Research On State Estimation And Sorting Of Retired Power Lithium Batteries

With the substantial increase of the electromobile,the retired power batteries have also been increased every year.Therefore,the reasonable sorting of retired power lithium battery can not only effectively apply retired power lithium battery to different scenarios,but also solve the problem of resource utilization and reduce the recovery cost.However,compared with the new power battery,the performance of the retired power battery will be reduced because of multiple charge-discharge cycles.Therefore,it is of great significance to study the performance estimation and sorting method of retired power lithium battery.This thesis will focus on the research of retired 18650 ternary lithium power battery.Firstly,the electrical parameters that affect the performance of the retired ternary lithium power battery are described in detail.The data of the retired lithium power battery is measured by the electronic load instrument and other experimental equipment to analyze the changes.The performance of retired ternary lithium power battery can be characterized by the fuzzy inference machine,and the performance is estimated to obtain the membership value of the performance state,which selects the single battery with poor electrical parameters in the retired lithium power battery.Secondly,in view of the differences that there will be have the inconsistent problem due to regrouping of the single retired power lithium batteries.According to the three parameters that affect the consistency of battery pack,k-means clustering algorithm optimized by genetic algorithm is used to regroup the retired power lithium batteries in thesis,and the remaining retired power lithium batteries after preliminary screening are uniformly sorted.Finally,on the basis of the consistency group in the retired power lithium battery packs,Vmin electrochemical characteristic model of the retired power lithium battery packs is built to estimate the SOC by using the method of no trace kalman filter under conditions of setting the single battery with the lowest capacity in the retired power lithium battery pack,which can obtain new performance status.The results show that the retired power batteries with poor performance can be preliminarily selected by using the method of fuzzy inference engine,and the performance of retired batteries can be characterized by membership value.The retired power batteries are regrouped by using the optimized k-means algorithm,which the consistency difference of the retired power batteries will be effectively reduced.The SOC of the battery packs can be accurately estimated by the unscented Kalman filter method to provide the reference for the sorting of retired power batteries.