State of charge estimation at different temperatures based on dynamic thermal model for lithium-ion batteries

Abstract

In the existing modeling and state of charge (SOC) estimation of lithium batteries, the effect of ambient temperature is often ignored, which brings large errors to the model parameter identification, so the battery modeling and SOC estimation under temperature variation are performed. In this paper, an improved dual-polarized dynamic thermal model is proposed, firstly, the thermal effect of lithium battery is analyzed, and the model parameters adopt the dynamic values of coupling temperature and SOC, and the model is improved in terms of both heat generation mechanism and heat transfer characteristics; then the model parameters are identified based on the least-squares method with variable forgetting factor (VFF-RLS). Finally, the SOC at different temperatures is evaluated based on the dynamic thermal model combined with the extended Kalman filtering (EKF) with nonlinear filtering capability. The results show that the temperature factor is an important and non-negligible influence factor in modeling, the proposed battery model in this paper has high computational accuracy, and the proposed SOC estimation method has better SOC estimation capability for the temperature variation problems.