Kalman Filter Estimation Of Lithium Battery SOC Based On Model Multi-parameter Updating

As the core of electric vehicles,lithium batteries are particularly important in the context of the current development of new energy markets.The battery management system monitors and protects the lithium battery and is the basis for stable,safe and high-efficiency output of the lithium battery.At the same time,the lithium battery SOC estimation is one of the core functions of the battery management system.High-precision estimation of lithium battery SOC can effectively optimize vehicle energy management,improve lithium battery safety protection,extend lithium battery cycle life,and reduce new energy vehicle cost.Therefore,this thesis studies to improve the accuracy of lithium battery SOC estimation,relying on the "China Postdoctoral Fund Project"(2017M613034): Coordinated Control of Electric Vehicle Composite Braking Based on LHMM/SVM Braking Intention Identification,"Shaanxi Key Industry Innovation Chain(Group)Project"(2018ZDCXL-GY-05-03-01): Research on Key Technologies of Distributed Drive Pure Electric Passenger Vehicles and Research on Key Technologies of Core Equipment and Data Fusion for Electric Vehicle Inspection Line(2019ZDLGY15-01),the following aspects have been studied:Firstly,in order to further understand the dynamic characteristics of the NCM lithium battery,the capacity and OCV performance tests at different temperatures and rates,three working conditions at different temperatures and the endurance cycle test were designed.By analyzing the equivalent circuit model,the offline parameter identification method based on data fitting and the online parameter identification method based on recursive least square with forgetting factor(FRLS)are applied in the Thevenin model and DP model parameter identification.The work is carried out in Simulink environment.The test data verification model parameter identification method and model accuracy,and the better model parameter identification method and model are selected for the next research.Then,based on Kalman filtering theory,extended Kalman filter(EKF)and singular value decomposition unscented Kalman filter(SVD-UKF)are proposed to overcome the disadvantages of Kalman filter linearization and covariance matrix.At the same time,based on Thevenin model combined with FRLS algorithm,the model parameters and state joint estimation algorithm are proposed to study SOC estimation.In the Simulink environment,the accuracy of the two SOC joint estimation algorithms is verified by using three test data at different temperatures.The SOC estimation algorithm with better estimation effect is selected for the next research.Finally,based on FRLS&SVD-UKF model parameter and state joint estimation algorithm,the model parameter identification fusion algorithm and model capacity update algorithm respectively solve the shortcomings of FRLS online parameter identification results under weak excitation conditions and the model capacity remains unchanged,and the SVD-UKF algorithm based on model multi-parameter updating is proposed.In the Simulink environment,the endurance cycle test data was used to verify that the algorithm estimates the SOC accuracy of the lithium battery under the condition that the maximum usable capacity of the lithium battery is continuously attenuated.