Systematic development of series-hybrid bus through modelling
Tarun HuriaGiovanni LutzembergerDipartimentodiSistemiElettricieAutomazioneUniversitadi PisaPisa ltalylutzemberger@dsea.unipi.it
emissions vehicle manufacturershave started developinghybrid and electric versions of theirvehicles.This paperversionof an existing busthrough power-trainmodelling.conventionalbus validatingitbymatchingthesimulation The approach involves accuratemodelling of theresults with the tests conducted on the actual bus and thendeveloping the hybrid version of the bus.
Keywords-hybrid vehicle modelling powertrains
1.NOMENCLATURE
AUX Auxiliaryelectrical loadBSFC Brake SpecificFuel ConsumptionED Electric DriveEGS Electrical GeneratingSystemEM Electric machine(either operating asmotororEPC ElectronicPowerConditioner generator)EV Electric VehicleHEV Hybrid ElectricVehicleICE Internal CombustionEnginePMM PowerManagementModuleRESS RechargeableEnergyStorageSystemSORTStandardisedOn-Road Test
II.INTRODUCTION
Withgrowing concernforenvironmentalproblemsamongstgovernmentsandinternationalpolicyformulation agencies more stringentstandardsforfuelconsumptionand emissionshavebeendeveloped.Vehicle manufacturershave focused their attention on development of hybrid andelectricversionsoftheirvehicles.Thesevehicleshaveadvanced power-trainsforefficient utilisation of energy.
Electric vehicles(EVs)appear tobe the best way out asthey implyreduced oil consumption andzeroin situemissions.However factors suchashighinitial cost shortdrivingrange andlongchargingtimearemajorlimitations.Hybrid electric vehicles(HEVs)were developed tooverethelimitationsofinternal bustionengine(ICE)vehicles and EVs.AnHEVbinesa conventionalelectric drive.When driveninthe electricmode HEVsarezero emission vehicles(ZEVs).HEVshave animprovedfueleconomy paredtoconventionalICEvehicles and
GiovanniPedeENEA
Giovanni SannaBredaMenarinibus
Bologna Italy
S.Maria diGaleria-Roma Italy
havealonger drivingrange thanEVs.Hybrid electricsystems. In the series hybrid system all the torque requiredforpropulsionisprovidedby anelectricmotor whileintheparallel hybrid system the torque obtained from theICE is mechanicallycoupledtothetorquefromtheelectricmotor[1] [2] [3]. The series HEV solution is monly chosenforhybridisingbuses[4] because of theinherentcharacteristics of thispowertrainscheme and willbepursued in this paper also.
Thispaper presents a systematic development of apletelineofseries-hybrid andelectricversions ofexistingcitybusesfortheItalianbusmanufacturer BredaMenarini Bus(BMB)[5] through modelling.The modelsweremadebothusingamercialsoftwareMSImagine.LabAMESim[6](AMESim) and a custom-builtpackage using Matlab Simulink? [7] (Matlab). AMESimperforming SIMulations ofengineering systems. standsforAdvancedModellingEnvironmentfor
Toinspireconfidenceinthequalityof themodels firsttheexisting buswasmodelled andvalidated by matchingthesimulationresultswiththeresultsonactual tests conductedontheexistingbus.After confirming theparameters modelswere developed forthe series-hybridversions of the bus.These were simulated on duty cycles ofthecities where thebuses areplanned tobe deployed.InSectionlll themodelfortheexistingbusanditsvalidationispresented.SectionIVpresentsthemodelfortheseries-hybrid bus.Finally in SectionsV&Vl the results of thesizingandconclusionofthestudyaresummarised.
III.M ODELLING& VALIDATION OF EXISTING BUSES
A.Model of existingbuses
Someof theexistingBMBbusesmodelled are showninusing AMESim.Sub-models were made forthe ICE theautomaticgearboxand theauxiliaryload.TheoutputfromtheICEisgiventothe automaticgearboxand theauxiliaryload.The automatic gearbox drives the wheels and thefeedbackfrom thewheels(i.e.thedifferencebetween thedesired speed and the actual speed) guides the response ofthe driver based on the required duty cycle.The model wasconfigurablefordifferent typesofbuseshaving differentauxiliaryloadsandrunning ondifferentdutycycles.
developedbytheInternationalAssociationofPublicTransport(UITP)BusCommittee [9]forbuses running onindispensablereferencepoint forbustransport whenmeasure put in place.
