Research On The Battery Pack Consistency Based On Running Data Of Electric Vehicle

Lithium-ion batteries have been widely used in electric vehicles(EVs)for its high voltage platform,large energy density,long cycle life and high safety.However,hundreds of cells are required to be connected in series and parallel structure to meet the power requirements of EVs in practical application.As for the battery pack used in domestic electric vehicles,there are still many variations between cells because of the initial difference in the process of cell manufacture and the cumulative difference in practical usage.The inconsistency of the battery pack not only reduces the performance of the battery,but also affects the monitoring accuracy of the battery management system(BMS).In extreme cases,it can even cause battery abnormalities and accidents.Therefore,based on the battery data collected in actual driving condition,this dissertation studies the inconsistency of battery pack’s temperature,voltage and capacity in detail.The raw data is firstly preprocessed to solve the problems of inconsistent format,missing data and dead data.And according to the(dis)charging state of the EV,the raw data is divided into discharging fragments and charging fragments.The statistical analysis and correlation analysis of the source data showed that the overall operating conditions of the battery are relatively good.It can be concluded that it has a high correlation between cell voltage,and the temperature of the battery module No.1 is quite different from other modules.Then,a method based on the hierarchical clustering analysis to judge the consistency of battery pack is proposed.Analysis result also indicates that the temperature inconsistency of battery module No.1 is the largest,and the position with the worst temperature consistency is accurately located.As for the consistency of cell voltage,it has been confirmed that the voltage consistency corresponding to the position with poor temperature consistency is also poor.Taking the clustering distance as the consistency evaluation index,the trend of battery inconsistency is further researched.It has been found that the battery temperature inconsistency is quite sensitive to the ambient temperature.Then,according to the temperature inconsistency,the data is divided into four stages.Moreover,it has been proved that the voltage inconsistency of the battery pack is greatly affected by the temperature inconsistency,and the voltage inconsistency is unrecoverable,which rises in steps.Finally,this paper innovatively proposes a method to evaluate the cell capacity consistency based on the improved incremental capacity(IC)curve under charging conditions.Taking the slow-charging conditions as research object,the characteristics of battery IC curve are analyzed.The traditional solution method is improved by polynomial filtering and probability frequency statistics(PDF)to smooth the IC curves of battery pack and cell respectively.The peak height of the IC curve is extracted to characterize the cell capacity,and the cell capacity consistency is evaluated by the standard deviation of the peak height.The results show that the cell capacity consistency is also greatly affected by the temperature consistency,but it is recoverable.Finally,it is confirmed that the cell capacity is basically normally distributed,and the cell capacity is graded by 3-method.