Year 2009 UCD- ITS- WP- 09- 03
Anticipating PHEV energy impacts in California
May 16, 2009
John Axsen1*, Ken Kurani1
1 Institute of Transportation Studies, University of California at Davis, 2028 Academic
Surge, One Shields Avenue, Davis, CA, 95616, U. S. A
* Corresponding author: U. S. A. Tel.: + 1 530 231 5007; fax: + 1 530 752 6572; E- mail address:
jaxsen@ ucdavis. edu
Institute of Transportation Studies ◦ University of California, Davis
One Shields Avenue ◦ Davis, California 95616
PHONE: ( 530) 752- 6548 ◦ FAX: ( 530) 752- 6572
WEB: http:// www. its. ucdavis. edu
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 1
EVS24
Stavanger, Norway, May 13- 16, 2009
Anticipating PHEV energy impacts in California
Jonn Axsen 1*, Ken Kurani 1
1 Institute of Transportation Studies, University of California at Davis, 2028 Academic Surge, One Shields Avenue,
Davis, CA, 95616, U. S. A
* Corresponding author: U. S. A. Tel.: + 1 530 231 5007; fax: + 1 530 752 6572; E- mail address: jaxsen@ ucdavis. edu
Abstract
To explore the potential energy impacts of widespread PHEV use, an innovative, three- part survey
instrument collected data from 877 new vehicle buyers in California. This analysis combines all the
available information from each respondent— driving, recharge potential, and PHEV design priorities— to
estimate the energy impacts of the respondents’ existing travel and understandings of PHEVs under a
variety of recharging scenarios. Results suggest that the use of PHEV vehicles could halve gasoline use
relative to conventional vehicles— the majority of this reduction being due to increases in charge sustaining
( CS) fuel economy. Using three scenarios to represent potential boundary conditions on PHEV driver
recharge patterns ( unconstrained, universal workplace recharging, and off- peak only charging), we estimate
tradeoffs between the magnitude and timing of PHEV electricity use. In the unconstrained “ Plug and Play”
recharge scenario, recharging peaks at 6: 15pm, following a far more dispersed pattern throughout the
earlier part of the day than anticipated by previous research. PHEV electricity use could be increased
through policies increasing non- home recharge opportunities ( e. g., the “ Enhanced Workplace Access”
scenario), but most of this increase occurs during daytime hours and could contribute to peak electricity
demand ( depending on a given region’s definition of “ peak”). We also demonstrate how deferring all
recharging to off- peak hours ( 8pm to 6am) could eliminate all additions to daytime electricity demand from
PHEVs. However, in such a scenario less electricity is used due to the elimination of daytime recharge
opportunities and less gasoline is displaced. Overall, policy, technology, and energy providers may use this
information to understand whether their plans, designs, and goals align with these present empirically-informed
understandings.
Keywords: PHEV, energy consumption, market, charging, environment
1 Introduction
As hybrid- electric vehicles ( HEVs) continue to
achieve significant commercial success in the
U. S. market, plug- in hybrid vehicles ( PHEVs)
are touted as the next step in electric drive
development [ 1]. PHEVs are one step closer to the
pure electric vehicle ( EV) initially envisioned by
California’s zero emissions vehicle mandate; users
can charge the battery from the electrical grid and
drive limited distances in charge- depleting ( CD)
mode. During this mode, the vehicle is powered
either by electricity only ( all- electric operation) or
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 2
by electricity and gasoline ( blended operation).
Once the battery is depleted to a minimum state
of charge, the PHEV uses only gasoline in charge
sustaining ( CS) mode, achieving gasoline- only
fuel efficiency typical of today’s HEV. Battery
size, degree of hybridization, and drivetrain
design all influence the overall operation of a
given PHEV [ 2].
PHEVs’ use of gasoline and electricity depends
on the interaction between vehicle design and
recharging and travel behaviours, creating
inherent uncertainty for policymakers,
automakers, electric utilities, researchers, and
other interest groups. Due to lack of direct data
on these interactions and behaviors, previous
impact and market analyses have assumed them
[ 1,3- 9], often drawing by proxy from data on
aggregate travel and housing stocks. The choice
of assumptions seriously affects results; Lemoine
et al. [ 1] illustrate how varying time of day
recharge assumptions can substantially influence
predictions of electricity grid impacts. In all,
there is a demonstrated lack of data on consumer
behavior and demand pertinent to PHEV markets
and subsequent environmental and energy
impacts.
We focus this paper on one question: what
energy impacts ( gasoline and electricity) can we
anticipate with significant PHEV adoption? To
empirically answer this question, we construct
recharge scenarios based on data collected by a
web- based survey of new vehicle buying
households. Results from this survey were
recently reported at the U. S. level [ 10], but this
paper focuses on respondents from California— a
sub- region that was purposely oversampled for
the purpose of conducting a representative
analysis. Three types of data were collected: ( 1)
time of day driving patterns, ( 2) recharge
potential, and ( 3) new- car buyers’ PHEV design
priorities, collected via design games. Taken
together, the collected information regarding
driving patterns, recharge potential and design
priorities allow the creation of realistic recharge
scenarios. We simulate grid impacts under three
scenarios to investigate potential tradeoffs
between overall gasoline and electricity use and
the timing of electricity use.
2 Methods
2.1 Survey design
Data was collected using a multi- part online
survey. Driving patterns and recharge potential
were elicited using a Plug- in Potential diary of
driving and parking for a vehicle purchased new
( model year 2002 or later) that is driven several
times per week by the respondent’s household.
