Impact of sensor accuracy of battery management system on SOC estimation of electric vehicle based on EKF algorithm

Abstract

State of Charge (SOC) estimation is one of the most important functions of the battery management system (BMS) for new energy vehicles. Extended Kalman Filter (EKF) algorithm has been widely used in SOC estimation of lithiumion batteries (LiBs). However, the input variables in the SOC estimation model, such as current and voltage, are affected by the accuracy of the sensor. When the measured values of the sensor drift, the SOC estimation model based on EKF can obtain accurate SOC estimation results. In view of these defects, based on the simulation model, the influence of current, terminal voltage measurement error and measurement noise on SOC estimation accuracy of EKF algorithm is quantitatively analyzed. The results show that the measurement errors of terminal voltage and current will affect the accuracy of SOC estimation by EKF algorithm, and the measurement noise of terminal voltage and current has little effect on the accuracy of SOC estimation by EKF algorithm. In the design of battery management system, in order to ensure SOC estimation error less than 5 %, the system error of terminal voltage sensor should be controlled within 5 ‰, and the system error of current sensor should be controlled within 4 %.