State-of-charge estimation for lithium-ion batteries based on modified unscented Kalman filter using improved parameter identification

The full and single hybrid pulse power characterization (HPPC) experiments are conducted on nickel cobalt manganese (NCM) and lithium iron phosphate (LFP) batteries to obtain the accurate relationship between state of charge (SOC) and open circuit voltage (OCV). The pseudo random number generated by the RAND function is used as the initial value of the parameter identification equation for the second-order equivalent circuit model (ECM), and the parameter identification of lithium-ion batteries is achieved through least squares fitting. The unscented Kalman filter (UKF) simulation model is modified by the first order low pass filtering (FOLPF) algorithm to improve the accuracy of SOC estimation of batteries. The results show that the goodness of fit between the real and estimated HPPC pulse voltage values at different SOC values is above 0.99 for both NCM and LFP batteries. Based on the parameter identification results, the maximum error of the HPPC voltage estimated by the second-order ECM is within 0.045 V under both full pulse and single pulse testing conditions while the maximum error of SOC estimation is within 0.025. The UKF model modified by the FOLPF algorithm provides a reference for the accurate SOC estimation of batteries. •The initial value of parameter identification is determined by the RAND function.•A UKF model is modified by first order low pass filtering algorithm.•The AE of voltage and SOC estimation are within 0.045 V and 0.025, respectively.•The modified UKF method is suitable for SOC estimation of NCM and LFP batteries.