Abstract:

A hybrid pulse power characteristic (HPPC) experiment was carried out on a battery pack composed of three lithium iron phosphate batteries to investigate the different effects of various factors causing the state of charge (SOC) estimation error. A second-order Thevenin battery model was used to construct a state equation, and SOC was estimated using the unscented Kalman filter (UKF) algorithm, and the influences of the errors of the model parameters, input current, terminal voltage, and open circuit voltage (OCV) on SOC estimation were analyzed. The results show that the SOC estimation error is 19.1%, 20.6% and less than 3%, when 1% noise is added to the OCV, terminal voltage, and input current, or when the parameters of the battery (ohm internal resistance, polarization resistance, polarization capacitance) are fixed, respectively, showing that the OCV and the terminal voltage have a great influence on the SOC estimation accuracy. The estimation error of the SOC of the lithium battery can be greatly reduced by the improvement of the measurement accuracy of the terminal voltage and the accuracy of the OCV curve fitting.

Key words:  battery model , unscented Kalman filter ,  estimation of state of charge (SOC) ,  open circuit voltage