Accurate state-of-health (SOH) prediction of lithium batteries is vital for guaranteeing the safe operation of battery systems and extending their service life. However, previous data-driven models were usually established based on ideal laboratory environments, which would severely restrict their generalization ability in real-world applications. This paper develops an SOH prediction method based on feature decoupling and a heterogeneous stacking ensemble learning framework (HSELF). First, the validation dataset was constructed by using accelerated cycle aging data with three levels of gradient noise. Second, features based on incremental capacity analysis (ICA) were extracted from a selected charging voltage interval to capture underlying battery degradation characteristics. Then, a feature decoupling strategy based on noise robustness decay rate (NRDR) and the corresponding differentiated feature input spaces were constructed for the heterogeneous stacking model. Additionally, a nested inner leave-one-battery-out cross-validation (LOBOCV) strategy was developed to achieve an unbiased evaluation of generalization under small-sample conditions. Finally, the ridge regression method was adopted as the meta-learner to integrate the outputs of the base models in the HSELF, leveraging their complementary strengths for mitigating prediction errors. Experimental results based on the LISHEN dataset successfully demonstrated the effectiveness of the developed SOH prediction method based on feature decoupling and HSELF. Under noise-free and typical industrial noise ($ \sigma \le 10\text{mV} $) conditions, the accuracy of the developed SOH prediction method would be significantly superior to that of individual base models. Furthermore, although the extreme noise ($ \sigma = 20\text{mV} $) causes severe performance degradation with individual base models, the developed SOH prediction method can consistently maintain high prediction stability ($ {R}^{2} > 0.93 $). In addition, experimental results based on an additional HNEI dataset also demonstrate that the developed SOH prediction method can achieve an effective balance among prediction accuracy, noise robustness, and cross-dataset robustness.
Citation: Yongxun Liu, Penglong Sun, Ke Zhao, Feng Zhao, Bin Wang. State-of-health prediction for lithium batteries based on feature decoupling and heterogeneous stacking ensemble learning framework[J]. AIMS Energy, 2026, 14(4): 923-954. doi: 10.3934/energy.2026037
Accurate state-of-health (SOH) prediction of lithium batteries is vital for guaranteeing the safe operation of battery systems and extending their service life. However, previous data-driven models were usually established based on ideal laboratory environments, which would severely restrict their generalization ability in real-world applications. This paper develops an SOH prediction method based on feature decoupling and a heterogeneous stacking ensemble learning framework (HSELF). First, the validation dataset was constructed by using accelerated cycle aging data with three levels of gradient noise. Second, features based on incremental capacity analysis (ICA) were extracted from a selected charging voltage interval to capture underlying battery degradation characteristics. Then, a feature decoupling strategy based on noise robustness decay rate (NRDR) and the corresponding differentiated feature input spaces were constructed for the heterogeneous stacking model. Additionally, a nested inner leave-one-battery-out cross-validation (LOBOCV) strategy was developed to achieve an unbiased evaluation of generalization under small-sample conditions. Finally, the ridge regression method was adopted as the meta-learner to integrate the outputs of the base models in the HSELF, leveraging their complementary strengths for mitigating prediction errors. Experimental results based on the LISHEN dataset successfully demonstrated the effectiveness of the developed SOH prediction method based on feature decoupling and HSELF. Under noise-free and typical industrial noise ($ \sigma \le 10\text{mV} $) conditions, the accuracy of the developed SOH prediction method would be significantly superior to that of individual base models. Furthermore, although the extreme noise ($ \sigma = 20\text{mV} $) causes severe performance degradation with individual base models, the developed SOH prediction method can consistently maintain high prediction stability ($ {R}^{2} > 0.93 $). In addition, experimental results based on an additional HNEI dataset also demonstrate that the developed SOH prediction method can achieve an effective balance among prediction accuracy, noise robustness, and cross-dataset robustness.
