Lumped semi-empirical model for lithium-ion batteries considering temperature effects

Constructing an accurate battery model is fundamental for state estimation in lithium-ion batteries. The traditional lumped semi-empirical model (LPM) combines the advantages of both electrochemical and equivalent circuit models (ECM). However, the influence of external temperature on model parameters is neglected during the modeling process, which leads to reduced accuracy in varying temperature environments. To address this issue, the temperature effects on the battery model are explored, the sources of model error are identified, and a lumped semi-empirical model considering temperature effects (LPM-T) is proposed. In the model construction process, battery characterization and dynamic load experiments are conducted on lithium-ion batteries at different temperatures. From these experiments, we extract battery capacity and open circuit voltage (OCV) at specific temperatures, and identify the corresponding internal resistance parameters at these temperatures based on the LPM. Additionally, mapping relationships between these model parameters and temperature are established using polynomial fitting and the Arrhenius equation. Finally, the proposed model is validated through a high-rate variable-temperature dynamic cycling load and compared with the other three battery models. The results show that the proposed model significantly improves the accuracy of terminal voltage predictions under variable-temperature conditions, offering enhanced applicability and robustness.