2025年基于自動(dòng)駕駛的未來(lái)交通與新型電力系統(tǒng)協(xié)同報(bào)告-2025-03-自動(dòng)駕駛

基于自動(dòng)駕駛的未來(lái)交通與新型電力系統(tǒng)協(xié)同PowerandTransportSynergyDrivenbyAutonomousElectricVehicles

HongcaiZhang

AssistantProfessor

StateKeyLabofInternetofThingsforSmartCity

uM澳大

UniversityofMacau

Oct15,2024

2

Content

Background&motivation

autonomousEVfleet

Routing&pricingofautonomousEVstopromote

renewablegenerationintegration

山

AutonomousEVsasmobilestoragesystemsto

enhancepowersystemresilience

Summary

EVsaredominatingfuturetransportationsystems

·EVstockhashit21Mandsalessharehasrisento30%inChinabytheendof

2023(over40%in2024)

Transportationnetwork

*Datasource:IEA,"GlobalEVOutlook2023,"2024.

growing

EV

charging

Impactof

load-aHainanexample

·EVchargingloadatmidnighthasreached450MW,witharapidincreaserateof75MW/min,significantlyhigherthanotherpeakperiods

·By2025,chargingloadcouldriseto800-1,000MW,furtherstressingthegridand

compromisingsystemstability

TimeofDay(15-minuteintervals)

Hainanchargingloadheatmap

AveragedailyEVchargingloadprofilesperstationinHainan

EVsasmobileenergystorageforpowersystem

·EVscomeinvarioustypeswithheterogeneousofworkingasmobileenergystoragetointeract

Small

Medium-

Large

Taxi(50-100kWh)

Private(50-100kWh)

Bus(120-300kWh)

Truck(130-180kWh)

DumperTruck(≥300kWh)Emergency(≥300kWh)

characteristics,butallarecapablewithpowersystem

Onemonth's

≈electricityconsumption

foraresident

Twoweeks'

≈electricityconsumption

forafamily

Oneweek's

≈electricityconsumption

forasix-storyapartment

5

Eraofautonomouselectricvehicles

·Globalautonomousvehiclemarketsizemayexceed2200billiondollarsby2023

·Over1kdesignmodelsforelectricverticaltake-offandlandingaircraftworldwidein2024,andalreadycommercializedinthedeliverybusiness

Globalautonomousvehiclemarketsize(billionUSdollars)

NumberofworldeVTOLaircraftdirectoryentries

*Datasource:STATISTA,"Sizeoftheglobalautonomousvehiclemarketin2021and2022,withaforecastthrough2030",2023.

**Datasource:VerticalFlightSociety,"WorldeVTOLAircraftDirectory",2024.

6

7

AutonomousEVswillstrengthpower&transportsynergy

·Fuelcostisthemajoroperationcost(timeisnotexpensive)

·Scheduleddriving&parkingbehaviors(nodrivertomakedecisions)

Operationscostsbreakdownforride-

hailingservices

■ICEV■AEV

AutonomousEVshavestrongermotivationtodetourforcheaperelectricity

Note:fuelefficiency0.32kWh/mileforAEVs,and30mi/gallonforICEVs;gasprice3.3$/gallon;averagedrivingspeed30mile/hour.

ResearchProblems

Planning

HowtooptimizefleetsizeandchargingsystemsforautonomousEVs?

Pricing

Howtodesignrouting&pricingforEVstoboostrenewableintegration?

Scheduling

HowcanautonomousEVsserveasmobilestoragetostrengthengridresilience?

凸

B

Transmissionnetwork

Large

曲

田

Commercialusers

Distributionrenewables

industrialusers

Centralized

windandPV

generationCharging

stations

Householdusers

Energystorage

乃

Chargingstations

3Charging.3y

Discharging

Electric

cars

和

Lo-

Electrictrucks

Electricbuses

向◎-◎

Electric

motorcycles

SmartV2Gservices8

Content

Background&motivation

Fleetsizing&chargingsystemplanningfor

autonomousEVfleet

Routing&pricingofautonomousEVstopromote

renewablegenerationintegration

AutonomousEVsasmobilestoragesystemsto

enhancepowersystemresilience

Summary

Fleetsizing&charginginfrastructureplanningforurban-scaleshared-useautonomousEVs

·Problemstatement:Howshared-useautonomousEVcompetewithtraditionalvehicles?

