Research On State Of Charge Estimation Method Of Battery

Battery is widely used in energy storage device of power systems and electric vehicles(EVs),as the key component,which improve the reliability of power system,and it is crucial for the power performance of EVs.In order to prevent the battery over-charge,over-discharge occurrence and keep battery in a safe,reliable,and high performance condition,it is necessary to estimate SoC of battery accurately,and has great practical significance.In view of the lithium-ion battery is widely used and is one of the most promising battery,lithium-ion battery is selected as the research object of this subject,and its estimation process of SoC and algorithm performances are studied.So,the following contents would be studied.Firstly,through the design and implementation of the related experiments of the lithium-ion battery,the characteristics of terminal voltage,resistance and capacity were analyzed.Based on the internal mechanism of the lithium-ion battery,these factors affecting the accuracy of SoC estimation were extracted,including resistance,current rate,self-discharge rate,ambient temperature and aging degree.Considering the importance of the relationship between SoC and open circuit voltage(OCV),the effects of experimental conditions(ambient temperature,current rate)and the state of the battery(aging degree,electrode material)on the SoC-OCV curve were analyzed.Secondly,the advantages and disadvantages of equivalent circuit models and the relationship between voltage and current were analyzed.Taking the Thevenin model and the two order RC model as examples,the parameters of the two models were identified by the least square method under the HPPC test conditions,and two models were built in Matlab/Simulink,their accuracy was verified.The results show that the Thevenin model and the two order RC model can accurately reflect the dynamic characteristics of the lithium-ion battery.The Thevenin model was selected as the model of the subject,since the requirement of accuracy and real-time performance.Thirdly,the principle and calculated process of the extended kalman filter(EKF)algorithm and the unscented kalman filter(UKF)algorithm were analyzed,the SoC of lithium-ion battery under constant current discharge condition was estimated using two algorithms,the estimated results were analyzed,and the feasibility of two algorithms was verified.And,the SoC of lithium-ion battery under DST condition was estimated,the accuracy and the reason of the error of two algorithms were analyzed.By analyzing the results of the two cases,which the initial SoC was inaccurate and the measured current had Gauss noise,the robustness performance and anti-interference performance of two algorithms were illustrated.The results show that compared with the EKF algorithm,the UKF algorithm has better performance in accuracy,robustness and anti-interference.