-66. [李杰,康龙云, 岳 睿, 等. 基 于 Transformer 的 锂 离 子 电 池SOC 估计方法[ J] . 电力电子技术, 2023, 57(7) : 64-66. ][21] Geng Xinyue, Hu Changhua, Zheng Jianfei, et al. Remaining useful life prediction of lithium batteries based onTransformer under the dual time scales [ J ] . AerospaceControl and Application, 2023, 49( 4) : 119 - 126. [ 耿鑫月, 胡昌华, 郑建飞, 等. 双时间尺度下基于 Transformer 的锂电池剩余寿命预测[ J] . 空间控制技术与应用, 2023, 49(4) : 119-126. ]
[22] Gu Xinyu, See K W, Li Penghua, et al. A novel state-of-health estimation for the lithium-ion battery using aconvolutional neural network and transformer model[ J] .Energy, 2023, 262: 125501.
[23] Jia Chenyu, Tian Yukai, Shi Yuanhao, et al. State ofhealth prediction of lithium-ion batteries based on bidirectional gated recurrent unit and transformer [ J] . Energy,2023, 285: 129401.
[24] Zhou Kate Qi, Qin Yan, Yuen Chau. Graph neural network-based lithium-ion battery state of health estimationusing partial discharging curve [ J ] . Journal of EnergyStorage, 2024, 100: 113502.
[25] Li Long, Yan Xumeng, Zhang Yusheng, et al. Lithiumbattery state of charge estimation method based on transferlearning[ J] . Journal of Xi′an Jiaotong University, 2023,57(11) : 142-150. [李龙, 燕旭朦, 张钰声, 等. 采用迁移学习的小样本锂电池荷电状态估算方法[ J] . 西安交通大学学报, 2023, 57(11) : 142-150. ]
[26] Chen Yuan, Li Jiale, Liu Yanzhong, et al. Fusion ofphysical model and deep learning for SOC estimation oflithium-ion batteries [ J ] . Journal of Wuhan University(Natural Science Edition) , 2025, 71 ( 6) : 782 - 792.[陈媛, 李佳乐, 刘炎忠, 等. 融合物理模型和深度学习算法的锂 离 子 电 池 SOC 估 计 [ J] . 武 汉 大 学 学 报(理学版) , 2025, 71(6) : 782-792. ]
[27] Yu Tianjian, Zeng Xiaoyan, Feng Enlai, et al. A Joint estimation of SOC-SOH for lithium batteries based on LSTM-Transformer multi-channel feature fusion [ J ]. Journal of Railway Science and Engineering, 2026, 23 ( 1): 301 -313. [于天剑, 曾笑颜, 冯恩来, 等. 基于 LSTM-Transformer 多通道特征融合的锂电池 SOC-SOH 联合估计[J]. 铁道科学与工程学报, 2026, 23(1): 301-313. ]
[28] Pang Songyan, Xiao Chuanliang, Peng Ke, et al. Jointestimation of lithium-ion battery state based on light gradient boosting machine [ J] . Automation of Electric Power Systems, 2025, 49(20) : 200-207. [庞松岩, 肖传亮,彭克, 等. 基于轻量级梯度提升机的锂离子电池状态联合估计[ J] . 电力系统自动化, 2025, 49(20) : 200-207. ]
[29] Wu Wentao, Qiao Dongdong, Wang Xueyuan, et al.Electrical and thermal characteristics of early stage internal short circuit of lithium-ion batteries based on electro-chemical mechanism model[ J] . Journal of University ofShanghai for Science and Technology, 2024, 46(1) : 10-20. [吴文涛, 乔冬冬, 王学远, 等. 基于电化学机理模型的锂离子电池早期内短路电热特征研究[ J] . 上海理工大学学报, 2024, 46(1) : 10-20. ]
[30] Kang Sangwan, Tu Hao, Fang Huazhen. BattBee: equivalent circuit modeling and early detection of thermal runaway triggered by internal short circuits for lithium-ion batteries[ J] . Applied Energy, 2026, 404: 127016.
