Research On SOC Estimation Of Power Battery Based On Unscented Kalman Filter

New energy vehicles have become the new darling of the automotive industry in the international and domestic environment.As its representative,electric vehicles occupy a major market share and are also the top priority of its development.The power battery management system in electric vehicles is the core guarantee for the efficient and safe operation of the vehicle,and the State of Charge(SOC)is one of the key variables for the power battery management system to mange and control the power battery.Accurate SOC estimation will help improve the performance and the safety of the vehicle.The focus of this article is about the estimation of power battery SOC,and the Unscented Kalman Filter algorithm is selected to estimate SOC according to the non-linear characteristics of the power battery.First,the IFR26650-33 A lithium iron phosphate power battery is placed in a temperaturecontrollable test box,and the electrical performance tests such as available capacity test,open circuit voltage test,hybrid pulse power characteristic test(HPPC),Dynamic Stress Test(DST)and Federal Urban Driving Schedual(FUDS)test are carried out respectively through BTS20 cell test system,which are prepared for the identification of the parameters of the equivalent circuit model and SOC estimation.Second,the Thevenin model of the equivalent circuit model is selected as the power battery model based on the accuracy of estimation and the amount of calculation.First,the online parameter identification of the model is carried out using the forgetting factor recursive least squares and the identification results show that the maximum error between the simulated voltage and the real voltage does not exceed 0.02 V.Then,the Simulink equvalent circuit model and the Simscape equivalent circuit model of the power battery are built.After compasion,the Simscape model that is easy to create,understand and maintain is selected.Based on this model,the Simulink parameter estimation toolbox Parameter Estimator is used to identify parameters offline.The identification results show that the simulated voltage curve and the experimental voltage curve have a high degree of coincidence,indicating the reliability of the offline parameter estimation toolbox Parameter Estimator.Again,based on the Simscape model of the power battery,the state transition equation and observation equation of the battery model are analyzed,and the Simulink function is defined separately after discretization,the Unscented Kalman Filter module is inserted,and the parameter configuration is completed.The current signals of the HPPC test,the DST and FUDS dynamic working condition are respectively imported for simulation.Finally,the simulation results show that when the initial value of SOC has a deviation of10%,after the curve converges,the deviation between the estimated SOC value and the real value of the HPPC test,DST and FUDS dynamic conditions is basically kept within 2%.In order to verify the robustness of the Unscented Kalman Filter module,the initial value deviation of SOC is artificially increased from 10% to 20%,30% under FUDS conditions,and it is found that with the increased of the initial value deviation of SOC,the error of SOC estimation increases in the initial period,but after the error converges,the estimated error of SOC can still be maintained within 2%.