Battery Pack Multi-Fault Diagnosis Algorithm Based on Dual-Perspective Spectral Attention Fusion

LIU Mingjun; GU Shenyu; YIN Jingde; ZHANG Yifan; DONG Zhekang; JI Xiaoyue

doi:10.11999/JEIT251156

Article Contents

Article Navigation > Journal of Electronics & Information Technology > 2025 >

LIU Mingjun, GU Shenyu, YIN Jingde, ZHANG Yifan, DONG Zhekang, JI Xiaoyue. Battery Pack Multi-Fault Diagnosis Algorithm Based on Dual-Perspective Spectral Attention Fusion[J]. Journal of Electronics & Information Technology. doi: 10.11999/JEIT251156

Citation:

LIU Mingjun, GU Shenyu, YIN Jingde, ZHANG Yifan, DONG Zhekang, JI Xiaoyue. Battery Pack Multi-Fault Diagnosis Algorithm Based on Dual-Perspective Spectral Attention Fusion[J]. Journal of Electronics & Information Technology. doi: 10.11999/JEIT251156

Citation:

LIU Mingjun, GU Shenyu, YIN Jingde, ZHANG Yifan, DONG Zhekang, JI Xiaoyue. Battery Pack Multi-Fault Diagnosis Algorithm Based on Dual-Perspective Spectral Attention Fusion[J]. Journal of Electronics & Information Technology. doi: 10.11999/JEIT251156

PDF( 5657 KB)

Battery Pack Multi-Fault Diagnosis Algorithm Based on Dual-Perspective Spectral Attention Fusion

doi: 10.11999/JEIT251156 cstr: 32379.14.JEIT251156

LIU Mingjun^{1, 2},
GU Shenyu^{1, 2},
YIN Jingde^{1, 2},
ZHANG Yifan^{1, 2},
DONG Zhekang^{1, 2
,
,},
JI Xiaoyue³

1.
School of Electronic Information, Hangzhou Dianzi University, Hangzhou 310018, China
2.
Zhejiang Province Key Laboratory of Intelligent Vehicle Electronics, Hangzhou 310018, China
3.
Department of Precision Instrument, Tsinghua University, Beijing 100084, China

Funds: China Postdoctoral Science Foundation under Grants 2024T170463 and 2024M751676, National Natural Science Foundation of China under Grant 62206062, Zhejiang Provincial Natural Science Foundation of China under Grant No. LZYQ25F020005, and Ministry of Science and Technology-Yangtze River Delta Science and Technology Innovation Program under Grant 2023CSJGG1300

