SOC Estimation Of High Power Lithium-ion Battery Based On EKF

Safe and stable operation of electric vehicles depends on the battery management system.The estimation of state of charge(SOC)and identification of model parameter are two core elements of battery management system for lithium ion battery.In this paper,15Ah-NCM high-power lithium-ion batteries(LIB)was investigated with Solartron 1470E-1455 electrochemical workstation platform,with emphasis on the identification of model parameter and SOC estimation.The following four main contents are discussed:(1)This section was focused on several fundamental concepts in battery management system for lithium-ion battery.Firstly,working principle of lithium ion battery was introduced;followed by an analysis of the popular estimation methods for SOC estimation and their features;finally,the typical models of lithium-ion battery were summarized.(2)Parameters of second order circuit model for high-power lithium-ion battery was identified based on current-step method in this section.Three questions were discussed in detail:how to choose data range of output terminal voltage for nonlinear fitting;how to implement nonlinear least square algorithm for fitting the testing data to the objective function;how to obtain and understand parameters of charging/discharging circuit model.(3)The resting time's effect on parameter identification of circuit model for LIB via current-step method was studied with theoretical calculation and experimental test.Firstly,an analytical solution for a second-order circuit model of LIB was obtained under bi-pulse excitement,followed by calculating transient voltage of interfacial capacitance.Subsequently,the depth of charge/discharge was formulated for interfacial capacitance,and effect of resting time on charge stored in interfacial capacitance was quantitatively analyzed.When pulse width equal to 600s and resting time ranging from 600-2400s,depth of discharging(DOD)for capacitance with smaller time constant is close to 100%,DOD for that with larger time constant perturbs in 88.36?99.54%.Therefore,if resting time larger than 600s,the charge stored in interfacial capacitance due to discharge pulse imposes an ignored effect on charge relaxing at the end of next charge pulse.(4)Based on second-order circuit model of LIB and integration equation of ampere-hour method,state-space equation for SOC estimation was built.Subsequently,extended Kalman filter(EKF)algorithm was used to predict SOC of LIB.Simulation results show the developed model has better adaptability,and the EKF algorithm can improve the precision of SOC estimation for LIB.The method presented in this work for model parameter identification and SOC estimation of LIB could also be used in other kinds of batteries.