Research On Estimation Method Of Soc For Power Battery

With the increasingly severe global environmental pollution and energy crisis,the development and utilization of new energy has become a new trend for future development.Electric vehicles are the application of new energy in the automotive industry.They are included in the seven strategic new industries in China with their high performance and low pollution advantages.As the core component of electric vehicles,batteries need a battery management system(BMS)to effectively manage them to improve the efficiency and stability of the entire vehicle system.The main functions of BMS include battery status analysis,data acquisition,energy control management,safety protection,balance control,and battery information management.Among them,the state of charge(SOC)prediction included in the battery state analysis is the focus and difficulty of the BMS.This function is of great significance for the control of the entire battery state and the prediction of vehicle mileage.From the perspective of practicability,this paper optimizes the accuracy and efficiency of SOC forecasting.The main research contents are as follows:(1)First,the development background of new energy vehicles is explained,the current research status of new energy vehicles is reviewed,and the current common battery models are summarized.The current mainstream SOC prediction algorithms are analyzed,and their respective applicable conditions,advantages and disadvantages are summarized.Secondly,several important types of power batteries are introduced,including their composition and working principles,and the University of Maryland CALCE Battery widely recognized in the battery field is selected.The data of the A123 battery provided by the Group is used as an experimental sample.Finally,starting from the working principle of the power battery,the battery performance parameters are explained,the definition of the battery SOC is explained,and the factors such as temperature,current and voltage are analyzed to predict the power battery SOC Impact.(2)The working principle of BP(Back-Propagation neural network)and its application in SOC are introduced in detail,and the Mind Evolutionary Algorithm(MEA)is introduced.Aiming at the problems of random distribution and repeated searching of MEA generated individuals,a Hybrid Mind Evolutionary Algorithm(HMEA)was proposed,which was used to optimize the initial weights and thresholds of BP neural networks.(3)The optimized BP neural network still has other problems such as high calculation cost and large error fluctuation in the prediction of SOC.The traditional Ampere-hour(Ah)integral method has high prediction efficiency,but it cannot eliminate the accumulated error.This paper chooses to combine the BP neural network with the Ah integral method,and proposes to improve the Ah integral method to reduce the computational cost of the neural network method,while reducing the prediction error of the Ah integral method.The simulation test using MATLAB shows that the improved accuracy of the Ah integral method is higher than that of the standard Ah integral method,which overcomes the shortcomings of the neural network method,such as large prediction error fluctuations and high calculation cost of the new intelligent algorithm,and has high practical value.There are 19 figures,4 tables,and 58 references.