Accepted Date: 2025-12-22
Rev Recd Date: 2025-12-22

Available Online: 2026-01-03

Abstract

Abstract

Objective With the rapid growth of electric vehicles and their widespread deployment, battery pack fault have become more frequent, highlighting the urgent need for efficient fault diagnosis methods. Although deep-learning-based approaches have made notable progress, existing studies remain limited in addressing multiple fault types—such as internal short circuit (ISC), sensor noise, sensor drift, and state-of-charge (SOC) inconsistency—and in modeling the coupling relationships among them. To overcome these limitations, this paper proposes a multi-fault diagnosis algorithm for battery packs based on dual-perspective spectral attention. A dual-perspective tokenization module is designed to comprehensively extract spatiotemporal features from battery data, while a spectral attention mechanism effectively addresses non-stationary time-series characteristics and captures long-term dependencies, thereby enhancing diagnostic performance. Experimental results under the Federal Urban Driving Schedule (FUDS), Urban Dynamometer Driving Schedule (UDDS), and Supplemental Federal Test Procedure (US06) demonstrate that the proposed method achieves average improvements of 10.98% in precision, 12.64% in recall, 13.84% in F1-score, and 13.45% in accuracy compared with existing multi-fault diagnosis methods. Furthermore, systematic ablation studies and robustness analyses are designed and implemented to compare the contribution mechanisms of core modules to the overall model performance, while fully validating the anti-interference capability and robustness of the proposed method in complex noisy environments. Overall, the dual-perspective spectral attention framework not only improves multi-fault diagnosis performance but also offers a new perspective for modeling complex spatiotemporal features, providing a promising solution for enhancing vehicle safety. Methods To enhance spatiotemporal feature extraction and fault diagnosis performance, this paper proposes a dual-view spectral attention fusion algorithm for battery pack multi-fault diagnosis. The overall architecture consists of four core modules (Fig. 3): a dual-view tokenization module, a spectral attention module, a feature fusion module and an output module. The dual-view tokenization module employs positional encoding to jointly model temporal and spatial dimensions, enabling comprehensive spatiotemporal feature extraction. When combined with the spectral attention mechanism, the model’s ability to handle non-stationary characteristics is significantly improved, leading to enhanced diagnostic performance. In addition, to address the lack of comprehensive publicly available datasets for battery pack fault diagnosis, a new dataset is constructed, including internal short circuit, sensor noise, sensor drift, and state-of-charge imbalance faults. The dataset covers three operating conditions—FUDS, UDDS, and US06—effectively alleviating data scarcity in this research field. Results and Discussions Experimental results indicate that the proposed method improves the average precision, recall, F1 score and accuracy by 10.98%, 12.64%, 13.84% and 13.45%, respectively, compared to the existing optimal fault diagnosis methods. The comparison experiments under different operating conditions (Table 7), validate this conclusion. While traditional CNN algorithms perform well in local feature extraction, their fixed-size convolutional kernels struggle to adapt to time features of varying frequencies, leading to insufficient long-term temporal dependency and global feature capture. RNN-based methods exhibit lower computational efficiency when processing large-scale datasets. Transformer-based models face limitations in spatial feature extraction and capturing temporal variations. In contrast, the proposed algorithm overcomes these shortcomings through an integrated architectural design. Ablation experiments demonstrate the contribution of each module to the model's performance (Table 8), the complete framework improves the average F1 score and accuracy by 9.30% and 9.26%, respectively, compared to its ablation variants. Robustness analysis under simulated noise environments (Table 9), reveals that the proposed method achieves an accuracy improvement of 49.95% to 124.34% over baseline methods at noise levels ranging from -2 dB to -8 dB, demonstrating superior noise resistance. Conclusions This paper presents a multi-fault diagnosis algorithm for battery packs that fuses dual-view tokenization with spectral attention to integrate spatiotemporal and spectral information. The dual-view tokenization module performs tokenization and positional encoding along both temporal and spatial axes, substantially improving the model’s spatiotemporal representation. The spectral attention mechanism enhances modeling of non-stationary signals and long-term dependencies. Experiments under FUDS, UDDS, and US06 driving cycles show that the proposed method outperforms existing multi-fault diagnosis techniques, with average gains of 13.84% in F1 score and 13.45% in accuracy. Ablation studies confirm that both modules make substantial contributions and that their combination enables superior handling of complex time-series features. Under high-noise conditions (–2 dB, –4 dB, –6 dB, –8 dB), the method also demonstrates enhanced robustness, with accuracy improvements over baselines of 49.95%, 90.39%, 112.01%, and 124.34%, respectively. Despite these promising results, this study has three limitations. First, the data are primarily derived from laboratory simulations; thus, the model’s generalization to real-world operating conditions requires further verification. Second, the current work has not fully considered the impact of fault severity on BMS hierarchical decision-making, and future work will focus on constructing a fault-severity grading mechanism. Third, the model’s physical interpretability requires improvement; subsequent research will attempt to incorporate equivalent circuit models or electrochemical mechanism models to balance high accuracy with greater interpretability.
- Multi-fault diagnosis,
- Lithium-ion battery pack,
- Spectral Attention

FullText(HTML)

References(35)

