This document summarizes research on connecting electric vehicles to electricity networks. It discusses battery charging behavior at home and during trips. A model found drivers try to avoid fully depleting batteries, with average inefficiencies of 1.5-1.8 kWh. Scenarios for Nagoya, Japan showed charging electric vehicles could increase peak demand but discharging vehicles at work could cut 0.1 GW of daytime demand. The research evaluates optimizing battery size, fast charger deployment, and using vehicle-to-grid to reduce peak electricity loads.

1. ENDORSING PARTNERS
Transport modelling and
simulation for next
generation infrastructure
development www.isngi.org
The following are confirmed contributors to the business and policy dialogue in Sydney:
•
Rick Sawers (National Australia Bank)
•
Nick Greiner (Chairman (Infrastructure NSW)
Monday, 30th September 2013: Business & policy Dialogue
Tuesday 1 October to Thursday, 3rd October: Academic and Policy
Dialogue
Presented by: Professor Toshiyuki Yamamoto, EcoTopia Science Institute,
Nagoya University
www.isngi.org

2. Transport modeling and simulation
for next generation infrastructure
development:
Connecting vehicle to electricity network
Toshiyuki Yamamoto
Nagoya University
2

3. Outline
• Background
– Next generation infrastructure and car in Japan
• Battery charging behavior
– At home
– Within trip
• Vehicle to grid
– Impact on electricity demand curve
• Conclusions
3

4. Next generation infrastructure
• Council for Science and Technology Policy, Japan
states the need for next generation infrastructure
• Features of next generation infrastructure
– Smart: information technology to forecast, control
and optimize infrastructure system
– System: value added as system in addition to strength
of products and technology itself
– Global: business strategy toward global deployment
4

5. Areas of next generation infrastructure
• Smart energy community
– Energy management system utilizing information technology
– Renewable energy, decentralized generating plant, etc.
• Intelligent transport system
– Communication networking among people, vehicles and road
utilizing information technology
– Navigation system, car sharing, LRT, etc.
• Next generation infrastructure in other areas
– Water supply, goods distribution, medical care, etc.
– Integrated system
5

6. Passenger car ownership in Japan
Light motor passenger car (~ 0.66L)
5
Small passenger car (~2L)
4
Ordinary passenger car (over 2L)
3
2
1
2010
2005
2000
1995
1990
1985
1980
1975
1970
0
1965
10 million vehicles
6
Year
Source: MLIT
6

7. Passenger car sales ranking in Japan in
2012
Rank
1
2
3
4
5
6
7
8
9
10
Model (Automaker)
Prius (Toyota)
Aqua (Toyota)
Mira (Daihatsu)
N BOX (Honda)
Fit (Honda)
Wagon R (Suzuki)
Tanto (Daihatsu)
Move (Daihatsu)
Alto (Suzuki)
Freed (Honda)
HV: hybrid vehicle
Sales
317,675
266,567
218,295
211,156
209,276
195,701
170,609
146,016
112,002
106,316
Engine type
HV
HV
Light motor
Light motor
Small / HV
Light motor
Light motor
Light motor
Light motor
Small / HV
Source: Nikkei Newspaper
7

8. Electric vehicles and Plug-in hybrid
vehicle in Japan
i-MiEV
2009
Leaf
2010
Prius plug-in hybrid
2012
More energy efficient, but
more electricity dependent
8

9. Battery charging at home
• Analysis on charge timing choice behavior of
plug-in hybrid vehicles in Toyota City, Japan
– This is a part of the results obtained by joint research
with Toyota Motor Corporation
9

10. Smart Melit（Smart Mobility & Energy
Life in Toyota City）project
–
–
–
–
Toyota City, Japan
67 new houses, some with plug-in hybrid Prius
HEMS （Home Energy Management System）
DRP (demand response point) system
10

11. Smart house
Visualization by HEMS
(home energy
management system)
PHV charger
PHV
DRP (demand
response point) portal
PHV charging monitor
11

12. Example of electricity demand curve
PHV charge
Summer
Air cond.
PHV charge
Winter
Air cond.
Heat pump
water heater Scheduled to fill-up at 4:00
12

13. DRP (demand response point)
• Peak pricing by point system
• Low at daytime (solar energy) &
high at evening (more activity at home)
Feb.
Example of DRP
May
Aug.
Nov.
13

14. Distribution of returning home timing
No charge
Charge
Time of day
Many cars return home at 18 to 20 o’clock,
which potentially cause peak demand
14

15. Charging time is shifted by
demand response point system
Other
Cheapest
period
Just after
came home
With demand
response point
W/O demand
response point
15

