Research On Online Joint Estimation Method Of Lithium-ion Power Battery State Based On GWO-SVM Algorithm

As a core component of battery management systems(BMS)in electric vehicle and energy storage applications,achieving accurate and fast battery status prediction is one of the important functions of BMS.State of charge(SOC)and state of health(SOH)are the two main indicators of battery status,which are essential criteria to ensure the security of electric vehicles and battery storage systems and to monitor the battery status.Therefore,carrying out research related to the accurate online estimation of SOC and SOH of lithium power batteries is beneficial to the safe operation of the system and has crucial academic significance and application prospects.The specific work of the research content of this paper is as follows:First,this paper offers an exhaustive review of existing methods for the estimation of battery state of charge as well as state of health.The problems of the battery SOC estimation and SOH estimation in the present stage,such as more complicated operation and long estimation duration,are analyzed to identify the influence of individual estimation methods and joint estimation methods on the battery charge state and health state estimation.On the basis of the laboratory experimental platform built in our subject group,the relevant experimental tests were carried out.Secondly,the support vector machine(SVM)algorithm is studied further,and the lithium-ion power battery experiments are carried out to analyze the coupling relationship between SOC and SOH,and to simplify the extractions of feature quantities and reduce unnecessary calculations.A method for joint estimation of lithium power batteries based on support vector machines is presented.The simulation verification shows that this method has excellent generalization performance,and the ability of calculation and accuracy of estimation have been improved in certain extent,which provides a certain foundation for the following online research.Nevertheless,this method is sensitive to the parameters,which still requires further optimizing the parameters to improve the accuracy and robustness of the estimation.Again,the joint battery state estimation method is proposed on the basis of the Gray Wolf Optimized Support Vector Machine(GWO-SVM)algorithm to overcome the problems that the parameters still need to be further optimized and the accuracy and robustness need to be improved when the SVM is jointly estimated.Through verification of working conditions and comparative tests with other algorithms,it is proved that the joint estimation method of battery state on the basis of Gray Wolf optimized support vector machine can be able to efficiently optimize the parameters and increase the accuracy and robustness of battery state estimation.Finally,the BMS of the lithium-ion power battery from hardware and software is designed to achieve the online application of joint estimation of lithium-ion power battery state basing on GWO-SVM algorithm.The experiments finally demonstrate that the platform can achieve accurate data collection and estimation of battery state online for lithium power batteries with high estimation accuracy and robustness,which has practical value.