Estimation the internal resistance of lithium-ion-battery using a multi-factor dynamic internal resistance model with an error compensation strategy

• Published in: Energy reports 2021-11, Vol.7, p.3050-3059
• Main Authors: Chen, Lin, Zhang, Mo, Ding, Yunhui, Wu, Shuxiao, Li, Yijing, Liang, Gang, Li, Hao, Pan, Haihong
• Format: Article
• Language: English
• Subjects: Cubic spline; Error compensation strategy; Fourth-order polynomial; Internal resistance modeling; Multi-factor
• ISSN: 2352-4847; 2352-4847
• DOI: 10.1016/j.egyr.2021.05.027

Battery thermal management (BTM) is essential to ensure the safety of the battery pack of electric vehicles. For a variety of BTM technologies, the battery’s internal resistance always plays a critical role in the heat generation rate of the battery. Many factors (temperature, SOC and discharge rate) impact on the internal resistance, however, scant research has explored the effect of battery discharge rate on the internal resistance. This study aims to establish a multi-factor dynamic internal resistance model (MF-DIRM) with error compensation strategy to accurately estimate the internal resistance. In the present study, the internal resistance is estimated using the MF-DIRM which fuses three parameters (the temperature, SOC and discharge rate) and the procedures are two-fold: firstly, the coefficients of the binary fourth-order polynomial about the internal resistance with temperature and SOC are obtained, and then the MF-DIRM is built through the cubic spline interpolation algorithm to interpolate these coefficients at different discharge rates. Rather than measure the internal resistance, this model can estimate the internal resistance, and the estimated error does not exceed 10mΩ. Secondly, to further improve the estimation accuracy of MF-DIRM, the error function of the internal resistance is constructed as a binary cubic spline with the temperature and SOC, thereby constructing a model with error compensation strategy. The MF-DIRM combined with error compensation strategy can better estimate resistance under various conditions and the maximum value of estimation error is less than 2.5 mΩ.