Research On Battery Management System Based On Lithium Battery

With the approaching of the fossil energy crisis and the increasingly prominent environmental problems,the storage and use of clean and sustainable new energy has become the trend of the times.Electricity is currently the hot and clean energy,and low-carbon zero emission has become its label,which coincides with the direction of the times.As the main storage carrier of electrical energy,the battery's energy storage efficiency,service life and safety have self-evident effects on electrical energy.Compared with other electrochemical energy storage batteries,lithium ion batteries have the advantages of high specific energy,high specific power,long cycle life,no memory effect,high environmental friendliness,and low storage difficulty.Design a battery management system(Battery Management System,BMS)for lithium-ion batteries,which aims to make an overall assessment of the battery's state of charge(State of Charge,SOC),state of health(State of Health,SOH),and safety performance.control.In previous studies of battery SOC estimation,the research object was often a single cell rather than a battery pack.In the estimation strategy,only the estimation accuracy was considered to be improved.Complex optimization algorithms and redundant data were often used to greatly increase the computational cost,and the accuracy was minimal.The improvement has an impact on the timeliness of the estimate,and it is difficult to promote it in practical applications.This paper considers the above factors comprehensively,and strives to improve the accuracy of battery SOC estimation while taking into account the computational complexity.The main research contents are as follows:1.Comprehensive analysis of the differences in the performance of the internal structure of lithium-ion batteries and the description of the electrochemical reaction process,establish a second-order RC equivalent circuit model of lithium batteries,and fit the open circuit voltage and SOC by HPPC experiments on the battery pack Function curve between them,from which the RC parameters in the equivalent circuit model are identified,a battery simulation model is built on the Simulink simulation platform,and the measured terminal voltage value and the estimated parameter estimation value are substituted into the model terminal voltage obtained by the equation The error between the predicted values verifies the accuracy of the model.2.Through the simulation model,the influence of the deviation between the measured value and the estimated value of the terminal voltage on the SOC estimation is analyzed.The Laplace transform method is used to differentiate the model parameters,which greatly reduces the amount of calculation in the parameter identification process.The Levenberg-Marquardt algorithm is used to identify the differentiated parameters,and then the inverse transform is used.Convert to model parameters and update the parameters in the SOC estimation algorithm to improve the estimation accuracy.3.The MATLAB / Simulink platform is used to build a SOC estimation simulation model based on the gradient boosting decision tree and extended Kalman filter joint(GBDT-EKF)algorithm.GBDT is used to fit the residuals of the prior and posterior estimates to the noise matrix of the current time.Relationship,considering the influence of temperature and capacity parameters on the accuracy of SOC estimation,compared with the EKF algorithm in the relationship between the terminal voltage and time curve,the open circuit voltage and SOC relationship curve,and the error between the SOC estimated value,and the noise matrix is used to estimate the SOC.Accuracy will have a greater impact.After iteratively updating the noise matrix,the SOC estimation accuracy is improved.By comparing the simulation results of the two algorithms,it is shown that the SOC estimation algorithm proposed in this paper is significantly improved in accuracy and can quickly converge to meet the requirements of complexity and accuracy.The effectiveness of the GBDT-EKF joint algorithm is verified.