| [1] |
Li P, Wu X, Grosu R, et al. (2025) Applying neural network to health estimation and lifetime prediction of lithium-ion batteries. IEEE Trans Transp Electrif 11: 4224–4248. https://doi.org/10.1109/TTE.2024.3457621 doi: 10.1109/TTE.2024.3457621
|
| [2] |
Goh HH, Lan Z, Zhang D, et al. (2022) Estimation of the state of health (SOH) of batteries using discrete curvature feature extraction. J Energy Storage 50: 104646. https://doi.org/10.1016/j.est.2022.104646 doi: 10.1016/j.est.2022.104646
|
| [3] |
Vignesh S, Che HS, Selvaraj J, et al. (2024) State of Health (SoH) estimation methods for second life lithium-ion battery—Review and challenges. Appl Energy 369: 123542. https://doi.org/10.1016/j.apenergy.2024.123542 doi: 10.1016/j.apenergy.2024.123542
|
| [4] |
Tang H, Chen J, Long Y, et al. (2025) Joint estimation of SOC and SOH for lithium-ion batteries via adaptive variable structure observers. IEEE Trans Ind Electron 72: 10605–10615. https://doi.org/10.1109/TIE.2025.3546355 doi: 10.1109/TIE.2025.3546355
|
| [5] |
Tian Y, Lin C, Xiong R (2025) Industry-oriented roadmap for lithium plating detection from short to long term. Green Energy Intell Transp 4: 100290. https://doi.org/10.1016/j.geits.2025.100290 doi: 10.1016/j.geits.2025.100290
|
| [6] |
Bavand A, Khajehoddin SA, Ardakani M, et al. (2022) Online estimations of li-ion battery SOC and SOH applicable to partial charge/discharge. IEEE Trans Transp Electrif 8: 3673–3685. https://doi.org/10.1109/TTE.2022.3162164 doi: 10.1109/TTE.2022.3162164
|
| [7] |
Pang Z, Yang K, Song Z, et al. (2023) A new method for determining SOH of lithium batteries using the real-part ratio of EIS specific frequency impedance. J Energy Storage 72: 108693. https://doi.org/10.1016/j.est.2023.108693 doi: 10.1016/j.est.2023.108693
|
| [8] |
Demirci O, Taskin S, Schaltz E, et al. (2024) Review of battery state estimation methods for electric vehicles-Part Ⅱ: SOH estimation. J Energy Storage 96: 112703. https://doi.org/10.1016/j.est.2024.112703 doi: 10.1016/j.est.2024.112703
|
| [9] |
Greenbank S, Howey D (2022) Automated feature extraction and selection for data-driven models of rapid battery capacity fade and end of life. IEEE Trans Ind Inf 18: 2965–2973. https://doi.org/10.1109/TII.2021.3106593 doi: 10.1109/TII.2021.3106593
|
| [10] |
Yang N, Song Z, Hofmann H, et al. (2022) Robust State of Health estimation of lithium-ion batteries using convolutional neural network and random forest. J Energy Storage 48: 103857. https://doi.org/10.1016/j.est.2021.103857 doi: 10.1016/j.est.2021.103857
|
| [11] |
Liu S, Nie Y, Tang A, et al. (2023) Online health prognosis for lithium-ion batteries under dynamic discharge conditions over wide temperature range. eTransportation 18: 100296. https://doi.org/10.1016/j.etran.2023.100296 doi: 10.1016/j.etran.2023.100296
|
| [12] |
Wang D, Li C, Wei Z, et al. (2025) A multi-forward-step lithium-ion battery SOH prediction method based on segmented gated recurrent unit and DualMamba. J Energy Storage 117: 116136. https://doi.org/10.1016/j.est.2025.116136 doi: 10.1016/j.est.2025.116136
|
| [13] |
Wang J, Wang D, Li H, et al. (2026) Parameter identification and state of charge estimation for lithium-ion batteries based on trust region least squares. AIMS Energy 14: 23–46. https://doi.org/10.3934/energy.2026002 doi: 10.3934/energy.2026002
|
| [14] |
Vennam G, Sahoo A (2023) A dynamic SOH-coupled lithium-ion cell model for state and parameter estimation. IEEE Trans Energy Convers 38: 1186–1196. https://doi.org/10.1109/TEC.2022.3218344 doi: 10.1109/TEC.2022.3218344
|
| [15] |
