1.沈阳建筑大学机械工程学院,辽宁 沈阳 110168
2.沈阳建筑大学电气与控制工程学院,辽宁 沈阳 110168
王馨璐(1998—),女,博士研究生,研究方向为锂离子电池建模,E-mail:wangixnlu012@163.com;
高治军,教授,博士生导师,研究方向为建筑能源系统管控,E-mail:gzj@sjzu.edu.cn。
收稿:2026-04-21,
修回:2026-05-28,
网络首发:2026-05-28,
移动端阅览
王馨璐, 高治军, 郭喜峰, 等. 基于健康特征增强与TCN-ASSA的锂离子电池RUL预测[J]. 储能科学与技术, XXXX, XX(XX): 1-13.
Wang Xinlu, Gao Zhijun, Guo Xifeng, et al. LITHIUM-ION BATTERY RUL PREDICTION BASED ON HEALTH FEATURE ENHANCEMENT AND TCN-ASSA[J]. Energy Storage Science and Technology, XXXX, XX(XX): 1-13.
王馨璐, 高治军, 郭喜峰, 等. 基于健康特征增强与TCN-ASSA的锂离子电池RUL预测[J]. 储能科学与技术, XXXX, XX(XX): 1-13. DOI: 10.19799/j.cnki.2095-4239.2026.0345.
Wang Xinlu, Gao Zhijun, Guo Xifeng, et al. LITHIUM-ION BATTERY RUL PREDICTION BASED ON HEALTH FEATURE ENHANCEMENT AND TCN-ASSA[J]. Energy Storage Science and Technology, XXXX, XX(XX): 1-13. DOI: 10.19799/j.cnki.2095-4239.2026.0345.
针对锂离子电池在循环老化过程中存在容量再生、局部波动及噪声干扰等问题,导致剩余使用寿命(Remaining Useful Life,RUL)预测精度下降,本文提出一种基于健康特征增强与时间卷积网络(Temporal Convolutional Network,TCN)-自适应稀疏自注意力(Adaptive Sparse Self-Attention,ASSA)的锂离子电池RUL预测方法。首先,基于充电阶段增量容量曲线,采用萨维茨基-戈莱滤波对原始曲线进行平滑增强,并提取峰值、峰值对应电压及特定电压区间包络面积等特征;同时结合等间隔充电时间和循环次数,构建多维健康特征集合,以提升对电池退化状态的表征能力。其次,构建融合TCN-ASSA的预测模型,利用TCN捕捉容量退化过程中的长期时序依赖关系,借助ASSA突出关键退化信息、抑制无关扰动和异常波动影响,从而提高RUL预测精度与模型鲁棒性。最后,基于Oxford和NASA锂离子电池数据集进行实验验证。在Oxford数据集上,TCN-ASSA的均方根误差(Root Mean Square Error,RMSE)为2.2726,平均绝对百分比误差(Mean Absolute Percentage Error,MAPE)为0.3634。NASA数据集的RMSE低至0.00734,MAPE低至0.00486。所提方法在消融实验、对比实验和跨数据集泛化实验中均取得了较优表现。表明该方法能够更准确地跟踪电池容量衰退趋势,具有较好的预测精度、鲁棒性和跨数据集适应性,可为锂离子电池健康管理与安全预警提供有效支持。
To address the decline in prediction accuracy of the remaining useful life (RUL) of lithium-ion batteries caused by capacity regeneration
local fluctuations
and noise interference during cyclic aging
this paper proposes a lithium-ion battery RUL prediction method based on health feature enhancement and Temporal Convolutional Network (TCN)-Adaptive Sparse Self-Attention (ASSA) framework. First
based on the incremental capacity curve during the charging process
Savitzky-Golay filtering is employed to smooth and enhance the original curve
and features such as peak value
the corresponding peak voltage
and the envelope area within a specified voltage interval are extracted. Meanwhile
equal-interval charging time and cycle number are incorporated to construct a multidimensional health feature set
thereby improving the representation of battery degradation states. Subsequently
a predictive model integrating TCN and ASSA is developed
in which TCN is used to capture long-term temporal dependencies in the capacity degradation process
while ASSA is introduced to emphasize critical degradation information and suppress irrelevant disturbances as well as abnormal fluctuations
thereby enhancing both RUL prediction accuracy and robustness. Finally
experiments are conducted on the Oxford and NASA lithium-ion battery datasets. On the Oxford dataset
TCN-ASSA achieved root mean square error (RMSE) of 2.2726 and mean absolute percentage error (MAPE) of 0.3634. On the NASA dataset
RMSE was as low as 0.00734 and MAPE as low as 0.00486. The results demonstrate that the proposed method achieves superior performance in ablation studies
comparative experiments
and cross-dataset generalization tests
enabling more accurate tracking of battery capacity degradation trends. The method exhibits high prediction accuracy
robustness
and adaptability
and can provide effective support for lithium-ion battery health management and safety warning.
