Closed-Loop Core Temperature Estimation for Lithium-Ion Batteries

Accurate estimation of the core temperature of lithium-ion batteries is critical for safe and reliable operation in modern energy storage systems. However, conventional lumped thermal models provide limited representation of internal temperature gradients, whereas high-fidelity thermal models may impose excessive computational demands for online applications. To address this issue, this work develops a closed-loop core-temperature estimation framework that integrates a physical multi-node thermal model with an extended Kalman filter (EKF). A Biot-number-based criterion is employed to determine the appropriate number of thermal nodes required to adequately represent the internal temperature distribution while avoiding unnecessary computational complexity. This multi-node structure is then embedded into an EKF, where measurable surface temperatures are used as observations, while the core temperature and heat-dissipation thermal resistances are estimated as state variables. This configuration enables surface-temperature feedback to continuously correct model predictions and compensate for variations in thermal boundary conditions. The method is experimentally validated on a 50-Ah prismatic lithium iron phosphate (LFP) cell equipped with an embedded core-temperature sensor. Results demonstrate that the proposed estimator achieves improved core-temperature estimation compared with open-loop model calculations while maintaining low computational cost and acceptable core-temperature estimation accuracy under the investigated variations in the entropy coefficient and thermal resistance. The proposed framework therefore provides a practical approach for real-time internal-temperature estimation under the investigated energy-storage operating conditions.