Respondents were assigned a day of the week and
instructed to record information for a 24 hour
period starting with their first trip of that day.
Information included the timing and distance of
each trip, parking locations, and the proximity of
those locations to an electrical outlet. Respondents
recorded data in a diary printed from a PDF
document and then input their data online. The
respondent’s diary day was immediately depicted
to them as a graph, using a technique similar to
that used by Kurani et al. [ 13,14] to help
respondents better understand their own driving
behaviour and how an electric- drive vehicle could
fit into their lifestyle.
The PHEV design priority data used in this
analysis were collected with a priority- evaluator
game. Commonly, researchers will infer
preferences for attributes of alternative fuelled
vehicles by presenting respondents with a
description of one or several new technologies,
followed with a set of hypothetical choice
scenarios in which respondents make several
choices from sets of vehicles of different attributes,
e. g. [ 15- 18]. However, Heffner et al. [ 19]
demonstrate that more in- depth research, such as
household interviews, can reveal important
information that choice experiments cannot. To
improve the quality of data gathered through a
nationwide survey, prior to the PHEV design
exercises, respondents were provided two types of
preparatory information: ( 1) the 24- hour diary
exercise described above served the additional role
of reflecting to respondents aspects of their travel
patterns and potential access to recharge spots, and
( 2) a PHEV buyers’ guide describing basic design
options for PHEVs. Respondents then completed a
Purchase Design game with design possibilities
priced in dollars; respondents could reject buying a
PHEV, retaining a conventional vehicle. This type
of in- depth research has previously been described
as reflexive design [ 14].
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 3
Table1: Price of upgrades for Purchase Deign game ( prices incremental to conventional vehicle)
“ Higher” price “ Lower” price
Attributes Attribute level Car Truck Car Truck
Base premium
over conventional
$ 3,000 $ 4,000 $ 2,000 $ 3,000
Added premiums:
Recharge time 8 hours
4 hours
2 hours
1 hour
0
+$ 500
+$ 1,000
+$ 1,500
0
+$ 1,000
+$ 2,000
+$ 3,000
0
+$ 250
+$ 500
+$ 750
0
+$ 500
+$ 1,000
+$ 1,500
CD mpg and type
Blended
75 mpg
100 mpg
125 mpg
All- electric
0
+$ 1,000
+$ 2,000
+$ 4,000
0
+$ 2,000
+$ 4,000
+$ 8,000
0
+$ 500
+$ 1,000
+$ 2,000
0
+$ 1,000
+$ 2,000
+$ 4,000
CD range
10 miles
20 miles
40 miles
0
+$ 2,000
+$ 4,000
0
+$ 4,000
+$ 8,000
0
+$ 1,000
+$ 2,000
0
+$ 2,000
+$ 4,000
CS mpg Conventional mpg + 10
Conventional mpg + 20
Conventional mpg + 30
0
+$ 500
+$ 1,000
0
+$ 1,000
+$ 2,000
0
+$ 250
+$ 500
0
+$ 500
+$ 1,000
Potential PHEV designs offered to respondents
were informed by previous analysis of early
PHEV drivers [ 20]. There were four PHEV
design attributes: ( 1) hours required for complete
recharge of a depleted battery, ( 2) gasoline use in
charge- depleting ( CD) mode, ( 3) miles of range
in CD mode, and ( 4) gasoline use in charge-sustaining
( CS) mode. 1 In each game, a base
PHEV design was offered with capabilities easily
achievable by current technology [ 2]: a PHEV
that requires up to 8 hours to completely
recharge, that can be driven for the first 10 miles
in CD mode using blended operation that
increases gasoline- only fuel economy to 75 mpg,
and that can improve fuel economy by 10 mpg
when operating in CS mode over a conventional,
i. e., gasoline- ICE, version of the same vehicle. 2
1 To ease comparison with other literature, we
report distances in miles, where 1 mile = 1.61 km.
2 These PHEV design games are meant to represent
designs that are technologically feasible, but not
necessarily with exact specifications. For instance,
the battery required for our base PHEV design
would likely require only 2 to 3 hours to fully
recharge with a 110 to 120 volt circuit. However,
with careful pre- testing, we consciously chose to
simplify attribute levels and ignore potential
interactions among attributes to create exercises
that are more likely to be understood by our
respondents than to adhere to experts’ knowledge.
Respondents were given opportunities to improve
each attribute under the different price conditions
depicted in Table 1. The PHEV design exercise
was framed in the context of the household’s next
new vehicle purchase. The questionnaire first
elicited information about the anticipated price,
make and model of the next new vehicle the
respondent’s household would buy. The
respondent then completed two PHEV purchase
exercises, each comparing their anticipated
conventional vehicle with a PHEV version of the
same. Respondents were presented with a “ higher”
price and “ lower” price PHEV purchase
conditions, where prices in both conditions also
depended on whether the vehicle was a car or
truck. Each exercise started with the same base
PHEV model, with additional upgrades available
for added price. In each exercise, the respondent
could choose to purchase their anticipated
conventional vehicle, the offered ( base) PHEV
version, or an upgraded PHEV version. The costs
in Table 1 are largely hypothetical, although they
are comparable to previous estimates [ 21- 23].
2.2 Data collection
Our target population is new vehicle buying
households in California. To qualify, respondents
had to own a new gasoline vehicle that they
purchased in 2002 or later, which they personally
drove at least 3 times per week. The respondent
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 4
also must have played a significant role in the
household’s decision to purchase this vehicle. In
limiting our study to this population, we imply
that the early market for PHEVs is limited to
households that tend to buy new vehicles in
general. In total, 877 California respondents
completed the entire survey in December of 2007
( Fig. 1).
