Study The Effect Of Charge Behavior On The Soh Of Electic Vehicle Battery

With the development and popularization of traditional fuel vehicles,vehicle exhaust emissions bring more and more pressure to the environment,but also face the problem of oil resource depletion.In order to cope with the problem of environmental protection and resource shortage,each country is actively promoting and guiding the development and production of new energy vehicles,pure electric vehicles because of its low system integration,environmental friendliness,energy conservation,low cost of use and other advantages of more and more users and car companies.As the future development direction of the automobile industry,the power of pure electric vehicles mainly depends on lithiumion power batteries.Lithium-ion battery has the advantages of high working voltage and high energy density compared to lead-acid battery,but lithium-ion power battery as one of the "three-electric" electric vehicles is relatively fragile,and the rapid decay of battery capacity leads to the reduction of vehicle mileage is still an urgent problem to be solved.Based on the above background,on the one hand,the problem of insufficient mileage of electric vehicles can be solved by developing lithium battery energy storage technology,such as the application of oxide or hydroxide cladding to the electrodes of the battery can improve the multipleperformance and cycle performance of the battery.On the other hand,in the current stage of battery technology without breakthrough progress,through the actual operation of battery work to study and analyze,mining bad charging behavior to slow down the rate of battery capacity decay,improve battery life is also very important,At the same time,this is also conducive to reducing the cost of electric vehicles and solving the user's mileage anxiety problem.The main research content of this article is as follows:(1)First of all,the original data is pre-processed and statistics of each charge information of the vehicle,based on density estimation method to quantify the description of charging behavior parameters and the time integration method to calculate the whole package capacity characterization battery health status,and the charging capacity linear regression to obtain the battery package capacity attenuation rate.(2)Secondly,from the perspective of statistical analysis,the correlation between charging behavior parameters and capacity attenuation rate is investigated,and the correlation coefficient is verified.Based on GBDT integrated learning algorithm,the regression rate of capacity attenuation is analyzed,the contribution rate of each charging behavior parameter to the model is analyzed from the perspective of the model evaluation,and the close relationship between each charging influence reverberating and the attenuation rate of battery capacity is explored by the predictive model.(3)Finally,the charging behavior is reduced to the dimensional processing,the behavior is clustered based on the machine learning K-Means cluster algorithm,and the cluster number is optimized according to the model evaluation index.According to the clustering results,the rate of decay of battery capacity between different categories is compared with the results of which show that the decay speed difference between different categories is obvious.The primary factor of capacity attenuation rate is current,followed by temperature,then frequency,and finally SOC.Through the above research work,the negative factors of power battery in actual use are analyzed,and it is expected to provide behavior suggestions for the application of power lithium batteries in electric vehicles,as well as the utilization of power batteries and traceability assessment in energy storage.