Modeling and SoC Estimation of Li-ion Batteries with an Improved Variable Forgetting Factor RLS Method Augmented with Extended Kalman Filter

Abstract

This paper presents an improved variable forgetting factor recursive least square (IVFF-RLS) and extended Kalman filter (EKF) based technique for accurate modeling and real-time state of charge (SoC) estimation of Li-ion batteries. In the proposed approach, the IVFF-RLS is used for an accurate estimation of varying battery parameters under abnormal change of operating states such as an abrupt shifting of the battery from charging to discharging state, data loss, etc. The IVFF-RLS is augmented with the extended Kalman filter (EKF) for real-time and improved SoC estimation of Li-ion batteries. Extensive validation studies are performed in the Matlab environment and then experimental studies have been carried out in the LabVIEW platform to validate the proposed IVFF-RLS-EKF technique. The outcomes of the experimental studies validate the higher accuracy and robustness of the proposed approach under a broad spectrum of operating temperature and system disturbances such as abrupt shifting from charging to discharging state and vice versa. The efficacy of the proposed approach has been compared against the coulomb counting technique (CCT) and traditional VFF-RLS-EKF approaches through experimental studies. The results show that the proposed IVFF-RLS-EKF technique outperforms the existing techniques ensuring highly accurate battery model parameters and SoC.