Multi-Task Deep Learning Framework for Battery Health Prognostics with Adaptive Transfer Learning 

Huang Linrun  Fang Weidong  Hu Jiacheng  ZhengYihan

Fujian Provincial Key Laboratory of Automotive Electronics and Electric Drive, Fujian University of Technology, Fuzhou, Fujian China，350118；

Abstract：Accurate state of health (SOH) estimation and remaining useful life (RUL) prediction are critical for lithium-ion battery management. Existing methods, however, often struggle with long-term prediction stability, domain shift, and joint optimization of SOH and RUL tasks. To address these issues, this paper proposes a multi-task attention-enhanced framework integrating fractional Brownian motion (FBM)-based degradation features and adaptive transfer learning for joint SOH and RUL prediction. Experiments on the CALCE dataset demonstrate that the proposed method achieves SOH MAE of 0.016±0.001 and RUL MAE of 41.0±3.1 cycles, with clear improvements over baseline methods. In addition, the framework improves long-horizon prediction reliability and cross-domain robustness. These results indicate its potential for practical battery health prognostics.

Keywords: Lithium-ion battery; State of health; Remaining useful life; Attention mechanism; Transfer learning; Fractional Brownian motion

References

[1]J. Lu, R. Xiong, J. Tian, C. Wang, and F. Sun, "Deep learning to estimate lithium-ion battery state of health without additional degradation experiments," Nature Communications, vol. 14, no. 1, Art. no. 2760, 2023,doi: 10.1038/S41467-023-38458-W.

[2]F.-M. Zhao et al., "Application of SOH estimation and RUL prediction of lithium-ion batteries based on AT-CNN-BiLSTM," Scientific Reports, vol. 14, no. 1, Art. no. 29026, 2024, doi:10.1038/s41598-024-80421-2.

[3]Y. Wang et al., "Enhanced vision-transformer integrating with semi-supervised transfer learning for state of health and remaining useful life estimation of lithium-ion batteries, " Energy & AI, vol.17, Art. no. 100405, 2024, doi: 10.1016/j.egyai.2024.100405.

[4]M. Lin, J. Wu, J. Meng, W. Wang, and J. Wu, "State of health estimation with attentional long short-term memory network for lithium-ion batteries," Energy, vol. 268, Art. no. 126706, 2023, doi: 10.1016/j.energy.2023.126706.

[5]J. Meng, D. Hu, M. Lin, J. Peng, J. Wu, and D.-I. Stroe, "A domain-adversarial neural network for transferable lithium-ion battery state of health estimation," IEEE Transactions on Transportation Electrification, vol. 11, no. 3, pp. 7732–7742, 2025, doi: 10.1109/TTE.2025.3530536.

[6]S. Sahoo et al., "Transfer learning based generalized framework for state of health estimation of Li-ion cells," Scientific Reports, vol. 12 , Art. no. 13173, 2022, doi: 10.1038/s41598-022-16692-4.

[7]Z. Ye and J. Yu, "State-of-health estimation for lithium-ion batteries using domain adversarial transfer learning," IEEE Transactions on Power Electronics, vol. 37, no. 3, pp. 3528–3543, Mar. 2022, doi: 10.1109/TPEL.2021.3117788.

[8]Y. Li et al., "SOH evaluation and RUL estimation of lithium-ion batteries using MC-CNN-TimesNet," Reliability Engineering & System Safety, vol. 261, Art. no. 111125, 2025, doi: 10.1016/j.ress.2025.111125

[9]P. Yang, H. Yang, and X. Meng, "Joint evaluation and prediction of SOH and RUL for lithium batteries based on a GBLS booster multi-task model," Journal of Energy Storage, vol. 75, Art. no. 109741, 2024, doi: 10.1016/j.est.2023.109741.

[10]N. Cai, X. Que, X. Zhang, W. Feng, and Y. Zhou, "A deep learning framework for the joint prediction of the SOH and RUL of lithium-ion batteries based on bimodal images," Energy, vol. 302, Art. no. 131700, 2024, doi: 10.1016/j.energy.2024.131700.

