Seite - 94 - in Emerging Technologies for Electric and Hybrid Vehicles

Energies 2016,9, 86 stations insuchawaythatminimize total costs. These total costs includestations-constructioncost, waiting timecost, and refueling cost. Olivella-Rosell et al. [56]proposeanagent-based simulation approachthatallowsforecastingtheEVchargingdemandinacertainurbanarea,andtheysuccessfully test theefficiencyof theirmodel in thecityofBarcelona,Spain. Inthespecificcaseofbatteryswappingstations,whenavehiclearrives, it requestsafullycharged batterypallet toreplace thenearlydepletedbatteries it currentlyholds. Therequestcouldeitherbe satisfiedbyafullychargedbatterypallet fromthestationstorage,orbyapallet that is just completing its charging. If the request is indeedsatisfied, thevehicle in turndeposits a fullyorpartially spent pallet. If thereare idlebatterypallet chargersat thestation, thespentbatterypallet isplacedononeof themandits rechargingbegins,otherwise it iskept inaqueueuntilabatterypalletcharger isavailable. If, instead, there isnofullychargedbatteryavailableat thestation, thenthevehiclecould leaveandgo toadifferentstation.Alternatively, it couldwait forabattery to fullycharge,whichmaytakesome time. Thevehiclecouldeventake, ifnecessary,a replacementbattery that isonlypartiallycharged anduse thatpartially chargedbattery to travel toanotherbatteryswapstationon its route. In this case, thevehiclewillhave tostopearlier thanplanned,andthis influences theroutesplanning,which estimatedsomestopsat stationsandsuddenly thevehicle is forcedtoperformothernotcoveredstops. Dependingonboth thenumberofbatterypallet chargers the stationholdsandnumberofbattery pallets thestationkeepsonhand, thesizeandattendantcostof thestationwill change. Theavailability ofchargedbatterypalletsatanygiventimedependsonthesizeof thestation, the inventoryofpallets, and thedemand for chargedpallets the station is experiencing. The station incursan indirect cost fromtheunavailabilityofchargedpalletswhenanEVarrives foranexchangebecause thedriverwill nothavetopayforabatteryswap,andtheremaybea lossofgoodwill fromtheunservedcustomer. Models toevaluate totaldirectand indirect costs forpossibledecisionsonstationsizingandinventory holdingwouldbevery important indesigningthebatteryswapping infrastructure. In the literature related to battery-swap station size, Zheng et al. [36] proposed amethod for locatingandsizingbatteryswapstations indistributionsystems,whichare twodeterminantskeys in the take-upofEVsasexplainedbefore. Theproblemismodeledasmaximizing thenetpresent valueof thebatteryswapstationproject,where thebatteryswapstationmodel, loadtype,network reinforcement,andreliabilityare taken intoconsideration. In the case of hydrogen-basedEVs, since theprice of hydrogen exhibits an inverse feedback interactionwith theadoptionrateof fuelcellvehiclesandcorrespondingdemandforhydrogen, this behavior has a compounding cyclical effect [57]. Existingmodels often fall shortwith respect to incorporatingthiseffect intocapacitydecisions. Further, thecapacitydecisiondependsonthedemand that isunknownapriori [58].Gametheorymayberequiredtodeterminetheoptimal timingofcapacity investment. Thus, for example, Qin et al. [59] uses an option-based approach todemonstrate the behavior or optimal capacitydecisions consideringavarietyof factors. StrubenandSterman [60] discussesrequirementsofsustainedadoptionofhydrogenEVs,andGnannandPlötz [61]providea reviewof integratedmarketandinfrastructuremodels. 3.4. FleetSizeandMix EVsare likely tobeused indeliveryfleetswithotherkindsofvehicles.Awell-studiedbranchof theVRPliterature ispreciselyaddressingtheproblemofheterogeneousfleets indeliveryfleets [62]. Asnoticed inLebeau et al. (2015) [63],mergingtheVRPresearchonelectricvehicleswith thefleet size andmixvehicle routingproblemis therefore relevant to comewith recommendations for logistics decisionmakers.Oneof thefirstattempts to investigate thespecificcharacteristicsofEVsaspartof the fleetofaVRPwasachievedbyGonçalves et al. [64]. TheyconsideredaVRPwithpickupanddelivery usingamixedfleet thatconsistsofEVsandvehiclesusing internal-combustionengines. Theobjective is tominimize total costs,whichconsistofvehicle relatedfixedandvariablecosts. Theyconsider time andcapacityconstraintsandassumeatimeforrechargingtheEVs,whichwerecalculatedfromthe totaldistancetravelledandtherangeusingonebatterycharge.Vehiclescanrechargeanywhereduring 94