An online model identification for state of charge estimation of lithium-ion batteries using extended kalman filter

Abstract

The accurate estimation of state of charge (SOC) of Lithium-ion battery is one of the most crucial issues for battery management system (BMS). The parameters of battery model varying at different status have a strong impact on SOC estimation. This paper presents an online model identification for Thevenin model using the forgetting factor recursive least squares (FRLS) algorithm. The extended Kalman Filter (EKF) estimator for SOC estimation of Lithium-ion battery is proposed based on the online parameter estimation. The federal urban driving schedule (FUDS) cycles is used to assess the superiority in improving the accuracy of the online model identification-based EKF algorithm with the offline model identification-based EKF method. Experimental results have demonstrated that the proposed online-EKF can not only achieve higher modelling accuracy but also improve the stability of SOC estimation. The maximun SOC estimation error is less than 1% under FUDS cycles.