An Integrated Transportation-Power System Model For a Decarbonizing World

Abstract

The increasing demand for electricity from electric vehicles (EVs) will require new paradigms to guarantee reliable and low-cost electricity. This study evaluated an approach to coupling an agent-based travel demand simulator and an electricity grid model to assess the economic costs of supplying power to meet this additional demand in Chicago. The study found that shifting from personal EVs to a fleet of shared, fully-automated all-electric vehicles (SAEVs) could lower per-mile emissions, congestion, and embodied vehicle and charging infrastructure emissions. The results should compel policymakers to shift the cost of providing power onto commercial customers, like electric ride-hail fleets, through price-indexed electricity prices, which can shift charging to off-peak periods or away from resource-scarce hours.

Full text

1
AN INTEGRATED TRANSPORTATION-POWER SYSTEM MODEL FOR
A DECARBONIZING WORLD
Matthew D. Deana, Krishna Murthy Gurumurthyb, Zhi Zhoub,
Omer Verbasb, Taner Cokyasarb, & Kara Kockelmanc*
a Department of Civil and Environmental Engineering, Institute of Transportation Studies,
University of California, Irvine, 4000 AIRB, Irvine, CA, 92697
b Transportation and Power Systems Division, Argonne National Laboratory
9700 S. Cass Ave, Argonne, IL, 60439
c Department of Civil, Architectural and Environmental Engineering, The University of Texas at Austin,
301 E. Dean Keeton St, Stop C1761, Austin, TX, 7871220
* Corresponding Author: kkockelm@mail.utexas.edu
“This is an Original Manuscript [pre-print] of an article published by Taylor & Francis in Transportation
Planning and Technology available at https://doi.org/10.1080/03081060.2024.2348721.
ABSTRACT
The increasing demand for electricity from electric vehicles (EVs) will require new paradigms to
guarantee reliable and low-cost electricity. This study evaluated an approach to coupling an agent-
based travel demand simulator and an electricity grid model to assess the economic costs of
supplying power to meet this additional demand in Chicago. The study found that shifting from
personal EVs to a fleet of shared, fully-automated all-electric vehicles (SAEVs) could lower per-
mile emissions, congestion, and embodied vehicle and charging infrastructure emissions. The
results should compel policymakers to shift the cost of providing power onto commercial
customers, like electric ride-hail fleets, through price-indexed electricity prices, which can shift
charging to off-peak periods or away from resource-scarce hours.
Keywords: integrated modelling, electrified mobility services, electrified transportation systems,
infrastructure systems, agent-based modelling, transportation simulations, electric vehicles, shared
autonomous electric vehicles

2
1. MOTIVATION
Speeding up the transition from gas- or diesel-powered personal, light-duty vehicles to zero-
emission vehicles is necessary for developed countries to meet voluntary climate change targets
(Bibra et al., 2022), provided that the power system transitions to low-carbon generation sources.
Variable renewable energy (VRE) technology, like solar and wind, with no fuel costs and small
variable costs, is economically advantageous to fossil fuel generation, thanks in part to production
and investment tax credits and a small levelized cost of electricity (U.S. EIA, 2022a). In addition
to supply-side changes, the electrification of heating sources, home appliances, and vehicles will
increase electricity demand and change its spatial-temporal distribution pattern. Under aggressive
electrification scenarios, the U.S. may need to double or triple its installed generation capacity
between 2018 and 2050 (Mai et al., 2018). Large-scale adoption of electric vehicles (EVs), to an
extent, also shifts decarbonization responsibility to another sector: the power system. The EV
transition increases interdependencies between the transportation and power sectors, which have
historically operated independently, despite both impacting household budgets at the individual
level and overlapping rights-of-way at the network level.
Charging infrastructure access, electricity rate design, and EV charging control are strategies that
can influence charging behavior and the expected rise in electricity demand (Powell et al., 2022).
Studies have investigated the effect of EV charging at different scales, from the power distribution
system effects of increased personal EV adoption (Clement-Nyns et al., 2010; Coignard et al.,
2019; Estandia et al., 2021; Muratori, 2018; Nacmanson et al., 2022; Wert et al., 2021) to
transmission and generation-level effects (Cheng et al., 2018; Forrest et al., 2016; Kintner-Meyer
et al., 2020; Sheppard et al., 2021; Szinai et al., 2020; van Triel and Lipman, 2020; Xu et al., 2020;
Zhang et al., 2020, 2019). The granular scale helps one understand how unmanaged charging may
increase co-incident peak loads in residential neighborhoods, age equipment faster, and lead to
grid capacity upgrades. On the other hand, macroanalysis shows the importance of load flexibility
in avoiding an increase in emissions, deferred capacity expansion to meet peak demand, reduced
investment in utility-scale energy storage, and increasing reliability in the system in the face of
climate change.
EVs could be important grid assets in smoothing the balance between supply and demand at
different time scales. Personal EV owners may allow their local power companies to stagger or
directly control charging at their workplaces and/or homes (Bailey and Axsen, 2015; Bauman et
al., 2016; Tarroja and Hittinger, 2021). If the cost of electricity is indexed to real-time marginal
generation costs (i.e., wholesale power prices), then EV ride-hailing fleets may seek to lower their
electric bills by charging during off-peak hours, assuming this coincides with periods of low
passenger demand (Dean et al., 2023; Iacobucci et al., 2021; Zhang and Chen, 2020). Price signals
are a form of smart or managed charging that can reshape the EV charging demand curve (Anwar
et al., 2022), and mitigate co-incident peak demand. The importance of studying the grid impacts
of large-scale EV adoption also requires understanding how emerging mobility, like shared,
autonomous, electric vehicles (SAEVs), could reshape urban mobility and change charging
profiles away from assumptions in previous studies that mainly focused on personal EVs (Dean
and Kockelman, 2022).
While many have studied the implications of personal, light-duty vehicle electrification on grid
resource adequacy (Anwar et al., 2022), studying EVs within the context of using on-demand
SAEVs is a challenge with technological delays in autonomous vehicle testing and deployment.

3
As recently as the 2016/2017 U.S. National Household Travel Survey (NHTS), 7.3% of
households used taxi services or ride-hail vehicles at least a few times a month (and 1.4% used
these services a few times a week)
1
(Crossland et al., 2021). Given the slow adoption of ride-hail
services and car-dependent land use policies in many American cities, the long-term adoption
levels of SAEVs are uncertain. The timing of the transition to SAEVs depends on technology
learning curves, human factors, falling costs, and regulations (Litman, 2022). Likewise, the
transition to low-carbon power generation depends on declining energy storage technology costs,
government mandates, grid interconnection reviews, land use policies, and public-private
investments (Susskind et al., 2022). Regardless of the timing, the scientific consensus is that a
quick transition from fossil-fuel-powered vehicles and power plants to zero- or low-carbon
alternatives is necessary to avoid the catastrophic effects of climate change (Rogelj et al., 2018;
Sperling et al., 2020).
SAEV fleet charging decisions, vehicle utilization, battery capacity, and infrastructure investments
all impact charging demand. Unlike personal EVs, which are parked for over 95% of an average
day (Shoup, 2011), passenger-serving fleet vehicles may be busy 8+ hours per day (Fakhrmoosavi
et al., 2023). Moreover, SAEV fleets will need to return to central hubs for cleaning, maintenance,
and maybe even charging (if wireless or automatic plug-in adapters at fleet-owned charging
stations are not in use) (Gurumurthy et al., 2022). Centralized charging infrastructure can reduce
land acquisition or leasing costs but increase deadheading between trip ends and charging hubs.
Investing in fewer charging outlets can reduce costs and increase charger use, but limits SAEV
fleet use for power grid benefits, like load shifting. In contrast, fleets could co-locate charging
infrastructure where personal vehicles are normally parked to align charging with periods of solar-
rich power and reduce empty travel between passenger-serving trips. In California and 10 other
western U.S. states, daytime charging of EVs provides significant economic and environmental
benefits (Powell et al., 2022).
SAEVs can have significant impacts on electricity systems, not only due to added power demands
– but also centralized fleet-control’s benefits, like charging modulation in response to transmission
congestion, VRE generation, and total load on the local transformer. Demand-side flexibility is an
important tool in increasing the power grid’s reliability, and removing the human driver from
charging decisions offers greater certainty when relying on EVs as a grid resource. Future SAEV
fleets may offer bi-directional charging (Dean et al., 2023; Iacobucci et al., 2021; Melendez et al.,
2020) at select chargers to provide a temporary power supply and help grid operators avoid rolling
power outages (Dean and Kockelman, 2022). As the transportation and electricity sectors converge
and new modes of transport emerge, integrated models are necessary to identify the opportunities
and challenges of this transition.
In this study, an agent-based transportation simulator estimates the daily charging demand for
personal (household-owned) EVs and a fleet of SAEVs under different adoption levels for the
Greater Chicago metro region, covering 20 counties with nearly 11 million people. The daily
charging demand is added to a state-wide power system operational model with least-cost unit
commitment and economic dispatch functions to understand changes in wholesale power prices
from large-scale vehicle electrification. The study first shows the effect of switching from internal
combustion engine vehicles (ICEVs) to EVs while holding Chicago’s destination and mode
1
Within the Census Bureau’s defined Chicago–Naperville–Elgin, IL–IN–WI metropolitan statistical area (MSA),
9.2% of households used taxi services at least a few times a month versus 2.8% using this mode a few times a week.

