Applied Energy 183 (2016) 1142–1156

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Impacts of plug-in electric vehicles in a balancing area

Ghazal Razeghi ⇑, Scott Samuelsen
Advanced Power and Energy Program (APEP), University of California, Irvine, CA 92697-3550, United States

h i g h l i g h t s

Unit commitment methodology is used to determine BEV impact on electricity market.

Roles of charging profile, dispatch strategy and interconnecting area are assessed.

Results show that impact of BEV on cost of electricity generation is small.

Controlled BEV charging can lower emissions intensity of the grid and MCP.

BEV deployment helps reduce overall criteria pollutant emissions.

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

Article history: High contributions of the electricity generation and transportation sectors to criteria pollutant and green-
Received 6 July 2016 house gas emissions have resulted in an increased interest and shift towards low to non-carbon genera-
Received in revised form 9 September 2016 tion options such as renewable wind and solar, and alternative transportation options including plug-in
Accepted 24 September 2016
electric vehicles. Since plug-in electric vehicles transfer the tailpipe emissions to the electric grid, it is
Available online 30 September 2016
important to study the interaction between the two sectors. In this paper, a previously developed spa-
tially and temporally resolved unit commitment model is used to determine the dispatch schedule of
Keywords:
resources with and without battery electric vehicles for 2050 in a fictitious balancing area located within
Plug-in electric vehicles
Grid modeling
the South Coast Air Basin of California. Cases studied include various charging profiles, penetration in
Cost of electricity light-duty fleet, imports mix, and grid dispatch strategies. Results of the analysis include average cost
Emissions of electricity production, market clearing price, temporal production of individual generators, and emis-
Unit commitment sions from electricity generation and the transportation sectors.
Economic impacts The results show that deploying battery electric vehicles (1) has little impact on the average cost of
electricity generation-maximum of $2.5 per MW h for the cases studied with 40% penetration in the
light-duty fleet, (2) reduces the overall criteria pollutant emissions except for one case, and (3) results
in a smoother load profile, reduces the use of peaking units, and reduces the average emission intensity
of the grid through controlled off-peak charging.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction tants and GHG emissions [3] with the transportation sector
accounting for 37 percent of GHG emissions in 2013 [4], and 49
Concerns about air quality, energy security, and global climate percent of NOx emissions in 2012 [5]. Thus, the transportation sec-
change have led to more stringent energy and environmental reg- tor is a major contributor to air pollution, a major contributing fac-
ulations to reduce energy consumption and emissions from both tor to chronic diseases and mortality impacting public health.
mobile and stationary sources. In 2014, the transportation sector Several pathways are available to address this issue: (1) increasing
accounted for 28 percent of energy consumption in the United the efficiency and thereby reducing tailpipe emissions of conven-
States [1] and it is estimated that the transportation energy use tional vehicles, (2) reducing the transportation demand by chang-
and greenhouse gas (GHG) emissions have increased 28 percent ing life style through public transit, a reduction in commute time
worldwide since 2000 [2]. by living closer to workplace, and consuming local produce, and
In the state of California, the electricity generation and trans- (3) implementing regulations to reduce emissions (such as Califor-
portation sectors are the main contributors to both criteria pollu- nia Assembly Bill 32 that requires a reduction to 1990 levels by
2020 in GHG emissions), and encourage alternative low and non-
⇑ Corresponding author. carbon transportation options (such as Assembly Bill 118 that sup-
E-mail address: [email protected] (G. Razeghi). ports alternative and renewable fuel and vehicle technologies [6]).

http://dx.doi.org/10.1016/j.apenergy.2016.09.063
0306-2619/Ó 2016 Elsevier Ltd. All rights reserved.
 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156 1143

Nomenclature

BEV Battery Electric Vehicle PHEV Plug-in Hybrid Electric Vehicle

CAISO California Independent System Operator SoCAB South Coast Air Basin
GHG Greenhouse Gas V2G Vehicle to Grid
LCOE Levelized Cost of Energy VMT Vehicle Miles Traveled
MCP Market Clearing Price
NERC North American Electric Reliability Corporation
PEV Plug-in Electric Vehicle

All these factors have resulted in regulatory initiatives to associated with generating units including part-load, start-up,
replace internal combustion engine vehicles with alternative, and ramping emissions.
lower emitting options. One of the options considered is plug-in
electric vehicles (PEVs) as a feasible, near-term option [7] that will 2. Approach
prepare consumers for a future that includes fuel cell vehicles, pub-
lic transit, and shared autonomous vehicles [8]. PEVs, which 2.1. Dispatch model
include plug-in hybrid electric vehicles (PHEV) and purely battery
electric vehicles (BEV), have the benefits of reduced liquid fuel A detailed market model including a dispatch model and multi-
usage [9], lower overall criteria pollutant emissions [10,11], ple modules was developed for the South Coast Air Basin (SoCAB)
improved air quality [12,13], utilization of idle and stranded as a balancing area to mimic the operations of the electricity mar-
resources, reduced GHG emissions [14], and a less expensive ket and business in the state of California while taking into account
source of mobility than gasoline on a per mile basis. the physical constraints of the system. In this methodology renew-
Since PEVs are directly connected to the grid, understanding the able resources are treated as must-take, and imports are
interaction between the transportation and electricity sectors is dispatched ahead of in-basin units to mimic actual market opera-
important to correctly characterize the impacts of deploying these tions. Note that in practice, imports are settled ahead of time and
vehicles. One group of studies, focused on the generation side of the negotiated price is usually lower than the spot market price
the electricity grid, suggest that large-scale deployment of PEVs (similar to bilateral contracts). This is simulated in this study by
will have limited negative impacts on the electric power system dispatching them first.
in terms of additional generation requirements [15,16], and that Studying historical data from the state of Californian and SoCAB
they will have positive impacts on emissions, the extent of which [30,31], it is concluded that the imports, more or less, follow the
depends on the charging profile, charging level, and the grid mix same profile as the demand [29]. Moreover, for the following rea-
[17]. Other groups of studies address the interaction of the PEVs sons it is assumed that transmission capacity for imports to SoCAB
with the distribution system [18], distribution transformer [19], remains unchanged: (1) Obtaining licenses required for building
and distribution substations [20], as well as the implications of transmission lines takes a long time, (2) environmental concerns,
vehicle-to-grid (V2G) [21,22] and using V2G to increase penetra- and (3) assessing an air quality episode in the basin requires a high
tion of intermittent resources [23,24]. generation inside the basin. This capacity is determined by assess-
Studies have also explored the impacts of PEVs on regional ing historical data [10].
electricity markets and ISOs, and focus on the appropriate regula- During off-peak hours, it is assumed that the imports are pro-
tions that need to be implemented to facilitate the integration of vided by load-following units with capacity factor projected for
PEVs [25,26]. However, there are few studies that assess the the year under study [32,33], and during peak hours, the extra
impact of PEVs on electricity market prices [27,28]. As a result, a import is assumed to be provided by peaking (and more expensive)
need exists to determine the interaction between the electricity units. In all cases, a 5% transmission loss in assumed for the
generation and transportation sectors in a manner that represents imports [34]. When the amount of imports is less than 10% of
real-life electricity market operations and captures the many the total import capacity, it is assumed that the generators provid-
physical constraints in order to (1) assess the impacts on the cost ing the imports are operating at minimum allowable capacity
of generation and operation of the grid, and (2) impact on overall factor, resulting in increase in the price of imports per MW h.
emissions. After dispatching must-take units and imports, conventional
A comprehensive spatially and temporally resolved dispatch in-basin units are dispatched using a unit commitment dispatch
model, based on unit commitment with market operations and model. The objective of the economic dispatch is to minimize the
associated physical constraints [29], is used in this paper. Various social cost of the market as shown in Eq. (1).
cases with different BEV charging profiles, and dispatch strategies
are studied. Three dispatch strategies are assessed, one with an X
Ng

