Research On Charge Control Of China Standard EMU Battery

As the energy storage device for rail trains,battery is widely used in emergency power supply and extended power supply of Electric Multiple Units(EMUs).The EMU battery has such problems as high costs,complicated replacement procedures and strict safety requirements.Therefore,it is of great research significance to control the battery charging process so as to monitor the battery working status,improve battery life,and ensure charging reliability.This thesis has a study on battery charging control of EMU from three aspects including the online estimation method of battery state of charge(SOC),the battery charging strategy and the control method of EMU battery charger.The main research contents of this study are as follows:This thesis takes the lithium titanate battery of the China standard EMU as the research object,selects the second-order RC equivalent circuit to build the battery model,and studies the two online parameter identification methods,namely,the identification accuracy of forgetting factor recursive least squares(FFRLS)and the extended Kalman filtering(EKF).This thesis compares the accuracy of parameter identification of the two methods through the battery working condition experiment.The results shows that when the EKF method is used for parameter identification,the error in predicting the battery voltage is less than 2%,which means the parameter identification accuracy is higher.Therefore,the EKF parameter identification method is selected for subsequent study.Furthermore,aiming at the problem of inaccurate estimation of battery SOC caused by fixed noise variance of EKF method,the adaptive extended Kalman filter(AEKF)is proposed to estimate battery SOC.Combined with EKF parameter identification method,the online real-time evaluation of model parameters and battery SOC status are completed.After that,the SOC estimation results of the AEKF method and the EKF method under the battery test working condition and the China standard EMU battery emergency working condition are compared respectively.The results shows that the SOC estimation method based on AEKF method can reduce the SOC estimation error to less than 2.5% under different working conditions,and can achieve rapid convergence of battery SOC estimation.Then,this thesis has a study of charging strategy of lithium titanate battery and builds a battery polarization voltage model based on the analysis of the battery internal polarization effect principle,A segmented constant current-constant voltage charging strategy based on battery SOC is proposed and the thesis uses genetic algorithm to optimize the current and voltage of the segmented charging method which takes average polarization voltage and charging time of the battery as the objective function.Compared the simulation results,it can be seen that the average polarization voltage of the new charging strategy is lower.Lastly,this thesis studies the working principle and control method of the China standard EMU battery charger and builds a simulation model.After verifying the feasibility of the model,fuzzy control and genetic algorithm are introduced to improve the PID control method of the battery charger.Simulation analysis shows that the output fluctuation of the optimized battery charger is below 2.5% under different working conditions,and the response time can be less than 0.1s,which has better output performance.