Data were collected with a web- based survey.
Relative to mail and telephone methods, this
mode improves the degree of design flexibility,
response interaction [ 24], response accuracy for
travel diaries [ 25], and data administration time
and cost [ 26]. In recent years, web- based surveys
were susceptible to non- coverage error, where a
significant portion of the target population, in
this case new car buyers, could be excluded if
they don’t have internet access. This concern is
declining in the U. S. as internet usage rates have
grown from 44% in 2000 to over 70% in 2007
[ 27]. Also, we suspect there is a positive
correlation between internet access and likeliness
to buy new vehicles, implying an even higher
usage rate among our target population.
However, non- response bias is still an important
concern because those without internet access
tend to be disproportionally older, with lower
incomes and less education [ 24, 28].
Figure1: Geographic distribution of California sample
Respondents for this survey were recruited by
Harris Interactive from their internet panel. To
counteract concerns of non- coverage and non-response
error, Harris estimates weights to better
match the realized sample to the target
population. Weights are based on geographic,
demographic and attitudinal data, and matched to
existing databases collected through multiple
survey modes ( including mail and telephone). All
results presented in this study use these weights to
match our sample to the California population of
new vehicle buyers.
To assess the external validity of our sample, we
compare our sample distributions with a sub-sample
of 389 California households owning new
vehicles drawn from the 2001 National Household
Travel Survey ( NHTS). We find that the income
levels of both samples are about 40 percent higher
than general population estimates from similar
years. Also, gender and age follow similar
distributions between the two samples of new
vehicle buying households. Our sample does have
fewer households without any college level
education ( 8.8 percent) relative to the NHTS
sample ( 22.1 percent), and fewer households living
in detached homes ( 68.1 percent) than the NHTS
sample ( 79.4 percent). Overall, we feel these
differences are not likely to be problematic. We
conclude that our sample matches well with one
other sample of new car owners on these socio-demographic
measures, strengthening claims that
our results can be extended to the California
population of new- car owning ( and therefore, new
car- buying) households.
3 Results
3.1 Recharge access
Results from the Plug- in Potential vehicle diary
indicate that more new vehicle buyers may be pre-adapted
for vehicle recharging than estimated in
previous constraints analyses. Following Graham
et al. [ 15], we consider a parking spot to be viable
for recharging if located within 25 feet of an
electrical outlet. Of the 877 respondents, 52.1
percent found at least one viable recharge location
during their 24- hour diary day, and 45.3 percent
identified one at their home. Only 4.4 percent of
respondents found outlets at work, and 9.1 percent
found outlets at other non- home locations ( e. g.
friend’s home, school, commercial site, etc.).
In Fig. 2, we represent driving and recharge
potential over a 24- hour cycle ( in 15 minute
intervals); the sample was proportionally assigned
a weekday or weekend- day to complete their diary.
On weekdays, the proportion of respondents’
driving follows an expected daily pattern ( the
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 5
black line in Fig. 2), peaking during common
commute hours at 7: 30am and 5: 00pm. In any
given 15- minute interval, total recharge potential
ranges from over 45 percent of respondents from
9: 00pm to 6: 00am, to under 20 percent from
10: 00am to 3: 00pm. Throughout the day, home is
by far the most frequent location of recharge
opportunities within respondents’ existing travel
and recharge potential. Neither work nor other
non- home locations have recharge potential that
surpass 4 percent of respondents for any 15
minute interval during the day. The general
pattern in Fig. 2 is consistent with driving
patterns; recharge potential drops when many
respondents are driving or parked at work or
other locations, and rises when vehicles are
parked at home. Driving patterns on weekend
days ( not shown) do not show morning and
afternoon peaks, but rather a single broad mid-day
rise to a peak at around 4: 00pm ( with a lesser
peak in the later evening). Weekend recharge
potential during any given 15 minute interval
ranges from a high of 55.1 percent to a low of
19.5 percent of all respondents; home also
dominates the potential recharge locations for
weekends.
0%
10%
20%
30%
40%
50%
60%
12am 4am 8am 12pm 4pm 8pm 12am
0%
3%
6%
9%
12%
15%
18%
Home Work
Other Drive
Figure2: Time of day driving and recharge potential
( weekdays only, n = 644)
3.2 PHEV design and value
Because recharge opportunities are relatively
sparse at work and other non- home locations, we
focus on home recharging as the key criteria to
characterize a potential early PHEV market in
this analysis. This constraint is substantiated by
the experience of early drivers of PHEV-conversions
[ 20]. Thus, for the remainder of this
study, we limit our consideration to the 45.3
percent of our sample that identified an electrical
outlet within 25 feet of their vehicle parking spot
at their home location at some time during their
24- hour diary. Among respondents with home
recharge potential, 73.3 percent designed a PHEV
for their next new vehicle in the “ higher” price
condition, and 84.0 percent did so in the “ lower”
price condition. We further constrain this segment
based on PHEV interest as indicated by purchase
intentions in the design games. Thus, we select the
respondents that demonstrate both access to
sufficient recharge infrastructure and PHEV
interest as a group best representing the early
PHEV market in California— 33.2 percent using
the “ higher” price scenarios, and 38.1 percent
using the “ lower” price scenarios. We will refer to
these subsets as the plausible early market
respondents.