[11]X. Chen et al., "A self-attention knowledge domain adaptation network for battery state-of-health estimation under shallow-cycle conditions," Journal of Energy Storage, vol.86, Art. no. 111197, 2024, doi: 10.1016/j.est.2024.111197

[12]K. Huang et al., " State of health estimation of lithium-ion batteries based on fine-tuning or rebuilding transfer learning strategies combined with new features mining," Energy, vol.282, Art. no.128739, 2023, doi: 10.1016/j.energy.2023.128739.

[13]F. Li et al., " State-of-health estimation for lithium-ion batteries using unsupervised deep subdomain adaptation," Energy, vol. 324, Art.no.135862, 2025, doi: 10.1016/j.energy.2025.135862.

[14]Z. E. Amogne, H. Bakai, and M. Frasca, "Transfer learning based on transferability measures for RUL and SOH prediction of lithium-ion batteries," Batteries, vol. 9, no. 5, Art. no. 280, 2023, doi: 10.3390/batteries9050280.

[15]F. Wang et al., "Physics-informed neural network for lithium-ion battery degradation stable modeling and prognosis," Nature Communications, vol. 15, no. 1, Art. no. 4332, 2024, doi: 10.1038/s41467-024-48779-z.

[16]Y. Guo et al., "State of health estimation for lithium-ion battery based on Bi-directional long short-term memory neural network and attention mechanism," Energy Reports, vol. 8, 2022, doi: 10.1016/j.egyr.2022.10.128.

[17]G. Xu et al., "LSTM-based estimation of state of health for lithium-ion batteries by incorporating spatio-temporal attention under different health stages," PLoS ONE, vol. 19, no. 12, Art. no.
e0312856, 2024, doi: 10.1371/journal.pone.0312856.

[18]X. Zhang et al., "A novel state-of-health prediction method based on long short-term memory network with attention mechanism for lithium-ion battery," Frontiers in Energy Research, vol. 10, Art. no. 972486, 2022, doi: 10.3389/fenrg.2022.972486.

[19]C. Zou ,X. Chen, and Y Zhang, "State of health prediction of lithium-ion batteries based on temporal degeneration feature extraction with deep cycle attention network," Journal of Energy Storage, vol. 65, Art. no. 107367, 2023, doi: 10.1016/j.est.2023.107367.

[20]T. He and Z. Gong, "State of health estimation for lithium-ion batteries using a hybrid neural network with multi-scale convolutional attention mechanism," Journal of Power Sources, vol. 609, Art. no. 234680, 2024, doi: 10.1016/j.jpowsour.2023.234680.

[21]T. Bai and H. Wang, "Convolutional transformer-based multiview information perception framework for lithium-ion battery SOH estimation," IEEE Transactions on Instrumentation and Measurement, vol. 72, Art. no. 2523312, 2023, doi: 10.1109/TIM.2023.3300451.

[22]L. Ren, J. Dong, X. Wang, Z. Meng, L. Zhao, and M. J. Deen, "A data-driven auto-CNN-LSTM prediction model for lithium-ion battery remaining useful life," IEEE Transactions on Industrial Informatics, vol. 17, no. 5, pp. 3478–3487, 2021, doi: 10.1109/TII.2020.3008223.

[23]X. Shu et al., "A flexible state-of-health prediction scheme for lithium-ion battery packs with LSTM and transfer learning," IEEE Transactions on Transportation Electrification, vol. 7, no. 4, pp. 2238–2248, 2021, doi: 10.1109/TTE.2021.3074638.

[24]T. Lin et al., "Investigating explainable transfer learning for battery lifetime prediction under state transitions," eScience,, vol. 4, , Art. no. 100280, 2024, doi: 10.1016/j.esci.2024.100280.

[25]A. Thelen et al., " Probabilistic machine learning for battery health diagnostics and prognostics—review and perspectives," npj Materials Sustainability, vol. 2, Art. no. 14, 2024, doi: 10.1038/s44296-024-00011-1.

[26]G. Hong et al., "An iterative model of the generalized Cauchy process for predicting the remaining useful life of lithium-ion batteries, " Measurement, vol. 187, Art. no. 110269, 2022, doi: 10.1016/j.measurement.2021.110269.