Estimation of state of charge for lithium-ion battery based on unscented Kalman filter

Abstract

The accurate estimation of state of charge (SOC) of lithium-ion battery is the key technology for restricting the development of electric vehicles. A method for estimating the SOC of lithium-ion batteries based on unscented Kalman filter (UKF) was proposed to estimate the SOC. First, this paper describes the structure and operation mechanism of the system simulation model. Then, based on the Thevinin model, the state-space equations were established, and the principle of the UKF was analyzed, avoiding the linearization of the nonlinear state equation through the way of unscented transform, achieving accurate estimation of lithium-ion battery SOC in nonlinear conditions. Finally, a simulation experiment was performed under NEDC driving cycles. The results show that the UKF algorithm can accurately track the charging and discharging changes of lithium-ion batteries. The SOC estimated error of the lithium-ion battery is about 2%, achieving accurate estimation of lithium-ion SOC in nonlinear conditions, effectively improving the accuracy of SOC estimation.