On-line battery state-of-charge estimation based on an integrated estimator

Abstract

Energy crises and environmental issues have promoted research into development of various types of electric vehicles (EVs). Since the control strategy of EVs is essentially dependent on the state-of-charge (SoC) estimation of the batteries, one of the most critical issues of battery management system (BMS) is to accurately estimate the SOC in real-time. This paper proposed an integrated SoC estimator based on the adaptive extended Kalman filter (EKF) and particle filter (PF). The adaptive EKF which can provides more accurate approximate distribution in PF, is used for on-line parameters estimation of polarization voltage and noise covariance. The PF is used for on-line SoC estimation based on the priori knowledge given by the adaptive EKF. In order to get accurate SoC estimation results, the cell model is established based on the integration of the equivalent circuit model and electrochemical model. The combined electrochemical model is employed to simulate the cell electrochemical characteristics and estimate the terminal voltage. What is more, the recursive least-squares (RLS) method is used for parameters identification to improve the model precision. Experiments are performed on the LiFePO4 cell at different temperatures and under dynamic current to verify the reliability and robustness of the proposed method. The results indicated that accurate and robust SoC estimation results can be obtained by the proposed method.