Focusing on the interests of these plausible early
market respondents, results of the PHEV design
games are summarized in Fig. 3. PHEV
performance priorities varied substantially; no
majority PHEV design emerged. A substantial
portion of plausible early market respondents
chose the base PHEV models with no upgrades—
31.5 percent in the higher price condition and 23.2
percent in the lower price condition. Among those
that chose to pay extra for upgrades, CS fuel
economy upgrades were chosen more often than
other upgrades, and there is no evidence of the
strong interest in all- electric CD operation
observed among some PHEV pioneers [ 20]. All-electric
upgrades were chosen by only 2.7 and 3.9
percent of respondents in the higher and lower cost
conditions, respectively. CD operation and range
improvements were chosen relatively less often
than CS upgrades.
0%
20%
40%
60%
80%
100%
Recharge
CD Type
CD Range
CS MPG
Recharge
CD Type
CD Range
CS MPG
% Choosing Upgrade
“ Higher” Price Scenario
( n = 295)
“ Lower” Price Scenario
( n = 339)
8h 75
mpg
10m + 10
mpg
8h 75
mpg
10m + 10
mpg
4h
2h
4h
2h
100
125
125
100
20m
20m
+ 20
+ 30
+ 20
+ 30 40m
1h AE 40m 1h AE
Figure3: Distribution of selected PHEV upgrades
( all plausible early market respondents)
3.3 PHEV energy use scenarios
To create scenarios of gasoline and electricity use
among early PHEV buyers, we integrate the % of Respondents with Access
% of Respondents Driving
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 6
information presented thus far from respondents
in the plausible early market segment: driving
behaviour and recharge potential as recorded by
their 24- hour diary, and PHEV design choices as
demonstrated in the Purchase Design game. In
other words, we create scenarios of gasoline use
and recharge patterns for each plausible early
market respondent as if they had driven their
PHEV design on their 24- hour vehicle diary day.
These scenarios rely on the following
assumptions:
• Gasoline use is modelled using the estimated
miles per gallon ( MPG) of the vehicle,
without accounting for potential variation in
driving patterns. In other words, if the
vehicle is rated at 20 MPG, we assume this
constant rate for each mile driven ( neglecting
potential for different drive patterns over a
given trip or across drivers).
• For charge depleting ( CD) operation,
electricity use ( kWh/ mile) and available
battery energy capacity ( kWh) is estimated
as in Table 2, based on previous estimates
[ 23, 28, 29]
Table2: Assumed PHEV energy requirements
CD mpg/ type Car Truck
75 MPG
CD kWh/ mile
kWh for 10 miles
kWh for 20 miles
kWh for 40 miles
0.12
1.2
2.3
4.6
0.15
1.5
3.0
5.9
100 MPG
CD kWh/ mile
kWh for 10 miles
kWh for 20 miles
kWh for 40 miles
0.14
1.4
2.7
8.0
0.17
1.7
3.5
7.0
125 MPG
CD kWh/ mile
kWh for 10 miles
kWh for 20 miles
kWh for 40 miles
0.18
1.8
3.6
7.3
0.23
2.3
4.7
9.3
All electric
CD kWh/ mile
kWh for 10 miles
kWh for 20 miles
kWh for 40 miles
0.30
3.0
6.0
12.0
0.38
3.8
7.7
15.4
• Each vehicle’s assumed battery state of
charge at the beginning of the day is a
function of the distance driven the previous
day ( assumed to be the same as the diary day
due to lack of multi- day data) and the
respondent’s estimated hours of recharge
potential from the previous day ( elicited
elsewhere on the survey).
• Following Lemoine et al.’ s [ 1] assumptions,
the minimum recharge rate for a PHEV battery
using a regular 110- 120 V outlet is 1 kWh per
hour. If the respondent’s chosen PHEV design
has a recharge rate faster than that required for
their battery size, we apply the shorter of the
two recharge times. For example, if the
respondent chose a PHEV requiring 8 hours
for complete recharge, yet their battery size is
only 1.2 kWh ( requiring a maximum of 1.2
hours for full recharge), we apply the 1.2 hour
time. In contrast, if the same respondent
selected a recharge time of one hour, we apply
the one hour time.
• Following Lemoine et al.’ s [ 1] assumptions,
vehicle recharging is approximately 83 percent
efficient— increasing the battery’s state of
charge by 1 kWh requires 1.2 kWh from the
electrical outlet.
• Each scenario is scaled up to represent 1
million vehicles. This value is not selected in
anticipation of a particular sales volume for a
particular year, but instead is a relatively
feasible market size that serves to normalize
energy use to allow comparisons across
scenarios ( with different sample sizes). 3
• Vehicles are recharged on a daily basis as
detailed in the scenario descriptions below.
• The PHEVs are used precisely as were their
non- PHEV variants; the scenarios are based on
replicating the travel- days as recorded in the
diaries and do not allow for households to
change the assignment of vehicles within the
household or otherwise change vehicle use in
response to the PHEV.
• We assume for this analysis that one- day
cross- sectional data are adequate to
characterize travel and therefore energy
impacts. One- day diaries systematically under-represent
longer trips unless the sampling is
conducted according to the frequency
distribution of travel- day or trip distances
across people and days. By sampling across all
seven days of the week we attempt to reduce
3 An alternative approach would be to estimate the
effect of each recharge scenario on the size of the
potential PHEV market, such as the addition of
potential PHEV buyers resulting from the expansion
of public vehicle recharge infrastructure, e. g. at the
workplace. However, we leave such analyses to
future research, and instead focus on “ what- if”
scenarios using a set market size.