economic dispatch strategy and two with environmental strategies Minimize ½C i ðPði; tÞÞIði; tÞ þ SðiÞIði; tÞfIði; tÞ  Iði; t  1Þg ð1Þ
i¼1
in which minimizing the overall emissions from generating units
is the objective instead of cost of the system. The environmental In this equation, Ng is the number of generators participating in
dispatch strategies (1) provide an opportunity to reduce the envi- the market, Ci is the cost function of generating unit i which is
ronmental impacts with neither investment nor change to the grid, equal to the Levelized Cost of Energy (LCOE) associated with that
and (2) indicate the maximum achievable reduction in a specific unit which is itself a function of the capacity factor. P(i,t) is the pro-
species associated with the grid mix and design under study, duction (generation) of unit i at time t, S(i) is the start-up cost of
thereby facilitating realistic roadmaps in the future. unit i, and finally I(i,t) is the commitment status of unit i at time t.
Role of imports in the overall results is also explored. The eco- Constraints of the system include matching demand and gener-
nomic metrics of each case-including average cost of generation, ation, minimum and maximum capacity factor of each generating
and market clearing price (MCP) are determined and compared. unit, ramping up and down limits, minimum on/off time and trans-
The dispatch schedules are then used to determine the emissions mission line constraints. A detailed description of the dispatch
1144 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156

model and assumptions can be found in Razeghi et al. [29]. Sum- 10%
mary of the methodology is provided in Appendix A. 9%
Another resource dispatch strategy is environmental dispatch
which, instead of minimizing the cost, a specific criteria pollutant 8%
or greenhouse gas is minimized as shown in Eq. (2). In this equa- 7%
tion EM(i,j) is the emission factor for generator (i) and associated
6%
with specific species j, and SE(i,j) is the start-up emission of that
generator replacing the start-up cost in Eq. (1). 5%
4%
X
Ng
Minimize ½EMði; jÞðPði; tÞÞIði; tÞ þ SEði; jÞIði; tÞfIði; tÞ  Iði; t  1Þg 3%
i¼1
2%
ð2Þ
1%
The following two environmental dispatch strategies are studied
0%
in this paper: NOx, and CO2 dispatch, minimizing the emission of 1 3 5 7 9 11 13 15 17 19 21 23
nitrogen oxides (NOx) and CO2e, respectively. Hour Ending
An important input of the model is the electricity demand asso- Uncontrolled Off-peak
ciated with future scenarios. The base electricity demand is fore-
casted based on the methodology discussed and used in Fig. 1. Normalized charging profile for a BEV fleet.
[10,12,29] which is based on population growth, census data, and
historical electricity demand and is consistent with projections
made by California Energy Commission [35,36]. National
Household
2.2. PEV electricity demand Travel
Year,Date Survey
In order to assess the impacts of PEVs which include both
PHEVs with lower all-electric range and equipped with a gasoline
engine, and BEVs with a greater all-electric range and thus greater
electricity demand, a transportation module is developed and inte- Vehicle PEV Charging
grated with the dispatch model. The characteristics of the future Type Penetration Profile
light-duty fleet are extracted from various sources including the
results of Air Resources Board (ARB) EMFAC model [37] which
takes into account the improvements in the conventional vehicles’
efficiency and tail-pipe emissions associated with future years.
These results include projections of the numbers of vehicles in each
fleet in future years, average daily miles traveled per vehicle and its Transportation
distribution throughout the day, and tailpipe emission factors. Tail- Module
pipe emission factors, daily Vehicle Miles Traveled (VMT), and VMT
daily distribution are used to calculate temporal emissions from
conventional vehicles (including both gasoline vehicles and
hybrids) for a specific fleet size and year.
Two charging profiles, used previously in various studies, are
used here: ‘‘uncontrolled charging” in which drivers have the PEV Emissions
option to charge at home and work and plug in their vehicles as Demand From
soon as they get to their destination; and an ‘‘off-peak charging Profile Vehicles
profile” where the charging is moved to hours with lower overall
electricity demand [10,12]. In all cases, home charging and 50%
Fig. 2. Transportation module.
of work charging is at Level 1, and the other 50% of work charging
is at Level 2. The charging profiles used are similar to those used in
other studies [10,12,38]. Travel survey data [39,40] are used to 40 miles or less, and 95% of drivers have a daily VMT of 100 miles
determine arrival and departure times, as well as dwelling dura- or less [39,40], and thus the 100 mile all electric range selected
tion, and the miles traveled. These data are used to determine here, can be used by the majority of drivers for their everyday
the charging profiles for uncontrolled and off-peak charging asso- use. In the base case, the penetration of PEVs is insignificant (less
ciated with each driver. The fleet electricity demand is then calcu- than 0.1%, equal to 2010 penetration), in all other cases BEVs pen-
lated. Normalized electricity demand associated with a fleet of etration in the light-duty fleet is increased by replacing same
BEVs in the balancing area under study is shown in Fig. 1. amount of conventional vehicles. The conventional vehicles
In the transportation module, the year of the scenario, penetra- replaced by BEVs have emission factors equal to the average emis-
tion of BEV, PHEV, or both, and the charging profile are given as sion factors of conventional vehicles in the fleet. For the majority of
inputs (see Fig. 2). The electricity demand of the PEVs (time the cases shown in this paper, BEV penetration is 40%. This pene-
resolved), and the emission associated with the fleet (also time tration of BEVs in light-duty fleet is consistent with several studies
resolved) are calculated. The electricity demand of these vehicles conducted on the feasible penetration of alternative electric vehi-
is then added to the base electricity demand for balancing supply cles [42,43].
and demand. In this paper, the results shown are associated with In this study, the alternative vehicles are assumed to be
BEVs with 100 mile (160 km) range and characteristics of a typical deployed uniformly throughout the area. This assumption does
BEV; 0.31 kW h/mi (DC) consumption and 0.85 charging efficiency not impact the results since this study focuses on the grid as a
[41]. In the state of California 70% of drives have a daily VMT of whole. This notwithstanding; a concentrated deployment of PEVs
 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156 1145

can have a negative impact on the distribution system (see, for 18

example, Razeghi et al. [19] and Ramos Muñoz et al. [44]). 16
14
3. Results and discussion 12
10