4
choices constant (a fuel-switching-only scenario). Then a range of scenarios are used to understand
how shared, fully-automated, all-electric vehicles (SAEVs), shared rides (with strangers), and
smart-charging price signals mitigate increases in power costs for everyone, while lowering per-
passenger mile emissions and excess investments in household EVs and chargers. The scenarios
simulate the transportation-energy transition in 2035 under different SAEV mode splits within a
2,360 square-mile geofence (service area). Seasonal effects are captured through EV energy
consumption, non-transportation power demand (i.e., load), and VRE generation to provide robust
results. The results reveal planning insights for future all-electric mobility solutions, like SAEVs,
and capture the interdependent relationship between additional EV charging load on wholesale
power prices and smart charging decisions.
The following sections describe the modeling framework, assumptions made, data sets used, and
charging strategies applied, before showing results and conclusions, along with recommendations
for policymakers and future research.
2. METHODS
The study’s first contribution is a framework for integrating power systems operation model and
a discrete-event transportation simulator. The transportation simulation covers the 20-county
Greater Chicago region and simulates a 24-hour period of typical weekday travel demand,
including urban, freight, and commercial vehicle trips that are internal or external to the region.
Charging demand was simulated for the region based on a day’s typical travel and assumptions on
charging frequency and access to charging stations. The underlying vehicle ownership choice
model, calibrated to present-day data and demographics (like household income), was adjusted to
increase the probability of choosing an electric powertrain vehicle. Charging control scenarios,
including different electric utility rates (i.e., prices) and charging equipment timers to delay
charging, were used to compare them to unmanaged charging. The total charging demand was
added to the baseload electricity demand curve. The power system model uses unit commitment
and economic dispatch to find the least-cost set of generators to dispatch to meet electricity demand
at all hours of the day while respecting transmission constraints. The model also determines the
marginal generation cost of each node at the transmission network, serving as the electricity
wholesale prices.
The workflow for this at-scale power analysis is shown in Fig. 1. The assumption taken was that
a large fleet of SAEVs may be price makers (or at least price influencers), with sufficient aggregate
charging demand to affect generation costs (and eventually electricity prices) in the power model.
The respective scenario inputs for personal EVs and SAEVs are shown in grey boxes on the left-
hand side. The results from the power system model (e.g., costs and generator status) are returned
to the transportation simulation if electricity prices are indexed to the wholesale market.

5
Fig. 1: Workflow of the at-scale power analysis.
2.1 Electricity Consumption
The state of Illinois has a goal of registering 1 million EVs in the state by 2030 (IDOT, 2022). An
estimated 10.4 million private and commercially owned automobiles and trucks were registered in
Illinois in 2017 (FHWA, 2021). This study simulates two on-road adoption levels for 2035: 8%
and 17% adoption, which would require a high EV sales share through 2035
2
. Since most early
adopters of EVs are wealthy residents and urban dwellers (Lee et al., 2019), these targets for the
Greater Chicago region may be attainable with accelerated vehicle turnover rates. Using adoption
targets and choice models to explain travel patterns, vehicle ownership, and charging options, the
transportation simulation tool, POLARIS, keeps track of vehicle movements throughout the
region’s travel network to reflect heterogeneity in energy consumption based on traffic levels. The
traffic flow model with dynamic traffic assignment provides the travel records for the region (Auld
et al., 2016). A machine-learning algorithm predicts energy consumption based on vehicle
trajectories across the synthetic road network (Moawad et al., 2021) and updates vehicle battery
levels throughout the simulation.
To estimate the effects of widespread EV adoption and increasing reliance on SAEVs, the
aggregate charging load is added to the non-transportation (or baseload) consumption profile.
Population growth, economic output, and cross-sector decarbonization (namely, switching from
fuel to electric for heating) will also increase baseload consumption of electricity. This study
applies a year-over-year growth factor that varies by month and hour based on estimates from the
National Renewable Energy Laboratory (NREL) Cambium model (Gagnon et al., 2023). The
average growth factor across twelve months and twenty-four hours is 1.19%, which is less than
the 1.23% assumed in the well-cited 2050 medium-rapid electrification scenario in Mai et al.
2
In Norway, the 8% personal vehicle fleet share was reached by 2019, when over 50% of new vehicles sold were
EVs. It is projected that 17% fleet share will be reached in 2022, when 80% of new vehicles sold will be EVs (Norsk
elbilforening, 2022).

6
(2018). See Fig. A.5 in the appendix for the monthly- and hourly-varying demand growth factors
used in this study.
The grid model, A-LEAF, is a combined least-cost unit commitment and economic dispatch model
written in Python that models the generation and transmission of power (Verbas and Zhou, 2020).
A mixed-integer linear program determines the commitment status of generators to balance supply
and demand for each hour of the day (i.e., unit commitment), followed by a sub-problem that
solves a linear program for power levels from each committed generator (i.e., economic dispatch).
This study used 2015 generator data to compare a business as usual (BAU) case and a predicted
future with coal retirement and VRE capacity generation exceeding natural gas, since wind and
solar capacity factors are less than those for thermal power plants (U.S. EIA, 2022b). The increase
in VRE can also reduce the cost of power, as shown in the merit order curves for the two feedstock
scenarios (see Figs. A.6 and A.7 in the Appendix).
2.2 Charging Strategies
Demand-side flexibility is critical to support VRE and avoid costly upgrades owing to steeper
demand for electricity. Thus, many power companies are using pricing, such as time-of-use (TOU)
rates, to avoid a co-incident peak in electricity demand (SEPA, 2019). To take advantage of these
off-peak prices, vehicles or charging equipment use timers to delay or stagger charging (if devices
are connected and the power company can shift timing by a few minutes). In contrast, fleets of
SAEVs will seek to minimize electricity purchasing costs and downtime of vehicles and increase
the utilization of chargers to avoid excess investments in infrastructure. Using wholesale power
prices to incentivize off-peak charging or TOU rates is expected instead of timer-based solutions.
This study uses an optimization-based control strategy that relays prices into within-day idle
vehicle dispatch decisions (Dean et al., 2023). Interested readers are referred to Dean et al. (2023)
for details on the methodological framework.The flat electricity price in this study, $76.36/MWh,
is based on the bundled rate available for medium-load commercial customers with secondary
voltage connections in ComEd’s service region (ComEd, 2022a). The study assumes the
residential TOU rate can apply to commercial customers, of which there are three time-varying
prices
3
: off-peak prices apply between the hours of 10 PM and 6 AM ($25.95/MWh), super-peak
prices apply between 2 PM and 7 PM ($68.58/MWh), and peak prices between 6 AM and 2 PM
and again between 7 PM and 10 PM ($34.19/MWh) (ComEd, 2022b). Customers can also choose
to pay a wholesale-indexed rate plan without any markup on the rate, which can reduce the
customer’s average power bill. Since the additional demand from EVs will require an additional
supply of power, the prices of the wholesale market cannot be assumed to be stable, as in Dean et
al. (2023). As a result, the stable wholesale price found in the coupled power-transportation model,
shown in Fig. 2, is used in the study results.
In contrast to SAEVs, personal EVs charge between activities, including at-home and at-work
activities, provided there are chargers at these locations (i.e., “top-off” charging). Once a personal
EV arrives at a destination, a heuristic determines whether to create an unplanned charging trip. If
an EV’s expected energy consumption for the next planned trip decreases the state of charge (SOC)
below an individual’s absolute minimum threshold (i.e., a random draw between 3% and 10%),
3
As of April 18, 2023, for non-summer months. The characteristic summer day in this study used the following
approved rates for summer days: $33.50/MWh for off-peak hours, $80.10/MWh for super-peak hours, and
$42.50/MWh for peak hours.

7
the driver searches for a public charging station that minimizes the total detour and wait delay.
Although there are a growing number of EV-managed charging pilots and EV-specific TOU rates
(SEPA, 2021), this study assumes drivers of EVs are not influenced by prices.
2.3 Scenario Definitions
The study compares the transportation and power system impacts of personal EVs and SAEVs.
Scenarios include electricity prices, charging strategies, adoption percentages of new technology,
ride-sharing acceptance (for SAEVs), and seasonal effects. Seasonality impacts include adjusting
the energy consumption of EVs based on auxiliary loads and the baseline demand for electricity,
both of which are driven primarily by heating, ventilation, and air conditioning (HVAC). Personal
EV and SAEV electricity consumption changes are assumed to follow real-world all-electric taxi
performance. Hao et al. (2020) estimated a 3.3% increase in energy consumption during the
summer from HVAC use and a 30% increase in the winter due to poor performance in colder
temperatures.
Personal light-duty EV adoption scenarios include low (8%) and high (17%) EV ownership as well
as charging infrastructure supply. The public charging station siting scenarios are defined as low
(1 charger per 35 vehicles) and high deployment (1 charger per 13 vehicles). Home charging
availability was randomly assigned based on adoption targets of 61% in single-family dwellings
and 5% in multi-family dwellings, such that 41% of households with light-duty EVs had a home
charger in the Chicago metro.SAEV scenarios include intra-geofence mode shares of 5%, 10%,
15%, and 20%. The effect of pooling strangers via dynamic ride-sharing (DRS) to increase vehicle
occupancy is tested in three scenarios: the SAEV fleet does not offer pooled rides, some customers
are willing to share rides (Gurumurthy and Kockelman, 2020), and all customers are willing to
share rides. Wholesale-indexed electricity prices are used as the smart charging control strategy,
where SAEVs centrally coordinate charging to lower power bill costs. The scenarios include retail
prices (e.g., fixed and TOU) and wholesale prices from ALEAF. The combination of scenarios
results in 144 SAEV simulations (Fig. 2).
Fig. 2: Combination of SAEV sensitivity analysis scenarios.
DRS
0% Acceptance
Partial
Acceptance
100%
Acceptance
Mode Share
5%
10%
15%
20%
Electricity Prices
Fixed
Time-of-Use
(TOU)
Wholesale
Seasons
Winter
Spring
Summer
Fall

8
2.4 Network Information
The synthetic transportation network includes 48,400 links and 35,800 nodes across the 20-county
metro. The geofenced SAEV fleet is limited to a 2,360 sq-mile region, shown in Fig. 3a, and
accounts for nearly 80% of the region’s population. The transportation data was obtained from the
Chicago Metropolitan Agency for Planning (CMAP). Fleet-owned charging stations with a power
rating of 120 kW were generated during a warm-up simulation run to distribute charging stations
5 miles apart (Gurumurthy et al., 2022). The synthetic power grid includes 20 nodes, 45 links, and
250 representative generators, with data inputs obtained from Form EIA-860 generator data. The
size of the node, shown in Fig. 3b, represents the share of the average annual daily load of the
state. The nodes in Fig. 3b are also the locations of generators in the system. The transmission
lines, shown in black, transport power to ensure that each node’s demand is met with supply, based
on 2015 data. As a result, demand in 2035 may exceed supply in some nodes when local VRE
output drops and there is not enough transmission capacity to import electricity. The economic
power system model reports the presence of congestion in the system through prices that exceed
the marginal cost curves of the power plants. This economic measure is designed to incentivize
demand response and energy efficiency measures in the short term and grid capacity expansion in
the long term. Readers interested in the development of synthetic bulk power and transmission
system networks are referred to Xu et al. (2017) and Birchfield et al. (2017).

