SOC Estimation Of Lithium Battery For Electric Vehicle

Environmental issues have become increasingly prominent,and new energy generation technology has become a global hotspot.The rapid development of new energy generation technology makes the price of electricity cheaper,but also promotes the development of electric vehicle technology.Power battery,the energy core of electric vehicle,is the key to the performance of electric vehicle.The state of charge(SOC)of the battery directly affects the driver's judgment on the cruising range of the electric vehicle.The accurate SOC has a direct impact on improving the user experience of the electric vehicle,the rationality of the vehicle control,and the reliable operation of the battery management system.So SOC estimation,as the key and difficult point of electric vehicle development,has important research significance.In this thesis,the charge and discharge characteristics,capacity characteristics,hysteresis effect,resistance characteristics and polarization characteristics of lithium-ion batteries are analyzed in detail through design experiments.Lithium-ion batteries are complex and highly nonlinear systems.Simple first-order models cannot accurately describe battery characteristics.After comprehensive analysis of common battery models,a second-order equivalent circuit model considering various battery characteristics is established.The improved open-circuit voltage model not only ensures the accuracy but also avoids the tedious high-order fitting formula.At the same time,the battery model parameters are identified based on the Hybrid Pulse Power Characterization Test(HPPC)data.Finally,the battery simulation model was built in MATLAB/Simulink,and the feasibility and accuracy of the built battery model were verified by dynamic stress test(DST).Aiming at the shortcomings of Extended Kalman Filter(EKF)algorithm,an SOC estimation algorithm based on Adaptive Cubature Kalman Filter(ACKF)is selected.Based on the third-order spherical-radial volume criterion,the algorithm avoids Taylor series expansion calculation,effectively avoids the influence of non-linear errors,and has high accuracy.Aiming at the unknown and time-varying measurement noise of non-linear systems,an Kalman filter algorithm based on innovation is applied to update the covariance of measurement noise in real time.Based on BMS test platform,the feasibility of the algorithm is verified by experiments.The results of data comparison and analysis show that: Adaptive Cubature Kalman Filter algorithm shows good performance in accuracy and convergence.The average error can be controlled within 2%,and the initial error can converge to the true value within 60 seconds when the initial error is 10%.