16. Charge timing choice model
• Multinomial logit model
No charge
Just after
came home
Cheapest
timing
Cheapest timing
before the next
vehicle use
– 12 Prius plug-in hybrid vehicles
– 2011/10/1 to 2012/10/31
– 4615 cases
Other
Didn’t change the
setting of ontimer previously
set, or by mistake
16

17. Charge timing choice model
Alternative
Variable
Constant
No charge Drive distance (<24 km)
Long distance dummy (>24 km)
Price for energy conscious person
Just after
Price for energy unconscious person
came home
Return home at daytime (9-16)
Constant
Price for energy conscious person
Cheapest
Price for energy unconscious person
time
Housewife dummy
Return home at evening (17-23)
Constant
Other
Return home at evening (17-23)
Same as the last charge dummy
Log-likelihood (0)
Log-likelihood at convergence
Adjusted rho-square
Coef.
1.34 **
-0.10 **
-0.38 **
-0.044 **
-0.065 **
0.70 **
-0.69 **
-0.016 **
0.001
0.66 **
1.41 **
-0.96 **
0.65 **
2.21 **
-5774
-4415
0.233
17
** 1%, * 5%

18. Sensitivity of the estimated model
Base case: High energy conscious male driver
returned home in evening after 10 km drive
Electricity price
No DRP (20.9 JPY)
Evening price
20.9 -> 28 JPY
+ Midnight price
20.9 -> 10 JPY
No charge
35%
36%
34%
Just after
came
home
Cheapest
timing
Other
47%
8%
38%
18%
19%
7%
42%
18%
Charge timing is easier to change than the timing
of air conditioner usage, etc.
18

19. Battery charging within trip
• The timing of mid-trip electric vehicle charging
– This is a part of the results obtained by the Project Consigning
Technology Development for Rational Use of Energy (Promotion
of aggregation and sharing of probe information)
– The dataset was provided by Japan Automobile Research
Institute (JARI)
19

20. Locations
2000
1800
1600
1400
1200
1000
800
600
400
200
0
2013.06
2013.07
2013.04
2012.06
2012.07
2012.08
2012.09
2012.10
2012.11
2012.12
2013.01
2013.02
2011.12
2012.01
2012.02
2012.03
2012.04
2011.09
2011.10
2011.03
2011.04
2010.12
2010.10
2010.04
2009.11
Fast charger deployment in Japan
Year.month
Source: CHAdeMO Association
20

21. Trade-off between battery size and fast
charger density
How to optimize battery size & fast charger
deployment?
• Drivers charge battery before empty
• Charging behavior should be understood
21

22. Data
•
•
•
•
Investigator: Japan Automobile Research Institute
Sample: 252 company cars & 247 private cars
Survey period: 2 years (2011.2-2013.1)
Survey area: 42 out of 47 prefectures in Japan
• Built-in data logger with GPS & communication
unit: clock time, location, vehicle state (driving,
normal charging, fast charging), odometer
reading, use of air-conditioner & heater, state of
charge
22

23. Distribution of SOC at normal charge
20%
Company cars
個人車両(N=492)
Private cars
法人平均(N=252)
18%
16%
14%
12%
10%
8%
6%
4%
2%
0%
~10%
~20%
~30%
~40%
~50%
~60%
~70%
~80%
~90%
90%~
SOC: state of charge
Company cars are charged at the end of the working hours
23
regardless of SOC

24. Distribution of SOC at fast charge
25%
20%
法人平均(N=252)
Company cars
個人車両(N=492)
Private cars
15%
10%
5%
0%
~10% ~20% ~30% ~40% ~50% ~60% ~70% ~80% ~90% 90%~
SOC: state of charge
Battery capacity is not fully utilized
24

25. Stochastic frontier model of SOC at
fast charging within trip
• Driver avoids running out of power
Actual remaining electricity
to start charging
=
≥
Subjective minimum electricity
+
• Inefficiency is added to minimum electricity
Actual remaining
electricity
Subjective minimum
electricity
Inefficiency
• Stochastic cost frontier model is applied
25

26. Distribution of subjective minimum
and actual remaining charge
personal-use vehicles on working day
9
actual
Percent of samples (%)
8
minimum
7
6
5
4
3
2
1
0.2
0.8
1.4
2.0
2.6
3.2
3.8
4.4
5.0
5.6
6.2
6.8
7.4
8.0
8.6
9.2
9.8
10.4
11.0
11.6
12.2
12.8
13.4
14.0
14.6
15.2
15.8
0
Remaining electricity (kWh)
• Subjective minimum remaining charge has peak at 3.6kWh
• 1.5kWh of average inefficiency is estimated
26

