Research On Remaining Driving Range Estimation Of Pure Electric Vehicles Based On Battery SOC And Operating Condition Identificatio

With the continuous rise of global car ownership,energy shortage,environmental pollution and other problems caused by cars are becoming increasingly serious.New energy vehicles,represented by pure electric vehicles,are more and more valued by countries.However,the remaining driving range information provided by pure electric vehicles is not accurate,which reduces consumers’ confidence in pure electric vehicles.In order to improve the estimation accuracy of remaining driving range of pure electric vehicles,the following studies were carried out:By building a power battery test platform,the characteristics of lithium ion power battery were tested and analyzed.According to the experimental data,the battery parameters were identified and the Simulink battery simulation model was established;In order to solve the problem that the traditional untracked Kalman filter(UKF)estimation of battery state of charge(SOC)had the risk of losing the positive quality of system covariance and low estimation accuracy,the adaptive singular value decomposition untracked Kalman filter(ASVD-UKF)algorithm was constructed and its effectiveness was verified;The typical working conditions were selected by hierarchical clustering algorithm,and the characteristic parameters were optimized by principal component analysis.The optimized data were input into the established BP neural network,support vector machine(SVM)and firefly algorithm optimization support vector machine(FA-SVM)for training.According to the training results,the FA-SVM condition recognition had the highest accuracy;By combining battery SOC estimation and working condition recognition,the vehicle SOC estimation simulation model,the remaining driving distance simulation model based on average energy consumption and the remaining driving distance simulation model based on working condition recognition were built in CRUISE,and the complex working conditions were established for simulation analysis.It was proved that the remaining driving range model based on driving condition recognition has certain advantages.In this paper,the ASVD-UKF algorithm was established to complete the highprecision estimation of lithium battery SOC,and on this basis,the remaining driving range estimation model based on FA-SVM condition recognition was established to achieve the accurate estimation of the remaining driving range of pure electric vehicles.