GW
3.1. Imports
8
Using the demand profiles, the imports profiles are determined 6
using the approach described in Section 2.1 and in [29]. The results
4
for 40% BEV penetration are shown in Fig. 3 for a representative
summer day in 2050 (a season selected as susceptible to high elec- 2
tricity demand, high temperatures, and associated susceptibility to 0
an air quality episode). As it can be seen in this figure, the majority 5 10 15 20
of the BEV electricity demand with off-peak charging will be sup- Hour Ending
ported by imports since they are dispatched first and the transmis- Available in-basin resources
sion capacity is not reached. In Fig. 3, a fourth profile (profile 4) is
In-basin generation required, Uncontrolled
shown which will be explored in cases with high wind penetration In-basin generation required, Off-peak
in order to reduce possible curtailment and will be discussed in fol-
lowing sections. Fig. 4. Comparison of available in-basin resources and additional in-basin gener-
ation required to serve 40% BEV penetration.

3.2. Additional generation

capacity factor, ramp rate, and min on/off times. For this estimate,
The methodology developed can also be used to determine the the transmission constraint is relaxed.
optimized required new generation and locating them based on Next, the BEV load is added and the import module previously
the future demand, and was previously used to determine the described is used to establish the associated increase in the output
additional generation required to serve the future electricity of in-basin generating units. Then 10% is added to this amount to
demand in the absence of BEV demand [29]. To ensure that suffi- account for the reserve requirements for this new load as previ-
cient reserves and ancillary services are available, the historical ously mentioned. In Fig. 4 the results are shown for 40% BEV pen-
operating reserves in the California Independent System Operator etration on a representative summer day in 2050, and are
(CAISO) [31] are studied together with North American Electric compared to the available in-basin generation.
Reliability Corporation (NERC) requirements for various types of Note that the deployment of BEVs does not result in a require-
ancillary services. Consistent with historical trends in CAISO and ment for additional in-basin generation at early hours of the day.
in order to have a conservative estimate at each time interval, 10 The extra demand is served instead by the imports to the balancing
percent is added to the demand to account for ancillary services area because (1) imports are dispatched first, and (2) and transmis-
and reserves. sion capacity is available. As a matter of fact, in the case with off-
With the addition of the BEV demand, it is necessary to first peak charging, the majority of the BEV load is served by imports. As
determine if the installed capacity is sufficient to serve the new a result, the emissions associated with the vehicles are transferred
electricity demand. To this end, the available resources in the bal- to outside of the basin.
ancing area over and above serving the base electricity demand and Between 2 pm and 5 pm, the available in-basin resources are
ancillary services (i.e., in the absence of BEV charging) are deter- not sufficient to provide all the extra demand associated with the
mined. This is accomplished by running the dispatch model for 40 percent BEV with uncontrolled charging. Therefore, another
the base case and then using the results to maximize the output two and half GWs of generation are added to the basin. Various
of the remaining generators subject to the constraints of min/max combinations of load-following and peaking units are added to
the model to determine which combination satisfies the electricity
demand completely, with the limiting factors being the ramping
rate, and minimum on-time. The location of these units will then
be determined based on land-use and footprint of the power plant
[10].
Included are 500 MW of combined cycle (one 500 MW) and one
and half GW of gas turbines (six 200 MW and three 100 MW). With
the addition of these generating units, the in-basin resources are
able to serve the load entirely. The dispatch model is then used
to develop dispatch schedules for the economic, NOx, and CO2 dis-
patch strategies for cases shown in Table 1 for year 2050.

3.3. Generation mix and schedule

The dispatch schedules associated with the eleven 2050 cases

are presented in Fig. 5 for a representative summer day. The results
shown have a temporal resolution of 10 min.
The uncontrolled charging of BEVs is shown to increase the use
of peaking units regardless of the dispatch strategy used or solar
penetration. This is expected since electricity demand of the vehi-
Fig. 3. Imports, 40% BEV penetration, 2050. cles in these cases coincides with the demand peak and thus the
1146 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156

Table 1
Description of cases.

Case number Dispatch strategy BEV Penetration Charging scenario Wind Solar Import profile (Fig. 3)
1 (Base Case) Economic Negligible Not applicable 7% 1% 1
2 Economic 40% Uncontrolled 7% 1% 2
3 Economic 40% Off-peak 7% 1% 3
4 NOx Negligible Not applicable 7% 1% 1
5 NOx 40% Uncontrolled 7% 1% 2
6 NOx 40% Off-peak 7% 1% 3
7 CO2 Negligible Not applicable 7% 1% 1
8 CO2 40% Uncontrolled 7% 1% 2
9 CO2 40% Off-peak 7% 1% 3
10 Economic 40% Uncontrolled 7% 15% 2
11 Economic 40% Off-peak 33% 1% 3
12 Economic 40% Off-peak 7% 1% 1
13 Economic 40% Off-peak 33% 1% 4