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 7
the effect on our analysis, but do not
represent that it is immune. It seems
plausible that we, and anyone using one- day
travel data, will underestimate total energy
use and gasoline use in particular. We leave
the estimation of the size of this potential
problem to future research.
Following these assumptions, we created four
scenarios using data from the plausible early
market respondents:
• No PHEVs ( Fig. 4a): In this scenario, we
estimate and aggregate the gasoline used by
the respondents on their actual diary days.
• Plug and play ( Fig. 4b): We simulate the
gasoline used for driving and the electricity
used for recharging, allowing that the
conventional vehicles are displaced by a
vehicle with the PHEV upgrades chosen in
the Purchase Design game. Drivers are
assumed to recharge whenever they are
parked within 25 feet of an electrical outlet.
In other words, there are no pricing
mechanisms, e. g., time of use electricity
tariffs, or technologies, e. g., smart charging
mechanisms, to divert recharging to off-peak.
• Enhanced workplace access ( Fig. 4c): This
scenario starts with the conditions in Plug
and Play, but further supposes that all
workers can and do recharge at work.
• Off- peak only ( Fig. 4d): Finally, using the
same recharge potential and PHEV designs
as Plug and Play, in this scenario no PHEV
recharging is allowed during daytime peak
hours ( 6am to 8pm). Smart charging
technology is used to optimize the timing of
electricity use over this period, represented
as a flat line ( the actual shape of this line
would likely vary according to the needs of a
particular electric utility).
Taken together, these scenarios are meant to
represent potential boundary conditions, that is,
where the entire market adheres to a selected
condition, i. e., no recharge regulation, enhanced
workplace access, or off- peak charging. Of
course, the early PHEV market may include
elements of more than one of these scenarios, as
well as other potential conditions we do not
consider here, all of which are likely to change
over time. However, the purpose of this exercise
is to present these boundary conditions to frame
discussions of the potential benefits and
drawbacks of different recharge strategies and
policies.
12am 4am 8am 12pm 4pm 8pm 12am
0
1000
2000
3000
4000
5000
( a) No PHEVs
0
100
200
300
400
500
12am 4am 8am 12pm 4pm 8pm 12am
0
1000
2000
3000
4000
5000
Home Work
Other Gas
( b) Plug and play
0
100
200
300
400
500
12am 4am 8am 12pm 4pm 8pm 12am
0
1000
2000
3000
4000
5000
Home Work
Other Gas
( c) Enhanced workplace access
0
100
200
300
400
500
12am 4am 8am 12pm 4pm 8pm 12am
0
1000
2000
3000
4000
5000
Home Work
Other Gas
( d) Off- peak only
Figure 4: Recharge profiles using “ higher” price
scenario ( weekdays only, n = 231)
Gal./ min per Gal./ min per Million Vehicles Gal./ min per Million Vehicles Gal./ min per Million Vehicles Million Vehicles
MW per Million PHEVs MW per Million PHEVs MW per Million PHEVs
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 8
Table3: Summary of recharge scenarios, scaled to one million PHEVs
PHEV Design Game:
“ Higher” price
PHEV Design Game:
“ Lower” price
Scenario Weekday
( n = 231)
Weekend
( n = 52)
Weekday
( n = 265)
Weekend
( n = 58)
No PHEVs Gasoline ( Gal.) 1,658,895 1,353,784 1,627,466 1,347,016
CS Upgrade Only Gasoline ( Gal.) 1,024,708 820,077 952,423 804,611
% Gas reduced 38.2% 39.4% 41.5% 40.3%
Plug and Play Gasoline ( Gal.) 870,444 690,669 778,571 678,475
% Gas reduced 47.5% 49.0% 52.2% 49.6%
Electricity ( MWh) 3,007 2,516 3,679 2,663
Peak ( MW) 416 300 513 327
Peak Time 6: 15pm 5: 15pm 6: 30pm 6: 30pm
Gasoline ( Gal.) 826,251 686,557 Enhanced Workplace 737,325 672,658
Access % Gas reduced 50.2% 49.3% 54.7% 50.1%
Electricity ( MWh) 3,843 2,655 4,481 2,843
Peak ( MW) 410 300 486 365
Peak Time 5: 45pm 5: 15pm 6: 15pm 6: 15pm
Off Peak Only Gasoline ( Gal.) 909,208 717,625 815,810 700,178
% Gas reduced 45.2% 47.0% 49.9% 48.0%
Electricity ( MWh) 2,199 1,917 2,873 2,217
Peak ( MW) 220 192 287 222
Peak Time 8pm- 6am 8pm- 6am 8pm- 6am 8pm- 6am
Figures 4a- d portray each scenario for
respondents who completed weekday diaries
given the PHEV designs they selected in the
“ higher” price conditions. Table 3 includes these
as well as results from respondents with weekend
day diaries, as well as “ lower” price conditions
PHEV designs. Figures 4a- d depict the time of
day gasoline use ( gallons per minute) and
electricity use ( MW) per million vehicles over a
24- hour period. The areas under the curves
represents the total gallons of gasoline, or MWh
of electricity, used over the day. In the Plug and
Play scenario, most recharging occurs at home,
peaking at 6: 15pm at 416 MW ( 513 MW in the
“ lower” vehicle price condition)— significantly
lower than the 1,200 MW peak anticipated by
Lemoine et al. [ 1] for 1 million PHEVs. Their
higher peak electricity demand estimate is due to
their assumptions about a uniform PHEV design
across the market ( 20 miles of all- electric CD
range) and relatively uniform recharging patterns
of PHEV drivers. 4 In contrast, the present study
allows for substantial variation in PHEV designs
and daily driving.