李炳金, 韩晓霞, 张文杰, 等. 锂离子电池剩余使用寿命预测方法综述[J]. 储能科学与技术, 2024, 13(04): 1266-1276.
Li B J, Han X X, Zhang W J, et al. Review of the remaining useful life prediction methods for lithium-ion batteries[J]. Energy storage science and technology, 2024, 13(04): 1266-1276.
Song K, Hu D, Tong Y, et al. Remaining life prediction of lithium-ion batteries based on health management: A review[J]. Journal of Energy Storage, 2023, 57: 106193.
周驰东, 李嘉宇, 余佩雯, 等. 机器学习在锂离子电池故障诊断中的应用综述[J]. 电池, 2025, 55(05): 1159-1165.
Zhou C D, Li J Y, Yu P W, et al. Review on application of machine learning in Li-ion battery fault diagnosis[J]. Dianchi (Battery Bimonthly), 2025, 55(05): 1159-1165.
Luo K, Chen X, Zheng H, et al. A review of deep learning approach to predicting the state of health and state of charge of lithium-ion batteries[J]. Journal of Energy Chemistry, 2022, 74: 159-173.
陈勇, 王俊磊, 王鹏, 等. 锂离子电池早期剩余寿命预测方法综述[J]. 电池, 2026, 56(01): 222-230.
Cheng Y, Wang J L, Wang P, et al. A review of early remaining life prediction methods for Li-ion battery[J]. Dianchi (Battery Bimonthly), 2026, 56(01): 222-230.
蔡艳平, 王新军, 姜柯, 等. 基于GRNN-GSA-ELM的在线锂离子电池RUL多步预测[J]. 机械工程学报, 2024,60(24): 296-308.
Cai Y P, Wang X J, J K, et al. Multi-step prediction of online lithium battery remaining useful life based on GRNN-GSA-ELM[J]. Journal of mechanical engineering, 2024,60(24): 296-308.
寇发荣, 杨天祥, 罗希, 等. 基于特征重构与多时间尺度的锂电池SOH和RUL联合估计[J]. 太阳能学报, 2025,46(06): 68-78.
Kou F R, Yang T X, Luo X, et al. Joint estimation of SOH and RUL for lithium batteries based on feature reconstruction and multiple times scales[J]. Acta energiae solaris sinica, 2025,46(06): 68-78.
Zhou Y, Wang S, Xie Y, et al. Remaining useful life prediction and state of health diagnosis for lithium-ion batteries based on improved grey wolf optimization algorithm-deep extreme learning machine algorithm[J]. Energy, 2023, 285: 128761.
卢婷, 杨文强. 锂离子电池全生命周期内评估参数及评估方法综述[J]. 储能科学与技术, 2020, 9(03): 657-669.
Lu T, Yang W Q. Review of evaluation parameters and methods of lithium batteries throughout its life cycle[J]. Energy storage science and technology, 2020, 9(03): 657-669.
陈金荣, 孙跃东, 邵裕新, 等. 基于等效电路模型的云端动力电池寿命估计[J]. 重庆理工大学学报(自然科学), 2024, 38(06): 11-20.
