2025年基于自动驾驶的未来交通与新型电力系统协同报告

Hongcai ZhangAssistant ProfessorState Key Lab of Internet of Things for Smart CityUniversity of Macau Oct 15, 2024 Content Background & motivation Fleet sizing & charging system planning forautonomous EV fleet Routing & pricing of autonomous EVs to promoterenewable generationintegration Autonomous EVs as mobile storage systems toenhancepower systemresilience Summary EVs are dominating future transportation systems EV stock has hit 21M and sales share has risen to 30% in China by the end of2023 (over 40% in 2024) Impact of growing EV charging load - a Hainan example EV charging load at midnight has reached 450 MW, with a rapid increase rate of75 MW/min, significantly higher than other peak periods By 2025, charging load could rise to 800-1,000 MW, further stressing the grid andcompromising system stability EVs as mobile energy storage for power system EVs come in various types with heterogeneous characteristics, but all are capableof working as mobile energy storage to interact with power system Era of autonomous electric yehicles : Global autonomous vehicle market size may exceed 2200 billion dollars by 2023 : Over 1k design models for electric vertical take-off and landing aircraft worldwidein 2024, and already commercialized in the delivery business Autonomous EVs will strength power & transport synergy Fuel cost is the major operation cost (time is not expensive) : Scheduled driving & parking behaviors (no driver to make decisions) Note: fuel efficiency 0.32 kWh/mile for AEVs, and 30 mi/gallon for ICEVs; gas price 3.3 $/gallon;averagedrivingspeed30mile/hour. Research Problems Content Background & motivation Fleet sizing & charging system planning forautonomous EV fleet Routing & pricing of autonomous EVs to promoterenewablegenerationintegration Autonomous EVs as mobile storage systems toenhancepowersystemresilience Summary Fleet sizing & charging infrastructure planning for urban-scaleshared-use autonomous EVs Problem statement: How shared-use autonomous EV compete with traditionalvehicles? ObjectiveFleetsize→ Charging infrastructureConstraints• Mobility demandsAEVdrivingrangeTechno-economic analysisVehicle battery capacity• Charger powerSocietal transportation system impact Methodology - vehicle shareability network Vehicle-shareability network (VSN)* AdoptdirectedacyclicgraphtodescriberelationshipsbetweentripsDescribefleetsizeproblemasaminimumpathcoveringproblemMinimum path covering problem can be solved as a maximum matching problem Methodology - vehicle shareability network with EV charging Describe charging range constraints by identifying charging behaviors andreconstructing vehicle shareability network Methodology - iterative algorithm with polynomial complexity : An iterative algorithm with complexity O(TE N2) Experiments & insights in New York City case New York City: Total trip 485,000 trips/day : Fleet size 13437 (real), 8100 (proposed, 40% reduction): Fleet size with EV charging 9,517 (15% increase because of downtime) Experiments & insights in New York City case Longer driving range & higher charger power may not be economic solution AnAEfleetof(50kW,50kWh)wasthemost cost-effective solutionLarge battery leads to high investmentand operation costsHigh charging power enhances vehicleutilization,buthasmarginaleconomicbenefit15 Experiments & insights in New York City case Automation leads to 45% VMT reduction, and 45% reduction on CO2 and PM2.5emissions (managedICEVvs unmanagedICEV) :Electrification leads to 84% reduction on CO2 (EVvs lcEV): Electrification and automation save over 90% CO2 emissions (AEV vs ICEV) Content Background & motivation Fleet sizing & charging system planning forautonomous EV fleet Routing & pricing of autonomous EVs to promoterenewable generationintegration Autonomous EVs as mobile storage systems toenhancepowersystemresilience Summary Intercity scenario: Routing autonomous EVs to promoteintegration of renewable generation Problem statement: Strategic EV fleet routing & charging on coupled power &transportation networks Powernetwork With power network: EVs may detour toconsumecheaperelectricity-choosecost-minimizing paths Withoutpowernetwork: EVstrytosavetime - choose the shortest paths Transportation network Method: optimization model : Optimize autonomous EV flow to minimize operational costs (quadratic) : Constraints : AC power flow (Second order cone) : Coupled constraints (Linear) : Driving range (expanded network) (Linear) Aarc538A≥0,538A Large-scale(may driveon anypaths) FrarcB.Fpath9Fpath≥ 0.g RequireEVsonlychooselimitedpaths Method: a column-generation like algorithm Iterative algorithm based on generalized locational marginal prices Experiments & insights on an interconnected power &transportation network . Results - distribution of autonomous EV traffic flow Experiments & insights on an interconnected power &transportation network : Results - operation costs (assume one driver in one c