4 In each recharge scenario presented by Lemoine et
al. [ 1], PHEV drivers are assumed to begin
recharging at approximately the same time of day
for the same duration.
Time of day gasoline use corresponds with the
rush hour periods observed in Fig. 2. These
simulations indicate that in the Plug and Play
scenario overall gasoline use is estimated to cut
gasoline use by half relative to the No PHEV
scenario ( Table 3). Notice that gasoline use is
reduced by a larger degree in the morning due to
the higher proportion of miles driven in CD mode
earlier in the day. Table 3 also shows that a large
portion of this gasoline reduction ( 75 to 85
percent) is due to upgrades to CS fuel economy
with CD capabilities eliminated. 5 For this reason,
overall gasoline savings varies little across the
three charging scenarios or the vehicle price levels
in the design game; in all instances, gasoline use is
cut in about half compared to the No PHEV
scenario.
However, the peak magnitude and timing of
recharging varies significantly across the
scenarios. Fig. 5 plots all three recharge scenarios.
The Enhanced Workplace Access scenario
increases overall electricity use by 28 percent
relative to Plug and Play, with much of the
5 However, simulating only CS fuel economy
upgrades may be inappropriate— respondents might
not have chosen the vehicle upgrades without plug- in
and CD capabilities.
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 9
addition occurring in the morning as drivers
arrive at work. In contrast, the Off Peak Only
scenario reduces electricity use by 27 percent,
largely due to the elimination of work and other
non- home recharge opportunities that occur
during peak hours. Of course, this scenario has
the benefit of eliminating all electricity use
during peak hours, with nightly demand balanced
at 220 MW. As noted, the specific balancing
strategy used in this scenario would likely vary
by electric utilities to flatten out overall off- peak
demand, as seen in Lemoine et al.’ s [ 1] “ optimal
charging” scenario. Our scenario merely
demonstrates the potential for shifting and
minimizing peak demand.
0
100
200
300
400
500
12am 4am 8am 12pm 4pm 8pm 12am
Plug and Play Worker
Off Peak
Figure4: Comparing PHEV recharge scenarios
(“ higher” price scenario, weekdays only, n= 231)
4 Discussion and Conclusions
Results from this analysis offer initial answers to
our research question: anticipating the energy
impacts of the early PHEV market. Our
simulated world contains far more variety of
PHEV designs than any prior study. This is an
intentional difference, allowing respondents to
design the PHEV they would most desire given
their current understanding and valuation of four
PHEV performance parameters. We believe this
is a more realistic representation of a plausible
near future than the imposition of one or a few
PHEV designs on the entire population of vehicle
drivers. Certainly as we analyze “ one- million
PHEV” scenarios— suggesting that we are
attempting to analyze a world existing a few
years after the introduction of PHEVs— a world
of greater variety is more plausible than a world
of one or a few PHEV designs. 6 Our scenario
6 Our simulated world may be too plastic, too
molded to the individual vehicle selections of our
respondents. We caution against strict adherence to
analyses remain susceptible to other threats
endemic to such efforts. Radically changing travel
behaviour— in response to fuel prices, competition
from other alternatives, or in response to PHEVs
themselves— could invalidate our use of data on
existing real travel. Rapid technology development
and cost reductions— or their delay— may render
our design games under-, or over- optimistic. And
as discussed in the description of our recharging
scenarios, none of them likely capture precisely
what will happen with workplace recharging,
efforts to control time of day of recharging, or
efforts to provide home recharging to the over one-half
of new car- buying households in California
who do not now find access to electricity where
they park their cars.
The present analysis is useful in providing a
plausible baseline for the early PHEV market; but
a baseline from which consumers, infrastructure
and vehicle providers, and policy makers can
create change. Research suggests that with the
right incentives, consumers might locate more
recharge locations, modify existing recharge
locations, e. g. clean up the home garage, and
adjust driving patterns and adapt vehicle use
among the household fleet to maximize electricity
use [ 14, 20]. Still much adaptation by consumers
may not occur until after they purchase a PHEV,
and their perceived recharge potential that may
lead to PHEV purchase may be based on existing
driving patterns, i. e., current perceptions of
recharge locations.
Still, it may be possible to lead PHEV purchases
by changing perceptions of the availability of
vehicle recharging, by actually increasing the
availability of recharging for those households
who do not now find it, and by improving the
visibility and viability of existing electrical
infrastructure for vehicle recharging. Recharge
infrastructure could expand to a higher percentage
of households with changes in building codes, as
well as increased employer and publicly installed
vehicle recharge outlets.
analogies to HEV markets to judge how quickly
makes and models of PHEVs will be introduced.
PHEVs are, as a set of design possibilities, more
plastic than HEVs, and certainly one of the viable
interpretations of the launch of HEVs is that a dearth
of makes and models slowed the market entry of
hybrid technology.
MW per Million PHEVs
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 10
Among the respondents with at- home vehicle
recharging, most constructed more expensive
vehicle designs that added plug- in capability to
their next vehicle purchase than did those
without access to recharging. Given access to
recharging and the distribution of PHEV designs
from the games, we estimate that about one- third
of new vehicle buying households in California
have both the required infrastructure and interest
to purchase a vehicle with plug- in capabilities.
