The SOC Estimation Of Polymer Lithium Ion Batteries Based On The Interactive Multimodel Extended Kalman Filter

Clean new energy has become the theme of today's society.The development of polymer lithium-ion batteries has also been greatly developed along with the trend of new energy sources.Therefore,the SOC estimation of polymer lithium-ion batteries is of great significance.Polymer lithium-ion batteries have been widely used in the market due to their high energy ratio,long cycle life,and high discharge rate.In this paper,the modeling of polymer lithium-ion batteries is discussed.After consulting the advantages and disadvantages of the electrochemical models,neural network models and equivalent circuit models used in previous batteries,the adaptive parameters based on the equivalent circuit model are selected.The battery is tested according to the test standard and test data is obtained.The parameters in the battery model were identified using the cftool tool in matlab,the modeling and simulation analysis was performed in simscape in matlab.Comparative analysis of test data and simulation results,which verifies the accuracy and reliability of the model established in this paper.The estimation of battery SOC is the core of battery management system.In this paper,the SOC of polymer lithium ion battery is studied.In this paper,the estimation of SOC is described,and its advantages and disadvantages are analyzed.Then based on the extended Kalman filter(EKF),a method of SOC estimation of polymer lithium-ion batteries based on the interactive multimodel extended Kalman filter is proposed.And using the parameter adaptive equivalent circuit model,the algorithm was simulated and analyzed.The simulation results show that the estimated error of the SOC of the polymer lithium-ion battery is within 3%,which proves that it has better estimation accuracy for the remaining battery power.Finally,the hardware and software of the battery management system are designed,and a battery management system prototype is developed.The experimental results show that the algorithm has good estimation accuracy.