load following units are not capable of picking up all of the addi- 3.4. Economic impacts
tional load due to the high ramping requirement. Off-peak charg-
ing on the other hand, reduces the use of peaking units. With the In this section, the impacts of BEV deployment on the cost of
addition of the BEV load in these cases, the overall load profile electricity in a balancing area are assessed through presenting
becomes smoother (with lower rate of change) and thus the the average cost of the grid and market clearing price (MCP). Elec-
load-following units will be capable of carrying this extra load tricity markets practice two methods for making payments to the
instead of firing new peaking units without violating their ramping generating units that are cleared in the market: (1) pay-as-bid:
constraints. The number of online units at each hour (for economic in which the generators are paid based on their bid which in a
dispatch), shown in Fig. 6, further demonstrates that while the competitive market should be close to the cost of generation, and
uncontrolled charging results in generators starting up (peaking (2) market clearing price in which all generator get paid the same
units), with off-peak charging the number of online generating price $/MW h. MCP is the price of the most expensive generators
units are even lower than the case without any BEVs. This is again that is serving the demand. In an economic dispatch strategy, this
due to the fact that, with a smoother load profile, load following will be the last generators that gets cleared in the market. Another
units are capable of serving the majority of the load without the way to explain the MCP is that it is the price that all supplier oper-
need to fire peaking units. Other dispatch strategies follow the ate at or below this price. Average cost is calculated by dividing the
same trend shown in Fig. 6 as well. The output of load following payments to all generators (including start-up costs) based on the
units increases in all cases with the addition of BEVs. bid they had placed, by the demand (pay-as bid method).
Environmental dispatch strategies result in increased use of In this study, it is further assumed that the electricity market is
some peaking units. This is due to the fact that advanced peaking a competitive market and bids represent the cost of generation and
units have lower emission factors compared to the aging units thus do not depend on time of day or electricity demand.
and the model takes into account the loss of efficiency of generat-
ing units with time. This has been further explained in Razeghi 3.4.1. Average cost
et al. [29]. The average electricity cost of conventional generating units
Fig. 5 also affirms that, as mentioned before, the majority of the inside the balancing area for cases with economic dispatch strategy
BEV demand with off-peak charging in Case 3 is served by the is shown in Fig. 9. For these cases with low renewable penetration,
imports. the cost shown does not include the payment to must-take resources
Before discussing the economic and environmental effects of and only includes generators that are competing in the market.
BEVs, the off-peak charging of vehicles is addressed (Case 12 and The average cost of electricity generation increases at peak
Case 13). In contrast to Cases 1–11, in which the import profile times in Case 2 as was expected due to BEV charging during peak
was determined using the method explained, in Case 12 it is hours resulting in generator start-ups. In Case 3, the price is
assumed that all the BEV demand is supported by in-basin units slightly increased when the vehicles are being charged. However,
resulting in the same imports profile as the base case. In Case 13, prices during on-peak hours are reduced because of reduced use
the amount of imports at early hours of the day is further lowered of peaking units and a smoother demand profile. Case 12, on the
(see Fig. 3) to reduce the amount of curtailed wind at these hours. other hand, results in greater increase in prices during off-peak
The dispatch schedules for Case 12 and Case 13 are shown in hours when the BEVs are being charged and all their demand is sat-
Fig. 7 with the economic dispatch strategy. It is shown that the isfied by in-basin generating units. The increase in prices during
electricity demand of the BEVs at early hours of the day results early hours of the day in Case 12 is also due to new generators
in an additional peak in the load daily profile. The use of peaking coming online as shown in Fig. 8.
units in Case 12 is reduced during peak demand times for reasons Overall, the average change of electricity generation costs in
previously discussed in regards to Case 3. The number of online one day with the addition of BEVs to the electricity demand is less
units for Case 12 is shown in Fig. 8 along with Case 1 for compar- than 1 $/MW h which is not significant with the Case 2 resulting in
ison. During off-peak hours when the BEVs are being plugged in, higher prices compared to both Case 1 and Case 3.
the number of online units is greater than Case 1 suggesting that The average cost for electricity generation is shown in Fig. 10 for
new generators are fired up to serve the added load. cases with CO2 dispatch. The average cost using this dispatch strat-
In Case 13, on the other hand, the use of peaking units is egy is higher than that of economic dispatch, since in economic
increased almost throughout the entire day. In this case, the use dispatch the cost in Eq. (1) is minimized where in environmental
of wind resources is increased compared to Case 11 at early hours dispatch strategies, units with lower emission factors that are usu-
of the day and thus more peaking units are required to handle the ally more expensive are dispatched and called upon first in the
intermittencies introduced to the grid. In this case, the load- market. With the addition of BEVs, the average cost of electricity
following units are also operating more dynamically. generation increases slightly during the most of the day. The
 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156 1147

10
4 Case 1 10
4 Case 2 10
4 Case 3

2 2 2

1.5 1.5 1.5

MW

MW
MW
1 1 1

0.5 0.5 0.5

0 0 0
5 10 15 20 5 10 15 20 5 10 15 20
Hour Ending Hour Ending Hour Ending

10
4 Case 4 10
4 Case 5 10
4 Case 6

2 2 2

1.5 1.5 1.5

MW

MW
1 MW 1 1

0.5 0.5 0.5

0 0 0
5 10 15 20 5 10 15 20 5 10 15 20
Hour Ending Hour Ending Hour Ending

10
4 Case 7 10
4 Case 8 10
4 Case 9

2 2 2

1.5 1.5 1.5

MW

MW
MW

1 1 1

0.5 0.5 0.5

0 0 0
5 10 15 20 5 10 15 20 5 10 15 20
Hour Ending Hour Ending Hour Ending

10
4
Case 10 10
4
Case 11

2 2

1.5 1.5
MW
MW

1 1

0.5 0.5

0 0
5 10 15 20 5 10 15 20
Hour Ending Hour Ending

Fig. 5. Dispatch schedule.

increase at peak hours with off-peak charging is due to the fact For cases with high renewable penetration, it is necessary to
that, with the addition of BEV charging, the overall electricity include the cost of these units since they have significant impact
demand becomes smoother as the contribution of cleaner and thus on the price. In Cases 10, 11, and 13, the cost of wind and solar
more expensive generating units to the grid mix becomes greater is included in determining the average cost by using the forecasted
compared to the case with no BEV, thus increasing the average cost LCOE [32,33,45]. The change in average cost of generation for cases
of generation. The average increase in cost of electricity generation with high renewable penetration compared to base case (Case 1) is
is slightly less 1.5 $/MW h for Case 8 and Case 9 compared to Case shown in Fig. 11. As expected, in Case 10 with high solar penetra-
7. Cases studied with NOx dispatch result in similar trends as tion the cost of generation is increased at peak times because: (1)
shown in Fig. 10 and Cases 5 and 6 result in an average 2.5 $/ solar generation is available and expensive, (2) uncontrolled charg-
MW h increase in cost of generation compared to Case 4. ing of vehicles results in increased use of peaking units. At other
1148 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156

250

200

Number
150

100

50

0
1 3 5 7 9 11 13 15 17 19 21 23
Hour Ending
Case 1 Case 2 Case 3

Fig. 6. Number of online generating units.

10 4 Case 12 unchanged or is slightly reduced because of reduced use of peaking

units as a result of off-peak charging of BEVs. In Case 13 where the
imports at early hours of the day are significantly reduced, the
2 average cost of generation is slightly higher than Case 11 because
of increased use of wind and reduced curtailment. Between 8 and
10 am, the cost of Case 13 is significantly higher than Case 11. At
1.5
these hours, because of the import profile selected (see Fig. 3),
MW

the imports require a high ramping rate and thus are more expen-
1
sive. Moreover, the contribution of peaking units is higher in Case
13 compared to Case 11 (Fig. 7). Both of these factors result in a
higher overall cost of generation. At other times, average cost of
0.5 generation associated with Case 13 is higher than Case 11 due to
increased use of peaking units.
To assess the impact of BEV penetration on average costs, the
0 simulations are repeated for BEV penetrations between 0 and
5 10 15 20
60% in increments of 10%. Changes in summer average cost of gen-
Hour Ending
eration are shown in Fig. 12 compared to the base case (Case 1) for
Case 13 the scenarios simulated as a function of BEV penetration.
10 4
Results from Fig. 12a indicate that average cost of generation
increases with BEV penetration, and for a specific dispatch strat-
2 egy, uncontrolled charging results in higher cost of generation
compared to off-peak charging except for a scenario in which all
the BEV demand is served by in-basin generating units (VII). In this
1.5 case at low BEV penetrations, additional in-basin generating units
need to come online to serve the load. At higher BEV penetrations,
MW