Li X, Ju L, Geng G, et al. (2023) Data-driven state-of-health estimation for lithium-ion battery based on aging features. Energy 274: 127378. https://doi.org/10.1016/j.energy.2023.127378 doi: 10.1016/j.energy.2023.127378
|
| [16] |
Xie Y, Wang S, Zhang G, et al. (2024) A review of data-driven whole-life state of health prediction for lithium-ion batteries: Data preprocessing, aging characteristics, algorithms, and future challenges. J Energy Chem 97: 630–649. https://doi.org/10.1016/j.jechem.2024.06.017 doi: 10.1016/j.jechem.2024.06.017
|
| [17] |
Zhang S, Liu Z, Xu Y, et al. (2025) An electrochemical aging-informed data-driven approach for health estimation of lithium-ion batteries with parameter inconsistency. IEEE Trans Power Electron 40: 7354–7369. https://doi.org/10.1109/TPEL.2025.3532588 doi: 10.1109/TPEL.2025.3532588
|
| [18] |
E L, Wang J, Yang R, et al. (2025) A physics-informed neural network-based method for predicting degradation trajectories and remaining useful life of supercapacitors. Green Energy Int Transp 4: 100291. https://doi.org/10.1016/j.geits.2025.100291 doi: 10.1016/j.geits.2025.100291
|
| [19] |
Wang S, Ma C, Gao H, et al. (2025) Improved hyperparameter Bayesian optimization-bidirectional long short-term memory optimization for high-precision battery state of charge estimation. Energy 328: 136598. https://doi.org/10.1016/j.energy.2025.136598 doi: 10.1016/j.energy.2025.136598
|
| [20] |
Wang S, Zhou J, Li L, et al. (2026) Review on data-driven state of health estimation methods for lithium-ion batteries. J Power Sources 690: 240783. https://doi.org/10.1016/j.jpowsour.2026.240783 doi: 10.1016/j.jpowsour.2026.240783
|
| [21] |
Wang F, Zhai Z, Liu B, et al. (2024) Open access dataset, code library and benchmarking deep learning approaches for state-of-health estimation of lithium-ion batteries. J Energy Storage 77: 109884. https://doi.org/10.1016/j.est.2023.109884 doi: 10.1016/j.est.2023.109884
|
| [22] |
Zhang Y, Wik T, Bergström J, et al. (2024) State of health estimation for lithium-ion batteries under arbitrary usage using data-driven multimodel fusion. IEEE Trans Transp Electrif 10: 1494–1507. https://doi.org/10.1109/TTE.2023.3267124 doi: 10.1109/TTE.2023.3267124
|
| [23] |
Bian J, Liu G, Chen J, et al. (2025) PSO-MLSt-LSTM: Multi-layer stacked ensemble model for lithium-ion battery SOH prediction via multi-feature fusion. J Energy Storage 125: 116825. https://doi.org/10.1016/j.est.2025.116825 doi: 10.1016/j.est.2025.116825
|
| [24] |
Sun J, Fan C, Yan H (2024) SOH estimation of lithium-ion batteries based on multi-feature deep fusion and XGBoost. Energy 306: 132429. https://doi.org/10.1016/j.energy.2024.132429 doi: 10.1016/j.energy.2024.132429
|
| [25] |
Kumar R, Das K, Krishna A (2024) Comparative analysis of data-driven electric vehicle battery health models across different operating conditions. Energy 309: 133155. https://doi.org/10.1016/j.energy.2024.133155 doi: 10.1016/j.energy.2024.133155
|
| [26] |
Chen L, Bao X, Lopes AM, et al. (2023) State of health estimation of lithium-ion batteries based on equivalent circuit model and data-driven method. J Energy Storage 73: 109195. https://doi.org/10.1016/j.est.2023.109195 doi: 10.1016/j.est.2023.109195
|
| [27] |
Mu G, Wei Q, Xu Y, et al. (2025) State of health estimation of lithium-ion batteries based on feature optimization and data-driven models. Energy 316: 134578. https://doi.org/10.1016/j.energy.2025.134578 doi: 10.1016/j.energy.2025.134578
|
| [28] |
Zhou Z, Duan B, Kang Y, et al. (2023) Practical state of health estimation for LiFePO4 batteries based on gaussian mixture regression and incremental capacity analysis. IEEE Trans Ind Electron 70: 2576–2585. https://doi.org/10.1109/TIE.2022.3167142 doi: 10.1109/TIE.2022.3167142
