| The passed years witnessed the rapid development of new energy vehicle,especially battery electric vehicles(BEVs),and the researches on technologies of motor,electronic control and power battery have made great progress.It is commonly recognized that battery management system(BMS)plays an essentional role in the development of BEVs,which is responsible to monitor and manage power batteries,and provides guarantee for vehicle’s endurance,service life and safety.However,one of the prominant problem is that BMS might facing the inaccuracy of the state of charge(SOC)estimation.In order to improve the accuracy of battery SOC estimation,more emphasises should be put on the simulation model of battery and SOC estimation algorithm.The objective of this study is to improve accuracy of SOC estimation for the power battery,which applied in FSEC,based on building battery simulation model and SOC estimation algorithms.The main work of this study is described brieftly as follows.Firstly,model the simulation of single battery and study the parameter identification of it.By analyzing the primary models of single battery,select the second-order RC equivalent circuit model to establish the simulation model for single battery.Meanwhile,the battery parameters test platform was builted to test the capacity,charge-discharge efficiency,open-circuit voltage and SOC curve,internal resistance and capacitance of single battery.Consequently,the battery parameters were identified according to the test data.Simulink software is applied to establish the simulation model of single battery.Compared the simulation results with the experimental values,the maximum error between them is 0.05 V,and the relative error is less than 0.23%.Secondly,the simulation model construction of battery pack is studied.The battery pack models were established by methods of monomer accumulation,the whole modeling and the minimum electromotive force,respectively.Compared the simulation results with the experimental values of these three methods,it is concluded that the minimum voltage battery modeling method got the highest accuracy,but it is quite time-consuming.In this dissertation,a hybrid method combining the minimum voltage modeling approach with the random sampling parameter identification was put forward to construct simulation model for battery pack and was implemented by use of Simulink software.Compared with the experimental values and the simulation results,the error of battery terminal voltage is less than 0.5%,and the single battery terminal voltage error is less than 0.7%.Furthermore,the method of SOC estimation algorithm of the battery pack is studied.Compared with the advantages and disadvantages of the traditional sliding mode algorithm and the extended Kalman algorithm,this dissertation attempt to propose an extended Kalmanadaptive sliding mode algorithm to estimate the SOC of battery pack,which solved the problems of the traditional sliding mode algorithm and the extended Kalman algorithm.Compared the experimental values with the simulation results,the error of the traditional sliding mode algorithm is less than 7%,the error of the extended Kalman algorithm is less than 3%,and the error of the extended Kalman-adaptive sliding mode algorithm is less than 2%.Finally,building the experimental platform to estimate the SOC of battery pack.Testing the battery pack,which is in NEDC operation condition.Compared the experimental results with the simulation results,which shows that the SOC error estimated by extended Kalmanadaptive sliding mode algorithm is less than 2%,which proves the effectiveness of the proposed algorithm. |
PDF Full Text Request