-
Accurate prediction of the state of health (SOH) of lithium-ion batteries in electric vehicles is crucial for ensuring the safety of drivers and passengers, optimizing battery management systems (BMS), and extending battery life. Reliable SOH prediction underpins essential BMS functions, including charge–discharge control, remaining useful life (RUL) prediction, and fault diagnosis. However, existing data-driven methods still face two long-standing challenges. First, most models rely excessively on a large number of handcrafted health features derived from voltage, current, and capacity data, resulting in feature redundancy and low computational efficiency. Second, SOH-related time series exhibit strong nonlinearity and non-stationarity, causing traditional neural networks to suffer from prediction drift, instability, and performance degradation under varying conditions. To address the first challenge, we propose a lightweight health feature selection mechanism that combines incremental capacity analysis (ICA) with correlation analysis to automatically identify compact and physically meaningful degradation features. Only four key health features that are highly correlated with capacity fading are selected, which effectively reduces model complexity and computational cost while maintaining high SOH prediction accuracy. To overcome the second challenge, we develop a hybrid neural network model (KanFormer) integrating the Kolmogorov–Arnold (KAN) representation theory with a Transformer-based temporal modeling framework for accurate and robust SOH prediction. Specifically, the proposed KanFormer framework consists of three hierarchical modules: (1) the local feature extraction module, which leverages the smooth interpolation capability of KAN to capture fine-grained degradation characteristics from voltage–capacity and incremental capacity (IC) curves, modeling local nonlinear behaviors in the degradation process; (2) the global feature extraction module, which employs a multi-head Transformer encoder to learn long-range dependencies and cross-scale temporal relationships, enabling the joint modeling of short-term dynamics and long-term aging evolution; and (3) the prediction output module, which uses nonlinear KAN layers to adaptively fuse local and global representations, producing numerically stable and highly accurate SOH prediction results. By combining the mathematical expressiveness of KAN with the temporal reasoning capability of the Transformer, KanFormer effectively mitigates prediction drift and oscillations induced by data nonlinearity and non-stationarity. Compared with conventional deep-learning models, the proposed method improves training efficiency by 15.32%. Experimental validation on three publicly available battery-aging datasets—Michigan Formation, HNEI, and NASA—demonstrates its superior performance, achieving MSE = 0.0045, MAE = 0.041, R2 = 0.978 on the Michigan dataset, MSE = 0.00055, MAE = 0.0175, R2 = 0.996 on the HNEI dataset, and MSE = 0.0056, MAE = 0.017, R2 = 0.984 on the NASA dataset. These results substantially outperform mainstream baselines, confirming the high accuracy and robustness of KanFormer. In summary, KanFormer unifies lightweight feature selection, nonlinear functional representation, and cross-scale temporal modeling, providing a scalable and interpretable solution for high-accuracy and high-efficiency SOH prediction.
-
Keywords:
- Lithium-ion battery /