The variety of PHEV designs created by
respondents suggests there is still ample
opportunity for automakers to explore and
develop different PHEV designs.
We observed a wide diversity of consumer
interests in PHEV design options. Starting with a
base PHEV design offering long recharge times,
short CD range, blended rather than all- electric
operation, but non- trivial increases in both CD
and CS gasoline fuel economy, the most popular
upgrade category was to further improve CS fuel
economy. Respondents also exhibited interest in
increasing vehicle range in CD mode, and
improving CD fuel economy. Fewer respondents
were willing to devote resources to reduce
recharge time; most plausible early market
respondents have access to periods of home-based
charging long enough to fully recharge
each day even at the slowest offered rate. Given
their present vehicle purchase and travel
behaviour, and their present understandings of
PHEVs ( as enhanced by the tutorial in their
questionnaire), almost no new car- buying
households in California design a PHEV with all-electric
CD operation.
The final analysis in this report combined all the
available information from each respondent—
driving, recharge potential, and PHEV design
priorities— to estimate the energy impacts of the
respondents’ existing travel and understandings
of PHEVs under a variety of recharging
scenarios. Results suggest that the use of PHEVs
could halve gasoline use relative to conventional
vehicles— the majority of this reduction being
due to increases in CS fuel economy. Using three
scenarios to represent potential boundary
conditions on PHEV driver recharge patterns
( unconstrained, universal workplace recharging,
and off- peak only charging), we estimate
tradeoffs between the magnitude and timing of
PHEV electricity use. In the unconstrained Plug
and Play recharge scenario, recharging peaks at
6: 15pm, following a far more dispersed pattern
throughout the earlier part of the day than
anticipated by previous studies [ 1,7]. The more
dispersed time- of- day recharging pattern in our
work is due to our ability to realistically account
for heterogeneity in driving and parking behaviour
and to allow for heterogeneity of PHEV designs.
PHEV electricity use could be increased through
policies increasing non- home recharge
opportunities ( e. g., the Enhanced Workplace
Access scenario), but most of this increase occurs
during daytime hours and could contribute to peak
demand ( depending on a given region’s definition
of “ peak”). We also demonstrate how deferring all
recharging to off- peak hours ( 8pm- 6am) could
eliminate all additions to daytime electricity
demand from PHEVs, similar to what Lemoine et
al. [ 1] call “ optimal charging.” However, as also
found by Kurani et al. [ 30] for EVs, in this
scenario less electricity is used due to the
elimination of daytime recharge opportunities and
thus less gasoline is displaced.
This analysis provides one measure of potential
threat and opportunity for electric utilities. The
threat is that without control, the majority of
recharging may occur during peak hours ( 6am-
8pm), with a peak at 6: 15pm during weekdays.
This spike coincides with seasonal peak electricity
demand periods in some California regions and
with a large enough PHEV market, overall
electricity generation requirements may be
increased [ 1]. However, the observed 12am- 6am
recharge potential in late evening and early
morning presents an opportunity for “ smart
charging” strategies in which PHEV recharging ( as
well as any other electrical load) can be shifted to
off- peak periods subject to varying levels of
control by electricity users and suppliers.
Our scenarios are limited in that we do not
represent recharge scenarios specific to the various
regions and electric utilities across California.
Instead we produce an aggregated state- wide
pattern without explicitly representing current
electricity demand patterns, i. e., without PHEVs.
Our intention is to represent energy use according
to general trends rather than to provide a specific
energy analysis for a given region.
Acknowledgments
The authors would like to thank the California
Energy Commission for funding this project and
the Social Sciences and Humanities Research
Council of Canada.
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 11
References
[ 1] D. Lemoine, et al., An innovation and policy
agenda for commercially competitive plug-in
hybrid electric vehicles, Environmental
Research Letters, 3( 2008), 1- 10
[ 2] J. Axsen et al., Batteries for plug- in hybrid
electric vehicles ( PHEVs): Goals and the
state of technology circa 2008. Institute of
Transportation Studies, University of
California, Davis, Research Report UCD-ITS-
RR- 08- 14, 2008
[ 3] R. Winkel et al., Global prospects of plug-in
hybrids, 22 nd International Battery,
Hybrid, and Fuel Cell Electric Vehicle
Symposium and Exhibition ( EVS- 22),
Yokohama, Japan, October 23- 28, 2006
[ 4] A. Vyas et al., Plug- in hybrid electric
vehicles: How does one determine their
potential for reducing U. S. oil dependence?,
23 rd International Electric Vehicle
Symposium and Exposition ( EVS- 23),
Anaheim, California, December 2- 5, 2007
[ 5] M. Duvall et al., Environmental Assessment
of Plug- In Hybrid Electric Vehicles, Volume
1: Nationwide Greenhouse Gas Emissions,
EPRI, Palo Alto, CA: 2002, Report
# 1015325, 2007
[ 6] J. Gondor et al. Using global positioning
system travel data to assess the real world
energy use of plug- in hybrid electric
vehicles, Transportation Research Record,
2017( 2007), 26- 32
[ 7] S. Hadley and A. Tsvetkova, Potential
Impacts of Plug- in Hybrid Electric Vehicles
on Regional Power Generation, Prepared
for the U. S. Department of Energy, Oak
Ridge National Laboratory, January, 2008
[ 8] C. Samaras and K. Meisterling, Life cycle
assessment of greenhouse gas emissions
from plug- in hybrid vehicles: Implications
for policy, Environmental Science and
Technology, 42( 9), 2008, 3170- 3176
[ 9] R. Sioshansi and P. Denholm, Emissions
impacts and benefits of plug- in hybrid
electric vehicles and vehicle- to- grid
services, Environmental Science and
Technology, 43( 4), 2009, 1199- 1204
[ 10] J. Axsen and K. Kurani, The early U. S.
