基于条件扩散噪声预测的锂离子电池车辆级故障检测
Condition-Guided Diffusion Noise Prediction for Vehicle-Level Fault Detection in Lithium-Ion Batteries
DOI: 10.12677/csa.2026.1611330, PDF,    科研立项经费支持
作者: 刘 红:安徽理工大学计算机科学与工程学院,安徽 淮南;合肥综合性国家科学中心能源研究院(安徽省能源实验室),安徽 合肥
关键词: 锂离子电池;故障检测;扩散模型;条件建模;噪声预测误差;Lithium-Ion Batteries; Fault Detection; Diffusion Model; Conditional Modeling; Noise Prediction Error
摘要: 针对真实车辆充电数据中故障样本稀缺,以及电压和温度响应随电流和荷电状态等运行条件变化,导致正常工况波动与故障响应偏离难以区分的问题,本文提出一种基于条件引导扩散模型(Condition-Guided Diffusion Model, CDM)的无监督车辆级故障检测方法。首先,将交叉注意力模块集成至U-Net的低分辨率层和瓶颈层,以充电过程中的电流和荷电状态(State of Charge, SOC)作为条件,引导模型学习给定运行条件下电压和温度的正常响应分布。其次,基于扩散模型的噪声预测误差,构建从片段级异常分数到车辆级异常分数的评估方法,从而在不使用故障样本训练的情况下识别潜在异常。实验结果表明,CDM在Dahu数据集上取得0.8744的AUROC和0.8546的AUPRC,整体优于GP、AE、Deep SVDD、GDN、CATCH、SARAD和DyAD等对比方法。在Dahu、Socea和Naobop三个数据集上,CDM的宏平均AUROC和宏平均AUPRC分别达到0.8806和0.8309,高于主要基准方法DyAD的0.8616和0.8089,表明所提出方法具有较好的车辆级故障综合判别能力。
Abstract: The scarcity of fault samples in real-world vehicle charging data, together with the dependence of voltage and temperature responses on operating conditions such as current and state of charge (SOC), makes it difficult to distinguish normal operating fluctuations from fault-induced deviations. To address this issue, this paper proposes an unsupervised vehicle-level fault detection method based on a Condition-Guided Diffusion Model (CDM). First, cross-attention modules are integrated into the low-resolution and bottleneck layers of a U-Net. Charging current and SOC are used as conditions to guide the model in learning the distribution of normal voltage and temperature responses under given operating conditions. Second, an evaluation method is developed based on the noise-prediction error of the diffusion model, aggregating segment-level anomaly scores into vehicle-level anomaly scores and thereby enabling potential anomalies to be identified without using fault samples for training. Experimental results show that CDM achieves an AUROC of 0.8744 and an AUPRC of 0.8546 on the Dahu dataset, outperforming the GP, AE, Deep SVDD, GDN, CATCH, SARAD, and DyAD baselines overall. Across the Dahu, Socea, and Naobop datasets, CDM achieves macro-averaged AUROC and AUPRC values of 0.8806 and 0.8309, respectively, exceeding the corresponding values of 0.8616 and 0.8089 obtained by the primary baseline, DyAD. These results demonstrate that the proposed method provides effective overall discrimination of vehicle-level faults.
文章引用:刘红. 基于条件扩散噪声预测的锂离子电池车辆级故障检测[J]. 计算机科学与应用, 2026, 16(11): 117-129. https://doi.org/10.12677/csa.2026.1611330

参考文献

[1] Shang, Y., Wang, S., Tang, N., Fu, Y. and Wang, K. (2024) Research Progress in Fault Detection of Battery Systems: A Review. Journal of Energy Storage, 98, Article 113079.
https://doi.org/10.1016/j.est.2024.113079
[2] Zhao, J., Liu, M., Zhang, B., Wang, X., Liu, D., Wang, J., et al. (2024) Review of Lithium-Ion Battery Fault Features, Diagnosis Methods, and Diagnosis Procedures. IEEE Internet of Things Journal, 11, 18936-18950.
https://doi.org/10.1109/jiot.2023.3324322
[3] Zhao, J., Feng, X., Wang, J., Lian, Y., Ouyang, M. and Burke, A.F. (2023) Battery Fault Diagnosis and Failure Prognosis for Electric Vehicles Using Spatio-Temporal Transformer Networks. Applied Energy, 352, Article 121949.