9
a. Geofenced SAEV service area
b. Power grid node-link network
Fig. 3: (a) Chicago metro’s synthetic road model centered on the geofenced SAEV service area
with fleet-owned charging stations (orange circles, n=66) and co-located maintenance depots
(white stars on orange circles, n=18), and the (b) Illinois’ synthetic power grid where each node
shows the locations of the generation sources and demand sinks (purple circles) and transmission
lines (black lines) connect the generation sources to demand sinks.RESULTS
The joint framework aggregated personal EV charging demand from the 20-county Greater
Chicago metropolitan area. The first half of the results presents the case where fuel switching
through fast turnover of personal EV fleets leads to less on-road emissions, but no other actions
are taken in the avoid-shift-improve decarbonization framework (Creutzig et al., 2018). The
second half of the results presents the case where the presence of on-demand SAEVs reduces
household vehicle ownership, leading to mode shift behavior and some avoided travel. There are
different scenarios where the fleet offers pooled rides with strangers, and either some or all
passengers are willing to share. Like the low and high personal EV adoption scenarios, the results
show different levels of SAEV mode share within the 2,360 sq-mile geofence to understand the
at-scale impacts of fleet electrification.
3.1 Transition to Personal Electric Vehicles
Personal EVs are assumed to charge as needed (i.e., unmanaged), leading to more EVs charging
later in the day. In other words, some motorists opportunistically charge their EVs at work and
public charging stations, while most EV owners charge after returning home. The baseline
demand
4
for electricity in Illinois varies by season (see Fig. 4). Colder months usually have a
morning and evening peak, while the summer has a single peak. The baseline demand in Illinois
4
Fig. 4 plots “baseline demand,” which is non-transportation demand minus generation from distributed rooftop solar
(seasonally adjusted), since the model does not consider distributed energy resources.

10
is highest in the summer
5
, and the demand curve peaks in the afternoon and early evening hours.
The peak demand for a characteristic 2035 summer day is 35.0 GW between 4 and 5 PM (high
public charging scenario). The additional demand for electricity from Chicago metro EVs can
increase the summer peak demand by 320–580 MW (low vs. high public charging scenario),
assuming unmanaged charging and drivers always charge when given the opportunity (i.e., “top-
off” charging behavior). In contrast, the winter peak demand between 6 and 7 PM may increase
by 410–760 MW. This change in electricity demand stems from worse performance in the winter
rather than seasonal changes in activities. Fig. 4 displays the new demand curve for Illinois if just
the 20-county Chicago metro were to quickly phase in EVs by 2035, such that the personal fleet
of LDVs was 17% EVs. The first column shows results for a high number of public charging
stations, and the second column reports results for a low number of charging stations. At 17% EV
adoption, unmanaged charging can increase the summer peak demand by 0.9% to 1.6% and the
new winter peak demand by 1.2% to 2.2%
6
.
The increase in peak demand does not necessarily guarantee a steep increase in the cost of
producing electricity. The marginal difference in wholesale energy prices due to Chicago personal
EV adoption depends on the season’s baseload, VRE output, and the additional demand from EV
charging. Wind output changes hourly and seasonally, resulting in a different supply curve. The
true cost of electricity in real-time is found at the intersection of the demand curve and the supply
curve (i.e., market clearing price or spot price). The shape of the supply curve is determined by
the merit order—the least-cost sequence of power plant output capacity based on the marginal cost.
VRE, with no fuel costs and very low operating costs, can shift the supply curve left or right,
depending on output (see Figs. A.3 and A.4 in the Appendix for the 2015 and 2035 merit order
plots for the state of Illinois). To a lesser extent, the new electricity demand from EVs can shift
the spot price by altering the operations of thermal power plants, which have minimum up/down
times and ramping constraints.
A characteristic winter day in 2035 is likely to see a spike in wholesale prices between 5 and 7
PM, while summer days see a spike at 6 PM due to additional daylight for solar generation. EV
charging, if left unmanaged, can increase peak demand during the natural ramp-up in the net load
curve. EV charging at home usually occurs at a lower power draw using Level 2 chargers; however,
at 17% EV adoption, this additional peak demand is spread out over more hours and may lead to
peak prices for an additional n hours. This study finds that winter prices may spike between 5 and
8 PM and summer prices may spike between 6 and 9 PM (or an increase in high prices for 1 hour
and 2 hours, respectively).
5
According to the 2020 Residential Energy Consumption Survey (RECS) and the 2019 American Community Survey,
around 95% of Illinois households use electric air conditioning, while less than 20% have an electric heating system
(U.S. Energy Information Administration, 2022c).
6
About half of the state’s population lives within the Greater Chicago metro area. Assuming that 20% personal EV
adoption increases peak demand by 1-2%, a 100% EV scenario might increase peak demand by 10%-20%.

11
Note: Assuming 17% fleet share and two public charging station density scenarios.
Fig. 4: Change in Illinois’ projected 2035 power demand curve due to Chicago’s unmanaged
personal EVs charging in winter, spring, summer, and fall.

12
The difference in projected statewide wholesale energy prices due to personal EV charging is
plotted in Fig. 5 using LOESS regression, where the solid purple line is the average change in price
from the four scenarios (2 EV adoption levels times 2 charging access scenarios). The shaded grey
region is a 95% confidence interval to account for the high uncertainty in energy price fluctuations
in 2035. The statewide wholesale electricity prices will almost certainly increase in the summer
and winter months in the evening hours (Fig. 5), when a peak in electricity demand coincides with
a drop in solar production. Spring and fall days, with a moderate temperature range, are less likely
to see large fluctuations in energy prices due to unmanaged EV charging. Figs. A.1–A.4 in the
appendix provide individual joint price and load curves without smoothing data, hence why there
are large spikes in prices in the summer and winter months. If transmission capacity is expanded
to reduce congestion, then the true increase in price due to the merit order of generators may more
closely resemble the projected increase in Fig. 5.
Note: Y-axis scale differs for each subfigure. LOESS regression-generated mean line in purple with 95% confidence
interval band captures wide estimate from 4 data points for each hour (2 EV adoption levels x 2 public charger
scenarios).
Fig. 5: Average change in Illinois’ projected 2035 hourly wholesale electricity prices due to
Chicago personal EV adoption by season.
3.2 Shared Autonomous Electric Vehicles
SAEV fleet operators can make centralized charging decisions and are flexible to price signals that
incentivize off-peak charging. The study reports findings from scenarios that used a fixed price,

13
which is equivalent to unmanaged charging, and two price-based managed charging strategies:
TOU and wholesale-indexed prices. Table 1 reports mobility and vehicle statistics for an SAEV
fleet under different mode share levels, DRS acceptance, and electricity prices for a characteristic
spring or fall day (where air temperatures do not increase vehicle energy consumption or increase
HVAC loads on the state’s electrical grid). Table 2 and Table 3 report results for the same metrics
for the characteristic summer and winter days, respectively.
The average SAEV drove between 290 and 310 miles each day (across all mode share scenarios),
with empty VMT adding upwards of 3 million VMT for each 5% increase in mode share. On
average, 40% of the VMT driven by each vehicle is without a passenger (i.e., empty VMT
[eVMT]). Higher duty cycles by fleet EVs lead to more energy consumption on a per-vehicle basis
compared to personal EVs. On average, an SAEV that completes over 22 person-trips per day will
require between 85 and 110 kWh
7
of electricity (i.e., the average person-trip consumes 3.9 to 5
kWh, including the additional eVMT consumption to support this trip; see Fig. A.9). In contrast,
the average personal EV requires less than 12 kWh of electricity each day to operate. The average
Chicago person-trip in a personal EV consumes 7.7 to 9.3 kWh, which is 1.9 times more energy
than with a fleet of shared SAEVs.
The average vehicle occupancy (AVO) across all vehicle-miles (revenue and non-revenue) for this
fleet of SAEVs ranges from 1.4 (5% mode share with no DRS) to 1.8 (20% mode share with DRS).
The revenue-miles adjusted AVO metrics are shown in the results tables, which aligns with public
data on driver-reported passenger counts in similar cities (i.e., New York City Taxi and Limousine
Commission’s trip dataset). If fleets offer a DRS option the expected revenue-miles AVO may
increase from 2.5 to 2.8. Although percent eVMT (%eVMT) is about 40% across all scenarios,
DRS scenarios have about a three-percentage point decrease. If no rides are shared, almost two-
thirds of empty travel occurs between customers. As DRS reduces deadheading between trip ends,
the relative share of support trips (e.g., charging, maintenance, cleaning, and rebalancing vehicles)
will increase from a third to over 40%. If riders take SAEVs in group settings or solo travelers are
willing to share rides, the congestion impact of large-scale SAEV service may be minimized. Even
with a high empty VMT, sharing vehicles, rides, and centralized dispatch of vehicles to passengers
with DRS can lead to more efficient energy use relative to the status quo of driving one’s personal
vehicle in sprawling regions, like Chicago.
To encourage households to give up their vehicles and use on-demand SAEV service, fleets should
minimize the percentage of unmet trips. A trip is declared unmet if a customer waits more than 18
minutes before their smartphone app notifies them of their vehicle assignment. The results in Table
1 indicate that fleet operators adjusting charging decisions to lower the fleet’s electric bill, either
through a set retail TOU rate or a wholesale-indexed rate, does not decrease the number of trips
served per vehicle per day, relative to the unmanaged case of flat prices. Aligning charging with
low-cost periods of electricity does not come at the expense of serving passenger trips. Further,
there is evidence that a fleet minimizing electricity expenditures could decrease percent empty
travel (Fig. A.10).
In comparison to Fig. 4, which shows an increase in the load curve due to unmanaged personal EV
charging, Fig. 6 shows the new load curve in the spring if between 5% and 20% of daily Chicago
7
Assuming a 33% increase in energy consumption during the winter (see methods). See Fig. A.8 for a boxplot of
daily energy consumption per SAEV by season.

