Research On Online Estimation Algorithm Of Remaining Power Of Electric Vehicle Power Battery Based On Data

As countries attach more importance to environmental protection,technological progress and energy security,the application of large-scale internal combustion engines that consume fossil energy in highway transportation is gradually being replaced by various types of power systems using other energy sources.The new energy automobile industry is welcoming opportunities for development.However,current electric vehicles still have problems such as short driving range,high initial cost and poor safety.The power battery system technology has become the main bottleneck for the development and industrialization of electric vehicles.The calculation of the state of charge(SOC)is an important technology in the battery management system.The main task of this paper is to study SOC estimation.The specific research work of this thesis includes:(1)Aiming at the contradiction between the accurate high dependence of the battery mathematical model and the difficulty in accurately obtaining the dynamic battery model in the model-based SOC estimation algorithm,a fully data-driven EKF-based power battery life cycle SOC is proposed.The estimation method realizes the SOC estimation of the life cycle of the power battery.(2)According to the EKF algorithm,the estimation error caused by the high-order Taylor term of the nonlinear system is neglected and the noise requirement is too high.The particle filter algorithm is used for SOC estimation,which effectively enhances the noise suppression capability and estimation accuracy of SOC estimation.At the same time,the least-squares method with forgetting factor is used to carry out on-line realtime identification of model parameters,realizing the dynamic online update of the life cycle of the power battery SOC.(3)Aiming at the accuracy,real-time and robustness requirements of power battery parameter identification,an intelligent online identification method based on improved GA algorithm is proposed.At the same time,for the problem that the particle diversity in the PF algorithm is easy to be sure,the UPF algorithm is used for SOC estimation,and the UT transform is used to generate sigma points of the particles,which expands the search space and diversity of the particles,and effectively improves the estimation accuracy of the SOC.(4)Aiming at the problem that the traditional unscented Kalman filter sampling space may cause SOC estimation accuracy and stability degradation due to the fixed sampling space,an adaptive unscented Kalman filter algorithm capable of selfadjusting feasible solution space search is proposed.Accurate acquisition of sampling points is achieved,which improves the estimation accuracy of SOC.At the same time,in order to improve the operational efficiency of SOC estimation,based on the improved GA algorithm,an intelligent online parameter identification method based on collaborative optimization mechanism is proposed to improve the efficiency of battery model parameter identification while improving its operating efficiency and reducing its efficiency.time complexity.Based on the above proposed intelligent online parameter identification method and AUKF algorithm joint estimation,the accuracy,robustness and stability of the power battery SOC estimation are improved,and the time complexity is also reduced.