Study On Estimation Method Of Charge State Of Lithium Battery Of Electric Vehicle

In the global environmental pollution and energy crisis two major issues increasingly serious background,the development of new energy vehicles has become a consensus.As the core component and power source of electric vehicle,the related technology of power battery is the bottleneck restricting the industrialization of electric vehicle.Among them,the accurate estimation of SOC is the focus and difficulty of the automotive industry and research institutes.Accurate SOC estimation is an important basis for the sustainable driving range of electric vehicles and can guarantee the safe use of power batteries,which has very important engineering value.For this purpose,this paper has done the following work:First analysis the principle and characteristic of the ternary lithium batteries,the working condition of research object battery monomer characteristic test,comparative analysis the advantages and disadvantages of different battery model,and from two aspects of model precision and computational complexity of the comprehensive consideration,based on the battery internal working characteristic,established the second-order resistance capacity lithium battery equivalent circuit model.Then,based on the second-order RC equivalent circuit model of lithium battery,the discretized state space equation is derived,the open-circuit voltage calibration test is carried out,and the fifth order ocv-soc relationship expression is obtained by means of the mean fitting of charge and discharge.The characteristics of lithium battery terminal voltage were analyzed,and the parameters of the battery model were identified by using the exponential fitting formula to fit the off-line test terminal voltage data.To solve the model parameter changes in a complex environment results in the decrease of SOC estimation accuracy problem,was proposed based on forgetting factor recursive least squares(FRLS)online identification method,form of battery model difference equation was deduced,using online identify DST and FUDS FRLS condition model parameters,and set up the Simulink model using the working condition of data on the verification,the verification results show that the average absolute error can keep within 3.5m V and verify the accuracy and effectiveness of the identification method.After that,the principle of classical kalman filter algorithm is expounded,and the utsvd transform based on singular value decomposition is proposed to solve the problem of non-positive determination of covariance matrix in the recursive process of SOC estimation of unscented kalman filtering(UKF),and the specific steps of realizing SOC estimation of UKF are expounded by combining the state space equation and parameters of the battery model.For electric car lithium-ion battery initial SOC error which resulted in increased SOC estimation error problems,memory no trace kalman filter is presented based on the attenuation of SOC estimation method,considering the noise statistical characteristics of the unknown and time-varying characteristics,the Sage-Husa adaptive filter theory is introduced to design consists of the SOC estimation method based on adaptive memory attenuated unscented kalman filtering(AMAUKF).Finally,by combining online identification of parameters with state,a joint online estimation method of model parameters and SOC based on FRLS and kalman filter was proposed,which realized real-time online estimation of model parameters and charged states,and built a Matlab/Simulink simulation model.From the initial SOC disturbance two error sources of error and working condition,design of four different test conditions on the joint estimation algorithm is verified,the results show that FRLS&AMAUKF joint estimation algorithm under the condition of 30% of the initial SOC error convergence time remained within the 80 s,SOC they can maintain absolute error and mean absolute error value within 1.5%,showed a faster convergence rate and high estimation precision;When the battery current condition changes,the FRLS&AMAUKF joint estimation algorithm shows strong robustness.Considering the complexity of operating conditions and environment of electric vehicles,a protection strategy of piecewise SOC estimation based on fusion algorithm is designed.