Online Parameter Identification And State Estimation Of Lithium Ion Battery Model For Electric Vehicles

Lithium-ion batteries are used in electric vehicles due to their high reliability,large battery capacity and long service life.The state of charge(SOC)and health status(SOH)of lithium-ion batteries are two important indicators of battery management system,and their accurate estimation is an important prerequisite to ensure the safe and efficient operation of electric vehicles.At present,the commonly used SOC estimation algorithm for lithium-ion batteries is the extended Kalman Filter(EKF)algorithm.Under complex battery operating conditions,it ignores the time-varying characteristics of model parameters and only uses the state data at the last moment as the prediction error,resulting in low SOC estimation accuracy.Aiming at the above problems,this paper combines the multi-information adaptive extended Kalman filter(MIAEKF)algorithm with the adaptive forgetting factor recursive least squares(AFFRLS)method,and proposes the MIAEKF-AFFRLS joint algorithm to update battery model parameters and SOC alternately.This paper takes 18650 ternary lithium ion battery as the research object.Firstly,the experimental platform of lithium-ion battery was built to conduct characteristic analysis and working condition experiment of lithium-ion battery,which laid the data foundation for the following battery model parameter identification and battery state estimation.Secondly,aiming at the problem that the off-line parameter identification algorithm is difficult to adapt to the complex operating conditions of batteries,this paper proposes AFFRLS algorithm to identify the model parameters online.The algorithm can adaptively adjust the forgetting factor according to the actual operating conditions,and identify the model parameters in real time.Moreover,the terminal voltage prediction ability of the offline algorithm and the AFFRLS online algorithm is compared and verified under DST and FUDS conditions.Thirdly,aiming at the problem that the EKF algorithm uses fixed system noise to cause low SOC estimation accuracy,this paper introduces the adaptive filtering algorithm to update the noise covariance in real time in the recursion process of EKF algorithm.Aiming at the problem that the EKF algorithm only uses single new prediction error in measurement update,which easily leads to the loss of measurement correction information.In this paper,the single information prediction error is extended to measure and update the vector matrix of multi-information prediction error,and the joint MIAEKF-AFFRLS algorithm is proposed to realize the alternating update of model parameters and SOC,and SOC estimation experiments are carried out under DST and FUDS conditions.The accuracy of MIAEKF-AFFRLS algorithm is verified by comparing the SOC errors of each algorithm.Finally,aiming at the problem of increasing SOC estimation error caused by the change of maximum available capacity,a joint estimation algorithm of SOC and SOH was proposed to characterize the battery SOH from the perspective of internal resistance.By establishing the relationship between battery capacity decay and ohmic internal resistance,the current maximum available capacity was obtained to realize the alternating update of maximum available capacity and SOC.The verification results of DST and FUDS conditions show that the MIAEKF-AFFRLS algorithm with SOH correction has higher SOC estimation accuracy.Figure [58] table [9] reference [82]...