|
| [29] |
He Y, Deng Z, Chen J, et al. (2024) State-of-health estimation for fast-charging lithium-ion batteries based on a short charge curve using graph convolutional and long short-term memory networks. J Energy Chem 98: 1–11. https://doi.org/10.1016/j.jechem.2024.06.024 doi: 10.1016/j.jechem.2024.06.024
|
| [30] |
Tang J, Li Y, Wang S, et al. (2024) Data-driven state of health estimation method of lithium-ion batteries for partial charging curves. IEEE Trans Energy Convers 39: 2230–2243. https://doi.org/10.1109/TEC.2024.3407136 doi: 10.1109/TEC.2024.3407136
|
| [31] |
He Y, Bai W, Wang L, et al. (2024) SOH estimation for lithium-ion batteries: An improved GPR optimization method based on the developed feature extraction. J Energy Storage 83: 110678. https://doi.org/10.1016/j.est.2024.110678 doi: 10.1016/j.est.2024.110678
|
| [32] |
Lyu Z, Tang Y, Wu Z, et al. (2024) Online state of health estimation for Li-ion batteries in EVs through a data-fusion-model method. J Energy Storage 100: 113588. https://doi.org/10.1016/j.est.2024.113588 doi: 10.1016/j.est.2024.113588
|
| [33] |
Sun J, Zhang X, Cao W, et al. (2025) State of health estimation of lithium-ion batteries based on data-driven methods with a selected charging voltage interval. AIMS Energy 13: 290–308. https://doi.org/10.3934/energy.2025012 doi: 10.3934/energy.2025012
|
| [34] |
Peng S, Zhu J, Wu T, et al. (2024) SOH early prediction of lithium-ion batteries based on voltage interval selection and features fusion. Energy 308: 132993. https://doi.org/10.1016/j.energy.2024.132993 doi: 10.1016/j.energy.2024.132993
|
| [35] |
Khaleghi Rahimian S, Tang Y (2023) A practical data-driven battery state-of-health estimation for electric vehicles. IEEE Trans Ind Electron 70: 1973–1982. https://doi.org/10.1109/TIE.2022.3165295 doi: 10.1109/TIE.2022.3165295
|
| [36] |
Sui X, He S, Vilsen SB, et al. (2021) A review of non-probabilistic machine learning-based state of health estimation techniques for lithium-ion battery. Appl Energy 300: 117346. https://doi.org/10.1016/j.apenergy.2021.117346 doi: 10.1016/j.apenergy.2021.117346
|
| [37] |
Feng X, Zhang Y, Xiong R, et al. (2024) Comprehensive performance comparison among different types of features in data-driven battery state of health estimation. Appl Energy 369: 123555. https://doi.org/10.1016/j.apenergy.2024.123555 doi: 10.1016/j.apenergy.2024.123555
|
| [38] |
Sun J, Wang H (2025) State of health estimation for lithium-ion batteries based on optimal feature subset algorithm. Energy 322: 135685. https://doi.org/10.1016/j.energy.2025.135685 doi: 10.1016/j.energy.2025.135685
|
| [39] |
Ma W, Xue J, Li Y, et al. (2025) A stable lithium-ion battery SOH estimation framework for suppressing measurement noise with unknown distribution. IEEE Trans Transp Electrif 11: 10336–10353. https://doi.org/10.1109/TTE.2025.3554735 doi: 10.1109/TTE.2025.3554735
|
| [40] |
Wang F, Zhai Z, Zhao Z, et al. (2024) Physics-informed neural network for lithium-ion battery degradation stable modeling and prognosis. Nat Commun 15: 4332. https://doi.org/10.1038/s41467-024-48779-z doi: 10.1038/s41467-024-48779-z
|
| [41] |
Ryu HH, Sun HH, Myung ST, et al. (2021) Reducing cobalt from lithium-ion batteries for the electric vehicle era. Energy Environ Sci 14: 844–852. https://doi.org/10.1039/D0EE03581E doi: 10.1039/D0EE03581E
|
| [42] |
Nan D, Wang P, Jia Y, et al. (2026) Multi-stress accelerated aging for cycle life evaluation of high-capacity, long-life lithium iron phosphate batteries. Appl Energy 404: 127126. https://doi.org/10.1016/j.apenergy.2025.127126 doi: 10.1016/j.apenergy.2025.127126
|
| [43] |