however, the smoother load profile results in reduced use of peak-

1
ing units and the cost of generation will be lower than the uncon-
trolled charging. Low penetrations of BEVs (20% and lower) have a
0.5 small impact on the cost of generation, as the penetration increases
so does the rate of change in the cost until 50% penetration is
reached after which the sensitivity of the cost to penetration is
0 reduced (the rate of change is decreased). CO2 dispatch results in
5 10 15 20
higher costs compared to NOx dispatch at low BEV penetration,
Hour Ending at higher penetrations, the costs associated with the two environ-
mental dispatch strategies are very close.
Fig. 12b includes scenarios with high penetration of renewable
resources. With high solar penetration, addition of BEVs with
uncontrolled charging results in increased use of peaking units
and the cost of generation increases with BEV penetration (VIII in
Fig. 7. Dispatch schedule Cases 12 and 13.
Fig. 12b). With higher wind penetration and off-peak charging, at
low penetrations, BEVs result in a reduction in cost of generation
times of the day, the impact on the average cost of generation is because they reduce the amount of curtailment and the use of
relatively small. peaking units. At higher BEV penetrations, however, the cost goes
In Case11 with 33% wind penetration, the average cost of gen- back up due to high contribution of wind in the mix and increased
eration is increased early in the morning and late at night when use of peaking units to handle the intermittencies. Scenario X
wind is available. At peak times, the average cost remains results in higher cost of generation due to high contribution of
 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156 1149

200
180
160
140
120

Number
100
80
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour Ending
Case 1 Case 12

Fig. 8. Number of online units in Case 12.

62

60

58
$/MWh

56

54

52

50
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour Ending
Case 1 Case 2 Case 3 Case 12

Fig. 9. Average cost associated with conventional generating units-Economic dispatch.

64

62

60
$/MWh

58

56

54

52

50
1 3 5 7 9 11 13 15 17 19 21 23
Hour Ending
Case 7 Case 8 Case 9

Fig. 10. Average cost associated with conventional generating units-CO2 dispatch.

wind to serving the load as well as more expensive imports which expensive generator being a load-following combined cycle unit
were discussed previously. with low capacity factor. With the addition of BEV electricity
demand, new generators (load-following units) come online but
3.4.2. Market clearing price they operate at high capacity factors and thus their cost of genera-
MCP for cases with economic dispatch is shown in Fig. 13. Over- tion (per MW h) is lower compared to Case 1. The reduction during
all, addition of BEVs with off-peak charging results in a reduction in peak demand times in Case 3 and Case 12, was expected because of
MCP with Case 12 having the biggest impact. Although in this case, reduction in peaking unit outputs, as well as reduced rate of change
the average cost of electricity generation is increased in early hours in the demand profile discussed before. With uncontrolled charg-
of the day (Fig. 9), the MCP is reduced during these hours. Before the ing, the MCP is increased at peak times due to increased use of peak-
addition of BEVs, low demand at these hours results in the most ing units that are operating at low capacity factors. Towards the end
1150 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156

32
28
24
20

$/MWh
16
12
8
4
0
-4 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour Ending
Case 10 Case 11 Case 13

Fig. 11. Change in average cost (including renewable) compared to Case 1- Economic dispatch.

(a) 8 As was previously mentioned, in the results shown so far, the

I:Economic,Uncontrolled
market is assumed to be competitive and the bids are equal or very
II:Economic,Off-peak
7 close to actual cost (i.e., the generators’ bids do not change with
III:NOx,Uncontrolled time of day). Another scenario is assessed where the generators
6 IV:NOx,Off-peak adjust their bids according to the demand profile with their bids
V:CO2,Uncontrolled never lower than the actual cost. The MCP results for this bidding
5 VI:CO2,Off-peak scenario (time-dependent bidding) and CO2 dispatch strategy are
$/MWh

VII:Economic, Off-peak, Import 1 shown in Fig. 14. With this bidding strategy, BEV deployment
4 results in an average increase in MCP of less than 15 $/MW h with
uncontrolled charging, and average reduction of less than 10 $/
3 MW h with off-peak charging. As expected, the MCP is higher when
the demand is high. With uncontrolled charging, the highest MCPs
2
are observed in the morning and in the afternoon coinciding with
the work and home charging (Fig. 1). Off-peak charging results in a
reduction in MCP during the majority of the day compared to Case
1
7 except at hours when units come online and once early in the day
and once in the afternoon.
0
0 10 20 30 40 50 60
BEV Penetration 3.5. Environmental impacts

(b) 18 VIII:15%Solar, Uncontrolled The results from the dispatch model include the temporally
IX:33%Wind, Off-peak resolved output of individual generators. These results are then
16
X:33%Wind, Off-peak, Import 4 used in an emissions module developed as part of the strategy
14 and described in [29]. This module includes the emission factors
for individual generators for criteria pollutants such as NOx, SOx,
12
and CO, and also greenhouse gas emissions (CO2e).
$/MWh

10 Emission factors for a specific generator depend on the capacity

factor. The adjustments to emissions as a function of capacity fac-
8
tor (power output) are done mainly using the results of the Wes-
6 tern Wind and Solar Integration Study (WWSIS) [46]. Overall, the
emission module includes part-load, ramping, and start-up emis-
4
sions. The efficiency and thus emissions degradation of generating
2 units with age is also taken into account and it is assumed that
retired units are replaced with units with advanced technology.
0 NOx emissions from generating units inside the balancing area,
0 10 20 30 40 50 60
associated with 4 cases with economic dispatch strategy for a rep-
BEV Penetration resentative summer day, are shown in Fig. 15.
From Fig. 15, it can be concluded that addition of BEVs with
Fig. 12. Change in summer average costs, compared to Base Case (Case 1).
uncontrolled charging results in an increase in criteria pollutant
emissions throughout all hours of the day except the first 6 h of
of the day, in Case 2, the use of load-following units is increased the days during which no change in emissions is observed. This
compared to Case 1 (Fig. 5) and the change in load is small, as a was expected since there is virtually no BEV charging at these
result the most expensive unit (combined cycle) is operating at hours. The rate of increase in emissions is higher with uncontrolled
higher capacity factor in Case 2. Overall, the uncontrolled charging charging and this is mainly because of the start-up and ramping
results is very small increase in average daily MCP, while the off- emissions associated with units that come online to serve the extra
peak charging result in decrease of up to 35 $/MW h in average load (Fig. 6). During peak time, although the output of peaking
daily MCP of the market (associated with Case 12). units is greater with uncontrolled charging, at the same time, the
 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156 1151

250

200

$/MWh
150

100

50

0
1 3 5 7 9 11 13 15 17 19 21 23
Hour Ending
Case 1 Case2 Case 3 Case 12

Fig. 13. Market clearing price.

400
350
300
250
$/MWh

200
150
100
50
0
1 3 5 7 9 11 13 15 17 19 21 23
Hour Ending
Case 7 Case 8 Case 9

Fig. 14. Market clearing price with time-dependent bidding.

is the fact that the average emissions intensity (lb MW h1) of

Case 1
5 the balancing area is highest at these hours because (1) only the
Case 2 least expensive units are online, and (2) units are operating at
Case 3 low capacity factors.
4 Case 12 From 7 to 10 am, in both Cases 3 and 12, the emissions increase
due to addition of BEVs mainly because there is significant charg-
ing occurring at these hours. The interesting point, however, is the
Metric Ton

3 reduction of criteria pollutant emissions during peak hours with

the addition of BEVs with off-peak charging. This is due to a
smoother load profile, resulting in fewer start-ups (Figs. 6 and 8),
2 increase in the capacity factor of online units, and ultimately low-
ering emissions.
The daily emissions associated with an average summer day in
1
2050 for all 13 cases are presented in Fig. 16 including the emis-
sions from the generating units inside the basin and increase in
the emissions from imports compared to cases with no BEV
0
5 10 15 20 deployment. To estimate the emissions from imports, one needs
Hour Ending to determine the mix of imports and their associated emissions.
In the state of California, imports are from coal or natural gas gen-
Fig. 15. NOx emissions from in-basin generating units. erators. Imports to the SoCAB are mostly from natural gas genera-
tors, biomass, and coal from out of state. Thus the average emission
factor of imports is higher or equal to that of the balancing area
load-following units are operating at much higher capacity factor under study. In order to show the impacts of imports, two scenar-
(and lower emissions factors) compared to the base case and thus ios are shown in Fig. 16, one assuming that the imports have an
the increase in in-basin emissions is lower at these hours. average emission factor equal to that of the balancing area
The impact of BEVs with off-peak charging is completely differ- (roughly equal to 0.2 lb MW h1 for NOx), the other scenario uses
ent from that of uncontrolled charging. In Case 3 at early hours of the average emission factor of the average U.S. grid (0.94 MW h1
the day the emissions do not change from the base case because as for NOx [47]).
was previously noted, the extra demand is served by imports. In The impacts of BEV deployment with economic dispatch strat-
Case 12, in which all the BEV demand is served by generating units egy, discussed at the beginning of this section and Fig. 16, suggest
inside the balancing area, the emissions increase significantly dur- that addition of BEVs results in an increase in overall grid emis-
ing the early hours of the day when the BEV charging is concen- sions (including the imports). The increase in emissions in off-
trated. The other reason for significant increase in the emissions peak charging, however, is not proportional to the demand
1152 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156

80 Economic Dispatch

60

NOx Dispatch CO2 Dispatch

40
Metric Ton/day

20

-20

-40
In-Basin Increase in imports (clean imports) Increase in imports (US average)

Fig. 16. Daily NOx emissions from in-basin units and imports.

increase suggesting that the overall emissions intensity in these grid emissions increase with addition of BEV even with off-peak
cases is lower compared to cases with no BEVs. Cases with high charging.
renewable penetration (Case 10, 11, and 13) have significantly To understand the environmental impacts of BEVs, the emis-
lower in-basin emissions as expected. In-basin emissions associ- sions from the transportation sector are included in the analysis.
ated with Case 13 is higher than Case 11 due to increased use of The results, shown in Fig. 17, include the emissions from the grid
peaking units as a result of higher use of intermittent wind energy (generating units inside the balancing area including the change
(see Figs. 5 and 7). However, this case results in reduction in in imports emissions) and the transportation sector. Note that Case
imports and has an overall lower emissions compared to Case 11. 12 is shown separately because it does not result in any change in
The results associated with the environmental dispatch strate- imports.
gies are quite different. Deployment of BEVs with uncontrolled In all cases except one (Case 3 with U.S. average grid import),
charging results in an increase in emissions, where the off-peak the addition of BEVs to the light-duty fleet results in an overall
charging has the potential to decrease the emissions if the imports emission reduction from the power generation and transportation
have the same emission factor as the balancing area. In these cases, sectors. The impact and importance of imports to the balancing
the reduction in peak hours is significant enough (not only start up area are also revealed in this figure. In scenarios using the U.S.
emissions are reduced, cleaner generating units are fired first) to average grid for the imports, the environmental benefits of the
offset the increase in the in-basin emissions during other hours. BEVs are reduced and in some cases marginal and in Case 3 the
In these cases, the reduction in the average grid emissions intensity overall emissions have increased. Thus, in order to fully benefit
with introduction of BEVs, is greater compared to economic dis- from the deployment of BEVs, it is not sufficient to ensure that they
patch, resulting in an overall reduction in grid emissions. However, are deployed in a relatively clean grid, the imports and surround-
if the imports mix is the same as the average U.S. grid, the overall ing interconnections have also a significant impact.

Clean Imports U.S. Grid Average Imports

100

80
Metric ton/day

60

40

20

Grid Transportation
Fig. 17. Overall daily NOx emission.
 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156 1153

Between Case 2 and Case 3, the case with off-peak charging In a grid with clean mix such as the balancing area studied in
results in increased overall emissions assuming that the imports this paper, the overall emissions from the electricity generation
have the same mix as the U.S. grid average. In this case, the deploy- and transportation sectors will be reduced with the deployment
ment of BEVs results in increased in-basin emissions and signifi- of BEVs. A resource dispatch strategy has a greater impact com-
cant increase in emissions from imports (Fig. 16) and thus the pared to BEV charging profile.
reduction in emissions from the transportation sector is not suffi- In all the cases studied except one (with imports mix of that of
cient to offset this increase. This is not observed in other dispatch average U.S), environmental benefits of off-peak charging are
strategies because of the reduction in emissions from in-basin gen- greater compared to uncontrolled charging. The main advantage
erating units, as previously discussed, along with reduction in of this charging strategy over the uncontrolled charging, is that it
transportation sector, help offset the increase in emissions from reduces the emission intensity of the grid i.e. kilogram of emission
imports. In Case 12 on the other hand, where all the extra demand per Megawatt-hour of electricity generation. In uncontrolled charg-
is supported by in-basin generators, the overall emissions are ing cases, the grid emission intensity increases due to increased use
reduced compared to Case 1. of peaking units and consequently more start-up emissions.
From the results presented in this section, it can be concluded
that BEVs have the potential to reduce the overall emissions; how-
With environmental dispatch strategy, it is possible to reduce
ever, the dispatch strategy of resources, vehicle charging profile, grid emissions with BEV deployment
renewable penetration, and imports mix have significant impacts
on these benefits. With implementation of environmental dispatch strategies, the
emissions intensity of the grid is decreased significantly that will
offset the increase in emissions from the BEV load and the emis-
4. Summary and conclusion sions from the electric grid decrease even with the addition of
new load.
In this paper, the economic and environmental impacts of
deploying BEVs in the light-duty fleet were assessed using a unit
It is not sufficient to analyze an isolated grid
commitment dispatch model capable of simulating various dis-
patch strategies and future scenarios. Various cases were modeled The role and importance of imports to a balancing area was
to study the impacts of various factors on the outcomes. Following studied in this paper showing that the environmental impacts of
are the conclusions of this paper: BEV deployment is also affected by interconnection to other bal-
ancing areas.

With controlled off-peak BEV charging, the need for installing
new capacity is eliminated Acknowledgments

One concern with the deployment of BEVs is that the electricity The authors thank the Advanced Power and Energy Program for
grid will not be able to handle the extra demand which is compa- financial support of this work.
rable to that of one household. This study suggests that the
installed generating units and reserves can handle a significant Appendix A
number of BEVs and maintain the reliability of the grid. With the
addition of BEVs, the transmission constraints where not violated A.1. Methodology
suggesting that the transmission system designed for the Base Case
can handle the BEV demand. The methodology developed to assess the economic and envi-
It was further shown that, with prior planning, sufficient gener- ronmental impacts of future scenarios include several modules
ation can be planned to satisfy the demand that includes the BEV which provide the necessary inputs to the unit commitment model
load. With off-peak charging and appropriate grid management, referred to as the dispatch model (Fig. A.1). Modules are developed
however, the need for adding more generation to the system is in MATLAB. The temporal resolution of the results depend on the
eliminated and the vehicles can be charged using already available inputs and data available and the methodology supports time res-
resources and thereby eliminate the costs associated with plan- olutions down to 5 min.
ning, licensing, building, and operating new power plants. The dispatch model is a unit commitment including an objective
function to minimize the social cost (Eq. (A.1)) and multiple con-

BEV impact on average cost of electricity generation is less than straints. The constraints are shown in Eq. (A.2)–(A.7). Eq. (A.2)
5%. establishes the relationship between the output of unit i, P(i), and
the commitment of that unit, I(i). Eqs. (A.3) and (A.4) are associated
In cases studied, the average cost of electricity production was with minimum and maximum capacity factor that unit i can oper-
increased with the addition of BEVs to the grid. However, this ate at. Eq. (A.5) is the necessary constraint to ensure that supply and
increase is on average less than 5 percent and it is greater with demand are equal. Eq. (A.6) is associated with the ramping limits of
environmental dispatch strategies compared to economic dispatch. generators, and Eq. (A.7) establishes the minimum on-time. The
transmission constraints are also added for transmission lines in

Off-peak charging result in MCP reduction locations that have historically been congested. The maximum
capacity of the line is considered in these constraints.
Deployment of BEVs with controlled off-peak charging results in
reduction in average MCP with all dispatch strategies. The impact of X
Ng

uncontrolled charging with economic dispatch is insignificant on Minimize ½C i ðPði; tÞÞIði; tÞ þ SðiÞIði; tÞfIði; tÞ  Iði; t  1Þg ðA:1Þ
i¼1
the average MCP, where with environmental dispatch strategies,
the uncontrolled charging results in increased average MCP.
Pði; tÞ ¼ 0 if and only if Iði; tÞ ¼ 0 ðA:2Þ

Off-peak charging helps reduce the emission intensity of the Pði; tÞ 6 Pgmax ðiÞ ðA:3Þ
grid
1154 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156

Fig. A.1. Methodology flow chart (from Razeghi et al. [29]).

Fig. A.2. State of California, SoCAB, and major utilities.

Iði; tÞPgmin ðiÞ 6 Pði; tÞ ðA:4Þ The set of equations introduced here make up a Mixed Integer
Non-Linear Program (MINLP) which is solved using AMPL language
sumfi in Ng gIði; tÞPði; tÞ ¼ DðtÞ ðA:5Þ using solvers such as Gurobi, and KNITRO [29].

absðPði; tÞ  Pði; t  1ÞÞ 6 RLðiÞ ðA:6Þ

A.2. South coast air basin
X
t
Iði; tÞ P Iði; t  1ÞT on ðiÞ ðA:7Þ The fictitious balancing area chosen in this study is the South
tT on ðiÞ
Coast Air Basin in California. This air basin includes Orange County,
 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156 1155

Fig. A.3. Location of generating units.

and part of Los Angeles, Ventura, San Bernardino, and Riverside [15] Letendre S, Watts R, Cross M. Plug-in hybrid vehicles and the vermont grid: a
scoping analysis. Burlington: UVM Transportation Center; 2008.
counties (Fig. A.2). This area is served by two Investor Owned Util-
[16] Kintner-Meyer M, Schneider K, Pratt R. Impact assessments of plug-in hybrid
ities (IOU), Southern California Edison (SCE), and San Diego Gas and vehicles on electric utilities and regional U.S. Power grids Part 1: technical
Electric (SDG&E), as well as several Publically Owned Utilities and analysis. Pacific Northwest National Laboratory; 2007.
municipalities, the largest one being Los Angeles Department of [17] Rangaraju S, De Vroey L, Messagie M, Mertens J, Van Mierlo J. Impacts of
electricity mix, charging profile, and driving behavior on the emissions
Water and Power (LADWP). SoCAB is chosen because it has been performance of battery electric vehicles: a Belgian case study. Appl Energy
a hub for alternative and new technologies, is subjected to strin- 2015;148:496–505.
gent California energy and environmental laws, and has one of [18] Clement-Nyns K, Haesen E, Driesen J. The impact of charging plug-in hybrid
electric vehicles on a residential distribution grid. IEEE Trans Power Syst
the worst air qualities in the nation and is facing challenges meet- 2010;25:371–80.
ing National Ambient Air Quality Standards. [19] Razeghi G, Zhang L, Brown T, Samuelsen S. Impacts of plug-in hybrid electric
Location of generating units in SoCAB is shown in Fig. A.3. The vehicles on a residential transformer using stochastic and empirical analysis. J
Power Sources 2014;252:277–85.
spatial resolution used in this study is 5 km by 5 km which is con- [20] Salah F, Ilg JP, Flath CM, Basse H, Dinther Cv. Impact of electric vehicles on
sistent with the air quality model used in the methodology. distribution substations: a Swiss case study. Appl Energy 2015;137:88–96.
[21] Clement-Nyns K, Haesen E, Driesen J. The impact of vehicle-to-grid on the
distribution grid. Electric Power Syst Res 2011;81:185–92.
References [22] Saber AY, Venayagamoorthy GK. Intelligent unit commitment with vehicle-to-
grid – a cost-emission optimization. J Power Sources 2010;195:898–911.
[1] U.S. Energy Information Administration. Use of Energy in the United States [23] Pecas Lopes JA, Rocha Almeida PM, Soares FJ. Using vehicle-to-grid to
Explained <http://www.eia.gov/Energyexplained/?page=us_ maximize the integration of intermittent renewable energy resources in
energy_transportation2015>. islanded electric grids. In: 2009 International Conference on Clean Electrical
[2] International Energy Agency. <http://www.iea.org/topics/transport/>. Power. p. 290–5.
[3] California Energy Commission. 2007 Integrated energy policy report. California [24] Lund H, Kempton W. Integration of renewable energy into the transport and
Energy Commission; 2007. electricity sectors through V2G. Energy Policy 2008;36:3578–87.
[4] California Air Resources Board. California greenhouse gas emission inventory – [25] Srivastava AK, Annabathina B, Kamalasadan S. The challenges and policy
2015 edition. <http://www.arb.ca.gov/cc/inventory/data/data.htm2015>. options for integrating plug-in hybrid electric vehicle into the electric grid.
[5] California Air Resources Board. 2012 Estimated annual average emissions. Electricity J 2010:23.
<http://www.arb.ca.gov/app/emsinv/2013/emseic1_query.php?F_DIV=-4&F_ [26] Hadley SW, Tsvetkova AA. Potential impacts of plug-in hybrid electric vehicles
YR=2012&F_SEASON=A&SP=2013&F_AREA=CA2013>. on regional power generation. Electricity J 2009;22.
[6] California Air Resources Board. Alternative and renewable fuel and vehicle [27] Villar J, Diaz C, Arnau J. Impact of plug-in-electric vehicles penetration on
technology program. <http://www.energy.ca.gov/altfuels/>. electricity demand, prices and thermal generation dispatch. In: 2012 9th
[7] Hadley SW. Impact pf plug-in hybrid vehicles on electric grid. In: Oak Ridge International Conference on the European Energy Market (EEM). p. 1–8.
National Laboratory, editor; 2006. [28] Weis A, Jaramillo P, Michalek J. Estimating the potential of controlled plug-in
[8] Fagnant DJ, Kockelman KM. The travel and environmental implications of hybrid electric vehicle charging to reduce operational and capacity expansion
shared autonomous vehicles, using agent-based model scenarios. Transport costs for electric power systems with high wind penetration. Appl Energy
Res Part C: Emerg Technol 2014;40:1–13. 2014;115:190–204.
[9] Zhang L, Brown T, Samuelsen GS. Fuel reduction and electricity consumption [29] Razeghi G, Brouwer J, Samuelsen S. A spatially and temporally resolved model
impact of different charging scenarios for plug-in hybrid electric vehicles. J of the electricity grid – economic vs environmental dispatch. Appl Energy
Power Sources 2011;196:6559–66. 2016;178:540–56.
[10] Razeghi G, Brown T, Samuelsen GS. The impact of plug-in vehicles on [30] California ISO. <http://www.caiso.com/green/renewableswatch.html>.
greenhouse gas and criteria pollutants emissions in an urban air shed using [31] California ISO. Open Access Same-Time Information System (OASIS).
a spatially and temporally resolved dispatch model. J Power Sources oasis.caiso.com.
2011;196:10387–94. [32] Tidball R, Bluestein L, Rodriguez N, Knoke S. Cost and performance
[11] Donateo T, Licci F, D’Elia A, Colangelo G, Laforgia D, Ciancarelli F. Evaluation of assumptions for modeling electricity generation technologies. National
emissions of CO2 and air pollutants from electric vehicles in Italian cities. Appl Renewable Energy Laboratory; 2010. NREL/SR-6A20-48595.
Energy 2015;157:675–87. [33] Black & Veatch. Cost and performance data for power generation technologies.
[12] Razeghi G, Carreras-Sospedra M, Brown T, Brouwer J, Dabdub D, Samuelsen S. Prepared for the National Renewable Energy Laboratory.
Episodic air quality impacts of plug-in electric vehicles. Atmos Environ [34] U.S. Energy Information Administration. <http://www.eia.gov/tools/faqs/faq.
2016;137:90–100. cfm?id=105&t=3>.
[13] Ferrero E, Alessandrini S, Balanzino A. Impact of the electric vehicles on the air [35] Kavalec C, Fugate N, Gorin T, Alcorn B, Ciminelli M, Gautam A, et al. California
pollution from a highway. Appl Energy 2016;169:450–9. energy demand final forecast 2012–2022 Volume 2: Electricity demand by
[14] Stephan CH, Sullivan J. Environmental and energy implications of plug-in utility planning area. California Energy Commission, Electricity Supply
hybrid electric vehicles. Environ Sci Technol 2008;42:1185–90. Analysis Division. CEC-200-2012-001-CMF-VII; 2012.
1156 G. Razeghi, S. Samuelsen / Applied Energy 183 (2016) 1142–1156

[36] Kavalec C, Fugate N, Gorin T, Alcorn B, Ciminelli M, Gautam A, et al. California [42] Santini DJ. How will light duty plug-in hybrid vehicle electric drive emerge and
energy demand final forecast 2012–2022 Volume 1: Statewide electricity evolve? An examination of recent and current developments. Argonne
demand and methods. California Energy Commission, Electricity Supply National Laboratory; 2007.
Analysis Division. CEC-200-2012-001-CMF-VI; 2012. [43] Olson T. Preliminary Electric Vehicle Demand Forecast Workshop on Revised
[37] California Air Resources Board. EMFAC2014. <https://www.arb.ca.gov/emfac/ Electricity and Natural Gas Demand Forecasts 2014–2024. <http://www.
2014/>. energy.ca.gov/2013_energypolicy/documents/2013-10-01_workshop/
[38] Yang J, He L, Fu S. An improved PSO-based charging strategy of electric vehicles presentations/06_Tim_Olson_Preliminary_Electric_Vehicle_Demand_Forecast.
in electrical distribution grid. Appl Energy 2014;128:82–92. pdf2013>.
[39] California Department of Transportation. California household travel survey. [44] Ramos Muñoz E, Razeghi G, Zhang L, Jabbari F. Electric vehicle charging
<http://wwwdotcagov/hq/tpp/offices/omsp/statewide_travel_analysis/ algorithms for coordination of the grid and distribution transformer levels.
chtshtml>. 2012. Energy 2016;113:930–42.
[40] U.S. Department of Transportation Federal Highway Administration. 2009 [45] California Energy Commission. Cost of generation. CEC-200-2010-002; 2010.
National Household Travel Survey. <http://nhts.ornl.gov/download. [46] Lew D. Western wind and solar integration study (WWSIS) phase 2:
shtml#2009>. introduction and input data. NREL; 2012.
[41] U.S. Department of Energy, U.S. Environmental Protection Agency. The official [47] Environmental Protection Agency. Emissions & generation resource integrated
U.S. government source for fuel economy information. <http:// database (eGRID). <https://www.epa.gov/energy/egrid>.
www.fueleconomy.gov/feg/findacar.shtml>.