27. Distribution of subjective minimum
and actual remaining charge
commercial-use vehicles on working day
Percent of samples (%)
8
actual
7
minimum
6
5
4
3
2
1
0.2
0.8
1.4
2.0
2.6
3.2
3.8
4.4
5.0
5.6
6.2
6.8
7.4
8.0
8.6
9.2
9.8
10.4
11.0
11.6
12.2
12.8
13.4
14.0
14.6
15.2
15.8
0
Remaining electricity (kWh)
• Same peak of minimum remaining charge
• Larger (1.8kWh) average inefficiency is estimated
27

28. Vehicle to grid
• Impact of electric vehicles on electricity
demand curve in Nagoya, Japan
– This is a part of the research results funded under the
Environmental Research and Technology Development
Fund by Ministry of the Environment
28

29. Time axis
Micro simulation model of individual’s
activity-travel pattern
【Staying】
<Activity Choice>
Staying
Commuting Commuting
to school
to work
Private
Business
Returning
home
【Returning home】
Railway
9 p.m.
【Private】
Walking
【Staying】
<Destination Choice>
Destination Destination
1
2
・・・・・
Destination
5 p.m.
【Staying】
s
2 p.m.
<Mode Choice>
Car
Railway
Route
Route
Bus
【Staying】
Bicycle &
Walking
Another
Office
<Route Choice>
1
2
・・・・・
Route
k
【Business】
Walking
8 a.m.
【Staying】
【to Work】
Restaurant
Railway
Work place
Home
• Nested logit model of activity type, destination,
mode and rail route choice at each time period
• Sequence of activity-travel pattern is simulated
29

30. Interaction between activity-based
model & dynamic traffic assignment
Time axis
Equilibrium state is calculated
【Staying】
OD demand
【Returning home】
Railway
9 p.m.
【Private】
Walking
【Staying】
5 p.m.
【Staying】
2 p.m.
【Staying】
【Business】
Walking
Another
Office
【to Work】
Restaurant
Railway
8 a.m.
【Staying】
Work place
Home
Travel pattern
OD travel time
Link travel speed
• Parking location, parking duration, and SOC of
each EV are simulated along time of day
30

31. Scenario analysis at Nagoya, Japan
 Nagoya Metropolitan Area
<Population in 2020>
Over 8.0 million
<Zone>
520 zones
(Nagoya City is divided
into 259
Average area is 1.3km2)
Osaka
Tokyo
Nagoya
<Road Network>
Link:22,466
Nagoya City
Node:7,600
10% of vehicles are assumed to be replaced by EV,
which means 472,000 EVs
31

32. Distribution of electricity demand
Electricity demand curve
2.5
Spatial distribution of electricity demand
GWh/h
Total
2.0
Home
Office
Store
1.5
Other
1.0
0.5
0.0
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Time of day
• Total demand is about 31GWh
• Higher demand at midday and at CBD
32

33. Scenarios for EV charging/discharging
• Case 0: No EVs
• Case 1: Charge at home immediately after
returning home
• Case 2: Charge at home during midnight
• Case 3: Charge at workplace immediately after
arriving at work
• Case 4: Charge at home during midnight and
discharge at workplace during daytime
(until the remaining charge at 5 kWh)
33

34. Impact on electricity demand curve
2.5
GWh/h
Case_0
No EVs
Case_1
Charge
at home
Case_2 at home midnight
Charge
Case_3 at workplace
Charge
Charge
Case_4 & discharge
2.0
1.5
1.0
0.5
0.0
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Time of day
• 0.1GW of daytime demand (9:00 to 16:00) can
be cut by vehicle to grid at workplace
34

35. Spatial distribution of impact
CBD area
Suburban area
500
-200
400
-400
300
-600
200
-800
100
0
-1,000
1,000
600
800
500
600
400
400
300
200
200
0
100
0
Charge/Discharge (kWh/h)
0
No. of the EV Parked (vehicles)
600
700
200
Charge/Discharge (kWh/h)
No. of the EV Parked (vehicles)
700
-200
3 4 5 6 7 8 9 1011121314151617181920212223242526 tiem period
3 4 5 6 7 8 9 1011121314151617181920212223242526 time period
Home
Home
Workplace
Other
Charge/Discharge
Workplace
Other
Charge/Discharge
• Impact is quite different between CBD and
suburban area
35

36. Conclusions
• Battery charging at home causes significant
electricity demand, but the timing can be
controlled by peak pricing
• Battery capacity is not fully utilized, and
measures to improve efficiency are needed
• Potential to cut down peak demand by vehicle
to grid at workplace
36