[31] Qiu Yishu, Dong Ti, Lin Da, et al. Fault diagnosis forlithium-ion battery energy storage systems based on localoutlier factor [ J ] . Journal of Energy Storage, 2022,55: 105470.
[32] Zhu Guangying, Sun Tao, Xu Yuwen, et al. Identification of internal short-circuit faults in lithium-ion batteriesbased on a multi-machine learning fusion[ J] . Batteries,2023, 9(3) : 154.
[33] Li Jinglun, Mao Ziheng, Gu Xin, et al. A quantitativefault diagnosis method for lithium-ion battery based onMD-LSTM [ J ] . IEEE Transactions on Transportation Electrification, 2025, 11(1) : 2266-2276.
[34] Chen B R, Kunz M R, Tanim T R, et al. A machinelearning framework for early detection of lithium platingcombining multiple physics-based electrochemical signatures [ J ] . Cell Reports Physical Science, 2021, 2(3) : 100352.
[35] Hu Taiping, Huang Haichao, Zhou Guobing, et al. Observation of dendrite formation at Li metal-electrolyte interface by a machine-learning enhanced constant potentialframework [ J ] . Nature Communications, 2025, 16:7379.
[36] Kim S Y, Kwon S W, Nasir Bashir M, et al. Predictingdendrite growth in lithium metal batteries through iterativeneural networks and voltage embedding[ J] . NPJ Computational Materials, 2025, 11: 337.
[37] Cai Xue, Zhang Caiping, Chen Jue, et al. Sensorlessbattery expansion estimation using electromechanical coupled models and machine learning[ J] . Journal of EnergyChemistry, 2025, 105: 142-157.
[38] Peng Jichang, Zhao Binting, Meng Jinhao, et al. Boundary evolution constrained Seq2Seq framework for lithium-ion batteries strain state estimation[ J] . Journal of EnergyStorage, 2026, 146: 120197.
[39] Kumar D, Sharma M K. Early prediction of battery swelling via delta resistance features and optimized machinelearning models to avoid thermal runaway [ J] . FranklinOpen, 2025, 12: 100389.
[40] Hong Jichao, Zhang Huanqin, Xu Xiaoming. Thermalfault prognosis of lithium-ion batteries in real-world electric vehicles using self-attention mechanism networks[ J] .Applied Thermal Engineering, 2023, 226: 120304.
[41] Naguib M, Chen Junran, Kollmeyer P, et al. Thermalfault detection of lithium-ion battery packs through an integrated physics and deep neural network based model[ J] . Communications Engineering, 2025, 4: 79.
[42] Tian Luyu, Dong Chaoyu, Wang Rui, 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.
[43] Zhang Jiaqi, Gao Dexin, Cheng Yuanming, et al. A newmethod for thermal runaway warning of electric vehiclecharging [ J ] . Measurement Science and Technology,2024, 35: 126217.
[44] Ding Mutao, Guo Shiwei, Shan Zhilin, et al. Thermalrunaway multi-data fusion detection of lithium batterybased on Transformer neural network [ J] . Fire Scienceand Technology, 2024, 43(7) : 937-945. [丁沐涛, 郭世伟, 单志林, 等. 基于 Transformer 神经网络的锂电池热失 控 多 数 据 融 合 探 测 [ J] . 消 防 科 学 与 技 术,2024, 43(7) : 937-945. ]
[45] Song Yuhang, Lyu Nawei, Shi Shuang, et al. Safetywarning for lithium-ion batteries by module-space air-pressure variation under thermal runaway conditions[ J] .Journal of Energy Storage, 2022, 56: 105911.
[46] Lyu Nawei, Jin Yang, Miao Shan, et al. Fault warningand location in battery energy storage systems via ventingacoustic signal[J]. IEEE Journal of Emerging and Selected Topics in Power Electronics, 2023, 11(1): 100-108.