Bus ModelsVivacity M Avancity L AvancitySWeight (trial) Length 9meters 9 67 tonnes 12meters 14 35 tonnes 20 99 tonnes 18metersWeight (full) 13 29 tonnes 17 08 tonnes 25.06 tonnesEngine Deutz Deutz MANCapacity4.8 158kW litres 7 2litres 213kW 235 kW 10.6 litresTorque MaxPower 800Nm 1200Nm 1600 NmConsumption 44 65 49 66 63 49SORT1 cycle litres/100 km litres/100 km litres/100 km（trialload)
Inorder toevaluate thetruebenefits of theseries-hybridand electric buses (which were to be developed) we neededexisting busesbased on a standard duty cycle.Fuel tobenchmark themagainst thefuel efficiency of theconsumptionvaluesof the existingbuses ontheStandardised On-Road Test SORT1 (shown inFigure 1)cycle were provided byBMB.
Fig. 1 The SORT1 standard cycle: speed profile
SORTprovidesrealisticmeasurementsastests arecarried out on-road and it applies toa bus and not asingleengine block in alaboratory.Different buses can beuVItalia[8] paredusingthesamebasictestmethodandprotocol.
These tests were carefully carried out byTalong with themanufacturer.The SORT cycles were
Fig. 2 The modlel of the existing bus on LMS Imagine.Lab AME Sim
power goes to the torque converter thento the gearbox andfinally to the wheels.The loss of power at each stage was clearlyemulatedbythe simulation.
B.Validating the model
BMBput at disposal of the studybasic data about thebus including some data onthe engine andautomaticgearbox.The buses employ diesel engines by majorEuropean automotive industries whileZF is the automaticgearbox supplier.It alsoprovided the actual duty cycles ofitsbusesrunningin cities where the hybridversions of itsbus are planned to run.However accurate data about thepower absorbed by the auxiliary load was not available.Thesimulationresultsof themodelwerevalidated againstthe actual SORT1 cycle test results supplied by BMB.All the existingBMB urban buses were modelled and theirsimulationresults pared with the testresults given byBMB.
Fig. 3 The power flows in the model from the ICE to wheels throughthe torque converter and gearbox
The notable feature of the model was its ability tosuccinctlyandaccuratelyreproducealltheresultsobservedin the actual tests including fuel consumptionfigures forthe SORT1cycle.Figure 3 shows the result of the powerflowfrom theICE to the wheels.Apart of the powerfromtheICE is absorbed by the auxiliaries.Theremaining
these results to a very good degree together with validation Itwasnoted thatthesimulationsonthemodelmatchedof the parameters used including the fuel map. Once
validated，thepowerrequirementoftheauxiliaryloadforeachofthebuseswasaccuratelyascertained.
IV. MODELLING THE SERIES HYBRID BUSES
TheICEof the conventionalvehicle is sized for theonly source to meet the powerand torque requirements of maximumpower and torquerequirementssinceit is thethewheels.Apartfromdisadvantageslikehighersize partsof the Brake SpecificFuel Consumption(BSFC) map thereby consuming higherfuel.It would bemore energyefficient if theICEoperatedonlyat themostefficientpartsof theBSFCmap.Secondly conventionalvehicles do nothavetheabilitytorecuperatebraking energies.Instead StorageSystem(RESS) which stores energy(either frombrakingorfromtheprimaryconverter)that islaterusedforthe primary converter which works to switch off theICEwhen thevehicleisstationary(bus stops traffic signals orjams etc) to enhance the fort of thepassengers byeliminating noise and pollutant emissions.In thispaper theauthors employed the ON/OFF strategy.
A.The series hybrid propulsion architecture
Theseries-hybridpropulsionarchitectureadoptedforthis paper is shown in Figure 4 with the arrows along thepowerfluxes showing the actual directions when thenumericalvaluesarepositive.
Fig.4 Principle scheme of a series-hybrid vehicle drive-train
The ICE is coupled to an electrical generator(EGS).The presence of a storage system(RESS) provides thevehicleflexibilityin sharing the powerrequired for propulsion.This management is carried out by the on-boardPowerManagement Module(PMM) which continuouslymonitorstheloadrequirement and decideshowtoshare thegoals (typical goal is vehicle efficiencymaximization).