Chen J R, Sun Y D, Shao Y X, et al. Battery life estimation of cloud-based based on equivalent circuit modeling[J]. Journal of Chongqing University of Technology (Natural Science), 2024, 38(06): 11-20.
徐鹏飞, 张睿, 丁飞. 基于电化学-热耦合模型的锂离子电池日历老化研究[J]. 电源技术, 2025, 49(08): 1638-1649.
Xu P F, Zhang R Ding F. Calendar aging study of lithium-ion batteries based on electrochemical-thermal coupling model[J]. Power supply technology, 2025, 49(08): 1638-1649.
Du Z, Zuo L, Li J, et al. Data-driven estimation of remaining useful lifetime and state of charge for lithium-ion battery[J]. IEEE Transactions on Transportation Electrification, 2021, 8(1): 356-367.
Zhang Y, Li Y. Prognostics and health management of Lithium-ion battery using deep learning methods: A review[J]. Renewable and sustainable energy reviews, 2022, 161: 112282.
王小明, 何叶, 王路路, 等. 基于DWD-SVR模型的锂离子电池剩余使用寿命预测[J]. 太阳能学报, 2025, 46(02): 52-59.
Wang X M, He Y, Wang L L, et al. RUL prediction for lithium-ion batteries based on DWD-SVR model[J]. Acta energiae solaris sinica, 2025,46(02): 52-59.
王伟亮, 刘会巧, 张天宇, 等. 基于多任务集成学习的储能电池剩余使用寿命预测[J]. 电力建设, 2024, 45(11): 25-33.
Wang W L, Liu H Q, Zhang T Y, et al. Multi-task ensemble learning-based prediction of remaining useful life of energy-storage batteries[J]. Electric Power Construction, 2024, 45(11): 25-33.
Li X, Yu D, Byg V, et al. The development of machine learning-based remaining useful life prediction for lithium-ion batteries[J]. Journal of Energy Chemistry, 2023, 82: 103-121.
余萍, 王浩年, 曹洁. 基于多尺度特征融合的并行神经网络剩余寿命预测方法[J]. 太阳能学报, 2026, 47(01): 785-796.
Yu P, Wang H N, Cao H, et al. Remaining useful life prediction method based on parallel neural networks with multi-scale feature fusion [J]. Acta energiae solaris sinica, 2026, 47(01): 785-796.
勇晔, 吴参林, 薛溟枫, 等. 基于多通道特征融合优化VMD-SCNN-LSTM的锂电池RUL预测[J]. 太阳能学报, 2025,46(12): 41-51.
Yong Y, Wu C L, Xue M F, et al. Multi-channel feature fusion-based VMD-SCNN-LSTM for lithiun-ion battery RUL prediction[J]. Acta energiae solaris sinica, 2025,46(12): 41-51.
Gomez W, Wang F, Chou J. Li-ion battery capacity prediction using improved temporal fusion transformer model[J]. Energy, 2024, 296: 131114.
Wang F, Zhai Z, Zhao Z, et al. Physics-informed neural network for lithium-ion battery degradation stable modeling and prognosis[J]. Nature Communications, 2024, 15(1): 4332.
武明虎, 岳程鹏, 张凡, 等. 多尺度分解下GRU-MLR组合的锂电池剩余使用寿命预测方法[J]. 储能科学与技术, 2023,12(07): 2220-2228.
Wu M H, Yue C P, Zhang F, et al. Combined GRU-MLR method for predicting the remaining useful life of lithium batteries via multiscale decomposition[J]. Energy storage science and technology, 2023,12(07): 2220-2228.
Howey, D., and C. Birkl. Oxford Battery Degradation Dataset 1. University of Oxford, 2017. https://ora.ox.ac.uk/objects/uuid:03ba4b01-cfed-46d3-9b1a-7d4a7bdf6fac
Wang X, Gao Z, Guo X, et al. Prediction of the Remaining Useful Life of Lithium-Ion Batteries Based on Health Features Extraction and Improved Stochastic Configuration Networks[J]. Batteries, 2026, 12(4): 114.
0
浏览量
15
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010102001997号