Certain investigation and implementation of Coulomb counting based unscented Kalman filter for state of charge estimation of lithium-ion batteries used in electric vehicle application

At various operating conditions, battery performance and attributes will change. By employing precise, effective circuit and battery models, designers can forecast and optimize battery run-time, the current state of charge (SOC), and circuit performance. Failure to anticipate SOC will result in overcharging or over-discharging, which may permanently harm the battery cells. In this paper, a Coulomb Counting-based Unscented Kalman filter (UKF) is proposed for the accurate estimation of SOC of the Lithium-ion battery. Unscented Kalman filter is chosen since it have better performance for non-linear system. Coulomb counter (CC) is employed to overcome the drawback of capacity degradation and power fading of the battery due to increase in degradation cycle. Thevenin equivalent circuit model comprising of a single RC network is considered as battery model. The proposed model is simulated in MATLAB software. The input to the UKF is temperature, voltage including gaussian noise and the battery capacity estimated by CC method. Simulation results obtained shows that degradation cycle have higher impact on rated capacity. Variation of temperature during charging and discharging cycle was also analyzed. The simulated output of UKF shows that, UKF along with CC estimate SOC of the battery with the minimized SOC estimation error of <1% compared to conventional method.