Search

Lithium‑ion Battery SOC Estimation under Broad‑Range Temperatures and Time‑Varying Operating Scenarios Using a Modified SR-UKF

Author(s)

Dan Zou

Affiliation(s)

College of Mechanical and Electrical Engineering, Yunnan Opening University, Kunming, Yunnan, China

Abstract

The State‑of‑Charge (SOC) constitutes a vital state indicator for the battery management system of power lithium‑ion batteries. Reliable State‑of‑Charge calculation provides prerequisites for safeguarding secure running of electrified automobiles and advancing onboard power allocation strategies. Faced with frequent model deviation, deteriorated computing precision and poor numerical robustness existing within power lithium‑ion battery SOC algorithms across broad‑range temperatures and time‑varying operating scenarios, this work investigates SOC calculation schemes for power lithium‑ion batteries that adapt to such complex thermal and operational circumstances. Based on the public LG‑18650‑HG2 dataset, the least‑squares method is adopted for offline identification of model parameters, and a parameter‑fitting correction model coupled with temperature and SOC is established. To overcome the limitations of the conventional square‑root unscented Kalman filter, such as fixed noise statistics and weak robustness under complicated operational environments, the Sage‑Husa adaptive noise strategy is introduced. A modified square‑root unscented Kalman filter is mathematically derived to realize dynamic updating of process and measurement noise covariance. This work completes a full set of theoretical designs covering battery modeling, parameter fitting and filter algorithm improvement. Theoretical analysis demonstrates that the modified algorithm can cope with model disturbances arising from broad‑range temperatures and time‑varying operating scenarios, effectively suppress calculation errors triggered by abrupt working‑condition shifts, and exhibits theoretical feasibility for high‑accuracy SOC calculation. The present work offers theoretical insights toward developing state‑calculation algorithms suited to power lithium‑ion batteries in complicated operational environments.

Keywords

SOC Estimation; Wide-Temperature-Range Parameter Identification; Square Root Unscented Kalman Filter; Sage-Husa Adaptive Noise

References

[1] Wang Q Y, LI H. Safety issues and failure analysis of lithium batteries. New Energy Technology, 2021(11): 15-19. [2] Wang M Z, XIAO Z L, ZHENG Y J. Sensitivity analysis of model parameters for lithium iron phosphate batteries at different temperatures. Journal of University of Shanghai for Science and Technology, 2022, 44(05): 449-456. [3] Jiang J C, SUN B X. Lithium-ion battery modeling and state estimation. Beijing: China Machine Press, 2021. [4] Chang S S. Research on modeling and state estimation of lithium-ion batteries under wide temperature range. Shanghai: Shanghai University of Electric Power, 2023. [5] Li X Y. Research on model construction and SOC estimation method of vehicle lithium-ion batteries. Harbin: Northeast Forestry University, 2025. [6] Zhang Z. Research on Accurate State of Temperature Estimation Method for Vehicle Lithium-ion Batteries under Full Temperature and Full Life Cycle. Kunming University of Science and Technology, 2024. [7] Xie Y, CHEN N, CHEN J, et al. Co-estimation of state of charge and model parameters for lithium-ion batteries based on variational Bayesian under time-varying noise and temperature conditions. Journal of Energy Storage, 2024, 96:112600. [8] Chen D M. Research on modeling and SOC estimation method of lithium-ion batteries for electric vehicles. Jiaozuo: Henan Polytechnic University, 2023. [9] Wang X D, DONG Z, WANG S H, et al. Wide-temperature and multi-condition SOC estimation of lithium-ion batteries based on improved open-circuit voltage model and adaptive square-root unscented Kalman filter. Transactions of China Electrotechnical Society, 2024, 39(24):7950-7964. [10] Huangfu H W, HAN A C. Research on equivalent model establishment and parameter identification method of lithium battery. Electric Switchgear, 2020, 58(03): 37-41. [11] Kollmeyer P, Vidal C, Naguib M, et al. LG 18650HG2 Li‑ion Battery Data and Example Deep Neural Network xEV SOC Estimator Script. Mendeley Data, 2020. DOI:10.17632/cp3473x7xv.3. [12] Fang J W. Research on Model Parameter Identification and State Estimation Method of Lithium Batteries. Yangzhou University, 2025. [13] Du B H, ZHANG Y, WU T Z, et al. Online identification method for equivalent model parameters of echelon utilization lithium-ion batteries. Energy Storage Science and Technology, 2021, 10(01): 342-348. [14] Julier S J, Uhlmann J K. Unscented filtering and nonlinear estimation. Proceedings of the IEEE, 2004, 92(3): 401-422. [15] Lu J T, PENG Z R. Structural response reconstruction with improved Sage-Husa Kalman filter based on adaptive window length. Chinese Journal of Computational Mechanics, 2025, 42(2):212-220.