●Objective

◆Fleetsize

◆Charginginfrastructure

●Constraints

◆Mobilitydemands

◆AEVdrivingrange

●Techno-economicanalysis

◆Vehiclebatterycapacity

◆Chargerpower

◆Societaltransportationsystemimpact

·H.Zhang,C.J.R.Sheppard,T.E.Lipman,andS.J.Moura,"JointFleetSizingandChargingSystemPlanningforAutonomousElectricVehicles,"IEEETransactionsonIntelligentTransportationSystems,vol.21,no.11,pp.4725-4738,November2020.DOI:10.1109/TITS.2019.2946152

·T.Zengs,H.Zhang.S.J.Moura,andzM.Shen,"EonomicandEnvronmentalBeneftsofAutomatedElecthicvehicleRide-HaingServicesinNewYorko

City,"ScientitIcReports,vOl.14,p.4180,2024.DOI:10.1038/S41598-024-54495-x

期

Starttime

EndtimeStarttime

a

b

Methodology-vehicleshareabilitynetwork

·Vehicle-shareabilitynetwork(VSN)*

·Adoptdirectedacyclicgraphtodescriberelationshipsbetweentrips

·Describefleetsizeproblemasaminimumpathcoveringproblem

·Minimumpathcoveringproblemcanbesolvedasamaximummatchingproblem

-

Endtime

Starttime

EndtimeStarttime

Starttimek

Endtime

自

a

b

g

1

h

m

i

n

f

e

C

C

d

Time

EndtimeStarttime

k

j

h

g

d

Endtime

n

m

i

f

e

Time

DirectedacyclicgraphMinimumpathcover

11

*M.M.Vazifeh,P.Santi,G.Resta,S.H.Strogatz,andC.Ratti,"Addressingtheminimumfleetprobleminon-demandurbanmobility,"Nature,vol.557,no.7706,pp.534-538,2018.

12

Methodology-vehicleshareabilitynetworkwithEVcharging

·Describechargingrangeconstraintsbyidentifyingchargingbehaviorsand

reconstructingvehicleshareabilitynetwork

k

k

:

j

n

n

m

a

m

a

I

b

h

i

g

h

b+g+ch

f

ed

C

f

e

c+d+ch

Currenttime=t

Time

Time

Currenttime=

AnewVSNgraph

chargingevents

Identifyfirst

Methodology-iterativealgorithmwithpolynomialcomplexity

·Aniterativealgorithmwithcomplexity0(TEN2)

Precomputednode-nodehourlytimematrix-NYCTLCdata

Constructroadnetwork,calculateroadODtravelingtime

VSNeraphl

Combinetripsbeforethe

earliestchargingeventforall

othertripchains

Considersecondary

ConstructVehicle-Shareabilitynetwork(directedacyclicgraph)

trafficspeedimpact

becauseofvehicle

automation

Identifytripsand

reconstructvehicle

sharablegraph

Linktravel

timeupdate

Combinetripsbeforethefirstchargingeventforeachtripchainwithachargingevent

ConstructbipartitegraphMaximummatching

Secondary

Addchargingevents

impact

Enumerateeverypathidentifiedupdated

Yes

dispatchtocharge

No,butlargedowntime

Rangeviolated?INo

ConvergenceCriteria?