https://doi.org/10.1016/j.apenergy.2023.121949
[4] Kong, D., Lv, H., Ping, P. and Wang, G. (2023) A Review of Early Warning Methods of Thermal Runaway of Lithium Ion Batteries. Journal of Energy Storage, 64, Article 107073.
https://doi.org/10.1016/j.est.2023.107073
[5] Yu, Q., Wang, C., Li, J., Xiong, R. and Pecht, M. (2023) Challenges and Outlook for Lithium-Ion Battery Fault Diagnosis Methods from the Laboratory to Real World Applications. eTransportation, 17, Article 100254.
https://doi.org/10.1016/j.etran.2023.100254
[6] Zou, B., Zhang, L., Xue, X., Tan, R., Jiang, P., Ma, B., et al. (2023) A Review on the Fault and Defect Diagnosis of Lithium-Ion Battery for Electric Vehicles. Energies, 16, Article 5507.
https://doi.org/10.3390/en16145507
[7] Xue, Q., Li, G., Zhang, Y., Shen, S., Chen, Z. and Liu, Y. (2021) Fault Diagnosis and Abnormality Detection of Lithium-Ion Battery Packs Based on Statistical Distribution. Journal of Power Sources, 482, Article 228964.
https://doi.org/10.1016/j.jpowsour.2020.228964
[8] Li, S., Zhang, C., Du, J., Cong, X., Zhang, L., Jiang, Y., et al. (2022) Fault Diagnosis for Lithium-Ion Batteries in Electric Vehicles Based on Signal Decomposition and Two-Dimensional Feature Clustering. Green Energy and Intelligent Transportation, 1, Article 100009.
https://doi.org/10.1016/j.geits.2022.100009
[9] Yi, L., Cai, X., Wang, Y., Luo, B., Liu, J. and Liu, B. (2023) Soft Short-Circuit Fault Diagnosis of the Lithium-Ion Battery Pack Based on an Improved EKF Algorithm. Recent Patents on Mechanical Engineering, 16, 138-149.
https://doi.org/10.2174/2212797616666230330103334
[10] Zhang, Y., Jiang, M., Zhou, Y., Zhao, S. and Yuan, Y. (2023) Towards High-Safety Lithium-Ion Battery Diagnosis Methods. Batteries, 9, Article 63.
https://doi.org/10.3390/batteries9010063
[11] Schaeffer, J., Lenz, E., Gulla, D., Bazant, M.Z., Braatz, R.D. and Findeisen, R. (2024) Gaussian Process-Based Online Health Monitoring and Fault Analysis of Lithium-Ion Battery Systems from Field Data. Cell Reports Physical Science, 5, Article 102258.
https://doi.org/10.1016/j.xcrp.2024.102258
[12] Spina, D.E., de O. Campos, L.F., de Arruda, W.F., Melo, A., Alves, M.F.D.S., Rabello, G.L., et al. (2024) Comparison of Autoencoder Architectures for Fault Detection in Industrial Processes. Digital Chemical Engineering, 12, Article 100162.
https://doi.org/10.1016/j.dche.2024.100162
[13] Zhu, J., Jiang, M. and Liu, Z. (2022) Fault Detection and Diagnosis in Industrial Processes with Variational Autoencoder: A Comprehensive Study. Sensors, 22, Article 227.
https://doi.org/10.3390/s22010227
[14] Kou, L., Chen, J., Qin, Y. and Mao, W. (2022) The Robust Multi-Scale Deep-SVDD Model for Anomaly Online Detection of Rolling Bearings. Sensors, 22, Article 5681.
https://doi.org/10.3390/s22155681
[15] Deng, A. and Hooi, B. (2021) Graph Neural Network-Based Anomaly Detection in Multivariate Time Series. Proceedings of the AAAI Conference on Artificial Intelligence, 35, 4027-4035.
https://doi.org/10.1609/aaai.v35i5.16523
[16] Yao, L., Zheng, J., Xiao, Y., Zhang, C., Zhang, L., Gong, X., et al. (2023) An Intelligent Fault Diagnosis Method for Lithium-Ion Battery Pack Based on Empirical Mode Decomposition and Convolutional Neural Network. Journal of Energy Storage, 72, Article 108181.
