SOC Estimation Based on Modified Covariance Extended Kalman Filter for Power Batteries of Electric Vehicles

摘要

Accurate estimation of state of charge (SOC) of power battery is vital for electric vehicles. Extended Kalman Filter (EKF) is well accepted to estimate SOC, but it is largely affected by measurement noise and the covariance is easy to appear morbidity during the recursive process. This paper proposes a modified covariance Extended Kalman Filter (MVEKF) algorithm, which recalculates the covariance with the modified estimation and updates the process gain. The new covariance is used for the next state estimation. Through pulse discharge and dynamic stress test (DST), the results show that the MVEKF is more accurate and tend to be stable faster compared with the EKF. Healthy covariances reduce operational error, so the MVEKF can improve estimation of SOC.