Yes

13

Outputresults

Experiments&insightsinNewYorkCitycase

·NewYorkCity:Totaltrip485,000trips/day

·Fleetsize13437(real),8100(proposed,40%reduction)

·FleetsizewithEVcharging9,517(15%increasebecauseofdowntime)

Autonomousconventionalvehicle(AV)tleet

operationstatusacrossa7-dayweekInfrastructureplanninginNYC14

Experiments&insightsinNewYorkCitycase

·Longerdrivingrange&higherchargerpowermaynotbeeconomicsolution

Battery

(a)Fleetpurchasecosts.

0

Battery

(c)Investmentcosts

Battery

(b)Infrastructureplacementcosts.

Battery

(d)Operationalcosts

(e)Totalcosts

·AnAEVfleetof(50kW,50kWh)wasthemostcost-effectivesolution

·Largebatteryleadstohighinvestmentandoperationcosts

·Highchargingpowerenhancesvehicle

utilization,buthasmarginaleconomic

benefit

15

Experiments&insightsinNewYorkCitycase

·Automationleadsto45%VMTreduction,and45%reductiononCO2andPM2.5emissions(managedICEVvsunmanagedICEV)

·Electrificationleadsto84%reductiononCO2(EVvsICEV)

·Electrificationandautomationsaveover90%CO2emissions(AEVvsICEV)

Density

CarbonemissionscomparingmanagedorunmanagedAEVandICEV

Density

Emissions(kgPM2.5)

PM2.5emissionscomparingmanagedor

unmanagedAEVandICEV16

Content

Background&motivation

Fleetsizing&chargingsystemplanningfor

autonomousEVfleet

Routing&pricingofautonomousEVstopromoterenewablegenerationintegration

AutonomousEVsasmobilestoragesystemsto

enhancepowersystemresilience

Summary

17

Intercityscenario:RoutingautonomousEVstopromoteintegrationofrenewablegeneration

·Problemstatement:StrategicEVfleetrouting&chargingoncoupledpower&transportationnetworks

oWithpowernetwork:EVsmaydetourtoconsumecheaperelectricity-choose

cost-minimizingpaths

Withoutpowernetwork:EVstrytosave

time-choosetheshortestpaths

·H.Zhang,Z.Hu,andY.Song,"PowerandTransportNexus:RoutingElectricVehiclesGrid,vol.11,no.4,pp.3291-3301,July2020.DOl:10.1109/TSG.2020.2967082|

submittedtoIEEETransactionsonEnergyMarket,PolicyandRegulation,2024.

·L.Pan,H.Zhang,andY.Xu,"OptimalPricingofElectricVehicleChargingonCoupled

Powernetwork

豐

5+

Transportationnetwork

toPromoteRenewablePowerIntegration,"IEEETransactionsonSmartPower-TransportationNetworkbasedonGeneralizedSensitivityAnalysis1"8

Method:optimizationmodel

·OptimizeautonomousEVflowtominimizeoperationalcosts(quadratic)

·Constraints

·ACpowerflow(Secondordercone)

·Coupledconstraints(Linear)

·Drivingrange(expandednetwork)(Linear)

Large-scale(maydriveonanypaths)

·Pathflowconstraints

19

=B?F,RequireEVsonlychooselimitedpaths

Method:a

column-generation

likealgorithm

·Iterativealgorithmbasedongeneralizedlocationalmarginalprices

Adoptgeneralizednodalelectricitypricesto

estimatetotaldriving

costs(time&electricity)

Initializepathset(shortestpath)foreachODpair

SolvethePEVroutingproblem

Solvepowerflow&calculate

generalizednodalelectricityprices

Identifyminimum-costpathfor

eachODpair

EVcanchangepathtoreducecosts?

No

Outputsolution

Remark:Thescaleoftheidentifiedpathsetismuchsmallerthanarcset;Thealgorithmconvergesinafinitenumberofiterations

Addthenewpathtotheset

Yes

Adoptshortestpath

algorithmtoidentifycost-minimizingpaths

*F.He,Y.Yin,andS.Lawphongpanich,Transp.Res.PartBMethodol.,2014.