14
trips were served by a fleet of SAEVs. These figures assume that all riders are willing to share
rides, which decreases energy consumption on a per-trip basis (Fig. A.9). Subfigures on the left
assume the region’s electricity provider does not charge time-varying electricity prices, while
subfigures on the right assume the fleet pays the prevailing wholesale price and adjusts charging
decisions based on day-ahead energy prices. At 20% SAEV mode share, the state’s springtime
peak demand for electricity could shift from 1 PM to 8 PM (representing a 7.9% increase in
demand at 8 PM) under flat energy prices. However, a fleet paying wholesale prices will try to
avoid charging in the evening hours when energy prices are higher, thereby shifting charging
earlier in the day and increasing the 1 PM peak by 8.9%. Shifting charging may also increase
empty travel if vehicles are not near charging stations and charging station locations are far from
trip pick-ups. The simulation results suggest that with uniform charger placement (see Fig. 3a),
the percent of empty VMT does not necessarily increase with time-varying energy prices (Fig.
A.10).
Shifts in EV charging patterns can lead to price fluctuations in statewide wholesale energy prices.
Fig. 7 plots projected price changes from a Chicago fleet serving 20% of trips in 2035, using
LOESS regression to reduce the magnitude of large price increases during hours where demand
exceeds supply due to transmission constraints. Prices will increase across all seasons due to an
increase in charging from SAEV adoption, which was not observed in the personal EV scenarios
(Fig. 5). Spring days have relatively low prices due to less electricity demand, but due to an
increase in charging after midnight (Fig. 6), the power grid operator dispatches additional
generators to meet this new demand, raising prices for all. However, if this charging demand
coincides with the midday peak, prices may rise substantially more than the midnight price
increase of nearly $13/MWh, indicating that smart charging decisions can mitigate price increases.
At the same time, there is an opportunity for more valley filling after midnight (Fig. 6), which
could reduce the expected price increase.

15
Table 1: SAEV Mobility and Vehicle Statistics by Mode Share, DRS Acceptance, and Electricity Price for a Spring (or Fall Day)
Metrics
Unmanaged Charging: Flat Electricity Prices (Spring/Fall)
SAEV (No DRS)
SAEV + DRS Acceptance Model
SAEV + 100% DRS Acceptance
Mode Share
5%
10%
15%
20%
5%
10%
15%
20%
5%
10%
15%
20%
Avg. Response Time
(min)
5.0 min
4.7 min
4.9 min
5.9 min
4.3 min
4.1 min
4.2 min
4.8 min
4.3 min
4.1 min
4.1 min
4.3 min
Avg Person-Trips/
SAV/day
24.3
trips
24.4
trips
22.7
trips
21.2
trips
26.3
trips
26.6
trips
26.1
trips
23.9
trips
26.2
trips
26.6
trips
26.3
trips
25.1
trips
Avg. SAV VMT/day
292 mi
308 mi
303 mi
298 mi
285 mi
301 mi
303 mi
300 mi
284 mi
300 mi
304 mi
295 mi
%eVMT
43.6%
41.4%
42.2%
43.0%
40.6%
38.0%
38.2%
40.0%
40.5%
37.9%
37.8%
37.9%
Charging
(kWh/SAV/day)
81.5
kWh
86.4
kWh
84.0
kWh
71.2
kWh
79.1
kWh
86.9
kWh
84.2
kWh
70.5
kWh
80.2
kWh
84.6
kWh
84.9
kWh
76.5
kWh
AVO
2.50
2.50
2.50
2.50
2.69
2.74
2.77
2.81
2.69
2.74
2.77
2.79
Metrics
Managed Charging: Time-of-Use (TOU) Electricity Prices (Spring/Fall)
SAEV (No DRS)
SAEV + DRS Acceptance Model
SAEV + 100% DRS Acceptance
Mode Share
5%
10%
15%
20%
5%
10%
15%
20%
5%
10%
15%
20%
Avg. Response Time
(min)
5.0
4.9
6.2
4.8
4.4
4.2
4.9
4.3
4.3
4.1
4.1
4.3
Avg Person-Trips/
SAV/day
25.0
24.7
21.7
23.5
27.1
27.1
24.9
25.7
26.2
27.3
26.3
25.5
Avg. SAV VMT/day
300
313
305
310
295
307
302
302
284
307
304
301
%eVMT
43.8%
41.5%
44.0%
41.2%
40.8%
38.0%
39.9%
37.7%
40.5%
37.6%
37.9%
38.1%
Charging
(kWh/SAV/day)
80.3
87.9
82.4
81.7
79.8
87.8
82.1
79.5
79.0
87.9
80.2
80.3
AVO
2.50
2.50
2.50
2.50
2.70
2.74
2.79
2.80
2.69
2.74
2.77
2.79
Metrics
Managed Charging: Wholesale-Indexed Electricity Prices, Endogenously Derived (Spring)
SAEV (No DRS)
SAEV + DRS Acceptance Model
SAEV + 100% DRS Acceptance
Mode Share
5%
10%
15%
20%
5%
10%
15%
20%
5%
10%
15%
20%
Avg. Response Time
(min)
5.3
5.0
5.3
4.8
4.5
4.2
4.2
4.5
4.5
4.1
4.2
4.2

16
Avg Person-Trips/
SAV/day
24.3
24.2
23.5
23.4
24.7
27.1
26.3
26.3
24.5
26.9
27.0
26.5
Avg. SAV VMT/day
296
306
317
311
291
300
310
305
270
300
305
307
%eVMT
44.4%
41.2%
42.9%
40.9%
41.5%
37.1%
39.1%
39.9%
41.8%
37.1%
37.5%
37.7%
Charging
(kWh/SAV/day)
92.6
84.0
83.7
81.9
87.3
84.3
81.6
81.0
73.2
85.0
80.2
81.8
AVO
2.50
2.50
2.50
2.50
2.70
2.74
2.77
2.80
2.69
2.74
2.78
2.80
Metrics
Managed Charging: Wholesale-Indexed Electricity Prices, Endogenously Derived (Fall)
SAEV (No DRS)
SAEV + DRS Acceptance Model
SAEV + 100% DRS Acceptance
Mode Share
5%
10%
15%
20%
5%
10%
15%
20%
5%
10%
15%
20%
Avg. Response Time
(min)
5.2
4.9
4.9
4.8
4.5
4.1
4.2
4.1
4.3
4.3
4.4
4.3
Avg Person-Trips/
SAV/day
24.6
23.2
24.3
23.6
26.9
27.3
26.8
26.7
27.2
26.9
26.3
24.7
Avg. SAV VMT/day
299
301
319
312
290
302
313
307
294
305
303
291
%eVMT
44.1%
40.9%
41.8%
41.4%
41.7%
37.3%
38.2%
37.4%
40.5%
38.2%
37.9%
38.9%
Charging
(kWh/SAV/day)
85.4
81.2
83.8
82.9
75.6
88.1
76.3
79.4
83.4
87.7
80.2
69.4
AVO
2.50
2.50
2.50
2.50
2.70
2.74
2.78
2.80
2.69
2.75
2.79
2.79
Note: Fleet size does not linearly scale with mode share within the geofence. The ratios of SAEVs to residents (assuming 5% to 20% mode shares) are 1:300, 1:140,
1:105, and 1:75, respectively. AVO = revenue-miles weighted average vehicle occupancy (accounting for party size), & DRS = dynamic ride-sharing (or pooling).
Table 2: SAEV Mobility and Vehicle Statistics by Mode Share, DRS Acceptance, and Electricity Price for a Summer Day
Metrics
Unmanaged Charging: Flat Electricity Prices (Summer)
SAEV (No DRS)
SAEV + DRS Acceptance Model
SAEV + 100% DRS Acceptance
Mode Share
5%
10%
15%
20%
5%
10%
15%
20%
5%
10%
15%
20%
Avg. Response Time
(min)
5.0
4.7
4.8
4.7
4.4
4.2
4.3
4.0
4.3
4.2
4.3
4.5
Avg Person-Trips/
SAV/day
24.3
24.1
23.6
23.7
26.2
26.6
26.7
26.7
26.2
26.5
26.5
25.2

