Research On Parameter Identification And State Of Charge Estimation Of Lithium-Ion Battery For Electric Vehicles

With the increasingly serious problems of energy shortage and environmental pollution,new energy vehicles,especially electric vehicles,have ushered in unprecedented development opportunities.The battery management system is one of the core technologies of electric vehicles,and accurate battery SOC estimation is very important for it.This paper takes the 18650 ternary lithium-ion battery as the research object,in order to improve the estimation accuracy of battery SOC,the following researches are carried out:Firstly,the working principle and key performance parameters of lithium-ion battery are analyzed,and the dynamic characteristics of the battery are analyzed through the test experiments of voltage,capacity and internal resistance.After comparative analysis of four commonly used equivalent circuit models,the second-order RC model is selected as the equivalent circuit model of lithium-ion battery used in this paper.The HPPC test of the battery are carried out,and the model parameters are obtained by offline identification using the exponential fitting method.The effectiveness of the established second-order RC model is verified in MATLAB/Simulink.Secondly,in order to solve the problem of low identification accuracy of traditional least squares method in uncertain noise environment,the bias compensation recursive least squares method with forgetting factor is proposed for online parameter identification,and the identification accuracy is verified by taking the battery terminal voltage as the observation value.On the basis of parameter identification,extended Kalman filter(EKF),unscented Kalman filter(UKF)and cubature Kalman filter(CKF)are used to estimate SOC respectively,and the accuracy of the three algorithms is compared.The results show that the estimation accuracy based on CKF algorithm is the best.In view of the problem that the system noise of CKF algorithm cannot be adaptively updated and the covariance matrix is easy to lose the positive definiteness,which may lead to the algorithm suspension,the Sage-Husa adaptive filtering algorithm and singular value decomposition method are introduced.The adaptive cubature Kalman filtering algorithm based on singular value decomposition is used to estimate the battery SOC,and the estimation accuracy is improved.Finally,the bias compensation recursive least square method with forgetting factor and the adaptive cubature Kalman filter algorithm based on singular value decomposition are combined to realize the joint estimation of battery model parameters and SOC,which further improves the estimation accuracy.Based on the existing battery test platform,the hardware verification platform of the algorithm is built,and the hardware and software of the battery voltage and current acquisition system are designed.The experimental results show that the hardware verification results of the proposed algorithm are consistent with the simulation results,which can realize the accurate estimation of SOC and prove the feasibility and accuracy in embedded devices.