20

Experiments&insightsonaninterconnectedpower&transportationnetwork

·Results-distributionofautonomousEV

Trafficflowdistribution(beforerouting)

trafficflow

Trafficflowdistribution(afterrouting)

21

22

Experiments&insightsonaninterconnectedpower&transportationnetwork

·Results-operationcosts(assumeonedriverinonecar)

Significantoperationcostsreduction(-20%)withmilddetour

Powergenerationandpurchase(MWh)Fuelingcosts(k$/h)

CaseElectricitypurchaseConventionalDGRenewableDGElectricityEmissionTotall

10.371.14105.37

110.21

Deourtime8.83

0.447.19

0-22.61

0.00422.56

Chargi3gtime

6.45

6.36

6.36

0.0990.012

0.66

0.66

2.370.29

15.58

15.53

6.050.86

5.45

0.64

94.65

113.98

1

2

3

4

一

Muchcleanerenergyconsumptionconsideringpower-transportnexus

Case

Electricitypurchase

(MWh)

ConventionalDG(MWh)Bus5

Renewable

DG(MWh)

Averagerenewable

Bus9

Bus10

Bus11

Bus13

powercurtailment(%)

1

10.36

6.05

21.51

24.11

21.75

27.27

21.13

2

1.14

0.86

27.69

30.0

26.28

30.0

5.03

23

Experiments&insightsonaninterconnectedpower&transportationnetwork

·BenefitsofroutingautonomousEVsaremoresignificantwith

·Morecongestedpowernetwork

·Lowerper-unitdrivingtimecost(autonomousvehicles!)

150%per-unittimeco

%

Openquestion:trade-offbetweendeliverytime&operationalcosts?

Intracityscenario:Pricingofurban-scaleautonomousEVcharging

·Problemstatement:Chargingserviceprovidersstrategicallypriceservicesconsideringfactors:

oEVs'routing&chargingbehaviorsareaffectedbybothchargingpricesand

congestionconditions

0EVs'routing&chargingbehaviorsinturnaffecteconomicoperationof

interconnectednetworks

oPricecompetitionexistsincharging

servicemarket

Powernetwork

豐

Transportationnetwork

25

Method:optimizationmodel

·Optimizechargingpricetomaximizechargingservices'profit(Bilinear)

·Constraints:Chargimaxtcan

·Pricingboundconstraints(Linear)

·Transportationflowconservationconstraints(Linear)

·Time-latencyconstraints(Powercone)

·Userequilibriumconstraint(Complementaryconstraints)

0≤f⊥v-u≥0Large-scale(equaltopathsetsize)

PathflowPathcostLowestpathcost

Method:gradientdescentalgorithm

·Decomposetheoriginalproblemintotwoconvexsub-problems(bilinearand

complementaryconstraintsareeliminated)

·Approximatethegradientaxev/aλoftheuserequilibriumsub-problem

·Solvetheproblemiterativelywithgradients

λ

User

equilibrium

a

Proposedmethod

Exponentialtime

complexity

Polynomialtime

complexity

User

lequilibrium

Originalmodel

Pricingproblem

Pricingproblem

27

Experiments-Algorithmicperformanceinlargenetworks

·Proposedmethodisabout50timesNETWORKCONFIGURATION

Network

OD

Node

Arc

FCS

EasternMassachusetts

1113

74

258

41

Winnipeg

1373

1057

2535

97

fasterthanconventionalmethod

·Solutiontimeismoresteadilyincreasing(Polynomialcomplexity)

·Capabletosolveurbanscalenetworks

ALGORITHMPERFORMANCEINWINNIPEG

Method

Pathsetsize

3996567874979200

MP

Profit($)

Solutiontime(s)

GDGSA

Profit($)

Solutiontime(s)

1941.7-2238.2-2270.6_2374.3

172.6375.8432.4493.5

ALGORITHMPERFORMANCEINEASTERNMASSACHUSETTS

set

size

2775

3182

4039

6085

Path

Method

MP

Profit($)

Solutiontime(S)

1410.2216.4

1443.1641.6

1543.51570.1

Profit($)