17
Avg. SAV VMT/day
292
306
303
314
285
302
303
308
284
302
302
300
%eVMT
43.8%
41.6%
41.6%
41.2%
40.7%
38.3%
36.5%
37.1%
40.6%
38.5%
36.6%
38.4%
Charging
(kWh/SAV/day)
80.5
86.3
83.4
101.1
79.2
89.6
90.2
101.1
79.1
89.5
90.2
93.2
AVO
2.50
2.50
2.50
2.50
2.69
2.74
2.78
2.80
2.69
2.74
2.78
2.50
Metrics
Managed Charging: Time-of-Use (TOU) Electricity Prices (Summer)
SAEV (No DRS)
SAEV + DRS Acceptance Model
SAEV + 100% DRS Acceptance
Mode Share
5%
10%
15%
20%
5%
10%
15%
20%
5%
10%
15%
20%
Avg. Response Time
(min)
5.1
4.9
5.0
4.6
4.3
4.2
4.1
4.5
4.3
4.2
4.3
4.0
Avg Person-Trips/
SAV/day
25.0
24.7
23.2
23.9
27.3
27.2
26.6
25.2
26.2
26.5
26.5
26.6
Avg. SAV VMT/day
301
314
304
316
295
308
302
304
284
302
302
309
%eVMT
44.0%
41.8%
40.8%
41.0%
40.6%
38.1%
36.4%
38.8%
40.6%
38.5%
36.6%
37.4%
Charging
(kWh/SAV/day)
80.2
90.5
91.5
101.6
79.8
90.8
90.6
94.2
79.1
89.5
90.2
101.7
AVO
2.50
2.50
2.50
2.50
2.70
2.75
2.77
2.81
2.69
2.74
2.78
2.80
Metrics
Managed Charging: Wholesale-Indexed Electricity Prices, Endogenously Derived (Summer)
SAEV (No DRS)
SAEV + DRS Acceptance Model
SAEV + 100% DRS Acceptance
Mode Share
5%
10%
15%
20%
5%
10%
15%
20%
5%
10%
15%
20%
Avg. Response Time
(min)
5.2
4.9
5.8
4.9
4.4
4.3
4.7
4.3
4.3
4.2
4.4
4.3
Avg Person-Trips/
SAV/day
24.6
24.9
21.4
23.9
26.9
26.7
26.9
26.1
27.2
26.8
26.5
25.4
Avg. SAV VMT/day
298
311
286
310
296
305
284
279
294
301
302
304
%eVMT
44.0%
40.8%
41.9%
40.0%
41.3%
41.3%
37.5%
37.5%
40.6%
37.6%
37.0%
36.2%
Charging
(kWh/SAV/day)
88.5
89.0
76.4
100.3
85.3
91.0
82.2
76.4
82.6
85.2
102.2
99.4
AVO
2.50
2.50
2.50
2.50
2.70
2.75
2.79
2.80
2.69
2.74
2.78
2.81

18
Note: Fleet size does not linearly scale with mode share within the geofence. The ratio of 1 SAEV to residents from 5% to 20% mode share is 1:300, 1:140, 1:105,
and 1:75, respectively. Abbreviations: AVO = revenue-miles weighted average vehicle occupancy (accounting for party size), DRS = dynamic ride-sharing (or
pooling).
Table 3: SAEV Mobility and Vehicle Statistics by Mode Share, DRS Acceptance, and Electricity Price for a Winter Day
Metrics
Unmanaged Charging: Flat Electricity Prices (Winter)
SAEV (No DRS)
SAEV + DRS Acceptance Model
SAEV + 100% DRS Acceptance
Mode Share
5%
10%
15%
20%
5%
10%
15%
20%
5%
10%
15%
20%
Avg. Response Time
(min)
6.2
5.8
6.1
5.6
5.1
4.6
4.6
4.4
5.1
4.6
4.6
6.4
Avg Person-Trips/
SAV/day
22.2
22.8
21.9
22.3
24.8
25.6
26.1
25.6
24.7
25.5
26.1
19.6
Avg. SAV VMT/day
288
305
303
310
284
299
305
306
285
299
304
261
%eVMT
46.9%
44.3%
43.4%
43.8%
43.5%
40.2%
38.5%
39.3%
43.6%
40.2%
38.4%
42.4%
Charging
(kWh/SAV/day)
97.9
105.7
107.0
121.6
98.1
108.3
109.6
121.8
99.4
108.4
113.5
93.2
AVO
2.50
2.50
2.50
2.50
2.70
2.75
2.78
2.80
2.70
2.75
2.78
2.82
Metrics
Managed Charging: Time-of-Use (TOU) Electricity Prices (Winter)
SAEV (No DRS)
SAEV + DRS Acceptance Model
SAEV + 100% DRS Acceptance
Mode Share
5%
10%
15%
20%
5%
10%
15%
20%
5%
10%
15%
20%
Avg. Response Time
(min)
6.2
6.1
6.2
6.0
5.3
5.1
4.7
4.8
5.1
4.6
4.6
4.9
Avg Person-Trips/
SAV/day
22.6
22.4
22.8
21.9
24.8
25.0
25.3
25.2
24.7
25.5
26.1
25.1
Avg. SAV VMT/day
291
304
311
310
287
299
299
305
285
299
304
305
%eVMT
46.7%
44.6%
43.7%
44.4%
43.8%
41.2%
38.9%
40.3%
43.6%
40.2%
38.4%
40.4%
Charging
(kWh/SAV/day)
96.3
103.9
114.7
119.8
95.8
104.3
113.8
120.2
98.2
108.4
113.5
119.7
AVO
2.50
2.50
2.50
2.50
2.70
2.75
2.78
2.82
2.70
2.75
2.78
2.82
Metrics
Managed Charging: Wholesale-Indexed Electricity Prices, Endogenously Derived (Winter)
SAEV (No DRS)
SAEV + DRS Acceptance Model
SAEV + 100% DRS Acceptance

19
Mode Share
5%
10%
15%
20%
5%
10%
15%
20%
5%
10%
15%
20%
Avg. Response Time
(min)
6.3
6.5
6.4
5.9
5.2
5.1
4.9
5.0
5.2
5.3
4.9
4.9
Avg Person-Trips/
SAV/day
22.0
22.0
22.9
22.2
24.8
25.1
24.6
18.9
24.4
24.1
25.4
23.8
Avg. SAV VMT/day
291
293
300
304
284
293
318
280
285
288
298
283
%eVMT
47.0%
44.6%
44.5%
43.0%
44.0%
40.9%
41.2%
39.3%
44.1%
41.2%
39.2%
39.4%
Charging
(kWh/SAV/day)
106.0
100.4
107.2
117.5
102.9
105.8
125.8
111.3
103.5
103.7
112.4
103.7
AVO
2.50
2.50
2.50
2.50
2.70
2.76
2.79
2.81
2.70
2.76
2.79
2.81
Note: Fleet size does not linearly scale with mode share within the geofence. The ratio of 1 SAEV to residents from 5% to 20% mode share is 1:300, 1:140, 1:105,
and 1:75, respectively. Abbreviations: AVO = revenue-miles weighted average vehicle occupancy (accounting for party size), DRS = dynamic ride-sharing (or
pooling).

20
Note: EV demand (purple) is added to the non-transportation demand (blue). Assuming 100% DRS acceptance.
Fig. 6: Change in Illinois’ projected 2035 springtime power demand profile due to Chicago
SAEVs charging, with and without wholesale prices (right and left panels), across different
SAV mode splits.

21
Note: Y-axis scale differs for each subfigure. LOESS regression-generated mean line in purpose with 95%
confidence interval band captures wide estimate from 12 data points for each hour (3 DRS scenarios x 4
feedback-loops per scenario).
Fig. 7: Average change in Illinois’ projected 2035 hourly wholesale electricity prices due to
a fleet of Chicago SAEVs charging by season, assuming 20% mode share.
4. DISCUSSION
4.1 Load Shape Impacts
The study finds that even at low EV ownership levels, like when 8% and 17% of personal
LDVs are EVs in 2035, the unmanaged charging demand from the 20-county Chicago metro
can increase the peak demand for the state by up to 2.2%. The added electricity demand
depends on weather conditions, with winter having the highest relative increase in peak
demand and does not consider EVs outside of this metro. If personal EVs charge between
activities or immediately when the driver returns home, assuming chargers are available, the
power grid must ramp up generation by 410–760 MW in the winter and 320–580 MW in the
summer during the maximum 1-hour peak in electricity demand.
At 20% mode share, a fleet of Chicago SAEVs can increase a spring day’s peak hour (at 1
PM) electricity demand by 2,100 MW. However, this additional increase in demand is met
with existing energy resources since the grid is designed around the annual peak (usually in
winter or summer). In winter, the 20% fleet scenario could increase peak hour demand (at 6
PM) by 1,800 MW (with DRS) to 2,400 MW (without DRS). Since the fleet expects higher
energy consumption during winter months and higher energy prices at peak hours, the

22
magnitude of the increase in energy demand during peak hours may be less in winter than
during more temperate months. On the other hand, higher non-transportation demand for
electricity and the additional demand from SAEVs may lead to a higher marginal increase in
wholesale prices than in the spring or fall (Fig. 7). This study finds the marginal increase
during temperate months may be as high as +$25/MWh, but summer and winter months may
see price increases of $250–$750/MWh for about five hours in the evening (adjusting down
in price to forecast transmission expansion).
The new peak from EVs in temperate months could be supplied by existing power plants, as
indicated by the negligible increases in hourly wholesale prices (Fig. 5). However, if other
sectors increase their electricity demand and there are additional retirements of dispatchable
fossil fuel-powered generators, capacity expansion may be warranted, especially within the
Chicago region if import/export remains constant. The new peak demand may require one to
two additional natural gas-fired peaker power plants with an estimated base capital cost of
$958/kW for a single 1.0 GW plant (U.S. EIA, 2020). Alternatively, utility-scale battery
storage systems with a four-hour duration requirement (see Fig. 7) may meet this temporary
shortfall in supply, especially since these assets are cost-competitive with peaker power
plants.
4.2 Effects of Vehicle- and Ride-sharing
The introduction of on-demand urban passenger mobility may motivate some households to
give up some or all of their personal vehicles. The household vehicle discontinuance model
by Menon et al. (2019) was adopted for the Chicago region and suggests that up to a quarter
of its personal vehicle fleet may be retired because of this new mobility option. Encouraging
households to scrap old and less efficient vehicles can greatly improve the personal vehicle
fleet’s efficiency. Governments could encourage households to scrap their oldest and least
efficient vehicle, based on registration data at each address, and give usable credits for
mobility options like SAEVs, public transit, and micromobility. Although some households
receiving this credit will likely have scrapped their vehicle anyway (i.e., the free-rider effect),
they still contribute to an important reduction in on-road emissions.
The 8% and 17% personal EV fleet adoption levels would replace over 500 thousand to a
million internal combustion engine vehicles with electric vehicles. In contrast, serving 20%
of trips in the sprawling 2,360 sq-mile region (see Fig. 3a) is handled with a fleet of 198
thousand SAEVs. As a result of reliable and fast vehicle assignment and dispatch (< 5 min.
response times), households let go of over 300–800 thousand personal vehicles. The
embodied vehicle emissions savings are equal to the avoided production emissions of
personal EVs. Production emissions depend on the carbon intensity of the grid where the
battery materials were processed, assembled, and packaged into the battery pack and the
vehicle, as well as supply chain transportation emissions. Dillman et al. (2020) found that an
average BEV produces 10.8 tons of CO2 equivalent (CO2e) during the production phase,
while an average gasoline-fueled vehicle produces 6.6 tons of CO2e. At 20% mode share, the
embodied emissions saved due to shared fleet vehicles relative to 20% personal EV fleet
adoption is 11.4 million tons of CO2e. Relying on shared vehicles instead of private vehicles
in the Chicago metropolitan area is the equivalent of avoiding over a quarter of a million U.S.
households’ annual carbon footprint (Jones and Kammen, 2011). Shifting people away from

