The Research And Implementation Of Battery Management Unit And Its SOC Estimation Algorithm

The environmental pollution and energy crisis brought by traditional fuel cars has become a difficult problem faced by all the countries in the world.Electric vehicle has been paid more and more attention at home and abroad because of its advantages of clean without any pollution,stable with low noise and so on.But the accurate model of battery,the high-precision estimation of SOC and the efficient management of battery are the key and difficult points in BMS.Therefore,the research of BMS related technology is of great significance.Based on actual project needs and related literature at home and abroad,the BMS is divided into battery management unit,data acquisition unit and active equalization unit according to the scheme of data decentralization and processing centralization.The battery management unit and the SOC estimation algorithm has been emphatically analyzed in this paper.The main works of this paper are as follows.First,by analyzing and comparing the advantages and disadvantages of the common battery model,the two order RC equivalent battery model is selected and the HPPC cycle charging and discharging test is used.The parameters of the data are fitted by the MATLAB/Curve Fitting tool to identify the model parameters.The battery model is built in MATLAB/Simulink,and the terminal voltage of the battery is simulated by fitting data.The result shows that the maximum error is not more than 0.015 V,therefore the model can accurately reflect the working characteristics of the battery.Secondly,by analyzing and comparing the advantages and disadvantages of the common battery SOC estimation algorithm,taking the two level RC equivalent circuit as the model,the EKF and CDKF algorithm are used to estimate the battery model respectively.Although the CDKF algorithm solves the problem that the nonlinear equations in EKF algorithm are only accurate to the first order expansion of Taylor and the computation of complex Jacobi matrix,the error of updating the observed information is large.Therefore,combined with iterative filtering theory,this paper proposes an improved LM-ICDKF algorithm,using Levenberg-Marquardt algorithm to optimize observation iteration process,so as to improve algorithm accuracy and algorithm stability.Thirdly,according to the main function of the battery management unit,the hardware design of each function module is completed.It mainly includes the MCU module,the power module,the CAN communication module,the total current detection module,the total voltage detection module and the data storage module.Among them,MCU is a Freescale series of MPC5604 BMLL chips.In the software section,the system master program and functional subroutine are designed using the C language.Finally,the test platform of the battery management unit is built,and all the function modules are tested.The accuracy of the battery total voltage and total current acquisition are tested,the test data indicate that the total voltage and total current acquisition accuracy is less than 0.5%.And then the battery SOC estimation accuracy is tested,the result shows that the SOC estimation error is less than 5%.