State-of-charge Estimation Of Li-ion Battery Based On Physical-data Fusion Model

In recent years,as a new generation of environmentally friendly energy storage battery,li-ion battery has been widely developed and applied in various fields.Both now and in the future,the application and development of li-ion battery always been a hot issue that the academic community has paid close attention to.Among them,the question of how to improve the accuracy of online estimation of the state of charge(SOC)of li-ion battery has been the focus of the research field of li-ion battery,and it is also a difficult problem.This thesis first introduce the working principle and characteristics of li-ion battery in detail,compare with commonly power battery,explain the advantages of li-ion battery as power battery,and for the problem of online estimation of the SOC of li-ion battery,this thesis analyzes and compares the present methods of estimating the SOC at home and abroad in the detail,and analyze the advantages and disadvantages of various methods.Subsequently,this thesis utilizes the li-ion battery model included in MATLAB \ Simulink to simulate the discharge test of the li-ion battery,and get two sets of different voltage,current,and SOC data of li-ion battery,which provide data for later methods to improve the accuracy of SOC estimation.This thesis focuses on the following studies to improve the estimation accuracy of the SOC of li-ion batterie.First,in order to improve the accuracy of online estimation of the SOC of li-ion batterie,the ampere-time integration method has been studied in this thesis.Aiming at the problem that the estimation error of ampere-time integration keeps accumulating with time,this thesis uses the Extreme Learning Machine algorithm in Machine Learning to establish an prediction model for estimation error for the ampere-time integration method,which took the battery current as input data and the estimated error value of the ampere-hour integration method as the output data.The output of the error prediction model is used as a correction term for the estimation result of the ampere-time integration method,and the error-correction is performed on the estimation result of the ampere-time integration method.Then,the amperetime integration method with error correction has been established.The simulation results show that compare with the traditional ampere-hour integration method,the ampere-hour integration method with an error correction term can effectively reduce the estimation error of the SOC of the li-ion battery and overcomes the problem that the SOC estimation error of the ampere-hour integration method increases with time.Secondly,in order to take into account the influence of battery voltage on the SOC estimation accuracy and make the estimation results more accurate,the equivalent circuit model method is further studied in this thesis.In view of the high dependence of the equivalent circuit model method on the battery model,and the influence of the errors generated by the voltage and current sensors during the measurement on the estimated state of charge,the parameter identification of the li-ion battery equivalent circuit model and the improvement of the Extended Kalman Filter algorithm are established,and a combined estimation method of recursive least squares and improved Extended Kalman Filter algorithm is established in this thesis.Finally,this thesis uses the voltage and current of the li-ion battery and the estimation error offline training error prediction model obtained by the joint estimation method to modify the estimation result of the equivalent circuit model,and a method for estimating the state of charge of li-ion battery based on a physical-data fusion model is established.The simulation result through MATLAB can prove that the improved Extended Kalman Filter algorithm effectively improves the estimation accuracy of the algorithm,and the method of on-line estimation of the SOC of the li-ion battery combined with the physical model and the data model reduces the estimation errors introduced by the voltage and current measurement.On the basis of ensuring the estimation accuracy,the physical-data fusion model overcome the computational difficulties caused by the equivalent circuit model,the on-line estimation accuracy of the li-ion battery SOC has been further improved.