- state of health /
- time series prediction /
- Kolmogorov–Arnold representation theory
-
[1] Gao M Y, Cai L H, Sun C C 2022 J. Electron. Inf. Technol. 44 3734 (in Chinese) [高明裕，蔡林辉，孙长城 2022 电子与信息学报 44 3734]
[2] Dong Z K, Ji X Y, Wang J Y, Gu Y T, Wang J F, Qi D L 2024 IEEE Trans. Consum. Electron. 70 4311
[3] Lucaferri V, Quercio M, Laudani A, Riganti Fulginei F 2023 Energies 16 7807
[4] Zhang Y L, Dong Z Q, Zheng L F, Wang K, Liang J F 2025 Acta Phys. Sin. 74 148803 (in Chinese) [张艳岚, 董泽庆, 郑利锋, 王凯, 梁君飞 2025 74 148803]
[5] Li X J, Yu Y T, Zhang Z W, Dong X R 2021 Acta Phys. Sin. 71 11401 (in Chinese) [李晓杰, 喻云泰, 张志文, 董小瑞. 2021 71 11401]
[6] Fahmy H M, Hasanien H M, Alharbi M, Ji H 2025 Sci. Rep. 15 14592
[7] Lai X, Yi W, Cui Y F, Qin C, Han X B, Sun T, Zhou L, Zheng Y J 2021 Energy. 216 119233
[8] Xiong R, Li L L, Li Z R, Yu Q Q, Mu H 2018 Appl. Energy 219 264
[9] Amir S, Gulzar M, Tarar M O, Naqvi I H, Zaffar N A, Pecht M G 2022 IEEE Access 10 18279
[10] Dong Z K, Gu S Y, Zhou S Q, Yang M J, Lai C S, Gao M Y 2025 IEEE Trans. Transp. Electr. 11 3655
[11] Dong Z K, Yang M J, Wang J F, Wang H, Lai C S, Ji X Y 2025 IEEE Trans. Transp. Electr. 11 2696
[12] Zhang M, Yang D F, Du J X, Sun H L, Li L W, Wang L C, Wang K 2023 Energies 16 3167
[13] Zhou D, Liang J L, Li F X, Cui Y X, Shan Y X, Zhang Y H, Chen M H, Li S 2024 J. Energy Storage 104 114579
[14] Huang J, He T, Zhu W L, Liao Y X, Zeng J H, Xu Q, Niu Y C 2025 Comput. Electr. Eng. 122 109930
[15] Zhang Y H, Li Z Y, Kong L Y, Xu H, Shen H W, Chen M 2025 J. Energy Storage 105 114538
[16] Qian C, Xu B H, Xia Q, Ren Y, Sun B, Wang Z L 2023 Appl. Energy 336 120793
[17] Zheng K, Meng J H, Yang Z P, Zhou F F, Yang K, Song Z X 2024 Appl. Energy 375 124077
[18] Li Y, Tao J 2020 Proceedings of the Chinese Control and Decision Conference (CCDC) Hefei, China, August 22–24, 2020 pp 5489–5494
[19] Zhang C L, Luo L J, Yang Z, Zhao S S, He Y G, Wang X, Wang H X 2023 Green Energy Intell. Transp. 2 100108
[20] Chen K Q, Pang H, Lin G Y, Wang F B, Nan W Z, Du J R 2024 Proceedings of the IEEE Transportation Electrification Conference and Expo, Asia-Pacific (ITEC Asia-Pacific) Chengdu, China, August 28–31, 2024 pp 423–428
[21] Ren L, Dong J B, Wang X K, Meng Z H, Zhao L, Deen M J 2021 IEEE Trans. Ind. Inf. 17 3478
[22] Gu X Y, See K W, Li P H, Shan K H, Wang Y P, Zhao L, Lim K C, Zhang N 2023 Energy 262 125501
[23] Li Z Y, Zhang X W, Gao W 2024 Energy 311 133418
[24] Liu W, Xu Y 2019 Proceedings of the IEEE Conference on Energy Internet and Energy System Integration (EI2) Beijing, China, November 8–10, 2019 pp 1203–1208
[25] Braco E, San Martin I, Sanchis P, Ursua A, Stroe D I 2022 J. Energy Storage 55 105366
[26] Ji X Y, Chen Y, Wang J F, Zhou G D, Lai C S, Dong Z K 2024 IEEE Internet Things J. 11 41941
[27] Chen X, Qin Y W, Zhao W D, Yang Q M, Cai N B, Wu K 2024 J. Energy Storage 86 111197
[28] Zhou W L, Lu Q, Zheng Y P 2022 Machines 10 512
[29] Krupp A, Ferg E, Schuldt F, Derendorf K, Agert C 2021 Batteries 7 2
[30] Weng A, Mohtat P, Attia P, Less G, Lee S, Stefanopoulou A 2021 Battery test data – fast formation study University of Michigan – Deep Blue Data https://doi.org/10.7302/pa3f-4w30
[31] Dubarry M, Beck D 2021 HNEI Long-Term Degradation Dataset Battery Archive https://batteryarchive.org
[32] Bodner A D, Tepsich A S, Spolski J N, Pourteau S 2024 arXiv preprint 2406.13155
Metrics
- Abstract views: 29
- PDF Downloads: 0
- Cited By: 0
下载:
