基于随机森林回归算法精确估计锂离子电池的放电容量.pdfVIP

中原工学院硕士学位论文

Accurateestimationofdischargecapacityoflithium-ionbatteries

basedonrandomforestregression

Abstract

Thebatterymanagementsystemisacrucialcomponentoftheelectricvehiclespower

system,andaccuratelyestimatingthedischargecapacityisakeyandchallengingdemandin

advancedbatterymanagementtechnology.Inpracticaluse,thedepthofchargeanddischarge

andthemagnitudeofdischargecurrentoftenvaryrandomlybasedontheusersactualneeds,

thusaccuratelyestimatingthebatterysdischargecapacityunderdifferentcharge-discharge

depthsandrandomcurrentconditionsishighlychallengingbutholdssignificantpractical

applicationvalue.Addressingfactorsthatinfluencetheaccuracyofdischargecapacity

estimation,suchasbatterycathodeandanodematerials,charge-dischargedepth,charging

currentmagnitude,dischargingcurrentmagnitude,featureextraction,andtheconstructionof

capacityestimationmodels,thisthesisconductsthefollowingresearchresults:

(1)Systematictestingandcollectionofcycleagingdatafor18650-typelithium-ion

batterieswiththreematerialsystems:F-series(NCM523/graphite),X-series(NCM811-

LMO/graphite),andB-series(NCM811/silicon-carbon).Electrochemicalimpedance

spectroscopy(EIS)dataweresystematicallycollectedunderdifferentcharge-discharge

depths,variouscycleagingstates,andstatesofcharge.Leveragingthedistributionof

relaxationtimes(DRT)technique,frequency-domaininformationwasconvertedintotime-

domaininformationtoexplorethecycleagingmechanismsofbatteriesinthesethreesystems.

(2)BasedonindependentlytestedandcollectedcycleagingdataforbatterieswithF-

series(NCM523/graphite),X-series(NCM811-LMO/graphite),andB-series

(NCM811/silicon-carbon)materialsystems,relaxationvoltagedata