Wassiliadis N, Schneider J, Frank A, et al. (2021) Review of fast charging strategies for lithium-ion battery systems and their applicability for battery electric vehicles. J Energy Storage 44: 103306. https://doi.org/10.1016/j.est.2021.103306 doi: 10.1016/j.est.2021.103306
|
| [44] |
Xie W, Liu X, He R, et al. (2020) Challenges and opportunities toward fast-charging of lithium-ion batteries. J Energy Storage 32: 101837. https://doi.org/10.1016/j.est.2020.101837 doi: 10.1016/j.est.2020.101837
|
| [45] |
Zhao J, Ling H, Liu J, et al. (2023) Machine learning for predicting battery capacity for electric vehicles. Etransportation 15: 100214. https://doi.org/10.1016/j.etran.2022.100214 doi: 10.1016/j.etran.2022.100214
|
| [46] |
Chai X, Li S, Liang F (2024) A novel battery SOC estimation method based on random search optimized LSTM neural network. Energy 306: 132583. https://doi.org/10.1016/j.energy.2024.132583 doi: 10.1016/j.energy.2024.132583
|
| [47] |
Xia F, Wang K, Chen J (2023) State of health and remaining useful life prediction of lithium-ion batteries based on a disturbance-free incremental capacity and differential voltage analysis method. J Energy Storage 64: 107161. https://doi.org/10.1016/j.est.2023.107161 doi: 10.1016/j.est.2023.107161
|
| [48] |
Mao K, Yao Y, Chen Y, et al. (2024) Fracture mechanisms of NCM polycrystalline particles in lithium-ion batteries: A review. J Energy Storage 84: 110807. https://doi.org/10.1016/j.est.2024.110807 doi: 10.1016/j.est.2024.110807
|
| [49] |
Zeng J, Liu S (2023) Research on aging mechanism and state of health prediction in lithium batteries. J Energy Storage 72: 108274. https://doi.org/10.1016/j.est.2023.108274 doi: 10.1016/j.est.2023.108274
|
| [50] |
Zhang Y, Liu Y, Wang J, et al. (2022) State-of-health estimation for lithium-ion batteries by combining model-based incremental capacity analysis with support vector regression. Energy 239: 121986. https://doi.org/10.1016/j.energy.2021.121986 doi: 10.1016/j.energy.2021.121986
|
| [51] |
Wu W, Wang S, Fan Y, et al. (2025) A hybrid squeeze excitation gate recurrent unit-autoregressive integrated moving average model for long-term state of health estimation of lithium-ion batteries with adaptive enhancement ability. J Energy Storage 131: 117600. https://doi.org/10.1016/j.est.2025.117600 doi: 10.1016/j.est.2025.117600
|
| [52] |
Hou J, Liu J, Chen F, et al. (2023) Robust lithium-ion state-of-charge and battery parameters joint estimation based on an enhanced adaptive unscented Kalman filter. Energy 271: 126998. https://doi.org/10.1016/j.energy.2023.126998 doi: 10.1016/j.energy.2023.126998
|
| [53] |
Wu M, Yue C, Zhang F, et al. (2024) State of health estimation and remaining useful life prediction of lithium-ion batteries by charging feature extraction and ridge regression. Appl Sci 14: 3153. https://doi.org/10.3390/app14083153 doi: 10.3390/app14083153
|
| [54] |
Wang X, Wang X, Ma B, et al. (2021) High precision error prediction algorithm based on ridge regression predictor for reversible data hiding. IEEE Signal Process Lett 28: 1125–1129. https://doi.org/10.1109/LSP.2021.3080181 doi: 10.1109/LSP.2021.3080181
|
| [55] |
Li X, Yuan C, Wang Z (2020) State of health estimation for Li-ion battery via partial incremental capacity analysis based on support vector regression. Energy 203: 117852. https://doi.org/10.1016/j.energy.2020.117852 doi: 10.1016/j.energy.2020.117852
|
Figures(15) / Tables(9)
Yongxun Liu, Penglong Sun, Ke Zhao, Feng Zhao, Bin Wang. State-of-health prediction for lithium batteries based on feature decoupling and heterogeneous stacking ensemble learning framework[J]. AIMS Energy, 2026, 14(4): 923-954. doi: 10.3934/energy.2026037
DownLoad: