Energy 115 (2016) 646e657

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Optimizing plug-in electric vehicle and vehicle-to-grid charge

scheduling to minimize carbon emissions
Christopher G. Hoehne*, Mikhail V. Chester 1
Civil, Environmental and Sustainable Engineering, Arizona State University, Tempe, AZ, 85281, United States

a r t i c l e i n f o a b s t r a c t

Article history: Electric vehicles are an emerging technology with significant potential for reducing carbon dioxide
Received 25 August 2015 emissions. Yet strategies to minimize carbon dioxide emissions by strategically charging during different
Received in revised form times of day have not been rigorously explored. To identify possibilities for minimizing emissions from
5 August 2016
plug-in electric vehicle use, daily optimized charging strategies over each electricity reliability region of
Accepted 7 September 2016
the United States are explored. Optimized schedules of plug-in electric vehicle charging for standard and
vehicle-to-grid use were compared with pre-timed charging schedules to characterize the potential for
carbon dioxide emission reductions across charging characteristics, regional driving, and marginal en-
Keywords:
Plug-in electric vehicle
ergy generation trends. It was found that optimized charging can reduce carbon dioxide emissions over
Vehicle-to-grid pre-timed charging by as much as 31% for standard use and 59% for vehicle-to-grid use. However, some
Charging scenarios of vehicle-to-grid participation were found to increase carbon dioxide emissions by up to 396 g
Optimization carbon dioxide per mile by displacing stored energy from more carbon-intense energy generation pe-
Carbon riods. Results also indicate that plug-in electric vehicle charging emissions can vary widely for a given
Emissions energy efficiency rating. Current energy efficiency ratings may lead to incorrect assumptions of plug-in
electric vehicles emissions compared to conventional gasoline vehicles due to varying regional and
temporal emissions. To coincide with the push for lower greenhouse gas emissions from transportation,
charging times for plug-in electric vehicles should target periods where charging promotes carbon di-
oxide reductions, and electric vehicle energy efficiency ratings should be reconsidered in order to pro-
mote sustainable plug-in electric vehicle use moving forward.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction
Electric vehicles (EVs) are an emerging technology with signifi-
cant potential for reducing carbon dioxide (CO2) emissions. Shifting
from conventional gasoline vehicles (CGVs) to plug-in hybrid electric
vehicles (PHEVs) could reduce the gasoline consumption by up to
52% of the US petroleum imports [1] and could reduce carbon
emissions in some cases by up to 60% in compared to conventional
hybrid vehicles [2]. EV registrations have exceeded 500,000 units
worldwide, and the registration rate was estimated to have doubled
every seven to eight months from December 2010 to May 2014 [3].
With technology advancements and financial incentives, EVs will
likely continue to see increased market penetration in the near
* Corresponding author.
E-mail addresses: christopher.g.hoehne@gmail.com (C.G. Hoehne), mchester@
asu.edu (M.V. Chester).
1
660 S College Avenue, P.O. Box 873005, Tempe, AZ, 85287-3005, United States.
future and become increasingly competitive with CGVs. Tax rebates
of up to $7500 dollars and fuel costs as low as $0.04 per mile make
EVs an attractive alternative to CGVs [4]. EVs use Lithium-ion bat-
teries almost exclusively and production costs are declining. Ac-
cording to a 2015 study, production costs of Lithium-ion batteries are
between $300/kWh to $410/kWh, and their production costs from
2007 to 2014 have dropped by approximately 14% annually with
projections to continue declining [5]. With the EV market share on
the rise and continuing developments in battery technology, EV use
is likely to continue to develop and expand. Yet EV deployment
strategies to ensure CO2 reductions remain largely undefined.
Emissions from EV use can vary widely depending on the region
and time-of-day in which charging occurs. Marginal energy pro-
duced during nighttime periods may have higher CO2 emissions than
peak periods for some regions by as much as 65% (southwest region
during summer) for generation-estimated marginal emissions [6] or
approximately 91% (southeast region) for consumption-estimated
marginal emission [7]. These extremes are due to high energy

http://dx.doi.org/10.1016/j.energy.2016.09.057
0360-5442/© 2016 Elsevier Ltd. All rights reserved.
 C.G. Hoehne, M.V. Chester / Energy 115 (2016) 646e657
Nomenclature
List of Acronyms
CGV Conventional Gasoline Vehicle
CO2 Carbon Dioxide
EV Electric Vehicle
kW Kilowatt
kWh Kilowatt-hour
MPG Mile Per Gallon
MPGe Mile Per Gallon equivalent
NERC North American Electric Reliability Corporation; NHTS
National Household Travel Survey; PEV Plug-in Electric
Vehicle; PHEV Plug-in Hybrid Electric Vehicle
UDDS Urban Dynamometer Driving Schedule
V2G Vehicle-to-grid
consumption during peak hours in summer months in the southern
Unites States. Energy generated for high demand periods is more
costly on the margin but less carbon intense [8], therefore causing
large variations in peak-to-nighttime marginal CO2 emissions. As a
result, EV charging during recommended nighttime hours can lead
to higher CO2 emissions. Zivin et al. [7] found that EV charging during
nighttime hours (12am - 4am) in the Midwestern United States may
result in higher CO2 emissions than similar use of a CGVs or PHEVs.
Tamayao et al. [9] found that nighttime charging of various EVs
mostly resulted in higher emissions than charging during the eve-
ning. Much discussion of EV charging has focused on questions of
when the technology breaks even with CGVs given regional grid
mixes. While valuable, there remains limited knowledge of how
charging should be scheduled across the US at different times of day
to minimize CO2 emissions. The opportunity remains to identify
optimal charging strategies to minimize CO2 emissions.
Much of the current literature investigating optimization of EV
charging focuses on minimizing costs. Cao et al. [10] proposed
methods to control EV charging loads with a focus on minimizing
charging cost, and Korolko and Sahinoglu [11] investigated opti-
mization of EV charging schedules in unregulated electricity mar-
kets to minimize overall costs. Yuksel and Michalek [12] estimated
EV CO2 emission variations from regional temperature differences,
driving patterns and grid emissions factors. Weis et al. [13] exam-
ined the cost implications of controlled EV charging in a single
region in the US. The findings indicate controlled charging using
the current grid mix produces overall negative net social benefits
due to increased coal generated energy utilized for vehicle charging
even though operator costs were reduced by 23%e34%. Khayyam
and Bab-Hadishar [14] modeled powertrain, driving factors and
environment factors for PHEVs to develop an intelligent energy
management system to improve the vehicles overall energy effi-
ciency. Due to growth in EV market penetration, EV charging
research remains focused on minimizing the strain on the electrical
grid by minimizing costs for operators. Consideration of environ-
mental impacts using grid emission factors has been considered in
some EV scheduling research. However, reducing environmental
impacts from EV charging has not been the motivation for opti-
mizing charge schedules, even though controlling EV charging to
reduce generation costs may produces negative net social benefits
[13]. As a result, optimized charge scheduling motivated by
reducing carbon emissions may expand the understanding of EV
647
Formulae terms
Bc EV battery capacity, in kWh
Ce The effective EV charge rate in kW
CL2 Max achievable charge rate of the EV in kW
dr Average urban daily miles driven in region r
ED,i Extra energy being charged to later be discharged to
the grid during peak hours
EV2G,i Total energy being charged or discharged by the V2G
EV in hour i
Ei Energy in kWh drawn from the grid to charge a given
EV in hour i
MERs,r,i Marginal emissions rate in grams CO2/kWh for season s
in region r for hour i
h Charging efficiency for level 2 charging
Ps EV powertrain efficiency in season s, in kWh per mile
Xs,r Daily CO2 emissions, in grams CO2 for a given EV in
season s in region r
YV2G,s,r CO2 emissions in grams per kWh discharged
charging related emissions and the practicality of optimized charge
scheduling.
In addition to EV charging, vehicle-to-grid (V2G) technology has
been cited as a method of integrating renewable energy to the grid
while also aiding in load balancing during peak hours by delayed
discharging of excess EV battery energy [15]. Emerging V2G tech-
nologies utilize EV battery energy by discharging excess capacity
during high demand hours to help regulate peak energy demands.
Research indicates that V2G use could be adopted in the near future
as a supplementary means for peak load regulation [16]. Fluctuating
energy mixes by region and time-of-day could cause highly varied
emissions when considering the energy displaced from V2G partic-
ipation. Factors such as regional energy mix, charge schedule, and EV
characteristics may all influence the emissions from V2G use.
Although there is potential for future V2G implementation in
EVs, there is little discussion on the environmental effects from V2G
charging. Most cases evaluate optimal scheduling of V2G use to
focus on maximizing economic benefit or minimizing the strain to
the energy providers, following the same trend as non V2G EV
scheduling research. Implications of controlled and intelligent
management of V2G use has been evaluated in a number of cases.
Honarmand et al. [17] modeled realistic V2G charge and discharge
schedules for implementation in an intelligent parking lot and
found that optimal charging occurs when energy prices are lower
and discharging occurs when energy prices are higher. With some
exceptions, most research does not focus on the environmental
impacts from the V2G implementation. Sioshansi and Denholm
[18] found that V2G use could reduce generator CO2 emissions
substantially in Texas. Cardoso et al. [19] modeled the effects of V2G
use at a large office building in San Francisco to find reductions in
energy cost and CO2 emissions when payback periods are small.
Therefore, investigating the optimization of V2G charging-
discharging by minimizing CO2 emissions may provide valuable
new insight.
If EVs can be positioned to minimize emissions from passenger
vehicle use, optimal charging strategies should be identified that
are sensitive to marginal time-of-day emissions and the regional
electricity mix. This paper addresses three specific research ques-
tions: 1) Can optimized or strategic EV charging schedules reduce
CO2 emissions for EV and V2G use? 2) What EV charging and
driving characteristics have the most influence on CO2 emissions?
and, 3) How might V2G use change CO2 emissions relative to EV
648 C.G. Hoehne, M.V. Chester / Energy 115 (2016) 646e657
use? This paper investigates the effects of optimizing charging
schedules to answer these questions.
2. Methods
Plug-in EV charging in multiple scenarios were assessed in each
electricity reliability region of the United States to characterize the
potential of charging characteristics, regional driving and marginal
energy generation trends that minimize CO2 emissions. First,
assuming all-electric use (modeled with a Nissan Leaf plug-in EV
and Chevrolet Volt PHEV) with regional urban driving character-
istics, CO2 emissions are estimated under three pre-timed daily
charging schedules within each North American Electric Reliability
Corporation (NERC) region for both temperate weather and
extreme weather cases. Next, an optimization framework is applied
to each scenario such that CO2 emissions are minimized. Last, CO2
emissions are estimated assuming V2G use for three pre-timed
charging schedules paired with two pre-timed discharge sce-
narios for the Nissan Leaf and Tesla Model S plug-in electric vehi-
cles (PEVs). The optimization framework was modified and applied
to minimize CO2 emissions for V2G use. Pre-timed and optimized
CO2 emissions for all scenarios is then compared to identify char-
acteristics that contribute most to emission reductions and assess
the potential gap in perceived versus actual emissions of EVs.
Due to possible expansion of EVs, all analyses of charging
schedules is first divided into two types of EV use scenarios for a
broader evaluation of CO2 reductions: standard EV use and V2G use.
Standard EV use is defined as an EV with average urban driving
characteristics, and V2G use is defined as a scenario in which a
properly equipped EV following average urban driving character-
istics discharges excess energy during peak demand hours. With
the known limited EV driving range coupled with non-residential
charging infrastructure most heavily found in the urban or subur-
ban setting [20], it is assumed that the majority of EV use occurs
within the urban or suburban environment. For V2G use, the
equipped EV discharges stored energy within the battery during
peak demand hours to help balance the load on the electrical grid.
The energy being stored in the EV battery to later be discharged is
defined as displaced energy. For the V2G scenario, three key
Fig. 1. Hierarchy flow chart outlining all charging cases examined. Standard EV use was evaluated for temperate and extreme weather scenarios. V2G use only considered temperate
weather use.
assumptions are necessary: the EV and grid operators have the
required technology to allow V2G discharging, and there is
adequate infrastructure at the EV user's home or destination to
allow discharging during peak hours, and V2G discharging is
assumed to occur at a uniform rate over peak demand periods. A
hierarchical summary of examined scenarios is shown in Fig. 1.
2.1. Data
The Nissan Leaf, the Chevrolet Volt, and the Tesla Model S are
chosen to represent the range in battery sizes and technologies for
EVs and PHEVs. They are applied to select charging scenarios due to
their popularity and representative charging characteristics. As of
January 2014, the Nissan Leaf holds the largest global market share
at 45% of all EVs sold [21]. According to a 2014 global review of EV
market penetrations, the Chevrolet Volt is the most popular EV in
the United States with around 55,000 cumulative units sold, fol-
lowed by the Nissan Leaf (over 42,000 cumulatively units) in sec-
ond, and the Tesla Model S (nearly 22,000 cumulatively units) in
third [22]. Together, these three EVs combine for over two thirds of
US market share [22]. The Leaf represents an economy PEV, the Volt
represents an economy PHEV and the Model S represents a large
battery pack luxury PEV. Displayed in Table 1 are the three main
characteristics when considering EV charging: battery capacity, Bc,
in kWh, charger rate, C, in kW and powertrain efficiency, Ps, in kWh/
mile. Two powertrain efficiency for each EV are evaluated to
represent temperate use (72
F ambient temperature with no
climate control) and extreme temperature use (20
F ambient
temperature with climate control set to 72
F) [23]. The Volt
powertrain efficiencies used are for all-electric driving. All vehicles
are assumed to use level 2 charging at 220 V and varying amperes
rates. Due to the lack of dynamometer data for the Tesla Model S, its
powertrain efficiencies are assumed to equal the Nissan Leaf for
analysis in the V2G case. The charger efficiency for all three EV
chargers is estimated to be 89.4% [24]. To apply these established
EVs to charging scenarios, hourly CO2 emissions rates are also
necessary.
To accurately estimate EV charging-related CO2 emissions from
different time-of-day charging scenarios, regional and temporal
 C.G. Hoehne, M.V. Chester / Energy 115 (2016) 646e657 649

Table 1
Summary of main charging-related EV characteristics.

EV Modela Charge rate Battery capacity UDDS powertrain efficiency (temperate)b UDDS powertrain efficiency (winter)c

Chevrolet Volt 4.4 kW (220 V, 20 A) 16 kWh 0.222 kWh/mile 0.415 kWh/mile

Nissan Leaf 6.6 kW (220 V, 20 A) 24 kWh 0.209 kWh/mile 0.418 kWh/mile
Tesla Model S 11 kW (220 V, 20 A) 60 kWh 0.209 kWh/mile 0.418 kWh/mile
a
All specifications are from 2012 year models [36e40].
b
UDDS, cold start, 72
F ambient temperature, climate control off [41].
c
UDDS, cold start, 20
F ambient temperature, climate control set to 72
F [41].

marginal emissions rates are utilized. Marginal emissions rates range in charge times. A charging schedule is defined as the daily
have been previously developed to estimate historical time-of- period in which an EV will recharge the battery capacity depleted
day changes in the marginal energy mix. Siler-Evans et al. [6] from the trips occurring in the last 24 h. The three pre-timed
estimated generation-based marginal emissions rates by schedules are defined as follows: day charging where EV
regressing historical hourly changes in energy generation from charging begins at 12pm, evening charging where the EV begins
2006 to 2011. Although this data set examines multiple emissions, charging at 6pm, nighttime charging where the EV begins
CO2 is the only greenhouse gas assessed. It is acknowledged that charging at 12am. Because the availability of EV charging infra-
other emissions from electricity generation and vehicle operation structure is not currently widely abundant [20], scenarios sup-
contribute to global warming potential, such as methane, nitrous porting times most common for residential charging is first
oxide, or automobile air-conditioning refrigerant emissions [25], formulated. The pre-timed evening charge scenario represents EV
however, it is not expected that their inclusion will significantly users performing a convenience charge at their own residence
alter results. These marginal emissions vary by region due to the after returning home immediately following the late afternoon
varying degree of carbon intensities in each region's electricity peak travel period. The pre-timed night charge scenario repre-
mix. To meet demand during peak periods, energy used on the sents a delayed charge starting at the end of the user's day. The
margin has lower carbon intensities relative to off-peak and pre-timed day charging scenario represents the opportunity of
baseload generation [6]. This occurs because off-peak and base- day charging consistent with the availability of non-residential
load generation is provided most heavily by burning coal, while charging infrastructure such as workplace or commercial
high demand periods typically are less carbon intense, mainly charging options. Daytime charging may increase the access and
from natural gas and small amounts of dispatchable renewables flexibility to charging times, therefore considering a pre-timed
[26]. These marginal CO2 emission rates were found to remain case with day charging is useful.
stable year-to-year [6]. To account for the effects of temperate and Charging-related CO2 emissions for pre-timed EVs are calcu-
extreme climate on EV powertrain performance, temperate sea- lated to compare with optimized profiles. The daily CO2 emissions,
son marginal emission rates are paired with temperate weather XEV,s,r, in grams CO2 for a given EV in season s in region r for pre-
powertrain efficiencies (72
F ambient temperature with no timed scheduling are estimated using
climate control), and winter season marginal emission rates are
paired with cold weather powertrain efficiencies (20
F ambient X
24
Ei
temperature with no climate control). The temperate season is Xs;r ¼ $MERs;r;i (1a)
defined as occurring in intermediate months without extreme 1
h
temperatures (March, April, September, and October). These two
pairing help achieve a range of possible energy efficiencies under
temperate and extreme weather scenarios to best characterize the Table 2
range of possible emissions. Table 2 shows these hourly marginal Temperate month marginal emission factors by NERC region (kilograms CO2/kWh)
[6]. Winter season marginal emission factors not shown.
CO2 emissions rates (in grams CO2/kWh) by NERC region for the
temperate season. Time ERCOT FRCC MRO NPCC RFC SERC SPP WECC
For all cases, it is assumed that the EVs are used for urban 12AM 585 530 916 474 794 795 755 512
driving with average daily regional driving distance of dr miles per 1AM 643 544 907 516 781 774 811 548
day. To estimate the average driving distance of EVs for each NERC 2AM 685 537 891 529 760 759 774 567
3AM 644 556 885 515 745 759 707 540
region, 2009 National Household Travel Survey (NHTS) data was
4AM 560 543 875 450 738 746 598 515
aggregated by driving characteristics and region [27]. Specifically, 5AM 499 506 902 447 736 743 602 491
the NHTS data is filtered to include only private, light-duty vehicle 6AM 594 531 885 454 739 790 605 478
urban trips to ensure driving characteristics would most nearly 7AM 599 557 855 463 720 686 665 506
match the EV driving under the Urban Dynamometer Driving 8AM 548 524 845 475 695 636 576 490
9AM 491 519 797 479 690 624 533 482
Schedule (UDDS). To avoid duplicate trip counts, only trips reported 10AM 498 504 809 484 691 620 530 473
by the driver were aggregated. Table 3 compares the average urban 11AM 491 470 802 483 693 645 546 485
daily driving distances compared to the urban driving cycle dis- 12PM 490 484 821 479 726 656 556 514
tances for the examined EVs. It should be noted that regional 1PM 479 492 819 468 699 700 573 499
2PM 474 492 809 437 733 715 584 511
roadway characteristics may alter vehicle emissions, but has been
3PM 507 458 795 455 736 709 602 482
found to be negligible [28], and therefore was not included in this 4PM 525 474 865 499 712 737 574 493
analysis. 5PM 512 525 872 482 728 735 564 486
6PM 517 526 868 507 744 739 584 500
7PM 519 517 850 488 723 669 577 513
2.2. Standard EV use case 8PM 500 511 849 474 719 627 551 511
9PM 480 528 881 428 713 593 569 467
Three pre-timed EV charging schedules for standard EV use are 10PM 526 537 893 424 756 619 588 416
11PM 546 562 905 453 804 730 658 456
developed as a baseline case to account for a flexible and realistic
650 C.G. Hoehne, M.V. Chester / Energy 115 (2016) 646e657

Table 3
Average urban daily driving distances by NERC regiona (2009 NHTS) [27].

NERC Average urban trip distance Average urban daily trips (trips/ Average urban daily distance driven UDDS drive cycle tested distance
region (miles) day) (miles) (miles)

FRCC 6.87 3.5 23.8 7.43 to 7.45

MRO 5.47 3.7 20.4
NPCC 6.96 3.6 25.0
RFC 6.74 3.6 24.1
SERC 6.90 3.6 24.6
SPP 5.86 4.0 23.7
TRE 7.04 3.7 25.9
WECC 6.94 3.6 25.0
a
NERC boundaries from 2010 were used to group NHTS data.

2.3. V2G use case

such that Ce  CL2
X
24
Ei
and ¼ dr $Ps
1
h stored battery energy to be later dispatched during peak hours.
where Ei denotes energy in kWh drawn from the grid to charge the
EV in hour i, MERs,r,i denotes the marginal emissions rate in grams
CO2/kWh for season s in region r for hour i (see Table 2), Ce denotes
the effective charge rate of the EV in kW (Table 1), CL2 denotes the
max achievable charge rate of the EV in kW (Table 1), dr denotes the
average urban light-duty vehicle distance driven per day in miles in
region r (see Table 2), Ps denotes the powertrain efficiency of the EV
for season s in kWh/mile (Table 1), and h denotes the charging ef-
ficiency for level 2 charging. This general equation is then applied to
each of the pre-timed schedules by assigning the specified hours of
charging. For example, an evening charging schedule (charging at
starting at 6pm) can be specified for a given region, season and EV
by:
Xs;r ¼ E18 $MERs;r;18 þ … þ Ej1 $MERs;r;j1 þ Ej $MERs;r;j
where X is the daily grams of CO2 emissions in grams and j is the
last hour charging is required. Note that variables E20 through Ej1
all must be less than or equal to the max hourly energy drawn such
that (1c) is satisfied. For this notation, j is a whole number and is
calculated by rounding up the product of dr, Ps, and h1. A complete
schedule is defined as one 24-h charge period, therefore the initial
result is daily emissions. Per mile emissions may be calculated by
dividing daily emissions, Xs,r, by the average daily driving distance,
dr. Applying these equations to each EV within each NERC region, a
comprehensive charging-related emissions profile for pre-timed
standard EV use was established. With these pre-timed emissions
calculated, optimization of standard EV use follows.
An optimization framework is applied to the Leaf and Volt under
temperate and extreme weather conditions to identify charging
characteristics that minimize CO2 emissions in each NERC region.
Similar to Equations (1a) through (1c), the general optimization
function to minimize CO2 emissions for a specific case was
formulated as
(1b)
To properly formulate CO2 emissions for both pre-timed and
optimized V2G scenarios, additional constraints and assumptions
are necessary. First, daily charging is increased to allow extra
(1c)
Peak hours available for discharge is limited to two scenarios
based on periods identified to be the most fit for stored energy
discharge [29]. These periods are the afternoon peak demand
period occurring from 10am to 4pm, and the evening peak de-
mand period occurring from 6pm to 12am. Based on recom-
mended charging practices, the battery packs are not allowed to
be discharged below 25% state of charge to avoid rapid degrada-
tion in battery performance and lifetime [30]. To keep consistency,
the Nissan Leaf was controlled to discharge 6 kWh evenly over the
6 h discharge periods, and the Tesla Model S was controlled to
discharge 30 kWh evenly over the 6 h discharge periods. Pre-
timed charging during the day with day discharging and pre-
timed charging during evening with evening discharging are not
considered due to conflicting charge-discharge times. These
specifications ensure that each EV will have enough battery ca-
pacity to continue driving if needed without depleting the battery
(2)
below recommended levels while allowing for comparisons of
high capacity vs small capacity V2G discharging events. Although
peak demand periods are consistent across regions, actual de-
mand will vary, and V2G discharging in practice will be controlled
to meet minute-by-minute demand based on supply and demand
in the current energy market. As such, emissions reductions from
discharging are determined assuming a uniform discharge rate
over peak demand periods. It is also assumed that V2G capable
vehicles to be utilized are connected during peak power demand
periods. With these justifiable assumptions and specifications,
CO2 emissions from V2G charging could be analyzed and
optimized.
Accounting for these extra constraints and assumptions, the
same three pre-timed schedules of charging (day, evening, and
night) are applied across the Leaf and Model S within each NERC
region to estimate the CO2 emissions displaced for the V2G sce-
narios. Fig. 1 shows the hierarchy of V2G cases examined. The CO2
emissions displaced per kWh for these pre-timed V2G cases are
estimated by using the equation

X
24
Ei
Xs;r ¼ minimize $MERs;r;i (3) " !
h X
24
EV2G;i
1 YV2G;s;r ¼ $MERs;r;i
1
h
following constraints (1b) and (1c). With optimal charging sched- #, (4a)
X
n
EV2G;i
ules that minimize CO2 emissions for each EV in each region,  XEV;s;r  $MERs;r;i ED;i
optimized results could be compared to pre-timed results. In i
h
addition, method to compare CGV emissions to standard EV use for
both optimized and pre-timed charging schedules was developed. following constraints (1b) and the new constraint
 C.G. Hoehne, M.V. Chester / Energy 115 (2016) 646e657 651

Fig. 2. Nissan Leaf optimized and pre-timed charging related CO2 emissions per mile for all-electric use (in grams per mile). Shaded bars represent CO2 emissions per mile for
temperate season use and positive error bars represent the maximum emissions during extreme weather use.

Midwest (MRO) region may emit more CO2 per mile than similar
X
24
E;i ED;i use of a 25 MPG CGV without optimized scheduling.
¼ dr $Ps þ (4b)  The utilization of optimized charging schedules under urban
1
h h
V2G use results in lower CO2 emissions than pre-timed V2G
schedules in most cases, with maximum reductions of 59%.
where EV2G,i equals the total energy being charged or discharged by
 V2G use may increase CO2 emissions per mile (by up to 396 g
the V2G EV in hour i (discharge rates are negative), ED,i equals the
CO2 per mile) when fully utilizing higher storage capacity bat-
extra energy being charged for later be discharge to the grid during
teries. Optimized charge scheduling may increase net CO2
peak hours, XEV,s,r is refers to the equivalent non-V2G case emis-
emissions when displacing stored battery energy from V2G
sions for the same EV, and i to n represent the start to end hours of
participation.
the discharge period (10am e 4pm or 6pme12pm). EV2G is equal to
6 kWh for Leaf, and 30 kWh for the Model S. The final result, YV2G,s,r,
may be positive or negative CO2 emissions in grams per kWh dis-
3.1. Standard EV use
charged. A positive emissions rate indicates displaced energy was
charged during a higher marginal emissions period leading to an
For standard EV use, optimized charging schedules would
increase in emissions per kWh discharged during the peak period.
result in lower CO2 emissions than pre-timed schedules in all
The previously outlined optimization framework was applied to
cases. Fig. 2 (Nissan Leaf) and Fig. 3 (Chevrolet Volt) display the
minimize CO2 emissions for V2G use by accounting for V2G con-
variations in CO2 emissions per mile for optimized and pre-timed
straints. For optimization of the V2G scenarios, the same restriction
schedules. Reductions in CO2 emissions per mile varied between a
for charge rate from the pre-timed V2G cases apply. Minimizing
minimum of 2.5% and a maximum of 31%. Over a 100,000 mile
YV2G,s,r in Equation (4a) with EV2G,i as the only decision variable
vehicle lifetime following average driving trends in all seasons,
following the charge rate constraint (1b) and the new V2G
optimized schedules could reduce a maximum of approximately 9
displaceable energy constraint (4b), the output determines the
short tons of CO2 (39% lifetime reduction), or a minimum of
lowest rate of displaced CO2 emissions per kWh. This optimization
approximately 1 short ton of CO2 (8% lifetime reduction). The
framework was applied individually to the Nissan Leaf and the Tesla
highest reduction cases observed are optimization of a Leaf and
Model S within each NERC region for temperate and extreme
Volt pre-timed night charging schedules in the Southwest Power
weather scenarios.
Pool (SPP). The lowest reduction cases are optimization of a Leaf
and Volt pre-timed day charging schedules in the Texas Reliability
Entity (TRE). In the Midwest and Southeast regions (MRO, SERC
3. Results and SPP), use of optimized EV charge schedules over pre-timed
night charging could be 59%e131% as effective as using
A summary of the major findings: optimized EV charge scheduling in place of a 35 MPG CGV (Figs. 4
and 5). Although CO2 emissions per mile in extreme weather
 Optimized scheduling under standard urban EV use can reduce approximately double largely in part to the near halved drive cycle
CO2 emissions over pre-timed schedules up to 31% in the short efficiencies under extreme weather conditions (20
F ambient
term and up to 39% in the long term. temperature and winter marginal energy generation emissions),
 In some scenarios, standard EV use may emit more CO2 per mile optimization of charging schedules is slightly less effective
than CGVs. In the most extreme case, urban winter EV use in the compared to the temperate season cases. Reductions in CO2
652 C.G. Hoehne, M.V. Chester / Energy 115 (2016) 646e657

Fig. 3. Chevrolet Volt optimized and pre-timed charging related CO2 emissions per mile for all-electric use (in grams per mile). Shaded bars represent CO2 emissions per mile for
temperate season use and positive error bars represent the maximum emissions during extreme weather use.

emissions per mile for extreme weather cases varied between a
minimum of 4.4% and a maximum 26%. Optimization improved
most over the night pre-timed charging schedule due to nighttime
mixes typically having the most carbon-intense electricity gen-
eration periods. Although optimized charge schedules can reduce
CO2 emissions in all cases, some scenarios provide negligible re-
ductions over pre-timed schedules. Pre-timed charging occurring
during day periods left the least room for CO2 emission reductions
through charge schedule optimization. As observed in Siler-Evans
et al. (and also in Table 2), regions have less carbon-intensive
mixes during periods with the highest demand for energy dur-
ing afternoon and evening peaks [6]. As a result, these hours are
highly targeted for optimal schedules. These trends were similarly
observed in Tamayao et al. [9] where delayed charging (12am
start) was found to produce more CO2 emissions in nearly all cases
than convenience charging (charging upon arriving home, most
commonly occurring around 5e8pm). In addition, these results of
emissions per mile agree with the range of estimated emissions
per mile found in Tamayao et al. with the major difference
stemming from inclusion of battery and vehicle manufacturing in
the system boundary (approximate range of 150e400 g CO2 mile).
As a result, the overall emissions per mile are slightly higher in the
Tamayao et al. case. Yuksel and Michalek also was found to have
similar emissions per mile (approximate range of 100e340 g CO2

Fig. 4. Nissan Leaf estimated lifetime charging related CO2 emission reductions from using optimized scheduling (in kilograms). Reductions are shown assuming optimized
scheduling of Nissan Leaf instead of four other scenarios: pre-timed day charging, pre-timed evening charging, pre-timed night charging, and use of a 35 MPG CGV. All cases assume
100,000 mile lifetime. All emissions are only operation realted.
 C.G. Hoehne, M.V. Chester / Energy 115 (2016) 646e657 653

Fig. 5. Chevrolet Volt estimated lifetime charging related CO2 emission reductions from using optimized scheduling (in kilograms). Reductions are shown assuming optimized
scheduling of Chevrolet Volt instead of four other scenarios: pre-timed day charging, pre-timed evening charging, pre-timed night charging, and use of a 35 MPG CGV. All cases
assume 100,000 mile lifetime. All emissions are only operation related.

per mile) when comprehensively considering regional tempera-
ture effects on EV efficiencies and emissions [12].
In most cases, the Nissan Leaf and Chevrolet Volt are estimated
to have lower CO2 emissions per mile from operation than the
majority of available CGVs and hybrid-electric vehicles during
temperate season use, but may often have similar or higher CO2
emissions per mile when considering extreme weather use. Figs. 4
and 5 compare the range of charging related CO2 emissions for the
Nissan Leaf and Chevrolet Volt in temperate and extreme weather
with benchmark CGV emissions rates shown for reference. The
MRO region has the highest estimated CO2 emissions per mile for
all scenarios due to the regions more carbon-intense energy mix

55 MPG

Fig. 6. Range of temperate season CO2 emissions per mile for the Nissan Leaf (left in green) and the Chevrolet Volt (right in yellow). Estimated CO2 emissions per mile for a 55 MPG
CGVs is shown for reference. The Nissan Leaf optimized scheduling realted emissions represented by light green; Nissan Leaf pre-timed scheduling related emissions range rep-
resented by dark green. The Chevrolet Volt optimized scheduling related emissions represented by light yellow; Chevrolet Volt pre-timed scheduling related emissions range
represented by dark yellow. Assumptions includes no vehicle climate control, ambient temperatures at 72
F and Chevrolet Volt is all-electric use only. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)
654 C.G. Hoehne, M.V. Chester / Energy 115 (2016) 646e657

25 MPG

35 MPG

55 MPG

Fig. 7. Range of extreme weather CO2 emissions per mile for the Nissan Leaf (left in green) and the Chevrolet Volt (right in yellow). Estimated CO2 emissions per mile for a 25, 35
and 55 MPG CGVs is shown for reference. The Nissan Leaf optimized scheduling related emissions represented by light green; Nissan Leaf pre-timed scheduling related emissions
range represented by dark green. The Chevrolet Volt optimized scheduling related emissions represented by light yellow; Chevrolet Volt pre-timed scheduling related emissions
rage represented by dark yellow. Assumptions includes vehicle climate control set at 72
F, ambient temperatures at 20
F and Chevrolet Volt is all-electric use only. Also note that
not all areas in each regions will achieve 20
F temperatures. Extreme weather estimates may be achieved in only some parts of each region. (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this article.)

relative to other regions. Under temperate season use of the Nissan
Leaf and Chevrolet Volt (Fig. 6), charging schedules in about half of
NERC regions already surpass the EPA's target for 55 mile per gallon
(MPG) (equivalent of approximately 162 g CO2 per mile2) ratings for
light-duty vehicles by 2025 [31]. However, extreme weather use of
these EVs (Fig. 7) in many regions may emit similar or more CO2
emissions per mile than current sold light-duty CGVs [32].
3.2. V2G use
The utilization of optimized EV charging schedules for V2G
use would result in lower estimated CO2 emissions than pre-
timed V2G schedules in most cases. Reductions in CO2 emis-
sions per mile varied between no reductions and a maximum of
59%. Although optimizing V2G charge schedules may reduce
emissions, displacing stored battery energy from V2G participa-
tion will still increase in CO2 emissions. V2G related CO2 emis-
sions are found to increase by a maximum of 396 g CO2 per mile
in the Midwest region when fully utilizing higher storage ca-
pacity batteries. Some optimization schedules were found to
increase CO2 per mile over pre-timed schedules due to some pre-
timed charging schedules occurring during periods with less
carbon-intense energy generation relative to its paired discharge
period, though this was uncommon. This indicates that even with
strategic charging, V2G use may increase CO2 emissions in some
regions. Fig. 8 gives an overview of these V2G results. The least
carbon intense periods are often the highest energy demand
periods due to energy generation on the margin using less
carbon-intense methods. These two periods (afternoon peak and
evening peak) are also the target periods for discharging stored
energy due to high demand. Because charging and discharging
2
Emissions rate calculated using 8887 g CO2 emitted from one gallon of gas [42].
are not allowed during the same time slots (Figure 8b and
d exclude conflicting charge scenarios), optimization to minimize
CO2 emissions occurs in the least carbon intense periods that are
not occupied for discharging. Therefore, if optimized charge
schedules are not restricted outside there discharge periods,
optimized charging gravitates towards peak generation periods
(evening discharge prefers to charge during the afternoon peak,
and day discharge prefers to charge during the evening peak).
Thus, reducing CO2 emissions under V2G use mostly occurs when
stored energy is charged during peak periods. When V2G
charging does not occur during any peak periods (Fig. 8, night-
time charging) emissions are found to increase. This will cause
optimization of V2G schedules to be counter-effective for aiding
in peak load reduction. These findings indicate although re-
ductions in CO2 emissions from V2G charging is possible, it may
be counter-intuitive and difficult for practical use. It also high-
lights that V2G use will, in most cases, increase overall CO2
emissions by using off-peak generated energy for peak load
reduction.
4. Discussion
The results indicate that EV users can reduce their overall
charging-related CO2 emissions by strategically adapting their
charging habits. A simple shift to reduce emissions could occur
from EV users avoiding nighttime charging and instead charging
during evening or day hours. Charging during evening and night
hours is more convenient via residential charging because EV users
are more likely to be home, but daytime charging for many EV users
could further reduce charging related emissions. This would
require the availability of non-residential charging infrastructure.
Thus, increased availability of non-residential charging infrastruc-
ture would greatly increase the flexibility of charging and help
promote strategic charge scheduling.
 C.G. Hoehne, M.V. Chester / Energy 115 (2016) 646e657
Fig. 8. Increase of CO2 emissions per mile for the Nissan Leaf and Tesla Model S under various V2G charge-discharge scenarios in temperate weather use. Values shown are in-
creases (or decreases when negative) due to displacing stored battery energy from V2G participation, not absolute emissions per mile. Assumptions include no vehicle climate
control, 72
F ambient temperature; 6 kWh for discharged per day for Nissan Leaf; 30 kWh discharged per day for Tesla Model S. Conflicting charge-discharge events are excluded
from the analysis.
Policy makers should consider additional subsidies or incentives
that promote strategic scheduling and less carbon-intense energy
use with high EV market penetration to reach goals of lower
emissions in the future automobile fleet. Persuading EV users to
alter charging habits for further emissions reductions may be
difficult without incentives or subsidies. Strategic and optimal
charging focusing on emission reductions may be inconvenient
(whether it be shifting charging times out of current habit or
increasing the marginal charge cost per mile) for the average EV
user without further incentive. Once EV market penetration in-
creases to the point of little to no tax rebates, EV targeted subsidies
could be transferred into EV energy purchases to shift charging
behavior. Alternatively, carbon taxes on vehicle emissions could be
implemented to incentivize strategic charging habits in EVs as well
as promote further EV use. This would allow for EV charging to be
effective at achieving high equivalent fuel economies and energy
efficiencies without drastically increasing the peak energy de-
mands. Future research may be needed to quantitatively estimate
the subsidies or taxes to promote the best application for carbon
emission reductions.
Due to a number of factors such as marginal energy generation,
regional energy mix composition, EV powertrain characteristics,
and user driving behavior, policy makers and EV users should
recognize emissions can vary widely from reported mile per gallon
equivalent (MPGe) ratings. Because the push for increased fuel
economies is driven by reducing greenhouse gas emissions, many
consumers and decision makers may incorrectly associate MPGe
ratings with carbon emissions rates [33]. This may lead to incorrect
assumptions concerning perceived EV emissions versus actual EV
emissions due to misunderstanding of the MPGe rating. As a result,
MPGe labels for EVs, which are often used to compare the vehicle's
energy use against CGVs, may not provide an effective comparison
655
for emissions in many cases. Policy and decision makers should
push for a more robust understanding of EV charging emissions,
such as exploring an alternative to rating to MPGe, to ensure EV
users are well informed and may realistically consider strategic
scheduling.
Adoption of V2G use is not a sustainable solution for reducing
carbon emissions due to the high potential for increased emis-
sions given current marginal energy generation trends. Without
strategic or optimized charging schedules or less carbon-intense
generation methods during off-peak hours, V2G use for load
balancing purposes will increase CO2 emissions in all cases. When
V2G discharging is near maximum capacity per day, longer pe-
riods of charging increase emissions by requiring increased
charging in more carbon-intense hours. For example, discharging
a fraction of the battery capacity for one or 2 h would require
much less charge time and therefore allow more concentrated
charging during less carbon-intense hours. As a result, selective,
targeted, or case-by-case V2G use with strategic scheduling would
be a more viable, supplementary solution to peak load regulation
rather than widespread adoption to ensure no large increase in EV
emissions.
5. Limitations
When considering long term policy goals for carbon and
greenhouse gas reductions, the results indicate that evaluating
strategic methods of EV charging schedules to reduce carbon
emissions is worthwhile. However, there are a few important
limitations and assumptions that should be highlighted.
The results focus on low EV market penetrations. It was
assumed that marginal energy generation would not be signifi-
cantly altered by a possible change in concentrated EV charging
656 C.G. Hoehne, M.V. Chester / Energy 115 (2016) 646e657
habits. If widespread EV adoption becomes a reality, energy de-
mand could substantially be increased during non-peak hours,
causing a change in off-peak energy generation trends. In addition,
this could affect V2G use which was not subject to load balancing
constraints. V2G may have more potential to balance grid loads and
emissions displacements with high EV market penetration. At low
market penetration, it was assumed V2G use would also not dras-
tically change the energy generation trends.
EV powertrain modeling and driving trends are also subject to
some limitations. With limited EV driving ranges and non-
residential charging infrastructure most heavily found in the ur-
ban or suburban setting [20], it is assumed that the majority of EV
use occurs within the urban or suburban environment. These re-
sults therefore may not accurately represent non-urban or long
distance EV use. However, it should be noted that EV drive cycle
powertrain efficiencies do not significantly decrease with longer
distance trips or higher average speeds [23]. Due to the lack of
powertrain modeling data for the Tesla Model S, V2G use of the
Model S was modeled assuming powertrain efficiencies equal to
the Nissan Leaf. However, the V2G analysis focuses more on energy
storage rather than powertrain efficiencies.
Lastly, energy efficiencies and total emissions in all cases
consider only operation-related emissions and do not encompass a
full life cycle assessment of EVs. To more completely evaluate the
long term sustainability of EVs, a life cycle assessment is necessary
[34,35].
6. Conclusion
Considering the growth of the EV market in addition to the
difficulty in accurately estimating regional vehicle emissions [28],
strategic or optimized charge scheduling has the potential to be
useful in reaching future emission reductions targets. In many
scenarios, optimization of EV charging schedules indicates that EV
users could practice methods of strategic charge scheduling to
further reduce CO2 emissions. Persuasion of EV users to change
their charging habits would likely require subsidies to either energy
generators or the EV users themselves to offset the increased cost
from strategic scheduling. Due to the many factors affecting EV
charging and use, charging related emissions can vary widely for a
given energy efficiency. If an individual is interested in purchasing
an EV because of their perceived reductions in emissions, MPGe
labels will likely not provide a meaningful indication for carbon
emissions. In addition, it would be greatly beneficial to better
educate EV users of projected time-of-day emission rates based on
regional and historical trends. EV users, manufactures and pro-
moters of sustainable transportation solutions should investigate
practical methods to more accurately reflect EV emissions
compared to CGVs and develop methods for better education of
charging-related EV emissions to promote smarter charging
practices.
References
[1] Scott MJ, Kintner-meyer M. Impacts assessment of plug-in hybrid vehicles on
electric utilities and regional US power grids part 2: economic assessment.
December 2007. p. 2016.
[2] Karabasoglu O, Michalek J. Influence of driving patterns on life cycle cost and
emissions of hybrid and plug-in electric vehicle powertrains. Energy Policy
2013;60:445e61.
[3] Vergis S, Turrentine TS. Plug-in electric Vehicles: a case study of seven mar-
kets. October, 2014.
[4] U.S. Department of Energy. All-electric vehicles. 2015 [Online]. Available:
Www.Fueleconomy.Gov. https://www.fueleconomy.gov/feg/evtech.
shtml#data-sources [accessed 28.07.15].
[5] Nykvist B, Nilsson M. Rapidly falling costs of battery packs for electric vehicles.
Nat Clim Chang 2015;5:100e3. no. April.
[6] Siler-Evans K, Azevedo IL, Morgan MG. Marginal emissions factors for the U.S.
Electricity system. Environ Sci Technol May 2012;46(9):4742e8.
[7] Graff Zivin JS, Kotchen MJ, Mansur ET. Spatial and temporal heterogeneity of
marginal emissions: implications for electric cars and other electricity-shifting
policies. J Econ Behav Organ 2014;107:248e68.
[8] Sieminski A. Fuels used in electricity generation - EIA. In: U.S. Nuclear infra-
structure council; 2013.
[9] Tamayao M-AM, Michalek JJ, Hendrickson C, Azevedo IML. Regional variability
and uncertainty of electric vehicle life cycle CO 2 emissions across the United
States. Environ Sci Technol Jul. 2015;49(14):8844e55.
[10] Cao Y, Tang S, Li C, Zhang P, Tan Y, Zhang Z, et al. An optimized EV charging
model considering TOU price and SOC curve. IEEE Trans Smart Grid 2012;3(1):
388e93.
[11] Korolko N, Sahinoglu Z. Robust optimization of EV charging schedules in
unregulated electricity markets. IEEE Trans Smart Grid 2015:1e9.
[12] Yuksel T, Michalek JJ. Effects of regional temperature on electric vehicle effi-
ciency, range, and emissions in the United States. Environ Sci Technol
2015;49(6):3974e80.
[13] Weis A, Michalek JJ, Jaramillo P, Lueken R. Emissions and cost implications of
controlled electric vehicle charging in the U.S. PJM interconnection. Environ
Sci Technol 2015;49(9):5813e9.
[14] Khayyam H, Bab-Hadiashar A. Adaptive intelligent energy management sys-
tem of plug-in hybrid electric vehicle. Energy 2014;69:319e35.
[15] Sovacool BK, Hirsh RF. Beyond batteries: an examination of the benefits and
barriers to plug-in hybrid electric vehicles (PHEVs) and a vehicle-to-grid
(V2G) transition. Energy Policy 2009;37(3):1095e103.
[16] Guille C, Gross G. A conceptual framework for the vehicle-to-grid (V2G)
implementation. Energy Policy 2009;37(11):4379e90.
[17] Honarmand M, Zakariazadeh A, Jadid S. Optimal scheduling of electric vehi-
cles in an intelligent parking lot considering vehicle-to-grid concept and
battery condition. Energy 2014;65:572e9.
[18] Sioshansi R, Denholm P. Emissions impacts and benefits of plug-in hybrid
electric vehicles and vehicle-to-grid services. Environ Sci Technol 2009;43(4):
1199e204.
[19] Cardoso G, Stadler M, Bozchalui MC, Sharma R, Marnay C, Barbosa-Po voa A,
et al. Optimal investment and scheduling of distributed energy resources with
uncertainty in electric vehicle driving schedules. Energy 2014;64:17e30.
[20] The Office of Energy Efficiency and Renewable Energy. Alternative fuels data
center: maps and data. 2015 [Online]. Available: http://www.afdc.energy.gov/
data/ [accessed 28.07.15].
[21] Nissan. Nissan LEAF global sales reach 100,000 units. 2014 [Online]. Available:
http://nissannews.com/en-US/nissan/usa/releases/nissan-leaf-global-sales-
reach-100-000-units?la¼1.
[22] Zhou Y, Wang M, Hao H, Johnson L, Wang H, Hao H. “Plug-in electric vehicle
market penetration and incentives: a global review. Mitig Adapt Strateg Glob
Chang 2014:777e95.
[23] Advanced Powertrain Research Facility at Argonne National Laboratory.
Downloadable dynamometer database. U.S. Department of Energy; 2015
[Online]. Available: http://www.transportation.anl.gov/D3/ [accessed
01.01.16].
[24] Sears J, Roberts D, Glitman K. A comparison of electric vehicle Level 1 and
Level 2 charging efficiency. In: 2014 IEEE conference on technologies for
sustainability (SusTech); 2014. p. 255e8.
[25] Li G, Eisele M, Lee H, Hwang Y, Radermacher R. Experimental investigation of
energy and exergy performance of secondary loop automotive air-
conditioning systems using low-GWP (global warming potential) re-
frigerants. Energy Apr. 2014;68:819e31.
[26] Sieminski A. Fuels used in electricity generation - EIA. In: U.S. Nuclear infra-
structure council; 2013.
[27] FHWA. 2009 national Household travel Survey. Federal Highway Adminis-
tration; 2009.
[28] Reyna JL, Chester MV, Ahn S, Fraser AM. Improving the accuracy of vehicle
emissions profiles for urban transportation greenhouse gas and air pollution
inventories. Environ Sci Technol 2015;49(1):369e76.
[29] U.S. Energy Information Administration. Variability in electricity demand
highlights potential roles for electricity storage. 2013 [Online]. Available:
http://www.eia.gov/todayinenergy/detail.cfm?id¼13131.
[30] Marano V, Onori S, Guezennec Y, Rizzoni G, Madella N. Lithium-ion batteries
life estimation for plug-in hybrid electric vehicles. In: Veh. Power propuls.
Conf. 2009. VPPC ’09. IEEE; 2009. p. 536e43.
[31] U.S. Environmental Protection Agency. EPA and NHTSA set standards to
reduce greenhouse gases and improve fuel economy for model years 2017-
2025 cars and light trucks. 2012. p. 1e10. no. August 2012.
[32] University of Michigan Transportation Research Institute. Sustainable
Worldwide Transportation - sales-weighted fuel economy (window sticker).
2015 [Online]. Available: http://www.umich.edu/~umtriswt/EDI_sales-
weighted-mpg.html [accessed 28.07.15].
[33] Attari S. Public perceptions of energy consumption and savings. Pnas 2010:
1e16.
[34] Hawkins TR, Singh B, Majeau-bettez G, Strømman AH. Assessment of con-
ventional and electric vehicles 2012;17(1):53e64.
[35] Michalek JJ, Chester M, Jaramillo P, Samaras C. Valuation of plug-in vehicle
life-cycle air emissions and oil displacement benefits 2011;108(40).
[36] Tesla motors, model S. 2015 [Online]. Available: http://www.teslamotors.
com/models [accessed 28.07.15].
[37] Nissan. Nissan LEAF electric car. 2015 [Online]. Available: http://www.
 C.G. Hoehne, M.V. Chester / Energy 115 (2016) 646e657 657

nissanusa.com/electric-cars/leaf/ [accessed 28.07.15]. https://www.chevrolet.com/content/dam/Chevrolet/northamerica/usa/

[38] Chevrolet. Chevrolet volt. 2015 [Online]. Available: http://www.chevrolet.
com/volt-electric-car.html [accessed 28.07.15].
[39] Tomic J, Kempton W. Using fleets of electric-drive vehicles for grid support.
J Power Sources 2007;168(2):459e68.
[40] Chevrolet. 2012 chevrolet volt owner manual. 2012 [Online]. Available:
nscwebsite/en/Home/Ownership/Manuals_and_Videos/02_pdf/2k12volt.pdf.
[41] Advanced Powertrain Research Facility at Argonne National Laboratory.
Downloadable dynamometer database. U.S. Department of Energy; 2015.
[42] U.S. Environmental Protection Agency. Greenhouse gas emissions from a
typical passenger vehicle. 2011. p. 1e5.