A multi-physical feature-driven LSTM network with dual attention for a wide-temperature-range SOC estimation of LFP batteries AITranslate
Abstract AITranslate
Accurate state of charge (SOC) estimation for lithium iron phosphate (LFP) batteries across a wide temperature range remains challenging, owing to their flat open-circuit voltage (OCV) plateau and pronounced temperature sensitivity. To address this, a Long Short-Term Memory (LSTM) network is proposed that integrates internal pressure features with a dual-attention mechanism. The measured internal pressure is introduced as a new physical observation to compensate for the lack of SOC-identifying information in the voltage plateau region of LFP batteries. The model jointly learns from internal pressure, temperature, voltage, and current. The LSTM's gating mechanism captures the dynamic coupling between these multi-source signals and SOC under wide-temperature conditions, while the dual-attention mechanism further improves estimation accuracy and interpretability. Experiments on the own dataset show that the method achieves root mean square errors (RMSE) of 1.44%, 1.14%, and 1.34% at 25℃, 35℃, and 45℃, respectively. In domain adaptation tests, fine-tuning with only 15% of the target temperature samples yields RMSE of 1.55% at 35℃ and 1.97% at 45℃. Cross-domain validation on a public National Aeronautics and Space Administration (NASA) dataset, using only 10% of target domain samples for fine-tuning, results in an RMSE of 6.25% and an r2 of 0.9339. Ablation studies confirm the unique contribution of the internal pressure feature. This work provides a robust, interpretable, and high-accuracy framework for LFP battery SOC estimation, paving the way for safer and more reliable battery management systems.
KeyWords AITranslate
1.Y. K. Hu, Y. F. Xie, L. Y. Zhao, et al., "AI-Enabled Cross-Scale Modeling and Parameter Transfer for Lithium-ion Batteries," Energy Storage Materials 89 (2026): 105222, https://doi.org/10.1016/j.ensm.2026.105222.
2.E. Sandrucci, S. Brutti, and F. Marini, "Chemometric Models and Electrochemical Techniques for Studying the Health of Li-Ion Batteries: A Review," Journal of The Electrochemical Society 173, no. 10 (2026): 100504, https://doi.org/10.1149/1945-7111/ae6893.
3.G. Liu, X. Tong, W. Ma, et al., "A Robust Non-Linear Method for the State-of-Health Estimation for Lithium-Ion Batteries Based on Dissipativity Theory for Electric Vehicle Applications," Multiscale and Multidisciplinary Modeling, Experiments and Design 7, no. 3 (2024): 2773–2781, https://doi.org/10.1007/s41939-023-00359-2.
4.L. S. Kham and N. Aunsri, "Bayesian Filtering Techniques for State of Charge and State of Health Estimation in Lithium-Ion Batteries for Electric Vehicles: Methods, Challenges, and Future Directions," Applied Energy 417 (2026): 128015, https://doi.org/10.1016/j.apenergy.2026.128015.
5.A. E. Alias and T. George, "Intelligent Estimation of Lithium-Ion Battery SOC and Temperature under Uncertainty: A Hybrid Approach for High-Performance Electric Vehicle Applications," Energy Technology 14, no. 3 (2026): e202502574, https://doi.org/10.1002/ente.202502574.
6.Y. Xu, J. Shi, X. Liu, et al., "Investigation of the Thermal Management Performance of a Lithium-Ion Battery Pack Using a Layered Cascaded Phase Change Coupled Dynamic Heat Transfer Model," Journal of Energy Storage 155 (2026): 121475, https://doi.org/10.1016/j.est.2026.121475.
7.Z. Li, H. Ni, W. Zhu, et al., "Adaptive SOC-OCV Mapping-Based Joint Estimation of SOC and SOH in Aging Lithium-Ion Batteries Using Extended Kalman Filtering," Journal of Power Electronics 26, no. 1 (2025): 1–13, https://doi.org/10.1007/s43236-025-01064-8.
8.R. Wang, B. X. Zhou, M. Li, and Z. H. Rao, "Dynamic Modeling of Lithium-Ion Batteries with Mechanism-Based and Data-Driven Approach," Journal of Energy Storage 156 (2026): 121684, https://doi.org/10.1016/j.est.|2026.121684.
9.K. X. Li and J. L. Li, "Feature Selection and State of Health Estimation Based on Segmented Voltage Data of Lithium-Ion Batteries," Journal of The Electrochemical Society 173, no. 12 (2026): 120508, https://doi.org/10.1149/1945-7111/ae73f4.
10.Y. Du, D. Liu, and Y. Song, "Remaining Useful Life Prediction of Lithium-Ion Battery Packs Using Prior Cell Degradation and Limited Pack-Level Data," Energy 360 (2026): 141948, https://doi.org/10.1016/j.energy.2026.141948.
11.N. K. Rajalwal, R. K. Upadhyay, D. Ghosh, and B. Khan, "Hybrid Support Vector and Gaussian Process Regression Approach for Secondary-Life State-of-Health Estimation of Lithium-Ion Batteries," Electric Power Systems Research 259 (2026): 113294, https://doi.org/10.1016/j.epsr.2026.113294.
12.J. Zhu, C. Peng, H. Lai, et al., "A Novel Aggregation Modeling Method for Lithium-Ion Batteries Based on Artificial Intelligence and Digital Twin," Energy 358 (2026): 141345, https://doi.org/10.1016/j.energy.2026.141345.