Solutiontime(s)

1473.9

144.6

1421.3

70.8

1473.6

134.5

1464.1102.9

GDGSA

MP:mathematicalprogramming

GDGSA:proposedmethod“-”:over7200seconds

28

Content

Background&motivation

Fleetsizing&chargingsystemplanningfor

autonomousEVfleet

Routing&pricingofautonomousEVstopromote

renewablegenerationintegration

AutonomousEVsasmobilestoragesystemsto

enhancepowersystemresilience

Summary

AutonomousEVsasmobilestoragesystemstoenhancepowersystemresilience

·Problemstatement:StrategicoperationofautonomousEVsincoupledpower&transportationnetworkstoenhancepowersystemresilience

Time

oPre-hazard:howtomaximizereservesupplyconsideringEVs'spatial-temporalcharacteristics?

oOn-hazard:howtomitigatesystemperformancelossconsideringthedynamiceffectsofhazard?

oPost-hazard:howtorestoresystemperformanceconsideringthedamagedpower&transportationnetworks?

·L.d,."ZIg,.ia,.i,"iiolelreiteiiioll:e13InectedPower-TransportationSystemUnderNaturalg

·L.Kong,H.Zhang,W.Li,H.Bai,andN.Dai,"Spatial-temporalSchedulingofElectricBusFleetinPower-TransportationCoupledNetwork,"IEEETransactionsonTransportationElectrification,vol.9,no.2,pp2969-2982,2023.DOI:10.1109/TTE.2022.3214335

Method:optimization

model

·OptimizeEVandpowernetworktomaximizetherestorationrevenue max[-CA+E(Rpe-cperD]

CAllo=COppo+CChg,pre-Rs

cOper,u=cGen,w+cChg,post-RS,

Pre-allocationcostOperationrevenueOperationcost

·Constraints

·ACpowerflow(Secondordercone)

·EVlocationconstraints(Bilinear)

·EVpowerandenergyconstraints(Integer)

·Hurricane-induceddamagemodel

CumulativedamagemodelofpowertowerPiece-wiseroaddamagemodel30

Experiments-restoredtopologyand

·UtilizingautonomousEVscansignificantly

restoredpowersupplyafterhazards

Restoredpowernetworktopologyundercase1

(proposed)and4(withoutEVs)

powersupply

increasetherestoredareaand

Time/h

Totalsuppliedowerprofilesatbuses23and30underdifferentcases

33

Experiments-revenue

·BenefitofautonomousEVs'participationinrestoration:

·Case1(Proposedmethod)hasasignificantrevenueincrease(369.53%)thanCase4(Benchmark)

·BenefitofspatialschedulingofautonomousEVs:

·Case1(spatial-temporalscheduling)hasmorerestorationrevenue(31.75%)thanCase2(onlytemporalscheduling),whichoffsetsextratripenergycost

Case

Costduringpre-hazardperiod

Costduringrestorationperiod

Costduringpost-hazardperiod

Restoration

revenue

PTCN'stotalrevenue

EVs'

charging

EVs'

opportunity

EVs'

re-dispatchtripenergy

DGs'

generation

EVs'energyrestoration

1

-2,715.85

-6,000.00

-158.12

-4,291.44

-3,533.78

61,059.40

44,518.34L

2

-2,843.49

-6,000.00

N/A

-4,833.00

-3,357.76

46,343.49

29,309.24I

3

-3,546.84

-6,000.00

N/A

-5,895.45

-4,535.41

38,741.24

18,763.54|

4

N/A

N/A

N/A

-6,628.50

N/A

16,110.00

9,481.50

34

Content

Background&motivation

Fleetsizing&chargingsystemplanningfor

autonomousEVfleet

Routing&pricingofautonomousEVstopromote

renewablegenerationintegration

AutonomousEVsasmobilestoragesystemsto

enhancepowersystemresilience

山

Summary

35

Summary

·Synergybetweenpowerandtransportationsystemsisamajorfeatureof