Page - 170 - in Emerging Technologies for Electric and Hybrid Vehicles

Energies 2017,10, 5 6R& IJVKRUW IJORQJ (a) 6R& 5VKRUW 5ORQJ (b) Figure 8. Parameter estimation results for the constant-current (CC) charging scenario: (a) time constant; (b) resistance. For thecaseof thedynamicdrivingscenario, thedischargingpulse-rest test is implemented,and thedata fromthe restperiodsare adopted in theparameter estimation. According to theanalysis in Section 3.2.2, different time constantswill be obtained from thefitted experimental datasets in different lengths. Firstly, inorder tocompare thebestfitperformances for themeasureddatasets in different lengths, themeasuredbattery terminalvoltageresponseat60%SoCduringa2-hrestperiod isadopted,andthecurvefittingresultsareshowninFigure9. It canbeobservedfromFigure9a that thefittingresultof thewholemeasuredvoltageresponseshowsabetterperformanceduringmostof therestperiod,especially in theequilibriumstate.Whereas for theperformanceof thefirst200s, the fittingresult throughtheprior0.5-hmeasuredvoltageresponseyields lesserrors,which is illustrated inFigure9b. Parameterestimationresults inFigure10showthe timeconstantsestimatedfromthe measuredvoltagedataset indifferent lengths, rangingfrom30min–2hwitha30-min interval. It can beobservedthat the timeconstants,bothfor the longtermandtheshort term, increasesimultaneously whenthe lengthof thefitteddataset increases. Inaddition,bycomparingFigure10withFigure8a, it canbe concluded that the time constants applied in theCCcharging scenario and thedynamic drivingscenarioshowdifferentvariationtendencies.Hence, it isessential toadoptdifferentsetsof modelparameters fordifferentoperatingscenarios. 7LPH PLQ PHDVXUHG W K W K W K (a) 7LPH V PHDVXUHG W K W K W K (b) Figure9.CurvefittingresultsofVP during the restperiodof thedischargingpulse-rest test at 60% SoC: (a) theoverall result; (b) aclose lookat the transientpartat thebeginning. 170