13.R. Ahmed, M. S. Hossen, N. T. Tithi, et al., "Role of Deep Learning in Battery Management System (BMS) for Electric Vehicles – A Review," Energy Reports 15 (2026): 109028, https://doi.org/10.1016/j.egyr.2025.109028.
14.X. Qiu, Y. Luo, and S. Wang, "A Method for Reconstructing the Charging Voltage Curve of Lithium Batteries Based on a Deep Time-Series Prediction Framework," Thermochimica Acta 762 (2026): 180384, https://doi.org/10.1016/J.tca.2026.180384.
15.Y. Li, R. Wang, Y. Jin, et al., "A Novel SOC Estimation Method for Lithium-Ion Batteries Based on Serial LSTM-UKF Fusion," Energies 19, no. 6 (2026): 1467, https://doi.org/10.3390/en19061467.
16.C. Tang, Z. Wei, H. Wang, H. Rong, Z. Wang, and J. Wu, "Remaining Useful Life Prediction of Lithium-Ion Batteries Using a Denoising Autoencoder and an Interval Attention LSTM Network," Engineering Research Express 8, no. 9 (2026): 095318, https://doi.org/10.1088/2631-8695/ae6045.
17.Z. Sun, W. Ye, Y. Mao, and Y. Sui, "A Hybrid CNN–LSTM–Attention Mechanism Model for Anomaly Detection in Lithium-Ion Batteries of Electric Bicycles," Batteries 11, no. 10 (2025): 384, https://doi.org/10.3390/batteries11100384.
18.B. Han, B. Q. Shi, H. M. Xie, and Y. L. Kang, "Evolution of Electrochemically Induced Stress and Its Effect on the Lithium-Storage Performance of Graphite Electrode," Science China Technological Sciences 66, no. 6 (2023): 1784–1796, https://doi.org/10.1007/s11431-022-2298-7.
19.C. Chang, X. Ye, X. Li, A. Tian, F. Mei, and J. Jiang, "Lithium-Ion Battery Electrochemical-Thermal-Mechanical-Impedance Coupling Model and Voltage-Impedance Dual-Objective Parameter Identification Method," Journal of Energy Storage 152 (2026): 120748, https://doi.org/10.1016/j.est.2026.120748.
Basic Information:
DOI：10.23919/CHAIN.2026.000019
Chinese Library Classification Number:
Citation Information:
Accurate state of charge (SOC) estimation for lithium iron phosphate (LFP) batteries across a wide temperature range remains challenging, owing to their flat open-circuit voltage (OCV) plateau and pronounced temperature sensitivity. To address this, a Long Short-Term Memory (LSTM) network is proposed that integrates internal pressure features with a dual-attention mechanism. The measured internal pressure is introduced as a new physical observation to compensate for the lack of SOC-identifying information in the voltage plateau region of LFP batteries. The model jointly learns from internal pressure, temperature, voltage, and current. The LSTM's gating mechanism captures the dynamic coupling between these multi-source signals and SOC under wide-temperature conditions, while the dual-attention mechanism further improves estimation accuracy and interpretability. Experiments on the own dataset show that the method achieves root mean square errors (RMSE) of 1.44%, 1.14%, and 1.34% at 25℃, 35℃, and 45℃, respectively. In domain adaptation tests, fine-tuning with only 15% of the target temperature samples yields RMSE of 1.55% at 35℃ and 1.97% at 45℃. Cross-domain validation on a public National Aeronautics and Space Administration (NASA) dataset, using only 10% of target domain samples for fine-tuning, results in an RMSE of 6.25% and an r2 of 0.9339. Ablation studies confirm the unique contribution of the internal pressure feature. This work provides a robust, interpretable, and high-accuracy framework for LFP battery SOC estimation, paving the way for safer and more reliable battery management systems.
quote
| GB/T 7714-2015 | [1] Wenju Ren, Maolin Guo, Jie Yang, et al. A multi-physical feature-driven LSTM network with dual attention for a wide-temperature-range SOC estimation of LFP batteries[J]. Chain, 2026, 3(3): 453-470. DOI:10.23919/CHAIN.2026.000019. |
| MLA | [1] Wenju Ren, et al., "A multi-physical feature-driven LSTM network with dual attention for a wide-temperature-range SOC estimation of LFP batteries." Chain, vol. 3, no. 3, 2026, pp. 453-470, https://doi.org/10.23919/CHAIN.2026.000019. |
| APA | [1] Wenju Ren, Maolin Guo, Jie Yang, Chang Liu, Sheng Lu, & Taixiong Zheng. (2026). A multi-physical feature-driven LSTM network with dual attention for a wide-temperature-range SOC estimation of LFP batteries. Chain, 3(3), 453-470. https://doi.org/10.23919/CHAIN.2026.000019 |
| IEEE | [1] Wenju Ren, Maolin Guo, Jie Yang, Chang Liu, Sheng Lu, and Taixiong Zheng, "A multi-physical feature-driven LSTM network with dual attention for a wide-temperature-range SOC estimation of LFP batteries," Chain, vol. 3, no. 3, pp. 453-470, 2026, doi: 10.23919/CHAIN.2026.000019. keywords: {attention mechanism;internal pressure;lithium-ion battery;Long Short-Term Memory network;state of charge} |