Study On Estimation Algorithm Of SOC Extended Kalman Filter Considering Time-varying Battery Parameters

With the rapid development of electric vehicles,power batteries as the core components of electric vehicles have been paid more and more attention.Electric vehicle power batteries must be equipped with battery management systems(BMS).State of charge(SOC)estimation for battery is the most basic and core function of BMS.Accurate estimation of SOC has always been the focus and difficulty of BMS development.In this paper,the method of battery SOC estimation is studied by taking lithium-ion battery as the research object and combining with the nonlinear characteristics of battery.The main contents of this paper are as follows:(1)Characteristic experiment and performance analysis of lithium-ion battery.By taking NCR18650 PF power battery pack as the research object,the experimental platform of battery pack is built.The working characteristics and influence factors for performance of lithium-ion batteries are studied experimentally.The definition of SOC is proposed by considering the influence factors of battery capacity comprehensively.It lays a theoretical foundation for the identification of rear battery parameters and SOC estimation.(2)Parameter identification and modeling of battery model.By comparing different types of equivalent circuit models,the second-order RC circuit equivalent model is selected.Because of the time-varying and nonlinear characteristics of the battery system,the RC parameters of the model will change greatly with the difference of charge and discharge ratio,SOC and temperature.In this paper,a simulated annealing algorithm is used to identify the RC parameters of the battery according to rebound voltage data under different discharge magnification and SOC.The variation of model parameters with discharge ratio and SOC is obtained by statistics in the offline state,and the parameters of the dynamic battery model of continuous charge and discharge state can be obtained by the method of table checking.Then,the on-line identification of the model parameters is completed.Based on MATLAB/Simulink,the power battery model is built,and the operating current value is used as the model input.Error between the experimental value and simulation value of the battery end voltage is small,which verifies the accuracy of the simulated annealing algorithm for the parameter identification of the power battery.(3)Battery pack SOC estimation according to extended Kalman filter(EKF)algorithm for parameter time-varying model.Based on the second-order RC equivalent circuit model,the EKF algorithm model is built in MATLAB/Simulink.In the different state of current and SOC,the system matrix parameters of EKF are updated in real time according to the parameters of the battery model.Error between the estimated SOC value and the theoretical value are small,which verifies the accuracy of the SOC estimation method.(4)Experiment based on hardware-in-the-loop simulation system of SOC estimation strategy.Based on MATLAB/Simulink and system of NI/PXI real-time measurement and control,platform based on hardware-in-loop simulation system of SOC estimation strategy is built.Deploying the battery model in a real-time processor and importing the SOC estimation strategy into a real controller,then the real-time and effectiveness of SOC estimation strategy are verified by semi-physical simulation technology.