23
personal vehicles to shared vehicles by providing reliable and convenient alternatives is a
decarbonization strategy that provides immediate benefits.
And while charging infrastructure has a negligible climate impact relative to in-use emissions
(Mulrow and Grubert, 2023), shifting to shared vehicles reduces the number of public (and
private) chargers that are needed. This study used a ratio of 5.87 SAEVs per 120 kW fast
charger (or port) and assumed a high supply of public chargers to achieve a ratio of 6.22
personal EVs per 50 kW fast charger (at 17% personal EV fleet adoption). At 20% mode
share, shared vehicles could avoid adding about 200 thousand public ports across 6–15
thousand public charging stations (depending on station density). The emissions reductions
come from installing the cables, wiring, and digital equipment and create other co-benefits,
like reducing impervious pavement across communities and the environmental damages from
sourcing metallic components in charging equipment.
The study compared the benefits of ride-sharing at two acceptance levels—all customers
versus some customers who were willing to share rides—relative to no sharing. Offering a
ride-sharing option decreased the average daily VMT for each SAEV by 7 to 15 miles, with
only a small difference in benefits between 100% ride-share acceptance and a choice model
by Gurumurthy and Kockelman (2020). This reduction in daily VMT per vehicle is
significant at scale, since 20% mode share is served with a fleet of 105,000 vehicles.
Assuming a BEV emits 190 g CO2eq per mile in Illinois from in-use activity, a fleet offering
ride-sharing trips could avoid 40 to 86 tons of CO2eq every weekday when serving 5% of the
region’s trips (Energy Systems 2023). Energy consumption on a per-trip basis also declined
with DRS (Fig. A.9). If all customers were willing to share rides and the fleet used a
directionality-based matching heuristic (Gurumurthy and Kockelman, 2022), the per-trip
energy savings may range from 0.27 to 0.52 kWh. At 20% mode share, Chicago’s fleetwide
energy savings from a single day due to DRS may be as high as 54.6 MWh
8
.
4.3 Limitations and Future Research
The unit commitment and economic dispatch power system model, called A-LEAF, was
developed from 2015 data. This paper adjusted generator information to simulate 2035
feedstocks, based on NREL’s Cambium model with an assumed business-as-usual scenario
with no phase-out of production or investment tax credits and no nascent technology.
Adjustments included retiring coal and oil-gas steam power plants, but due to modeling
limitations the paper did not add utility-scale battery storage, which constitutes 6.1% of the
state’s projected generation capacity (or 6.28 GW). Despite these limitations, the paper
estimated that a 17% Chicago personal EV fleet scenario would increase winter peak demand
by at most 760 MW, which could be met by a new gas-fired peaker power plant or utility-
scale battery storage systems. However, the model predicted that electricity demand growth
in the summer and winter months would lead to high prices and possibly curtailed demand,
highlighting the need for transmission system expansion, beyond 2015 capacity, to meet
decarbonization goals.
8
This is equal to the average daily electricity consumption of 1,900 U.S. residential customers, according to
2021 data from the U.S. Energy Information Administration (EIA).

24
This paper did not consider energy imports from other states or transmission expansion. The
results indicate that these limitations underscore the importance of investing in transmission
system expansion and reducing barriers to upgrading or building new transmission
infrastructure. To provide insight into projected price changes in the state's wholesale market
from Chicago EVs, the paper uses LOESS regression, which can mimic price changes under
a limited build-out of transmission capacity because the modeling technique uses a weight
function to smooth out large differences in data.
There are several challenges of modeling into 2035, namely estimating the energy increase
from the electrification of transportation, building, and industrial sectors. For example, the
adoption of heat pumps and electrification of space heating will further increase the region’s
winter baseline demand curve, which may overlap with a higher EV load due to battery
performance issues in colder weather. Although wind generation performance is above the
yearly median in the winter months in Illinois, solar generation capacity is lower. Potential
shifts in peak demand from summer to winter will require power grid expansion planning
that considers the region’s seasonally adjusted VRE generation capacity, which is not within
the scope of this paper.
This paper acknowledges the difficulty of fully capturing the variability in VRE output by
hour and across days, as well as the variability in non-transportation load across days within
the season. Future work could use more characteristic days, as well as extreme winter and
summer days, to develop a more comprehensive view of the transition to a clean
transportation and power system. Nonetheless, the paper concludes that the results indicate
the value of using EVs to solve problems caused by EVs (via smart charging), especially
during summer and winter peak hours.
Finally, there are several other assumptions in this study that may impact results, including
uniform plug density at fleet-owned charging stations and the assumption of 50 kW public
chargers. Future work could use optimization-based charging station siting algorithms and
include heterogeneity in public chargers to develop more behaviorally realistic models for
charging station selection based on one's value of travel time, trip purpose, and flexibility in
the next activity.
5. CONCLUSION
This study aimed to investigate the impacts of vehicle electrification and new mobility
technology, like shared autonomous electric vehicles (SAEVs), on the power grid using an
integrated transportation-power system model. The power system model examined shifts in
energy demand and wholesale power prices in Chicago due to adoption levels of EVs,
seasonal effects on energy consumption, and electricity price rates. The results showed that
at relatively low EV penetration levels (8% to 17%), an increase in demand may require at
most 1 GW of additional generation capacity, and the state's transition to intermittent VREs
and phase-out of coal power plants will likely not substantially increase wholesale power
prices due to unmanaged personal EV charging at peak hours.
However, the simulation results found that wholesale power prices will increase during
winter (+$100/MWh) and summer (+$300/MWh) peak hours due to higher energy fees and
steep congestion fees on the 2015-era transmission system. Even without this new demand

25
for electricity from EVs, the model results indicate that prices would still spike due to
inadequate transmission capacity to send power from VRE sources to demand centers like
Chicago.
Additionally, a fleet of SAEVs serving 20% of Chicago's regional trips is more energy
efficient and avoids several hundred thousand vehicles' embodied emissions, but the higher
daily use of fleet vehicles and reliance on fast charging equipment increases electricity
demand and thus energy prices. Although a fleet paying wholesale prices uses these price
signals to reduce electricity demand during peak hours, spreading charging demand in hours
before and after the baseline peak creates new "ridges" in energy demand, which raises prices
for all.
The modeling framework and relative scale of the findings are relevant to both transportation
and power system audiences. The results reveal the pathways that result in expected emission
and cost reductions, indicating that investing in transmission lines can reduce congestion fees
in wholesale markets and address spatiotemporal imbalances in energy supply and demand.
At the same time, some energy customers, like EV fleets, could alter their charging behavior
to avoid adding to peak demand. Finally, ride-sharing with electric vehicles avoids new
embodied emissions from personal vehicles and charging equipment and can reduce energy
consumption on a per-trip basis, even with high empty travel between passengers and fleet-
owned maintenance and charging depots.
6. CRediT AUTHORSHIP STATEMENT
Matthew D. Dean: Conceptualization, Methodology, Software, Formal Analysis, Writing –
original draft, Visualization. Krishna Murthy Gurumurthy: Conceptualization,
Methodology. Zhi Zhou: Conceptualization, Software. Omer Verbas: Software. Taner
Cokyasar: Software. Kara M. Kockelman: Supervision. All authors reviewed and
approved the final version of the manuscript.
7. DECLARATION OF COMPETING INTEREST
The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this study.
8. ACKNOWLEDGEMENTS
This material is based upon work supported by the National Science Foundation Graduate
Research Fellowship Program under Grant No. DGE-1610403. Any opinions, findings, and
conclusions or recommendations expressed in this material are those of the author(s) and do
not necessarily reflect the views of the National Science Foundation.
The work done in this paper was sponsored by the U.S. Department of Energy (DOE) Vehicle
Technologies Office (VTO) under the Systems and Modeling for Accelerated Research in
Transportation (SMART) Mobility Laboratory Consortium, an initiative of the Energy
Efficient Mobility Systems (EEMS) Program. The U.S. Government retains for itself, and
others acting on its behalf, a paid-up nonexclusive, irrevocable worldwide license in said

26
article to reproduce, prepare derivative works, distribute copies to the public, and perform
publicly and display publicly, by or on behalf of the Government.
9. APPENDIX
9.1 Personal Light-duty Vehicle Electrification
Personal light-duty EV adoption levels and charging infrastructure are projected using
estimates from the National Renewable Energy Laboratory tool, Electric Vehicle
Infrastructure – Projection (EVI-Pro).
The effect of personal EV charging demand is studied across the four seasons at two fleet
share adoption levels and public charging infrastructure scenarios. Fig. A.1 adds the change
in market clearing price at the state level due to this shift in demand to the plots in Fig. 4 for
a characteristic spring day. The previous, or “baseline,” price line is in solid green while the
new prices are shown with a dotted green line. Figs. A.2 to A.4 show similar plots for
summer, fall, and winter seasons, respectively.
Note: EV demand (purple) is added to the baseload demand (blue) in the area plot. The market clearing price
from the baseload demand (solid green) is compared to the price with the additional EV demand (dashed
green) in the line plot.
Fig. A.1: Change in Illinois’ projected spring 2035 power demand curve overlaid with the
change in the market clearing price due to Chicago’s unmanaged personal EVs charging at
8% vs 17% household-EV fleet adoption and charging station density levels.