References

[1]	SUN Zhenyu, WANG Zhenpo, LIU Peng, et al. Relative entropy based lithium-ion battery pack short circuit detection for electric vehicle[C]. Proceedings of the 2020 IEEE Energy Conversion Congress and Exposition (ECCE), Detroit, USA, 2020: 5061–5067. doi: 10.1109/ECCE44975.2020.9235755.
[2]	ZHEN Chanzwen, CHEN Ziqiang, and HUANZ D. A novel sensor fault diagnosis method for lithium-ion battery system using hybrid system modeling[C]. Proceedings of the Condition Monitoring and Diagnosis (CMD), Perth, Australia, 2018: 1–5. doi: 10.1109/CMD.2018.8535711.
[3]	YANG Ruixin, XIONG Rui, HE Hongwen, et al. A fractional-order model-based battery external short circuit fault diagnosis approach for all-climate electric vehicles application[J]. Journal of Cleaner Production, 2018, 187: 950–959. doi: 10.1016/j.jclepro.2018.03.259.
[4]	YANG Ruixin, XIONG Rui, and SHEN Weixiang. Experimental study on external short circuit and overcharge of lithium-ion battery packs for electric vehicles[C]. Proceedings of the 4th International Conference on Green Energy and Applications (ICGEA), Singapore, Singapore, 2020: 1–6. doi: 10.1109/ICGEA49367.2020.241506.
[5]	SCHMID M, KNEIDINGER H G, and ENDISCH C. Data-driven fault diagnosis in battery systems through cross-cell monitoring[J]. IEEE Sensors Journal, 2021, 21(2): 1829–1837. doi: 10.1109/JSEN.2020.3017812.
[6]	LIU Hanxiao, LI Liwei, DUAN Bin, et al. Multi-fault detection and diagnosis method for battery packs based on statistical analysis[J]. Energy, 2024, 293: 130465. doi: 10.1016/j.energy.2024.130465.
[7]	ZHU Xiaoqing, WANG H, WANG Xue, et al. Internal short circuit and failure mechanisms of lithium-ion pouch cells under mechanical indentation abuse conditions: An experimental study[J]. Journal of Power Sources, 2020, 455: 227939. doi: 10.1016/j.jpowsour.2020.227939.
[8]	ZHENG Yuejiu, LU Yifan, GAO Wenkai, et al. Micro-short-circuit cell fault identification method for lithium-ion battery packs based on mutual information[J]. IEEE Transactions on Industrial Electronics, 2021, 68(5): 4373–4381. doi: 10.1109/TIE.2020.2984441.
[9]	LI Fang, MIN Yongjun, ZHANG Yong, et al. Towards general and efficient fault diagnosis: A novel framework for multi-fault cross-domain diagnosis of lithium-ion batteries in real-world scenarios[J]. Energy, 2025, 334: 137825. doi: 10.1016/j.energy.2025.137825.
[10]	HU Xiaosong, ZHANG Kai, LIU Kailong, et al. Advanced fault diagnosis for lithium-ion battery systems: A review of fault mechanisms, fault features, and diagnosis procedures[J]. IEEE Industrial Electronics Magazine, 2020, 14(3): 65–91. doi: 10.1109/MIE.2020.2964814.
[11]	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.
[12]	张照娓, 郭天滋, 高明裕, 等. 电动汽车锂离子电池荷电状态估算方法研究综述[J]. 电子与信息学报, 2021, 43(7): 1803–1815. doi: 10.11999/JEIT200487. ZHANG Zhaowei, GUO Tianzi, GAO Mingyu, et al. Review of SoC estimation methods for electric vehicle Li-ion batteries[J]. Journal of Electronics & Information Technology, 2021, 43(7): 1803–1815. doi: 10.11999/JEIT200487.
[13]	ZHAO Yiwen, DENG Junjun, LIU Peng, et al. Enhancing battery durable operation: Multi-fault diagnosis and safety evaluation in series-connected lithium-ion battery systems[J]. Applied Energy, 2025, 377: 124632. doi: 10.1016/j.apenergy.2024.124632.
[14]	NAHA A, KHANDELWAL A, HARIHARAN K S, et al. On-board short-circuit detection of Li-ion batteries undergoing fixed charging profile as in smartphone applications[J]. IEEE Transactions on Industrial Electronics, 2019, 66(11): 8782–8791. doi: 10.1109/TIE.2018.2889623.
[15]	SEO M, GOH T, PARK M, et al. Detection method for soft internal short circuit in lithium-ion battery pack by extracting open circuit voltage of faulted cell[J]. Energies, 2018, 11(7): 1669. doi: 10.3390/en11071669.
[16]	XIONG Rui, SUN Wanzhou, YU Quanqing, et al. Research progress, challenges and prospects of fault diagnosis on battery system of electric vehicles[J]. Applied Energy, 2020, 279: 115855. doi: 10.1016/j.apenergy.2020.115855.
[17]	高明裕, 蔡林辉, 孙长城, 等. 一种基于斯皮尔曼秩相关结合神经网络的电池组内部短路故障检测算法[J]. 电子与信息学报, 2022, 44(11): 3734–3747. doi: 10.11999/JEIT210975. GAO Mingyu, CAI Linhui, SUN Changcheng, et al. An internal short circuit fault detecting of battery pack based on spearman rank correlation combined with neural network[J]. Journal of Electronics & Information Technology, 2022, 44(11): 3734–3747. doi: 10.11999/JEIT210975.
