Emissions Due to Plug-in Hybrid Electric Vehicle Charging in High Wind Systems Allison Weis Roger Leuken Jeremy Michalek Paulina Jaramillo Carnegie Mellon University USAEE Annual Meeting July 29, 2013
Feb 22, 2016
Emissions Due to Plug-in Hybrid Electric Vehicle Charging in High Wind Systems
Allison WeisRoger Leuken
Jeremy MichalekPaulina Jaramillo
Carnegie Mellon University
USAEE Annual MeetingJuly 29, 2013
Carnegie Mellon University
Motivation
• Electric vehicles are forecasted to make up 2%-15% of vehicle fleet by 2025
• Significant wind generation is expected in states with Renewable Portfolio Standards– Fluctuations in wind generation will require additional grid flexibility,
some of which could come from controlled electric vehicle charging• Electricity sector emissions are key to understanding if electric
vehicles cause lower emissions overall compared to gasoline vehicles
Carnegie Mellon University
Related Work
• Operational emissions studies with a simplified model for the electricity system– Stephan and Sullivan– McCarthy and Yang
• Operational emissions studies with dispatch assuming controlled charging and existing electricity grid– Sioshansi and Denholm– Sioshansi and Miller
• Full life cycle analysis studies with a simplified model for the electricity system– Michalek et. al.– Hawkins et. al.– Argonne National Lab
Carnegie Mellon University
Research Questions
• Can controlled charging reduce the impact of having electric vehicles on the grid compared to uncontrolled charging?– Cost– Emissions– Damages from emissions
• How does a high wind penetration change the impacts of electric vehicles and controlled charging?
Carnegie Mellon University
System Overview
Conventional Power Plants
Non-vehicle LoadPlug-in
Vehicles
Energy Balance Wind Power
All Vehicles
Carnegie Mellon University
Power Grid Data
• 2010 PJM power plants and 2010 fuel prices by state• 5 transmission regions with power limited connections
Region 1
TI1-5
Region 3
Region 5
Region 4
Region 2
Carnegie Mellon University
Wind Plant Data
Eastern Wind Integration and Transmission Study on-shore production data at a 10-minute resolution
Added by capacity factor (high to low) within PJM region until wind plants capable of producing 20% of the load
7
Carnegie Mellon University
Electric Vehicle Profiles• Driving profiles from National Household Travel Survey
– Charge at home at the end of the day• Uncontrolled charging based on all passenger vehicles• Controlled charging based on 20 representative profiles
• 16 kWh battery PHEV (Chevy Volt)
• 10% of passenger vehicles in PJM (2.4 million)
Aggregated
Optimized 20
Carnegie Mellon University
Unit Commitment and Economic Dispatch
Subject to:• Generation = Load• Spinning reserves• Power plant constraints
o Minimum and maximum generations levelso Ramp-rate limitso Minimum runtime and downtime
• Vehicle battery chargingo Battery state of chargeo Charge rate limits
minimize ∑𝑡 𝑖𝑚𝑒
∑𝑝 𝑙𝑎𝑛𝑡
StartupCosts+ShutdownCosts+FuelCostsMixed Integer Linear Program:
Carnegie Mellon University
Controlled charging significantly reduces the cost of PHEV charging
Controlled Charging Annual Cost Savings
Net Savings % of System Costs % of Vehicle Costs
PJM Base Wind $127 million 0.72% 41%
20% Wind $144 million 1.05% 52%
Carnegie Mellon University
Controlled charging of PHEVs increases generation from coal plants
Coal Coal Coal Coal
Wind
Wind
Combined Cycle
Combined Cycle
Combustion Turbine
Combustion Turbine
Combined Cycle
Combined Cycle
Nuclear
Oil/Gas Steam
Combustion Turbine
Nuclear
PJM Base Wind 20% Wind
Carnegie Mellon University
Increased wind resources help keep controlled charging from increasing emissions
Con
trol
led
Unc
ontr
olle
d
Con
trol
led
Unc
ontr
olle
d
Con
trol
led
Unc
ontr
olle
d
Con
trol
led
Unc
ontr
olle
d
Con
trol
led
Unc
ontr
olle
d
Con
trol
led
Unc
ontr
olle
d
Con
trol
led
Unc
ontr
olle
d
Con
trol
led
Unc
ontr
olle
d
Con
trol
led
Unc
ontr
olle
d
Unc
ontr
olle
d
Con
trol
ledU
ncon
trol
led
Con
trol
led
Con
trol
led
Unc
ontr
olle
d
Carnegie Mellon University
SO2
SO2
SO2SO2
CO2
CO2
CO2
CO2
PM25 PM25PM25 PM25
NOX
SO2
VOCVOC
Damages due to emissions increase with controlled charging, even with large wind resources
• Marginal monetary health damages calculated using the APEEP (Air Pollution Emission Experiments and Policy) model
• CO2 damages from $35/metric ton social cost of carbon• Increase due to SO2 damages
PJM Base Wind 20% Wind
Carnegie Mellon University
Key FindingsControlled Charging
• Cuts the costs of integrating EVs by 40%-50%
• Reduces energy consumption by about 8% by reducing use of inefficient storage
• Increases the utilization of low cost plants (particularly coal)
• Change in Emissions– higher CO2, PM, SO2 and NOX
emissions– lower VOC and NH3 emissions
• Increases damages from EV-integration emissions.
20% RPS• Slightly decreases costs of
integrating EVs• Greatly reduces emissions
damages of EV integration• With controlled charging
– higher wind utilization – lower CO2, PM, VOC, NOX and
NH3 emissions– higher SO2 emissions
• Controlled charging still increases damages
Carnegie Mellon University
Future Work
• Optimize controlled charging using full social costs by including emission prices in objective function
• Evaluate the emissions and damages from the full life cycle of the vehicles and compare to gasoline vehicles
• Emissions given future fuel prices and power plant fleet
Carnegie Mellon University
Acknowledgements
Funding by:• Doris Duke Charitable Foundation• Richard King Mellon Foundation• Electric Power Research Institute• Heinz Endowment• National Energy Technology Laboratory• National Science Foundation CAREER Award #0747911• Toyota Motor Corporation• National Science Foundation Graduate Research Fellowship Program• Carnegie Mellon Electricity Industry Center through the RenewElec
project
Carnegie Mellon University
Thank you!
Carnegie Mellon University
Carnegie Mellon University
Unit Commitment and Economic Dispatch
C
SDC SUC F G
Start-up costs Shut-down costs Fuel costs
minimize x x xitit it i
ti
i
c h
Subject to:• Generation = Load• Minimum and maximum generations levels• Ramp-rate limits• Minimum runtime and downtime• Vehicle battery charging
minimize ∑𝑡 𝑖𝑚𝑒
∑𝑝 𝑙𝑎𝑛𝑡
StartupCosts+ShutdownCosts+FuelCosts