Research On SOC Online Estimation And SOP Prediction Method Of Lithium-ion Battery

In electric vehicles(Evs),the State of Charge(SOC)and peak power(SOP)of the power battery are two important parameters and design basis for the battery management system(BMS)to monitor and manage Evs.The accurate estimation of SOC and SOP is of great theoretical and practical value for the batteries safety use and cycle life,as well as the optimization and matching of the vehicle energy.Traditional SOC estimation methods of power battery can not meet the real-time requirements,and the accuracy of SOP estimation is relatively low,which affects the batteries safety and proper use.This dissertation focus on lithium-ion battery,improving the estimation precision of SOC and SOP and aims to realize online estimation of them.Firstly,the research background and significance of the state estimation of lithium ion battery is clarified,then the SOC definition is discussed with comprehensive review of overseas and domestic research status of SOC and SOP estimation.Secondly,lithium-ion battery working mechanism and its main technical indicators are described.In order to better understand the charging and discharging characteristics of lithium-ion battery,a battery test platform is constructed and the key performance parameters of lithium ion battery are tested.Based on the analysis of testing datas,the firstorder RC equivalent circuit model of the battery is established under Simulink environment,then experiment is performed to verify the model accuracy.Thirdly,forgetting factor recursive least squares(FFRLS)is employed to realize the on-line parameters identification of lithium-ion battery model,then a SOC estimation method based on forgetting factor recursive least squares(FFRLS)and extended Kalman filter(EKF)is proposed,corresponding simulation experiment is carried out under different conditions,through analyzing the experimental results,some shortcomings of this algorithm are found.In order to improve these deficiencies,a novel approach based on FFRLS and double kalman filtering for SOC estimation is designed.By this novel method,online SOC estimation is achieved,as well as battery model parameters identification in real time.Simulation experiment is conducted to verify this new approach and results show that SOC estimation precision is considerably increased with better algorithm stability and robustness.Finally,a multi-parameter constraint method based on BP neural network is proposed,it combines the SOP prediction method based on battery dynamic model?SOP prediction method based on battery SOC and HPPC method,fully considering the battery voltage,current,SOC and other design constraints.The BP neural network model is established in the MATLAB environment for predicting the peak power of the battery,and the simulation results show that the SOP prediction error of this method is within 2%.