Research On On-line Estimation Method Of Lithium Battery SOC Based On Combination Model

Battery management system(BMS)is one of the important components of new energy vehicles,and its research has become a hot topic in recent years.State of charge(SOC)is an important indicator in BMS and has an important significance for its accurate estimation.However,SOC is a state variable inside the battery,which cannot be directly measured,so it needs to be estimated by a certain nonlinear relationship through various indirect methods.To this end,this dissertation first analyzes the performance of the battery and compares the various models of the battery.Common battery models usually only consider the basic characteristics of the battery,ignoring the hysteresis effect between the open circuit voltage and the SOC.In this dissertation,a new model is obtained by combining the Preisach discrete model and the first-order RC model,and the corresponding state space equation is established.Then,the various SOC estimation methods are compared and analyzed,and the accuracy obtained by the model-based estimation method can be obtained more accurately.In this paper,four common algorithms,Extended Kalman Filter(EKF),Unscented Kalman Filter(UKF),Particle Filter(PF)and PI observation,are studied in detail.The UKF is the SOC estimation method.In addition,the advantages and disadvantages of recursive least squares and EKF for parameter identification are discussed.Finally,the EKF algorithm is selected as the parameter identification method.In addition,the double filtering algorithm of Frisch scheme is proposed in this dissertation.The basic principle of noise estimation by Frisch scheme is introduced in the dissertation.Then the implementation process of the double filtering algorithm of Frisch scheme is introduced.In this dissertation,the Frisch scheme is used to estimate the charge and discharge current and terminal voltage of the battery,and then filter out.Then,UKF combined with EKF is used for parameter identification and SOC online joint estimation.Finally,in order to verify the effectiveness of the proposed algorithm,the experimental platform is used to verify the algorithm.Under the same experimental conditions,the EKF-UKF algorithm before denoising and the EKF-UKF algorithm after denoising and the RLS-UKF algorithm before denoising and the RLS-UKF algorithm after denoising were tested separately.The experimental results show that the accuracy of the parameter identification and SOC online joint estimation using EKF-UKF after noise processing by Frisch scheme is the highest,and the algorithm is robust.