https://doi.org/10.1016/j.est.2023.108181
[17] Sun, J., Ren, S., Shang, Y., Zhang, X., Liu, Y. and Wang, D. (2023) A Novel Fault Prediction Method Based on Convolutional Neural Network and Long Short-Term Memory with Correlation Coefficient for Lithium-Ion Battery. Journal of Energy Storage, 62, Article 106811.
https://doi.org/10.1016/j.est.2023.106811
[18] Liu, X., Cai, H., Zhou, Z., Kong, Y., Zhou, X., Han, X., et al. (2024) Enhancing Multi-Type Fault Diagnosis in Lithium-Ion Battery Systems: Vision Transformer-Based Transfer Learning Approach. Journal of Power Sources, 624, Article 235610.
https://doi.org/10.1016/j.jpowsour.2024.235610
[19] Zhang, J., Wang, Y., Jiang, B., He, H., Huang, S., Wang, C., et al. (2023) Realistic Fault Detection of Li-Ion Battery via Dynamical Deep Learning. Nature Communications, 14, Article No. 5940.
https://doi.org/10.1038/s41467-023-41226-5
[20] Cao, R., Zhang, Z., Shi, R., Lu, J., Zheng, Y., Sun, Y., et al. (2025) Model-Constrained Deep Learning for Online Fault Diagnosis in Li-Ion Batteries over Stochastic Conditions. Nature Communications, 16, Article No. 1651.
https://doi.org/10.1038/s41467-025-56832-8
[21] Ho, J., Jain, A. and Abbeel, P. (2020) Denoising Diffusion Probabilistic Models. 2020 Advances in Neural Information Processing Systems, Online, 6-12 December 2020, 6840-6851.
https://proceedings.neurips.cc/paper/2020/hash/4c5bcfec8584af0d967f1ab10179ca4b-Abstract.html
[22] Wyatt, J., Leach, A., Schmon, S.M. and Willcocks, C.G. (2022) AnoDDPM: Anomaly Detection with Denoising Diffusion Probabilistic Models Using Simplex Noise. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), New Orleans, 19-20 June 2022, 649-655.
https://doi.org/10.1109/cvprw56347.2022.00080
[23] Zhang, H., Wang, Z., Zeng, D., Wu, Z. and Jiang, Y.G. (2025) DiffusionAD: Norm-Guided One-Step Denoising Diffusion for Anomaly Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 47, 7140-7152.
https://doi.org/10.1109/tpami.2025.3570494
[24] Xiao, C., Gou, Z., Tai, W., Zhang, K. and Zhou, F. (2023) Imputation-Based Time-Series Anomaly Detection with Conditional Weight-Incremental Diffusion Models. Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Long Beach, 6-10 August 2023, 2742-2751.
https://doi.org/10.1145/3580305.3599391
[25] Chen, Y., Zhang, C., Ma, M., Liu, Y., Ding, R., Li, B., et al. (2023) ImDiffusion: Imputed Diffusion Models for Multivariate Time Series Anomaly Detection. Proceedings of the VLDB Endowment, 17, 359-372.
https://doi.org/10.14778/3632093.3632101
[26] Zhong, G., Wang, P., Yuan, J., Li, Z. and Chen, L. (2025) Multi-Resolution Decomposable Diffusion Model for Non-Stationary Time Series Anomaly Detection. International Conference on Learning Representations, Singapore, 24-28 April 2025, 33527-33553.
https://proceedings.iclr.cc/paper_files/paper/2025/hash/9f7f063144103bf6debb09a3f15e00fb-Abstract-Conference.html
[27] Wu, X., Qiu, X., Li, Z., Wang, Y., Hu, J., Guo, C., et al. (2025) CATCH: Channel-Aware Multivariate Time Series Anomaly Detection via Frequency Patching. International Conference on Learning Representations, Vancouver, 24-28 April 2025, 17017-17045.
https://proceedings.iclr.cc/paper_files/paper/2025/hash/2b25c39788e5cf11d3541de433ebf4c0-Abstract-Conference.html
[28] Dai, Z., He, L., Yang, S.H. and Leeke, M. (2024) SARAD: Spatial Association-Aware Anomaly Detection and Diagnosis for Multivariate Time Series. Advances in Neural Information Processing Systems 37, Vancouver, 10-15 December 2024, 48371-48410.
https://doi.org/10.52202/079017-1533