Advanced Plug-in Electric Vehicle Travel and Charging Behavior

Michael Nicholas, Ph.D. Gil Tal, Ph.D.

Thomas Turrentine, Ph.D. (ARB Contract 12-319 – Funding from CARB and CEC)

Advanced Clean Cars Symposium September 27, 2016

1

Overview

• Project Background • Baseline statistics on participating households • Household travel changes over time • Engine starts of plug-in hybrids • Overall motivation to plug in

– Charging behavior of participating households

2

Project Background • Project consists of 3 components:

1. Intro survey to recruit households statewide • Over 6000 completed surveys from throughout California

2. High resolution on-board data collection of PEV and ICE vehicles in selected households for 12 months • PEV models: Leaf, Volt, Prius Plug-in, C-Max, i3 REx, Tesla • PEV parameters: battery SOC, speed, RPM, GPS, charging level and

kWh, etc. • ICE parameters: speed, RPM, GPS, fuel economy variables, refueling • 264 Households: 72 Completed, 60 in progress, 132 forthcoming

beginning Fall 2016 3. Exit survey of selected households

• Project rolled out in different phases – First wave began summer 2015

3

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

0 20 40 60 80 100 120 140

Annu

al e

VMT

EPA Electric Range

Observed BEV and PHEV eVMT (OEM Data and UCD Survey Data)

INL eVMT UCD EVMT BEV eVMT j2841 Utility Factor (12,000 miles)

eVMT Overview: UCD Data Consistent with OEM Data

Source: Nicholas, Michael and Tal Gil. 2016, January 10-14. EVMT in the Household Fleet: Integrating Battery Electric Vehicles into Household Travel. In Transportation Research Board. Washington DC.

4

INL Volt eVMT Higher Than UCD

Our Leaf Households Have More Cars and Drivers than Prius Households

5

0

0.5

1

1.5

2

2.5

3

Mean Number of Vehicles Mean Number of Drivers

Prius Energi Volt Leaf

Mean Number of Vehicles Mean Number of Drivers

PEV Households Similar in Travel Needs (per Driver) But Shift Miles to Higher MPG

Vehicles

6

10425

12268

2829

9610

11778

4982

9697

11122

8186

1060910230 10230

0

2000

4000

6000

8000

10000

12000

14000

Mean Annual HH VMT PerDriver

Mean Annual PEV VMT Mean Annual eVMT

Prius Energi Volt Leaf

Mean Annual HH VMT Per Driver

Mean Annual PEV VMT

Mean Annual eVMT

14000

12000

10000

8000

6000

4000

2000

0

eVMT Percentages Three Ways: By PEV, Household, Avg Driver VMT

7

23%

15%

27%

42%

25%

52%

74%

45%

84%

100%

43%

96%

0%

20%

40%

60%

80%

100%

120%

Mean Percent eVMT (PEV) Mean Percent eVMT (HH) Mean eVMT as Share ofAverage HH Driver

Prius Energi Volt Leaf

Mean Percent eVMT (PEV)

Mean Percent eVMT (HH)

Mean eVMT as Share of Avg HH Driver

100%

80%

60%

40%

20%

0%

Accounting for All HH Vehicle GHG Emissions per Mile, Prius HH Most Efficient

8

HH Miles Relatively Constant

9

VMT Down for Fusion and Prius

10

“Other Vehicle” ICE Miles Increasing Slightly for PHEVs

11

Seasonal Effects? % eVMT Down in Winter

12

Blended PHEVs Long Soak IC Starts Often High Power

Soak Time (minutes)

eTrips

13

Preliminary

Overall Motivation to Plug In

• You must plug in to get eVMT – Is it worth it to the driver? What is the cost/benefit?

• Analysis of survey respondents shows PHEVs more likely to be plugged in when more range can be recovered – Longer range PHEVs less likely to never be plugged in

because they provide greater potential for more miles to be recovered per charging event

– Every mile recovered from a charge event increases the likelihood of plugging in by 1.4%

14

PHEVs With Longer Range Are Unlikely to Never be Plugged in. Cost-Benefit is Always Higher. • 481 PHEV owners (Plug-in Prius, Energi, Volt) have free

charging at work. Do they plug in? • If one-way distance is 10 miles: All act statistically

similar • Likelihood of plugging in is a function of range

recovered • Longer range PHEVs eventually plug in

PHEV 10

PHEV 20

PHEV 40

10 miles 20 miles 40 miles

PHEV 10

PHEV 20

PHEV 40

PHEV 10

PHEV 20

PHEV 40

Source: Nicholas, Michael and Tal Gil. 2017 (Forthcoming), January 8-12. You Can’t Take It With You: Examining The Role Of PHEV Range In The Decision To Plug In. In Transportation Research Board. Washington DC.

$

$

$

$

$$

$$

$

$$

$$$

15

Because Low-Range PHEVs Provide Little Cost-Benefit, They are More Likely to Never Be Plugged in

Source: Nicholas, Michael and Tal Gil. 2017 (Forthcoming), January 8-12. You Can’t Take It With You: Examining The Role Of Phev Range In The Decision To Plug In. In Transportation Research Board. Washington DC.

0%

5%

10%

15%

20%

25%

30%

35%

0 5 10 15 20 25 30 35 40

Perc

ent N

ot P

lugg

ing

In

PHEV Vehicle Range

Not Plugging in Regularly (4 or Fewer Events)

2015 Survey 2016 Survey 2016 Used Survey

Power (2015 Survey) Power (2016 Survey) Power (2016 Used Survey)

16

PHEV Vehicle Range

Perc

ent N

ot P

lugg

ing

In

2015 - New

2016 - New

2016 - Used

Even BEVs Value Not Plugging in at Home (Survey Preferences)

17

Many Will Plug in at Home if Battery SOC is > 18% and <80%.

18

Longer Range BEVs Wait 50 Miles Before Plugging in

19

Volt Drivers Plug in the Most. Not Normalized for Charger Access

20

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

0 10 20 30 40 50 60 70 80 90

Numb

er of

Even

ts

Vehicle Range

Plug In Events

Vehicle Range

Num

ber o

f Eve

nts

Home

Away

Home + Away Plug In Events

Conclusions

• Many…but for this project: • The lower the electric range the fewer eVMT

(duh!) – Technical potential is lower – Customers are not as willing to plug in

• Volts are similar in behavior and potential to Leafs for our sample, but slightly lower eVMT

21

More Conclusions • Household data show declining use of PEV

– Older vehicles are naturally used less – Gas prices decrease

• High power cold starts more likely on smaller battery PHEVs

• GHG per household mile shows inefficient household miles (especially in BEV HH) need to be replaced with efficient miles. – Vehicle replacement is key – Substitute gasoline miles in an efficient PHEV – Zero carbon electricity is necessary

22

Thank You

• Michael Nicholas mianicholas@ucdavis.edu • Gil Tal gtal@ucdavis.edu • Thomas Turrentine tturrentine@ucdavis.edu

23

Participant Locations

24

Linear Regression Model TABLE 3 Linear Regression Model

Dependent variable: percent of commute days plugging in Term Estimate Std Error t Ratio Prob>|t| Intercept 0.4105088 0.052351 7.84 <.0001* Weekly congestion frequency (days) -0.047059 0.010829 -4.35 <.0001* Time restrictions dummy -0.0801766 0.019353 -4.14 <.0001* Income fraction from 500000 -0.172085 0.082624 -2.08 0.0378* Recoverable One Way Miles 0.0144503 0.002104 6.87 <.0001* R2 = 0.165702 R2 (adj) = 0.158691

25