27
Note: EV demand (purple) is added to the baseload demand (blue) in the area plot. The market clearing price
from the baseload demand (solid green) is compared to the price with the additional EV demand (dashed
green) in the line plot.
Fig. A.2: Change in Illinois’ projected summer 2035 power demand curve overlaid with the
change in the market clearing price due to Chicago’s unmanaged personal EVs charging at
different EV fleet adoption and charging station density levels.

28
Note: EV demand (purple) is added to the baseload demand (blue) in the area plot. The market clearing price
from the baseload demand (solid green) is compared to the price with the additional EV demand (dashed
green) in the line plot.
Fig. A.3: Change in Illinois’ projected fall 2035 power demand curve overlaid with the
change in the market clearing price due to Chicago’s unmanaged personal EVs charging at
different EV fleet adoption and charging station density levels.

29
Note: EV demand (purple) is added to the baseload demand (blue) in the area plot. The market clearing price
from the baseload demand (solid green) is compared to the price with the additional EV demand (dashed
green) in the line plot.
Fig. A.4: Change in Illinois’ projected winter 2035 power demand curve overlaid with the
change in the market clearing price due to Chicago’s unmanaged personal EVs charging at
different EV fleet adoption and charging station density levels.
9.2 Power System Projections for Future-Year 2035
The demand for electricity is set to rise given population growth, economic outlooks, and the
electrification of end-uses (e.g., transportation vehicles, heating and cooling equipment, and
appliances). Instead of applying a uniform growth factor that does not account for additional
heating or cooling demands by season, the authors opted to use estimated generation data
from NREL’s Cambium tool. The estimated annual growth rate between 2015 and 2035 was
obtained and averaged by month and hour for all 8,760 hours of the year (Fig. A.5). Since
electricity generation must match demand at all hours, the growth factors are appropriate
even if transmission losses overestimate demand. While total demand is higher in the evening
hours in the summer months, the growth factor is higher in the morning hours, which might
capture the pre-cooling load-shifting behavior that is increasingly common with
programmable air conditioning systems.

30
Fig. A.5: Predicted electricity demand growth factors year-over-year for Illinois (by hour of
day and month of year, from NREL’s Cambium tool).
Power plants are dispatched in order of the marginal cost of producing power (i.e., merit
order). The baseline power system in 2015 is plotted in Fig. A.6, and the estimated 2035
system is plotted in Fig. A.7.
Note: Solar is not plotted due to its small generation capacity in 2015 but should appear to the left of wind.
Fig. A.6: Merit order curve for Illinois power grid in 2015.

31
Note: This plot ignores distributed rooftop solar, biofuel, and battery storage resources.
Fig. A.7: Predicted merit order curve for Illinois power grid in 2035.
The seasonal capacity factors for utility-scale wind and solar in Illinois are shown in Table
A.1 (U.S. EIA, 2022b).
Table A.1: Capacity Factors for Utility-Scale Wind and Solar.
Source
Month of the Year
1
2
3
4
5
6
7
8
9
10
11
12
Wind
33.8
39.7
46.8
39.6
34.0
24.2
15.0
17.1
31.5
33.2
46.0
45.9
Solar
7.3
11.7
20.3
24.5
23.5
27.5
24.4
26.6
24.5
15.0
14.2
10.0

32
Fig. A.8: Daily energy consumption per SAEV by mode share and season.
Note: Spring and fall have no energy consumption multiplier due to auxiliary loads and normal battery
performance in temperate weather, thus flat and TOU rates show no variability in fall.
Fig. A.9: Energy consumption per person-trip by season, energy price, and DRS scenario.

33
Fig. A.10: Empty travel by mode share and energy price.
Smart charging using price signals is most effective when fleets pay wholesale-indexed
energy rates. Fig. A.11 displays the expected change in statewide electricity prices in 2035
due to a fleet of SAEVs in Chicago serving 20% of daily trips. The existing rise in prices is
already accounted for and any large increase reflects an extension of already high prices.
Wholesale-indexed rates can avoid an extra $50/MWh increase in the evening peak.
Although TOU rates can lead to a decrease in prices in the morning hours the fleet’s charging
behavior with this structure leads to higher increases in state electricity prices than flat rates.

34
Fig. A.11: Change in state electricity prices due to additional electricity demand from a
Chicago SAEV serving 20% of daily trips in the winter.
Similarly, Fig. A.12 displays the change in state electricity prices for the summer months.
Wholesale-indexed rates are only marginally better than TOU or flat rates during the
afternoon through evening hours. If fleets ignore energy prices they may end up charging
more of their fleet vehicles overnight when trip demand is lower. However, unmanaged
charging at scale may lead to higher increases for all energy customers (+$300/MWh).
Fig. A.12: Change in state electricity prices due to additional electricity demand from a
Chicago SAEV serving 20% of daily trips in the summer.
10. REFERENCES
Anwar MB, Muratori M, Jadun P, Hale E, Bush B, Denholm P, et al. Assessing the value of
electric vehicle managed charging: a review of methodologies and results. Energy
Environ Sci 2022. https://doi.org/10.1039/D1EE02206G.
Auld J, Hope M, Ley H, Sokolov V, Xu B, Zhang K. POLARIS: Agent-based modeling
framework development and implementation for integrated travel demand and
network and operations simulations. Transportation Research Part C: Emerging
Technologies 2016;64:101–16. https://doi.org/10.1016/j.trc.2015.07.017.

35
Bailey J, Axsen J. Anticipating PEV buyers’ acceptance of utility controlled charging.
Transportation Research Part A: Policy and Practice 2015;82:29–46.
https://doi.org/10.1016/j.tra.2015.09.004.
Bauman J, Stevens MB, Hacikyan S, Tremblay L, Mallia E, Mendes CJ. Residential Smart-
Charging Pilot Program in Toronto: Results of a Utility Controlled Charging Pilot.
World Electric Vehicle Journal 2016;8:531–42.
https://doi.org/10.3390/wevj8020531.
Bibra EM, Connelly E, Dhir S, Drtil M, Henriot P, Hwang I, et al. Global Electric Vehicle
Outlook 2022: Securing supplies for an electric future. International Energy
Agency; 2022.
Birchfield AB, Xu T, Gegner KM, Shetye KS, Overbye TJ. Grid Structural Characteristics
as Validation Criteria for Synthetic Networks. IEEE Transactions on Power
Systems 2017;32:3258–65. https://doi.org/10.1109/TPWRS.2016.2616385.
Cheng AJ, Tarroja B, Shaffer B, Samuelsen S. Comparing the emissions benefits of
centralized vs. decentralized electric vehicle smart charging approaches: A case
study of the year 2030 California electric grid. Journal of Power Sources
2018;401:175–85. https://doi.org/10.1016/j.jpowsour.2018.08.092.
Clement-Nyns K, Haesen E, Driesen J. The Impact of Charging Plug-In Hybrid Electric
Vehicles on a Residential Distribution Grid. IEEE Transactions on Power Systems
2010;25:371–80. https://doi.org/10.1109/TPWRS.2009.2036481.
Coignard J, MacDougall P, Stadtmueller F, Vrettos E. Will Electric Vehicles Drive
Distribution Grid Upgrades?: The Case of California. IEEE Electrification
Magazine 2019;7:46–56. https://doi.org/10.1109/MELE.2019.2908794.
ComEd. Current Rates & Tariffs. Rates & Tariffs 2022a.
https://www.comed.com/MyAccount/MyBillUsage/Pages/CurrentRatesTariffs.aspx
(accessed May 10, 2022).
ComEd. Residential Time of Use Pricing Pilot Electricity Charges: Supplement to Rate
RTOUPP (1). Chicago, IL: 2022b.
Creutzig F, Roy J, Lamb WF, Azevedo IML, Bruine de Bruin W, Dalkmann H, et al.
Towards demand-side solutions for mitigating climate change. Nature Climate
Change 2018;8:260–3. https://doi.org/10.1038/s41558-018-0121-1.
Crossland C, Brakewood C, Cherry C. Investigating the Service of App-Based Rideshare
and Transportation Network Companies in Tennessee. Tennessee Department of
Transportation; 2021.

36
Dean MD, Kockelman KM. Are Electric Vehicle Targets Enough? The Decarbonization
Benefits of Managed Charging and Second-Life Battery Uses. Transportation
Research Record 2022.
Dean MD, de Souza F, Gurumurthy KM, Kockelman KM. Multi-stage charging and
discharging of electric vehicle fleets. Transportation Research Part D: Transport and
Environment 2023;118:103691. https://doi.org/10.1016/j.trd.2023.103691.
Dillman KJ, Árnadóttir Á, Heinonen J, Czepkiewicz M, Davíðsdóttir B. Review and Meta-
Analysis of EVs: Embodied Emissions and Environmental Breakeven.
Sustainability 2020;12:9390. https://doi.org/10.3390/su12229390.
Estandia A, Schiffer M, Rossi F, Luke J, Kara EC, Rajagopal R, et al. On the Interaction
between Autonomous Mobility on Demand Systems and Power Distribution
Networks -- An Optimal Power Flow Approach. IEEE Trans Control Netw Syst
2021;8:1163–76. https://doi.org/10.1109/TCNS.2021.3059225.
Fakhrmoosavi F, Gurumurthy KM, Kockelman KM, Hunter C, Dean MD. On- and Off-
Street Parking Strategies and Outcomes for Shared Autonomous Vehicle Fleet
Operations, Washington, D.C.: 2023.
Federal Highway Administration (FHWA). State Motor-Vehicle Registrations - 2017.
Highway Statistics 2017 2021.
https://www.fhwa.dot.gov/policyinformation/statistics/2017/mv1.cfm (accessed
October 5, 2022).
Forrest KE, Tarroja B, Zhang L, Shaffer B, Samuelsen S. Charging a renewable future: The
impact of electric vehicle charging intelligence on energy storage requirements to
meet renewable portfolio standards. Journal of Power Sources 2016;336:63–74.
https://doi.org/10.1016/j.jpowsour.2016.10.048.
Gagnon P, Cowiestoll B, Schwarz M. Cambium 2022 Scenario Descriptions and
Documentation. Golden, CO: National Renewable Energy Laboratory; 2023.
https://doi.org/10.2172/1915250.
Gurumurthy KM, Dean MD, Kockelman KM. Sensitivity of Charging and Maintenance
Trips of Shared Fully-Automated Electric Vehicle Fleets in a Large-Scale Model
2022.
Gurumurthy KM, Kockelman KM. Dynamic ride-sharing impacts of greater trip demand
and aggregation at stops in shared autonomous vehicle systems. Transportation
Research Part A: Policy and Practice 2022;160:114–25.
https://doi.org/10.1016/j.tra.2022.03.032.
Gurumurthy KM, Kockelman KM. Modeling Americans’ autonomous vehicle preferences:
A focus on dynamic ride-sharing, privacy & long-distance mode choices.

