Research On Capacity Estimation Method For Rapid Detection Of Power Battery Package

In recent years,electric vehicles have been flourishing with the ideas of saving energy and protecting the environment.However,due to inconsistencies in the production and use of electric vehicle power batteries,the capacity decline rate of the battery cells and the battery SOC are different,which limits the performance of the battery pack,reduce capacity utilization,and shorten the battery pack service time.For electric vehicles in service,use the data of one deep charge and the capacity estimation method to obtain the available capacity of each cell in the battery pack,the SOC distribution of each cell,and the maximum available capacity of the entire battery pack.These can provide data reference for the balanced maintenance of the battery pack,improve the energy and capacity utilization of the battery pack,and extend the service life of the battery pack.Due to the high proportion of power batteries in the entire vehicle,power battery status assessment,especially rapid detection of capacity,is a problem that needs to be solved urgently in various fields such as new energy vehicle insurance value assessment,used car identification assessment and vehicle battery ladder utilization screening.Therefore,this paper aims at the integrated platform of electric vehicle charging and inspection,using the primary charging data of electric vehicles to study the capacity estimation of single cells in the battery pack.The specific contents of the study are as follows:Four ternary batteries with a rated capacity of 42 Ah disassembled from an electric automobile vehicle were used as a unit capacity test experiment and a group charge-discharge cycle experiment.The battery’s discharge and charge curves were used to estimate the low-end capacity and high-end capacity of the cell,which in turn gives the maximum usable capacity of the monomer.Increment capacity analysis is based on the battery’s charge curve and is an in-situ non-destructive analysis method.Adopting Two algorithms,wavelet multi-scale analysis and Gaussian process regression,according to the selection of the voltage interval and the selection interval of the capacity increment curve data,a total of 12 methods are selected to filter the increment capacity curve made by the cyclic data of two adjacent ternary batteries in the same batch for accurately obtaining the characteristic parameters of the increment capacity curve.Use the determined filtering method to conduct increment capacity analysis on the aging process data of a ternary and manganese oxide lithium battery,and obtain the six characteristics of the maximum peak position,peak height,right slope,left slope,peak area,and SOC at the peak of the curve.The Pearson correlation coefficient between the parameters and the maximum available capacity was analyzed by principal component regression to establish a principal component regression capacity estimation model based on different characteristic parameters.Considering that the deep charge and discharge are difficult to achieve in the test of electric vehicles and the test time is too long,and the vertical analysis of the battery cells may not be suitable for the horizontal analysis of the cells in the battery pack,finally based on the main peak of the battery incremental capacity curve to estimate the maximum available capacity of the battery based on the capacity between full charge.