market for PHEVs: Anticipating consumer
awareness, recharge potential, design
priorities and energy impacts. Institute of
Transportation Studies, University of
California, Davis, Research Report UCD-ITS-
RR- 08- 22, 2008
[ 11] K. Nesbitt et al., Home recharging and the
household electric vehicle market: A
constraints analysis, Transportation Research
Record, 1366 ( 1992), 11– 19
[ 12] B. Williams and K. Kurani, Estimating the
early household market for light- duty
hydrogen- fuel- cell vehicles and other
“ Mobile Energy” innovations in California:
A constraints analysis, Journal of Power
Sources, 160 ( 2006), 446- 453
[ 13] K. Kurani et al., Demand for electric vehicles
in hybrid households: an exploratory
analysis, Transport Policy, 1( 4), 1994, 244-
256
[ 14] K. Kurani et al., Testing electric vehicle
demand in ‘ hybrid households’ using a
reflexive survey, Transportation Research Part
D 1( 2), 1996, 131- 150
[ 15] R. Graham et al., Comparing the Benefits and
Impacts of Hybrid Electric Vehicle Options,
EPRI, Palo Alto, CA, Report # 1000349, 2001
[ 16] D. Bunch et al., Demand for clean- fuel
vehicles in California: A discrete choice
stated preference survey. Transportation
Research A, 27( 1993), 237- 253
[ 17] G. Ewing and E. Sarigollu, Assessing
consumer preferences for clean- fuel vehicles:
A discrete choice experiment. Journal of
Public Policy and Marketing, 19( 2000), 106-
118
[ 18] D. Potoglou and P. Kanaroglou, Household
demand and willingness to pay for clean
vehicles, Transportation Research Part D,
12( 2007), 264- 274
[ 19] R. Heffner et al., Symbolism in California’s
early market for hybrid electric vehicles.
Transportation Research Part D, 12( 2007),
396- 413
[ 20] K. Kurani et al., Driving plug- in hybrid
electric vehicles: Reports from U. S. drivers of
HEVs converted to PHEVs, circa 2006- 07,
23 rd International Electric Vehicle
Symposium and Exposition ( EVS- 23),
Anaheim, California, December 2- 5, 2007
[ 21] F. Kalhammer et al.,. Status and Prospects for
Zero Emissions Vehicle Technology: Report
of the ARB Independent Expert Panel 2007,
Prepared for State of California Air Resources
Board, Sacramento, California, April 13,
2007
[ 22] T. Markel et al., Plug- in hybrid vehicle
analysis, Milestone Report, National
Renewable Energy Laboratory, NREL- MP-
540- 40609, November, 2006,
EVS24 International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium 12
http:// www. nrel. gov/, accessed on 2008- 09-
22.
[ 23] M. Kromer and J. Heywood, Electric
Powertrains: Opportunities and Challenges
in the U. S. Light- Duty Vehicle Fleet. LFEE
2007- 03 RP. Sloan Automotive Laboratory,
Massachusetts Institute of Technology,
Cambridge, MA, 2007
[ 24] S. Rhodes et al., Collecting behavioural
data using the world wide web:
Considerations for researchers, Journal of
Epidemiology and Community Health,
57( 2003), 68- 73
[ 25] T. Adler et al., Use of internet- based
household travel diary survey instrument,
Transportation Data and Information
Technology Research, 1804( 2002), 134- 143
[ 26] M. Couper, Web surveys: A review of issues
and approaches, Public Opinion Quarterly,
64( 4), 2000, 464- 494
[ 27] Internet World Stats. United States of
America: Internet Usage and Broadband
Usage Report, July 2007,
http:// www. internetworldstats. com/ am/ us. ht
m, accessed on 2008- 01- 27
[ 28] M. Couper et al., Noncoverage and
nonresponse in an internet survey, Social
Science Research, 36( 2007), 131- 148
[ 28] M. Duvall et al., Comparing the Benefits
and Impacts of Hybrid Electric Vehicle
Options for a Compact Sedan and Sport
Utility Vehicles, EPRI, Palo Alto, CA,
Report # 1006892, 2002
[ 29] A. Burke, and E. Van Gelder, Plug- in
hybrid- electric vehicle powertrain design
and control strategy options and simulation
results with lithium- ion batteries,
Proceeding ( CD) from the EET- 2008
European Ele- Drive Conference,
International Advanced Mobility Forum,
Geneva, Switzerland, March, 2008
[ 30] K. Kurani et al. Where, when, how fast and
how much? Questions about consumer
demand for home, away from home, time of
day, and speed of recharging for electric
vehicles, 14 th International Electric Vehicle
Symposium and Exposition ( EVS- 14), Palm
Beach, Florida, October, 1997
Authors
Jonn Axsen is a PhD candidate in the
Transportation Technology and Policy
( TTP) program at the University of
California, Davis. He holds degrees in
resource management and business
administration, and is currently
conducting market research for
electric- drive vehicles.
Dr. Ken Kurani is a research engineer
at the Institute of Transportation
Studies at UC Davis. He’s earned an
international reputation for innovative
approaches to analyzing demand for
transportation and communication
technologies. His contributions are
methodological, theoretical, and
empirical, intersecting with several
disciplines, including economics,
sociology, and engineering.