B.Sizing of ponents of the series hybrid propulsion system
Anefficientseries-hybrid electricpropulsionsystemrequires optimal sizing of allits ponents. Models of theseries-hybridbus were built on bothAMESim andMatlab. Theresults of simulationwere parable whichgave the
All subsystems weremodelled weighting the accuracyand plexity for the purpose considered.In particular since thefastest transients useful to globally size thehybrid
drive trainhave constant times of the order of100 ms much fasterphenomena such as bustion dynamics orvalveswitchinginsideelectronicconverters wereconsidered tobealgebraic.Figure5depictsthesimulationschemeinAMESim.Themain subsystemswere:
Internal bustionengine(ICE) themain source forused the BSFC maps.The model could emulate theengine torque mechanicalpower engine efficiency fuel consumption emissions and engine speed.?Electric generator(EGS) coupled to the ICE thatgenerated electricity availableforpropulsion.Rechargeable energy storage system(RESS)that couldbeposed of devices such as electrochemicalbatteries supercapacitors fly-wheels etc.The authors modelled theRESSwitha lithiumionbattery density and energy density. consideringitsgoodperformanceintermsofpowerAuxiliaryelectrical load(AUX) that could belights ! the control system the air-conditioner etc;ElectricDrive(ED) to provide power and torque to thewheels and also to produceregenerative power consistingofanelectricmotorandapowercontroller;and？ The vehicle which also contained sub-systems formechanical transmission and vehicle dynamics.Thefinaldrive consisted of a fixed gearratiodriven by thetractionmotor.Mechanical equations describing thevehiclelongitudinalbehaviourwasused.
Eachsub-modelwascarefullyconstructedtobeable toaccurately emulate experimental results.e.g.themodel fortheelectricmotor/generatorneededparametersinafiledefining the lost power vs.thetorque and therotaryvelocity.Themodel of thebatteryneededdata fileslistingtheinternal resistance vs.the depth of discharge.Similarly the plete BSFC maps were fed to the ICE sub-model and themodelcoulddetermine themostefficient operatingline forthe each level ofpower demanded.Efforts werefor efficiencies. madetomodelrealisticsub-systems building inparameters
ThePMMwas themost criticalelement in the model.It was programmed to determine the most optimal way tomeet the driver's requirementsforpower while utilizing thetheforward approachforthismodel which was: two energy sources of thevehicle.The authors employed
?a driving duty cycle(calledmission profilein theprogram)fed to the driverblock.?the driverblockconverted thereferencespeed intomandsforthePMM.?thePMM determined the mostoptimal strategybased ontheinstantaneous power demand levels and feedbackfrom key parameters of the ICE EGS RESS and ED.
Fig. 5 Series-Hybrid bus model in AME Sim
C.Possible energy management strategies
In a series hybrid solution allpowerfortractioniselectric.The sum of energies of the two(or more)powersources in the series-hybrid vehicle isusually depicted as a DCbus.Powerneededforthetraction and auxiliariesistakenfrom thisDCbus.
Thefundamentalroleof thePMMistointerpretdriver′s mands and accordingly determine which partof therequested propulsivepowerwouldbedeliveredbythe EGS and which by the RESS.In other words todetermine how to depose the quantityPep (t) into PeGs(t) and PREss(t):
This degree of freedom could be used to minimize anobjectivefunctionthatcouldbefuelconsumption.Therelationship(1) is guaranteed by thephysics of drive train.A possiblecontrol strategy could be:
Pep(t)is determined to answerthe driveras closely aspossible.It could be considered a directconsequence of trip characteristics vehiclemass andpower losses in the ED.Peas(t) is determined byPMM according to someoptimization rule(that will be discussed later).？ PREss(t) is automatically determined by difference.
Theuserload Pep draws the main focus in this controlstrategy while the power generation from fuel;Peas;isgiven a supporting role.This control strategy(described inandis also adoptedfor thispaper.Theusefulpowerthatgoesinto theload Pep(t) could be imagined to beconstituted by an average value and a ripple.Eq. (1) is thusmodified as follows:
It ispossibletocontrol the systemsuch that thequantityr(t) is pletely delivered by PRess (t) and does not formpart of theprimaryconverter:
Hence theICEdelivers only the average powerrequested bypropulsion leaving the RESSto deliver therest.
