| New energy electric vehicles(EVs)have achieved rapid development recently because it unique advantages like no emissions,low noise,high energy utilization rate and renewable energy,and are the mainstream of the future develop of the automotive industry.Battery Management System(BMS),as a key technology of new energy electric vehicles,plays up a important role in online monitoring of the working conditions of power batteries under complex conditions,as well as the control and management of batteries.Real-time prediction of state of charge(SOC)and state of health(SOH)of power batteries is a key technology in BMS.At present,the mainstream SOC prediction methods have prediction according to the basic parameters of the battery,data-driven algorithms,and model-based prediction,which have the advantages of simplicity,good training effect and strong stability,but have problems like big error,low accurate,and high requirements for hardware equipment.Precise prediction of the SOH value of the power battery is the key to the smooth and security operation of electric vehicles,and it is also the main basis for replacing the battery of electric vehicles,which will also affect the prediction accuracy of battery SOC.With the rapid development of electric vehicles,the prediction accuracy of SOC and SOH in BMS is increasingly high.Therefore,on the basis of analyzing and summarizing the prediction way of power battery status domestic and abroad,this paper conducts special research on power battery SOC and SOH real-time prediction methods.Firstly,the characteristics of power batteries are analyzed,the kinds of batteries and their inner structure and working theory are introduced,and the voltage characteristics of power batteries,the related between open circuit voltage(OCV)and SOC,the polarized phenomenon of power batteries,the capacity characteristics of power batteries,the internal resistance characteristics of power batteries,and the definition of power battery SOC and SOH are described in detail.Secondly,the set up of the equivalence model of the power battery is researched,and the second-order RC equivalence circuit model with high accuracy and moderate calculation quantity,and meantime can better reflect the static and dynamic characteristics of the battery,is chose as the battery model in the paper.Because the complex operation conditions of the power battery,if the traditional offline method is used to identify the model parameters,the matching of parameters and models cannot be achieved,resulting in big errors.Therefore,the paper introduces Forgetting Factor Recursive Least Squares(FFRLS)to identify model parameters online,which solution the problem of big error result by mismatch between parameters and models,and the algorithm has high accuracy in identifying model parameters.Then,the method of SOC prediction of power battery is studied.For the traditional Extended Kalman Filter(EKF)algorithm to perform linear transformation of the system linearization error and Jacobian matrix is difficult to calculate,at the same time,the Unscented Kalman Filter(UKF)algorithm can’t guarantee the non-negative of the deviation covariance matrix when filtering,resulting in the issue of filter divergent.Therefore,the paper improves and optimizes the UKF algorithm to obtain the Square Root Unscented Kalman Filter(SR-UKF)algorithm to estimation the SOC of power batteries.Finally,according to the Matlab simulation software,constant current and Dynamic Stress Test(DST)conditions are used to validate the accuracy of SR-UKF algorithm in predicting SOC,and the results indicate which compared with the usual EKF algorithm and UKF algorithm,SR-UKF algorithm can achieve better results in predicting SOC.Finally,the method of SOH prediction of power battery is studied.After analyzing the battery SOH characterization parameters,the ohmic internal resistance is used as the parameter to evaluate the battery SOH.Based on the established second-order RC equivalent circuit model combined with SR-UKF algorithm,real-time prediction of ohmic internal resistance is realized,and battery SOH is predicted from the perspective of internal resistance.Finally,simulation experiments are carried out under two different discharge conditions to verify the accuracy of the algorithm in predicting SOH. |