[18]	LAO Zhenpeng, HE Deqiang, JIN Zhenzhen, et al. Few-shot fault diagnosis of turnout switch machine based on semi-supervised weighted prototypical network[J]. Knowledge-Based Systems, 2023, 274: 110634. doi: 10.1016/j.knosys.2023.110634.
[19]	JI Xiaoyue, CHEN Yi, WANG Junfan, et al. Time-frequency hybrid neuromorphic computing architecture development for battery state-of-health estimation[J]. IEEE Internet of Things Journal, 2024, 11(24): 39941–39957. doi: 10.1109/JIOT.2024.3448350.
[20]	YU Quanqing, LI Jianming, CHEN Zeyu, et al. Multi-fault diagnosis of lithium-ion battery systems based on correlation coefficient and similarity approaches[J]. Frontiers in Energy Research, 2022, 10: 891637. doi: 10.3389/fenrg.2022.891637.
[21]	CAI Linhui, WANG Han, DONG Zhekang, et al. A multi-fault diagnostic method based on category-reinforced domain adaptation network for series-connected battery packs[J]. Journal of Energy Storage, 2023, 60: 106690. doi: 10.1016/j.est.2023.106690.
[22]	SHEN Xiaowei, LUN Shuxian, LI Ming. Multi-fault diagnosis of electric vehicle power battery based on double fault window location and fast classification[J]. Electronics, 2024, 13(3): 612. doi: 10.3390/electronics13030612.
[23]	ZHOU Juan, WU Zonghuan, ZHANG Shun, et al. A fault diagnosis method for power battery based on multiple model fusion[J]. Electronics, 2023, 12(12): 2724. doi: 10.3390/electronics12122724.
[24]	ZHAO Hongyu, ZHANG Chengzhong, XU Liang, et al. A deep neural network for multi-fault diagnosis of battery packs based on an incremental voltage measurement topology[J]. Energy, 2025, 316: 134590. doi: 10.1016/j.energy.2025.134590.
[25]	WANG Xin, MAO Dongxing, and LI Xiaodong. Bearing fault diagnosis based on vibro-acoustic data fusion and 1D-CNN network[J]. Measurement, 2021, 173: 108518. doi: 10.1016/j.measurement.2020.108518.
[26]	YANG Guanghua, LIU Yuexiao, LI Na, et al. Intelligent fault diagnosis method of capacitor voltage transformer based on recurrent neural network[C]. Proceedings of the 4th International Conference on Computer Engineering and Application (ICCEA), Hangzhou, China, 2023: 412–416. doi: 10.1109/ICCEA58433.2023.10135307.
[27]	LIU Xuyang, CAI Hongchang, ZHOU Zihan, et al. Enhancing multi-type fault diagnosis in lithium-ion battery systems: Vision transformer-based transfer learning approach[J]. Journal of Power Sources, 2024, 624: 235610. doi: 10.1016/j.jpowsour.2024.235610.
[28]	YUAN Haitao, LI Changlong, ZHOU Mingyang, et al. Multi-fault diagnosis for lithium-ion batteries under diverse operating conditions based on multi-source domain generalization[J]. Energy, 2025, 335: 138230. doi: 10.1016/j.energy.2025.138230.
[29]	MA Mina, LI Xiaoyu, GAO Wei, et al. Multi-fault diagnosis for series-connected lithium-ion battery pack with reconstruction-based contribution based on parallel PCA-KPCA[J]. Applied Energy, 2022, 324: 119678. doi: 10.1016/j.apenergy.2022.119678.
[30]	DONG Zhekang, GU Shenyu, ZHOU Shiqi, et al. Periodic segmentation transformer-based internal short circuit detection method for battery packs[J]. IEEE Transactions on Transportation Electrification, 2025, 11(1): 3655–3666. doi: 10.1109/TTE.2024.3444453.
[31]	WANG Jiayang, ZHANG Xinhao, HAI Yifeng, et al. MDGN: Circuit design of memristor-based denoising autoencoder and gated recurrent unit network for lithium-ion battery state of charge estimation[J]. IET Renewable Power Generation, 2024, 18(3): 372–383. doi: 10.1049/rpg2.12809.
[32]	KANG Yongzhe, DUAN Bin, ZHOU Zhongkai, et al. A multi-fault diagnostic method based on an interleaved voltage measurement topology for series connected battery packs[J]. Journal of Power Sources, 2019, 417: 132–144. doi: 10.1016/j.jpowsour.2019.01.058.
[33]	GRABOW J, KLINK J, ORAZOV N, et al. Triggering and characterisation of realistic internal short circuits in lithium-ion pouch cells—a new approach using precise needle penetration[J]. Batteries, 2023, 9(10): 496. doi: 10.3390/batteries9100496.
[34]	毛琳, 张海新, 何志伟, 等. 一种电池相平面结合Conformer-BiGRU网络的电池内短路故障诊断方法[J]. 电子与信息学报, 2025, 47(10): 4031–4043. doi: 10.11999/JEIT250313. MAO Lin, ZHANG Haixin, HE Zhiwei, et al. A battery internal-short-circuit fault diagnosis method combining battery phase plane with Conformer-BiGRU network[J]. Journal of Electronics & Information Technology, 2025, 47(10): 4031–4043. doi: 10.11999/JEIT250313.
[35]	DONG Zhekang, YANG Mengjie, WANG Junfan, et al. PFFN: A parallel feature fusion network for remaining useful life early prediction of lithium-ion battery[J]. IEEE Transactions on Transportation Electrification, 2025, 11(1): 2696–2706. doi: 10.1109/TTE.2024.3427334.