The strategypresented aboverequiresa consideration(even approximate) of the future system load i.e.thefuture behaviour of the power demand Peo (t) which is a Srequestfortorqueand thevehiclefunctionofthedriver duty cycle.The approximatelevel ofpower neededby thevehicleinfuture could beobtained by multiplying the pasthistory of Pep(t) with a simple filter e.g:
？ Pegs(t) is the output of a filter having as inputPeD (t)
and as a transfer function 1/(1e).
smandsTnbothcases a suitablevaluefotneeds tobechosen.
values of the ICE angular velocity are reported directly on the engine map Fig. 6 Output of the internal optimisation algorithm: the optimal
Afterdetermining Peas(t) aninternal algorithm wasused tochoosetheoptimalvaluesof theICErotaryvelocity(Fig.6) corresponding to the minimum fuel consumption.More detailsonpossiblehybridvehicleenergymanagementstrategiescanbefound in[13] [17].
V.RESULTS
The objective of the study was the sizing of a pleteconventional buses.Thepropulsion system sizing was sized lineof series-hybridbusesforBMB to augment theirinaccordancewiththeboundaryconditionsregardingperformance of the vehicle at fulload.
1.Level road
b)pure electric mode on urban cycles(22%of thetotal time) as a zero emission vehicle (ZEV)
a) max speed:80 km/h
2.Road with16% gradient
a)max speed:10 km/h (range 0 5 km)b)start-up acceleration:0 3m/s2
The RESS was sized considering condition 1b)(pureelectricmode onlevel gradient) according to the specifiedrange and maximum current limitsfor theLi-ion battery.
Conditions2a)and2b) respectively identified themaxtractivepowerand themax(starting)tractiveeffortforthepropulsionsystem:therefore they pletely definedcharacteristicsfortheED.
The sizing of theICEcouldbeevaluatedwithreferenceto the maximum power i.e. the max power needed by the ICE or theefficient power i.e. the power at which theengineefficiency ishighest.
Due to the powermanagement algorithms selected theefficientpowerwouldalsobetheaverageusefulpowerevaluatedbyconsidering theON/OFFstrategyfortheprimary converter.Infact the two diferent sizing criteriacould be summarised as:
constant speed drive at50km/h ICEalwaysON fullyloadedvehicle:theconstantICEpowerobtainedisthe
maximumpowerrequired;Vehiclerunning theSORT1cycleonlevel gradientwith theON/OFFstrategy:when thevehicleisstationary orinZEVmode theICEis switched off.TheICEpowerwashence evaluated throughthe
following energybalance equation:
where Pet isthe efficientpowergeneratedfrom ICEduringthe ON-statePavgistheaveragepowerrequestedfromthepropulsion.Emax and Emin represent the switch ON-OFFPef could be termed asfficient power. limits(also indicated inFigure7).The calculated power
Fig. 7 Meaning of the quantities used for describing ICE ON-OFFstrategy correction
The main characteristics of the sizing arelisted inTable11.
Bus ModelsVivacity M Avancity L Avancity SWeight (fuload) Length 9meters 13 95 tonnes 12 meters 17 96 tonnes 18meters 26 67 tonnesWeight (partial load) 11 3 tonnes 14 3 tonnes 21 1 tonnesAuxiliaries 6 kW 9kW 12 kWICEmaxpower 48 kW 61 kW 84kWRESS Energy ICE efficient power 36kW 31 1 kWh 48 kW 51 8kWh 69kW38 9kWh
From themainresults it wasinteresting tonote that thein the conventional bus.The authors did not have access to sizeof theICEneededreducedsignificantlyfromtheonespecificfuel consumptionmapsforsmallerengines andhence toevaluate thefuel consumptions programmed themodeltodownscaletheexistingBSFCmaps.Theenginewasdownsizedwithreferencetothetorque keeping thethus avoiding thepoorefficiencyzones of theoriginal sameBSFCvaluesasthebigger engineforthe smaller engine.Fuel consumptions aresummarised inTable Ill paringperformanceofthehybridbusestotheconventional ones.This however wasonlyanad-hocsolution but the authors are confident that theresultsobtained through this approach would also be similar.
TABLE II. FUEL CONSUMPTION
Bus ModelsVivacity M Avancity L Avancity SLength 9meters 12 meters 18metersHEV (fuload) 43 57 57 89 83 68SORT1 cycle Conv.(full load) litres/100km 55 98 75.89 litres/100km litres/100kmSORT1 cycle litres/100km litres/100km litres/100km
Figure8showstheresultsof the simulationforthemodel for theAvancity L bus on the SORT1 cycle.ltclearlyshowsthedistributionof thepowerfromthetwosources.TheRESS ED EGS AUXpowerprofiles arerelatedby the following equation:
