Research On Modeling And State Of Estimation Of Lithium-ion Power Battery

Lithium-ion power battery is one of the core EIC system of Blade Electric Vehicle,that can provide enough energy for Blade electric vehicle.In recent years,the accidents of Blade electric vehicles are have occurred frequently in the domestic and overseas,part of them are lead to lithium-ion power battery.A good battery management system(BMS)can guarantee the safety and reliability of Blade electric vehicles,and state estimation of lithium-ion power battery is one of the most core technology of BMS.Firstly,advantages and disadvantages of Lithium-ion power battery are introduced.The characteristics performance tests,structural and working principle of lithium-ion power batteries are introduced in the domestic and overseas.Analyzed the effects of temperature on battery capacity,Open circuit voltage(OCV)and AC impedance.Analyzed the performance degradation characteristics of lithium ion batteries.Secondly,the common equivalent circuit model(ECM)and Pseudo two-dimensional(P2D)model of Lithium-ion power battery are introduced.Based on the simplified electrochemical model,the nonlinear relationship of OCV-SOC was fitted.The Recursive least square(RLS)algorithm with forgetting factor is used to identify the battery model parameters,established lithium-ion power battery model.Thirdly,introduced common methods of state of charge(SOC)estimation and analyzed the difficulties of SOC estimation.The advantages and disadvantages of kalman filtering(KF)algorithm are expounded.Considering the complexity of lithium-ion battery operating conditions,a simulation model of SOC estimation was built based on MATLAB.The estimation of battery SOC is achieved,and the on-line estimation of lithium-ion battery SOC is achieved at different temperatures,with sufficient accuracy and robustness.Finally,the SOH estimation of lithium-ion power battery was studied.The factors affecting the capacity attenuation of lithium-ion power battery are analyzed.Aiming at the difficulty of SOH estimation,a neural network-based SOH estimation method is proposed.The SOH estimation accuracy of three neural network algorithms is compared.Considering the limitation of the above methods in practical application,an improved incremental capacity method is proposed.The accuracy of the algorithm is verified based on six groups of aging data of lithium-ion power batteries.The system modeling,SOC and SOH estimation of lithium-ion power battery at different temperatures were studied,which improved the accuracy and robustness of SOC estimation in BMS.The average error of SOC was less than 0.9%,and the convergence time of the algorithm was less than 100 s.The accuracy and application range of SOH estimation are improved,and the average absolute percentage error is less than 1.4%.