[47] Hou Jidong, Zhu Xiayu, Song Yuhang, et al. Faultwarning and localization for lithium-ion batteries by lasersignals [ J ] . Journal of Energy Storage, 2025, 133:118044.
[48] Guo Xingyu, Song Yuhang, Lyu Nawei, et al. A novelpassive wireless safety early warning technique based ontransient pressurization relief energy harvesting of batterysafety valves[ J] . Journal of Energy Storage, 2025, 129:117394.
[49] Zhan Yichen, Yan Kehuan, Zheng Xianghan. Integratingtransformers into physics-informed neural networks: anapproach to lithium-ion battery state-of-health prognostics[ J] . International Journal of Electrical Power & EnergySystems, 2025, 172: 111173.
[50] Wang Fujin, Zhai Zhi, Zhao Zhibin, et al. Physics-informed neural network for lithium-ion battery degradationstable modeling and prognosis [ J ] . Nature Communications, 2024, 15: 4332.
[51] Deng Weikun, Le Hung, Nguyen K T P, et al. A genericphysics-informed machine learning framework for batteryremaining useful life prediction using small early-stagelifecycle data[ J] . Applied Energy, 2025, 384: 125314.
[52] Yuan Zhuchen, Pan Yue, Wang Huaibin, et al. Faultdata generation of lithium ion batteries based on digitaltwin: a case for internal short circuit[ J] . Journal of Energy Storage, 2023, 64: 107113.
[53] Ji Chunhui, Jin Guang, Zhang Ran. State of charge estimation for lithium-ion batteries based on a digital twin hybrid model[ J] . Energy Reports, 2025, 13: 2174-2185.
[54] Reeta, Bala Anju, Singh M. Smart cloud-integrated battery management system for real-time state of charge estimation of lithium-ion battery pack in electric vehicles[ J] . Journal of Energy Storage, 2026, 145: 119894.
[55] Yi Yahui, Xia Chengyu, Feng Chao, et al. Digital twin-long short-term memory ( LSTM ) neural network basedreal-time temperature prediction and degradation modelanalysis for lithium-ion battery [ J ] . Journal of EnergyStorage, 2023, 64: 107203.
[56] Hu Jincheng, Fu Pengyu, Wei Zhongbao, et al. Earlyprediction of lithium-ion battery degradation with a generative pre-trained transformer [ J ] . Nature Communications, 2026, 17: 126.
[57] Tuncel S, Cinar H, Gucyetmez M, et al. Large languagemodel-based methodology for data-driven health predictionof Lithium-ion batteries[ J] . Energy Reports, 2025, 13:6425-6434.
[58] Zhang Wenjie, Feng Haohao, Ren Mifeng, et al. LLMMSIformer: a method for lithium-ion battery state ofhealth estimation based on a large language model andmulti-time scale interval transformer[ J] . Energy, 2026,342: 139417.

相似文献/References:

[1]秦东晨,赵鸿飞,武红霞,等.基于可重构电路的串联电池动态分组主动均衡方法[J].郑州大学学报(工学版),2025,46(2):90.[doi:10.13705/j.issn.1671-6833.2024.05.003]
　QIN Dongchen,ZHAO Hongfei,WU Hongxia,et al.Dynamic Grouping Active Equalization Method for Series Connected Batteries Based on Reconfigurable Circuits[J].Journal of Zhengzhou University (Engineering Science),2025,46(6):90.[doi:10.13705/j.issn.1671-6833.2024.05.003]
[2]秦东晨,罗庆洲,杨俊杰,等.基于PSO-MPC的锂离子电池快速安全充电策略[J].郑州大学学报(工学版),2025,46(5):90.[doi:10.13705/j.issn.1671-6833.2025.05.005]
　QIN Dongchen,LUO Qingzhou,YANG Junjie,et al.Fast and Safe Charging Strategy for Lithium-ion Batteries Based on PSO-MPC[J].Journal of Zhengzhou University (Engineering Science),2025,46(6):90.[doi:10.13705/j.issn.1671-6833.2025.05.005]

更新日期/Last Update: 2026-11-13