Multipoint Temperature Estimation for Large-Format Prismatic Lithium-Ion Batteries With Ultralow Sensor Requirements

Accurate temperature monitoring is crucial for the safe operation of large-format prismatic lithium-ion batteries. However, the significant thermal heterogeneity across the battery surface and localized heating caused by the high internal resistance of the positive electrode make it difficult to characterize the thermal characteristics of the battery through a single point temperature. Furthermore, the limited number of temperature sensors in practical applications presents a huge challenge to high-precision temperature monitoring. To address these challenges, a novel sensorless state of temperature (SOT) estimation framework based on a dual-heat-source 8-node thermal network model is developed in this study. Independent heat sources are strategically positioned at both the cathode interface and the core of the battery, and an 8-node simplified thermal transfer model architecture is established, which effectively balances computational accuracy with model complexity. Then, with the terminal voltage serving as the observation variable, a closed-loop observation system for state of charge (SOC) and SOT joint estimation is constructed based on the extended Kalman filter (EKF) algorithm, achieving real-time high-precision estimation of multi-point temperature in lithium-ion batteries under dynamic environmental conditions. The proposed method is evaluated under various operating conditions, including multiple noise environments, different driving simulation conditions. Results show that the RMSE of the SOT estimation after system stabilization remains around 1.67 °C.