Research On SOH Estimation Of Lithium Batteries For Electric Vehicles

With the current shortage of international oil resources,the world is facing more and more serious energy and environmental problems.The research and development of energy-saving,economical and environmentally friendly electric vehicles have become one of the important ways to solve energy and environmental problems all over the world.As the main driving force of electric vehicles,the technology development of BMS will lead to more efficient use of lithium ion batteries,which will take the performance of electric vehicles to a higher level.As one of the performance parameters of various lithium batteries that need to be tested by BMS,the accurate evaluation of SOH is not only an important basis for the calculation of key parameters such as SOC and SOP,but also a significant reference for the evaluation of when the power battery system needs to be replaced,whether it can be degraded for use and the utilization value evaluation after degradation.Firstly,the charging and discharging characteristics of lithium-ion battery are studied.Secondly,the influencing factors of performance degradation of power battery are analyzed.Through the comparative analysis of the actual maximum capacity of battery and the internal resistance of battery,the actual maximum capacity of battery is determined as the evaluation index of SOH.Thirdly,the constant current charging voltage curve of lithium battery is processed and transformed by capacity increment analysis method,and the simplified d/d(1 processing method is selected to calculate the capacity increment curve,and the median filtering method is used to filter the curve.Then,the peak intensity and peak position voltage of No.2 and No.3 capacity peaks on the curve are mined by the feature vector on the capacity increment curve As the initial eigenvector,the peak intensity and peak position voltage of No.2 peak are further determined as the feature vector of SOH estimation model through the grey correlation analysis method,and the input parameters of SOH estimation model are also determined.Then,based on the introduction of the basic principles of support vector regression and grey wolf optimizer algorithm,the differential evolution method is integrated into the grey wolf optimizer algorithm for improvement,and then a joint prediction model based on the improved grey wolf optimization algorithm and support vector regression is constructed with two feature vectors as input and lithium battery SOH as output,which is verified by simulation experiments with MATLAB.Finally,the experimental verification of estimation of SOH of lithium battery was completed in BMS.On the basis of setting up the experimental platform,a real-time mathematical model of capacity increment curve was established in BMS,two feature vectors of capacity increment curve were extracted to realize the estimation function of SOH,and the measures to improve the accuracy of charge and discharge data collection were proposed.The battery cycle aging test data were used to train the estimation model,and the relevant model parameters were written into BMS.The test verified that the feature vector extraction and the estimation function of SOH of lithium battery could be successfully carried out in BMS.By comparing the SOH estimation results with the actual results,it can be seen that in the whole cycle life cycle of lithium battery,the BMS system can achieve the accurate estimation of SOH in different aging stages of battery,and the estimation error is no more than 4%,which has good application value.