37
Technological Forecasting and Social Change 2020;150:119792.
https://doi.org/10.1016/j.techfore.2019.119792.
Hao X, Wang H, Lin Z, Ouyang M. Seasonal effects on electric vehicle energy
consumption and driving range: A case study on personal, taxi, and ridesharing
vehicles. Journal of Cleaner Production 2020;249:119403.
https://doi.org/10.1016/j.jclepro.2019.119403.
Iacobucci R, Bruno R, Schmöcker J-D. An Integrated Optimisation-Simulation Framework
for Scalable Smart Charging and Relocation of Shared Autonomous Electric
Vehicles. Energies 2021;14:3633. https://doi.org/10.3390/en14123633.
Illinois Department of Transportation (IDOT). Drive Electric Illinois 2022.
https://idot.illinois.gov/home/drive-electric-illinois (accessed October 5, 2022).
Jones CM, Kammen DM. Quantifying Carbon Footprint Reduction Opportunities for U.S.
Households and Communities. Environ Sci Technol 2011;45:4088–95.
https://doi.org/10.1021/es102221h.
Kintner-Meyer M, Davis S, Sridhar S, Bhatnagar D, Mahserejian S, Ghosal M. Electric
Vehicles at Scale - Phase 1 Analysis: High EV Adoption Impacts on the Western
U.S. Power Grid. Richland, Washington: Pacific Northwest National Laboratory;
2020.
Lee JH, Hardman SJ, Tal G. Who is buying electric vehicles in California? Characterising
early adopter heterogeneity and forecasting market diffusion. Energy Research &
Social Science 2019;55:218–26. https://doi.org/10.1016/j.erss.2019.05.011.
Litman T. Autonomous Vehicle Implementation Predictions: Implications for Transport
Planning. Victoria, Canada: Victoria Transport Policy Institute; 2022.
Mai T, Jadun P, Logan J, McMillan C, Muratori M, Steinberg D, et al. Electrification
Futures Study: Scenarios of Electric Technology Adoption and Power Consumption
for the United States. Golden, CO: National Renewable Energy Laboratory; 2018.
Melendez KA, Das TK, Kwon C. Optimal operation of a system of charging hubs and a
fleet of shared autonomous electric vehicles. Applied Energy 2020;279:115861.
https://doi.org/10.1016/j.apenergy.2020.115861.
Menon N, Barbour N, Zhang Y, Pinjari AR, Mannering F. Shared autonomous vehicles and
their potential impacts on household vehicle ownership: An exploratory empirical
assessment. International Journal of Sustainable Transportation 2019;13:111–22.
https://doi.org/10.1080/15568318.2018.1443178.
Moawad A, Li Z, Pancorbo I, Gurumurthy KM, Freyermuth V, Islam E, et al. A Real-Time
Energy and Cost Efficient Vehicle Route Assignment Neural Recommender
System. ArXiv:211010887 2021.

38
Mulrow J, Grubert E. Greenhouse gas emissions embodied in ev charging infrastructure: A
method and case study of Georgia, US 2021-2050 2023.
Muratori M. Impact of uncoordinated plug-in electric vehicle charging on residential power
demand. Nat Energy 2018;3:193–201. https://doi.org/10.1038/s41560-017-0074-z.
Nacmanson WJ, Zhu J, Ochoa LF. Assessing the unmanaged EV hosting capacity of
Australian rural and urban distribution networks. CIRED Porto Workshop 2022: E-
mobility and power distribution systems, Hybrid Conference, Porto, Portugal: IET;
2022. https://doi.org/10.1049/icp.2022.0795.
Norsk elbilforening. Norwegian EV market 2022. https://elbil.no/english/norwegian-ev-
market/ (accessed March 10, 2023).
Powell S, Cezar GV, Min L, Azevedo IML, Rajagopal R. Charging infrastructure access
and operation to reduce the grid impacts of deep electric vehicle adoption. Nat
Energy 2022:1–14. https://doi.org/10.1038/s41560-022-01105-7.
Rogelj J, Shindell D, Jiang K, Fifita S, Forster P, Ginzburg V, et al. Mitigation Pathways
Compatible with 1.5°C in the Context of Sustainable Development. Global
Warming of 1.5°C. An IPCC Special Report on the impacts of global warming of
1.5°C above pre-industrial levels and related global greenhouse gas emission
pathways, in the context of strengthening the global response to the threat of climate
change, sustainable development, and efforts to eradicate poverty, 2018, p. 82.
Sheppard CJR, Jenn AT, Greenblatt JB, Bauer GS, Gerke BF. Private versus Shared,
Automated Electric Vehicles for U.S. Personal Mobility: Energy Use, Greenhouse
Gas Emissions, Grid Integration, and Cost Impacts. Environ Sci Technol
2021;55:3229–39. https://doi.org/10.1021/acs.est.0c06655.
Shoup D. The High Cost of Free Parking, Updated Edition. 1st edition. Chicago:
Routledge; 2011.
Smart Electric Power Alliance (SEPA). The State of Managed Charging in 2021. 2021.
Smart Electric Power Alliance (SEPA). A Comprehensive Guide to Electric Vehicle
Managed Charging 2019.
Sperling D, Fulton LM, Arroyo V. Accelerating Deep Decarbonization in the U.S.
Transportation Sector 2020.
Susskind L, Chun J, Gant A, Hodgkins C, Cohen J, Lohmar S. Sources of opposition to
renewable energy projects in the United States. Energy Policy 2022;165:112922.
https://doi.org/10.1016/j.enpol.2022.112922.
Szinai JK, Sheppard CJR, Abhyankar N, Gopal AR. Reduced grid operating costs and
renewable energy curtailment with electric vehicle charge management. Energy
Policy 2020;136:111051. https://doi.org/10.1016/j.enpol.2019.111051.

39
Tarroja B, Hittinger E. The value of consumer acceptance of controlled electric vehicle
charging in a decarbonizing grid: The case of California. Energy 2021;229:120691.
https://doi.org/10.1016/j.energy.2021.120691.
van Triel F, Lipman TE. Modeling the Future California Electricity Grid and Renewable
Energy Integration with Electric Vehicles. Energies 2020;13:5277.
https://doi.org/10.3390/en13205277.
U.S. Energy Information Administration (EIA). Levelized Costs of New Generation
Resources in the Anuual Energy Outlook 2022 2022a.
U.S. Energy Information Administration (EIA). Illinois Electricity Profile 2021 2022b.
https://www.eia.gov/electricity/state/illinois/index.php (accessed February 16,
2023).
U.S. Energy Information Administration (EIA). State Profile and Energy Estimates. Illinois
2022c. https://www.eia.gov/state/analysis.php?sid=IL#49 (accessed March 1, 2023).
U.S. Energy Information Administration (EIA). Capital Costs and Performance
Characteristics for Utility Scale Power Generating Technologies. Washington, D.C.:
U.S. Department of Energy; 2020.
Verbas O, Zhou Z. Mobility, Energy, and Grid Integrated Modeling Using POLARIS,
Autonomie, and ALEAF 2020.
Wert JL, Shetye KS, Li H, Yeo JH, Xu X, Meitiv A, et al. Coupled Infrastructure
Simulation of Electric Grid and Transportation Networks. 2021 IEEE Power &
Energy Society Innovative Smart Grid Technologies Conference (ISGT),
Washington, DC, USA: IEEE; 2021, p. 1–5.
https://doi.org/10.1109/ISGT49243.2021.9372175.
Xu L, Yilmaz HÜ, Wang Z, Poganietz W-R, Jochem P. Greenhouse gas emissions of
electric vehicles in Europe considering different charging strategies. Transportation
Research Part D: Transport and Environment 2020;87:102534.
https://doi.org/10.1016/j.trd.2020.102534.
Xu T, Birchfield AB, Gegner KM, Shetye KS, Overbye TJ. Application of Large-Scale
Synthetic Power System Models for Energy Economic Studies, 2017.
https://doi.org/10.24251/HICSS.2017.386.
Zhang C, Greenblatt JB, MacDougall P, Saxena S, Jayam Prabhakar A. Quantifying the
benefits of electric vehicles on the future electricity grid in the midwestern United
States. Applied Energy 2020;270:115174.
https://doi.org/10.1016/j.apenergy.2020.115174.

40
Zhang J, Jorgenson J, Markel T, Walkowicz K. Value to the Grid From Managed Charging
Based on California’s High Renewables Study. IEEE Transactions on Power
Systems 2019;34:831–40. https://doi.org/10.1109/TPWRS.2018.2872905.
Zhang TZ, Chen TD. Smart charging management for shared autonomous electric vehicle
fleets: A Puget Sound case study. Transportation Research Part D: Transport and
Environment 2020;78:102184. https://doi.org/10.1016/j.trd.2019.11.013.